Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Mar 1.
Published in final edited form as: Drug Alcohol Depend. 2024 Jan 22;256:111099. doi: 10.1016/j.drugalcdep.2024.111099

Mobile Medication Adherence Platform for Buprenorphine (MAP4BUP): A Phase I Feasibility, Usability and Efficacy Pilot Randomized Clinical Trial

Sterling M McPherson 1,2,3, Crystal L Smith 1,2,3, Luke Hall 1,2,3, André Miguel 1,2,3, Theresa Bowden 2,3,5, Abigail Keever 1,2,3, Alex Schmidt 1,2,3, Katie Olson 6, Nicole Rodin 2,3,4, Michael G McDonell 1,3, John M Roll 1,3,5, Jeff LeBrun 7
PMCID: PMC10923156  NIHMSID: NIHMS1963639  PMID: 38306822

Abstract

Background/Aim:

Poor medication adherence is one of the main barriers to the long-term efficacy of buprenorphine/naloxone (BUP/NAL). The aims of this pilot investigation were to examine if a Bluetooth-enabled pill cap and mobile application is a feasible, usable tool for increasing BUP/NAL adherence among people with an opioid use disorder.

Methods:

This pilot randomized clinical trial (RCT; total n=41) lasted 12 weeks and was conducted in two office-based BUP/NAL provider locations in Spokane, WA and Coeur d’Alene, ID from January 2020 to September of 2021 with an 11-month gap due to COVID-19. Patients receiving BUP/NAL who consented to participate were randomized to receive the pill cap device (PLY group; n=19) or a service as usual (SAU group; n=22) group that included an identical but inactive cap for their bottle. The PLY group received reminders via text and voice, and the support of a “helper” (e.g., friend) to monitor pill cap openings.

Results:

Most participants in PLY group found the device both feasible (92.86%) and usable (78.57%). Most participants liked using the device (92.86%) and were satisfied with the device (85.71%). While not statistically different from one another, medication adherence per the Medication Possession Ratio was 75% in the SAU group and 84% in the PLY group. Pill cap openings were significantly higher in the PLY group with an average of 91.8 openings versus the SAU group's average of 56.7 (p<0.05).

Conclusion:

The devices was feasible, usable, and patients had high levels of satisfaction. The device was associated with increased pill openings.

Keywords: mobile-technology, medication adherence, buprenorphine, naloxone, opioid use disorder, medications for opioid use disorder

1. INTRODUCTION

In the last decade, an opioid epidemic has been acknowledged by six federal agencies including the National Institute on Drug Abuse, Substance Abuse and Mental Health Services Administration, Centers for Medicare and Medicaid Services, Food and Drug Administration, Centers for Disease Control, and the Office of National Drug Control Policy. (Food and Drug Administration, 2014; Volkow et al., 2014) The opioid use crisis in the US has resulted in nearly 650,000 overdose deaths since 1999, with more than half of them occurring since 2016.(2021; 2022; Stringfellow et al., 2022; U.S. Department of Health and Human Services Substance Abuse and Mental Health Services Administration; U.S. Drug Enforcement Administration, 2021) The continued increases in opioid associated deaths has occurred despite the federal government spending billions of dollars to increase access to medications for opioid use disorder (MOUD), including buprenorphine/naloxone (BUP/NAL).(U.S. Department of Health and Human Services Substance Abuse and Mental Health Services Administration) BUP/NAL can assist in the reduction of overdose deaths, decrease opioid misuse, decrease usage of costly health services and increase outpatient services usage.(Fudala et al., 2003; Gopaldas et al., 2023; Liao et al., 2023; Mattick et al., 2014; Sordo et al., 2017)

However, poor medication adherence remains one of the main barriers to the long-term efficacy of BUP/NAL. We define medication adherence to BUP/NAL using the medication possession ratio, similar to what others have done (MPR; pills taken/pills prescribed)(Saloner et al., 2017; Velez et al., 2022; Velez et al., 2021; Williams et al., 2023). In a large multicenter randomized trial comparing Methadone to BUP/NAL for treatment-seeking individuals with an OUD (n=1,267), treatment retention to BUP/NAL was 46% and 12.3% at the end of the trial (24th week) and at the 60-month follow-up assessment, respectively.(Hser et al., 2016; Hser et al., 2014) Other previous studies have found adherence levels between 32% and 43% in patients(Ronquest et al., 2018; Ruetsch et al., 2017; Tkacz et al., 2014) and between 21% and 41% in samples of Medicaid patients (Kinsky et al., 2019; Ronquest et al., 2018; Samples et al., 2018). This represents a sub-optimal situation in need of solutions to ameliorate BUP/NAL non-adherence.

There are a several patient-level correlates of non-adherence (e.g., age, co-occurring psychiatric condition, environment of living with someone who is using illicit substances) that have been investigated (Dayal and Balhara, 2017; Fareed et al., 2014; Rosenblum et al., 2011). There is also research on social desirability in the context of substance use (Krasnoff, 1976; Yoshino and Kato, 1995; Zemore, 2012) has found higher social desirability to be associated with greater engagement with treatment. However, to date there remains surprisingly little evidence in the literature on how to increase adherence to BUP/NAL (Peter et al., 2023). Based on these data suggesting that self-image impacts treatment response, implications are that treatment retention might be maximized by capitalizing on this through engaging people in treatment with more interpersonal interactions leading to positive behaviors related to promoting social desirability. Given that BUP/NAL is a schedule III drug, improving medication adherence and treatment attendance is critical to decreasing relapse and diversion.(Hadland et al., 2018; Lofwall and Walsh, 2014)

The intent of the current investigation was to examine the ability of an inexpensive (i.e., $40, without any insurance reimbursement), Bluetooth-enabled pill bottle cap and associated mobile application (app; i.e., Pillsy, an Optimize Health product; see www.pillsy.com for more information) designed to monitor and increase BUP/NAL adherence with the assistance of a designated “helper” (e.g., friend, family member) to increase medication adherence, improve treatment engagement and reduced use of illicit substances. Pillsy acts like a digital medication coach, providing reminders using a mobile application, text messages, and automated phone calls.(Park et al., 2022) The platform automatically tracks timing of doses and sends intelligent reminders to create a unique feedback loop, which allows the developers to constantly optimize reminder messages to increase adherence and meet the needs of the person in treatment. More specifically, the app will learn which messages are more or less effective for specific people at specific times, and over time, the app will learn which messages to send and when in order to optimize adherence behavior that is detectable (e.g., opening the Pill cap) for a given patient.

