Financial incentives and real-time adherence monitoring to promote daily adherence to HIV treatment and viral suppression among people living with HIV: a pilot study

Abstract

Background:

Interventions to promote medication adherence and viral suppression are needed among HIV-positive individuals. We aimed to determine the feasibility, acceptability, and preliminary impact of daily financial incentives linked to real-time adherence monitoring among treatment-experienced individuals.

Methods:

At an HIV clinic in Philadelphia, we conducted a pilot randomized trial among treatment-experienced HIV-positive adults with unsuppressed viral loads (>400 copies/ml). Participants randomized to the intervention group were eligible for daily lottery-based financial rewards dependent on antiretroviral therapy (ART) adherence, measured by a wireless-enabled electronic pill bottle. Participants also received a financial incentive for achieving viral suppression at 3 months. The control group received the standard of care. We measured acceptance and feasibility through follow-up survey at 3 months, viral suppression at 3 months, and adherence.

Results:

Among 29 participants, 28 (93%) completed 3-month follow-up and 24 (83%) completed a 3-month laboratory visit. Electronic pill bottles were highly acceptable to participants, with most strongly agreeing that they worked well, were reliable, and easy to use. Among those who received the intervention, 77% were very satisfied with their experience. Among those who completed the 3-month lab visit, viral suppression was achieved by 40% in the intervention group and 29% in the control group. ART adherence ≥80% was achieved by 36% and 25% in the intervention and control groups, respectively.

Conclusions:

Daily financial incentives coupled with real-time adherence monitoring is a promising strategy to support ART adherence among HIV-positive individuals who are not virally suppressed. This novel approach warrants testing in a larger trial.

INTRODUCTION

Adherence to antiretroviral therapy (ART) is critical to achieving viral suppression and lowering the risk of morbidity and mortality among people living with HIV (PLWH).^1–3 However, in the US, less than 65% of PLWH have achieved viral suppression.⁴ Interventions to improve adherence to ART are therefore an important clinical and public health priority.

One intervention to promote medication adherence is the use of financial incentives. By offering a monetary incentive such as a cash payment or lottery-based reward for intended behavior, such an intervention can motivate health behaviors typically associated with immediate costs, such as inconvenience or time, and delayed benefits. Financial incentives have been shown to be effective in promoting multiple health behaviors, including smoking cessation, weight loss, and uptake of HIV prevention services.^5–8 However, financial incentives to improve HIV treatment outcomes have shown mixed results.⁹ In a trial of incentives linked to viral suppression conducted in the US, viral suppression rates were modestly higher among those who received incentives compared with standard-of-care.¹⁰ In addition, a trial in Uganda found that lottery prizes tied to clinic visits and daily pill bottle openings increased ART adherence.¹¹ On the other hand, in a trial of PLWH with substance use, offering financial incentives did not affect viral suppression.¹² Similarly, a trial of financial incentives for viral suppression among PLWH in Uganda had no impact on viral suppression.¹³ The efficacy of financial incentives to promote health behaviors may thus be sensitive to multiple variables, such as how incentives are framed, timed, and implemented in the context of a desired outcome.

Delivering timely incentives may be required to promote daily adherence needed for viral suppression, yet few studies have linked incentives to real-time monitoring devices that enable frequent rewards for adherence behavior.^14,15 Wireless pill bottles can provide real-time monitoring and can be coupled with daily incentives.^16,17 This may frame financial incentives as a daily adherence-dependent reward, promoting medication adherence habit formation. Therefore, we examined the acceptability, feasibility, and preliminary impact of a daily lottery-based incentive coupled with real-time feedback to promote ART adherence among treatment-experienced PLWH.

METHODS

Study setting and population

Individuals were screened for eligibility between June and September 2019 at the Partnership Comprehensive Care Practice in Philadelphia, a Ryan White-funded clinic that provides comprehensive services to over 1,700 PLWH. Eligible individuals had a diagnosis of HIV established in 2017 or earlier; ≥18 years old; prescribed first- or second-line ART medications; and had unsuppressed viral loads>400 copies/ml for the two most recent lab readings. ART regimens were determined to be fully active by the clinical investigators. Individuals were excluded from the study if they were pregnant; did not have stable residence; were unable to provide consent; were a non-English speaker; or had a cognitive impairment as determined by the clinical investigators. In addition, individuals were excluded with high medication burden, including those with more than 5 additional medications other than ART medications; prescribed insulin; or prescribed maraviroc or enfuvirtide.

