The Impacts of Social Support and Relationship Characteristics on Commitment to Sobriety Among People in Opioid Use Disorder Recovery

Abstract

Objective:

Despite evidence that social support is beneficial for people living with opioid use disorders (OUDs), research has yet to investigate whether social support within certain relationships is more or less effective. The current study examined whether social support, relationship closeness with a disclosure partner, and/or the history of joint substance use between participants and disclosure partners affect commitment to sobriety among people receiving medications for OUD.

Method:

Over a period of 3 months (two time points), participants taking OUD medications took part in a mixed-methods egocentric social network study exploring their relationships with disclosure partners before and following OUD disclosure (i.e., first telling a disclosure partner about one's OUD history or treatment). Data included 131 disclosure events/relationships clustered within 106 participants.

Results:

Greater relationship closeness was associated with increased commitment to sobriety over time. Further, significant interactions were found between social support and disclosure partner closeness, partner closeness and history of joint substance use, and social support and history of joint substance use. Higher social support was associated with greater commitment to sobriety among those disclosing to close partners. In contrast, receiving social support or disclosing to a close partner with whom there was a history of joint substance use was associated with decreased commitment to sobriety.

Conclusions:

Findings highlight the complexities of social support among people in treatment for OUD and demonstrate that relationship closeness and a history of joint substance use with a disclosure partner may be important factors to consider before disclosure.

Strengthening and supporting commitment to sobriety can promote positive recovery trajectories among people living with opioid use disorder (OUD; Kelly & Greene, 2014b). Commitment to sobriety combines aspects of motivation with self-efficacy and comprises the intention to abstain from substance use while prioritizing recovery (Dingle et al., 2019; Jonas et al., 2019; Kelly & Greene, 2014a, 2014b). Previous research supports associations between commitment to sobriety with long-term recovery from substance use and positive aspects of identity transformation such as enhanced well-being (Bergman et al., 2015; Dingle et al., 2019; Kelly & Greene, 2014a). Therefore, greater insight into predictors of commitment to sobriety can inform intervention efforts to support recovery outcomes among people living with OUD. Social support has also been identified as a crucial component of recovery that is associated with positive outcomes such as abstinence and treatment retention (Dobkin et al., 2002).

Yet, to access the potential benefits of social support from others, people living with OUD must first disclose their OUD status to another person (i.e., disclosure partner) through the act of OUD disclosure (Chaudoir & Fisher, 2010; Earnshaw et al., 2013). OUD disclosure to a disclosure partner (e.g., family member or friend) can be a stressful event, the results of which may play an important role in OUD recovery; yet little research has focused on disclosure, social relationships, and recovery-related outcomes, such as commitment to sobriety (Earnshaw et al., 2021). It is possible that people who receive greater social support following disclosure experience greater commitment to sobriety over time, especially in the presence of certain relationship characteristics, such as disclosing to a partner with whom they feel greater interpersonal closeness.

Previous literature and theory on substance use disorders have suggested that associations between social support and health outcomes may depend, in part, on relationship characteristics (DiMatteo, 2004; Weston et al., 2018). Although there has been ample evidence supporting the role of social support on recovery outcomes (e.g., Brown et al., 2013), researchers tend to combine multiple sources of social support and do not assess whether there are differential impacts related to relationship characteristics. Berkman and colleagues’ (2000) model on how social networks affect health provides a conceptual framework that highlights relationship characteristics as complex mechanisms through which social support can impact health outcomes important for OUD recovery. Applying Berkman and colleagues’ model (2000) within the context of OUD recovery suggests that it is crucial to not only understand how social support directly impacts health outcomes, such as commitment to sobriety, but also for research to consider which types of relationship characteristics play a facilitative or detrimental role for people living with OUD.

According to Berkman and colleagues’ model (2000), both interpersonal closeness and social influence, such as a history of joint substance use (i.e., having used substances in the past with the disclosure partner), may have an impact on the association between social support and commitment to sobriety. Interpersonal closeness reflects the strength of an emotional bond between two people and/or the degree of shared idiosyncratic knowledge (Dibble et al., 2012). Previous research supports associations between closeness and health. For example, supportive close relationships enhance mental and physical well-being, whereas unsupportive close relationships create stress that can undermine health and well-being (Uchino et al., 1996). Although less explored, similar associations have been shown in relation to opioid use and treatment outcomes (Nam & Fukui, 2017), where closer marital relationships were predictive of less heroin use (Heinz et al., 2009).