Since Pillsy only nominally increases the cost of oral BUP/NAL treatment and providers can bill for monitoring time (CPT code 99091), it is a potentially attractive solution to patients, providers, and payers. However, there are no studies available that have systematically investigated this technology’s impact on medication adherence in this population. This pilot investigation reports on the use of this device in a phase I randomized clinical trial among patients receiving BUP/NAL for an OUD. This trial has relatively low power to detect a signal for efficacy, which was intentional given that the primary targets were feasibility, acceptability and usability. The primary aim of the Mobile Medication Adherence Platform for Buprenorphine (MAP4BUP) pilot randomized clinical trial (RCT) was to determine whether patients randomized to receive the Pillsy device (PLY) found it feasible and usable, and to determine whether it could improve adherence to BUP/NAL, compared to those in the services as usual (SAU) group.

2. METHOD

2.1. Study Overview

This RCT was conducted in Washington state and Idaho in two office-based BUP/NAL providers located in Spokane, WA and Coeur d’Alene, ID from January 2020 to September of 2021 (with an 11-month gap due to COVID-19 related study closures). While the trial opened in January of 2020, no participants were successfully recruited prior to COVID restrictions being put in place. Thus, while there was a delay in recruitment, no participant’s data collection was interfered with due to COVID. The goal of this RCT was to determine whether the use of the PLY intervention was usable and feasible, and if those randomized to the PLY group could improve adherence to BUP/NAL compared to the SAU group. The PLY group received automated reminders via text and phone call after the app was initially set up with a research coordinator to help them understand how to program different doses into the app to tailor the reminder system to their needs. This group also included the recruitment of a “helper” (i.e., anyone who they chose that would help them in their treatment) who could monitor pill cap openings through an app on their own phone and remind the person in treatment using the device as needed, or check-in with them, etc. Notably, the SAU group received an identical cap to use for their medication bottle, but it only tracked closings and openings. It did not come with an app setup to monitor and remind the device user to take their medication (i.e., did not provide reminders or any other messaging), nor did they receive access to a “helper” through the app like the PLY group did. This allowed the team to ask SAU participants about their experience with the pill cap on the feasibility, acceptability and usability questions for a comparison between the two groups. There were four occasions during the 12-week period where participants met with a research coordinator to complete a schedule of assessments: the initial screening, at 1 week, 6 weeks, and 12 weeks. Study procedures were classified as exempt by the Institutional Review Board (see NCT04656899 for additional detail).

2.2. Participants

The study was conducted in partnership with Ideal Option outpatient treatment centers that routinely prescribe BUP/NAL and provides behavioral counseling in conjunction with BUP/NAL. Participants were identified either by the outpatient clinicians, nurse managers, or medical assistants, and referred to the onsite study coordinators, or saw fliers in the waiting room, self-identified and requested an appointment with the coordinator. Those that were interested in participating were required to have been initiated on BUP/NAL at least 30 days prior to the initial baseline visit. There was a total of n=83 patients screened, n=41 of which were randomized and included in the final, intent-to-treat sample. Participants were between 18 and 60 years of age (mean of 36.9 years). Inclusion criteria were: Self-reported OUD diagnosis and received BUP/NAL treatment for their OUD at Ideal Option; owned a working smartphone; ability to read and speak English; and could identify one study “helper” who could agree to participate (i.e., as described above, this "helper” could be a friend, family member, provider, caregiver, etc.).

Exclusions included:

Taking opioids for the treatment of chronic, non-acute pain; being prescribed a chronic opioid agonist treatment; possessing any pending legal issues that could affect the participant’s participation; having an allergy or previous serious reaction to BUP/NAL; currently being treated for an SUD other than OUD or nicotine. Notably, while taking BUP/NAL required a diagnosis of moderate or severe OUD, some participants in this study self-reported symptoms consistent with mild OUD likely due to the treatment they were receiving at Ideal Option. Having a mild OUD did not disqualify a participant as their provider was still treating their OUD with BUP/NAL. People were excluded if they were taking opioids for the treatment of chronic, non-acute pain because these patients often have a more complicated set of prescribing patterns for BUP/NAL and other opioids. This could have made accurate assessment of adherence to BUP/NAL more challenging in a pilot study due to the increased heterogeneity in prescription amount, type and duration. See Figure 1 for a complete CONSORT diagram of how participants moved through the clinical trial over time.

Figure 1.

Figure 1.

Open in a new tab

CONSORT Diagram for Mobile Medication Adherence Platform for Buprenorphine (MAP4BUP) Phase I Pilot Randomized Clinical Trial.

2.3. Procedures

After initial screening for inclusion and exclusion criteria and ensuring that the potential participant understood their role in the study, participants were required to provide written informed consent before any other study activities occurred.

2.3.1. Randomization

Participants were centrally registered in the REDCap database system maintained by the Analytics and PsychoPharmacology Laboratory and randomly assigned 1:1 to the PLY or SAU group. We used a permuted block randomization and stratified on age, sex, and severity of the OUD (mild, moderate, severe) based on DSM 5 criteria.

