Attitudes Toward Pharmacy-based HCV/HIV Testing Among People Who Use Drugs in Rural Kentucky

Abstract

Purpose:

Rural areas of the US have experienced outbreaks of Human Immunodeficiency Virus (HIV) and Hepatitis C Virus (HCV) infections among people who use drugs (PWUD). Pharmacy-based interventions may play a crucial role in prevention and entry into care, especially when traditional health care access is limited. The willingness of rural PWUD to use pharmacies for HIV/HCV-related services remains unknown. The purpose of this study was to describe the factors associated with the perceived likelihood of participating in free pharmacy-based HIV and HCV testing among PWUD living in rural Kentucky.

Methods:

Baseline data from the CARE2HOPE study in 5 Appalachian counties in eastern Kentucky were used. Participants were recruited using respondent-driven sampling and completed interviewer-administered surveys. Guided by the Andersen and Newman Framework of Health Services Utilization, we examined distributions and correlates of items regarding willingness to participate in free pharmacy-based HIV/HCV testing using logistic regression. Analyses included individuals who reported being HIV (N=304) or HCV (N=185) negative.

Findings:

Seventy-five percent of PWUD reported being “very likely” to participate in free pharmacy-based HIV testing and 80% for HCV testing. Two factors were associated with being less willing to participate in free HIV testing: PWUD who previously tested for HIV (OR: 0.47, CI: 0.25–0.88) and PWUD who obtained a high school diploma or equivalent compared to those who completed less (OR: 0.50, CI: 0.26–0.99).

Conclusion:

Free pharmacy-based HIV and HCV testing was invariably acceptable among most of the rural PWUD in our sample, suggesting that pharmacies might be acceptable testing venues for this population.

Injection-related infectious diseases, including Human Immunodeficiency Virus (HIV) and Hepatitis C Virus (HCV), have increased in the US, especially in rural areas among people who inject drugs (PWID).^1,2 Between 2006 and 2012, the largest reported increases in HCV cases occurred in predominantly rural areas of central Appalachia, including the states of Kentucky, Tennessee, Virginia, and West Virginia.³ In the US (2017), the incidence of new HCV cases attributed to PWID was 86.6%.⁴ In 2015, rural Scott County, Indiana, experienced the fastest spreading HIV outbreak ever documented in the US.⁵ The outbreak was linked to misuse of the prescription opioid oxymorphone via injection, and approximately 84% of those with HIV also had HCV.⁵ In Kentucky (2017), 19.8% of the new HIV diagnoses among males and 38.2% among females were attributed to injection drug use.⁴

Early detection of HIV/HCV facilitates a person’s access to care, allows more patient treatment autonomy, increases the effectiveness of anti-retroviral medications, and prevents further transmission.^6,7 As such, expanding opportunities for screening and early detection of HIV/HCV is an important public health response. Efforts to engage pharmacists as partners in public health for people who use drugs (PWUD) have primarily focused on increasing access to naloxone and facilitating non-prescription syringe sales.^8–10 For example, in Kentucky, pharmacists have had statutory authority to initiate the dispensing of naloxone since 2015 and to sell syringes without a prescription for over 2 decades.^11,12 Similar efforts to promote pharmacy-based HIV/HCV testing have commenced and have been shown effective in reaching at-risk individuals.^13,14 Pharmacists’ involvement includes medication adherence counseling, HIV/HCV treatment education, and identification and mitigation of drug-drug interactions.^15–18 Expanded pharmacy services for PWID may decrease injection-related risk for infection.¹⁹

Pharmacies may be accessible locations for HIV/HCV-related services in rural areas. They are effective in providing health services that target health promotion and disease prevention in rural areas.²⁰ Though location to primary health care remains an issue for rural populations, an estimated 70% of rural residents live within 15 miles of a pharmacy.^21–23 Increasing access to pharmacists may improve engagement of at-risk populations in the cascades of care needed to prevent and treat HIV and/or HCV.²⁴

Despite the documented risk of HIV/HCV in rural areas among PWID and the benefits of pharmacist-provided services, there is little information on PWID willingness to engage in HIV/HCV testing at pharmacies in rural areas. The purpose of this brief report, guided by the Andersen and Newman Framework of Health Services Utilization,²⁵ is to identify the characteristics of PWUD living in rural Kentucky that are associated with the likelihood of using free pharmacy-based HIV and HCV testing.