Recruitment and Randomization

Eligible participants were identified through a monthly review of patient records. Potential participants were reviewed by study staff, including a clinical nurse, clinical pharmacist, and three HIV care providers who were also study investigators. Investigators then approached participants for study participation. After providing informed consent, participants completed a baseline survey that included questions about socio-demographics, mental health, HIV history, medication adherence, and risk preferences.

Block randomization was conducted by the Way to Health platform (WTH).¹⁸ Participants were randomized 1:1 to an intervention and control group, with half of the control group receiving an electronic pill bottle (Digital Media Technologies, LLC). The electronic pill bottle measures adherence through bottle openings, capped at one opening daily, sending the data wirelessly to the WTH platform. All participants received $100 at enrollment and an additional $100 for completing an end-of-study lab visit three months after enrollment.

Intervention group

To implement the lottery-based incentives for adherence, a two-digit number was randomly assigned to all intervention participants, with a random two-digit number drawn daily on the WTH platform. If participants opened the bottle the previous day, they were entered into a daily lottery with a chance to win $10 or $100 if one or two digits matched with the lottery number, respectively. If participants were fully adherent, expected average lottery earnings were $2.55 per day, with an expected total earning of $250 over three months. Participants also received an additional $100 if viral suppression (VL<400 copies/mL) was achieved at three months. Compensation was provided to participants through a virtually replenished cash gift card. Participants received one of four daily feedback messages, depending on adherence the day prior and their lottery number. Through use of lotteries instead of fixed incentives, the intervention leveraged behavioral economics principles of regret aversion to amplify the power of financial incentives.^19,20

Control group

Participants in both the intervention group and the control group received the standard of care for adherence counseling and outreach from clinic staff. Half the control group were given electronic pill bottles to monitor medication adherence, but did not receive any communication or feedback regarding adherence.

Follow-up

Study participants were followed for three months. All participants were also asked to complete a follow-up survey and a lab visit at three months. The follow-up survey included questions about medication adherence, including self-reported measures, and about participants’ experience in the study.

Study Measures

Feasibility and acceptability measures were assessed through follow-up survey at three months. The clinical outcome of interest was viral suppression (viral load<400 copies/ml) at three months and was reported using per protocol and intention-to-treat analyses. A Wang 95% confidence interval for the difference between groups was estimated using two-sided 0.05 significance level.²¹ Confidence intervals were presented around the primary clinical outcome only given small sample size as a pilot study and inability to specify precision. Secondary clinical outcomes included medication adherence, measured with electronic pill bottles; self-reported adherence; completion rates of the follow-up lab visit; and participant usage of the pill bottles. Analyses were conducted using Stata 15.1 and R version 3.5.2.

Ethics Statement

This study was approved by the Institutional Review Board of the University of Pennsylvania with a reliance agreement signed by Drexel University. The trial was registered on clinicaltrials.gov with the identifier NCT03995745.

RESULTS

Between June and September 2019, approximately 1,700 patients were screened for eligibility. The majority were excluded due to being virally suppressed, and 255 were assessed for eligibility, 226 were excluded for not meeting inclusion criteria or declining to participate, and 29 participants were ultimately enrolled in the study. Fourteen were randomized to the intervention group and 15 to the control group, with 8 participants in the control group receiving the pill bottle.

Participants’ median age was 44 years, 12 self-identified as women, 16 as men, and 1 as transgender (Table 1). Most were single (62%) and identified as Black (76%). By comparison, about 64% of PLWH in Philadelphia are Black.²² The majority had completed high school (79%), had an annual income of less than $24,999 (86%), and 31% had experienced homelessness in the last 12 months.