2.3.2. Study Interventions

Ideal Option was responsible for all BUP/NAL prescriptions, dose adjustments, behavioral counseling, and other aspects of each participant’s treatment plan. Optimize Health, the creator and distributor of the Pillsy bottle cap, through the trained research coordinators, was responsible for training each participant on use of the Pillsy smart bottle cap, the mobile application, and the SMS texting platform. Participants in the PLY group received both the Pillsy cap with the Pillsy application that sent messages and reminders. These reminders included a light on the smart pill cap that would flash if a participant did not take their medicine at the correct time. If the bottle was not opened within 20 minutes, an SMS text reminder would be sent to the participant’s phone. Should the participant not open the bottle within 60 minutes, an automated phone call would be sent. The “Pillsy Helper” feature would prompt the helper via a text message notification one hour after a missed dose, which would allow the helper to contact the participant and offer motivation or setup a brief appointment with someone from their care team or provide other support that is consistent with remaining adherent to their treatment plan. Participants in the SAU group received an “inactive” Pillsy smart cap that tracked openings but did not provide any reminders or messages, nor did this group receive the benefit of a helper. Notably, participants in the SAU group were informed that their smart cap would track all openings.

2.3.3. Measures and Outcomes

During the visits, the participants were requested to complete a series of assessments that collected information on demographics, feasibility, usability, acceptability, smoking history, alcohol and opioid use history, and quality of life. The primary outcomes were feasibility (i.e., how easy the device is to use), adherence to BUP/NAL assessed by medication possession ratio (MPR; pills taken/pills prescribed)(Saloner et al., 2017; Velez et al., 2022; Velez et al., 2021; Williams et al., 2023), retention (defined as a continuous variable: number of days in treatment, and a binary variable: completed treatment), non-prescribed opioid use, and pill cap openings.

The system's feasibility, acceptability and usability, among other items, were assessed with a number of questions, including some questions inspired by the System Usability Scale (SUS).(Brooke, 1996) Notably, while our questions were inspired by the SUS, we included twice as many questions as the SUS and we tailored our questions to be specific to the adherence system under investigation. This is also why we do not report a single measure score, but instead include more granular data at the question level. Response options for feasibility, acceptability and usability range from 1= strongly agree, 4 = neutral, to 7= strongly disagree. Options 1-3 represented strongly agree, moderately agree, and agree. These three categories were collapsed for the purposes of presentation in Table 2 to indicate the percentage of participants who agreed with a given statement.

Table 2.

Percentage of patient agreement on device system feasibility, acceptability, and usability from baseline to the end of the study.

Device System
Questions		 Week 1 Week 6 Week 12
SAU PLY SAU PLY SAU PLY
1. Overall, I am satisfied with how easy it is to use this system 100.00% 91.67% 89.47% 82.35% 85.71% 78.57%
2. It was simple to use this system 100.00% 91.67% 89.47% 82.35% 92.86% 85.71%
3. I can effectively take my medication using this system 90.91% 91.67% 84.21% 88.24% 92.86% 92.86%
4. I feel comfortable using this system 90.91% 91.67% 100.00% 82.35% 85.71% 92.86%
5. It was easy to learn to use this system 100.00% 91.67% 89.47% 88.24% 100.00% 85.71%
6. I believe I became more adherent using this system 81.82% 91.67% 68.42% 64.71% 71.43% 78.57%
7. I can easily view my medication history at any time 100.00% 91.67% 89.47% 87.50% 85.71% 100.00%
8. The information (such as online help, onscreen messages) provided with this system is clear 72.73% 100.00% 84.21% 88.24% 85.71% 92.86%
9. It is easy to find the information I needed 90.91% 91.67% 84.21% 88.24% 85.71% 92.86%
10. The information provided for the system is easy to understand 100.00% 91.67% 94.74% 88.24% 92.86% 92.31%
11. The organization of information on the system screens is clear 81.82% 83.33% 89.47% 88.24% 85.71% 92.86%
12. The interface of this system is pleasant 90.91% 75.00% 94.74% 94.12% 92.86% 85.71%
13. I like using the interface of this system 90.91% 58.33% 84.21% 64.71% 78.57% 92.86%
14. This system has all the functions and capabilities I expect it to have 90.91% 75.00% 78.95% 82.35% 85.71% 92.31%
15. Overall, I am satisfied with this 100.00% 83.33% 89.47% 76.47% 78.57% 85.71%
Open in a new tab

Note: SAU = service as usual group, PLY = pillsy group.

MPR was calculated as the sum of days’ supply for all fills in treatment divided by total days in treatment, and then multiplied by 100. This measure was then examined as a continuous measure and as a binary measure (>79.9% adherence was coded as adherent, <80.0% was coded as non-adherence). Retention was also examined as a continuous variable (number of days in the study and still in treatment) and a binary variable (completed the study or not). We also analyzed repeated measures urine analysis drug test (UA’s) for non-prescribed opioid use and stimulant use. Lastly, we examined the number of total pill cap openings across the two groups as a measure of treatment engagement. The MPR and pill cap openings are measures of adherence and treatment engagement, respectively. They are orthogonal measures derived from different data sources, i.e., the MPR comes from prescription data obtained from the clinic where the patients were seen, whereas pill cap openings is obtained directly from the system they were using as part of this RCT. The MPR is a direct, commonly used method of measuring adherence, as noted above. A score of 100% would indicate perfect adherence regardless of prescription type, dose, or other characteristics. Pill cap openings cannot be directly tied to adherence, i.e., people can open their pill bottle 100 times, for example, and never take a single dose. For the purposes of this RCT, we view pill cap openings as a measure of treatment engagement that is useful, but not definitive, and the results should be interpreted relatively (i.e., comparing the two group to each other) and not as absolute measures. Data to calculate the MPR and UA data was obtained from the partnering clinics with the consent of participants at baseline.