Methods

Study Design

In the Kentucky Communities and Researchers Engaging to Halt the Opioid Epidemic (CARE2HOPE) project, a sample of PWUD (N=325) was created to inform the development of evidence-based responses to prevent and treat the consequences of opioid-related syndemics, including HCV and HIV, in 5 rural counties in eastern Kentucky. According to the 2013 Rural–Urban Continuum Codes and the 2010 US Census, all 5 counties were classified as nonmetropolitan and/or rural.^26,27 Eligible participants were ≥18 years old, resided in 1 of the 5 Appalachian Kentucky counties, and have either used opioids to get high or injected any drug to get high in the past 30 days.

Participants were recruited through respondent-driven sampling (RDS) between February 2018 and November 2019. The seeds for the RDS were recruited from 2 sources: (1) a sample of well-connected individuals derived from an online survey study previously conducted in the region, and (2) community outreach, including cookouts, advertisements, and community partners. The RDS method details can be found in a previous study.²⁸ This study was approved by the University of Kentucky IRB, and all participants consented.

Measures

We created 2 dependent variables to examine the likelihood of participating in 1) HIV testing and 2) HCV testing at the pharmacy. These variables were derived from Likert-type question items (eg, “How likely would you be to participate in free HIV [HCV] testing if it was offered by a local pharmacy?”). Responses were on a 4-point scale ranging from “very unlikely” to “very likely.” Since the vast majority (75% and 80%) of the responses to the HIV and HCV dependent variable, respectively, were “very likely,” we dichotomized for modeling (Table 1).

Table 1.

Summary Statistics for Respondents by Question Type

	HIV (n=304)a		HCV (n=185)a
Outcome Variable	n	%	n	%
Likely to participate in free testing
Very unlikely	11	3.6	5	2.7
Somewhat unlikely	12	4.0	8	4.3
Somewhat likely	53	17.4	24	13.0
Very likely	228	75.0	148	80.0
Predisposing Factors
Age (years)
18–24	25	8.2	15	8.1
25–34	124	40.8	71	38.4
35–44	101	33.2	59	31.9
45–64	54	17.8	40	21.6
Highest Level of Education
Less than high school	96	31.6	59	31.9
High school diploma or GED	137	45.1	89	48.1
Some college and above	71	23.4	37	20.0
Genderb
Male	168	55.3	100	54.0
Female	136	44.7	85	46.0
Ethnicityb
Not Hispanic or Latino	303	99.7	185	100.0
Race
Self-identified as White	296	97.4	178	96.2
Self-identified as anything else	8	2.6	7	3.8
Cohabitationc
No	174	57.2	101	54.6
Yes	130	42.8	84	45.4
Enabling Factors
Homeless in past six months
No	192	63.2	118	63.8
Yes	112	36.8	67	36.2
Currently have health insurance
No	51	16.8	33	17.8
Yes	253	83.2	152	82.2
Total Income in past 30 days
<$1000	217	71.4	125	67.6
≥$1000	87	28.6	60	32.4
Needs Factors
History of testing for HIV
No	104	34.2	-	-
Yes	200	65.8	-	-
History of testing for HCV
No	-	-	83	44.9
Yes	-	-	102	55.1
Shared needles, syringes, cookers, cottons, or rinse water in past 30 days
No	177	58.2	125	67.6
Yes	127	41.8	60	32.4

^aMissing values in any variable excluded the observation.

^bOther response categories had very few responses.

^cThose who responded “married” or “living with partner” were categorized as cohabiting. All other responses were characterized as no cohabitation.

The independent variables of interest were chosen based on the Andersen and Newman Framework of Health Services Utilization.²⁵ Briefly, this model posits that 3 domains of characteristics may affect a person’s access to and use of health services: predisposing factors, enabling factors, and needs factors. The predisposing components existed prior to the onset of the illness (eg, age and sex). Enabling factors describe the logistical aspect of obtaining care and using services. The needs factors include perceived risk from the person and the evaluated judgment from the professional.