Table 1:

Baseline characteristics of study participants (N = 29)

Characteristic	Overall (%)	Intervention (N=14)	Control (N=15)
Gender
Male	16 (55.2)	6 (42.9)	10 (66.7)
Female	12 (41.4)	7 (50)	5 (33.3)
Transgender	1 (3.4)	1 (7.1)	0 (0)
Age, mean (SD)	44.2 (10.5)	46.8 (8.4)	41.4 (12.0)
Marital Status
Single	18 (62.1)	8 (57.1)	10 (66.7)
Married	2 (6.9)	2 (14.3)	0 (0)
Unmarried partner	7 (24.1)	3 (21.4)	4 (26.7)
Divorced	2 (6.9)	1 (7.1)	1 (6.7)
Race/Ethnicity
Hispanic	3 (10.3)	1 (7.1)	2 (13.3)
Non-Hispanic Black	22 (75.9)	11 (78.6)	11 (73.3)
Non-Hispanic White	1 (3.4)	1 (7.1)	0 (0)
Non-Hispanic Other	3 (10.3)	1 (7.1)	2 (13.3)
Education
Some high school or less	6 (20.7)	2 (14.3)	4 (26.7)
High school graduate	14 (48.3)	9 (64.3)	5 (33.3)
Some college	8 (27.6)	2 (14.3)	6 (40)
College graduate	1 (3.4)	1 (7.1)	0 (0)
Household Income
Less than $24,999	25 (86.2)	12 (85.7)	13 (86.7)
$25,000-$49,999	3 (10.3)	1 (7.1)	2 (13.3)
$50,000-$74,999	1 (3.4)	1 (7.1)	0 (0)
Homelessness in last 12 months
Yes	9 (31.0)	3 (21.4)	6 (40)
No	20 (69.0)	11 (78.6)	9 (60)
Year of HIV diagnosis
Before 1995	9 (31.0)	6 (42.9)	3 (20.0)
1995–2004	13 (44.8)	4 (28.6)	9 (60.0)
2005–2014	6 (20.7)	4 (28.6)	2 (13.3)
2015–2019	1 (3.4)	0 (0)	1 (6.7)
Year of ART initiation
Before 1995	8 (27.6)	5 (35.7)	3 (20.0)
1995–2004	10 (34.5)	3 (21.4)	7 (46.7)
2005–2014	9 (31.0)	5 (35.7)	4 (26.7)
2015–2019	2 (6.9)	1 (7.1)	1 (6.7)
Missed days of ART, past 2 weeks
0	11 (37.9)	5 (33.3)	6 (42.9)
1–3	10 (34.5)	5 (33.3)	5 (35.7)
4–10	8 (27.6)	5 (33.3)	3 (21.4)

All measures were obtained through self-report during the baseline survey.

Using the pill bottles and obtaining feedback was very acceptable to study participants in the intervention group (Table 2). Most participants strongly agreed that the pill bottle worked well (83%), was easy to use (75%), and was reliable (83%). Among those who received the intervention, 77% indicated they were very satisfied with receiving messages from the study team and from their experience.

Table 2.

End of study feasibility and acceptability measures

		N (%)
Question	Overall	Intervention	Control
Received a study issued, wireless pill bottle	20 (71.4)	12 (92.3)	8 (53.3)
The wireless pill bottle worked well.†
Not at all/Not really	0 (0)	0 (0)	0 (0)
Neutral/Somewhat	6 (30.0)	2 (16.7)	4 (50.0)
Very much	14 (70.0)	10 (83.3)	4 (50.0)
It was easy to learn to use my pill bottle.†
Not at all/Not really	0 (0)	0 (0)	0 (0)
Neutral/Somewhat	5 (25.0)	3 (25.0)	2 (25.0)
Very much	15 (75.0)	9 (75.0)	6 (75.0)
The pill bottle was reliable.†
Not at all/Not really	0 (0)	0 (0)	0 (0)
Neutral/Somewhat	5 (25.0)	2 (16.7)	3 (37.5)
Very much	15 (75.0)	10 (83.3)	5 (62.5)
Getting help/support about using the pill bottle was simple.†
Not at all/Not really	0 (0)	0 (0)	0 (0)
Neutral/Somewhat	5 (25.0)	2 (16.7)	3 (37.5)
Very much	14 (70.0)	10 (83.3)	4 (50.0)
Overall, how satisfied are you with your experience receiving messages from the study (text, phone, email)?
Not at all/Not really	2 (7.1)	1 (7.7)	1 (6.7)
Neutral/Somewhat	2 (7.1)	1 (7.7)	1 (6.7)
Very much	21 (75.0)	10 (76.9)	11 (73.3)
Overall, how satisfied are you with your experience in the study?
Not at all/Not really	0 (0)	0 (0)	0 (0)
Neutral/Somewhat	5 (17.9)	2 (15.4)	3 (20.0)
Very much	22 (78.6)	10 (76.9)	12 (80.0)

Note: Responses may not add to 100% due to missing responses.