The impact of social support and closeness on recovery outcomes may further depend on whether the source of support also has a history of substance use. Although social support is often viewed as beneficial to recovery, there is evidence to suggest that social support within certain types of relationships may promote or inhibit recovery behaviors (Buchanan & Latkin, 2008; Nam & Fukui, 2017). For example, mixed evidence suggests positive correlations between the quantity of people in recovery within one's social network and quality of life (Muller et al., 2017), whereas other findings demonstrate that social network members who have engaged in previous problematic substance use can be either helpful or detrimental for recovery (i.e., unsupportive of abstinence; Tracy et al., 2012).

Review of the relevant previous findings indicates that the association between social support and commitment to sobriety among people in recovery from OUDs is likely complex, with relationship characteristics playing a pivotal role. Using dual frameworks, including the Disclosure Process Model (Chaudoir & Fisher, 2010) and the Berkman and colleagues’ (2000) model, the current study explored the association between social support and commitment to sobriety within the context of a disclosure event or particular relationship (i.e., disclosure partner). The Disclosure Process Model theorizes that a single disclosure event can have important implications for long-term health outcomes.

Our first research question was to assess the effects of social support and relationship characteristics (i.e., closeness and a history of joint substance use between a participant and disclosure partner) on commitment to sobriety. It was expected that more social support received from a disclosure partner and disclosure in the context of closer relationships would result in increased commitment to sobriety from baseline. In contrast, it was expected that history of joint substance use with a disclosure partner would result in decreased commitment to sobriety from baseline.

Our second research question aimed to examine the interaction effects of (a) social support and closeness, (b) social support and disclosure to a partner with a history of joint substance use, and (c) closeness and disclosure to a partner with a history of joint substance use on commitment to sobriety. Interaction effects were hypothesized such that high social support within a close relationship would result in increased commitment to sobriety, yet social support and/ or closeness among relationships with a history of joint substance use would have a negative effect on commitment to sobriety.

Method

Sample and procedures

The sample includes participants from an original parent study of 146 people receiving treatment for OUDs and located in Delaware, a state that has demonstrated some of the highest rates of opioid-related overdose deaths (Vivolo-Kantor, 2018). The parent study, the UDisclose Project, is a longitudinal, egocentric social network analysis that lasted from March 2018 to May 2019. Participants were eligible to participate if they were 18 years or older; were seeking treatment for a substance use disorder at a well-established Delaware substance use disorder treatment facility; and were planning to disclose or tell one or more disclosure partners about their substance use history, recovery, or treatment within approximately 3 months following their baseline appointment. All participants were actively in OUD treatment reporting a primary substance of opioids (93.4%) or some combination of opioids and another substance (6.6%; e.g., opioids and cocaine).

Participants were recruited with purposeful sampling using flyers, word of mouth, and in-person recruitment at two treatment locations: suburban and urban environments. Throughout the recruitment and study period, two members of the study team would visit the respective clinics daily to answer questions about the study and to perform screening and assessments. Eligible participants were asked to take part in a quantitative survey at two time points (approximately 3 months apart), where the full study lasted from 1 to 2 hours. Computerized surveys using Qualtrics were administered by the study team members and took place oneon-one in private rooms within the participant's respective treatment location.

Each session included quantitative surveys spanning substance use, treatment environments, and social networks. Between time points (approximately 3 months), the research team called and texted respondents with appointment reminders and established a phone line unique to the study and accessible only by study team members to allow for a private means of communication to reduce attrition (Nicks et al., 2017). Participants answered similar measures at both time points and were compensated with $50 in gift certificates to local establishments for each interview. Participants were asked to describe their relationship with the disclosure partner and identify descriptive characteristics about the disclosure partner, including their gender, race/ethnicity, age, whether the partner was in recovery, and their role (e.g., immediate family member) in relation to the participant. The study received institutional review board approval from the University of Delaware, and all participants provided informed consent.