2.3.4. Statistical Analysis

We computed means and standard deviations for continuous variables, and frequencies and percentages for categorical variables (see Table 1). To test if participants varied across baseline characteristics, we used between-subjects t-tests for continuous variables and chi-square tests for categorial variables. For feasibility and usability, we analyzed several questions related to these outcomes descriptively, but we were specifically interested in questions three (i.e., "I can effectively take my medication using this system") and six ("I believe I became more adherent using this system"; see Table 2), respectively for these outcomes, as they are related to medication adherence. We were also specifically interested in the outcomes for questions 13 (i.e., "I like using the interface of this system") and 15 (i.e., "Overall, I am satisfied with this system") as they are most closely related to the acceptability of using this system.

Table 1.

Demographics and descriptive statistics of key outcomes (n=41) for Mobile Medication Adherence Platform for Buprenorphine (MAP4BUP) Phase I Pilot Randomized Clinical Trial.

SAU PLY Total p
Total N (%) Category/Descriptor 22 (53.7%) 19 (46.3%) 41 (100%)
Age Mean (SD) 37.7 (11.0) 36.1 (8.1) 36.9 (9.7) 0.597
Biological Sex Male 12 (54.5) 7 (36.8) 19 (46.3) 0.412
Female 10 (45.5) 12 (63.2) 22 (53.7)
Employment Unemployed 13 (59.1) 11 (57.9) 24 (58.5) 1.000
Employed 9 (40.9) 8 (42.1) 17 (41.5)
Prefer not to answer 0 (0.0) 0 (0.0) 0 (0.0)
Marital Status Single 15 (68.2) 5 (26.3) 20 (48.8) 0.019
Married or cohabitating 3 (13.6) 9 (47.4) 12 (29.3)
Divorced or Separated 4 (18.2) 5 (26.3) 9 (22.0)
Race American Indian 0 (0.0) 2 (10.5) 2 (4.9) 0.154
White 20 (90.9) 16 (84.2) 36 (87.8)
Black 2 (9.1) 0 (0.0) 2 (4.9)
Prefer not to answer 0 (0.0) 1 (5.3) 1 (2.4)
BUP/NAL Initiation Not Reported 1 (4.5) 1 (5.3) 2 (9.8) 0.703
0-30 days ago 8 (36.4) 5 (26.3) 13 (31.7)
21-60 days ago 1 (4.5) 2 (10.5) 3 (7.3)
61-90 days ago 1 (4.5) 2 (10.5) 3 (7.3)
91-120 days ago 11 (50.0) 9 (47.4) 20 (48.8)
Baseline DSM Diagnosis Mild 5 (22.7) 9 (47.4) 14 (34.1) 0.139
Moderate 2 (9.1) 0 (0.0) 2 (4.9)
Severe 15 (68.2) 10 (52.6) 25 (61.0)
Continuous MPR Mean (SD) 75.2 (33.4) 83.9 (28.4) 79.2 (31.1) 0.379
Binary MPR < 80 8 (36.4) 4 (21.1) 12 (29.3) 0.465
> 80 14 (63.6) 15 (78.9) 29 (70.7)
Continuous Retention Mean (SD) 52.6(36.4) 71.3 (28.2) 61.0 (35.3) 0.465
Binary Retention Did Not Complete 10 (45.5) 6 (31.6) 16 (39.0) 0.412
Completed Study 12 (54.5) 13 (68.4) 25 (60.9)
Pill Cap Openings Mean (SD) 56.7 (54.3) 91.8 (69.3) 72.5 (63.2) 0.080
Opioid Positive UAs Mean (SD) 1.7 (4.4) 2.1 (7.1) 1.9 (5.7) 0.839
Stimulant Positive UAs Mean (SD) 3.1 (5.0) 1.9 (4.2) 2.6 (4.6) 0.397
Open in a new tab
*

Note: PLY=Group of participants assigned to receive the Pillsy intervention; SAU=Group of participants assigned to receive Services As Usual. MPR=Medication Possession Ratio, OUD=opioid use disorder. BUP/NAL=buprenorphine/naloxone.

A linear regression was used for the continuous MPR outcome variable, and a logistic regression for the binary MPR outcome variable. For retention, a linear regression was used with number of days in treatment as the outcome variable, and a logistic regression was used with the binary indication of whether the patient completed the study as the outcome variable. A linear regression was also used with number of pill cap openings during the study as the outcome. Generalized estimating equations (GEE) were used to analyze the repeated urine analyses (UAs) over time. Providers determined when they felt it was appropriate to require their patients (our participants) to submit UAs based on several factors, including but not limited to length of time in treatment, concerns about medication diversion, and illicit substance use. We used UAs our participants submitted (by request of their provider) in our analysis. Notably, this is not a primary outcome of this trial, and have no reason to believe that either of our randomized groups was more or less likely to have a UA requested. We first created two drug categories: stimulants and opioids. The stimulant category included amphetamines, cocaine, and methamphetamines. The opioid category included opiates, oxycodone, norfentanyl, monoacetylmorphine, codeine, desmethyltapentadol, desmethyltramadol, fentanyl, hydrocodone, hydromorphone, meperidine, morphine, norhydrocodone, noroxycodone, oxymorphone, tapentadol, and tramadol. For each category, we then created a binary variable that indicated if the patient had at least one positive UA for the week. The patient was coded negative if they had no positive UAs for the week.

After the initial, unadjusted models were run for our efficacy outcomes, all outcomes were then analyzed again with age, sex, and baseline DSM score included as covariates. There was no missing data in any of the predictors and there was minimal missing data for one of the outcomes (i.e., retention, 5% or less; see Table 1), and no missing data for the other outcomes. An alpha threshold of 0.05 was used to determine statistical significance.

3. RESULTS

3.1. Descriptive Statistics

Of the 41 participants enrolled in this study, 22 were randomized to the SAU group and 19 were randomized to the PLY group (see Figure 1). The average age of the sample was 37 years of age and 54% of the sample’s biological sex was female. Most of the sample was unemployed (59%), were white (88%) and had a severe OUD at baseline (61%). At baseline, the only variable that differed significantly between the two treatment groups was marital status (p=0.019; see Table 1).