In this analysis, predisposing factors were age, gender, education, race/ethnicity, and cohabitation; enabling factors were homelessness, health insurance, and income; and needs factors were prior HIV/HCV testing and sharing injection equipment (Table 1). The prior access to care factor was determined from the question “Have you ever been tested for [“HIV, the AIDS virus,” Hepatitis C] before today?”

Before beginning analyses, we categorized age and income because they were non-normally distributed and re-categorized education and marital status due to wide response distributions. Participants with missing data were excluded (including individuals who responded “don’t know” on select variables). Because 99.8% of the sample self-identified as non-Hispanic White, race/ethnicity was not included in the regression analysis.

Analysis

Descriptive statistics were used to summarize key variables. We examined the correlates of HIV testing and of HCV testing in separate regression models. The 2 samples were not mutually exclusive. Participants with a previous positive test for HIV (n=1) or HCV (n=128) were excluded from each analytic sample. Before performing logistic regression, bivariate analysis using Pearson chi-square was used to assess the association between each predictor and each outcome variable. Although none of the variables reached statistical significance in their association with the outcome (alpha ≥ 0.05), the variables were still included in both models due to a priori theoretical criteria. Additionally, variables in both multivariate models did not exhibit multicollinearity (conditional indices < 30). SAS v9.4 (SAS Institute, Inc., Cary, NC) was used for all analyses.²⁹

Results

Descriptive Statistics

HIV testing model.

After excluding those who tested positive for HIV (n=1, 0.31%) and participants with missing data (n=20, 6.15%), 304 people were included in this model. The majority of the respondents were aged 25–34 years (40.8%), completed a high school diploma or equivalent (45.1%), were male (55.3%), lived alone (57.2%), were homeless (63.2%), had health insurance (83.2%), and had a total income of less than $1000 in the past 30 days (71.4%). The majority had tested for HIV previously (65.8%) and had shared injection equipment within the past 30 days (58.2%; Table 1).

Most respondents chose “very likely” to the question pertaining to the likelihood of participating in free pharmacy-based HIV testing (75.0%). After adjusting for considered covariates, those with a high school diploma or equivalent were less likely to be willing to participate in free pharmacy-based HIV testing than those who had less education (OR: 0.50, CI: 0.26–0.99; Table 2). Those tested for HIV previously were also less likely to be willing to participate in free pharmacy-based HIV testing (OR: 0.47, CI: 0.25–0.88). All other covariates were statistically insignificant. In supplemental analyses, we examined participants that had never tested. Those with no history of HIV testing had an estimate suggesting higher odds of endorsing “very likely” to participate in free testing. However, sample sizes were too small to draw confident conclusions.

Table 2.

Adjusted Odds Ratios (AOR) and 95% Confidence Intervals (CI) for PWUD Reporting to Be “Very Likely” to Participate in Free HIV/HCV Testing at Pharmacies

	HIV (n=304)			HCV (n=185)
Demographic	AOR	95% CI		AOR	95% CI
Age
25–34 vs. 18–24	1.22	0.45	3.33	0.79	0.16	3.36
35–44 vs. 18–24	1.59	0.56	4.53	0.96	0.21	4.33
45–64 vs. 18–24	1.46	0.46	4.58	1.03	0.21	4.98
Education Status
High school diploma or GED vs. Less than high school	0.50	0.26	0.99	0.54	0.22	1.34
Some college and above vs. Less than high school	0.58	0.25	1.31	0.76	0.23	2.51
Gender
Female vs. Male	1.02	0.58	1.82	0.93	0.41	2.12
Cohabitationa
Yes vs. No	0.75	0.43	1.31	0.55	0.25	1.20
Homeless
Yes vs. No	1.14	0.64	2.04	1.95	0.81	4.69
Have insurance
Yes vs. No	1.49	0.72	3.06	1.87	0.68	5.17
Income
≥$1000 vs. <1000	0.76	0.42	1.44	1.27	0.54	3.00
History of testing for HIV
Yes vs. No	0.47	0.25	0.88	-	-	-
History of testing for HCV
Yes vs. No	-	-	-	0.91	0.41	1.99
Shared needles, syringes, cookers, cottons, or rinse water in past 30 days
Yes vs. No	1.08	0.62	1.88	0.91	0.41	2.02

^aThose who responded “married” or “living with partner” were categorized as cohabiting. All other responses were characterized as no cohabitation.