Of the original 146 participants, 208 disclosure events were planned. One hundred and twenty-four participants came back for Time 2 (85%), and 143 of the planned disclosures were completed from Time 1 to Time 2. Of the remaining, 18 participants were not included within the current analysis because of incomplete or missing data and other extenuating circumstances that arose following study initiation (e.g., participant death). Further, participants who planned to disclose to Person X during Time 1 but disclosed to Person Y instead were excluded from the current analysis because of incongruent Time 1 and Time 2 relationship data. More participants recruited from the urban clinic, χ²(2) = 11.605, p = .003, and who were older, t(145) = 3.27, p = .001, were lost to follow-up. The final sample included 131 disclosure events or disclosure partner relationships nested within 106 participants.

Measures

Covariates assessed at baseline were participant age, gender (dichotomized 1 = men, 0 = women), race/ethnicity (dichotomized 1 = non-Hispanic White, 0 = other [i.e., composite of Black, Asian, Latinx, Native American, Native Hawaiian, and multiple]), treatment facility location (dichotomized 1 = urban, 0 = suburban), participant medication type (1 = methadone), and if the disclosure partner was in recovery (1 = yes). Participant ID was used as a cluster to match participants within disclosure relationships.

Commitment to sobriety was assessed using the Commitment to Sobriety Scale at Times 1 and 2. Developed as a brief measure assessing an individual's willingness and motivations to stay away from active substance use and continue in recovery (Kelly & Green, 2014a), the measure consisted of five items assessing the following: (a) staying sober is the most important thing in my life, (b) I am totally committed to stay off drugs, (c) I will do whatever it takes to recover, (d) I never want to return to drug use, and (e) I have had enough. The original response categories were recoded to scale ranging from 0 (strongly disagree) to 5 (strongly agree), where a higher score corresponded to better commitment to sobriety. Scores were averaged per participant for one composite commitment to sobriety score per time point. The dependent variable was then assessed as a change score, taking the difference between the two time points (Cronbach's α = .93).

Disclosure partner relationship variables

Social support was assessed at Time 2 using items from the modified Medical Outcomes Study Social Support Survey (mMOS-SS; Moser et al., 2012). The mMOS-SS was created to reflect enacted forms of social support rated by participants who are dealing with stressful life circumstances. Social support items were included as part of an egocentric social network assessment to rate the acts of social support provided by specific disclosure partner(s) named by the participant. The social support scale at Time 2 consisted of three items asking how often the participant turned to their disclosure partner for help when needed, for suggestions on personal problems, and for fun. The original response categories were recoded to range from 0 to 4 (where 0 = none of the time and 4 = all of the time) and averaged at the level of the disclosure relationship for one composite social support score per disclosure. Scores closer to 0 corresponded with lower social support, and scores closer to 4 corresponded with higher social support (Cronbach's α = .87).

Closeness was adapted from the Unidimensional Relationship Closeness Scale (URCS), designed to assess closeness in social and personal relationships (Dibble et al., 2012). The current study used an abbreviated version of the URCS at Time 2 to establish self-reports of closeness between study participants and their disclosure partners(s) (Dibble et al., 2012). The closeness scale consisted of three items (i.e., [a] This person [disclosure partner] and I talk about important personal things, [b] my relationship with this person is close, and [c] this person and I do a lot together). Original response categories were recoded to scale ranging from 0 to 4 (where 0 = not true at all and 4 = completely true) and averaged to create composite closeness scores per disclosure. Scores closer to zero corresponded with less closeness between the participant and disclosure partner, whereas scores closer to four corresponded with more closeness between the participant and disclosure partner (Cronbach's α = .92).

History of joint substance use between the participant and the disclosure partner was assessed at Time 2, with one question asking whether the participant had ever used alcohol or other substances with their disclosure partner(s). Response categories included yes or no, and the final measure was dichotomized (1 = yes).

Analysis

Before we tested research questions, data were screened for univariate normality and outliers. The regression model was tested using Mplus 8 with an added cluster accounting for participants (i.e., participant ID). A robust maximum likelihood parameter estimator was also used to further account for the non-independence of observations and to give unbiased parameter estimates due to missing data. T tests were used to determine whether missing data were missing at random. Interactions were tested by creating product terms. Before the regression analysis, it was planned to examine relevant two-way interactions regardless of main effect significance in order to identify group differences and allow for a better understanding of when social support can be most effective for commitment to sobriety within a disclosure context (Fairchild & MacKinnon, 2009).