3.2. Primary Outcomes

Most participants found the cap to be feasible (see Table 2; Question 3 "I can effectively take my medication using this system": SAU = 92.86%, PLY = 92.86%) and usable (Question 6 "I believe I became more adherent using this system": SAU = 71.43%, PLY = 78.57%) at the end of the study period (Week 12). Additionally, most participants found the cap to also be acceptable (Question 13 "I like using the interface of this system": SAU = 78.57%, PLY = 92.86%; Question 15 "Overall, I am satisfied with this system": SAU = 78.57%, PLY = 85.71%) at Week 12. Notably, the SAU participants were reporting on an inactive cap that did not provide reminders and other supports that the PLY group received. Regarding the device's efficacy at increasing BUP/NAL adherence, despite descriptive differences between the two groups, the difference between the SAU group’s average MPR (75.2%) and the PLY group’s average MPR (83.9%) was not statistically different on either the continuous or binary versions of the outcome (p=0.379, p=0.465, respectively). These differences remained non-significant in the adjusted models as well (see Table 3).

Table 3.

Adjhusted model results for clinical outcomes in full sample (n=41).

Continuous MPR B (95% CI) p
Treatment - PLY group 7.87 (−13.3, 29.0) 0.455
Age 0.85 (−0.25, 1.95) 0.126
Sex – Female 4.05 (−17.1, 25.1) 0.699
Baseline DSM – Moderate −2.14 (−52.0, 47.8) 0.931
Baseline DSM - Severe −8.46 (−30.3, 13.4) 0.437
Binary MPR OR (95% CI) p
Treatment – PLY group 1.54 (0.30, 8.41) 0.603
Age 1.09 (0.99, 1.21) 0.085
Sex – Female 1.94 (0.37, 10.1) 0.417
Baseline DSM – Moderate 0.11 (0.002, 5.26) 0.250
Baseline DSM - Severe 0.38 (0.05, 2.19) 0.306
Continuous Retention B (95% CI) p
Treatment - PLY group 18.7 (−2.33 – 39.6) 0.080
Age 1.72 (0.62 – 2.82) 0.003
Sex - Female 2.08 (−19.4 – 25.5) 0.845
Baseline DSM – Moderate −13.3 (−62.2, 35.6) 0.585
Baseline DSM - Severe −15.6 (−37.3, 6.22) 0.156
Binary Retention OR (95% CI) p
Treatment - PLY Group 2.18 (0.33 – 17.2) 0.420
Age 1.20 (1.07 – 1.43) 0.011
Sex – Female 1.67 (0.24 – 10.7) 0.585
Baseline DSM – Moderate 0.04 (0.0002, 3.37) 0.187
Baseline DSM - Severe 0.17 (0.02, 1.13) 0.089
Pill Cap Openings B (95% CI) p
Treatment – PLY group 48.7 (10.6, 86.8) 0.014
Age 3.12 (1.16, 5.08) 0.003
Sex – Female 8.46 (−29.4, 46.4) 0.653
Baseline DSM – Moderate 29.7 (−57.2, 116.6) 0.492
Baseline DSM - Severe 35.0 (−3.58, 73.5) 0.074
Open in a new tab

Note: For retention, MPR, and pill cap openings output, the reference groups for treatment, sex, and baseline DSM are control group, male, and mild, respectively.

3.3. Secondary Outcomes

The difference between the SAU group’s study retention days (mean of 53 days in study) and the PLY group’s study retention days (mean of 71 days in study) was not statistically different (p=0.096; see Table 1 and Figure 2), nor was the difference between the SAU group’s study completion percentage (54.5%) versus the PLY’s group’s completion percentage (68.4%; p=0.412). These differences remained non-significant in the adjusted models (see Table 3). Lastly, after controlling for age, sex, and baseline DSM score, there was a statistically significant difference in average pill cap openings between the two treatment groups, with the PLY group having an average of 49 more pill cap openings compared to the SAU group (B=48.7; p=0.014, see Table 3, Figure 3).

Figure 2.

Figure 2.

Open in a new tab

Median (with 25th and 75th quartile at either end of the box) Number of Days Enrolled by Treatment Group During the 12-week Trial Period.

Figure 3.

Figure 3.

Open in a new tab

Median (with 25th and 75th quartile at either end of the box) Number of Pill Cap Openings by Treatment Group During the 12-week Trial Period.

For the outcome of illicit opioid UAs during the study period, the time by treatment interaction was not statistically significant in the unadjusted or adjusted model (p=0.466, p=0.796, respectively). For the outcome of illicit stimulant use, time was nearly significant in the unadjusted model (B=−0.374, p=0.057) and was significant in the adjusted model (B=−0.042, p=0.046). This suggests that both the SAU and PLY groups were less likely to submit a positive illicit stimulant UA over time. The time by treatment interaction was also nearly significant in both the unadjusted (B=0.068, p=0.057) and adjusted model (B=0.070, p=0.053). Notably, there was no statistically significant difference between the groups on the number of opioids (SAU mean=5.20, PLY=5.09; p=0.842) and stimulants (SAU mean=5.59, PLY=4.89; p=0.157) urines submitted for testing during the study period.

4. DISCUSSION

The primary aim of this study was to determine whether patients randomized to the PLY group found it feasible and usable, and to determine whether it could improve adherence to BUP/NAL, compared to those in the SAU group. Most of the participants in PLY group found the Bluetooth-enabled pill cap used to track their BUP/NAL both feasible (92.86%) and usable (78.57%), and most participants liked using the device (92.86%) and were satisfied with the device (85.71%), indicating a high degree of feasibility. These are early, yet promising, indicators that this device could be deployed at scale for a test of its efficacy in a larger, more diverse sample.

