Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
The interplay between social connection and substance use
(USC Thesis Other)
The interplay between social connection and substance use
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
THE INTERPLAY BETWEEN SOCIAL CONNECTION AND SUBSTANCE USE
by
Nina Caitlin Christie
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PSYCHOLOGY)
August 2023
Copyright 2023 Nina Caitlin Christie
ii
Epigraph
Love and compassion are necessities, not luxuries. Without them humanity cannot survive –
Dalai Lama
iii
Dedication
For Kanon – you are loved beyond measure.
iv
Acknowledgements
I would first like to thank my primary mentor, John Monterosso, who guided me through
graduate school. John provided relentless support for each endeavor I took on as a researcher,
advocate, teacher, leader, and student. John – you have truly shaped me as a researcher, and
more importantly, as a person. Your mentorship has provided me with a roadmap of how to be a
thoughtful, rigorous, and kind leader. I am truly grateful to have had the opportunity to learn
from you over the last six years, and am looking forward to working with you for the years to
come.
I would like to thank Jordan Davis, who took me on as a student in his lab during his first
year at USC. Jordan, working with you showed me the immense power in collaborative
relationships, perfecting statistical modeling, and enjoying life as an academic. Being in your lab
inspired me to think bigger than the individual – to really think about the impacts of systems and
environments, and ultimately led me to complete a dual-degree MPH during my time at USC.
I would like to thank Antoine Bechara, who provided me with support, encouragement,
and the opportunity to grow as a researcher and scholar. Antoine, working with you has given me
the perspective of a leader, and has shown me how to look at the bigger picture in all of this.
I would like to thank my committee members, Gale Lucas and Mark Lai. You have
provided me with a different point of view through which to view my research, both statistically
and conceptually. Your perspectives gave me valuable insights that I will take with me as I move
forward in my career.
I would like to thank Eric Pedersen who mentored me first as a Research Assistant and
then as a collaborator. Your mentorship inspires me to lead with kindness, dignity, and respect. I
am so fortunate to have had the opportunity to work with you these last few years.
v
Thank you to my Research Assistants, particularly Vanya Vojvodic and Myra Moghal.
Vanya – thank you for taking on the role as lead research assistant for our project in Skid Row.
You are a stellar researcher; your leadership, organization, and gumption truly was the glue that
held that project together. I am confident that your journey through medical school will lead you
to an amazing career. Your sense of purpose is inspiring, and the medical field is better off for
having you there. Thank you for being a friend. Myra – you have done incredible work on the
study presented here in Chapter 3, as well as on the neuroimaging project. You have been my
right-hand for over a year and I am truly so excited for you to begin medical school, although we
are sad to see you move on from the lab. It has been a true pleasure to work with you, and I
cannot wait to see all that you go on to do in the future.
I have also had the tremendous fortune of working with Leslie Berntsen as my teaching
guide and leadership inspiration throughout graduate school. Leslie, from the first day we met,
you inspired me to be a better instructor and leader. You lead by example, and you have shown
me the intangible value in supporting others, lifting up students and mentees, and being a force
for positive change. You inspired me to take on the role as Co-President of GASP, and you
supported me in developing and teaching my own course just one year into graduate school. I am
lucky to call you a mentor and friend.
In completing this work, I have relied heavily on institutional support. I want to thank
three departmental leaders in particular: Twyla Ponton, Jennifer Vo, and Jo Ann Farver. Twyla,
thank you for sharing your organizational skills, communication, and kind presence with me, the
graduate student body, and the department; I have appreciated knowing I could always rely on
you. Jennifer, thank you for your relentless efforts to guide me through the psychology
department here at USC, including during my pursuit of the dual-degree. Your support has been
vi
instrumental in my career. Jo Ann, thank you for advocating for the graduate student body here
at USC. Most all of the studies in my dissertation were completed with departmental funds
through a mechanism that you implemented; I recognize and appreciate your efforts.
Lastly, I want to acknowledge my tribe: the family and friends who have given me
endless love and support throughout my life. Mom and Dad, thank you for encouraging me to
achieve my goals, and reminding me that life outside of the ivory tower continues day after day.
I strive for my work to have impact that reaches far outside of these university walls, because of
you. To my siblings – Cole, Kanon, and Serena – you mean the entire world to me. I cannot say
where I would be now if it were not for you leading the way; you were the first to show me that
human connection is what makes a life worth living. To my partner, Will – you have brought
consistent love and joy into my life. Thank you for sharing your world with me (and for cooking
so many dinners during my late-night writing sessions). I am eternally grateful to have found you
and to be building a life with you. I want to thank my friends and labmates who have provided
me with moral support, technical support, and perspective – it truly takes a village. Thank you
Sharon Saltoon, Claudia Okuniewski, Yvette Cabrera, Alex Prince, Jackson Trager, Esthelle
Uwusi-Boisvert, Gülnaz Kiper, Tony Vaccaro, Kaitlyn Yakura, Milad Kassaie, Shaddy Saba,
Sheila Pakdaman, and Graham DiGuiseppi.
vii
Table Of Contents
Epigraph ………………………..…………………………………………………………………ii
Dedication………………………………………………………………………………...………iii
Acknowledgements ………………………………………………………………………………iv
List of Tables …………………………………..…………………………………………..……viii
List of Figures ………………………….…………………………………………………………ix
Abstract ………………...……………………………………………………………..…..………x
Introduction …………………………..…………………………...………………………………1
Chapter 1: A Sense of Belonging: Opioids Mimic the Emotional Experience of Social
Connection…………..…………………………………………….…………………22
Introduction.………………………….……………….………………..………23
Method .……………………………...…………………….…………..………26
Results………………………………………………………...…..……………29
Discussion………………………………………………………….……..……40
Conclusion ..………………………….………………………...………………43
Chapter 2: COVID-19 and Substance Use Behaviors……………………………...……………..44
Introduction ……...………………………………………………………………45
Method ….………………………………….……………………………………47
Results……………………………..……..………………………………………49
Discussion…………………………………..……………………………………58
Conclusion..…………………………...........……………………………………62
Chapter 3: The Perceived Role of Social Connection in Relapse Risk and Relapse
Prevention……………………..……………..……………..………………………..64
Introduction …………………………………...…………………………………65
Method ……….…………………………………………….……………………68
Results……………………………………………………………………………73
Discussion………………………………………………………………..………85
Conclusion ……..………………..……………………………………….………88
Conclusions …………………………………………………………………...…………………88
References ..…………………………………………………….………………….....………… 90
Appendices ……….…………………………………………………………………………….111
Appendix A: Chapter 1 Appendix………………………………………………111
Appendix B: Chapter 3 Appendix ………………………………………………114
viii
List Of Tables
Chapter 1
Table 1: Demographic characteristics of the sample……………..………………...……………30
Table 2: Exploratory factor analysis loadings and variance explained …………………………31
Table 3: Confirmatory factor analysis standardized loadings for each latent factor ……………32
Table 4: Confirmatory factor analysis covariances between latent factors………….………..…33
Table 5: Confirmatory factor analysis latent factor loadings for each item………………..……33
Chapter 2
Table 1. Sample demographic characteristics .……………..……………………………………51
Chapter 3
Table 1: Demographic characteristics of the sample, .……………..……………………………74
Table 2: Relapse risk score predicted from a three-way interaction between Protagonist Drug
. of Choice, Valence, and Sociality………………………………………..……………..77
Table 3: Perceived protective effect against relapse following positive events predicted by a
two-way interaction between Protagonist Drug of Choice, and Sociality with
Emotional Response as a covariate ...……………………………………………......…..79
Table 4: Perceived relapse risk following negative events predicted by a two-way
interaction between Protagonist Drug of Choice, and Sociality with Emotional
Response as a covariate…......…………………………..………………………………..81
Table 5: Proportion of people who perceive an increased risk of relapse following
positive nonsocial events predicted by protagonist drug of choice * group (Hx-Pu
vs. No Hx-PU) with emotional response included as a covariate ………...……………..83
Table 6: Proportion of people who perceive an increased risk of relapse following positive
social events predicted by protagonist drug of choice * group (Hx-Pu vs.
No Hx-PU) with emotional response included as a covariate. …………………….……...85
ix
List Of Figures
Introduction
Figure 1: A conceptual model of the cyclical nature of opioid use and social connection..………3
Figure 2: A linear model depicting the correlation between drug overdose mortality rates and
social capital among 35 counties in West Virginia …………………….…………………11
Chapter 1
Figure 1: Factor analysis plot with factor loadings from Wave 1……………………………….34
Figure 2: Violin plot of endorsement of Hug/Belong items by Drug of Choice ………………..36
Figure 3: Violin plot of endorsement of Secure/Loved items by Drug of Choice ……..………..37
Figure 4: Violin plot of endorsement of Content/Satisfied items by Drug of Choice…...………38
Figure 5: Violin plot of endorsement of Excited/Happy items by Drug of Choice …………..…39
Figure 6: Endorsement of Hug/Belong items by Drug of Choice and DAST Score.........………40
Chapter 2
Figure 1. Social Connectedness Scale scores across drug of choice ...............................………50
Figure 2. An ANOVA depicting changes in social connection before and during COVID……. 52
Figure 3: Difference in reported social isolation by drug of choice …………………………….58
Chapter 3
Figure 1: Factor analysis of EFA with Target Rotation.....................................................………76
Figure 2: Marginal means of perceived protective effect of positive life events………………...79
Figure 3: Marginal means of perceived relapse risk following negative life events...…………..80
Figure 4: Proportion of people who perceived an increased risk of relapse to at least one
positive nonsocial event (of the 4 nonsocial positive events depicted in vignettes) ….... 82
Figure 5: Proportion of people who perceived an increased risk of relapse to at least one positive
social event (of the 4 social positive events depicted in vignettes) ……………………………..84
x
Abstract
This three-chapter dissertation examines the association between social connection and
substance use. In the introduction, I use excerpts from my previously published work (Christie,
2021) to delineate the theoretical foundations and public health impetus for the present work.
Following this introduction to the topic, I present a series of studies that assess how substance
use influences feelings of social wellbeing, which then influences substance use behaviors, and
how this iterative process impacts the trajectory of recovery from problem substance use, with an
emphasis on opioid use.
Chapter 1. I evaluate the hypothesis that opioid use is associated with the subjective
experience of positive feelings that are typically elicited by social connection. I aim to answer
the question: Is the acute opioid high associated with feelings that normally accompany positive
social connection and intimacy - above that of other drugs?
Chapter 2. I share my published work that assesses how social distancing measures
aimed at reducing the spread of COVID-19 impacted substance use behaviors (Christie et al.,
2021). I aim to answer the question: Did the implementation of large-scale social distancing
measures impact drug use behaviors?
Chapter 3. I measure perceptions of relapse risk to emotionally significant life events
that vary by: valence (positive/negative event), sociality (social/nonsocial event), drug of choice,
and personal history of problem substance use (yes/no). Here, I aim to answer the questions: 1)
Do people perceive social events to be particularly impactful in terms of mitigating /
exacerbating relapse risk relative to nonsocial events?
Conclusions. Across these three chapters, I focus on the critical role of a commonly
overlooked factor in substance use: social connection. This work has implications for both
xi
clinicians and policy-makers to improve the prognosis of substance use disorders through the
inclusion of social connection at each stage from prevention to treatment.
1
Introduction
Overview
The impact of social isolation on human behavior and psychology has been gaining
attention from scholars and the public alike. Over the last few decades, the size of social
networks has decreased, and the number of close friends and family that people confide in is
shrinking; this is coupled with a growing proportion of older adults in the USA living alone
(McPherson et al., 2006; Portacolone, 2013). Concurrently, drug overdose has become the
number one cause of accidental death in the USA, with most drug-related deaths resulting from
opioid use (Schiller & Mechanic, 2017). With the onset of the COVID-19 pandemic, these
numbers reached previously unfathomable rates: from 2019 to 2020, there was a 31% increase in
fatal overdoses in the United States, with a 56% increase in fatal overdose from synthetic opioids
(Hedegaard et al., 2021). In 2021, the number continued to climb with over 106,000 fatal
overdoses (National Institute on Drug Abuse, 2023). There is some hope; preliminary 2022 data
suggest a modest drop in overdose death rates compared with 2021 (Overdose Deaths Declined
but Remained Near Record Levels During the First Nine Months of 2022 as States Cope with
Synthetic Opioids, 2023).
I posit that there is a connection between the rise of opioid use and increasing social
isolation; researchers in psychology and neuroscience have developed a convincing literature
pointing to the critical role of endogenous opioids in social attachment. There is substantial
evidence that the endogenous opioid system plays a central role in the formation and
maintenance of social bonds in humans and other primates(Machin & Dunbar, 2011; Panksepp et
al., 1980, 1997). Additionally, there is overlap in the brain regions implicated in opioid use
disorder, pain, and social-emotional functioning. Specifically, the anterior insula and the dorsal
2
anterior cingulate cortex are activated both during the experience of ‘physical’ pain, such as a
mild electric shock, and ‘social’ pain, such as social exclusion (Eisenberger, 2012; Inagaki &
Eisenberger, 2013). People who experience social pain often use the same language as they
would for physical body insults, and for good reason, a ‘broken heart’ and a broken arm are
represented through the same neural pathways in the brain (Eisenberger & Lieberman, 2004).
The human need for social connectedness is deeply rooted in our biology, and our
endogenous opioid system appears to contribute significantly to the regulation of that need.
Social isolation is linked to an increase in substance use, and I argue that increasing social
cohesion and the feeling of social belongingness among individuals with a substance use
disorder – especially an opioid use disorder – is a key component to addressing the overdose
crisis in the USA today. In Figure 1, I present a conceptual model for the cyclical, bidirectional
associations between social isolation/connection and opioid use. I begin the cycle with acute
opioid use. I hypothesize that this induces a temporary sense of social–emotional well-being,
which reduces the need for a person to seek external social connection, leading to an increased
perception and experience of social isolation, which once more leads a person to use opioids.
Over time, I posit that this iterative cycle becomes chronic substance use with its own set of
unique psychobiological ramifications.
3
Figure 1: A conceptual model of the cyclical nature of opioid use and social connection.
Part I: Humans Are Highly Social Creatures
Humans are social beings; as infants we cry to get attention, as toddlers we play with
others and form friendships, as adolescents we form tight-knit peer groups, and as adults we
form friendships, colleague relationships and family units. Sociality is arguably a central element
of our evolutionary niche.
Evolutionarily, individuals who were able to form trust and long-lasting bonds with a
group were more likely to reproduce and survive. Social isolates would be selected out – they
were less likely to live successful, long lives without the support of a social network to hunt
prey, gather food or raise offspring. The risks of starvation, attack from a predator or death via
injury were all mitigated by integrating into a social group. When humans experience social
isolation, a stress response serves as an adaptive signal for heightened vigilance and social
motivation promoting group inclusion is heightened (Grant et al., 2009; Leary et al., 1995; G. A.
Matthews & Tye, 2019). However, if isolation is persistent, chronic stress takes a large toll on
Acute opioid use
Temporary
pharmacologically-
induced sense of social
connection
Decreased motivation to
experience social
connection
Increased perception &
experience of social
isolation
4
our bodies, making social isolation risky in and of itself. Thus, despite the fact that the
environmental risks accompanying social isolation are mostly gone in our modern world (people
can order food using an app on their phone, remain at home under shelter, and even work
remotely behind a screen all day and survive in American society), there are still clear physical
and psychological benefits of strong positive social bonds. Individuals who are more socially
integrated live longer lives and adolescents with strong social ties are less likely to experience
mental illness (Lamblin et al., 2017).
There are decades of observational research suggestive of a causal link between social
relationships and health. Yang and colleagues employed a life course approach to identify
potential mechanisms for this association (Yang et al., 2016). Combined longitudinal data from
four different nationally representative datasets covering adolescence through late adulthood
revealed that social relationships impact health through changes in physiological functioning.
Outcomes included physiological functioning and incidence of physical disorders. Positive social
relationships are protective against physical disease, and promote better physiological
functioning in a dose–response manner. Across the life span, those with few or weak social
bonds are more likely to experience chronic stress, inflammation and obesity (Yang et al., 2016).
Humans are not unique in their reliance on social relationships – baboons with stronger
lifelong social bonds live longer than their lonelier peers (Silk et al., 2010). In rodents, social
neglect after birth is linked to a compromised immune response to stress over the life span,
increasing vulnerability to disease in adulthood (Hermes et al., 2006). Additionally, there is some
evidence of sex differences in vulnerability to isolation stress. Male rodents’ immune response
was weakened even more than the immune response for female rodents in the socially isolated
condition. This mirrors what we see in the human literature – men who are socially isolated are
5
at higher risk of mortality than isolated females, with some studies linking this difference to
heightened chronic inflammation in men (House et al., 1982; Yang et al., 2013).
A series of experiments in the 80s that have since been dubbed ‘Rat Park’ reported that
social isolation contributes causally to addiction: isolated rodents were more likely to become
addicted to substances than socially housed animals (Alexander et al., 1981). While the power of
social rewards is, of course, not limitless (Bozarth et al., 1989), recent rodent work using an
operant choice paradigm directly demonstrates that drug self-administration (methamphetamine
and heroin) is dramatically reduced by the availability of a competing social reward (opportunity
for physical contact; Venniro et al., 2018). In humans, researchers argue that “addiction can be
understood and potentially ameliorated via a contextualized reinforcer pathology model in which
lack of alternative reinforcement is a major risk factor for addiction” (Acuff et al., 2023). In all,
the quality and quantity of social relationships for humans and other mammals have profound
impacts on life, health and well-being: for meta-analysis and synthesis, see (Holt-Lunstad et al.,
2010).
Part II: The Opioid Crisis is Temporally (And Possibly Neurobiologically) Linked to Social
Isolation
Are Social Isolation and Addiction Linked?
There is a well-documented connection between social isolation and addiction. Many
people have had the unfortunate experience of watching a friend or family member struggle with
addiction and have seen the toll addiction takes on social connections and relationships. A
person’s motivation to use a substance erodes social ties over time as they begin to miss social
obligations, behave in secrecy to obtain or use the substance, and become less invested in
relationships that exist outside of the drug-use sphere – behaviors which physically and socially
6
isolate the person from their family and friends (Daley et al., 2018; gili et al., 2017; Volkow et
al., 2011). This bidirectional association creates a cycle in which an individual may cope with
feelings of isolation by engaging in drug use, which then further isolates them from society and
their loved ones, leading them to engage in more drug use and so on.
While isolation is associated with substance use disorders in general, there is suggestive
evidence that social wellbeing is particularly impaired among those with an opioid use disorder.
People who use opioids are more likely than those who use other substances to have several risk
factors for poorer social wellbeing including: unstable social networks (Bohnert et al., 2009;
Buchanan & Latkin, 2008; Costenbader et al., 2006; Saladin et al., 1995), using alone (Barman-
Adhikari et al., 2015; De Pirro et al., 2018; Wojcicki, 2019), unstable employment (Becker et al.,
2008; Segest et al., 1990), lower educational achievement (Chatterji, 2006; Johnston et al., 2016;
Martins et al., 2015; Register et al., 2001) and experience stigma. Stigma plays a large role in the
ostracization of people who use drugs, especially drugs which are deemed less socially
acceptable, such as illicit opioids or methamphetamine (Brown, 2015). Stigma against opioids is
multifaceted: stigma comes from the public, from family, and from health practitioners (Olsen &
Sharfstein, 2014). The general public often expresses disdain, disgust and contempt for
individuals with an opioid use disorder for their ‘moral failings’ and inability to quit using drugs.
Additionally, those who seek medication-assisted treatment (which some physicians are reluctant
to prescribe) are at higher risk of being ostracized from the recovery community, as many peer
group programs reject the use of opioid medications to treat opioid use disorders (Olsen &
Sharfstein, 2014). This complex issue is compounded by drug criminalization: among all
individuals released from prison, those with a history of a substance use disorder are the least
socially integrated, with unstable housing and low levels of employment (Western et al., 2015).
7
Individuals who use opioids are among the most stigmatized by the public, peers, and health
practitioners.
Social isolation is associated with psychological states that are commonly comorbid with
drug use, specifically depression. Depression is diagnosed in about 8% of the U.S. general
population but is present in 25–30% of people who use heroin (Brody et al., 2009; Darke &
Ross, 1997; Dinwiddie et al., 1992; Havard et al., 2006). Recent work has found that while there
is a bidirectional relationship between loneliness and depressive symptoms, loneliness is a
stronger predictor of later depressive symptoms than the other way around (Vanhalst et al.,
2012). This indicates that in most cases, experiences of isolation and loneliness are risk factors
for a depressive episode later on in life. The role of loneliness in the onset of depressive
symptoms was demonstrated in a cohort of older adults (50–68 years old at study onset), which
concluded that loneliness predicted future depressive symptomology but not vice versa
(Cacioppo et al., 2010). According to research from the UK, feelings of loneliness are highest
among young adults (18–24 years old) compared to older adults, a new trend that is troublesome
considering that endorsement of loneliness was associated with increased risk of depression,
unemployment and poor health later in life (Matthews et al., 2019).
Are People Who Use Opioids at Higher Risk of Suicide?
Individuals with a substance use disorder have a higher risk of suicide than the rest of the
population (Schneider, 2009; Vijayakumar et al., 2011). A 1997 meta-analysis found that there is
a 4-fold increased risk of suicide among people who use cannabis, a 6-fold increase in risk
among people who use alcohol, a 14-fold increase in risk among people who use opioids and a
20-fold risk among people reporting polysubstance use (E. C. Harris & Barraclough, 1997).
Suicide risk for those using opioids may actually be underestimated as some deaths may be
8
misattributed as accidents, rather than suicide, due to the uncertainty around the circumstances
and the isolated lives many people who use opioids come to lead. The proportion of suicide
deaths among people who use heroin ranges from 3% to 35%, with most studies reporting a
proportion between 3% and 10% (Darke & Ross, 2002). When comparing the predictors of
suicide among those who use opioids to the predictors of suicide among a community sample,
they are mostly the same, but the prevalence of these risk factors (including social isolation) is
much higher among people who use opioids. A more recent study assessed a range of suicidal
behaviors and ideations among people who use heroin and matched controls: they report that
those who use heroin are more likely to report remorse over not dying as well as a resolute intent
to commit suicide than their matched controls (Maloney et al., 2007).
Some argue that there is no specific link between opioids and suicide and that
confounding variables explain the relationship. One such argument hinges on the high rate of
polysubstance use: use of more than one substance is normative among those who use opioids
and also presents a larger risk factor for suicide (Darke & Ross, 2002). Yet, when comparing risk
among those who report single-drug use, there is an increased risk of suicide among those who
only use opioids when compared to people who report single-drug use of other drugs,
demonstrating plausibility of a specific link between opioid use and suicide (E. C. Harris &
Barraclough, 1997).
The argument for the specific link between opioid use and suicide is more convincing
when we compare a similarly stigmatized drug – methamphetamine. A study assessing suicide
risk over a 10-year period from 1999 to 2009 found that any drug and alcohol use among youth
is associated with a higher odds ratio of suicide risk (Wong et al., 2013). The associated risk of
suicide based on 10 different classes of substances shows that people who use heroin have the
9
highest odds ratio for suicide risk, with methamphetamines use yielding the second highest risk
(Wong et al., 2013). The difference in risk (shown in univariate odds ratios) associated with
heroin and methamphetamine – when controlling for other predictors of suicide – grows as
suicide risk becomes more severe: suicidal ideation (5.0 vs 4.3), suicidal plans (5.9 vs 4.5),
suicide attempts (12.0 vs 7.1) and serious suicide attempts (23.6 vs 13.1).
Researchers have applied the Brain Opioid Theory of Social Attachment (BOTSA)
framework (described in detail in Part IV) as a mechanistic explanation of the link between
opioids and suicidality, positing that lifetime experiences of social pain are associated with
dysfunction in the opioid system and that this dysfunction can lead to specific psychopathologies
such as depression and suicidality (P. Lutz et al., 2020; Panksepp et al., 1980). Additionally,
opioid misuse – but not opioid use – is associated with a greater risk of later suicide attempt
(Samples et al., 2019). Recently, public health researchers have deemed that the opioid and
suicide epidemics in America are a syndemic, rather than independent epidemics, and that they
bidirectionally exacerbate one another (Fornili, 2018). Individuals may be driven to consume
opioids by the need to relieve physical and social pain and that when the pain relief is blunted
from chronic use people may turn to suicide as a means to alleviate the pain (Nobile et al., 2020).