Model fit was saturated with a nonsignificant chi-square test for the baseline model. All continuous variables and their interaction terms were mean centered to improve the interpretation of the coefficients. We controlled for an inflated type 1 error and possible confounds by running the Benjamini–Hochberg correction (Benjamini & Hochberg, 1995). The Benjamini–Hochberg correction helps to control for a false discovery rate and has been deemed more powerful in comparison to Bonferroni's correction (Williams et al., 1999). The Benjamini–Hochberg correction is reflected as an adjusted (adj.) p value within the generalized linear mixed model.

Results

Common demographic and descriptive information is presented for the current sample of participants living with OUD at both the participant and relationship levels (Table 1). The participant-level variables evaluated included gender, location, race/ethnicity, age, medication type, and commitment to sobriety change scores. Overall, the majority of the sample included White people (77%) in treatment for OUDs at the suburban clinic (76%). Among all the participants, there was a mean age of 38 years, and the average commitment to sobriety change score increased from Time 1 to Time 2, indicating growth over time. Within disclosure relationships, relevant variables assessed included history of joint substance use, social support received from the disclosure partner following the disclosure event, the level of closeness within the relationship, and the covariate assess ing if the disclosure partner was in recovery. The majority of participants (78%) reported no history of joint substance use (i.e., the participant never used substances with the disclosure partner). Further, participants indicated a moderate amount of social support and closeness with their disclosure partner.

Table 1.

Descriptive statistics

Variable	n (%) or M (sD), range
Participant variables (n = 106)
Gender
Female	62 (58.8%)
Male	44(41.2%)
Location
Suburban	81 (76.4%)
Urban	25 (23.6%)
Race/ethnicity
Other	24 (22.9%)
White	82 (77.1%)
Medication type
Methadone	16 (14.7%)
Buprenorphine	91 (87.3%)
Age, in years	37.84(11.23), 18.0-76.0
Commitment to sobriety,a Δ	0.116 (0.68), −2.0−2.4
Relationship variables (n = 131)
History of joint substance use
No	102 (78.0%)
Yes	29 (22.0%)
Social supportb	1.55 (1.10), 0.0-4.0
Closenessc	2.61 (1.13), 0.0-4.0

Notes: N = 131 relationships (disclosure relationships), n (clusters) = 106 participants. Within race/ethnicity, other includes responses of Black, Asian, Latino, Native American, Native Hawaiian, and mixed.

^a Commitment to sobriety (measured as a change score between Times 1 and 2) was averaged per participant, in which a higher score reflects more commitment.

^b Social support from the participant's disclosure partner(s) was measured via the modified Medical Outcomes Study Social Support Survey, in which a higher score indicates more social support.

^c Closeness between the participant and his/her disclosure partner(s) was adapted from the Unidimensional Relationship Closeness Scale, in which a higher score indicates more closeness between participant and disclosure partner. Relationship variables (i.e., social support, closeness) were averaged at the level of the disclosure relationship for one composite score per disclosure relationship.

Table 2 reflects the demographic characteristics for the disclosure partners, or those to whom the participants disclosed, reported by participants. Fifty-eight percent of the disclosure partners were female, with an average age of 49.9 years. The majority of disclosure partners were classified as White (78.4%) and were considered members of the participant's immediate family (41.2%; e.g., parent) or extended family (24.3%; e.g., cousin). Only a small percentage of the disclosure partners were known to be in recovery from substance use (5%). The mean length of time for participant disclosure conversations was 53.2 minutes (SD = 64.3 minutes), the majority of participants (66%) disclosed to only one person, and 79% chose to have their disclosure conversation in person versus over the phone or via an alternative method (e.g., letter).

Table 2.