As to the efficacy of the device, these data show promising signals that indicate a need for additional investigation based on the descriptive, not inferential, results. For example, the adherence per the MPR was 75% in the SAU group and 84% in the PLY group. While not a statistically significant difference, both numbers are markedly higher than most national reports of BUP/NAL adherence that typically report adherence between 21% and 43%, albeit over 12 months of treatment but inclusive of both commercially insured patients and Medicaid patients.(Kinsky et al., 2019; Pizzicato et al., 2020; Ronquest et al., 2018; Ruetsch et al., 2017; Samples et al., 2018; Tkacz et al., 2014) While there is no directly comparable reports on adherence across 12 weeks like our investigation, other studies report six weeks and 180 days of BUP/NAL adherence levels of 67% and 27%, respectively; both of which are also substantially lower than reported for either group in our investigation.(Muruganandam et al., 2019; Pizzicato et al., 2020) Moreover, the level of treatment retention was descriptively higher in the PLY group with an average of 71 days in the study (68.4% study completion) compared to the SAU group's average of 53 days in the study (54.5% study completion). While not statistically significant, this approximately 2.5 week increase in retention is in the expected direction.

Pill cap openings, another measure of patient engagement and not necessarily adherence was significantly higher in the PLY group with an average of 91.8 openings versus the SAU group's average of 56.7. As noted above, the data for pill cap openings came from the system used in this investigation, while data for adherence via the MPR came directly from the clinics where patients were receiving their care. It is important to note that SAU participants for this study were informed at the beginning of the trial that their cap openings were going to be recorded. Anecdotally, several participants in the SAU group reported that knowing that the cap openings were being tracked increased their accountability and adherence to the medication protocol. As noted above, higher social desirability has been found to be associated with greater treatment engagement.(Davis et al., 2014; Zemore, 2012) The implications could be that treatment retention could be maximized by capitalizing on this through engaging people in treatment through more interpersonal interactions leading to positive behaviors related to promoting social desirability.(Lopez et al., 2019) Future studies would do well to assess therapeutic alliance before, during and after investigations of similar devices. It is possible that the differences observed in this trial would have been more pronounced if SAU participants would have used a standard cap. While only one efficacy outcome achieved statistical significance, the PLY group was associated with a better response in all outcome measures. Due to our relatively small sample (n = 41), this lack of statistical significance is likely due to statistical power. We suggest that the observed associations for the efficacy outcomes be considered by their effect sizes and not just their respective significance levels.

4.1. Strengths and Limitations

Some strengths of this study include the representation of women (54%), the high rate of end-of-treatment assessment (95%), and the inclusion of the number of pill cap openings as an additional adherence outcome. This measure may prove to be a useful tool when evaluating pharmacological adherence. Nevertheless, this study has some important limitations to consider. First, our sample was predominantly White (88%), limiting our ability to generalize these findings to other races/ethnicities. Second, this study had a relatively small sample (n = 41), which may have impacted our ability to detect smaller effect size differences. Third, the control, or SAU group, was too active as a realistic comparator. While we labeled this group "services as usual", patients experienced an enhanced level of service by virtue of receiving a cap they knew was monitoring their pill cap openings. While this aspect of the study was necessary to measure pill cap openings, this is beyond what services are normally provided to patients. As a result, this aspect of the study may have increased the level of adherence in the SAU group to be markedly higher than the nationally reported levels. Future investigation of this device should consider a more realistic comparator that does not receive an active cap the way our SAU group did, as this could make it more difficult to determine the true efficacy of the Pillsy cap. Fourth, this study did not evaluate the role of the ‘helper’. Future research should not only engage enlisted helpers in evaluation of this therapeutic from their perspective but should also query people in treatment using the device as to whether and how their helper engaged with them or impacted their treatment outcomes. Lastly, the use of the MPR is not as strong of a measure of adherence as some other options, and while many of those options require medication adherence tools that are equally difficult or expensive to deploy, limiting its feasibility usage in real treatment settings, it’s possible that the use of such tools may have improved our ability to determine the efficacy of PLY in promoting BUP/NAL adherence. However, it is important to note that, when working with BUP/NAL, tracking individual levels of adherence (beyond dispensed medications) can be difficult because it is highly individualized with widely varying dosages across individuals and across time. Thus, the MPR remains a good option to estimate medication adherence while also being a commonly reported measure, enabling our results to be compared effectively to nationally reported data.

4.2. Conclusion

This study shows that Pillsy is a feasible and usable device that may increase adherence to BUP/NAL, while assisting clinicians in monitoring treatment engagement from a distance. As more and more telehealth options become available for the treatment of substance use disorders, Pillsy is a potentially low-cost technology that shows promise in improving adherence to BUP/NAL among individuals seeking treatment for OUD. Future studies with larger samples and conducted over longer periods of time are necessary to determine the efficacy of Pillsy in promoting long-term BUP/NAL adherence. Additionally, while the device that we are reporting on is consistent with a report published by our team that asked patients receiving BUP/NAL about several features,(Smith et al., 2023) there are some features (e.g., reinforcement for meeting adherence goals) that were not included. If some additional features could be integrated, this would only further enhance the potency of such a device.

Highlights.

  • Poor medication adherence is a barrier to the efficacy of buprenorphine/naloxone (BUP/NAL).

  • The aim of this study was to test if an electronic cap (i.e., Pillsy) was usable, feasible, and efficacious.

  • Participants in Pillsy group found the cap used to be both feasible (92.86%) and usable (78.57%).

  • Medication adherence was 75% in the control group and 84% in the PIllsy group.

  • The device was associated with increased pill openings.

Disclosures and Acknowledgments

Jeff LeBrun is the co-founder and former CEO of Optimize Health, and the inventor of the Pillsy device used in this report. Dr. McPherson has also received research funding from Orthopedic Specialty Institute and consulted for Consistent Care company and the Department of Justice. This funding is in no way related to the investigation reported here.

This project was supported by the following grants from the National Institutes of Health: R44DA977631. The funding source had no role other than financial support.