This combined body of work suggests that chronic opioid use is intimately tied in with chronic
social pain, both of which can lead to dysfunction in the endogenous opioid system. This
psychobiological dysfunction may contribute to the high rate of suicide in this population.
Part III: Temporal Correlation Between Social Isolation and Opioid Use
Is Social Connectedness Declining in the USA?
Social connectedness typically refers to an individual persons’ relationships and social
wellbeing. Social capital is a measure of networks, relationships and resources in a community.
10
There has been a decline in social capital in the USA, documented by Robert Putnam first in his
1995 essay and then soon after in his book “Bowling Alone: The Collapse and Revival of
American Community” (Putnam, 1995, 2000). Longitudinal data suggest that there has been a
decline across community and political membership (Putnam, 2000). People in modern society
have smaller core networks and fewer non-kin individuals in those networks (Hampton et al.,
2011). Additionally, more adults in the USA are living alone compared with people in most other
countries (Pew Research Center, 2020). Reports on human loneliness argue that societally, we
should be concerned about the global epidemic of loneliness that is taking a massive toll on
human life and health, including contribution to substance use disorders (Snell, 2017). Overall,
loneliness and social isolation in the USA appears to be increasing at the same time that opioid
use is rising.
Are There Regional Variations in Patterns of Isolation and Addiction?
Geographic trends can help to identify the relationship between social capital and opioid
use. There is evidence that social capital is a protective factor against overdose as a cause of
mortality (Zoorob & Salemi, 2017). The CDC published maps charting trends of opioid overdose
rates across the USA. There are clear geographic differences in terms of where the opioid
epidemic is affecting communities the most. Similarly, geographic data exist for measuring
social capital in the USA. The Joint Economic Committee of the US Congress published data
collected between 2013 and 2016 showing a composite social capital index indicating which
states have high and low social capital. The index is composed of seven subscales: (i) family
unity, (ii) family interaction, (iii) social support, (iv) community health, (v) institutional health,
(vi) collective efficacy and (vii) philanthropic health.
11
Looking at local communities, there is evidence of an association between social capital
and opioid overdose rates. We should interpret this with caution, as many of the metrics to
calculate social capital are correlated with socioeconomic status and poverty. West Virginia has
the highest rates of drug overdose in the country, with 51.5 deaths per 100 000 (age-adjusted
death rates) and also the highest rate of opioid-specific overdose with 86% of all fatal drug
overdoses attributed to opioids in 2016 (2017 Drug Overdose Death Rates, 2017; Drug Overdose
Deaths in West Virginia | County Health Rankings & Roadmaps, 2017). In West Virginia, there
is a relationship between social capital and drug mortality: counties with the highest drug
overdose rates are those with the lowest social capital.
Figure 2: A linear model depicting the correlation between drug overdose mortality rates and social
capital among 35 counties in West Virginia.
In Figure 2, I present a novel analysis of publicly available data: I carried out a linear
regression model using R to assess the relationship between social capital and overdose rates
using overdose data from West Virginia County Health Rankings 2017; the report used data from
12
2013 to 2015 CDC WONDER mortality data, which produces age-adjusted death rates for each
county in the USA(Drug Overdose Deaths in West Virginia | County Health Rankings &
Roadmaps, 2017; SCP Index - Social Capital Project - United States Senator Mike Lee, 2018).
For this analysis, I used the overdose death rates by number per 100,000 deaths in each county,
available at the website in the reference list. An increase of one standard deviation of social
capital (publicly available data from the Social Capital Project; range is 10–67) corresponded
with a 10-point reduction in overdose death rates (range for fatal overdose rate per 100 000
deaths is 11–93; P < 0.001). The county data are not specific to opioids, but as noted above, most
overdose deaths in West Virginia (86%) were attributed to opioids (Drug Overdose Deaths in
West Virginia | County Health Rankings & Roadmaps, 2017). This provides support for the
claim that there is a community-level association between fatal drug overdoses and social capital.
Recent work has looked into social capital as a direct protective factor against overdosing
on opioids. At the individual level, the critical factor in social capital is a person’s own social
networks; social capital can be predicted at the community level, looking at the density of the
community, the level of civic engagement, and a sense of belonging, trust and reciprocity within
the community (Zoorob & Salemi, 2017).
Overall, societal and cultural contexts may be more associated with opioid use and
overdose than previously thought. The combined inputs of a person’s social capital (as described
above) contribute to the opioid epidemic. The characteristics of social capital, including the
ability of an individual to have healthy social networks, a sense of belongingness and
participation in the community, are intimately tied in with the psychobiological argument that
endogenous opioids play a critical role in a person’s ability to experience motivation to pursue—
and pleasure from—social connection and bonding.
13
Part IV: The Endogenous Opioid System and Social Bonds
Is There Support for the Brain Opioid Theory of Social Attachment?
The Brain Opioid Theory of Social Attachment (BOTSA) was proposed two decades ago
as an explanatory model of an organism’s capacity to form social attachments through a
neurobiological lens (Panksepp, 1998). This theory asserts that the endogenous opioid system is
one of the neurobiological substrates underlying primates’ capacities to form lifelong social
bonds. I apply the BOTSA framework as a mechanistic explanation of the link between opioid
use disorders and social isolation. The neural underpinnings of our ability to form relationships
with intimate partners, parents and children has largely been studied with a focus on oxytocin,
which is present in a range of species from rodents to primates. Oxytocin is critical for the
formative period of pair bonds with parents and romantic partners (Borrow & Cameron, 2012;
Ferguson et al., 2000; Carter et al., 1995). In Machin and Dunbar’s review of BOTSA, they state
that a reliance on rodent models has led us to overstate the role of oxytocin, while
simultaneously stunting research into complementary neurobiological substrates that may
underlie the complex and enduring social relationships seen in primate species—particularly the
endogenous opioid system (Machin & Dunbar, 2011). They argue that oxytocin is crucial for the
‘onset’ of relationships but that the opioid system plays a larger role when it comes to the
‘maintenance’ of pair bonds across the lifetime. Primates, including humans, are among the few
species with the capacity to form and maintain lifelong bonds between non-kin individuals. Soon
after endorphins (a class of endogenous opioids) were discovered in the 1970s, they were
proposed as a neurochemical substrate for parental and romantic relationships (Bell & Malick,
1976; Herman, 1979; Panksepp et al., 1980). This was in part due to the emotional and
behavioral similarities observed in those with an opioid addiction and those in serious romantic
14
relationships. Individuals who became addicted to opioids were seen as obsessive in the way that
a teenager is obsessive about a ‘first love’. Interestingly, the language employed by individuals
who use opioids fits with the empirical claim that opioids produce feelings of social warmth and
connectedness. In Heilig’s popular book ‘The Thirteenth Step’, he discusses how heroin
subjectively makes a person feel like they are getting a ‘hug from mum’ (Heilig, 2015). Absolute
advocacy, an organization focused on mental health services and drug education, talks about
heroin like this (The Heroin Hug | Absolute Advocacy, 2016):
Imagine being wrapped in the world’s biggest, warmest, most welcoming hug. Now, imagine
having access to that hug at almost any time. Imagine it being on demand for good and bad days
alike. If you could wrap yourself in a needed or wanted hug whenever you wanted, would you?
And a first responder working in North Carolina says his patients describe a heroin high like this
(Stapleton, 2018):
And they said you know, the first time you do it, you just get this secure feeling. It’s almost like a
warm embrace, like a hug from your grandma. That’s the way it’s been explained to me. And
they said once you feel that you crave it constantly.
For comparison, the Drug Policy Alliance talks about the cocaine high like this :
People who use cocaine describe a feeling of alertness, power and energy. They are likely to feel
more confident and excited. They may also experience anxiety, paranoia and agitation.
These anecdotes demonstrate the idea that opioids (unlike other drugs such as stimulants) are
specifically related to the experience of belongingness and inclusion. Recent work assessing the
role of the endogenous opioid system in social bonding has found these claims to be supported.
From a neuropsychological perspective, chronic exogenous opioid use leads to
downregulation of the mu-opioid system, and may make it more difficult for individuals to
experience the rewarding feeling of ‘natural’ rewards, such as positive social interaction
15
(Goodman et al., 1996; Lutz et al., 2014; Stafford et al., 2001). It is important to note that there is
complexity within the system, as chronic use leads to upregulation in the kappa-opioid system,
which is thought to play a role in the ‘dark side’ of addiction as it is associated with stress-
induced relapse and the downregulation of the mesolimbic dopamine system (Karkhanis et al.,
2017). Thus, chronic use may exacerbate the need to continue to use higher doses to achieve a
sense of social well-being, while also further reducing motivation to pursue ‘natural’ social
rewards.
Is the Endogenous Opioid System Associated with Social Bonding?
Is the Endogenous Opioid System Associated with Maternal/Infant Bonding?
BOTSA predicts that the administration of an opioid agonist would decrease maternal behaviors
toward the infant, such as orienting to a crying pup. It has been hypothesized that the binding of
exogenous opioids provides the organism with a sense of warmth and contentment, thus
diminishing the need to fill this desire through physical touch and maternal bonding, which
release endorphins – an endogenous opioid (Bridges & Grimm, 1982; Grimm & Bridges, 1983;
Panksepp et al., 1980). This hypothesis is supported by the primate and rodent literature:
administration of morphine, an opioid agonist, decreases maternal bonding behaviors, whereas
the concurrent administration of morphine and naloxone (an opioid antagonist) eliminated this
reduction in maternal responsiveness to rodent pups (Bridges & Grimm, 1982; Grimm &
Bridges, 1983). Social touch also increases the release of endorphins; ‘kangaroo-care’ wherein
parents are encouraged to have skin-to-skin contact with their infants increases endorphin levels,
sleep quality and reciprocity in the mother and infant dyad.
Is the Endogenous Opioid System Associated with Social Bonding in Non-Human
Animals? The maintenance of non-kin relationships is also mediated by endogenous opioid
16
systems, in part through rough and tumble play. There is an increase in the release of endogenous
opioids when animals are partaking in rough and tumble play, and this increase is seen most
prominently in regions associated with social behavior and reward, particularly the amygdala and
nucleus accumbens (Panksepp et al., 1985; Trezza et al., 2010; Vanderschuren et al., 2016). The
opioid system is one of the very few neurochemical systems that have been found to increase the
subjective ‘liking’ of a stimulus, rather than just the observable ‘wanting’ component of
motivation (Berridge et al., 2009). Beyond that, the nucleus accumbens is one of the few brain
regions with a ‘hedonic hotspot’ that is activated by mu-opioids: social play is not just about
‘wanting’ or motivational salience, it is hedonically pleasurable (Berridge et al., 2009; Trezza et
al., 2010). Blocking mesolimbic dopamine—a core element of motivated or ‘wanting’
behavior—does not always affect social play behavior; however, blocking opioid receptor
activity with naltrexone in the nucleus accumbens does reduce social play behaviors (Trezza et
al., 2011). In primates, grooming releases endogenous opioids; experimental research has shown
that exogenous opioid agonists (such as morphine) reduce grooming behaviors as the opioid
system has reduced sensitivity due to higher receptor occupancy, whereas naloxone (an opioid
antagonist) increases such behaviors. When given naloxone, the primate increases grooming
behaviors; researchers have posited that this behavior ensues in order to increase the release of
endogenous opioids to fight the effects of the naloxone, activate the system and feel the
rewarding aspects of social touch (Niesink et al., 1996). While the administration of naloxone
and its subsequent effect on social and maternal behaviors are both interpretable from the
BOTSA perspective, it is not clear why blocking of opioid signaling appears to cause a reduction
in maternal reactivity and social play (a disruption of function) but an increase in social
grooming behavior (a compensatory response).
17
Is the Endogenous Opioid System Associated with Social and Romantic Bonding in
Humans? Social touch also plays a large role in the endogenous opioid system in humans.
Massages increase endorphin levels and a subjective sense of well-being (Kaada & Torsteinb,
1989; McKechnie et al., 1983). Additionally, social touch increases the availability of mu-opioid
receptors in the thalamus, striatum and frontal cortices—including the orbitofrontal cortex
(OFC)—key regions in reward and sociality (Nummenmaa et al., 2016). Recent evidence from a
double-blind experiment on touch where participants were given either naltrexone (an opioid
antagonist) or morphine (an opioid agonist) revealed no significant impact of opioid drug
condition on perceived pleasantness of touch (Løseth et al., 2019). However, the touch was a
brush stroke on the arm delivered by the experimenter who the participant could not see (e.g.
non-social touch); thus, the authors concluded that while the endogenous opioid system is not a
necessary component to feel pleasure from touch, it may play a role when the ‘context’ of social
touch is taken into account. These studies provide mixed results in support of the claim that the
endogenous opioid system modulates non-kin bonding through touch and play.
Lastly, the endogenous opioid system plays an important role in the development of
romantic relationships. As previously mentioned, part of the rationale for studying opioids as a
neural mechanism underlying social attachment was the clinical similarities between a budding
romance and a budding opioid addiction. Touch and the associated feelings of comfort, safety
and well-being are core elements of healthy romantic relationships. Indeed, endorphin levels
increase with sexual behavior (Jain et al., 2019). Opioid addiction negatively impacts
relationships across the board with detrimental outcomes for familial, social and romantic ties.
This disruption is somewhat more complex for romantic partners: individuals who are addicted
to opioids—males in particular—tend to lose acute sexual interest in their partners, with
18
impairments in both psychological and physiological arousals (Khajehei & Behroozpour, 2018).
Opioids produce physiological and hormonal changes that result in a reduction in sexual
behavior—effects which have not been found from non-opioid substance use. People who use
heroin experience opioid-induced hypogonadism, and males who use opioids have lower
testosterone levels (Khajehei & Behroozpour, 2018; Rasheed & Tareen, 1995). Qualitative
research has found that people who use opioids report feeling that the drug replaces the need for
sex, and the need to be around friends: chronic use appears to both diminish the value of
interpersonal relationships (which typically become strained in the course of a substance use
disorder), and simultaneously increase the value of the drug which provides feelings associated
with positive social relationships (that over time have become less available – and less positive –
to the individual; Albertín & Íñiguez, 2008). A study on women who use opioids reports that
people who use opioids tend to remain in relationships mostly for the functional purposes a
partner serves, rather than the emotional support and love that may otherwise be important
factors (Rosenbaum, 1981).
Both the acute and chronic effects of opioids have deleterious effects on sexual
functioning, although the acute effects are more varied: there is some evidence that individuals
may use opioids as a self-medication for sexual dysfunction, such as premature ejaculation in
males or dyspareunia (pain during intercourse) in females (Peugh & Belenko, 2001). This is in
direct opposition to the acute effects of other psychoactive substances, including cocaine,
methamphetamine, and alcohol – all of which increase libido and risky sexual behavior (DeVido,
2020; Simons et al., 2018). In fact, researchers are assessing the interrelated nature of sexual
behavior and meth addiction: there is evidence that incorporating therapy focused on sexual
behavior may improve drug treatment outcomes for meth-dependent individuals (Knight et al.,
19
2019). Overall, the endogenous opioid system is responsible for more than pain regulation; it
plays a role in feelings of well-being and comfort that come from social bonds, romantic bonds
and the hedonic reward system.
In the human brain, neural pathways for social integration/exclusion and endogenous
opioids overlap in several key areas related to addiction: the insula, the amygdala and the
striatum. The following is a brief summary of the role these systems play in substance use and
social connection. First, the insula was linked to substance addiction through studies on patients
with insular brain lesions; among patients with this lesion who smoked, most reported marked
decrease in cravings to use cigarettes and overall cigarette smoking (Naqvi & Bechara, 2009).
The insula plays a role in the human ability to perceive interoceptive cues, including feelings of
craving as well as physical and social pain (Eisenberger, 2015; Heilig et al., 2016). Second, the
amygdala is associated with stress-induced drug-seeking behavior and is critical for responding
to natural rewards, including social connection (Sharp, 2017). Third, the striatum has been
established as a key region associated with reward. Both social integration and addictive
substances produce subjective feelings of reward, which are represented in the brain via the
release of endogenous opioids and mesolimbic dopamine. For a more detailed review on the role
that these key brain regions play in social inclusion and addiction, see Heilig et al. (2016).
Conclusion
The endogenous opioid system provides a psychobiological mechanistic explanation for
the role of social connectedness in addiction, specifically opioid use disorder. The opioid
epidemic can be viewed through the BOTSA framework as an epidemic of social isolation and a
lack of belongingness, which people seek to mitigate through the use of opioids. To the degree
that opioids directly mimic the positive feelings experienced through close social connection,
20
opioid use may become more compelling as a result of increasingly impoverished authentic
social connectedness within society. Chronic opioid use affects an individual’s psychological and
neurological well-being, impairing his or her ability to participate as a member of a cohesive
social group. I posit that there is a connection between the rise of opioid use and increasing
social isolation; policy-makers, psychologists and clinicians should consider the impact of social
capital, social connectedness and social isolation when addressing issues faced by individuals,
families and communities impacted by substance use disorders. The endogenous opioid system
may be a mechanistic basis for bidirectional causal links between social isolation and opioid use
disorder.
While more work is needed to evaluate several pieces of the puzzle surrounding the
endogenous opioid system, social isolation and addiction, there is enough evidence to claim that
social isolation and substance use are bidirectionally exacerbating one another. An opinion
article in Nature Reviews Neuroscience states that we already have enough knowledge of the
association between the addiction and social belongingness to implement effective treatment
interventions (Heilig et al., 2016):
Improving the social integration of drug users through opportunities for housing, jobs and
meaningful relationships is therefore not merely a nonspecific intervention but rather a
neurobiologically specific and critically important way to decrease drug use.
In recent years, there have been more calls to reframe the study and treatment of substance use
disorders, altering the perception of these conditions as an individual ‘disease’ or moral failing
toward a new conceptualization of addiction as a community- and cultural-level issue that can be
better addressed through larger-scale social changes as opposed to solely relying on individual
therapy and treatment (Alexander, 2012). Integrating social inclusion into the framework of
research will increase the clinical utility of neurobiological studies aimed at evaluating the
21
mechanisms and potential treatments for substance use disorders. Identifying and accounting for
relevant social variables in both research and treatment practices is a critical element in treating
those who suffer from chronic, debilitating, and lethal substance use disorders.
22
Chapter 1: A Sense of Belonging: Opioids Mimic the Emotional Experience of Social
Connection
Abstract
Chapter 1 will focus on evaluating the hypothesis that acute opioid use is associated with
subjective feelings of social connection. Chapter 1 addresses the question: is the “heroin hug” a
common experience among those who use opioids? Social connection is critical for wellbeing,
and people with a substance use disorder are more likely to experience isolation and loneliness.
The Brain Opioid Theory of Social Attachment integrates evidence that the endogenous opioid
system is critical for maintaining social bonds, and for the positive feelings associated with
social connections. The present study used a retrospective self-report survey design to evaluate
the subjective emotional experience of the acute high from alcohol, marijuana,
methamphetamine, and opioids. The hypotheses were as follows: 1) Compared to the other drugs
considered, greater feelings of social wellbeing and connection would be attributed to the high
from opioid use (consistent with BOTSA), and 2) among opioid users, greater reported social
feelings derived directly from drug use would be associated with more severe reported drug use
problems. Data was collected online through Prolific.co. Participants were 325 people with a
history of problematic substance use. We collected demographic information and used a novel
measure to evaluate the subjective emotional experience of the acute drug high. Positive
emotional experiences that we considered to be associated with social connection (like a hug,
sense of belonging, loved, secure) were evaluated along with other positive emotional states.
Those who used opioids were significantly more likely to endorse hug/belong items, but there
was no effect of drug of choice on loved/secure items (although those who use opioids scored
directionally higher on these items). Interestingly, while reported drug use severity was
23
associated with more social feelings from initial drug use (hug/belong), that relationship was not
specific to people who used opioids; across participants, those that more strongly endorsed these
experiences when they initially used their drug of choice (prior to problematic use) also reported
subsequent development of more severe addiction drug use problems.
We provide empirical evidence for the “heroin hug”; people who used opioids report that
the high feels like a hug and elicits a sense of belonging. People who use opioids may be at
heightened risk for problems related to social isolation, and are likely to benefit from additional
social supports.
Introduction
Social Behavior and Substance Use
Humans are hyper-social beings (Baumeister, 2010). Strong, positive social relationships
improve longevity, health, and quality of life (Silk et al., 2010; Yang et al., 2016). Social
isolation and loneliness are linked with negative physical and mental health outcomes including
early mortality, inflammation, and depression (Cacioppo & Cacioppo, 2014; Holt-Lunstad et al.,
2015; Social Isolation And Health | Health Affairs Brief, n.d.). In the United States, people have
fewer close friends than they did 30 years ago (Cox, 2021; McPherson et al., 2006), continuing a
trend that appears to extend back at least to the mid 20
th
century (Putnam, 1995). COVID-19
exacerbated this trend as people reported increased social isolation and feelings of loneliness
(Kovacs et al., 2021).
Social wellbeing influences a wide array of health-related behaviors, including substance
use. People use drugs for many reasons, including pursuit of pleasure and facilitation of positive
social interactions, or to relieve pain or escape social distress (Cooper et al., 2016; Deckman et
al., 2014; Patrick et al., 2018; Votaw & Witkiewitz, 2021). For example, social rejection from a
24
close friend or family member is associated with an increased likelihood of drinking later that
day (Laws et al., 2017). People who felt lonelier than the average person during COVID-19
drank more drinks per day than their less lonely counterparts (Bragard et al., 2021). People who
report problematic substance use commonly experience the erosion of social bonds, which can
lead to increased drug use (Dingle et al., 2015; Ingram et al., 2020; Wesselmann & Parris, 2021).
As substance use becomes clinically significant, decision-making increasingly favors the
acquisition and use of substances over other obligations and opportunities, including social
connection (Robinson et al., 2022; Volkow et al., 2011).
Clinical interventions for substance use often target social isolation, either by providing
new social connections within a recovery community (Laudet et al., 2002), or by focusing on
repairing damaged relationships (Mericle, 2014; Meyers & Miller, 2001). Peer-run treatments
focus on human connection through shared experiences of addiction, including peer support
groups like Alcoholics Anonymous (Stone et al., 2017), which works in part through increasing
pro-abstinence social network ties (J. F. Kelly et al., 2011).
The Opioid System and Social Functioning
The endogenous opioid system plays an important role in the emotions that accompany
social attachment. According to the Brain Opioid Theory of Social Attachment (BOTSA)
(Machin & Dunbar, 2011; Panksepp et al., 1980), this system contributes significantly to the
reward associated with social ties across the life span. People with an opioid use disorder
commonly experience strained relationships across familial, social and romantic ties (Strang et
al., 2020). Consistent with this, opioid dependence is (relative to other drugs) more closely
linked with a fearful-avoidant attachment style, which is characterized by a fear of reliance on
25
others but a simultaneous strong unmet desire for social connection (Bartholomew & Horowitz,
1991; Kyte et al., 2020).
Given that the release of endogenous opioids appears to contribute significantly to
feelings of socioemotional wellbeing, it is possible that social connection is particularly
important in addiction to, and recovery from, opioids (Christie, 2021; Panksepp et al., 1997).
Academics, advocates, and front-line workers describe a phenomenon that has been dubbed the
“Heroin Hug”. In his book “The Thirteenth Step”, Markus Heilig describes how heroin makes a
person feel like they are getting a “hug from mum” (Heilig, 2015). Absolute Advocacy says that
using heroin feels like “being wrapped in the world’s biggest, warmest, most welcoming hug”
(The Heroin Hug | Absolute Advocacy, 2016). A first responder who has reversed dozens of
overdoses says his patients describe the heroin high as “a warm embrace, like a hug from your
grandma” (Stapleton, 2018). For comparison, those who use cocaine say that high gives them “a
feeling of alertness, power and energy” and “feel more confident and excited” (What Does It
Feel like to Use Cocaine? | Drug Policy Alliance, n.d.). People describe the methamphetamine
high as “an intense rush that produces high energy, focus, and euphoria” (Bacsi, 2022).
Social Facilitation Vs. Sociodelic Effects
An important distinction can be made between 1) drug-effects that facilitate social
engagement, indirectly leading to more of the positive feelings attendant to social interaction, vs.
