Search

Background: Some patients prescribed opioid analgesic (OA) medications for pain experience serious side effects, including dependence, sedation, and overdose. As most patients are at low risk for OA-related harms, risk reduction interventions requiring multiple counseling sessions are impractical on a large scale., Objective: This study evaluates whether an intervention based on reinforcement learning (RL), a field of artificial intelligence, learned through experience to personalize interactions with patients with pain discharged from the emergency department (ED) and decreased self-reported OA misuse behaviors while conserving counselors' time., Methods: We used data representing 2439 weekly interactions between a digital health intervention ("Prescription Opioid Wellness and Engagement Research in the ED" [PowerED]) and 228 patients with pain discharged from 2 EDs who reported recent opioid misuse. During each patient's 12 weeks of intervention, PowerED used RL to select from 3 treatment options: a brief motivational message delivered via an interactive voice response (IVR) call, a longer motivational IVR call, or a live call from a counselor. The algorithm selected session types for each patient each week, with the goal of minimizing OA risk, defined in terms of a dynamic score reflecting patient reports during IVR monitoring calls. When a live counseling call was predicted to have a similar impact on future risk as an IVR message, the algorithm favored IVR to conserve counselor time. We used logit models to estimate changes in the relative frequency of each session type as PowerED gained experience. Poisson regression was used to examine the changes in self-reported OA risk scores over calendar time, controlling for the ordinal session number (1st to 12th)., Results: Participants on average were 40 (SD 12.7) years of age; 66.7% (152/228) were women and 51.3% (117/228) were unemployed. Most participants (175/228, 76.8%) reported chronic pain, and 46.2% (104/225) had moderate to severe depressive symptoms. As PowerED gained experience through interactions over a period of 142 weeks, it delivered fewer live counseling sessions than brief IVR sessions (P=.006) and extended IVR sessions (P<.001). Live counseling sessions were selected 33.5% of the time in the first 5 weeks of interactions (95% CI 27.4%-39.7%) but only for 16.4% of sessions (95% CI 12.7%-20%) after 125 weeks. Controlling for each patient's changes during the course of treatment, this adaptation of treatment-type allocation led to progressively greater improvements in self-reported OA risk scores (P<.001) over calendar time, as measured by the number of weeks since enrollment began. Improvement in risk behaviors over time was especially pronounced among patients with the highest risk at baseline (P=.02)., Conclusions: The RL-supported program learned which treatment modalities worked best to improve self-reported OA risk behaviors while conserving counselors' time. RL-supported interventions represent a scalable solution for patients with pain receiving OA prescriptions., Trial Registration: Clinicaltrials.gov NCT02990377; https://classic.clinicaltrials.gov/ct2/show/NCT02990377., (©John D Piette, Laura Thomas, Sean Newman, Nicolle Marinec, Joel Krauss, Jenny Chen, Zhenke Wu, Amy S B Bohnert. Originally published in the Journal of Medical Internet Research (https://www.jmir.org), 11.07.2023.)

Purpose: To test whether technology-facilitated self-management support improves depression in primary care settings. Methods: We randomized 204 low-income primary care patients who had at least moderate depressive symptoms to intervention or control. Intervention participants received 12 months of weekly automated interactive voice response telephone calls that assessed their symptom severity and provided self-management strategies. Their patient-nominated supporter (CarePartner) received corresponding guidance on self-management support, and their primary care team received urgent notifications. Those randomized to enhanced usual care received printed generic self-management instructions. Results: One-year attrition rate was 14%. By month 6, symptom severity on the Patient Health Questionnaire-9 (PHQ-9) decreased 2.5 points more in the intervention arm than in the control arm (95% CI -4.2 to -0.8, p  = 0.003). This benefit was similar at month 12 ( p  = 0.004). Intervention was also over twice as likely to lead to ≥50% reduction in symptom severity by month 6 (OR = 2.2 (1.1, 4.7)) and a decrease of ≥5 PHQ-9 points by month 12 (OR = 2.3 (1.2, 4.4)). Conclusions: Technology-facilitated self-management guidance with lay support and clinician notifications improves depression for primary care patients. Subsequent research should examine implementation and generalization to other chronic conditions. clinicaltrials.gov, identifier NCT01834534.

Background: Patients' engagement in mobile health (m-health) interventions using interactive voice response (IVR) calls is less in low- and middle-income countries (LMICs) than in industrialized ones. We conducted a study to determine whether automated telephone feedback to informal caregivers ("CarePartners") increased engagement in m-health support among diabetes and hypertension patients in Bolivia., Materials and Methods: Patients with diabetes and/or hypertension were identified through ambulatory clinics affiliated with four hospitals. All patients enrolled with a CarePartner. Patients were randomized to weekly IVR calls including self-management questions and self-care education either alone ("standard m-health") or with automated feedback about health and self-care needs sent to their CarePartner after each IVR call ("m-health+CP")., Results: The 72 participants included 39 with diabetes and 53 with hypertension, of whom 19 had ≤6 years of education. After 1,225 patient-weeks of attempted IVR assessments, the call completion rate was higher among patients randomized to m-health+CP compared with standard m-health (62.0% versus 44.9%; p < 0.047). CarePartner feedback more than tripled call completion rates among indigenous patients and patients with low literacy (p < 0.001 for both). M-health+CP patients were more likely to report excellent health via IVR (adjusted odds ratio [AOR] = 2.60; 95% confidence interval [CI], 1.07, 6.32) and less likely to report days in bed due to illness (AOR = 0.42; 95% CI, 0.19, 0.91)., Conclusions: In this study we found that caregiver feedback increased engagement in m-health and may improve patients' health status relative to standard approaches. M-health+CP represents a scalable strategy for increasing the reach of self-management support in LMICs.