All authors contributed in a significant way to the manuscript and all authors have read and approved the final manuscript.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Declarations of Interests

Jeff LeBrun is the co-founder and former CEO of Optimize Health, and the inventor of the Pillsy device used in this report. He was also the contact Principal Investigator for this NIDA SBIR grant. Dr. Sterling McPherson (lead author) was a Multiple Principal Investigator and has no vested interest in the company or the device. The Pillsy device is no longer in production and Jeff LeBrun is no longer an employee of Optimize Health, but he does sit on the Board of Directors.

REFERENCES

  1. 2021. National Center for Health Statistics, Mortality Multiple Cause-of-Death Public Use Data Files and Documentation. www.cdc.gov/nchs/data_access/vitalstatsonline.htm#Mortality_Multiple. 2021). [Google Scholar]
  2. 2022. Centers for Disease Control and Prevention National Vital Statistics System, Mortality Multiple Cause-of-Death. www.cdc.gov/nchs/data_access/vitalstatsonline.htm#Mortality_Multiple. 2022). [Google Scholar]
  3. Brooke J, 1996. SUS-A quick and dirty usability scale. Usability evaluation in industry 189(194), 4–7. [Google Scholar]
  4. Davis CG, Doherty S, Moser AE, 2014. Social desirability and change following substance abuse treatment in male offenders. Psychol Addict Behav 28(3), 872–879. [DOI] [PubMed] [Google Scholar]
  5. Dayal P, Balhara YPS, 2017. A naturalistic study of predictors of retention in treatment among emerging adults entering first buprenorphine maintenance treatment for opioid use disorders. Journal of Substance Abuse Treatment 80, 1–5. [DOI] [PubMed] [Google Scholar]
  6. Fareed A, Eilender P, Ketchen B, Buchanan-Cummings AM, Scheinberg K, Crampton K, Nash A, Shongo-Hiango H, Drexler K, 2014. Factors affecting noncompliance with buprenorphine maintenance treatment. Journal of Addiction Medicine 8(5), 345–350. [DOI] [PubMed] [Google Scholar]
  7. Food and Drug Administration, 2014. Prescription Opioid Abuse and Misuse Issues: FDA Briefing for Stakeholders. https://wayback.archive-it.org/7993/20170111003429/http://www.fda.gov/NewsEvents/Speeches/ucm391767.htm. (Accessed 10-24-23. [Google Scholar]
  8. Fudala PJ, Bridge TP, Herbert S, Williford WO, Chiang CN, Jones K, Collins J, Raisch D, Casadonte P, Goldsmith RJ, Ling W, Malkerneker U, McNicholas L, Renner J, Stine S, Tusel D, Buprenorphine/Naloxone Collaborative Study, G., 2003. Office-based treatment of opiate addiction with a sublingual-tablet formulation of buprenorphine and naloxone. N Engl J Med 349(10), 949–958. [DOI] [PubMed] [Google Scholar]
  9. Gopaldas M, Wenzel K, Campbell ANC, Jalali A, Fishman M, Rotrosen J, Nunes EV, Murphy SM, 2023. Impact of Medication-Based Treatment on Health Care Utilization Among Individuals With Opioid Use Disorder. Psychiatr Serv 74(12), 1227–1233. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Hadland SE, Park TW, Bagley SM, 2018. Stigma associated with medication treatment for young adults with opioid use disorder: A case series. Addiction Science & Clinical Practice 13(1), 1–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Hser YI, Liang D, Lan YC, Vicknasingam BK, Chakrabarti A, 2016. Drug Abuse, HIV, and HCV in Asian Countries. J Neuroimmune Pharmacol 11(3), 383–393. [DOI] [PubMed] [Google Scholar]
  12. Hser YI, Saxon AJ, Huang D, Hasson A, Thomas C, Hillhouse M, Jacobs P, Teruya C, McLaughlin P, Wiest K, Cohen A, Ling W, 2014. Treatment retention among patients randomized to buprenorphine/naloxone compared to methadone in a multi-site trial. Addiction 109(1), 79–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Kinsky S, Houck PR, Mayes K, Loveland D, Daley D, Schuster JM, 2019. A comparison of adherence, outcomes, and costs among opioid use disorder Medicaid patients treated with buprenorphine and methadone: A view from the payer perspective. J Subst Abuse Treat 104, 15–21. [DOI] [PubMed] [Google Scholar]
  14. Krasnoff A, 1976. Differences between alcoholics who complete or withdraw from treatment. J Stud Alcohol 37(11), 1666–1671. [DOI] [PubMed] [Google Scholar]
  15. Liao S, Jang S, Tharp JA, Lester NA, 2023. Relationship between medication adherence for opioid use disorder and health care costs and health care events in a claims dataset. J Subst Use Addict Treat 154, 209139. [DOI] [PubMed] [Google Scholar]
  16. Ling W, Shoptaw S, Goodman-Meza D, 2019. Depot buprenorphine injection in the management of opioid use disorder: from development to implementation. Substance abuse and rehabilitation 10, 69. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Ling W, Wesson DR, Charuvastra C, Klett CJ, 1996. A controlled trial comparing buprenorphine and methadone maintenance in opioid dependence. Arch Gen Psychiat 53(5), 401–407. [DOI] [PubMed] [Google Scholar]
  18. Lofwall MR, Walsh SL, 2014. A review of buprenorphine diversion and misuse: The current evidence base and experiences from around the world. Journal of Addiction Medicine 8(5), 315. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Lopez A, Schwenk S, Schneck CD, Griffin RJ, Mishkind MC, 2019. Technology-Based Mental Health Treatment and the Impact on the Therapeutic Alliance. Curr Psychiatry Rep 21(8), 76. [DOI] [PubMed] [Google Scholar]
  20. Mattick RP, Breen C, Kimber J, Davoli M, 2014. Buprenorphine maintenance versus placebo or methadone maintenance for opioid dependence. Cochrane Database Syst Rev(2), CD002207. [DOI] [PubMed] [Google Scholar]