2) drug-effects that directly generate feelings normally caused by social connection (which may
come to replace actual social connection). We will refer to these as “social facilitation” and
“sociodelic” effects respectively. Social facilitation is commonly reported in connection to
alcohol, ecstasy, methamphetamine, and marijuana, with people reporting easier and/or deeper
connections with others during drug use (Dumbili et al., 2021; B. C. Kelly et al., 2013; Kilwein
26
et al., 2022; Morean et al., 2013; Sumnall et al., 2006). While both impact substance use
behaviors, we are particularly interested in the latter as likely more relevant to addiction.
Sociodelic effects may lead to a downward spiral in which drug use ameliorates feelings of
isolation in the short-term – but, because it replaces rather than facilitates social connection,
exacerbates isolation in the long-term. And with greater isolation, the value of any sociodelic
effects the individual derives from drug use may increase. To the extent that people who use
drugs strongly experience such effects, they may promote and/or reflect a downward spiral
pattern of use.
Study Aims
Although it is established that social connection is generally important in substance use
disorders (McGaffin et al., 2017; Pettersen et al., 2019), there is no empirical evidence that the
opioid high produces more subjectively positive social feelings than other drugs, or that the
severity of opioid use is related to subjective sociodelic effects. Our hypotheses were that 1)
opioid use will be associated with greater sociodelic effects relative to other substances, and 2)
among people who use opioids, those reporting the strongest sociodelic effects will tend to have
had the most severe drug-related problems.
Method
Sample Selection
We used a retrospective survey design to evaluate feelings associated with the acute drug
high among people with a history of problematic substance use. The present study used a
convenience sample; we collected data through Prolific Academic Ltd., an online data collection
platform (Prolific.co) that has good transparency, functionality, and a relatively high minimum
hourly payment for participants (Palan & Schitter, 2018). We invited adults to complete a
27
screening survey to determine eligibility. Individuals screened into the full study if they (a) had a
history of problems with substance use, (b) currently do not engage in problematic substance use
(i.e., are not using any substances, or use occasionally/casually with no problems), and (c) were
adults currently residing in the United States who have Prolific accounts. We evaluated
individuals in recovery (asking about their initial experiences) rather than individuals who were
currently experiencing substance use disorder symptomology in order to minimize altered
emotional responses due to variability in tolerance and to mitigate response variability related to
active intoxication. For our first hypothesis, we were interested in comparing people whose drug
of choice (DOC) was non-prescription opioids, methamphetamine, marijuana, and alcohol to one
another. We oversampled those who used opioids and methamphetamine to ensure that we had
an adequate sample size to statistically compare between groups. We oversampled by selectively
inviting those who indicated that their primary drug of choice was opioids or methamphetamine
in the screener to complete the full survey until we had ~50 participants in each category for
DOC. Participants completed the full study between June 6, 2021 and August 14, 2021. We
invited all participants to take the survey a second time between November 18, 2022 to March 6,
2022. In the present study, the second survey timepoint data is only used for confirmatory factor
analysis as part of the psychometric analyses.
Measures
The full survey included demographic questions and a measure of the subjective
emotional experience of being high. For the latter, participants were instructed to indicate the
degree to which the high induced each of eight positive emotional experiences, four of which we
a priori considered to be associated with social connection (Like a Hug, Sense of Belonging,
Loved, Secure) and four of which we did not consider especially linked to social connection
28
(Happy, Excited, Content, Satisfied). Some items were taken directly from the Positive and
Negative Affect Scale (Watson et al., 1988). Responses were made using a Likert-scale where 1
indicated, “strongly disagree” and 5 indicated, “strongly agree.” The measures were the same the
second time participants were invited to complete the survey, with the exception that
demographic questions were not asked in duplicate, and the emotional response questions had
additional text in the instructions. in the original survey, the instructions read as follows: please
think about the time just when you initiated use of [DOC], during the initial phases or the first
few time(s) you tried it. In the second survey, the following was added to the instructions: please
try to think only about the direct chemical effect [DOC] had on your experience (apart from
anything else that might have contributed to how it felt). All study measures were approved by
the local Institutional Review Board. Statistical analyses were completed using R version 4.2.2.
Analytic Procedures
Demographic differences. We used chi-square tests and ANOVA to assess whether
there were any significant differences between groups (based on drug of choice) in demographic
characteristics and drug use characteristics.
Psychometric properties of positive feelings associated with the high. We conducted
an Exploratory Factor Analysis (EFA) to determine the factor structure of the 8 items that
characterized participants’ drug experiences (Costello & Osborne, 2019; Fabrigar et al., 1999).
First, using only the primary data collected in the first wave of the survey, we completed a
parallel analysis with the factor analysis function in R. Using these parallel analysis results, we
conducted an EFA (again only with the data from the first wave of the survey) with the principal
axis factoring method and oblimin rotation. Then, to validate the resulting EFA structure, we
29
conducted a Confirmatory Factor Analysis (CFA) on the subset of participants who completed
the second wave of the survey.
Emotional response by drug of choice. We first tested a two-way interaction between
DOC (4 levels) and emotion category (as determined via factor analysis) to predict endorsement
of subjective emotions - using the mean scores for each category. We then used ANCOVA to
determine if there was a significant effect of DOC on positive feelings for each of the unique
emotion categories (one model for each emotion category). We included the following covariates
in each model: DAST (Drug Abuse Screening Test - severity of use), education (ordinal variable,
8 levels), and Age of First Intoxication. Five individuals were missing all demographic data and
were excluded from the analyses.
Emotional response by severity of use. To assess our hypothesis that sociodelic effects
of the drug high emerge specifically among people who use opioids, we modeled DAST scores
(a measure of severity of use) as a function of 1) DOC, 2) each of the four emotion category
composites pairs, and 3) interaction terms between DOC and each of the four emotion category
composite pairs, resulting in nine independent variables. We were particularly interested in
DAST as a function of DOC and reported acute subjective drug effects on social wellbeing
during initial use as potential support for the hypothesis that for opioids in particular, endorsing
sociodelic effects would be associated with heightened severity, as discussed above.
Results
Demographic Characteristics of the Sample
Demographic information on study participants (n = 325) is presented in Table 1.
Participants across DOC categories did not differ significantly in income, age, ethnicity, or
gender. However, the groups varied on education, DAST, and age of first intoxication. Those
30
whose DOC was alcohol or marijuana tended to be more educated and have a lower DAST score
than those whose DOC was either opioids or methamphetamine. Those whose DOC was alcohol
also tended to have had their first intoxication experience at a later age than other groups.
Table 1
Demographic characteristics of the sample
31
Psychometric Analyses
We evaluated whether our data was suited for an exploratory factor analysis, first with a
Kaiser-Meyer-Olkin (KMO) test to measure whether we had sampling adequacy: the KMO value
was 0.82. Next, we used Bartlett’s test of Sphericity to assess whether the variables in the
correlation matrix were correlated; the chi-square value was 2104.63 which is above the critical
value (p < 0.001). Both tests indicate that factor analysis is appropriate for our data. A parallel
analysis (using the fa.parallel function in the R package psych) determined that a 4-factor
solution fit the data best. We report the primary factor loadings in Figure 1, and the full loadings
in Table 2. The factor structure separated the eight items into four categories, each of which were
predominantly linked with two items. Two of the pairs consisted of emotions which we a priori
considered to be associated with positive social experiences (Hug/Belong, and Loved/Secure).
The other two pairs consisted of positive emotional experiences that we did not consider to be
especially tied to the social domain (Happy/Excited, and Content/Satisfied).
Table 2
Exploratory factor analysis loadings and variance explained
Factor 1 Factor 4 Factor 2 Factor 3
Belong -0.25 0.26 0.15 0.75
Hug 0.20 -0.12 -- 0.89
Secure 0.18 0.83 -- --
Loved -- 0.91 -- --
Happy 0.46 -- 0.65 --
Excited -- -- 0.95 --
Content 0.86 -- -- --
Satisfied 0.78 0.17 -- --
32
Factor 1 Factor 4 Factor 2 Factor 3
SS Loadings 1.72 1.63 1.38 1.37
Proportion Of
Variance
Explained
0.22 0.20 0.17 0.17
Cumulative
Variance
Explained
0.22 0.42 0.59 0.76
We then used wave 2 data (n = 183) to complete a confirmatory factor analysis to assess
the stability of the EFA results. The hypothesized user model showed a good fit to the data, as
indicated by the comparative fit indices (CFI = .996, TLI = .992), suggesting a satisfactory
model fit. The model did not significantly deviate from the observed data (χ²(14) = 16.547, p =
.281).
The Root Mean Square Error of Approximation (RMSEA) was .032 (90% CI [0.000,
0.082]), indicating a reasonable fit of the model to the data within acceptable bounds. The
following table presents the standardized parameter estimates and covariances for the latent
variables in the confirmatory factor analysis.
Table 3
Confirmatory factor analysis standardized loadings for each latent factor
Latent Variable Standardized Loading
Hug/Belong 0.77
Secure/Loved 0.82
Happy/Excited 0.94
Content/Satisfied 0.79
33
Table 4
Confirmatory factor analysis covariances between latent factors
Covariances Value
Hug/Belong - Secure/Loved 0.61
Hug/Belong - Happy/Excited 0.38
Hug/Belong - Content/Satisfied 0.45
Secure/Loved - Happy/Excited 0.41
Secure/Loved - Content/Satisfied 0.49
Happy/Excited - Content/Satisfied 0.50
Table 5
Confirmatory factor analysis latent factor loadings for each item
ITEM STD.LV STD.ALL
Belong 0.86 0.77
Hug 1.07 0.76
Secure 0.93 0.82
Loved 1.09 0.87
Happy 0.83 0.94
Excited 0.60 0.51
Content 0.83 0.79
Satisfied 1.27 0.95
Note. Standardized Factor Loading (std.lv): The standardized factor loading represents the correlation between the
observed indicator (item) and the latent variable, after taking into account the measurement error. Standardized
Factor Loading with respect to the Latent Variable (std.all): The standardized factor loading with respect to the
latent variable represents the correlation between the observed indicator and the latent variable, considering both the
measurement error and the variance of the latent variable.
34
Latent variable names have been replaced with more descriptive labels. The findings
from the CFA support the hypothesized model, suggesting significant relationships between the
latent variables and their covariances. Based on these results, we computed mean scores for each
pair of items grouped by the factor analysis.
We used the four resulting mean ratings in subsequent analyses rather than latent factor
scores to allow greater interpretability of between-category comparisons. The mean ratings are
fairly reliable with correlations from wave 1 to wave 2 ranging from 0.30 and 0.50 for each
emotion category.
Figure 1: Factor analysis plot with factor loadings from Wave 1.
Endorsement of Emotions by Drug of Choice
We examined whether endorsement of positive emotions differed as a function of DOC
and Emotion Category, with DAST, education, and age of first intoxication included as
35
covariates. We included a DOC X Emotion Category interaction term because we were
particularly interested in whether opioids might evidence sociodelic effects. We observed a
significant main effect of Emotion Category (F(3, 1229) = 33.41, p < .01). Here we report the
mean and standard deviation for each pair. This effect was driven by generally higher
endorsement of the Content/Satisfied (3.81 + 1.03) and Happy/Excited pairs (3.87 + 0.94),
followed by Hug/Belonging (3.53 + 1.06), and finally Secure/Loved (3.10 + 1.10). We observed
an overall main effect of DOC (F(3, 1229) = 5.19, p < .01) with somewhat higher emotion
ratings (across categories) for opioids (3.78 + 1.18) and methamphetamine (3.66 + 1.09) than for
marijuana (3.54 + 1.00) and alcohol (3.52 + 1.05). We observed a main effect of education:
people with the highest educational attainment reported higher ratings of emotions (3.77 + 1.28)
than those with the lowest educational attainment (2.75 + 0.87). There was no effect of age of
first intoxication. Most importantly, we observed a significant interaction between DOC and
Emotion Category (F(9,1229) = 3.05) indicating that the pattern of experienced emotions
differed significantly as a function of DOC. See Appendix Table 1 for complete model results.
To characterize the nature of the observed interaction between DOC and Emotion
Category, we separately modeled 1) Hug/Belonging, 2) Secure/Loved, 3) Content/Satisfied, and
4) Happy/Excited. Models were identical to that described above, apart from the omission of
Emotion Category and the Emotion Category X DOC interaction term, since each of these
models isolated a single emotion category. Below we present results of each of the four models
in turn: each model uses ANCOVA with DOC as the independent variable of interest, and DAST,
education, and age of first intoxication as covariates predicting the response to the Emotion
Category of interest.
Hug And Belong: Opioids Create a Sense of Belonging
36
Drug of Choice significantly predicted Hug/Belong responses (F(3, 258) = 4.04, p < .01);
see Appendix Table 2. In order to follow up on the nature of this effect of DOC, a post-hoc
Tukey HSD test was carried out for the Hug/Belong responses. Participants reporting on the high
from opioid use scored significantly higher than those who used alcohol (mean difference = 0.44,
95% CI [0.02, 0.85], p < 0.05) and marijuana (mean difference = 0.64, 95% CI [0.17, 1.12], p <
0.01). There was no significant pairwise difference for those who used methamphetamine (mean
difference = 0.41, 95% CI [-0.15, 0.96], p = 0.23). No pair-wise comparison that did not involve
the opioid group approached significance (each p-values > 0.5). See Figure 2. Neither education,
nor age of first intoxication were significant predictors of Hug/Belong. However, the DAST
score was a significant covariate in the model (β = 0.08, t(270) = 3.06, p = 0.002), which we
further explore in the section below reporting on models of DAST scores.
Figure 2: Violin plot of endorsement of Hug/Belong items by Drug of Choice. People who used opioids
report significantly higher endorsement of Hug/Belong feelings than those who used alcohol and
marijuana.
37
Secure and Loved
There was no significant effect from any variables included in the model as a predictor of
Loved/Secure items. See Appendix Table 3 and Figure 3.
Figure 3: Violin plot of endorsement of Secure/Loved items by Drug of Choice. There is no significant
relationship between Drug of Choice and endorsement of these items.
Content and Satisfied
Drug of Choice significantly predicted Content/Satisfied responses (F(3, 259) = 3.45, p <
.05); see Appendix Table 4. Neither age of first intoxication, education, nor DAST score were
significant predictors. In a post-hoc TukeyHSD test, participants reporting on the high from
marijuana scored significantly higher than those reporting on alcohol intoxication (mean
difference = 0.45, 95% CI [0.08, 0.83], p < 0.05), but there was no significant pairwise
38
difference for methamphetamine (mean difference = 0.35, 95% CI [-0.86, 0.17], p = 0.31) or
opioids (mean difference = 0.06, 95% CI [-0.53, 0.39], p = 0.98). See Figure 4.
Figure 4: Violin plot of endorsement of Content/Satisfied items by Drug of Choice. People who used
marijuana endorse significantly higher feelings for Content/Satisfied than those who used alcohol.
Excited and Happy
Drug of Choice significantly predicted Excited/Happy responses (F(3, 259) = 3.56, p <
.05); see Appendix Table 5. Neither education, age of first intoxication, nor DAST score were
significant predictors. In a post-hoc TukeyHSD test, participants reporting on the high from
methamphetamine scored significantly higher than those reporting on alcohol (mean difference =
0.44, 95% CI [0.02, 0.86], p < 0.05) and marijuana (mean difference = 0.70, 95% CI [0.23, 1.17],
p < 0.01). There was no significant pairwise difference for those reporting on opioids (mean
difference = -0.45, 95% CI [-0.93, 0.03], p = 0.08). See Figure 5.
39
Figure 5: Violin plot of endorsement of Excited/Happy items by Drug of Choice. Participants who used
methamphetamine endorse Happy/Excited items significantly more than those who used alcohol or
marijuana.
The Association Between Sociodelic Effects and Severity of Addiction
Drug of choice significantly predicted DAST scores (F(3, 251) = 19.37, p < .001), with
the highest scores reported among the opioid group (7.88 + 2.00) followed by the
methamphetamine group (7.41 + 2.00), and lower scores for the marijuana (5.76 + 1.96) and
alcohol groups (5.71 + 2.12). We also observed a main effect of Hug/Belonging: higher
endorsement was associated with greater severity of substance use (F(1, 251) = 4.80, p < .05).
No other emotion response category significantly predicted DAST scores (all p-vales > 0.35); see
Appendix Table 6 for full model results. Contrary to expectation, we did not observe a
significant interaction between DOC and Hug/Belonging response (F(3, 251) = 0.47, p = 0.70).
See Figure 6.
40
Figure 6: Endorsement of Hug/Belong items by Drug of Choice and DAST Score (substance use
severity) showing a main effect of DAST score on Hug/Belong items with no significant interaction by
Drug of Choice
Discussion
Our primary aim was to determine if, consistent with the BOTSA framework, opioids
have a greater sociodelic effect than other drugs. We observed partial support for this hypothesis.
In particular, among the items conceived a priori to reflect positive social emotions, one item
pair (Hug/Belonging) was differentially associated with opioid use, while the other item pair
(Loved/Secure) was not differentially associated with opioid use. People who used opioids were
more likely to report that the acute high felt like a hug, and gave them a sense of belonging than
those who used alcohol and marijuana (controlling for severity of drug use and other
demographic characteristics), with a similar (though not significant) difference observed in
comparison with methamphetamine use. This finding is in agreement with prior work on the
41
“heroin hug”, and on the overlap between the endogenous opioid system and feeling a sense of
belonging.
We had predicted that opioid use would be differentially associated with endorsement of
the four social-feeling terms in the present study (including loved/secure); while those who use
opioids did report directionally higher feelings on these items, there was no statistically
significant effect by drug of choice. One reason for this finding may be the influence of context.
While we instructed participants to respond to the survey thinking about the effects of the drug
high, it is possible (and even likely) that people may conflate the social-emotional context of
their use with the emotional experience induced by the drug itself. For example, a frequent
pairing of methamphetamine use with sexual intimacy (a social facilitation effect) might have
affected the degree to which participants indicated that the drug itself elicited a feeling of being
loved (Flores-Aranda et al., 2019; Lasco & Yu, 2023). Moreover, the unequal degree of stigma
differentially associated with the four substances we considered might have influenced the
feelings attributed to the drug effect. Future research should be directed at strategies to better
isolate sociodelic drug effects, apart from any contextual factors that may influence subjective
emotional experiences.
Our second aim was to assess whether a greater level of reported sociodelic effects
among people who use opioids was associated with greater problem severity. We anticipated that
for this group, people who report more positive sociodelic effects from the drug would be more
likely to fall into the pernicious feedback loop wherein the drug replaces social connections, and
through continued use, they become less social, leading to the desire to further self-medicate, and
so on (Christie, 2021). Regarding this feedback loop, we observed no support for this hypothesis
about a relationship specifically linked to opioid use (p > 0.45 for all interactions between DOC
42
and emotion responses as predictors of DAST score). However, we did observe a main effect of
Hug/Belonging as a predictor of severity: participants who reported greater feelings of
Hug/Belonging while high also reported greater severity of drug use problems (β = 0.08, t(270) =
3.06, p = 0.002). This is consistent with a more general idea (not specific to opioid use) that
when drug use replaces social connection, it may reflect or even promote a downward spiral.
Although this was not anticipated, it is consistent with the idea that many psychoactive
substances can alleviate social pain. While this effect may be most pronounced for opioids,
which are effective in reducing the pain associated with loneliness (Christie, 2021; P. Lutz et al.,
2020), alcohol also reduces self-reported distress following a social exclusion task (Hales et al.,
2015), and the relationship between feeling lonely and feelings of low self-worth is attenuated
for those who report frequent marijuana use (Deckman et al., 2014). Although this finding is
post-hoc, we think the possibility that sociodelic drug effects are connected with problem
severity deserves further investigation.
Limitations
There are several limitations in the present study. This data is retrospective; people are
responding about their feelings when they first started using their drug of choice (which for most
people was many years ago, and does not vary significantly by drug of choice). This may result
in biased reports of emotional states when using a substance that later became problematic for
the people in the sample. While these retrospective reports may be biased, there is no evidence
that they would be differentially biased based on drug of choice. A second limitation is that we
are asking about emotional states evoked by specific substances, but, as noted above, emotional
states may differ by drug of choice for non-pharmacological reasons (i.e., context of drug use).
The present study is designed to assess the subjective emotional experience as it is recalled,
43
rather than the strictly pharmacological experience. With that said, our findings are in line with
prior research on the subjective experiences of different substances (e.g., methamphetamine use
was differentially linked to feeling Excited/Happy and marijuana use with feeling
Content/Satisfied; (Bacsi, 2022; Kish, 2008; Zeiger et al., 2010).
Conclusion
The acute opioid high produces short-lived feelings that, to some degree, mimic feelings
people get when experiencing deep social connection (consistent with the “heroin hug”
characterizations discussed above). This may reflect the role that the endogenous opioid system
plays in the formation and maintenance of social bonds. However, across substances (i.e., not
opioid-specific), people who report greater sociodelic emotions while high were more likely to
have a severe substance use problem. When drugs replace rather than facilitate social
connection, people may be at higher risk for a downward spiral of decreasing social
disconnection, and increasing drug use. People who experience sociodelic drug effects may need
additional supports to prevent problematic use and/or aid their transition to recovery.
44
Chapter 2: COVID-19 and Substance Use Behaviors
Abstract
Communities across the world responded to the COVID-19 pandemic in part through
initiating social distancing measures. These measures, which in many places in the United States
included Safer at Home messaging, present a unique opportunity to look at the impact of
population-level social stressors and isolation on substance use behaviors. The global pandemic
brought about significant changes in how people interact with one another, work, and
communicate. This societal level shift offered a quasi-experimental look at how social isolation
impacts behavior and wellbeing. This chapter will address the question: Did the implementation
of social distancing measures impact drug use behaviors?
Social distancing measures were widely adopted in response to the COVID-19 pandemic.
High levels of social connection are positively associated with beneficial health outcomes, while
social isolation is associated with poor long-term health outcomes including reduced life
expectancy. The present study evaluates the impact of social distancing measures during the
early period of COVID-19 on substance use behaviors among those in the United States. We
used an internet-based survey with participants (n = 157; 86 male) reporting a history of
problems related to drug use. We relied on ANOVA and logistic regression techniques to assess
the associations between social connection and substance use. People with more severe drug use
problems reported feeling more socially isolated during social distancing. Those who primarily
use alcohol reported higher global feelings of social connection than those who primarily use
opioids. During social distancing, participants reported an increase in alcohol and cigarette
consumption, and a decrease in cocaine use. Lastly, those who reported using drugs for social
reasons were less likely to have decreased substance use during social distancing. The current
45
study provides evidence that social distancing guidelines have impacted both substance use
behaviors and feelings of social and physical connection. Further, there are differential impacts
based on drug of choice. These results advance delineation of the connection between sociality
and drug use.
Introduction
Social distancing guidelines were developed by public health officials to reduce the
burden of the COVID-19 pandemic. These policies called for physical distancing from others in
an effort to reduce the spread of disease, specifically by limiting physical contact with others. An
unfortunate byproduct of this policy is that many individuals have also experienced social
isolation—or a decreased sense of social connection—due to limitations on physical
engagement. Social distancing measures have prevented people from physically seeing one
another (e.g., sporting events or cafes), and have simultaneously transitioned in-person
opportunities for social connection to remote ones (e.g., social media or video calls).
Prior work has demonstrated that a lack of social connection is specifically associated
with drug use. Factors related to feelings of diminished social connection (e.g., living alone)
increase the probability of prescription opioid and benzodiazepine misuse within older adults
(Day & Rosenthal, 2019). From a societal level, communities that have lower social capital have
higher rates of drug overdoses per capita (Zoorob & Salemi, 2017). Random assignment of social
connection is not viable; research on the topic has necessarily relied on correlational data, and is
inconclusive with regards to directionality . It is plausible that a lack of social connection is a
consequence of chronic drug use, rather than an antecedent. Given the barrier to experimental
work on the topic, social distancing guidelines put forth due to the coronavirus pandemic have
presented a unique opportunity to observe how social connection and drug use change during a
46
period of en masse social distancing.
The social distancing guidelines widely adopted to reduce the spread of COVID-19 may
impact drug use in several distinct ways, some of which are directly related to social connection.