HCV testing model.

After excluding those who tested positive for HCV (n=128, 39.38%) and those who had incomplete responses (n=12, 3.69%), 185 people were included in the model. The sample had a similar demographic and behavioral profile as those included in the HIV model (Table 1).

The vast majority of respondents chose “very likely” to the question pertaining to the likelihood of participating in free pharmacy-based HCV testing (80.0%). None of the covariates were significantly associated with the likelihood of participating in free pharmacy-based HCV testing.

Discussion

Free, pharmacy-based HIV and HCV testing was acceptable among the vast majority of PWUD in our study (75.0% and 80.0%, respectively). These numbers are similar to those reported by Brewer and colleagues, who found that 71% of pharmacy patron respondents born between 1945 and 1965 and at risk for HCV were willing to receive HCV screening in a community pharmacy.³⁰ To our knowledge, this is the first paper to describe the likelihood of PWUD to participate in free testing at pharmacies. We found little evidence suggesting that any of the health utilization factors enabled or impeded perceived participation in free HIV or HCV testing in rural pharmacies. The utilization of pharmacies as a site to provide free HIV and HCV testing may be effective in reaching these high-risk populations and may be a cost-effective public health response.^31,32

We found that PWUD who previously tested for HIV expressed a lower perceived likelihood of participating in free HIV testing in a pharmacy. This finding contrasts with the Andersen and Newman Framework of Health Services.²⁵ A few hypotheses may be worth exploring in the future. First, if prior testing experiences have been poor (eg, stigmatizing), respondents may be concerned that future testing experiences might be challenging, regardless of the provider. Although our sample did not permit a robust analysis of PWUD new to HIV testing, this group should remain a focus for prevention efforts. Second, a prior negative HIV test may have provided a false sense of security about on-going risky behavior. Finally, respondents may be comfortable with their own provider (many had insurance in our sample) and feel no need to visit the pharmacy for free testing services.

PWUD in this rural sample who had a high school diploma or equivalent were less likely than PWUD who did not to report being likely to participate in free HIV testing at a pharmacy. Thus, PWUD with less formal education may be more likely to accept free HIV testing offered at pharmacies. The role of education, prior testing behavior and future testing preferences among rural PWUD remains unclear but warrants further research.

Limitations

Several limitations in this study are noteworthy. First, those who previously tested positive for HIV or HCV skipped the question that was used to determine the outcome variable. This is a limitation because those who test positive for HCV might be vulnerable to re-infection after they are cured, and therefore, pharmacy-based HCV testing remains relevant for them. Second, the findings may not be generalizable to rural PWUD in other regions. Research in other rural areas is needed to understand acceptability of pharmacy-based HIV/HCV testing among PWUD in these communities. Finally, we used willingness as a measure of likelihood, which we acknowledge is an imperfect measure of likelihood.

Our study has several strengths. First, interviewers were trained to minimize the potential for recall and social desirability bias. Second, this analysis was developed using a theory-based approach. The framework for health utilization allowed us to choose key variables that may affect an individual’s choice in expressing likelihood of visiting pharmacies to receive testing services.

Future Directions

Pharmacy-based interventions may play a crucial role in preventing HIV/HCV infections, especially in rural areas where traditional health care access points are limited. By increasing geographic coverage of HIV/HCV related services, PWUD have a higher chance of practicing harm reducing activities that reduce HIV and HCV transmission and protect their own health, as studies have shown that areas with poor proximity to injection-related HIV prevention activities have higher risk for HIV.^33–35 The addition of HIV/HCV testing at pharmacies may increase spatial access of HIV/HCV services to PWUD, especially in rural areas where health care is limited.

Further assessments of barriers to implementing evidence-based harm reduction services in pharmacies could provide greater insights in developing an effective mode of delivery for PWUD in Kentucky and other rural areas.³⁶ Our findings suggest that the perceived likelihood of participating in free pharmacy-based HIV or HCV testing services is not associated with multiple predisposing, enabling, or needs factors. Overall, PWUD are likely to participate in free pharmacy-based HIV and/or HCV testing regardless of their individual characteristics.