Disclosure partner descriptive statistics

Variable	n (%) or M (SD), range
Gender
Female	76 (58.1%)
Male	55 (41.9%)
Role
Immediate family	54 (41.2%)
Extended family	32 (24.3%)
Significant other	9 (6.6%)
Friend	21 (16.2%)
Work relationship	11 (8.8%)
Other	4 (2.9%)
Race/ethnicity
Other	28 (21.6%)
White	103 (78.4%)
In recovery
No/unsure	124 (94.9%)
Yes	7 (5.1%)
Age, in years	49.99 (17.89), 10.0-96.0

Notes: N = 131 relationships (disclosure relationships), n (clusters) = 106 participants. Within race/ethnicity, other includes responses of Black, Asian, Latino, Native American, Native Hawaiian, and mixed.

Results of the generalized linear mixed model are included in Table 3. Significant results retained their significance following the Benjamini–Hochberg correction reflected as adj. p values. The first research question was partially supported in that a significant effect was found for closeness and history of joint substance use. However, there was no statistically significant association between social support and change in commitment to sobriety. Closeness was positively associated with a change in commitment to sobriety (b = 0.135, p = .05, adj. p = .063), such that participants who disclosed to partners with whom they felt a closer relationship showed an increase in commitment to sobriety. There was also a significant association between history of joint substance use and commitment to sobriety (b = -0.498, p = .009, adj. p = .013), such that previous substance use with a disclosure partner predicted a decreased commitment to sobriety from baseline.

Table 3.

Unstandardized results predicting change in commitment to sobriety

Variables	b	SE b	95% CI		p	Adj. p
			LL	UL
Predictor variables
Social support	-0.072	0.086	-0.244	0.099	.405	.125
Closeness	0.135	0.078	0.017	0.289	.050	.063
Joint substance use	−0.498	0.187	−0.871	−0.126	.009	.013
Interactions
Social Support × Closeness	0.091	0.047	0.002	0.191	.037	.038
Social Support × Joint Substance Use	−0.407	0.187	−0.787	−0.054	.018	.025
Closeness × Joint Substance Use	−0.239	0.123	−0.483	−0.058	.042	.050
Covariates
Partner in recovery	0.033	0.301	-0.563	0.631	.930	.150
Gender	-0.227	0.122	-0.470	0.016	.067	.088
Location	-0.147	0.140	-0.508	0.214	.295	.113
Age	-0.003	0.006	-0.017	0.011	.587	.138
Race/ethnicity	-0.043	0.162	-0.460	0.375	.293	.100
Medication type	0.252	0.135	-0.013	0.517	.062	.075
Intercept
Intercept	0.152	0.127	−0.100	0.405	.234	-

Notes: N = 131 relationships (disclosure relationships), n (clusters) = 106 participants. Unstandardized results are reported (b). The dependent variable is the Commitment to Sobriety change score from Time 1 to Time 2. All continuous variables were mean centered (significant effects bolded). b = unstandardized regression coefficient; SE b = standard error for the regression coefficient; CI = confidence interval; LL = lower limit; UL= upper limit; adj. p = Benjamini–Hochberg adjusted p values.

Significant interactions were found supporting the second hypothesis. There was a significant two-way interaction between social support and closeness (b = 0.091, p = .037, adj. p = .038) on participant change in commitment to sobriety. Thus, participants reported an increase in their commitment to sobriety when disclosing to a close partner who was highly supportive. In contrast, participant commitment to sobriety decreased when disclosing to a partner with whom they had used substances (history of joint substance use) and was also either supportive (social support by history of joint substance use interaction b = −0.407, p = .018, adj. p = .025) or close (closeness by history of joint substance use interaction b = −0.239, p = .042, adj. p = .050).

Discussion

Findings from the current study provide insight into the nuanced role of social support and relationship variables (i.e., closeness and history of joint substance use) among people in recovery from OUD within the context of disclosure events/relationships. On the surface, preliminary results suggest that disclosing to and receiving social support from a close disclosure partner is likely beneficial for one's commitment to OUD recovery. However, a deeper examination of these disclosure events revealed a further distinction of note. Namely, that the directionality of social support's impact on commitment to sobriety depends on both closeness and a history of substance use with the intended disclosure partner. Findings from the interaction between social support and closeness suggest that disclosing to a supportive and close disclosure partner can help to increase participant commitment to sobriety and thus may be beneficial for recovery. However, significant interactions between (a) history of joint substance use and social support and (b) history of joint substance use and closeness suggest that disclosing to a partner with whom there has been joint substance use can have a negative effect on commitment to sobriety. Taken together, these results suggest that optimal targets for OUD disclosure are partners with whom there is a close relationship and no history of joint substance use.