Daily routines have been disrupted and replaced by increased isolation and uncertainty. Changes
in social connection may lead to either increases or decreases in drug use behaviors. Those with
more unscheduled time as a result of lessened work commutes, partial or total job loss, and the
limited availability of social outlets may increase their drug consumption. Drug use may also
serve as a form of stress relief and escapism. Others who typically use drugs in social settings,
such as bars or parties, may reduce their drug consumption. However, there are additional factors
that have changed due to COVID-19 that are not directly related to social connection, such as the
economic impact. Studies on the Great Recession of 2008 reported an increased use of outpatient
mental health services and increased utilization of medications such as sleep aids, opiates,
antidepressants, and anxiolytics in years following the profound economic decline (Modrek et
al., 2015). Mental health problems and substance use were also found to be exacerbated by
recession-based experiences within financial, job-related, and housing domains (Forbes &
Krueger, 2019). While it would be ideal to examine the role of changes in social connection on
drug use without co-occurring features, the pandemic has altered many facets of everyday life all
at once, limiting the ability to distinguish between such features.
More recently, the American Medical Association released a brief detailing the increased
opioid-related mortality during the COVID-19 pandemic (COVID-19 May Be Worsening Opioid
Crisis, but States Can Take Action, 2020). Social distancing measures have impacted treatment
options, leading to ineffective delivery of medications like methadone and naloxone, as well as
limited access to harm-reduction centers (Khatri & Perrone, 2020; LeSaint & Snyder, 2020).
47
Risk for drug use is compounded by poverty, mental illness, and education level (Kurti et al.,
2016), which may have been also significantly changed as a result of COVID-19 and social
distancing measures. These measures may have exacerbated problems for those with an
addiction, considered as a “disease of isolation” by some in the medical community (Despite the
Challenges, We Must Fight Harder to Address the Nation’s Opioid Epidemic, 2020; Harris et al.,
2020). There is a great deal of evidence that isolation is associated with addiction (Copeland et
al., 2018; Harris et al., 2020; Zoorob & Salemi, 2017). Yet, the connections between isolation
and substance use are complex, in part because substance use is often a social activity. Moreover,
the connection between isolation and drug use is difficult to establish because it is not feasible to
experimentally manipulate social isolation in humans. A primary aim in the present study is to
draw inferences about the impact of social isolation by treating COVID-19 related restrictions on
social gatherings as an exogenous source of increased isolation. We were particularly interested
in the associations between feelings of isolation and drug use. The current study aims to examine
the impact of feelings of social isolation and connection on substance use patterns among a
sample of adults in the United States (U.S.) with a history of problems associated with substance
use, using the social distancing guidelines put into place during the COVID-19 pandemic as a
proxy for the experimental manipulation of social connection on a global scale.
Method
The current study employed a cross-sectional design to assess people who currently use
drugs and those with a history of drug use. Participants were recruited via Prolific Academic
Ltd., an online data collection platform (prolific.co) that has good transparency, functionality,
and a relatively high minimum hourly payment for participants (Palan & Schitter, 2018). Prolific
allows researchers to select participants using a two-part survey in which participants complete a
48
screening survey and are subsequently invited to complete the full study. All data were collected
during a period of encouraged social distancing (between April 17
th
and May 4
th
, 2020) and thus
all data pertaining to pre-social distancing are retrospective. The current study used the Drug
Abuse Screening Test (DAST-10; a validated clinical measure), as a screening survey to identify
participants with a history of at least a low level of problems related to drug use—defined as
scoring a 2 or above on the scale (Villalobos-Gallegos et al., 2015; Yudko et al., 2007). The full
survey included: a modified time-line follow back assessing number of days on which a
substance was used in the last 14-day period, as well as number of days used in a typical 14-day
period before the onset of social distancing measures (for alcohol, cigarettes and vaping,
participants were asked an additional question asking for the number of drinks/cigarettes/vape
pods per day); the Social Connectedness Scale (SCS)—a validated global measure which
assesses “cognition of enduring interpersonal closeness with the social world” (Lee et al., 2001);
Functions of Substance Use Scale—a validated measure of the motivation to use a substance
(e.g., for social reasons, physical reasons, or to enhance mood); single item Likert-scale
questions assessing physical and social isolation (one item for physical isolation, one each for
social isolation both pre- and during social distancing); and demographic questions on age,
ethnicity, socioeconomic status, sobriety status, and membership in Anonymous communities
(i.e., Alcoholics Anonymous, Narcotics Anonymous). Current feelings of social isolation and
social connection were measured using the Social Connectedness Scale. In addition to the SCS,
participants were asked to rate their pre-social distancing feeling of connectedness on a single
Likert-scale anchored at “socially isolated” and “socially connected.” Social isolation and
connection in this study are defined as opposing ends of a single spectrum. All study measures
were approved by the local Institutional Review Board.
49
Sample Selection
Participants were eligible for the current study if they reported a history of a minimum of
a low level of problems associated with drug use, as assessed via the DAST-10 (measure of
problems related to substance use). Previous literature has found that a cutoff score of three or
higher is associated with a DSM-3R diagnosis of substance abuse or substance dependence
(Bohn et al., 1991). A literature review of effective behavioral health screening tools within
primary care settings reinforced the use of the DAST-10 with a cutoff score of three or higher for
a substance use disorder, with a discussion of using two as a lower cutoff score for primary care
applications (Mulvaney-Day et al., 2018).
Analytic Plan
First, we evaluate how global feelings of social connection are associated with a) drug of
choice and b) severity of drug use. Next, we address how social distancing guidelines are
associated with a) substance use behaviors and b) feelings of social and physical connection.
Third, we evaluate how changes in feelings of social connection related to social distancing are
associated with a) changes in drug use behaviors, and b) whether this relationship varies by
severity of use. Fourth, we assess how specific motivations to use substances were associated
with whether use increased or decreased with the onset of social distancing guidelines. Last, we
present an exploratory analysis of the association between drug of choice and changes in feelings
of social connection (this analysis is exploratory due to the small sample size within specific
subsamples of participants).
Results
Sample Characteristics
A total of 795 individuals completed the DAST-10 screening. About half of the sample
50
screened in by scoring exactly a two on the DAST-10; the other half scored mostly a three, with
fewer participants scoring a four through a ten. The 220 participants who scored a 2 or above
were invited to take the full survey. Among those invited, 157 participants completed the study.
Among those who participated, 126 participants reported current drug use, and 31 reported being
currently in sobriety or belonging to a recovery community (i.e., not currently using substances).
Sample characteristics are reported in Table 1 below.
51
Table 1
Sample demographic characteristics
52
Global Feelings of Social Connection
Drug of Choice. Participants who reported current substance use (n = 126) were asked to
select their primary drug of choice (DOC) from the following: Alcohol (n = 56), Cocaine (n = 0),
Marijuana (n = 52), Methamphetamine (n = 2), Opioids (n = 3), and Other (n = 13; participants
wrote in answers including kratom, nicotine, ecstasy, caffeine, LSD, and mushrooms for the
“other” category). With the Social Connectedness Scale (higher scores represent feeling more
connected), we replicated findings of good reliability statistics (α = 0.94; Lee & Robbins, 1995).
An ANOVA examining the association between DOC and scores on the Social Connectedness
Scale indicated that not every group was equal, and a post-hoc Tukey test determined that there
is a significant pair-wise difference between people who use alcohol and those who use opioids;
people who use alcohol had a mean score of 79.95 on the SCS, whereas people who use opioids
had a mean score of 41.5 (see Figure 1). This result should be interpreted with caution, as only
three people in the sample identified opioids as their primary drug of choice.
Figure 1. ANOVA comparing the Social Connectedness Scale scores for those whose primary drug of
choice is alcohol, marijuana, opioids, and other. Black dot displays the mean, lines display the standard
deviation. There is a significant difference between those who use alcohol compared with those who use
opioids in terms of global feelings of social connection. Drugs with fewer than 3 respondents were
collapsed into the “Other” column, including methamphetamine (n =2).
53
Severity of Drug Use
We evaluated whether global feelings of social connection during social distancing are
related to severity of drug use among those currently using drugs (n = 126). We dichotomized
DAST-10 responses: individuals with a score of 3 or higher are categorized as “higher drug use
problems” and those who scored a 2 are categorized as “lower drug use problems.” We report no
statistically significant difference between the two groups in feelings of social connection
assessed by using the SCS.
Impacts of Social Distancing
Changes in Substance Use. Among those who reported smoking cigarettes pre-social
distancing (n = 31), individuals reported smoking 2.7 more cigarettes per day during the
implementation of social distancing (p < 0.05; t = −2.61, df = 30, p-value = 0.014; mean of
differences = −2.65, 95% CI [−4.71, −0.58].
Similarly, among those who reported drinking pre-social distancing (n = 103), individuals
reported drinking on 1.7 more days per 14-day period during social distancing compared with
their drinking pre-social distancing (p < 0.001; t = −4.45, df = 102, p-value = 2.20; mean of
differences = −1.65; 95% CI [−2.39, −0.91]. Among individuals who reported any heavy
drinking days pre-social distancing, there was a trend in the direction of more heavy drinking
days during social distancing, (p <.05; t = −1.73, df = 63, p-value = 0.09; mean of the differences
= −0.66; 95% CI [−1.41, 0.10]).
In contrast, those who used cocaine pre-social distancing (n = 12) reported less cocaine
use during social distancing. These individuals reported using cocaine 1.2 fewer days out of a
54
typical 14-day period during implementation of social distancing guidelines (p < 0.05; t = 2.65, df
= 11, p-value = 0.02; mean of the differences = 1.16, 95% CI [0.20, 2.14].
We found no statistically significant results for the impact of social distancing measures
on differences of marijuana use (p = 0.61; t = −0.51, df = 86, 95% CI [−0.80, 0.47]), opioid use
(p = 0.94; t = −0.08, df = 14, 95% CI [-1.94 1.81], vape use (p = 0.18; t = −1.36, df = 28, 95% CI
[−0.24, 0.05]), and methamphetamine use (p = 0.68; t = −0.43, df = 7, 95% CI [−3.25, 2.25]).
Additionally, we observed no significant differences between changes in cannabis use among
those who reside in states where cannabis is recreationally legal compared with states where
recreational use is illegal (p = 0.74; t = −0.33, df = 85, 95% CI [−1.05, 0.75]).
Changes in Feelings of Social Connection. There was a significant increase in both
physical and social isolation (low social connection) among the entire sample (n = 157) during
social distancing. For social connection, the average response differed from a score of 3.68 to a
score of 2.72 out of 6, or a reduction of 0.96 units on the Likert-scale (p < 0.001; 95% CI [0.75,
1.17]). For physical connection, the average response differed from 3.68 to 2.30, or a reduction
of 1.38 units on the Likert-scale (p < 0.001; 95% CI [1.16, 1.61]). This effect is similar when we
look only at people currently using substances, excluding those who are currently not using any
(n = 126).
Factors Associated With Changes in Feelings of Social Connection During Social Distancing
Changes in Drug Use. Cocaine use was marginally associated with change in social
connection; those reporting smaller reductions in social connection during social distancing
reported a lower decrease in cocaine use relative to those reporting larger reductions in social
connection (i.e., became relatively more socially isolated) with the onset of social distancing
(r = 0.62, p = 0.055, df = 8, 95% CI [−0.01, 0.90]. Those who became relatively more isolated
55
decreased their cocaine consumption more than those whose level of social connection did not
change as dramatically. This correlation was among those who reported any cocaine
use before the onset of social distancing guidelines. There were no statistically significant
differences between change in drug use and change in feelings of social or physical isolation for
any other drug category.
Severity of Drug Use. Figure 2 depicts a two-way ANOVA; there is a significant main
effect of social distancing measures (pre vs. during; F-value = 35.0, p < 0.001), and a significant
main effect of DAST score (high vs. low; F-value = 4.25, p < 0.05), but no interaction effect (F-
value = 0.59, p > 0.05). We then assessed whether social connection was significantly different
during social distancing between those scoring high and low on the DAST. The variance in the
groups was not equal, so a Wilcoxon-Mann-Whitney test was used. Those with more severe drug
use problems were more likely to report lower feelings of social connection during social
distancing (mean = 2.48) than those with less severe drug use problems (mean = 2.95; p < 0.05).
56
Figure 2. An ANOVA showing how self-report feelings of social connection differ when participants
report about connection/isolation before and after the implementation of social distancing measures, as
well as how feelings of connection vary for those who have low or high problems associated with their
drug use. There is no significant interaction.
Motivations for drug use
Changes in Drug Use. The Functions of Substance Use Scale was used to assess motives
to engage in drug use. The scale is divided into five subscales: changing mood, physical effects,
social purposes, to facilitate activity, or to manage effects from other substances. Confirmatory
factor analysis was consistent with the five-subscale structure. We carried out two exploratory
logistic regressions to examine the association between the functions of substance use scale and
changes in drug use. Specifically, one logistic regression assessed whether Function of Substance
Use subscale scores (z-transformed) were predictive of participants who reported increased drug
use during social distancing (n = 57), and a second logistic regression to assess whether the same
subscales predicted which participants reported decreasing drug use during social distancing
57
(n = 23). No subscale scores significantly predicted which participants would increase their drug
use during social distancing. However, higher scores on the social reasons subscale did predict
reduced chance of reporting a decrease in use: for every unit increase in drug use for social
reasons, the log odds of reporting a decrease in substance use decreased by 0.72 (p < 0.05, 95%
CI [−1.42, −0.06]). There were no significant differences among the other four subscales.
Exploratory Analysis
Is Drug of Choice Associated With Changes in Social Connection During Social
Distancing? In addition to the global Social Connection Scale measure, participants answered a
single-item question assessing their feelings of social connection both 1) before, and 2) during
the implementation of social distancing measures. This allows for evaluation of how social
distancing measures impacted within-person feelings of social connection, rather than relying
solely on the SCS. A t-test of the change scores between feelings of isolation pre- and during
social distancing shows that those who prefer using alcohol and those who prefer using
opioids—based on their stated DOC—are significantly different; people using alcohol-DOC are
more likely to report feeling a decrease in social connection during social distancing (relative to
pre-social distancing; mean change score = − 1.02), whereas people using opioids-DOC report
no differences in feelings of social isolation pre- and during social distancing (mean change
score = 0.00; p < 0.01; see Figure 3). There may have been a floor effect with people who use
opioids exhibiting lower social connection even before social distancing guidelines were
implemented. We did not observe a significant effect of severity of problems associated with
drug use between those using alcohol and those using opioids, although there was a trend in the
direction of people who use opioids reporting more problems associated with drug use compared
to people who use alcohol (p = 0.06; t = 2.05, df = 12, 95%CI [-0.13, 4.0]). We report no
58
statistically significant differences in feelings of social connection between people who use
opioids and those who use alcohol either pre- or during social distancing; the only significant
difference is for change scores. This is a preliminary, exploratory analysis, as the small sample
size is an insufficient basis for reliable statistical inference.
Figure 3. Alcohol and opioid users are significantly different in their change scores of social connection
pre and during social distancing. There is no significant difference of reported isolation either in pre or
during social distancing, but the change in pre to post is steeper for those who primarily use alcohol,
while there is no significant difference among those using opioids.
Discussion
The current study explored the impact of social distancing guidelines due to COVID-19
on substance use behaviors. We examined changes in social connection and physical isolation as
a result of social distancing protocols in conjunction with substance use behaviors, including the
use of nicotine, alcohol, marijuana, and illicit substances. Participants reported increased feelings
59
of both physical and social isolation during the implementation of social distancing measures.
This was found to be true across the entire sample of people who currently and formerly used
drugs.
Several changes in drug use behaviors were reported upon the implementation of social
distancing guidelines. The sample reported increased cigarette smoking and alcohol consumption
(both in number of drinking days as well as number of binge-drinking episodes). Interestingly,
cocaine use in our sample decreased with the implementation of social distancing guidelines.
However, it is important to note that cocaine was not the primary drug of choice for any
participant in this sample, and thus this finding may not generalize to individuals who primarily
use cocaine. These changes in substance use during social distancing were found to be unrelated
to higher self-reported feelings of social connection, although sample sizes were modest for
correlational analyses within particular drug categories.
We used the SCS as a trait measure of social connection; the SCS has an item range of
1 − 6 with a total score range of 20 − 120. In previous samples, average reported SCS scores are
typically higher than the scores found in our sample: SCS score of 88.02 (SD = 16.82) in
undergraduate students (Lee & Robbins, 1995); SCS score of 89.84 (SD = 15.44) in another
population of undergraduate students (Lee & Robbins, 1995); and item averages at 4.66 (SD =
1.12) in an international student population (Yeh & Inose, 2003). The average SCS score in our
sample was 76.75 (SD = 19.10), and the item average score reported was 3.84 (SD = .95). While
this is a highly imperfect comparison between college student samples and the current sample of
adults with a minimum of a low level of problems associated with substance use during a
pandemic, it is worthwhile to note that SCS scores are different in our sample than in the sample
with which the SCS measure was developed and validated. There were no differences in SCS
60
scores between people currently using drugs and those who formerly engaged in substance use.
We did not find a significant correlation between SCS score and change in substance use for any
drug class.
There are several alternative explanations that could account for differences between
drug classes in regards to how substance use changed during social distancing. The availability
of certain substances has changed. For example, nicotine and alcohol are available and legal in
every state, while the availability of substances like heroin has been reduced dramatically with
recent restrictions on international trade (Coronavirus Effects on Global Drug Traffic, Mexican
Cartels - Los Angeles Times, 2020), and our data shows that people increased consumption
during social distancing specifically for substances that are more accessible. Alcohol sales during
the third week of March 2020 (immediately after the implementation of social distancing
measures in the US) were up 55% compared with the same week in March 2019 (Valinsky,
2020). The economic impact of the pandemic may have also contributed to changes in social
connection and substance use patterns, as recession-based hardships have been associated with
increased substance use and mental health struggles. Tens of millions of people lost their jobs as
a result of the pandemic, including almost 20% of our sample, with unemployment rates at the
time the survey was completed reaching levels comparable to those of the Great Depression
(Kochhar, 2020). However, in our sample, feelings of social connection during COVID-19 and
global feelings of social connection measured via the SCS were unrelated to loss of employment.
Additionally, many individuals have engaged in social distancing and are staying at home with
family members, which may influence substance use due to varying levels of stigma ascribed to
different substances. These environmental changes may have interacted with decreased social
connection to influence substance use behaviors. Relatedly, we found that individuals who
61
reported using drugs for social reasons were less likely to have reduced substance use during
social distancing, pointing to the importance of addressing motives and reasons for substance
use. Substance use patterns, and their relationship with social and cultural norms, may vary
depending on why the individual is using the substance. These results highlight the importance of
assessing social connection and motivations for substance use.
Of note, two participants who selected the “other” category for primary drug of choice
reported using kratom – a tropical tree used for centuries in Africa and Southeast Asia for its
medicinal properties, which is now gaining attention from regulatory bodies due to its potential
for adverse health outcomes. In the U.S., it is currently touted by some for use as a harm
reduction tool due to its opioid-like effects and potential to reduce withdrawal and cravings for
those with an opioid use disorder. It is possible that some individuals may have replaced opioids
with kratom during COVID-19 if their usual supply of illicit opioids became more difficult to
obtain. For a brief history on the use kratom in the West and its pharmacological properties, see
the following articles (Smith & Lawson, 2017; Veltri & Grundmann, 2019; Warner et al., 2016).
Communities and individuals who experience a sense of belongingness or connection
with their family, social circles, and broader communities are at a lower risk of overdose (Zoorob
& Salemi, 2017). With the onset of social distancing measures in March 2020, the U.S. saw an
alarming increase in the rate of opioid overdoses, pointing to the need for public health reform
(Despite the Challenges, We Must Fight Harder to Address the Nation’s Opioid Epidemic,
2020). Several changes have already been made during the COVID-19 pandemic in order to
increase access to medically assisted treatment and harm reduction programs—including waiving
the need for patients hoping to initiate buprenorphine maintenance to schedule an in-person visit
with a physician prior to receiving medication (C. S. Davis & Samuels, 2020). These kinds of
62
legal and policy changes around access to treatment have the potential to have a positive long-
lasting impact, as policy-makers and researchers advocate for these changes be maintained even
after “normal” in-person activities resume. Enhancing social connection is one piece of a larger
picture that has the potential to change patterns of substance use.
Limitations
There are several important limitations to note in the present study. First, there is a small
sample size within drug subgroups, we note specifically that only three participants specified
opioids as their drug of choice. The comparisons made between drug groups serve as a
suggestion for a future research direction that can incorporate a larger sample size. Second, these
data are a cross-sectional snapshot of individuals’ self-reported substance use and feelings of
social and physical connection both pre-social distancing, and at the time of the survey during-
social distancing; thus, directionality cannot be inferred. However, this study also has several
strengths, including the timing of data collection which occurred when there was a national call
for social distancing across the U.S. Additionally, participants were screened in based on a prior
history of problems related to substance use, ensuring that we could make comparisons about
severity across individuals who use different substances.
Conclusion
We find that social distancing guidelines are, perhaps not surprisingly, associated with
increased feelings of social and physical isolation in this sample. Concurrently, we report
increases in cigarette consumption, number of overall drinking days and heavy drinking days,
and a decrease in cocaine consumption. Overall, substance use patterns and feelings of social
connection have changed during COVID-19. Similar studies have been conducted globally
during the pandemic to assess how substance use behaviors have changed. It is important to view
63
this work in the context of this research. At the time of this report, the global pandemic has been
associated with increases in anxiety, depression, substance use, and suicidal ideation (Czeisler et
al., 2020; Panchal et al., 2021). Additionally, in the U.S., higher levels of loneliness has been
linked to social distancing guidelines (Killgore et al., 2020). Future work should expand upon the
preliminary analyses regarding differences in feelings of social connection between those who
primarily use alcohol and those who primarily use opioids. Additionally, future work should
consider social motivations for use, as well as overall feelings of social connection or isolation,
as these have a demonstrated impact on substance use behaviors.
64
Chapter 3: The Perceived Role of Social Connection in Relapse Risk and Relapse
Prevention
Abstract
How do people view emotionally significant life events as impacting relapse risk?
Investigating perceived risk is important for at least two reasons. First, individuals may be
insightful about their recovery – identifying perceived risk can inform investigation of actual
risk. Second, right or wrong, perceived relapse risk may guide behavior. We examined relapse
risk perception using vignettes depicting emotionally significant events (4 positive-social, 4
negative-social, 4 positive-nonsocial, and 4 negative-nonsocial). We recruited 265 people with a
history of problem substance use (with alcohol, marijuana, methamphetamine, or opioids as their
drug of choice) and 291 participants with no reported history of problem substance use. Positive
social (relative to positive non-social) events were perceived to be more protective against
relapse (t = 8.37, p < 0.001) and negative social (relative to negative non-social) events were
perceived to be riskier (t = 12.33, p < 0.001). People generally expected positive events would
reduce relapse risk. However, those with a history of problem opioid use (t = -4.87, p < .001) or
methamphetamine use (t = -3.84, p < 0.001) were more likely to view at least one positive
nonsocial event as increasing risk than those with no history of problem use. People with a
history of problem opioid use were also more likely to report a perceived increased in risk of
relapse following a positive social event than those with no history of problem use (-3.13, p <
0.01). These findings provide evidence that emotionally evocative social events are perceived to
be particularly powerful risk and protective factors for those in recovery, and highlight that
positive life experiences are not universally perceived to confer a protective effect against
relapse risk.
65
Introduction
Substance Use and Relapse
Relapse is a common experience for individuals in recovery from a substance use
disorder: between 40 – 60 % of people return to substance use post-treatment, and 30% of people
who initiate treatment drop out before the treatment is complete (Andersson et al., 2019;
Brandon et al., 2007; Lappan et al., 2020). People are even more likely to drop out of treatment if
they do not experience positive social interactions with their peers, therapists, and staff in the
treatment center (Nordfjærn et al., 2010). The term relapse is included in the National Institute
on Drug Abuse definition of addiction as a “chronic, relapsing disorder” (National Institute on
Drug Abuse). Various factors influence relapse risk for people who are in remission from a
substance use disorder. Environmental factors such as high levels of neighborhood-level violent
crime increase the likelihood that someone will return to substance use post-treatment (J. P.
Davis et al., 2022), while access to buprenorphine and methadone provide a protective effect
against relapse risk (Clark et al., 2015). Individual factors such as having positive social
relationships, high self-efficacy, and meaning in life provide protection against relapse risk
(Pettersen et al., 2019; Sliedrecht et al., 2019; Zaidi, 2020), whereas factors like younger age, co-
morbid psychiatric conditions, and symptom severity increase relapse risk (J. P. Davis et al.,
2022; Sliedrecht et al., 2019).