^†Denominator is number of participants who answered “Yes” to “Received a study issued, wireless pill bottle.”

Regarding follow-up, 28/29 (93%) individuals completed the follow-up survey, and 10/14 (71%) in the intervention group and 14/15 (93%) in the control completed a lab visit at three months. Of those with available lab data, 4/10 (40%) in the intervention group and 4/14 (29%) in the control group achieved the primary endpoint of VL<400 (estimated difference of 11.4%; exact 95% CI, −28.7, 49.9). Under an intention-to-treat analysis, 4/15 (27%) in the intervention group and 4/14 (29%) in the control group achieved a VL<400 (estimated difference of 1.9%; exact 95% CI −34.5, 30.7). All participants in the intervention group received the incentives as planned, with mean lottery winnings of $170 with a standard deviation of $115.

Medication adherence data from the electronic pill bottles was available for all 14 participants in the intervention and all 8 participants in the control with the pill bottle. Adherence declined in both intervention and control groups during the study period, with mean adherence of 47% in the control group and 62% in the intervention group. In the intervention group, 5/14 (36%) had medication adherence ≥80% over the three months of the study, compared with 2/8 (25%) in the control group. According to self-reported data at the end of the study, 33% in the control group reported no missed days and 57% reported 1–3 missed days of ART over the past two weeks, compared with 39% and 54% in the intervention group, respectively.

DISCUSSION

This pilot study evaluated the acceptability, feasibility, and preliminary impact of real-time adherence monitoring linked to daily financial incentives to support ART adherence among treatment-experienced PLWH. We found usage of the wireless pill bottle to be highly acceptable among study participants, with most participants noting the bottle worked very well, was easy to use, and reliable. Although our study was not designed to determine efficacy, we observed higher levels of adherence and viral suppression in the intervention group. This real-time adherence feedback strategy may be valuable for promoting daily ART adherence in HIV treatment.

This intervention was particularly innovative in leveraging daily adherence monitoring to deliver daily text messages. This “automated hovering” approach uses behavioral economics principles and automates the provision of daily adherence support.²³ While text messages have been used to improve ART adherence,^24,25 prior interventions have not used a real-time feedback strategy, and have sent messages less frequently with the concern that reminders that are too frequent may lose efficacy.^9,26 On the other hand, daily reminders may be advantageous when informed by prompt, real-time, adherence feedback.^27,28 We found that three-quarters of study participants were very satisfied with receiving daily messages, which suggests that a more frequent feedback approach may be acceptable when tailored messaging is provided.

Our pilot study also used wireless pill bottle data to deliver lottery-based financial incentives. Most individuals who received the intervention were highly satisfied with their experience and none were dissatisfied. Previously, incentives to improve ART adherence were less frequently deployed because incentives were contingent upon clinic visits or viral load monitoring, potentially limiting the effectiveness of incentives for behaviors that require frequent reinforcement.^10,11,13 In addition, a lottery-based financial incentive may have advantages in behavior change compared to fixed financial reward systems, particularly for individuals who are more risk tolerant.^19,20

We note several limitations. First, our sample size was small, and we were unable to evaluate the efficacy of the intervention. Furthermore, our exclusion criteria limit the generalizability of our findings, and a larger trial may use less restrictive eligibility criteria. Second, pill bottle openings are a proxy for adherence, and we cannot know if the medication was ingested. Additional strategies for measuring adherence, including therapeutic drug monitoring,²⁹ could inform future adherence interventions leveraging incentives. Third, we did not evaluate the implementation or cost-effectiveness of our intervention, which may have an impact on whether the intervention is adopted if found to be efficacious. Qualitative data could provide additional insight into participants’ experience with the intervention. Finally, because the study was conducted for only three months, we were unable to evaluate longer term endpoints. One notable strength of our study was that it was conducted among treatment-experienced PLWH with unsuppressed viremia, a key group for whom adherence interventions are urgently needed.

In summary, daily financial incentives coupled with real-time ART adherence monitoring and feedback was acceptable and feasible among treatment-experienced PLWH. Preliminary results from this novel intervention indicate potential for increased adherence and viral suppression among PLWH, and warrant evaluation in a larger randomized controlled trial.