DSM-5 substance use disorder symptom clusters and HIV antiretroviral therapy (ART) adherence

Abstract

This study examines self-reported 30-day antiretroviral therapy (ART) adherence among 101 people living with HIV and substance use disorders (SUD) in New York City in terms of Diagnostic and Statistical Manual – 5^th Edition (DSM-5) SUD symptom clusters: impaired control, social impairment, risky use, and pharmacological criteria. Overall, 60.4% met DSM-5 criteria for stimulant, 55.5% for alcohol, 34.7% for cannabis, and 25.7% for opioid SUD. Of the 76 participants with a current ART prescription, 75.3% reported at least 90% 30-day adherence. Participants with vs. without alcohol SUD were significantly less likely to report ART adherence (64.3% vs. 88.2%, p=.017). Endorsement of social impairment significantly differed among adherent vs. non-adherent participants with alcohol SUDs (74.1% vs. 100%, p=.038) and with opioid SUDs (94.1% vs. 50.0%, p=.040). Understanding specific SUD symptom clusters may assist providers and patients in developing strategies to improve ART adherence.

Introduction

Substance use disorders (SUD) and HIV are common synergistic epidemics (Hartzler et al., 2017). Among people living with HIV (PLWH), SUD may reduce linkage to and retention in HIV care (Metsch et al., 2009) and adherence to antiretroviral therapy (ART) (Arnsten et al., 2002; Lucas, 2011). Nevertheless, universal ART is recommended for all PLWH regardless of SUD (DHHS, 2016; WHO, 2012). Although providers are urged to identify and intervene on potential SUD adherence challenges using harm reduction strategies (WHO, 2007, 2012), they may be reluctant to prescribe ART to patients with SUD (Volkow & Montaner, 2010).

Understanding how patients’ specific substance use patterns and precise manifestation of SUD are associated with ART non-adherence (Bonn-Miller, Oser, Bucossi, & Trafton, 2014; Malta, Magnanini, Strathdee, & Bastos, 2010) may assist providers in developing tailored harm reduction strategies (Lehavot et al., 2011). Some providers already informally assess for SUD-related symptoms linked to non-adherence (e.g., lack of daily routines) (Campbell, Tross, Wolff, Weaver, & Des Jarlais, 2018). Most studies have focused on more global assessment of SUD (Bonn-Miller et al., 2014; Hinkin et al., 2007; Lehavot et al., 2011) and how the amount (Chibanda, Benjamin, Weiss, & Abas, 2014; Hendershot, Stoner, Pantalone, & Simoni, 2009) and frequency (Mellins et al., 2009) of substance use influence adherence, without examining specific SUD symptomatology.

Purpose

Recognizing the opportunity for developing harm reduction strategies based on specific Diagnostic and Statistical Manual of Mental Disorders, 5^th Edition (DSM-5) SUD symptoms to improve ART adherence, the current study aimed to examine both: a) if ART adherence differed among those with and without specific SUD (i.e., alcohol, cannabis, opioids, and sedative/hypnotic/anxiolytics) and b) if endorsing specific clusters of DSM-5 SUD symptoms differed among ART-adherent and non-adherent PLWH.

Materials & methods

Participants

This quantitative analysis was part of a larger, mixed methods evaluation of universal ART implementation in New York City (NYC) (Campbell et al., 2016) among 101 PLWH recruited 2014 to 2017. Eligible participants: a) reported problem substance use (any illicit drug use in the past year and/or heavy drinking in the past month [at least four daily drinks if assigned male; at least three daily drinks if assigned female at birth] (DHHS, 2015)); b) were living with HIV; c) were able to speak and understand English; and d) lived in NYC.

Procedures

Potential participants were recruited from NYC Department of Health and Mental Hygiene (DOHMH) sexual health clinics, in a hospital-based drug detoxification (detox) unit, and through flyers and word-of-mouth. Verbal consent was obtained for a brief screening interview either in person or over the phone. If eligible, participants gave written consent for five face-to-face interviewer-administered 90-minute computer-assisted interviews, at baseline and four follow-up assessments six months apart. Institutional Review Board approval was obtained from the Icahn School of Medicine at Mount Sinai.

Measures

Current (past-year) alcohol, cannabis, opioid, and sedative/hypnotic/anxiolytic use disorders were assessed using DSM-5 criteria (American Psychiatric Association, 2013; Grant et al., 2016).