While relapse is a common feature across substances, people in recovery from an opioid
use disorder are even more likely to carry compounding risk factors: they are more likely to be
unemployed and have a lower socioeconomic status than those who are in recovery from other
substances – both of which increase the risk of relapse (Christie, 2021; Kadam et al., 2017). The
Brain Opioid Theory of Social Attachment (BOTSA) provides a possible explanation for this
66
difference – the endogenous opioid system is directly related to rewarding properties of social
ties and relationships (Machin & Dunbar, 2011; Panksepp et al., 1980). It is important to
evaluate the subjective experience of social connection and social isolation among people in
recovery, and specifically among people in recovery from an opioid use disorder.
Perceived Risk of Relapse
Perceptions of relapse risk are important to understand for two reasons. First, they are
associated with actual relapse risk (Walton et al., 2000). In a study that assessed mood states and
relapse (both prospectively and retrospectively), people in recovery prospectively reported that
future negative emotional states would likely be strong risk factors for relapse: their perceptions
were accurate. In the follow-up, participants frequently reported a negative mood state prior to a
relapse event – with no evidence to support the hypothesized retrospective bias in reports of
emotional states (Hodgins et al., 1995). In this same study, negative mood states were associated
with a prolonged return to substance use, while social pressures were associated with a brief
return to substance use – indicating that people accurately perceived what types of situations
conferred the most significant relapse risk (Hodgins et al., 1995). In another study, 90% of youth
who have experienced a post-treatment relapse reported that their emotional state was a
contributing factor in their relapse (Gonzales et al., 2012), which is in line with prior work on
clinical models of relapse prevention (Larimer et al., 2003).
The second reason it is important to understand perceived risk is because perceptions
influence behavior; people with different perceived relapse risks differentially engage in relapse
prevention efforts. For example, one common expression in 12-step programs is to “avoid
people, places, and things” that trigger relapse – implying that individuals can identify triggers
and subsequently avoid these perceived triggers. Qualitative work in a population of women who
67
were formerly incarcerated found that this was a commonly endorsed strategy for relapse
prevention (Johnson et al., 2013). Additionally, among those who use opioids, a higher perceived
risk of relapse is associated with higher odds of desiring medication assisted treatment (Bailey et
al., 2013). Given that perception is linked both to actual risk, and to risk mitigation behaviors, it
is important to investigate perceived risk of relapse in order to more effectively prevent relapse.
Currently, little is known about potential differences in perceptions of relapse risk between
the general population and people in recovery. These groups may have different beliefs about
what kind of situations/events increase or decrease the risk of relapse. This difference in
perception may have downstream implications for relapse. When people complete substance use
treatment, they are not isolated from the outside world; they continue to interact with friends,
family, coworkers, and peers, some of whom do not have direct experience with addiction.
Individuals committed to recovery may structure their lives to avoid situations that put them at
risk, and cultivate those that bolster their recovery. Those around them who care may also make
decisions with their loved-one’s recovery in mind. If there are systematic differences in how
those with vs. without addiction experience understand relapse risk, it may lead to discord. If, as
seems plausible, individuals in recovery have a better understanding of relapse risk than those
who have no history of addiction, then the divergence may imply areas in which opportunities
for loved ones to be supportive are missed. We know of no work to date comparing the perceived
risk of relapse following emotionally significant life events among those in recovery from
problem substance use and those who have no history of problem use.
The present study aims to address this gap in the literature and identify the perceived risk and
protective effect of various emotionally evocative events, with a comparative analysis between
different substances (i.e., alcohol, marijuana, methamphetamine, and opioids) and different
68
populations (i.e., people with a history of problem substance use, and those with no history of
problematic substance use). Our study design and hypotheses were pre-registered via the Open
Science Foundation. Full results for any hypotheses and planned analyses that are not presented
in the main text are included in the Appendix. Any results/analyses in the present work that were
not pre-registered are described as exploratory.
Method
Sample Selection
The present study used a convenience sample; we collected data from adults who reside
in the United States and who have an account with Prolific Academic Ltd., an online data
collection platform (Prolific.co) that has good transparency, functionality, and a relatively high
minimum hourly payment for participants (Palan & Schitter, 2018). We invited adults who use
Prolific to complete a screening survey to describe their substance use behaviors, including their
drug of choice and the severity of their substance use. We used the screening survey data to
separate the selected participants into two groups: individuals with a history of problematic
substance use (Hx-PU) and individuals with no prior history of problematic substance use (No
Hx-PU). Individuals screened into the full study as part of the history of problematic substance
use group if they had (a) a self-reported history of problems associated with drug use, (b)
currently do not engage in problematic substance use (i.e., are not using any substances, or use
occasionally/casually with no problems), (c) scored a 3 or above on the DAST-10 screener (a
clinical measure of severity of use), (d) indicated that their substance of choice was either
alcohol, marijuana, methamphetamine, or opioids, and (e) were adults currently residing in the
United States who speak English and have Prolific accounts. Individuals screened into the full
study for the no prior history of problematic use group if they had (a) no self-reported history of
69
problems associated with drug use, (b) currently do not engage in problematic substance use (i.e.,
are not using any substances, or use occasionally/casually with no problems), and (c) were adults
currently residing in the United States who speak English and have Prolific accounts. We
oversampled those who used opioids and methamphetamine to ensure that we had an adequate
sample size to statistically compare between groups. We oversampled by selectively inviting
those who indicated that their primary drug of choice was opioids or methamphetamine in the
screener to complete the full survey until we had ~50 participants in each category for drug of
choice. Participants completed the full study between March 20, 2022 and December 22, 2022.
Measures
We used a novel measure – the Narrative-Based Inventory of Relapse Risk (NIRR) – that
is composed of vignettes that depict unique experiences of protagonists in recovery. Using the
NIRR, we evaluate a) strength of perceived emotional responses to hypothetical positive and
negative life events, and b) perceived reduction or increase in the risk of relapse following those
events. Participants were asked to read 16 vignettes that are split into four categories with four
unique vignettes in each category: 1) Social Positive (e.g., reconnecting with a childhood friend,
getting engaged), 2) Social Negative (e.g., being rejected by a romantic partner, losing an
important friendship), 3) Non-Social Positive (e.g., buying a house, getting a promotion), and 4)
Non-Social Negative (e.g., major setback in starting your own business, being in danger of
eviction). Within the social vignettes, two scenarios focus on romantic partner relationships and
two focus on familial/friend relationships. Within the non-social vignettes, two focus on work-
related scenarios, and two focus on other life scenarios (e.g., legal issues, housing situations).
Each participant read all 16 vignettes. Because perceived relapse risk may not be identical across
different substances, the participants in the No Hx-PU group were randomly assigned to read the
70
vignettes about people in recovery from one of the following substances: alcohol, marijuana,
methamphetamine, or opioids. We chose these substances in order to compare “protagonist drug
of choice” for people in the Hx-PU group (who self-reported their own prior drug of choice,
which was matched to the protagonists’ drug of choice in the vignettes) to people who have no
history of problematic substance (who were randomly assigned to read about the protagonists’
drug of choice).
For participants with No Hx-PU, the instructions read as:
In the following brief stories, please imagine that the person described has had moderate
problems with using [Substance X]. Imagine that this person has been able to quit using
it for the past few months, though not without some continued struggle.
Please evaluate the degree to which you think the situations described would put them at
any higher or lower risk for relapse. Of course, everyone is different, so just give us your
“gut feeling” based on the little information you are given.
In the instructions, “[Substance X]” was randomly assigned as either alcohol, marijuana,
methamphetamine, or opioids.
For participants with a Hx-PU, the instructions read as:
In the following brief stories, please imagine that the person described had similar levels
of problems with [Substance X] as you had. Imagine that this person has been able to
quit using it for the past few months, though not without some continued struggle.
Please evaluate the degree to which you think the situations described would put them at
any higher or lower risk for relapse. Of course, everyone is different, so just give us your
“gut feeling” based on the little information you are given.
In the instructions, “[Substance X]” - alcohol, marijuana, methamphetamine, or opioids -
was based on the participant’s self-reported DOC from the screener, such that their own
identified drug of choice matched the protagonists’ drug of choice in the survey.
Participants were first instructed to evaluate the intensity of predicted emotional
responses for each of these 16 vignettes, which included both positive and negative scenarios.
They responded on an 11-point unipolar scale ranging from a “not strong at all” emotional
71
response (1) to an “extremely strong” emotional response (11). Participants were then asked to
evaluate their perception of the risk of relapse for the protagonist in the vignette. They responded
on an 11-point bipolar scale ranging from a significant decrease in the risk of relapse (1) to a
significant increase in the risk of relapse (11). The midpoint of the scale is 6: scores above 6
indicate a perceived increased risk of relapse, scores below a 6 indicate a perceived protective
effect against relapse, and a score of 6 indicates no perceived impact on the risk of relapse. All
study measures were approved by the local Institutional Review Board. All statistical analyses
were completed using R version 4.4.2.
Analytic Procedures
Demographic Differences. We used chi-square tests and ANOVA to evaluate if there
were any significant differences in demographic characteristics between those with a Hx-PU, and
those with No Hx-PU.
Psychometric Properties of the NIRR. As noted above, we used a novel assessment to
evaluate the perceived risk of relapse following emotionally significant life events. We
conducted an Exploratory Factor Analysis (EFA) with target rotation (using our hypothesized 4
subscale structure) to evaluate the factor structure of the 16-item measure (Costello & Osborne,
2019; Fabrigar et al., 1999).
Model Building and Covariate Selection. Before the full analyses were conducted, we
first determined which demographic covariates to include in the final models. We ran a stepwise
regression model with forward and backward selection to identify demographic covariates that
may predict NIRR responses.
Perceived Risk of Relapse by Valence, Sociality, and Drug of Choice. We used
ANCOVA to predict judged relapse risk with a 3-way interaction between a) protagonist drug of
72
choice b) valence (positive/negative), and c) sociality (social/nonsocial) and d) a covariate term:
magnitude of the emotional response to the vignette (to account for strength of emotional
response to the depicted scenario). Any significant ANCOVAs were followed by subgroup
analyses.
Perceived Risk of Relapse by Drug of Choice and Sociality (Among those with A
History of Problem Use). We used ANCOVA to determine if there was a two-way interaction
between drug of choice and sociality (social/nonsocial) when predicting perceived relapse risk
among the population of people with a Hx-PU. In our pre-registered analytic plan, we had stated
that we would use ANOVA to determine if there were interactions between the two groups (Hx-
PU / No Hx-PU) across the four levels of the NIRR vignettes (positive/negative, and
social/nonsocial). We also stated that we would conduct a t-test of the overall global scores (with
negative items reverse coded) on the NIRR to assess differences in responses to the risk of
relapse and the rated emotional response to the scenarios among these two groups, as well as
four additional t-tests to determine if there are differences across each of the four types of
vignettes between the groups. These analyses from the pre-registration are reported in the
Appendix.
While we anticipated a uniform association between valence and risk (negative scenarios
perceived as higher risk, positive scenarios perceived as decreased risk), the initial analyses
indicated that that was not uniformly true, as discussed below. So, we separately model positive
and negative scenarios for the final analyses. We tested two models: 1) an ANCOVA model
predicting relapse risk for the eight positive NIRR vignettes with an interaction between
protagonist drug of choice*sociality among the group with no history of problem use, with
magnitude of emotional response as a covariate, and 2) an identical model for the eight negative
73
NIRR vignettes. To improve the interpretability of between-model comparisons, we reverse
coded the positive items, such that higher scores indicated a higher protective effect against
relapse (to match the negative items, for which a higher score indicated a higher risk of relapse).
Perceived Increased Risk of Relapse Following Positive Events. The following is an
exploratory analysis (i.e., not included in the pre-registration). We used two logistic regression
models to predict whether people perceived any positive a) social or b) nonsocial events as
increasing the risk of relapse using a protagonist drug of choice*group interaction as our
predictor. We created two binary outcomes, one for positive social events, and the second one for
positive nonsocial events. For both outcomes, responses were coded as a 0 if all positive events
were rated as protective (i.e., a 6 or above on the reverse coded 1-11 bipolar scale), or as a 1 if at
least one positive event was rated as increasing relapse risk (i.e., a 5 or below on the reverse
coded 1-11 bipolar scale). Because the majority of participants rated all positive events as
protective, the distribution of perceived risk from the positive scenarios was dichotomized
(separately for social and non-social scenario groups) to classify participants as those who
viewed at least one positive event as risk-enhancing vs. those that did not. This allowed data to
be analyzed using logistic models.
Results
Demographic Characteristics of the Sample
There are a total of 443 participants in the present study, 216 of whom self-identify as being in
recovery from problem substance use. Full details of the demographic makeup of the sample,
including differences between groups, are presented in Table 1.
74
Table 1
Demographic characteristics of the sample, including a column on the far right indicating any
significant differences between the Hx-PU and No Hx-PU groups
Exploratory Factor Analysis
First, we evaluated whether our data was suited for an exploratory factor analysis. We ran
a Kaiser-Meyer-Olkin (KMO) test to measure whether we had sampling adequacy for each
variable in the model: the KMO value was 0.88, indicating that our data is factorable. Next, we
used Bartlett’s test of Sphericity to determine whether the variables in the correlation matrix are
correlated (null hypothesis is that the variables are uncorrelated). The chi-square value is
3024.98 (p < 0.001). Both tests indicate that factor analysis is appropriate for our data. Using the
75
factor analysis function in R package psych, with a principal axis factoring method and target
rotation, we report the following factor loadings for the hypothesized 4-factor solution.
The model had 437 observations and 62 degrees of freedom. The chi-square test of model
fit was not statistically significant, χ²(62) = 69.46, p > .24. The root mean square of the residuals
(RMSA) was .02, and the df corrected root mean square of the residuals was .03. The Tucker
Lewis Index of factoring reliability was .993, suggesting good reliability. The RMSEA index
was .016, with a 10% confidence interval ranging from 0 to .034. The Bayesian Information
Criterion (BIC) value was -307.5.These findings indicate a good model fit to the data.
Table 2
Correlation Table Between Latent Factors
Nonsocial Positive Social Positive Nonsocial Negative Social Negative
Nonsocial Positive 1.00 0.67 0.12 0.07
Social Positive 0.67 1.00 -0.14 -0.15
Nonsocial Negative 0.12 -0.14 1.00 0.39
Social Negative 0.07 -0.15 0.39 1.00
76
Figure 1: Factor analysis showing factor loadings for each item on its primary factor, as well as
correlations between the latent factors. In the figure, “PA” indicates a factor obtained via the principal
axis factoring method.
Model Building and Covariate Selection
The stepwise regression model selected no demographic characteristics as significant
predictors, with an Adjusted R-squared value of 0.004 and an AIC value of 2383.8 (relative to an
AIC value of 2404.85 with all covariates included).
Predicting Perceived Relapse Risk by Drug of Choice, Valence, and Sociality
We used ANCOVA to determine whether the interaction between protagonist drug of
choice, sociality (social vs. nonsocial), and valence (positive/negative) significantly predicted
perceived relapse risk as hypothesized; see Table 2. There is a significant main effect of
protagonist drug of choice (F(3, 1732) = 3.77), a significant main effect of valence (F(1,1732) =
5321.12), and no main effect of sociality (F(1, 1732) = 1.04). There is a significant two-way
interaction between Protagonists’ DOC and Sociality (F(1,1732) = 169.38). There is no evidence
77
of the predicted three-way interaction effect between DOC, valence, and sociality. For the main
effect of protagonist drug of choice, participants rated the global risk of relapse (across
sociality/valence categories) as slightly higher for alcohol (6.03 + 0.89) than methamphetamine
(5.72 + 0.78). For the main effect of valence, positive events were perceived as being protective
against relapse (i.e., below a 6 on the 11-point bipolar scale; 3.26 + 1.70), and negative events
were perceived to increase risk of relapse (i.e., above a 6 on the 11-point bipolar scale; 8.52 +
1.45).
Table 2
Relapse Risk Score Predicted From A Three-Way Interaction Between Protagonist Drug Of Choice,
Valence, And Sociality
Df Sum Sq Mean Sq F Value Pr(>F)
Protagonist Drug of Choice 3 26 9 3.77 0.01
Valence 1 12107 12107 5321.12 <0.001
Sociality 1 2 2 1.04 0.31
Protagonist Drug of Choice
X
Valence
3 3 1 0.42 0.74
Protagonist Drug of Choice
X
Sociality
3 4 1 0.65 0.58
Valence
X
Sociality
1 385 385 169.38 <0.001
Protagonist Drug of Choice
X
Valence
X
Sociality
3 1 0 0.09 0.96
Residuals 1732 3941 2 NA NA
78
Protective Effect Against Relapse Following Positive Life Events
The following analyses separately modeled responses to positively valenced items, and
negatively valenced items. To increase ease of interpreting visualizations, ratings for relapse risk
were reverse coded so that higher numbers now indicate an increased protective effect against
relapse. When predicting the protective effect against relapse following the experience of
positive events, there is a significant main effect of Protagonist’s Drug of Choice (which
matched participants’ former drug of choice for the Hx-PU group; F(3, 421) = 2.69, p < 0.05);
people perceived a lower protective effect from positive events for protagonists who were in
recovery from opioid use. There is also a significant main effect of the magnitude of the
emotional response (F(1,421) = 44.26, p < 0.001), such that for each unit increase in emotional
response, there is a 0.34 higher perceived protective effect against relapse (p < 0.001). Finally,
there is a significant main effect of sociality (F(1,421) = 39.40, p < 0.001); people reported a
higher perceived protective effect against relapse from positive social events compared to
positive nonsocial events, even when including the strength of the emotional response in the
model (mean difference = 0.99, 95% CI[0.65, 1.31], p <0.001). See Figure 2.
79
Figure 2: Marginal means of perceived protective effect of positive life events (controlling for the
perceived magnitude of the positive emotion induced by the event). X-axis indicates the Protagonist Drug
of Choice for the protagonist in the vignette; colors indicate sociality of the event depicted in the NIRR
vignette; circles indicate the predicted marginal means of the perceived protective effect in the model,
error bars indicate the confidence interval.
Table 3
Perceived Protective Effect Against Relapse Following Positive Events Predicted By A Two-Way
Interaction Between Protagonist Drug Of Choice, And Sociality With Emotional Response As A
Covariate
Df Sum Sq Mean Sq F Value Pr(>F)
Protagonist Drug of Choice 3 21.62 7.21 2.69 0.046
Sociality 1 105.80 105.80 39.40 <0.001
Emotional Response 1 44.26 44.26 16.49 <0.001
Protagonist Drug of Choice
X
Sociality
3 2.25 0.75 0.28 0.84
Residuals 421 1130 2.684 NA NA
Increased Risk of Relapse Following Negative Life Events
80
When predicting relapse risk following the experience of a negative life event, there is a
significant main effect of the magnitude of the emotional response (F(1,421) = 67.85, p < 0.001);
such that for each unit increase in emotional response, there is a 0.55 higher perceived risk of
relapse (p < 0.001). See Table 3. There is also a significant main effect of sociality: people
reported a higher perceived risk of relapse from negative social events compared to negative
nonsocial events (F(1,421) = 80.79, p < 0.001), even when including the strength of the
emotional response in the model (mean difference = 1.07, 95% CI[0.82, 1.32], p <0.001). See
Figure 3.
Figure 3: Marginal means of perceived risk of relapse following the experience of negative life events. X-
axis indicates the Drug of Choice for the protagonist in the vignette; colors indicate sociality of the event
depicted in the NIRR vignette; circles indicate the predicted marginal means of the perceived risk in the
model, error bars indicate the confidence interval.
81
Table 4
Perceived Relapse Risk Following Negative Events Predicted By A Two-Way Interaction Between
Protagonist Drug Of Choice, And Sociality With Emotional Response As A Covariate.
Df Sum Sq Mean Sq F Value Pr(>F)
Protagonist Drug of Choice 3 9.63 3.21 2.10 0.10
Sociality 1 123.30 123.30 80.79 <0.001
Emotional Response 1 103.50 103.50 67.85 <0.001
Protagonist Drug of Choice
X
Sociality
3 4.25 1.42 0.93 0.43
Residuals 421 642.4 1.526 NA NA
Increased Risk of Relapse Following Positive Life Events
Nonsocial Events. In model 1, we evaluated the proportion of people who endorse an
increased risk of relapse following a positive nonsocial event with Group (Hx-Pu vs. No Hx-PU)
and emotional response as predictor variables. There is a significant main effect of both Group (t
= -6.17. p < 0.001) and no main effect of emotional response. People with a history of problem
use were more likely to report a perceived increased risk of relapse as a consequence of one or
more positive nonsocial events than those with no history of use.
In the second model, we added in an interaction term: Group*Protagonist DOC to
determine if the relationship between group and likelihood of reporting an increased risk of
relapse following a positive event varied based on the protagonist DOC. In model 2, there is a
significant main effect of DOC; people (across both groups) were more likely to report an
increased risk from at least one non-social positive event for vignettes where the protagonists’
drug of choice was opioids (t = 5.32, p < 0.001). We report a Group*Protagonist DOC
interaction: those who have No Hx-PU were less likely to report an increased risk of relapse
82
following positive events for scenarios where the protagonists’ drug of choice was opioids (t = -
4.87, p < .001) or methamphetamine (t = -3.84, p < 0.001) than those who with a Hx-PU (where
their DOC matched protagonist’s DOC). See Figure 4. For the group who is in recovery from
problem opioid use (and was reading vignettes about people in recovery from opioid use), 75%
reported an increased risk of relapse from at least one positive non-social event. Fewer than 50%
of people with no history of problem use reported an increased risk of relapse from any single
positive non-social event.
Figure 4: Proportion of people who perceived an increased risk of relapse as a consequence of at least one
positive nonsocial event (of the 4 nonsocial positive events depicted in vignettes). X-axis indicates the
Drug of Choice for the protagonist in the vignette; colors indicate group membership (Hx-PU and No Hx-
PU); dots represent the predicted proportion of people in each group who perceived an increased relapse
risk for at least one positive event, error bars indicate the 95% confidence interval.
83
Table 5
Proportion Of People Who Perceive An Increased Risk Of Relapse Following Positive Nonsocial Events
Predicted By Protagonist Drug Of Choice * Group (Hx-Pu Vs. No Hx-PU) With Emotional Response
Included As A Covariate
Estimate Standard Error T Value Pr(>|t|)
Intercept 0.51 0.12 4.14 <0.001
Marijuana Drug of Choice -0.11 0.04 -2.89 0.004
Methamphetamine Drug of Choice 0.08 0.05 1.43 0.15
Opioids Drug of Choice 0.26 0.05 5.32 <0.001
No Hx-PU Group -0.01 0.04 -0.28 0.78
Emotional Response -0.004 0.01 -0.40 0.69
Marijuana Drug of Choice X No Hx-PU
Group
-0.09 0.06 -1.48 0.14
Methamphetamine Drug of Choice X
No Hx-PU Group
-0.27 0.07 -3.84 <0.001
Opioids Drug of Choice X No Hx-PU
Group
-0.33 0.07 -4.87 <0.001
Social Events. In model 1, we evaluated the proportion of people who endorse an
increased risk of relapse following a positive social event with Group (Hx-Pu vs. No Hx-PU) and
emotional response as predictor variables. There is a significant effect of emotion response (t = -
5.65, p < 0.001) and no main effect of Group. People who viewed events as eliciting a stronger
emotional response were significantly less likely to report a perceived increased risk of relapse in
response to one or more positive social events.
In the second model, we added in an interaction term: Group*Protagonist DOC to
determine if the relationship between group and likelihood of reporting an increased risk of
relapse following a positive event varied based on the protagonist DOC. In model 2, there is a
significant main effect of the strengths of the emotional response when predicting the proportion
of people who indicate an increased risk of relapse following the experience of one or more of
the social positive events; people (across Hx-PU and No Hx-PU groups) were more likely to
84
report an increased risk from at least one non-social positive event if the event was perceived to
produce a weaker emotional response (t = -6.01, p < 0.001). See Table 6. For each unit increase
in emotional response, the log odds of reporting any positive social event as increasing risk
decrease by 0.37 (p < 0.001) . We report a Group*Protagonist DOC interaction: those who have
no history of problem use were less likely to report an increased risk of relapse following
positive events for scenarios about opioid use (t = -3.13, p < .01) than those with a Hx-PU
(where their DOC matched protagonist's DOC). See Figure 5. Overall, fewer than 20% of any
participants reported an increased risk of relapse following any positive social scenario.