These conclusions confirm that aspects of social support facilitate OUD recovery (Buchanan & Latkin, 2008) and further research by designating which types of relationships may be the most appropriate targets for disclosure. Aligning with previous research, results suggest that the quality and nature of interpersonal relationships can have an influential impact on recovery-related outcomes and motivations (Polcin & Korcha, 2017). Confirming Berkman and colleagues’ conceptual model (Berkman et al., 2000), closeness and history of joint substance use played a critical role in how sources of social support influenced commitment to sobriety. Results also align with the Disclosure Process Model, which underscores the importance of disclosure on health outcomes and suggests that disclosure is a complex process through which multiple aspects of disclosure interact, including interactions between social support and disclosure partner relationship characteristics, to impact important health outcomes (Chaudoir & Fisher, 2010).

The current study also provides empirical evidence for what substance use disorder specialists have long maintained (Buchanan & Latkin, 2008): A supportive relationship with someone with whom one has previously engaged in problematic substance use can be adversely influential on recovery efforts. Therefore, healthcare providers should consider encouraging patients to pursue social support from and consider disclosing to those without a shared history of problematic substance use (Buchanan & Latkin, 2008; Dennis et al., 2014; Dingle et al., 2015).

The results of this study should be considered in the light of its strengths and limitations. Although the data were collected over two time points, the longitudinal nature of the research extended over a short period of time. Additional research is needed to examine the impacts of social support over a more extended period in order to encompass broader recovery trajectories. Further, more work is needed to ensure that these findings generalize to a larger and more diverse participant population, including marginalized populations who may not be currently accessing substance use disorder treatment. Likewise, the findings from this study suggest that a deeper understanding is warranted to fully comprehend and use the many facets of social support systems for people in OUD recovery.

As such, future research should investigate more relationship-level variables and their effects on multiple types of social support (e.g., functional support) and recovery outcomes, including disclosure experiences with partners who are in stable recovery versus currently engaging in substance use or assessing disclosure reactions specific to recovery behaviors (e.g., disclosing to someone who encourages abstinence). Such research may extend how aspects of social support influence recovery behaviors and could inform recommendations to enhance recovery outcomes (Cloud & Granfield, 2001; Granfield & Cloud, 1999).

It is also of note that participants may choose to disclose to those they perceive as more socially supportive. More research is needed to understand how individuals living with OUD select disclosure partners. In addition, future research should consider additional features of recovery among more diverse samples of people in recovery, such as history of OUD treatment, the time enrolled in the current treatment, and how specific forms of social support may affect recovery. Future research should continue to advance disclosure research surrounding people living with OUD, including by characterizing the content of disclosures, as disclosure is an intimate interpersonal facet of recovery that has the potential to influence both positive and negative outcomes.

The findings provide a foundation for actionable strategies that can be used to assist interventions for people living with OUDs.Although many recovery-related interventions already aim to enhance social support, the current findings can provide a deeper understanding of the nuance of support systems. As such, interventions aimed at cultivating certain sources of social support, such as close others with no history of joint substance use, may enhance recovery-related outcomes. An example of a relevant intervention that can be updated and further adapted to a variety of disclosure relationships was created by Kidorf and colleagues (1997). This behavioral intervention was designed to encourage people who are receiving methadone treatment to include supportive and abstinent family members in their treatment process and may be used to further enhance disclosure experiences (Kidorf et al., 1997). Likewise, the current results have implications for people who are considering OUD disclosure. Namely, disclosure to close others where there is a history of joint substance use may undermine a patient's commitment to sobriety.

Taken together, the findings of this study support that disclosure is a uniquely intimate act that can influence recovery, and more specifically close others can play an important role by impacting future recovery-related characteristics. Such positive forms of social support should be encouraged to aid in patients’ recovery efforts, and patients should be advised as to which types of relationships may best facilitate their recovery. As such, optimal strategies warrant additional research to identify disclosure partners who have the potential to enhance versus disrupt long-term OUD recovery outcomes.

Conflict-of-Interest Statement

The authors declare that they have no conflicts of interest. They alone are responsible for the content and writing of the article.