Objective: The goal of this trial is to evaluate a novel intervention designed to improve post-hospitalization support for older adults with chronic conditions via: (a) direct tailored communication to patients using regular automated calls post discharge, (b) support for informal caregivers outside of the patient's household via structured automated feedback about the patient's status plus advice about how caregivers can help, and (c) support for care management including a web-based disease management tool and alerts about potential problems., Methods: 846 older adults with common chronic conditions are being identified upon hospital admission. Patients are asked to identify a "CarePartner" (CP) living outside their household, i.e., an adult child or other social network member willing to play an active role in their post-discharge transition support. Patient-CP pairs are randomized to the intervention or usual care. Intervention patients receive automated assessment and behavior change calls, and their CPs receives structured feedback and advice via email and automated calls following each assessment. Clinical teams have access to assessment results via the web and receive automated reports about urgent health problems. Patients complete surveys at baseline, 30 days, and 90 days post discharge; utilization data is obtained from hospital records. CPs, other caregivers, and clinicians are interviewed to evaluate intervention effects on processes of self-care support, caregiver stress and communication, and the intervention's potential for broader implementation. The primary outcome is 30-day readmission rates; other outcomes measured at 30 days and 90 days include functional status, self-care behaviors, and mortality risk., Conclusion: This trial uses accessible health technologies and coordinated communication among informal caregivers and clinicians to fill the growing gap between what discharged patients need and available resources. A unique feature of the intervention is the provision of transition support not only for patients but also for their informal caregivers.

Objective: To understand patient participation in interactive voice response (IVR) depression monitoring and self-management support calls and estimate the workload from clinical alerts based on patients' IVR reports., Study Design: Observational study from program implementation in 13 community- and university-based primary care practices., Methods: Patients with depression were identified using electronic records and enrolled by telephone. Patients were asked to complete IVR assessments weekly; those with significantly improved symptoms had the option of completing assessments monthly. Patients could enroll with an informal caregiver who received automated feedback based on patients' IVR reports. Clinicians received alerts regarding significant changes in the scores on the Patient Health Questionnaire 9-item depression measure, antidepressant adherence problems, or suicidal ideation., Results: A total of 387 patients were followed for 12,042 weeks. More than half (59%) opted to participate with a caregiver. Patients completed 68% of 7912 attempted IVR assessments. Assessment completion was unrelated to patients' depressive symptoms and was higher among those who participated with a caregiver, were married, had more comorbidities, or reported missing a prior appointment. Assessment completion was lower when patients received monthly versus weekly assessment attempts. Clinical alerts were generated during 4.9% of follow-up weeks; most represented medication adherence problems (2.8%). Alerts indicating suicidal ideation were rare (0.2% of patient-weeks)., Conclusions: IVR support calls represent a viable strategy for increasing access to depression monitoring and self-management assistance in primary care. These programs generate a manageable number of alerts, most of which can be triaged with limited physician involvement.

Objective: Hypertension and other noncommunicable diseases represent a growing threat to low/middle-income countries (LMICs). Mobile health technologies may improve noncommunicable disease outcomes, but LMICs lack resources to provide these services. We evaluated the efficacy of a cloud computing model using automated self-management calls plus home blood pressure (BP) monitoring as a strategy for improving systolic BPs (SBPs) and other outcomes of hypertensive patients in two LMICs., Subjects and Methods: This was a randomized trial with a 6-week follow-up. Participants with high SBPs (≥140 mm Hg if nondiabetic and ≥130 mm Hg if diabetic) were enrolled from clinics in Honduras and Mexico. Intervention patients received weekly automated monitoring and behavior change telephone calls sent from a server in the United States, plus a home BP monitor. At baseline, control patients received BP results, hypertension information, and usual healthcare. The primary outcome, SBP, was examined for all patients in addition to a preplanned subgroup with low literacy or high hypertension information needs. Secondary outcomes included perceived health status and medication-related problems., Results: Of the 200 patients recruited, 181 (90%) completed follow-up, and 117 of 181 had low literacy or high hypertension information needs. The median annual income was $2,900 USD, and average educational attainment was 6.5 years. At follow-up intervention patients' SBPs decreased 4.2 mm Hg relative to controls (95% confidence interval -9.1, 0.7; p=0.09). In the subgroup with high information needs, intervention patients' average SBPs decreased 8.8 mm Hg (-14.2, -3.4, p=0.002). Compared with controls, intervention patients at follow-up reported fewer depressive symptoms (p=0.004), fewer medication problems (p<0.0001), better general health (p<0.0001), and greater satisfaction with care (p≤0.004)., Conclusions: Automated telephone care management plus home BP monitors can improve outcomes for hypertensive patients in LMICs. A cloud computing model within regional telecommunication centers could make these services available in areas with limited infrastructure for patient-focused informatics support.