Figure 5: Proportion of people who perceived an increased risk of relapse as a consequence of at least one
positive social event (of the 4 social positive events depicted in vignettes). X-axis indicates the Drug of
Choice for the protagonist in the vignette; colors indicate group membership (Hx-PU and No Hx-PU);
dots represent the predicted proportion of people in each group who perceived an increased relapse risk
for at least one positive event, error bars indicate the 95% confidence interval.
85
Table 6
Proportion Of People Who Perceive An Increased Risk Of Relapse Following Positive Social Events
Predicted By Protagonist Drug Of Choice * Group (Hx-Pu Vs. No Hx-PU) With Emotional Response
Included As A Covariate
Estimate Standard Error T Value Pr(>|t|)
Intercept 0.59 0.08 7.23 <0.001
Marijuana Drug of Choice 0.01 0.03 0.48 0.63
Methamphetamine Drug of Choice -0.04 0.04 -1.12 0.27
Opioids Drug of Choice 0.03 0.03 0.93 0.35
No Hx-PU Group 0.02 0.03 0.84 0.40
Emotional Response -0.05 0.01 -6.01 <0.001
Marijuana Drug of Choice: No Hx-
PU Group
0.01 0.04 0.31 0.75
Methamphetamine Drug of Choice:
No Hx-PU Group
-0.05 0.05 -1.03 0.30
Opioids Drug of Choice: No Hx-PU
Group
-0.14 0.05 -3.13 0.002
Discussion
Increasing knowledge about relapse risk (both real and perceived) is important to better
support people in recovery and ultimately to reduce relapse rates. Our hypotheses that negative
events would be perceived as increasing relapse risk, and positive events would be perceived as
decreasing relapse risk was, not surprisingly, supported. Both people with a Hx-PU and those
with No Hx-PU perceived negative events to significantly increase relapse risk, and positive
events to significantly protect against relapse. While most people reported protective effects (i.e.,
reduction in relapse risk) for all four positive events, we evaluated which subsets of participants
were more likely to report an increase in relapse risk from any one of these positive events in our
exploratory analyses.
86
Some people did perceive that the experience of one or more positive life events may put
individuals at an increased risk for experiencing a relapse. For positive nonsocial events (e.g.,
buying a house, getting a promotion), between 25 - 75% of people reported that at least one of
these events could put a person at an increased risk of relapse. The proportion was higher for
people with a history of problematic use (particularly in the group where protagonist’s drug of
choice was opioids or methamphetamine): this group was more likely than those with no history
of problem use to report an increase in relapse risk following the experience of a positive
nonsocial events. For the positive social events, we found a similar pattern of results, but the
difference between the groups (Hx-PU, No Hx-PU) was only significant for the opioid condition.
Additionally, the overall proportion of people reporting an increased risk of relapse following a
positive social event was much lower than the nonsocial events; only 3 – 17% of the sample
perceived that at least one positive social event would increase relapse risk, and those who
perceived the events to be particularly emotionally evocative were more likely to report a
protective effect than a risky effect. In all, this highlights the critical role of positive social
relationships in relapse prevention efforts, and indicates at the same time that not all positive life
events are protective against relapse: people with a history of problem substance use (particularly
with heavily stigmatized substances) reported that some positive non-social events can be
triggering for relapse. Positive life events or changes, such as starting a new job, or achieving a
major goal may create risk for those who are trying to maintain a substance-free lifestyle. Many
people use substances to celebrate life achievements, and these celebratory times may trigger
people to celebrate their accomplishments in a way that they have done before – with drugs. This
is in line with recent evidence that experiencing stronger than usual positive emotions increases
87
enhancement motives to use substances, which may trigger craving and increase the odds of a
lapse or relapse event (Votaw & Witkiewitz, 2021).
While avoiding all negative experiences is an impossible feat, the influence that risk
factors exert is not immutable. Therapeutic models such as Mindfulness-Based Relapse
Prevention are effective in mitigating relapse risk through several mechanisms, one of which is
encouraging people to learn to sit with discomfort and craving through therapeutic practices like
urge surfing (Bowen et al., 2014; J. P. Davis et al., 2019; Roos et al., 2020). Models like this are
well suited to help people avoid unwanted substance use behaviors, as they focus on how deal
with in-the-moment emotions and feelings that accompany everyday life events (rather than
avoiding specific life events altogether). At the same time, seeking out positive social
interactions can serve as an additional layer of support and buffering against risky events
(Havassy et al., 1991; Hunter-Reel et al., 2009). Treatment settings, relapse-prevention
programs, and even health policies should make a concerted effort to integrate the protective role
of peers, family, and social networks in their clinical programming and policies.
Additionally, wide-scale awareness of the importance of social wellbeing in the
prevention of and treatment of substance use disorders could reduce the demonstrated
discrepancy in perception of relapse risk following positive events between people with a history
of problem use, and those with no history of problem use. People who use heavily stigmatized
substances (i.e., illicit opioids and methamphetamine) are more likely to report an increased risk
of relapse following a positive nonsocial life event than those with no history of problem use.
People trying to help loved ones may underestimate the risk associated with positive events in
particular, and thus they may be less effective in helping mitigate relapse risk. For example, if a
parent has a child who is in recovery and they just got a promotion, the parent may incorrectly
88
think that this good thing is sure to be protective, and may fail to provide support at a critical
time. Future work should evaluate under what conditions this difference in risk perception
following positive events may be associated with supportive behaviors among loved ones.
Conclusion
Emotionally evocative social events are perceived to be particularly powerful risk /
protective factors when it comes to relapse prevention. We also identify a difference in perceived
risk of relapse following the experience of positive events by protagonists’ drug of choice
between people with a history of problem use, and those with no history of problem use. People
with history of problem use (particularly opioid or methamphetamine use) are more likely to
view positive events as increasing relapse risk than those who are in recovery from alcohol or
marijuana use. This study evaluated perceived (rather than actual) relapse risk. However, to the
extent that people with direct experience are correct, these data suggest that relapse prevention
measures could benefit from creating environments where people are likely to experience
positive social interactions, and developing coping skills to counteract the negative social
interactions people inevitably encounter.
Conclusions
Throughout this dissertation, I first delineated a theoretical foundation for the relationship
between substance use and social connection – with an emphasis on stigmatized substances,
particularly opioids. In Chapter 1, I provided evidence that psychoactive substances produce
sociodelic feelings during the acute drug high, and that these feelings are endorsed more often by
those who use opioids. Additionally, those who go on to develop a more severe substance use
disorder retrospectively endorse higher levels of sociodelic emotions during the initial stages of
their use – regardless of drug of choice. This is in line with a self-medication hypothesis
89
framework: some people appear to be alleviating feelings of isolation or loneliness through
substance use.
In Chapter 2, I described how social distancing guidelines impacted substance use
behaviors across various drug classes. Social distancing guidelines were associated with
increased feelings of social and physical isolation. People reported increased cigarette
consumption, increased number of overall drinking days and heavy drinking days, and decreased
cocaine consumption. In all, this study points to the prominent role of social motivations and
stress-reduction motivations in substance use behaviors. When the social landscape changed
rapidly and dramatically, we saw associated changes in substance use behaviors.
In Chapter 3, I demonstrate that people perceive emotionally evocative social events to be
particularly powerful risk / protective factors when it comes to relapse risk / prevention. This
study highlights that not all positive events confer a perceived protective effect against relapse:
people with a history of problem use are more likely to report a perceived increase in relapse risk
following positive nonsocial events. It is important to increase awareness that positive life events
and milestones may increase risk of relapse for some people; social supports (which are well-
documented protective factors) should mobilize during these periods, and not only during
periods following a negative life experience.
This work has implications both for research and for clinical practices. There is an urgent
need to incorporate social relationships and social wellbeing into the continuum of care for
substance use disorders from prevention, to acute treatment, and long-term recovery. Over time,
I see this work as the first step to building a foundation for a research track focused on the
critical role of human connection in mental health and wellbeing.
90
References
2017 Drug Overdose Death Rates. (2017). CDC Injury Center.
https://www.cdc.gov/drugoverdose/data/statedeaths/drug-overdose-death-2017.html
Acuff, S. F., MacKillop, J., & Murphy, J. G. (2023). A contextualized reinforcer pathology
approach to addiction. Nature Reviews Psychology, 1–15. https://doi.org/10.1038/s44159-
023-00167-y
Albertín, P., & Íñiguez, L. (2008). Using drugs: The meaning of opiate substances and their
consumption from the consumer perspective. Addiction Research and Theory, 16(5),
434–452. https://doi.org/10.1080/16066350802041455
Alexander, B. K. (2012). Addiction: The Urgent Need for a Paradigm Shift. Substance Use &
Misuse, 47(13–14), 1475–1482. https://doi.org/10.3109/10826084.2012.705681
Alexander, B. K., Beyerstein, B. L., Hadaway, P. F., & Coambs, R. B. (1981). Effect of early
and later colony housing on oral ingestion of morphine in rats. Pharmacology,
Biochemistry and Behavior, 15(4), 571–576. https://doi.org/10.1016/0091-
3057(81)90211-2
Andersson, H. W., Wenaas, M., & Nordfjærn, T. (2019). Relapse after inpatient substance use
treatment: A prospective cohort study among users of illicit substances. Addictive
Behaviors, 90, 222–228. https://doi.org/10.1016/J.ADDBEH.2018.11.008
Bacsi, K. (2022). What Does Meth Feel Like? Understanding How A Meth Addict Thinks and
Feels. The Recovery Village Drug and Alcohol Rehab.
https://www.therecoveryvillage.com/meth-addiction/understanding-meth-addict-thinks-
feels/
Bailey, G. L., Herman, D. S., & Stein, M. D. (2013). Perceived Relapse Risk and Desire for
Medication Assisted Treatment among Persons Seeking Inpatient Opiate Detoxification.
Journal of Substance Abuse Treatment, 45(3), 302–305.
https://doi.org/10.1016/j.jsat.2013.04.002
Barman-Adhikari, A., Rice, E., Winetrobe, H., & Petering, R. (2015). Social Network Correlates
of Methamphetamine, Heroin, and Cocaine Use in a Sociometric Network of Homeless
Youth. Journal of the Society for Social Work and Research, 6(3), 433–457.
https://doi.org/10.1086/682709
91
Bartholomew, K., & Horowitz, L. M. (1991). Attachment Styles Among Young Adults: A Test
of a Four-Category Model. Journal of Personality and Social Psychology, 61(2), 226–
244. https://doi.org/10.1037/0022-3514.61.2.226
Baumeister, R. F. (2010). The Cultural Animal: Human Nature, Meaning, and Social Life. In The
Cultural Animal: Human Nature, Meaning, and Social Life. Oxford University Press.
https://doi.org/10.1093/ACPROF:OSO/9780195167030.001.0001
Becker, W. C., Fiellin, D. A., Merrill, J. O., Schulman, B., Finkelstein, R., Olsen, Y., & Busch,
S. H. (2008). Opioid use disorder in the United States: Insurance status and treatment
access. Drug and Alcohol Dependence, 94(1–3), 207–213.
https://doi.org/10.1016/j.drugalcdep.2007.11.018
Bell, R. M. S., & Malick, J. B. (1976). Enkephalins and Endorphins: A Major Discovery? JAMA:
The Journal of the American Medical Association, 236(25), 2887–2888.
https://doi.org/10.1001/jama.1976.03270260043030
Berridge, K. C., Robinson, T. E., & Aldridge, J. W. (2009). Dissecting components of reward:
‘Liking’, ‘wanting’, and learning. Current Opinion in Pharmacology, 9(1), 65–73.
https://doi.org/10.1016/J.COPH.2008.12.014
Bohn, M. J., Babor, T. F., & Kranzler, H. R. (1991). Validity of the Drug Abuse Screening Test
(DAST-10) in inpatient substance abusers. In Problems of Drug Dependence.
Bohnert, A. S. B., Bradshaw, C. P., & Latkin, C. A. (2009). A social network perspective on
heroin and cocaine use among adults: Evidence of bidirectional influences. Addiction,
104(7), 1210–1218. https://doi.org/10.1111/j.1360-0443.2009.02615.x
Borrow, A. P., & Cameron, N. M. (2012). The role of oxytocin in mating and pregnancy.
Hormones and Behavior, 61(3), 266–276. https://doi.org/10.1016/j.yhbeh.2011.11.001
Bowen, S., Witkiewitz, K., Clifasefi, S. L., Grow, J., Chawla, N., Hsu, S. H., Carroll, H. A.,
Harrop, E., Collins, S. E., Lustyk, M. K., & Larimer, M. E. (2014). Relative Efficacy of
Mindfulness-Based Relapse Prevention, Standard Relapse Prevention, and Treatment as
Usual for Substance Use Disorders: A Randomized Clinical Trial. JAMA Psychiatry,
71(5), 547–556. https://doi.org/10.1001/jamapsychiatry.2013.4546
Bozarth, M. A., Murray, A., & Wise, R. A. (1989). Influence of housing conditions on the
acquisition of intravenous heroin and cocaine self-administration in rats. Pharmacology,
92
Biochemistry and Behavior, 33(4), 903–907. https://doi.org/10.1016/0091-
3057(89)90490-5
Bragard, E., Giorgi, S., Juneau, P., & Curtis, B. L. (2021). Loneliness and Daily Alcohol
Consumption During the COVID-19 Pandemic. Alcohol and Alcoholism (Oxford,
Oxfordshire), 57(2), 198–202. https://doi.org/10.1093/alcalc/agab056
Brandon, T. H., Vidrine, J. I., & Litvin, E. B. (2007). Relapse and Relapse Prevention. Annual
Review of Clinical Psychology, 3(1), 257–284.
https://doi.org/10.1146/annurev.clinpsy.3.022806.091455
Bridges, R. S., & Grimm, C. T. (1982). Reversal of morphine disruption of maternal behavior by
concurrent treatment with the opiate antagonist naloxone. Science, 218(4568), 166–168.
https://doi.org/10.1126/science.7123227
Brody, A. L., Mandelkern, M. A., Olmstead, R. E., Allen-Martinez, Z., Scheibal, D., Abrams, A.
L., Costello, M. R., Farahi, J., Saxena, S., Monterosso, J., & London, E. D. (2009).
Ventral Striatal Dopamine Release in Response to Smoking a Regular vs a Denicotinized
Cigarette. Neuropsychopharmacology, 34(2), 282–289.
https://doi.org/10.1038/npp.2008.87
Brown, S. A. (2015). Stigma towards Marijuana Users and Heroin Users. Journal of
Psychoactive Drugs, 47(3), 213–220. https://doi.org/10.1080/02791072.2015.1056891
Buchanan, A. S., & Latkin, C. A. (2008). Drug use in the social networks of heroin and cocaine
users before and after drug cessation. Drug and Alcohol Dependence, 96(3), 286–289.
https://doi.org/10.1016/j.drugalcdep.2008.03.008
Cacioppo, J. T., & Cacioppo, S. (2014). Social Relationships and Health: The Toxic Effects of
Perceived Social Isolation. Social and Personality Psychology Compass, 8(2), 58–72.
https://doi.org/10.1111/spc3.12087
Cacioppo, J. T., Hawkley, L. C., & Thisted, R. A. (2010). Perceived social isolation makes me
sad: 5-year cross-lagged analyses of loneliness and depressive symptomatology in the
chicago health, aging, and social relations study. Psychology and Aging, 25(2), 453–463.
https://doi.org/10.1037/a0017216
Chatterji, P. (2006). Illicit drug use and educational attainment. Health Economics, 15(5), 489–
511. https://doi.org/10.1002/hec.1085
93
Christie, N. C. (2021). The role of social isolation in opioid addiction. Social Cognitive and
Affective Neuroscience, 16(7), 645–656. https://doi.org/10.1093/scan/nsab029
Christie, N. C., Vojvodic, V., & Monterosso, J. R. (2021). The Early Impact of Social Distancing
Measures on Drug Use. Substance Use and Misuse, 56(7), 997–1004.
https://doi.org/10.1080/10826084.2021.1901934
Clark, R. E., Baxter, J. D., Aweh, G., O’Connell, E., Fisher, W. H., & Barton, B. A. (2015). Risk
Factors for Relapse and Higher Costs Among Medicaid Members with Opioid
Dependence or Abuse: Opioid Agonists, Comorbidities, and Treatment History. Journal
of Substance Abuse Treatment, 57, 75–80. https://doi.org/10.1016/j.jsat.2015.05.001
Cooper, M. L., Kuntsche, E., Levitt, A., Barber, L. L., & Wolf, S. (2016). Motivational models
of substance use: A review of theory and research on motives for using alcohol,
marijuana, and tobacco. In The Oxford handbook of substance use and substance use
disorders, Vol. 1 (pp. 375–421). Oxford University Press.
Copeland, M., Fisher, J. C., Moody, J., & Feinberg, M. E. (2018). Different Kinds of Lonely:
Dimensions of Isolation and Substance Use in Adolescence. Journal of Youth and
Adolescence, 47(8), 1755–1770. https://doi.org/10.1007/s10964-018-0860-3
Coronavirus effects on global drug traffic, Mexican cartels—Los Angeles Times. (n.d.).
Retrieved August 4, 2020, from https://www.latimes.com/world-nation/story/2020-04-
20/cartels-are-scrambling-virus-snarls-global-drug-trade
Costello, A., & Osborne, J. (2019). Best practices in exploratory factor analysis: Four
recommendations for getting the most from your analysis. Practical Assessment,
Research, and Evaluation, 10(1). https://doi.org/10.7275/jyj1-4868
Costenbader, E. C., Astone, N. M., & Latkin, C. A. (2006). The dynamics of injection drug
users’ personal networks and HIV risk behaviors. Addiction, 101(7), 1003–1013.
https://doi.org/10.1111/j.1360-0443.2006.01431.x
COVID-19 may be worsening opioid crisis, but states can take action. (2020, May 28). American
Medical Association. https://www.ama-assn.org/delivering-care/overdose-
epidemic/covid-19-may-be-worsening-opioid-crisis-states-can-take-action
Cox, D. (2021). The state of American friendship: Change, challenges, and loss—The Survey
Center on American Life. Survey Center on American Life.
94
https://www.americansurveycenter.org/research/the-state-of-american-friendship-change-
challenges-and-loss/
Czeisler, M. É., Lane, R. I., Petrosky, E., Wiley, J. F., Christensen, A., Njai, R., Weaver, M. D.,
Robbins, R., Facer-Childs, E. R., Barger, L. K., Czeisler, C. A., Howard, M. E., &
Rajaratnam, S. M. W. (2020). Mental Health, Substance Use, and Suicidal Ideation
During the COVID-19 Pandemic—United States, June 24–30, 2020. MMWR. Morbidity
and Mortality Weekly Report, 69(32), 1049–1057.
https://doi.org/10.15585/mmwr.mm6932a1
Daley, D. C., Smith, E., Balogh, D., & Toscaloni, J. (2018). Forgotten but Not Gone: The Impact
of the Opioid Epidemic and Other Substance Use Disorders on Families and Children.
Commonwealth, 20(2–3). https://doi.org/10.15367/com.v20i2-3.189
Darke, S., & Ross, J. (1997). Polydrug dependence and psychiatric comorbidity among heroin
injectors. Drug and Alcohol Dependence, 48(2), 135–141. https://doi.org/10.1016/S0376-
8716(97)00117-8
Darke, S., & Ross, J. (2002). Suicide among heroin users: Rates, risk factors and methods.
Addiction, 97(11), 1383–1394. https://doi.org/10.1046/j.1360-0443.2002.00214.x
Davis, C. S., & Samuels, E. A. (2020). Opioid Policy Changes During the COVID-19
Pandemic—And Beyond. Journal of Addiction Medicine, 14(4).
https://doi.org/10.1097/ADM.0000000000000679
Davis, J. P., Barr, N., Dworkin, E. R., Dumas, T. M., Berey, B., DiGuiseppi, G., & Cahn, B. R.
(2019). Effect of Mindfulness-Based Relapse Prevention on Impulsivity Trajectories
Among Young Adults in Residential Substance Use Disorder Treatment. Mindfulness.
https://doi.org/10.1007/s12671-019-01164-0
Davis, J. P., Rao, P., Dilkina, B., Prindle, J., Eddie, D., Christie, N. C., DiGuiseppi, G., Saba, S.,
Ring, C., & Dennis, M. (2022). Identifying individual and environmental predictors of
opioid and psychostimulant use among adolescents and young adults following outpatient
treatment. Drug and Alcohol Dependence, 233, 109359.
https://doi.org/10.1016/j.drugalcdep.2022.109359
Day, B. F., & Rosenthal, G. L. (2019). Social isolation proxy variables and prescription opioid
and benzodiazepine misuse among older adults in the U.S.: A cross-sectional analysis of
data from the National Survey on Drug Use and Health, 2015-2017. Drug and Alcohol
Dependence, 204, 107518. https://doi.org/10.1016/j.drugalcdep.2019.06.020
95
De Pirro, S., Galati, G., Pizzamiglio, L., & Badiani, A. (2018). The affective and neural
correlates of heroin versus cocaine use in addiction are influenced by environmental
setting but in opposite directions. Journal of Neuroscience, 38(22), 5182–5195.
https://doi.org/10.1523/JNEUROSCI.0019-18.2018
Deckman, T., DeWall, C. N., Way, B., Gilman, R., & Richman, S. (2014). Can Marijuana
Reduce Social Pain? Social Psychological and Personality Science, 5(2), 131–139.
https://doi.org/10.1177/1948550613488949
Despite the Challenges, We Must Fight Harder to Address the Nation’s Opioid Epidemic. (n.d.).
Retrieved August 6, 2020, from https://morningconsult.com/opinions/despite-the-
challenges-we-must-fight-harder-to-address-the-nations-opioid-epidemic/
DeVido, J. J. (2020). Stimulants: Caffeine, Cocaine, Amphetamine, and Other Stimulants. In
Absolute Addiction Psychiatry Review (pp. 185–203). Springer International Publishing.
https://doi.org/10.1007/978-3-030-33404-8_12
Dingle, G. A., Cruwys, T., & Frings, D. (2015). Social Identities as Pathways into and out of
Addiction. Frontiers in Psychology, 6.
https://www.frontiersin.org/articles/10.3389/fpsyg.2015.01795
Dinwiddie, S. H., Reich, T., & Cloninger, C. R. (1992). Psychiatric comorbidity and suicidality
among intravenous drug users. Journal of Clinical Psychiatry, 53(10), 364–369.
Drug overdose deaths in West Virginia | County Health Rankings & Roadmaps. (2017).
https://www.countyhealthrankings.org/app/west-
virginia/2017/measure/factors/138/data?sort=sc-3
Dumbili, E. W., Hanewinkel, R., Hannah M., Degge, Ezekwe, E. C., & Nnajiofor, M. (2021).
Cannabis use motivations: A study of young adults in Nigeria. Drugs: Education,
Prevention and Policy, 28(6), 585–594. https://doi.org/10.1080/09687637.2020.1834514
Eisenberger, N. I. (2012). The neural bases of social pain: Evidence for shared representations
with physical pain. Psychosomatic Medicine, 74(2), 126–135.
https://doi.org/10.1097/PSY.0b013e3182464dd1
Eisenberger, N. I. (2015). Social Pain and the Brain: Controversies, Questions, and Where to Go
from Here. Annual Review of Psychology, 66(1), 601–629.
https://doi.org/10.1146/annurev-psych-010213-115146
96
Eisenberger, N. I., & Lieberman, M. D. (2004). Why rejection hurts: A common neural alarm
system for physical and social pain. Trends in Cognitive Sciences, 8(7), 294–300.
https://doi.org/10.1016/j.tics.2004.05.010
Fabrigar, L. R., Wegener, D. T., MacCallum, R. C., & Strahan, E. J. (1999). Evaluating the use
of exploratory factor analysis in psychological research. Psychological Methods, 4, 272–
299. https://doi.org/10.1037/1082-989X.4.3.272
Ferguson, J. N., Young, L. J., Hearn, E. F., Matzuk, M. M., Insel, T. R., & Winslow, J. T.