  21. Muruganandam P, Shukla L, Sharma P, Kandasamy A, Chand P, Murthy P, 2019. 'Too little dose - too early discontinuation?'-Effect of buprenorphine dose on short term treatment adherence in opioid dependence. Asian J Psychiatr 44, 58–60. [DOI] [PubMed] [Google Scholar]
  22. Park HR, Kang HS, Kim SH, Singh-Carlson S, 2022. Effect of a Smart Pill Bottle Reminder Intervention on Medication Adherence, Self-efficacy, and Depression in Breast Cancer Survivors. Cancer Nurs 45(6), E874–E882. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Peter SC, Murphy JG, Witkiewitz K, Hand SB, Thomas F, Johnson KC, Cowan R, Harris M, Derefinko KJ, 2023. Use of a sequential multiple assignment randomized trial to test contingency management and an integrated behavioral economic and mindfulness intervention for buprenorphine-naloxone medication adherence for opioid use disorder. Trials 24(1), 237. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Pizzicato LN, Hom JK, Sun M, Johnson CC, Viner KM, 2020. Adherence to buprenorphine: An analysis of prescription drug monitoring program data. Drug Alcohol Depend 216, 108317. [DOI] [PubMed] [Google Scholar]
  25. Ronquest NA, Willson TM, Montejano LB, Nadipelli VR, Wollschlaeger BA, 2018. Relationship between buprenorphine adherence and relapse, health care utilization and costs in privately and publicly insured patients with opioid use disorder. Subst Abuse Rehabil 9, 59–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Rosenblum A, Cleland CM, Fong C, Kayman DJ, Tempalski B, Parrino M, 2011. Distance traveled and cross-state commuting to opioid treatment programs in the United States. Journal of Environmental and Public Health 2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Ruetsch C, Tkacz J, Nadipelli VR, Brady BL, Ronquest N, Un H, Volpicelli J, 2017. Heterogeneity of nonadherent buprenorphine patients: subgroup characteristics and outcomes. Am J Manag Care 23(6), e172–e179. [PubMed] [Google Scholar]
  28. Saloner B, Daubresse M, Alexander GC, 2017. Patterns of buprenorphine-naloxone treatment for opioid use disorder in a multi-state population. Medical care 55(7), 669. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Samples H, Williams AR, Olfson M, Crystal S, 2018. Risk factors for discontinuation of buprenorphine treatment for opioid use disorders in a multi-state sample of Medicaid enrollees. J Subst Abuse Treat 95, 9–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Smith CL, Keever A, Bowden T, Olson K, Rodin N, McDonell MG, Roll JM, Smoody G, LeBrun J, Miguel AQ, McPherson SM, 2023. Patient Feedback on a Mobile Medication Adherence App for Buprenorphine and Naloxone: Closed and Open-Ended Survey on Feasibility and Acceptability. JMIR Form Res 7, e40437. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Sordo L, Barrio G, Bravo MJ, Indave BI, Degenhardt L, Wiessing L, Ferri M, Pastor-Barriuso R, 2017. Mortality risk during and after opioid substitution treatment: systematic review and meta-analysis of cohort studies. Bmj 357, j1550. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Stringfellow EJ, Lim TY, Humphreys K, DiGennaro C, Stafford C, Beaulieu E, Homer J, Wakeland W, Bearnot B, McHugh RK, Kelly J, Glos L, Eggers SL, Kazemi R, Jalali MS, 2022. Reducing opioid use disorder and overdose deaths in the United States: A dynamic modeling analysis. Sci Adv 8(25), eabm8147. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Tkacz J, Volpicelli J, Un H, Ruetsch C, 2014. Relationship between buprenorphine adherence and health service utilization and costs among opioid dependent patients. J Subst Abuse Treat 46(4), 456–462. [DOI] [PubMed] [Google Scholar]
  34. U.S. Department of Health and Human Services Substance Abuse and Mental Health Services Administration, HHS releases $1.5 billion to states, tribes to combat opioid crisis. www.samhsa.gov/newsroom/press-announcements/202008270530.
  35. U.S. Drug Enforcement Administration, 2021. “2020 National Drug Threat Assessment (NDTA)”. www.dea.gov/sites/default/files/2021-02/DIR-008-212020NationalDrugThreatAssessment_WEB.pdf. [Google Scholar]
  36. Velez FF, Anastassopoulos KP, Colman S, Shah N, Kauffman L, Murphy SM, Ruetsch C, Maricich YA, 2022. Reduced healthcare resource utilization in patients with opioid use disorder in the 12 months after initiation of a prescription digital therapeutic. Advances in Therapy 39(9), 4131–4145. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Velez FF, Colman S, Kauffman L, Ruetsch C, Anastassopoulos K, 2021. Real-world reduction in healthcare resource utilization following treatment of opioid use disorder with reSET-O, a novel prescription digital therapeutic. Expert Review of Pharmacoeconomics & Outcomes Research 21(1), 69–76. [DOI] [PubMed] [Google Scholar]
  38. Volkow ND, Frieden TR, Hyde PS, Cha SS, 2014. Medication-assisted therapies—tackling the opioid-overdose epidemic. New England Journal of Medicine 370(22), 2063–2066. [DOI] [PubMed] [Google Scholar]
  39. Williams AR, Aronowitz S, Gallagher R, Behar E, Gray Z, Bisaga A, 2023. A Virtual-First Telehealth Treatment Model for Opioid Use Disorder. Journal of General Internal Medicine 38(3), 814–816. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Yoshino A, Kato M, 1995. Influence of social desirability response set on self-report for assessing the outcome of treated alcoholics. Alcoholism, clinical and experimental research 19(6), 1517–1519. [DOI] [PubMed] [Google Scholar]
  41. Zemore SE, 2012. The effect of social desirability on reported motivation, substance use severity, and treatment attendance. J Subst Abuse Treat 42(4), 400–412. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES