The impact of using reinforcement learning to personalize communication on medication adherence: findings from the REINFORCE trial.

Text messaging can promote healthy behaviors, like adherence to medication, yet its effectiveness remains modest, in part because message content is rarely personalized. Reinforcement learning has been used in consumer technology to personalize content but with limited application in healthcare. We tested a reinforcement learning program that identifies individual responsiveness ("adherence") to text message content and personalizes messaging accordingly. We randomized 60 individuals with diabetes and glycated hemoglobin A1c [HbA1c] ≥ 7.5% to reinforcement learning intervention or control (no messages). Both arms received electronic pill bottles to measure adherence. The intervention improved absolute adjusted adherence by 13.6% (95%CI: 1.7%-27.1%) versus control and was more effective in patients with HbA1c 7.5- < 9.0% (36.6%, 95%CI: 25.1%-48.2%, interaction p < 0.001). We also explored whether individual patient characteristics were associated with differential response to tested behavioral factors and unique clusters of responsiveness. Reinforcement learning may be a promising approach to improve adherence and personalize communication at scale.

Clinicians' and Patients' Perspectives on Hypertension Care in a Racially and Ethnically Diverse Population in Primary Care.

Importance: Hypertension control remains suboptimal, particularly for Black and Hispanic or Latino patients. A need exists to improve hypertension management and design effective strategies to efficiently improve the quality of care in primary care, especially for these at-risk populations. Few studies have specifically explored perspectives on blood pressure management by primary care providers (PCPs) and patients. Objective: To examine clinician and patient perspectives on barriers and facilitators to hypertension control within a racially and ethnically diverse health care system. Design, Setting, and Participants: This qualitative study was conducted in a large urban US health care system from October 1, 2020, to March 31, 2021, among patients with a diagnosis of hypertension from a racially and ethnically diverse population, for a range of hypertension medication use hypertension control, as well as practicing PCPs. Analysis was conducted between June 2021 and February 2022 using immersion-crystallization methods. Main Outcomes and Measures: Perspectives on managing blood pressure, including medication adherence and lifestyle, considerations for intensification, and experiences and gaps in using health information technology tools for hypertension, were explored using semistructured qualitative interviews. These cycles of review were continued until all data were examined and meaningful patterns were identified. Results: Interviews were conducted with 30 participants: 15 patients (mean [SD] age, 58.6 [16.2] years; 10 women [67%] and 9 Black patients [60%]) and 15 clinicians (14 PCPs and 1 medical assistant; 8 women [53%]). Eleven patients (73%) had suboptimally controlled blood pressure. Participants reported a wide range of experiences with hypertension care, even within the same clinics and health care system. Five themes relevant to managing hypertension for racially and ethnically diverse patient populations in primary care were identified: (1) difficulty with self-management activities, especially lifestyle modifications; (2) hesitancy intensifying medications by both clinicians and patients; (3) varying the timing and follow-up after changes in medication; (4) variation in blood pressure self-monitoring recommendations and uptake; and (5) limited specific functionality of current health information technology tools. Conclusions and Relevance: In this qualitative study of the views of PCPs and patients on hypertension control, the participants felt that more focus should be placed on lifestyle modifications than medications for hypertension, particularly for patients from racial and ethnic minority groups. Participants also expressed concerns about the existing functionality of health information technology tools to support increasingly asynchronous hypertension care. More intentional ways of supporting treatment intensification, self-care, and follow-up care are needed to improve hypertension management for racially and ethnically diverse populations in primary care.

Reducing ethnic and racial disparities by improving undertreatment, control, and engagement in blood pressure management with health information technology (REDUCE-BP) hybrid effectiveness-implementation pragmatic trial: Rationale and design.

BACKGROUND: While racial/ethnic disparities in blood pressure control are documented, few interventions have successfully reduced these gaps. Under-prescribing, lack of treatment intensification, and suboptimal follow-up care are thought to be central contributors. Electronic health record (EHR) tools may help address these barriers and may be enhanced with behavioral science techniques. OBJECTIVE: To evaluate the impact of a multicomponent behaviorally-informed EHR-based intervention on blood pressure control. TRIAL DESIGN: Reducing Ethnic and racial Disparities by improving Undertreatment, Control, and Engagement in Blood Pressure management with health information technology (REDUCE-BP) (NCT05030467) is a two-arm cluster-randomized hybrid type 1 pragmatic trial in a large multi-ethnic health care system. Twenty-four clinics (>350 primary care providers [PCPs] and >10,000 eligible patients) are assigned to either multi-component EHR-based intervention or usual care. Intervention clinic PCPs will receive several EHR tools designed to reduce disparities delivered at different points, including a: (1) dashboard of all patients visible upon logging on to the EHR displaying blood pressure control by race/ethnicity compared to their PCP peers and (2) set of tools in an individual patient's chart containing decision support to encourage treatment intensification, ordering home blood pressure measurement, interventions to address health-related social needs, default text for note documentation, and enhanced patient education materials. The primary outcome is patient-level change in systolic blood pressure over 12 months between arms; secondary outcomes include changes in disparities and other clinical outcomes. CONCLUSION: REDUCE-BP will provide important insights into whether an EHR-based intervention designed using behavioral science can improve hypertension control and reduce disparities.

Effect of a Biopsychosocial Intervention or Postural Therapy on Disability and Health Care Spending Among Patients With Acute and Subacute Spine Pain: The SPINE CARE Randomized Clinical Trial.

Importance: Low back and neck pain are often self-limited, but health care spending remains high. Objective: To evaluate the effects of 2 interventions that emphasize noninvasive care for spine pain. Design, Setting, and Participants: Pragmatic, cluster, randomized clinical trial conducted at 33 centers in the US that enrolled 2971 participants with neck or back pain of 3 months' duration or less (enrollment, June 2017 to March 2020; final follow-up, March 2021). Interventions: Participants were randomized at the clinic-level to (1) usual care (n = 992); (2) a risk-stratified, multidisciplinary intervention (the identify, coordinate, and enhance [ICE] care model that combines physical therapy, health coach counseling, and consultation from a specialist in pain medicine or rehabilitation) (n = 829); or (3) individualized postural therapy (IPT), a postural therapy approach that combines physical therapy with building self-efficacy and self-management (n = 1150). Main Outcomes and Measures: The primary outcomes were change in Oswestry Disability Index (ODI) score at 3 months (range, 0 [best] to 100 [worst]; minimal clinically important difference, 6) and spine-related health care spending at 1 year. A 2-sided significance threshold of .025 was used to define statistical significance. Results: Among 2971 participants randomized (mean age, 51.7 years; 1792 women [60.3%]), 2733 (92%) finished the trial. Between baseline and 3-month follow-up, mean ODI scores changed from 31.2 to 15.4 for ICE, from 29.3 to 15.4 for IPT, and from 28.9 to 19.5 for usual care. At 3-month follow-up, absolute differences compared with usual care were -5.8 (95% CI, -7.7 to -3.9; P < .001) for ICE and -4.3 (95% CI, -5.9 to -2.6; P < .001) for IPT. Mean 12-month spending was $1448, $2528, and $1587 in the ICE, IPT, and usual care groups, respectively. Differences in spending compared with usual care were -$139 (risk ratio, 0.93 [95% CI, 0.87 to 0.997]; P = .04) for ICE and $941 (risk ratio, 1.40 [95% CI, 1.35 to 1.45]; P < .001) for IPT. Conclusions and Relevance: Among patients with acute or subacute spine pain, a multidisciplinary biopsychosocial intervention or an individualized postural therapy intervention, each compared with usual care, resulted in small but statistically significant reductions in pain-related disability at 3 months. However, compared with usual care, the biopsychosocial intervention resulted in no significant difference in spine-related health care spending and the postural therapy intervention resulted in significantly greater spine-related health care spending at 1 year. Trial Registration: ClinicalTrials.gov Identifier: NCT03083886.

Association Between Cost-Saving Prescription Policy Changes and Adherence to Chronic Disease Medications: an Observational Study.

BACKGROUND: Pharmacy benefit design is one tool for improving access and adherence to medications for the management of chronic disease. OBJECTIVE: We assessed the effects of pharmacy benefit design programs, including a change in pharmacy benefit manager (PBM), institution of a prescription out-of-pocket maximum, and a mandated switch to 90 days' medication supply, on adherence to chronic disease medications over time. DESIGN: We used a difference-in-differences design to assess changes in adherence to chronic disease medications after the transition to new prescription policies. SUBJECTS: We utilized claims data from adults aged 18-64, on ≥ 1 medication for chronic disease, whose insurer instituted the prescription policies (intervention group) and a propensity score-matched comparison group from the same region. MAIN MEASURES: The outcome of interest was adherence to chronic disease medications measured by proportion of days covered (PDC) using pharmacy claims. KEY RESULTS: There were 13,798 individuals in each group after propensity score matching. Compared to the matched control group, adherence in the intervention group decreased in the first quarter of 2015 and then increased back to pre-intervention trends. Specifically, the change in adherence compared to the last quarter of 2014 in the intervention group versus controls was - 3.6 percentage points (pp) in 2015 Q1 (p < 0.001), 0.65 pp in Q2 (p = 0.024), 1.1 pp in Q3 (p < 0.001), and 1.4 pp in Q4 (p < 0.001). CONCLUSIONS: In this cohort of commercially insured adults on medications for chronic disease, a change in PBM accompanied by a prescription out-of-pocket maximum and change to 90 days' supply was associated with short-term disruptions in adherence followed by return to pre-intervention trends. A small improvement in adherence over the year of follow-up may not be clinically significant. These findings have important implications for employers, insurers, or health systems wishing to utilize pharmacy benefit design to improve management of chronic disease.

REinforcement learning to improve non-adherence for diabetes treatments by Optimising Response and Customising Engagement (REINFORCE): study protocol of a pragmatic randomised trial.

INTRODUCTION: Achieving optimal diabetes control requires several daily self-management behaviours, especially adherence to medication. Evidence supports the use of text messages to support adherence, but there remains much opportunity to improve their effectiveness. One key limitation is that message content has been generic. By contrast, reinforcement learning is a machine learning method that can be used to identify individuals' patterns of responsiveness by observing their response to cues and then optimising them accordingly. Despite its demonstrated benefits outside of healthcare, its application to tailoring communication for patients has received limited attention. The objective of this trial is to test the impact of a reinforcement learning-based text messaging programme on adherence to medication for patients with type 2 diabetes. METHODS AND ANALYSIS: In the REinforcement learning to Improve Non-adherence For diabetes treatments by Optimising Response and Customising Engagement (REINFORCE) trial, we are randomising 60 patients with suboptimal diabetes control treated with oral diabetes medications to receive a reinforcement learning intervention or control. Subjects in both arms will receive electronic pill bottles to use, and those in the intervention arm will receive up to daily text messages. The messages will be individually adapted using a reinforcement learning prediction algorithm based on daily adherence measurements from the pill bottles. The trial's primary outcome is average adherence to medication over the 6-month follow-up period. Secondary outcomes include diabetes control, measured by glycated haemoglobin A1c, and self-reported adherence. In sum, the REINFORCE trial will evaluate the effect of personalising the framing of text messages for patients to support medication adherence and provide insight into how this could be adapted at scale to improve other self-management interventions. ETHICS AND DISSEMINATION: This study was approved by the Mass General Brigham Institutional Review Board (IRB) (USA). Findings will be disseminated through peer-reviewed journals, clinicaltrials.gov reporting and conferences. TRIAL REGISTRATION NUMBER: Clinicaltrials.gov (NCT04473326).

Randomised controlled trial targeting habit formation to improve medication adherence to daily oral medications in patients with gout.

INTRODUCTION: Medication adherence for patients with chronic conditions such as gout, a debilitating form of arthritis that requires daily medication to prevent flares, is a costly problem. Existing interventions to improve medication adherence have only been moderately effective. Habit formation theory is a promising strategy to improve adherence. The cue-reward-repetition principle posits that habits are formed by repeatedly completing an activity after the same cue and having the action rewarded every time. Over time, cues become increasingly important whereas rewards become less salient because the action becomes automatic. Leveraging the cue-reward-repetition principle could improve adherence to daily gout medications. METHODS AND ANALYSIS: This three-arm parallel randomised controlled trial tests an adaptive intervention that leverages the repetition cue-reward principle. The trial will began recruitment in August 2021 in Boston, Massachusetts, USA. Eligible patients are adults with gout who have been prescribed a daily oral medication for gout and whose most recent uric acid is above 6 mg/dL. Participants will be randomised to one of three arms and given electronic pill bottles. In the two intervention arms, participants will select a daily activity to link to their medication-taking (cue) and a charity to which money will be donated every time they take their medication (reward). Participants in Arm 1 will receive reminder texts about their cue and their charity reward amount will be US$0.50 per day of medication taken. Arm 2 will be adaptive; participants will receive a US$0.25 per adherent-day and no reminder texts. If their adherence is <75% 6 weeks postrandomisation, their reward will increase to US$0.50 per adherent-day and they will receive reminder texts. The primary outcome is adherence to gout medications over 18 weeks. ETHICS AND DISSEMINATION: This trial has ethical approval in the USA. Results will be published in a publicly accessible peer-reviewed journal. TRIAL REGISTRATION NUMBER: NCT04776161.

Design of the Spine Pain Intervention to Enhance Care Quality And Reduce Expenditure Trial (SPINE CARE) study: Methods and lessons from a multi-site pragmatic cluster randomized controlled trial.