(2000). Social amnesia in mice lacking the oxytocin gene. Nature Genetics, 25(3), 284–
288. https://doi.org/10.1038/77040
Flores-Aranda, J., Goyette, M., Aubut, V., Blanchette, M., & Pronovost, F. (2019). Let’s talk
about chemsex and pleasure: The missing link in chemsex services. Drugs and Alcohol
Today, 19(3), 189–196. https://doi.org/10.1108/DAT-10-2018-0045
Forbes, M. K., & Krueger, R. F. (2019). The Great Recession and Mental Health in the United
States. Clinical Psychological Science: A Journal of the Association for Psychological
Science, 7(5), 900–913. https://doi.org/10.1177/2167702619859337
Fornili, K. (2018). The Opioid Crisis, Suicides, and Related Conditions. Journal of Addictions
Nursing, 29(3), 214–220. https://doi.org/10.1097/JAN.0000000000000240
gili, m, Zanganeh Motlagh, F., & Taghvayi, D. (2017). Prediction of Drug Use Tendency Based
on Psychological Loneliness and Cognitive Emotion Regulation among Addicts under
Abstinence. Research on Addiction, 11(43), 145–160.
Gonzales, R., Anglin, M. D., Beattie, R., Ong, C. A., & Glik, D. C. (2012). Understanding
Recovery Barriers: Youth Perceptions About Substance Use Relapse. American Journal
of Health Behavior, 36(5), 602–614. https://doi.org/10.5993/AJHB.36.5.3
Goodman, C. B., Emilien, B., Becketts, K., Cadet, J. L., & Rothman, R. B. (1996).
Downregulation of mu-opioid binding sites following chronic administration of
neuropeptide FF (NPFF) and morphine. Peptides, 17(3), 389–397.
https://doi.org/10.1016/0196-9781(96)00002-2
Grant, N., Hamer, M., & Steptoe, A. (2009). Social Isolation and Stress-related Cardiovascular,
Lipid, and Cortisol Responses. Annals of Behavioral Medicine, 37(1), 29–37.
https://doi.org/10.1007/s12160-009-9081-z
97
Grimm, C. T., & Bridges, R. S. (1983). Opiate regulation of maternal behavior in the rat.
Pharmacology, Biochemistry and Behavior, 19(4), 609–616.
https://doi.org/10.1016/0091-3057(83)90336-2
Hales, A. H., Williams, K. D., & Eckhardt, C. I. (2015). A Participant Walks Into a Bar…. Social
Psychology, 46(3), 157–166. https://doi.org/10.1027/1864-9335/a000235
Hampton, K. N., Sessions, L. F., & Her, E. J. (2011). CORE NETWORKS, SOCIAL
ISOLATION, AND NEW MEDIA. Information, Communication & Society, 14(1), 130–
155. https://doi.org/10.1080/1369118X.2010.513417
Harris, E. C., & Barraclough, B. (1997). Suicide as an outcome for mental disorders. British
Journal of Psychiatry, 170(3), 205–228. https://doi.org/10.1192/bjp.170.3.205
Harris, P. A., Taylor, S., & Reznikoff, C. (2020). Issue brief: Reports of increases in opioid-and
other drug-related overdose and other concerns during COVID pandemic *Updated.
https://www.abc15.com/news/rebound/keeping-you-safe/amid-covid-19-pandemic-the-
Havard, A., Teesson, M., Darke, S., & Ross, J. (2006). Depression among heroin users: 12-
Month outcomes from the Australian Treatment Outcome Study (ATOS). Journal of
Substance Abuse Treatment, 30(4), 355–362. https://doi.org/10.1016/j.jsat.2006.03.012
Havassy, B. E., Hall, S. M., & Wasserman, D. A. (1991). Social support and relapse:
Commonalities among alcoholics, opiate users, and cigarette smokers. Addictive
Behaviors, 16(5), 235–246. https://doi.org/10.1016/0306-4603(91)90016-B
Hedegaard, H., Miniño, A., Spencer, M. R., & Warner, M. (2021). Drug Overdose Deaths in the
United States, 1999–2020. National Center for Health Statistics ( U.S.).
https://doi.org/10.15620/cdc:112340
Heilig, M. (2015). The Thirteenth Step: Addiction in the Age of Brain Science. Columbia
University Press. http://cup.columbia.edu/book/the-thirteenth-step/9780231172363
Heilig, M., Epstein, D. H., Nader, M. A., & Shaham, Y. (2016). Time to connect: Bringing social
context into addiction neuroscience. Nature Reviews Neuroscience, 17(9), 592–599.
https://doi.org/10.1038/nrn.2016.67
Herman, H. (1979). AN EXPLORATION OF BRAIN SOCIAL ATTACHMENT SUBSTRATES IN
GUINEA PIGS. Bowling Green State University.
98
Hermes, G. L., Rosenthal, L., Montag, A., & McClintock, M. K. (2006). Social isolation and the
inflammatory response: Sex differences in the enduring effects of a prior stressor.
American Journal of Physiology - Regulatory Integrative and Comparative Physiology,
290(2), R273-R282. https://doi.org/10.1152/ajpregu.00368.2005
Hodgins, D. C., Link to external site, this link will open in a new window, el-Guebaly, N., &
Armstrong, S. (1995). Prospective and retrospective reports of mood states before relapse
to substance use. Journal of Consulting and Clinical Psychology, 63(3), 400–407.
https://doi.org/10.1037/0022-006X.63.3.400
Holt-Lunstad, J., Smith, T. B., Baker, M., Harris, T., & Stephenson, D. (2015). Loneliness and
Social Isolation as Risk Factors for Mortality: A Meta-Analytic Review. Perspectives on
Psychological Science, 10(2), 227–237. https://doi.org/10.1177/1745691614568352
Holt-Lunstad, J., Smith, T. B., & Layton, J. B. (2010). Social Relationships and Mortality Risk:
A Meta-analytic Review. PLoS Medicine, 7(7), e1000316.
https://doi.org/10.1371/journal.pmed.1000316
House, J. S., Robbins, C., & Metzner, H. L. (1982). The association of social relationships and
activities with mortality: Prospective evidence from the tecumseh community health
study. American Journal of Epidemiology, 116(1), 123–140.
https://doi.org/10.1093/oxfordjournals.aje.a113387
Hunter-Reel, D., McCrady, B., & Hildebrandt, T. (2009). Emphasizing interpersonal factors: An
extension of the Witkiewitz and Marlatt relapse model. Addiction, 104(8), 1281–1290.
https://doi.org/10.1111/j.1360-0443.2009.02611.x
Inagaki, T. K., & Eisenberger, N. I. (2013). Shared Neural Mechanisms Underlying Social
Warmth and Physical Warmth. Psychological Science.
https://doi.org/10.1177/0956797613492773
Ingram, I., Kelly, P. J., Deane, F. P., Baker, A. L., Goh, M. C. W., Raftery, D. K., & Dingle, G.
A. (2020). Loneliness among people with substance use problems: A narrative systematic
review. Drug and Alcohol Review, 39(5), 447–483. https://doi.org/10.1111/dar.13064
Jain, A. K., Mishra, A., Shakkarpude, J., & Lakhani, P. (2019). Beta endorphins: The natural
opioids. IJCS, 7(3), 323–332.
Johnson, J. E., Schonbrun, Y. C., Nargiso, J. E., Kuo, C. C., Shefner, R. T., Williams, C. A., &
Zlotnick, C. (2013). “I know if I drink I won’t feel anything”: Substance use relapse
99
among depressed women leaving prison. International Journal of Prisoner Health, 9(4),
169–186. https://doi.org/10.1108/IJPH-02-2013-0009
Johnston, L. D., O’Malley, P. M., Miech, R. A., Bachman, J. G., & Schulenberg, J. E. (2016).
Monitoring the future: National survey results on drug use, 1975-2016: Overview, key
findings on adolescent drug use.
Kaada, B., & Torsteinb, O. (1989). Increase of plasma β-endorphins in connective tissue
massage. General Pharmacology, 20(4), 487–489. https://doi.org/10.1016/0306-
3623(89)90200-0
Kadam, M., Sinha, A., Nimkar, S., Matcheswalla, Y., & De Sousa, A. (2017). A Comparative
Study of Factors Associated with Relapse in Alcohol Dependence and Opioid
Dependence. Indian Journal of Psychological Medicine, 39(5), 627–633.
https://doi.org/10.4103/IJPSYM.IJPSYM_356_17
Karkhanis, A., Holleran, K. M., & Jones, S. R. (2017). Dynorphin/Kappa Opioid Receptor
Signaling in Preclinical Models of Alcohol, Drug, and Food Addiction. In International
Review of Neurobiology (Vol. 136, pp. 53–88). Academic Press Inc.
https://doi.org/10.1016/bs.irn.2017.08.001
Kelly, B. C., LeClair, A., & Parsons, J. T. (2013). Methamphetamine Use in Club Subcultures.
Substance Use & Misuse, 48(14), 1541–1552.
https://doi.org/10.3109/10826084.2013.808217
Kelly, J. F., Stout, R. L., Magill, M., & Tonigan, J. S. (2011). The role of Alcoholics Anonymous
in mobilizing adaptive social network changes: A prospective lagged mediational
analysis. Drug and Alcohol Dependence, 114(2–3), 119–126.
https://doi.org/10.1016/J.DRUGALCDEP.2010.09.009
Khajehei, M., & Behroozpour, E. (2018). Endorphins, oxytocin, sexuality and romantic
relationships: An understudied area. World Journal of Obstetrics and Gynecology, 7(2),
17–23. https://doi.org/10.5317/wjog.v7.i2.17
Khatri, U. G., & Perrone, J. (2020). Opioid Use Disorder and COVID-19: Crashing of the Crises.
Journal of Addiction Medicine, 10.1097/ADM.0000000000000684.
https://doi.org/10.1097/ADM.0000000000000684
100
Killgore, W. D. S., Cloonan, S. A., Taylor, E. C., Lucas, D. A., & Dailey, N. S. (2020).
Loneliness during the first half-year of COVID-19 Lockdowns. Psychiatry Research,
294, 113551. https://doi.org/10.1016/j.psychres.2020.113551
Kilwein, T. M., Wedell, E., Herchenroeder, L., Bravo, A. J., & Looby, A. (2022). A qualitative
examination of college students’ perceptions of cannabis: Insights into the normalization
of cannabis use on a college campus. Journal of American College Health, 70(3), 733–
741. https://doi.org/10.1080/07448481.2020.1762612
Kish, S. J. (2008). Pharmacologic mechanisms of crystal meth. CMAJ : Canadian Medical
Association Journal, 178(13), 1679–1682. https://doi.org/10.1503/cmaj.071675
Knight, R., Karamouzian, M., Carson, A., Edward, J., Carrieri, P., Shoveller, J., Fairbairn, N.,
Wood, E., & Fast, D. (2019). Interventions to address substance use and sexual risk
among gay, bisexual and other men who have sex with men who use methamphetamine:
A systematic review. Drug and Alcohol Dependence, 194, 410–429.
https://doi.org/10.1016/j.drugalcdep.2018.09.023
Kochhar, R. (2020, 11). Unemployment rose higher in three months of COVID-19 than it did in
two years of the Great Recession. Pew Research Center.
https://www.pewresearch.org/fact-tank/2020/06/11/unemployment-rose-higher-in-three-
months-of-covid-19-than-it-did-in-two-years-of-the-great-recession/
Kovacs, B., Caplan, N., Grob, S., & King, M. (2021). Social Networks and Loneliness During
the COVID-19 Pandemic. Socius, 7, 2378023120985254.
https://doi.org/10.1177/2378023120985254
Kurti, A. N., Keith, D. R., Noble, A., Priest, J. S., Sprague, B., & Higgins, S. T. (2016).
Characterizing the intersection of Co-occurring risk factors for illicit drug abuse and
dependence in a U.S. nationally representative sample. Preventive Medicine, 92, 118–
125. https://doi.org/10.1016/j.ypmed.2016.09.030
Kyte, D., Jerram, M., Science, R. D.-M., & 2020, U. (2020). Brain opioid theory of social
attachment: A review of evidence for approach motivation to harm. Motivation Science,
6(1), 12–20.
Lamblin, M., Murawski, C., Whittle, S., & Fornito, A. (2017). Social connectedness, mental
health and the adolescent brain. Neuroscience and Biobehavioral Reviews, 80, 57–68.
https://doi.org/10.1016/j.neubiorev.2017.05.010
101
Lappan, S. N., Brown, A. W., & Hendricks, P. S. (2020). Dropout rates of in-person
psychosocial substance use disorder treatments: A systematic review and meta-analysis.
Addiction, 115(2), 201–217. https://doi.org/10.1111/add.14793
Larimer, M. E., Palmer, R. S., & Marlatt, G. A. (2003). Relapse Prevention: An Overview of
Marlatt’s Cognitive-Behavioral Model. In Psychosocial Treatments. Routledge.
Lasco, G., & Yu, V. G. (2023). Pampalibog: Chemsex, desire and pleasure in the Philippines.
Culture, Health & Sexuality, 0(0), 1–16. https://doi.org/10.1080/13691058.2023.2192256
Laudet, A. B., Savage, R., & Mahmood, D. (2002). Pathways to long-term recovery: A
preliminary investigation. Journal of Psychoactive Drugs, 34(3), 305–311.
https://doi.org/10.1080/02791072.2002.10399968
Laws, H. B., Ellerbeck, N. E., Rodrigues, A. S., Simmons, J. A., & Ansell, E. B. (2017). Social
Rejection and Alcohol Use in Daily Life. Alcoholism: Clinical and Experimental
Research, 41(4), 820–827. https://doi.org/10.1111/acer.13347
Leary, M. R., Schreindorfer, L. S., & Haupt, A. L. (1995). The Role of Low Self-Esteem in
Emotional and Behavioral Problems: Why is Low Self-Esteem Dysfunctional? Journal of
Social and Clinical Psychology, 14(3), 297–314.
https://doi.org/10.1521/jscp.1995.14.3.297
Lee, R. M., Draper, M., & Lee, S. (2001). Social connectedness, dysfunctional interpersonal
behaviors, and psychological distress: Testing a mediator model. Journal of Counseling
Psychology, 48(3), 310–318. https://doi.org/10.1037/0022-0167.48.3.310
Lee, R. M., & Robbins, S. B. (1995). Measuring belongingness: The Social Connectedness and
the Social Assurance scales. Journal of Counseling Psychology, 42, 232–241.
https://doi.org/10.1037/0022-0167.42.2.232
LeSaint, K. T., & Snyder, H. R. (2020). Impact of Social Distancing on Individuals Who Use
Drugs: Considerations for Emergency Department Providers. Western Journal of
Emergency Medicine, 21(5), 1102–1104. https://doi.org/10.5811/westjem.2020.7.47896
Løseth, G. E., Eikemo, M., & Leknes, S. (2019). Effects of opioid receptor stimulation and
blockade on touch pleasantness: A double-blind randomised trial. Social Cognitive and
Affective Neuroscience, 14(4), 411–422. https://doi.org/10.1093/scan/nsz022
102
Lutz, P., Courtet, P., & Calati, R. (2020). The opioid system and the social brain: Implications
for depression and suicide. Journal of Neuroscience Research, 98(4), 588–600.
https://doi.org/10.1002/jnr.24269
Lutz, P. E., Ayranci, G., Chu-Sin-Chung, P., Matifas, A., Koebel, P., Filliol, D., Befort, K.,
Ouagazzal, A. M., & Kieffer, B. L. (2014). Distinct Mu, Delta, and Kappa opioid
receptor mechanisms underlie low sociability and depressive-like behaviors during heroin
abstinence. Neuropsychopharmacology, 39(11), 2694–2705.
https://doi.org/10.1038/npp.2014.126
Machin, A. J., & Dunbar, R. I. M. (2011). The brain opioid theory of social attachment: A review
of the evidence. Behaviour, 148(9–10), 985–1025.
https://doi.org/10.1163/000579511X596624
Maloney, E., Degenhardt, L., Darke, S., Mattick, R. P., & Nelson, E. (2007). Suicidal behaviour
and associated risk factors among opioid-dependent individuals: A case-control study.
Addiction, 102(12), 1933–1941. https://doi.org/10.1111/j.1360-0443.2007.01971.x
Martins, S. S., Kim, J. H., Chen, L. Y., Levin, D., Keyes, K. M., Cerdá, M., & Storr, C. L.
(2015). Nonmedical prescription drug use among US young adults by educational
attainment. Social Psychiatry and Psychiatric Epidemiology, 50(5), 713–724.
https://doi.org/10.1007/s00127-014-0980-3
Matthews, G. A., & Tye, K. M. (2019). Neural mechanisms of social homeostasis. Annals of the
New York Academy of Sciences, 1457(1), 5. https://doi.org/10.1111/nyas.14016
Matthews, T., Danese, A., Caspi, A., Fisher, H. L., Goldman-Mellor, S., Kepa, A., Moffitt, T. E.,
Odgers, C. L., & Arseneault, L. (2019). Lonely young adults in modern Britain: Findings
from an epidemiological cohort study. Psychological Medicine, 49(2), 268–277.
https://doi.org/10.1017/S0033291718000788
McGaffin, B. J., Deane, F. P., Kelly, P. J., & Blackman, R. J. (2017). Social support and mental
health during recovery from drug and alcohol problems.
Https://Doi.Org/10.1080/16066359.2017.1421178, 26(5), 386–395.
https://doi.org/10.1080/16066359.2017.1421178
McKechnie, A. A., Wilson, F., Watson, N., & Scott, D. (1983). Anxiety states: A preliminary
report on the value of connective tissue massage. Journal of Psychosomatic Research,
27(2), 125–129. https://doi.org/10.1016/0022-3999(83)90088-0
103
McPherson, M., Smith-Lovin, L., & Brashears, M. E. (2006). Social Isolation in America:
Changes in Core Discussion Networks over Two Decades. American Sociological
Review, 71(3), 353–375. https://doi.org/10.1177/000312240607100301
Mericle, A. A. (2014). The role of social networks in recovery from alcohol and drug abuse. The
American Journal of Drug and Alcohol Abuse, 40(3), 179–180.
https://doi.org/10.3109/00952990.2013.875553
Meyers, R. J., & Miller, W. R. (2001). A Community Reinforcement Approach to Addiction
Treatment. In R. J. Meyers & W. R. Miller (Eds.), A Community Reinforcement
Approach to Addiction Treatment. Cambridge University Press.
https://doi.org/10.1017/CBO9780511570117
Modrek, S., Hamad, R., & Cullen, M. R. (2015). Psychological Well-Being During the Great
Recession: Changes in Mental Health Care Utilization in an Occupational Cohort.
American Journal of Public Health, 105(2), 304–310.
https://doi.org/10.2105/AJPH.2014.302219
Morean, M. E., Corbin, W. R., & Treat, T. A. (2013). The Subjective Effects of Alcohol Scale:
Development and psychometric evaluation of a novel assessment tool for measuring
subjective response to alcohol. Psychological Assessment, 25(3), 780–795.
https://doi.org/10.1037/a0032542
Mulvaney-Day, N., Marshall, T., Downey Piscopo, K., Korsen, N., Lynch, S., Karnell, L. H.,
Moran, G. E., Daniels, A. S., & Ghose, S. S. (2018). Screening for Behavioral Health
Conditions in Primary Care Settings: A Systematic Review of the Literature. Journal of
General Internal Medicine. https://doi.org/10.1007/s11606-017-4181-0
Naqvi, N. H., & Bechara, A. (2009). The hidden island of addiction: The insula. Trends in
Neurosciences, 32(1), 56–67. https://doi.org/10.1016/J.TINS.2008.09.009
National Institute on Drug Abuse. (2023, February 9). Drug Overdose Death Rates. National
Institute on Drug Abuse. https://nida.nih.gov/research-topics/trends-statistics/overdose-
death-rates
National Institute on Drug Abuse. (--). Drug Misuse and Addiction. National Institute on Drug
Abuse. https://nida.nih.gov/publications/drugs-brains-behavior-science-addiction/drug-
misuse-addiction
104
Niesink, R. J. M., Vanderschuren, L. J. M. J., & Van Ree, J. M. (1996). Social play in juvenile
rats after in utero exposure to morphine. NeuroToxicology, 17(3–4), 905–912.
Nobile, B., Lutz, P. E., Olie, E., & Courtet, P. (2020). The Role of Opiates in Social Pain and
Suicidal Behavior. Current Topics in Behavioral Neurosciences, 46, 197–210.
https://doi.org/10.1007/7854_2020_167
Nordfjærn, T., Rundmo, T., & Hole, R. (2010). Treatment and recovery as perceived by patients
with substance addiction. Journal of Psychiatric and Mental Health Nursing, 17(1), 46–
64. https://doi.org/10.1111/j.1365-2850.2009.01477.x
Nummenmaa, L., Tuominen, L., Dunbar, R., Hirvonen, J., Manninen, S., Arponen, E., Machin,
A., Hari, R., Jääskeläinen, I. P., & Sams, M. (2016). Social touch modulates endogenous
μ-opioid system activity in humans. NeuroImage, 138, 242–247.
https://doi.org/10.1016/j.neuroimage.2016.05.063
Olsen, Y., & Sharfstein, J. M. (2014). Confronting the stigma of opioid use disorder—And its
treatment. JAMA - Journal of the American Medical Association, 311(14), 1393–1394.
https://doi.org/10.1001/jama.2014.2147
Overdose Deaths Declined but Remained Near Record Levels During the First Nine Months of
2022 as States Cope with Synthetic Opioids. (2023, March 13).
https://doi.org/10.26099/b912-4124
Palan, S., & Schitter, C. (2018). Prolific.ac—A subject pool for online experiments. Journal of
Behavioral and Experimental Finance, 17, 22–27.
https://doi.org/10.1016/j.jbef.2017.12.004
Panchal, N., Kamal, R., Feb 10, R. G. P., & 2021. (2021, February 10). The Implications of
COVID-19 for Mental Health and Substance Use. KFF. https://www.kff.org/coronavirus-
covid-19/issue-brief/the-implications-of-covid-19-for-mental-health-and-substance-use/
Panksepp, J. (1998). Affective neuroscience: The foundations of human and animal. In New
York: Oxford University Press.
Panksepp, J., Herman, B. H., Vilberg, T., Bishop, P., & DeEskinazi, F. G. (1980). Endogenous
opioids and social behavior. Neuroscience & Biobehavioral Reviews, 4(4), 473–487.
https://doi.org/10.1016/0149-7634(80)90036-6
105
Panksepp, J., Jalowiec, J., DeEskinazi, F. G., & Bishop, P. (1985). Opiates and play dominance
in juvenile rats. Behavioral Neuroscience, 99(3), 441–453. https://doi.org/10.1037/0735-
7044.99.3.441
Panksepp, J., Nelson, E., & Bekkedal, M. (1997). Brain Systems for the Mediation of Social
Separation-Distress and Social-Reward Evolutionary Antecedents and Neuropeptide
Intermediaries. Annals of the New York Academy of Sciences, 807(1 Integrative N), 78–
100. https://doi.org/10.1111/j.1749-6632.1997.tb51914.x
Patrick, M. E., Fairlie, A. M., & Lee, C. M. (2018). Motives for simultaneous alcohol and
marijuana use among young adults. Addictive Behaviors, 76, 363–369.
https://doi.org/10.1016/j.addbeh.2017.08.027
Pettersen, H., Landheim, A., Skeie, I., Biong, S., Brodahl, M., Oute, J., & Davidson, L. (2019).
How Social Relationships Influence Substance Use Disorder Recovery: A Collaborative
Narrative Study. Substance Abuse: Research and Treatment, 13.
https://doi.org/10.1177/1178221819833379
Peugh, J., & Belenko, S. (2001). Alcohol, drugs and sexual function: A review. Journal of
Psychoactive Drugs, 33(3), 223–232. https://doi.org/10.1080/02791072.2001.10400569
Pew Research Center. (2020). Older people are more likely to live alone in the U.S. than
elsewhere in the world | Pew Research Center. https://www.pewresearch.org/fact-
tank/2020/03/10/older-people-are-more-likely-to-live-alone-in-the-u-s-than-elsewhere-in-
the-world/
Portacolone, E. (2013). The notion of precariousness among older adults living alone in the U.S.
Journal of Aging Studies, 27(2), 166–174. https://doi.org/10.1016/j.jaging.2013.01.001
Putnam, R. D. (1995). Bowling Alone: America’s Declining Social Capital. Journal of
Democracy, 6(1), 65–78. https://doi.org/10.1353/jod.1995.0002
Putnam, R. D. (2000). Bowling Alone: The Collapse and Revival of American Community: New
York: Simon und Schuster, 2001. ISBN. Culture and Politics, 223–234.
https://doi.org/10.2307/3089235
Rasheed, A., & Tareen, I. A. (1995). Effects of heroin on thyroid function, cortisol and
testosterone level in addicts. Polish Journal of Pharmacology, 47(5), 441–444.
106
Register, C. A., Williams, D. R., & Grimes, P. W. (2001). Adolescent Drug Use and Educational
Attainment. Education Economics, 9(1), 1–18. https://doi.org/10.1080/09645290124529
Robinson, M. J. F., Zumbusch, A. S., & Anselme, P. (2022, January 28). The Incentive
Sensitization Theory of Addiction. Oxford Research Encyclopedia of Psychology.
https://doi.org/10.1093/acrefore/9780190236557.013.715
Roos, C., Bowen, S., & Witkiewitz, K. (2020). Approach Coping and Substance Use Outcomes
Following Mindfulness-Based Relapse Prevention Among Individuals with Negative
Affect Symptomatology. Mindfulness, 11(10), 2397–2410.
https://doi.org/10.1007/s12671-020-01456-w
Rosenbaum, M. (1981). When drugs come into the picture, Love flies out the window: Women
addicts’ love relationships. Substance Use and Misuse, 16(7), 1197–1206.
https://doi.org/10.3109/10826088109039173
Saladin, M. E., Brady, K. T., Dansky, B. S., & Kilpatrick, D. G. (1995). Understanding
comorbidity between ptsd and substance use disorders: Two preliminary investigations.
Addictive Behaviors, 20(5), 643–655. https://doi.org/10.1016/0306-4603(95)00024-7
Samples, H., Stuart, E. A., & Olfson, M. (2019). Opioid Use and Misuse and Suicidal Behaviors
in a Nationally Representative Sample of US Adults. American Journal of Epidemiology,
188(7), 1245–1253. https://doi.org/10.1093/aje/kwz061
Schiller, E. Y., & Mechanic, O. J. (2017). Opioid overdose. Challenging Cases and
Complication Management in Pain Medicine, 3–7. https://doi.org/10.1007/978-3-319-
60072-7_1
Schneider, B. (2009). Substance Use Disorders and Risk for Completed Suicide. Archives of
Suicide Research, 13(4), 303–316. https://doi.org/10.1080/13811110903263191
SCP Index—Social Capital Project—United States Senator Mike Lee. (2018).
https://www.lee.senate.gov/public/index.cfm/scp-index
Segest, E., Mygind, O., & Bay, H. (1990). The influence of prolonged stable methadone
maintenance treatment on mortality and employment: An 8-year follow-up. Substance
Use and Misuse, 25(1), 53–63. https://doi.org/10.3109/10826089009056200
107
Sharp, B. M. (2017). Basolateral amygdala and stress-induced hyperexcitability affect motivated
behaviors and addiction. Translational Psychiatry, 7(8), e1194.
https://doi.org/10.1038/tp.2017.161
Silk, J. B., Beehner, J. C., Bergman, T. J., Crockford, C., Engh, A. L., Moscovice, L. R., Wittig,
R. M., Seyfarth, R. M., & Cheney, D. L. (2010). Strong and consistent social bonds
enhance the longevity of female baboons. Current Biology, 20(15), 1359–1361.
https://doi.org/10.1016/j.cub.2010.05.067
Simons, J. S., Simons, R. M., Maisto, S. A., Hahn, A. M., & Walters, K. J. (2018). Daily
associations between alcohol and sexual behavior in young adults. Experimental and
Clinical Psychopharmacology, 26(1), 36–48. https://doi.org/10.1037/pha0000163
Sliedrecht, W., de Waart, R., Witkiewitz, K., & Roozen, H. G. (2019). Alcohol use disorder
relapse factors: A systematic review. Psychiatry Research, 278, 97–115.
https://doi.org/10.1016/j.psychres.2019.05.038
Smith, K. E., & Lawson, T. (2017). Prevalence and motivations for kratom use in a sample of
substance users enrolled in a residential treatment program. Drug and Alcohol
Dependence, 180, 340–348. https://doi.org/10.1016/j.drugalcdep.2017.08.034
Snell, K. D. M. (2017). The rise of living alone and loneliness in history. Social History, 42(1),
2–28. https://doi.org/10.1080/03071022.2017.1256093
Social Isolation And Health | Health Affairs Brief. (n.d.). Retrieved February 21, 2023, from
https://www.healthaffairs.org/do/10.1377/hpb20200622.253235/full/
Stafford, K., Gomes, A., Shen, J., & Yoburn, B. C. (2001). μ-Opioid receptor downregulation
contributes to opioid tolerance in vivo. Pharmacology Biochemistry and Behavior, 69(1–
2), 233–237. https://doi.org/10.1016/S0091-3057(01)00525-1
Stapleton, E. (2018, November 9). A Night Near Death | The ‘Heroin Hug:’ The Warmth That
Fuels Addiction | wfmynews2.com. WFMY News.
https://www.wfmynews2.com/article/news/local/a-night-near-death-the-heroin-hug-the-
warmth-that-fuels-addiction/83-612770666
Stone, D., Conteh, J. A., & David Francis, J. (2017). Therapeutic Factors and Psychological
Concepts in Alcoholics Anonymous. Journal of Counselor Practice, 8(2), 120–135.
https://doi.org/10.22229/nav074629
108
Strang, J., Volkow, N. D., Degenhardt, L., Hickman, M., Johnson, K., Koob, G. F., Marshall, B.
D. L., Tyndall, M., & Walsh, S. L. (2020). Opioid use disorder. Nature Reviews Disease
Primers, 6(1), Article 1. https://doi.org/10.1038/s41572-019-0137-5
Sue Carter, C., Courtney Devries, A., & Getz, L. L. (1995). Physiological substrates of
mammalian monogamy: The prairie vole model. Neuroscience and Biobehavioral
Reviews. https://doi.org/10.1016/0149-7634(94)00070-H
Sumnall, H. R., Cole, J. C., & Jerome, L. (2006). The varieties of ecstatic experience: An
exploration of the subjective experiences of ecstasy. Journal of Psychopharmacology,
20(5), 670–682. https://doi.org/10.1177/0269881106060764
The Heroin Hug | Absolute Advocacy. (2016). https://www.absoluteadvocacy.org/the-heroin-
hug/
Trezza, V., Baarendse, P. J. J., & Vanderschuren, L. J. M. J. (2010). The pleasures of play:
Pharmacological insights into social reward mechanisms. Trends in Pharmacological
Sciences, 31(10), 463–469. https://doi.org/10.1016/j.tips.2010.06.008
Trezza, V., Damsteegt, R., Marijke Achterberg, E. J., & Vanderschuren, L. J. M. J. (2011).
Nucleus accumbens μ-opioid receptors mediate social reward. Journal of Neuroscience,
31(17), 6362–6370. https://doi.org/10.1523/JNEUROSCI.5492-10.2011
Valinsky, J. (2020, 01). Booze sales are booming as people stockpile alcohol ... But it may not
last | CNN Business. CNN. https://www.cnn.com/2020/04/01/business/alcohol-sales-
coronavirus-trnd/index.html
Vanderschuren, L. J. M. J., Achterberg, E. J. M., & Trezza, V. (2016). The neurobiology of
social play and its rewarding value in rats. Neuroscience and Biobehavioral Reviews, 70,
86–105. https://doi.org/10.1016/j.neubiorev.2016.07.025
Vanhalst, J., Luyckx, K., Teppers, E., & Goossens, L. (2012). Disentangling the longitudinal
relation between loneliness and depressive symptoms: Prospective effects and the
intervening role of coping. Journal of Social and Clinical Psychology, 31(8), 810–834.
https://doi.org/10.1521/jscp.2012.31.8.810
Veltri, C., & Grundmann, O. (2019). Current perspectives on the impact of Kratom use.
Substance Abuse and Rehabilitation, 10, 23–31. https://doi.org/10.2147/SAR.S164261
109
Venniro, M., Zhang, M., Caprioli, D., Hoots, J. K., Golden, S. A., Heins, C., Morales, M.,
Epstein, D. H., & Shaham, Y. (2018). Volitional social interaction prevents drug
addiction in rat models. Nature Neuroscience, 21(11), 1520–1529.
https://doi.org/10.1038/s41593-018-0246-6
Vijayakumar, L., Kumar, M. S., & Vijayakumar, V. (2011). Substance use and suicide. Current
Opinion in Psychiatry, 24(3), 197–202. https://doi.org/10.1097/YCO.0b013e3283459242
Villalobos-Gallegos, L., Pérez-López, A., Mendoza-Hassey, R., Graue-Moreno, J., & Marín-
Navarrete, R. (2015). Psychometric and diagnostic properties of the Drug Abuse
Screening Test (DAST): Comparing the DAST-20 vs. The DAST-10. Salud Mental.
https://doi.org/10.17711/sm.0185-3325.2015.012
Volkow, N. D., Baler, R. D., & Goldstein, R. Z. (2011). Addiction: Pulling at the Neural Threads
of Social Behaviors. Neuron, 69(4), 599–602.
https://doi.org/10.1016/j.neuron.2011.01.027
Votaw, V. R., & Witkiewitz, K. (2021). Motives for Substance Use in Daily Life: A Systematic
Review of Studies Using Ecological Momentary Assessment. Clinical Psychological
Science, 9(4), 535–562. https://doi.org/10.1177/2167702620978614
Walton, M. A., Blow, F. C., & Booth, B. M. (2000). A comparison of substance abuse patients’
and counselors’ perceptions of relapse risk: Relationship to actual relapse. Journal of
Substance Abuse Treatment, 19(2), 161–169. https://doi.org/10.1016/S0740-
5472(00)00115-X
Warner, M. L., Kaufman, N. C., & Grundmann, O. (2016). The pharmacology and toxicology of
kratom: From traditional herb to drug of abuse. International Journal of Legal Medicine,
130(1), 127–138. https://doi.org/10.1007/s00414-015-1279-y
Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and Validation of Brief Measures
of Positive and Negative Affect: The PANAS Scales. Journal of Personality and Social
Psychology. https://doi.org/10.1037/0022-3514.54.6.1063
Wesselmann, E. D., & Parris, L. (2021). Exploring the Links Between Social Exclusion and
Substance Use, Misuse, and Addiction. Frontiers in Psychology, 12.
https://www.frontiersin.org/articles/10.3389/fpsyg.2021.674743
Western, B., Braga, A. A., Davis, J., & Sirois, C. (2015). Stress and hardship after prison.
American Journal of Sociology, 120(5), 1512–1547. https://doi.org/10.1086/681301
110
What does it feel like to use cocaine? | Drug Policy Alliance. (n.d.). Retrieved January 9, 2020,
from http://www.drugpolicy.org/drug-facts/cocaine/what-cocaine-feels-like
Wojcicki, J. M. (2019). Dying alone: The sad irrelevance of naloxone in the context of solitary
opiate use. Addiction, 114(3), 574–575. https://doi.org/10.1111/add.14508
Wong, S. S., Zhou, B., Goebert, D., & Hishinuma, E. S. (2013). The risk of adolescent suicide
across patterns of drug use: A nationally representative study of high school students in
the United States from 1999 to 2009. Social Psychiatry and Psychiatric Epidemiology,
48(10), 1611–1620. https://doi.org/10.1007/s00127-013-0721-z
Yang, Y. C., Boen, C., Gerken, K., Li, T., Schorpp, K., & Harris, K. M. (2016). Social
relationships and physiological determinants of longevity across the human life span.
Proceedings of the National Academy of Sciences of the United States of America,
113(3), 578–583. https://doi.org/10.1073/pnas.1511085112
Yang, Y. C., McClintock, M. K., Kozloski, M., & Li, T. (2013). Social isolation and adult
mortality: The role of chronic inflammation and sex differences. Journal of Health and
Social Behavior, 54(2), 183–203. https://doi.org/10.1177/0022146513485244
Yeh, C. J., & Inose, M. (2003). International students’ reported English fluency, social support
satisfaction, and social connectedness as predictors of acculturative stress. Counselling
Psychology Quarterly, 16, 15–28. https://doi.org/10.1080/0951507031000114058
Yudko, E., Lozhkina, O., & Fouts, A. (2007). A comprehensive review of the psychometric
properties of the Drug Abuse Screening Test. Journal of Substance Abuse Treatment,
32(2), 189–198. https://doi.org/10.1016/j.jsat.2006.08.002
Zaidi, U. (2020). Role of Social Support in Relapse Prevention for Drug Addicts.
Zeiger, J. S., Haberstick, B. C., Corley, R. P., Ehringer, M. A., Crowley, T. J., Hewitt, J. K.,
Hopfer, C. J., Stallings, M. C., Young, S. E., & Rhee, S. H. (2010). Subjective effects to
marijuana associated with marijuana use in community and clinical subjects. Drug and
Alcohol Dependence, 109(1–3), 161–166.
https://doi.org/10.1016/j.drugalcdep.2009.12.026
Zoorob, M. J., & Salemi, J. L. (2017). Bowling alone, dying together: The role of social capital
in mitigating the drug overdose epidemic in the United States. Drug and Alcohol
Dependence, 173, 1–9. https://doi.org/10.1016/J.DRUGALCDEP.2016.12.011
111
Appendix A: Chapter 1
Appendix Table 1: Overall emotional response predicted by Drug of Choice, Emotion Category,
and the interaction term Drug of Choice*Emotion Category
Df Sum Sq Mean Sq F Value Pr(>F)
Emotional
Category
3 104.60 34.88 33.41 0.001
Drug of
Choice
3 16.27 5.42 5.19 0.001
Emotional
Category:
Drug of
Choice
9 28.7 3.19 3.05 0.001
Residuals 1229 1283 1.04 NA NA
Appendix Table 2: Endorsement of Hug/Belong items predicted by Drug of Choice; covariates
are DAST, Age, Income, and Education
Df Sum Sq Mean Sq F Value Pr(>F)
Drug of
Choice
3 12.66 4.22 4.11 0.007
DAST 1 4.04 4.04 3.94 0.048
Age 1 0.57 0.57 0.56 0.46
Income
(Ordinal)
11 7.46 0.68 0.66 0.77
Gender 3 2.83 0.94 0.92 0.43
Education
(Ordinal)
7 12.45 1.78 1.73 0.10
Residuals 245 251.30 1.03 NA NA
112
Appendix Table 3: Endorsement of Secure/Loved items predicted by Drug of Choice; covariates
are DAST, Age, Income, and Education
Df Sum Sq Mean Sq F Value Pr(>F)
Drug of
Choice
3 6.66 2.22 1.93 0.12
Total DAST 1 0.15 0.15 0.13 0.72
Age 1 0.01 0.01 0.01 0.92
Income
(Ordinal)
11 14.97 1.36 1.19 0.30
Gender 3 17.86 5.95 5.19 0.002
Education
(Ordinal)
7 5.77 0.82 0.72 0.66
Residuals 245 281.20 1.15 NA NA
Appendix Table 4: Endorsement of Content/Satisfied items predicted by Drug of Choice;
covariates are DAST, Age, Income, and Education
Df Sum Sq Mean Sq F Value Pr(>F)
Drug of
Choice
3 9.77 3.26 3.30 0.02
Total DAST 1 0.02 0.02 0.02 0.88
Age 1 2.46 2.46 2.49 0.12
Income
(Ordinal)
11 21.49 1.95 1.98 0.03
Gender 3 9.15 3.05 3.09 0.03
Education
(Ordinal)
7 12.53 1.79 1.81 0.09
Residuals 246 242.80 0.99 NA NA
Appendix Table 5: Endorsement of Excited/Happy items predicted by Drug of Choice;
covariates are DAST, Age, Income, and Education
Df Sum Sq Mean Sq F Value Pr(>F)
Drug of
Choice
3 9.46 3.152 4.01 0.008
Total DAST 1 0.42 0.42 0.54 0.46
Age 1 0.37 0.37 0.47 0.50
Income
(Ordinal)
11 12.67 1.15 1.46 0.15
Gender 3 8.31 2.77 3.52 0.02
Education
(Ordinal)
7 5.43 0.78 0.99 0.44
Residuals 246 193.60 0.79 NA NA
113
Appendix Table 6: DAST Score predicted by Drug of Choice, each emotion category
(independently), and the interaction between Drug of Choice and each emotion category
Df Sum Sq Mean Sq F Value Pr(>F)
Drug of Choice 3 246.6 82.21 19.368 < 0.001
Content/Satisfied 1 0.10 0.12 0.03 0.87
Loved/Secure 1 1.30 1.33 0.312 0.58
Excited/Happy 1 3.60 3.60 0.85 0.36
Hug/Belong 1 20.40 20.38 4.80 0.03
DOC:Content/Satisfied 3 5.50 1.83 0.43 0.73
DOC:Loved/Secure 3 10.40 3.48 0.82 0.48
DOC: Excited/Happy 3 7.10 2.38 0.56 0.64
DOC: Hug/Belong 3 6.00 1.99 0.47 0.70
Residuals 251 1065.4 4.24 NA NA
114
Appendix B: Chapter 3
Hypothesis testing as listed in the pre-registration that was not addressed in the main text:
H5: We predict that those in the recovery group will report different emotional responses
/ risk of relapse overall than those in the no history of substance use disorder group. (We
do not have a prediction about directionality).
We will do a t-test of the overall global scores on the NIRR to assess differences in
responses to the risk of relapse and the rated emotional response to the scenarios among
these two groups. We will also perform t-tests to determine if there are differences across
each of the four types of vignettes (pos/neg & social/nonsocial) between the groups. We
will then use ANOVA to determine if there are any interactions between the two groups
across the four levels of the NIRR vignettes.
Planned Analyses (Emotional Response)
In the model predicting emotional response by Protagonist DOC*Group*Sociality, we
see a main effect of all three predictors. There is a main effect of Protagonist DOC (F(3,1732) =
5.48, p < 0.001) such that for the marijuana vignettes, people are reporting a higher perceived
emotional response than for the opioid and meth vignettes. For Group (F(1,1732) = 13.35, p <
0.001), people with a Hx-PU report higher perceived emotional responses than those with No
Hx-PU. Lastly, for Sociality (F(1,1732) = 18.44, p < 0.001), social events elicited a higher
perceived emotional response than nonsocial events. There is also a significant interaction
between DOC*Group (F(3,1732) = 2.81, p < 0.05).
115
Appendix Figure 2: Emotional response predicted by a three-way interaction between
protagonist DOC*sociality*group.
Planned Analyses (T-tests of Relapse Risk Scores)
Global Relapse Risk: T-test of global NIRR relapse risk scores between groups. There is
no significant difference in global NIRR scores of relapse risk between those with a Hx-
PU and those with No Hx-PU (t = 1.19, p = 0.23, 95% CI [-1.14, 4.66]).
Positive Social: T-test of positive social relapse risk scores between groups. There is no
significant difference in perceived relapse risk between those with a Hx-PU and those
with No Hx-PU (t = -0.09, p = 0.93, 95% CI [-0.29, 0.26]).
Positive Nonsocial: T-test of positive nonsocial relapse risk scores between groups.
There is no significant difference in perceived relapse risk between those with a Hx-PU
and those with No Hx-PU (t = 1.37, p = 0.17, 95% CI [-0.10, 0.58]).
116
Negative Social: T-test of negative social relapse risk scores between groups. There is no
significant difference in perceived relapse risk between those with a Hx-PU and those
with No Hx-PU (t = 1.34, p = 0.18, 95% CI [-0.08, 0.41]).
Negative Nonsocial: T-test of negative nonsocial relapse risk scores between groups.
There is no significant difference in perceived relapse risk between those with a Hx-PU
and those with No Hx-PU (t = 0.37, p = 0.71, 95% CI [-0.22, 0.32]).
Planned Analysis (ANOVA of Relapse Risk Scores)
We predicted perceived risk of relapse to life events (using the reverse coded positive
items) with a three-way interaction between group*sociality*protagonist DOC. There was a main
effect of sociality (F(1,1731) = 181.9, p < 0.001);, and a main effect of emotional response
(F(1,1731) = 146.2, p < 0.001). There were no other significant effects in the model.
Appendix Figure 1: Perceived risk / protective effects of life events predicted by
group*sociality*protagonist DOC interaction and emotional response. X-axis indicates the Drug of
Choice for the protagonist in the vignette; colors indicate sociality of the event depicted in the NIRR
vignette; circles indicate the predicted marginal means of the perceived risk in the model, error bars
indicate the confidence interval.
Abstract (if available)
Abstract
This three-chapter dissertation examines the association between social connection and substance use. In the introduction, I use excerpts from my previously published work (Christie, 2021) to delineate the theoretical foundations and public health impetus for the present work. Following this introduction to the topic, I present a series of studies that assess how substance use influences feelings of social wellbeing, which then influences substance use behaviors, and how this iterative process impacts the trajectory of recovery from problem substance use, with an emphasis on opioid use.
Chapter 1. I evaluate the hypothesis that opioid use is associated with the subjective experience of positive feelings that are typically elicited by social connection. I aim to answer the question: Is the acute opioid high associated with feelings that normally accompany positive social connection and intimacy - above that of other drugs?
Chapter 2. I share my published work that assesses how social distancing measures aimed at reducing the spread of COVID-19 impacted substance use behaviors (Christie et al., 2021). I aim to answer the question: Did the implementation of large-scale social distancing measures impact drug use behaviors?
Chapter 3. I measure perceptions of relapse risk to emotionally significant life events that vary by: valence (positive/negative event), sociality (social/nonsocial event), drug of choice, and personal history of problem substance use (yes/no). Here, I aim to answer the questions: 1) Do people perceive social events to be particularly impactful in terms of mitigating / exacerbating relapse risk relative to nonsocial events?
Conclusions. Across these three chapters, I focus on the critical role of a commonly overlooked factor in substance use: social connection. This work has implications for both clinicians and policy-makers to improve the prognosis of substance use disorders through the inclusion of social connection at each stage from prevention to treatment.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
The moral foundations of needle exchange attitudes
PDF
Social self-control and adolescent substance use
PDF
Homelessness and substance use treatment: using multiple methods to understand risks, consequences, and unmet treatment needs among young adults
PDF
Opioid withdrawal symptoms, opioid use, and injection risk behaviors among people who inject drugs (PWID)
PDF
Anthropomorphic sociality theory: how connections to nonhumans connect us to humans
PDF
A multi-system evaluation of medication for opioid use disorder for people who use opioids
PDF
The impact of childhood trauma on substance use and mental health during the SARS-CoV-2 pandemic among young adults
PDF
Adolescent conduct problems and substance use: an examination of the risk pathway across the transition to high school
PDF
A qualitative study of street fentanyl in Dayton, Ohio: drug markets, trajectories, and overdose risk reduction
PDF
Using integrative data analysis to evaluate gender differences in effects of multisystemic therapy for justice-involved youth
PDF
Integrative care strategies for older adults experiencing co-occurring substance use and mental health disorders (I-CARE)…
PDF
Social exclusion decreases risk-taking
PDF
Friendship network position on adolescent behaviors: an examination of a broker position and the likelihood of alcohol and cigarette use
PDF
Technology enhanced substance use disorder treatment
PDF
The dynamic relationship of emerging adulthood and substance use
PDF
HUB_mv: an evidence-based non-police led MH/SUD diversion program in rural MA
PDF
Interagency collaboration: cultivating resources and strengthening substance use service delivery in child welfare
PDF
Essays in opioid use and abuse
PDF
Distress tolerance and mindfulness disposition: associations with substance use during adolescence and emerging adulthood
PDF
Climate change communication: challenges and insights on misinformation, new technology, and social media outreach
Asset Metadata
Creator
Christie, Nina Caitlin
(author)
Core Title
The interplay between social connection and substance use
School
Dual Degree
Degree
Doctor of Philosophy / Master of Public Health
Degree Program
Psychology
Degree Conferral Date
2023-08
Publication Date
06/14/2023
Defense Date
04/12/2023
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
Addiction,BOTSA,Isolation,OAI-PMH Harvest,opioid,opioids,social connection,substance use,substance use disorder
Format
theses
(aat)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Monterosso, John R. (
committee chair
), Bechara, Antoine (
committee member
), Davis, Jordan P. (
committee member
), Lai, Mark (
committee member
), Lucas, Gale (
committee member
)
Creator Email
ncchrist@usc.edu,ninacchristie@gmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-oUC113170533
Unique identifier
UC113170533
Identifier
etd-ChristieNi-11959.pdf (filename)
Legacy Identifier
etd-ChristieNi-11959
Document Type
Dissertation
Format
theses (aat)
Rights
Christie, Nina Caitlin
Internet Media Type
application/pdf
Type
texts
Source
20230616-usctheses-batch-1056
(batch),
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Repository Email
cisadmin@lib.usc.edu
Tags
BOTSA
opioid
opioids
social connection
substance use
substance use disorder