For each SUD, symptoms were classified into four symptom clusters (Table 1): impaired control (e.g., using more than planned; cravings); social impairment (e.g., use interfered with family, school, or work); risky use (e.g., driving while intoxicated); and pharmacological criteria (e.g., tolerance or withdrawal) (National Institute on Drug Abuse, 2016). Endorsement was defined as indicating at least one symptom within a cluster.

Table 1.

DSM-5 Criteria for Substance Use Disorders (SUD) in Four Symptom Clusters (National Institute on Drug Abuse, 2016)

In the past 12 months, presence of: 2–3 symptoms = mild SUD; 4–5 symptoms = moderate SUD; 6 or more symptoms = severe SUD
Impaired Control	1. Drinking/using for more, or longer, than intended 2. More than once, wanted to cut down or stop drinking/using, or tried to, but couldn’t. 3. Spent a lot of time drinking/using, or being sick or getting over other aftereffects 4. Wanted a drink/to use so badly, couldn’t think of anything else
Social Impairment	5. Found that drinking/using—or being sick from drinking/using—often interfered with taking care of home or family, or caused job troubles, or school problems 6. Continued to drink/use even though it was causing trouble with family or friends 7. Given up or cut back on activities that were important or interesting, or gave pleasure, in order to drink/use
Risky Use	8. More than once gotten into situations while or after drinking/using that increased chances of getting hurt (such as driving, swimming, using machinery, walking in a dangerous area, or having unsafe sex) 9. Continued to drink/use even though it was making you feel depressed or anxious or adding to another health problem, or after having had a memory blackout
Pharmacological Criteria	10. Had to drink/use much more than you once did to get the effect you want, or found that usual number of drinks/amount of substance had much less effect than before 11. Found that when the effects of alcohol/drug were wearing off, you had withdrawal symptoms, such as trouble sleeping, shakiness, restlessness, nausea, sweating, a racing heart, or a seizure, or sensed things that were not there

Past-30-day ART adherence was assessed via self-report using the Visual Analog Scale (VAS) of zero to 100% adherence (Giordano, Guzman, Clark, Charlebois, & Bangsberg, 2004). Adherence was defined as ≥90% within the past month across all prescribed ART medications (Yager et al., 2017).

Statistical analyses

Descriptive analyses examined baseline differences in adherence among participants with and without each SUD. For each SUD-positive subsample, we compared baseline differences in endorsement of the four SUD symptom clusters among adherent and non-adherent participants using χ²or Fisher’s exact tests. All hypothesis tests were two-sided using 5% level of significance and performed in SAS® 9.3 [SAS institute, Cary, NC].

Results

Demographic characteristics are presented in Table 2. A significantly lower proportion of participants with vs. without alcohol SUD (64.3% vs. 88.2%, p=.017) reported ART adherence (Table 3). Among those with alcohol SUD and prescribed ART, the proportion endorsing social impairment was significantly lower among adherent vs. non-adherent participants (74.1% vs. 100%; p=.038; Table 4). Among those with opioid SUD and prescribed ART, however, the proportion endorsing social impairment was significantly higher among adherent vs. non-adherent participants (94.1% vs. 50.0%; p=.040; Table 4).

Table 2.

Demographic Characteristics (N = 101 PLWH)

	n	%
Recruitment Location
DOHMH Sexual Health Clinics	38	37.6
Detox Unit	54	53.5
Flyers/Word-Of-Mouth	9	8.9
Gender
Male	87	86.1
Female	12	11.9
Transgender Woman	2	2.0
Race/Ethnicity
Black/African American, Non-Hispanic	51	50.5
Hispanic/Latino (any race)	36	35.6
White, Non-Hispanic	13	12.9
Other/Multiracial, Non-Hispanic	1	1.0
Employed full-time*
Yes	27	26.7
No	73	72.3
Housing
Own Apartment/House	60	59.4
Someone Else’s Apartment/House	20	19.8
Unstable Housing (e.g., shelter)	21	20.8
	M (SD)	Range
Age	43.9 (13.2)	21–65
Years of Education	13.1 (2.4)	6–19

^*Total does not add up to 100% due to missing responses

Table 3.

Substance Use Disorders (SUD) & Antiretroviral Therapy (ART) Adherence (N = 101 PLWH)

		All PLWH	Currently Prescribed ART		Currently Prescribed ART & Adherent		Currently Prescribed ART & Non-Adherent
			n	% of Total N	n	% of adherent on ART	n	% of non-adherent on ART
			76	75.25	57	75.00	19	25.00
					Substance Use Disorders by ART adherence and non-adherence
Substance Use Disorders	n	% of N	n	% Prescribed ART	n	% adherent/on ART	n	% non-adherent/on ART	P-value
Alcohol									.017
SUD	56	55.5	42	75.0	27	64.3	15	35.7
No SUD	45	44.6	34	75.6	30	88.2	4	11.8
Cannabis									.888
SUD	35	34.7	25	71.4	19	76.0	6	24.0
No SUD	66	65.4	51	77.3	38	74.5	13	25.5
Opioids									.885
SUD	26	25.7	23	88.5	17	73.9	6	26.1
No SUD	75	74.3	53	70.7	40	75.5	13	24.5
Sedatives/Hypnotics									.636ⱡ
SUD	7	6.9	6	85.7	4	66.7	2	33.3
No SUD	94	93.1	70	74.5	53	75.7	17	24.3
Stimulants									.051ⱡ
SUD	61	60.4	50	82.00	34	68.0	16	32.0
No SUD	40	39.6	26	65.0	23	88.5	3	11.5

^ⱡResults are based on Fisher’s Exact tests

Table 4.

ART Adherence by Substance Use Disorder Symptom Clusters (N = 101 PLWH)^*

	With SUD		With SUD & Prescribed ART		With SUD, Prescribed ART, & Adherent		With SUD, Prescribed ART, & Non-Adherent
				Endorsement of Clusters
Alcohol Use Disorder	n = 56		n = 42		n = 27		n = 15
DSM-5 SUD Symptom Clusters	n	% Endorsed	n	% Endorsed	n	% Endorsed	n	% Endorsed	P-value
Impaired Control	52	92.9	38	90.5	25	92.6	13	86.7	.608ⱡ
Social Impairment	45	80.4	35	83.3	20	74.1	15	100.0	.038ⱡ
Risky Use	41	73.2	31	73.8	20	74.1	11	73.3	1.00ⱡ
Pharmacological Criteria	35	62.5	22	52.4	12	44.4	10	66.7	.167
Cannabis Use Disorder	n = 35		n = 25		n = 19		n = 6
DSM-5 SUD Symptom Clusters	n	% Endorsed	n	% Endorsed	n	% Endorsed	n	% Endorsed	P-value
Impaired Control	32	91.4	25	100.0	19	100.0	6	100.0	1.00#
Social Impairment	22	62.9	15	60.0	11	57.9	4	66.7	1.00ⱡ
Risky Use	29	82.9	19	76.0	17	89.5	5	83.3	1.00ⱡ
Pharmacological Criteria	22	62.9	16	64.0	12	63.2	4	66.7	1.00ⱡ
Opioid Use Disorder	n = 26		n = 23		n = 17		n = 6
DSM-5 SUD Symptom Clusters	n	% Endorsed	n	% Endorsed	n	% Endorsed	n	% Endorsed	P-value
Impaired Control	26	100.0	23	100.0	17	100.0	6	100.0	1.00#
Social Impairment	22	84.6	19	82.6	16	94.1	3	50.0	.040ⱡ
Risky Use	25	96.2	22	95.7	17	100.0	5	83.3	.261ⱡ
Pharmacological Criteria	24	92.3	22	95.7	17	100.0	5	83.3	.261ⱡ
Stimulant Use Disorder	n = 61		n = 50		n = 34		n = 16
DSM-5 SUD Symptom Clusters	n	% Endorsed	n	% Endorsed	n	% Endorsed	n	% Endorsed	P-value
Impaired Control	61	100.0	50	100.0	34	100.0	16	100.0	1.00#
Social Impairment	61	100.0	50	100.0	34	100.0	16	100.0	1.00#
Risky Use	60	98.5	49	94.0	33	97.1	16	100.0	1.00ⱡ
Pharmacological Criteria	51	83.6	42	84.0	28	82.4	14	87.5	1.00ⱡ

^*Sedative/hypnotic symptom clusters by adherence were not examined given that only seven participants met criteria for sedative/hypnotic SUD.