BACKGROUND: Low back and neck pain (together, spine pain) are among the leading causes of medical visits, lost productivity, and disability. For most people, episodes of spine pain are self-limited; nevertheless, healthcare spending for this condition is extremely high. Focusing care on individuals at high-risk of progressing from acute to chronic pain may improve efficiency. Alternatively, postural therapies, which are frequently used by patients, may prevent the overuse of high-cost interventions while delivering equivalent outcomes. METHODS: The SPINE CARE (Spine Pain Intervention to Enhance Care Quality And Reduce Expenditure) trial is a cluster-randomized multi-center pragmatic clinical trial designed to evaluate the clinical effectiveness and healthcare utilization of two interventions for primary care patients with acute and subacute spine pain. The study was conducted at 33 primary care clinics in geographically distinct regions of the United States. Individuals ≥18 years presenting to primary care with neck and/or back pain of ≤3 months' duration were randomized at the clinic-level to 1) usual care, 2) a risk-stratified, multidisciplinary approach called the Identify, Coordinate, and Enhance (ICE) care model, or 3) Individualized Postural Therapy (IPT), a standardized postural therapy method of care. The trial's two primary outcomes are change in function at 3 months and spine-related spending at one year. 2971 individuals were enrolled between June 2017 and March 2020. Follow-up was completed on March 31, 2021. DISCUSSION: The SPINE CARE trial will determine the impact on clinical outcomes and healthcare costs of two interventions for patients with spine pain presenting to primary care. TRIAL REGISTRATION NUMBER: NCT03083886.

Preferences for mHealth Technology and Text Messaging Communication in Patients With Type 2 Diabetes: Qualitative Interview Study.

BACKGROUND: Individuals with diabetes need regular support to help them manage their diabetes on their own, ideally delivered via mechanisms that they already use, such as their mobile phones. One reason for the modest effectiveness of prior technology-based interventions may be that the patient perspective has been insufficiently incorporated. OBJECTIVE: This study aims to understand patients' preferences for mobile health (mHealth) technology and how that technology can be integrated into patients' routines, especially with regard to medication use. METHODS: We conducted semistructured qualitative individual interviews with patients with type 2 diabetes from an urban health care system to elicit and explore their perspectives on diabetes medication-taking behaviors, daily patterns of using mobile technology, use of mHealth technology for diabetes care, acceptability of text messages to support medication adherence, and preferred framing of information within text messages to support diabetes care. The interviews were digitally recorded and transcribed. The data were analyzed using codes developed by the study team to generate themes, with representative quotations selected as illustrations. RESULTS: We conducted interviews with 20 participants, of whom 12 (60%) were female and 9 (45%) were White; in addition, the participants' mean glycated hemoglobin A1c control was 7.8 (SD 1.1). Overall, 5 key themes were identified: patients try to incorporate cues into their routines to help them with consistent medication taking; many patients leverage some form of technology as a cue to support adherence to medication taking and diabetes self-management behaviors; patients value simplicity and integration of technology solutions used for diabetes care, managing medications, and communicating with health care providers; some patients express reluctance to rely on mobile technology for these diabetes care behaviors; and patients believe they prefer positively framed communication, but communication preferences are highly individualized. CONCLUSIONS: The participants expressed some hesitation about using mobile technology in supporting diabetes self-management but have largely incorporated it or are open to incorporating it as a cue to make medication taking more automatic and less burdensome. When using technology to support diabetes self-management, participants exhibited individualized preferences, but overall, they preferred simple and positively framed communication. mHealth interventions may be improved by focusing on integrating them easily into daily routines and increasing the customization of content.

The Internet is valid and reliable for child-report: an example using the Activities Scale for Kids (ASK) and the Pediatric Quality of Life Inventory (PedsQL).

OBJECTIVE: This study tested the impact of web administration on well-established measures of children's physical function and quality of life. STUDY DESIGN AND SETTING: Participants were recruited from clinics at six hospitals. They completed the Activities Scale for Kids (ASK) and the Pediatric Quality of Life Inventory (PedsQL) questionnaires twice, in a crossover design that used paper and web-based modes of administration. Intraclass correlation coefficients were used to assess the validity of the new web formats relative to the original paper formats and their test-retest reliability. RESULTS: Sixty-nine children ranging in age from 8.0 to 13.4 years (mean=11.0 years) completed the study. The sample included children with cerebral palsy (19), spina bifida (23), and cystic fibrosis (27). The mean ASK score was 77.5 and the mean PedsQL score was 69.1. The intermethod intraclass correlation coefficients were 0.98 (lower limit 0.94) for the ASK and 0.64 (lower limit 0.35) for the PedsQL. These compare to intraclass correlation coefficients of 0.99 and 0.94, respectively, for traditional paper formats. CONCLUSION: The web ASK was valid in comparison to the original paper format. Consistency in mode of administration may be more important when using the PedsQL. Both measures were highly reliable on paper and on the web.