^ⱡP-values are based on Fisher’s Exact tests.

^#P-values are 100% by definition (comparing 100% endorsed vs. 100% endorsed)

Discussion

We observed two key findings: first, a significantly smaller proportion of those with vs. without alcohol SUD reported adherence to ART. Second, significant associations between social impairment and ART adherence were observed among those with alcohol or opioid SUD.

Although this study involved a small sample size and was exploratory by nature, our findings may hold potential for informing clinical practice among HIV medical providers (Bonn-Miller et al., 2014; Lehavot et al., 2011; Malta et al., 2010). Those with alcohol SUD and non-adherent (vs. adherent) to ART, but those with opioid SUD and adherent (vs. non-adherent) were more likely to endorse social impairment symptoms. Social impairment (e.g., problems with family or employment) may be associated with ART non-adherence in people with alcohol SUD given periods of binge use and disruption to daily routines (Campbell et al., 2018; Cofrancesco Jr et al., 2008; Reback, Larkins, & Shoptaw, 2003). People with opioid SUD, however, may still maintain ART adherence even in when experiencing opioid-related social impairment. As providers have previously reported, individuals with opioid SUD often have routine patterns of use that can be leveraged to support HIV medication adherence (for example, taking ART medications right before a specific time each day that opioids are used) (Campbell et al., 2018).

Given that discrete symptoms of SUD appear to distinctly impact ART adherence, providers may need to draw on different strategies to support patients’ adherence. For patients with alcohol SUD and social impairment, social support interventions may be effective (Lehavot et al., 2011), including linkage to peer support groups, peer recovery coaches, or individual mental health counseling. Providers could also offer adherence-related support (Lehavot et al., 2011)—such as directly observed ART and one-on-one adherence counseling (Thompson et al., 2012)—for those who may have lost support from family or friends as a result of their alcohol use. For patients with opioid SUD, adherence strategies may involve harm reduction techniques, such as drawing on opioid injection routines as possible anchors for taking ART (Campbell et al., 2018). This may prove particularly effective for patients who are not ready to decrease or stop their use (Altice, Springer, Buitrago, Hunt, & Friedland, 2003).

HIV primary care providers could benefit from continuing education and training in addiction (Campbell et al., 2018; Montague et al., 2015) to facilitate more comprehensive assessment, SUD treatment referrals, and use of innovative collaborative care models to assist in managing care for patients living with HIV and SUD. Integrating SUD treatment within HIV primary care settings is an ideal approach to closing gaps in linkage, retention, and adherence to both SUD treatment and HIV medical care (Lucas et al., 2010).

Limitations

There are several limitations to the current study. First, adherence for this analysis was collected via self-report; however, self-report measures such as the VAS appear to be significantly associated with HIV viral load (Giordano et al., 2004). Second, the majority of participants reported poly-substance use (90.0%) and disorders (61.1%), but we were unable to explore how multiple SUDs may have jointly influenced ART adherence given the small sample size. Further, we could not assess how combined symptomatology across multiple SUD may have influenced adherence outcomes. Because of the exploratory nature, small sample size, and lack of variability in SUD symptom cluster endorsement (particularly among opioid SUD) additional research would be necessary to confirm the direction of observed associations. A larger, more diverse sample of PLWH with SUD could permit multivariable analytic approaches and more robust consideration of SUD subtypes, as well as greater variability in SUD severity. Finally, there may be additional constructs beyond DSM-5 SUD clusters that could also be clinically meaningful for ART adherence.

Conclusions

Despite the aforementioned limitations, our findings suggest the importance of assessing discrete SUD symptom clusters to gain a richer, more patient-centered understanding of how SUD may be associated with ART non-adherence. PLWH with SUD may maintain adherence even if achieving abstinence from substances is not their desired outcome. Precise adherence barriers (and strategies for facilitating adherence) may vary by SUD type and symptom cluster. Linking PLWH to SUD treatment can be challenging for multiple reasons, including: ambivalence about seeking treatment; scarcity of readily available, accessible, or acceptable services; and a dearth of services sensitive to the diverse lived experiences of PLWH (Durvasula & Miller, 2014). As such, considering DSM-5 SUD symptom clusters to determine the most effective harm reduction strategies tailored to the unique needs of PLWH with SUD may contribute to improved HIV care and may support HIV providers in following the universal ART treatment guidelines with all patients.