A Co-Designed and Bilingual WhatsApp-Based Chatbot Intervention (PflegeBot) for Informal Caregivers in Germany: Single-Arm Nonrandomized Feasibility Study
JMIR Formative Research · 2026-11-11 · 10 · e95268-e95268
DOI: 10.2196/95268 ↗Abstract
Abstract Background Informal caregivers play a central role in long-term care but experience elevated risks of stress, depressive symptoms, and caregiver burden, alongside barriers to accessing preventive support. Low-threshold digital public health interventions may offer scalable support; however, evidence on their feasibility, usability, and contextual fit remains limited, particularly for heterogeneous caregiver populations. Objective This study describes the development and implementation of PflegeBot, an intersectionality-informed WhatsApp-based chatbot intervention, and evaluates its feasibility, usability, and acceptability among informal caregivers in Germany. Methods A single-arm, nonrandomized feasibility study with pre-post assessments was conducted between December 2023 and March 2024. Informal caregivers aged 18 years and older were recruited via community-based, institutional, and online channels. PflegeBot was a co-designed, bilingual (German and Turkish) WhatsApp-based chatbot that delivered 84 scheduled messages over 12 weeks. Content included caregiving information, self-care, and cognitive behavioral therapy–informed self-help, plus optional frequently asked questions and question functions. Baseline and full follow-up assessments were administered online. Feasibility was assessed using indicators of recruitment, chatbot initiation and message delivery, opt-out, and provision of follow-up measures. Usability was assessed using the System Usability Scale (SUS), and acceptability was assessed using study-specific items. Exploratory outcomes included perceived stress (4-item Perceived Stress Scale [PSS-4]), caregiver burden (4-item Zarit Burden Interview [ZBI-4]), depressive symptoms (Center for Epidemiologic Studies Depression Scale [CES-D]), and loneliness (6-item De Jong Gierveld Loneliness Scale [DJGLS-6]). Analyses were descriptive; no hypothesis tests were conducted. Open-ended follow-up feedback was examined using a systematic inductive descriptive approach. Results Of 102 participants included in the baseline analysis sample, 84 (82%) were women. According to chatbot log data, message delivery was initiated for 67 participants. Forty-nine participants provided at least one follow-up measure, and complete SUS data were available for 46 participants. The mean SUS score was 82.45 (SD 14.86, 95% CI 78.03‐86.86; range 47.5‐100), and 35 out of 46 (76%) participants scored at or above the benchmark of 68. Overall satisfaction with PflegeBot had a mean rating of 4.02 out of 5 (SD 0.98; n=46). Among participants responding to the relevant item, 1 out of 45 (2%) reported using the optional FAQ function. Mean changes were −0.57 (95% CI −1.24 to 0.10; n=35) for perceived stress, 0.81 (95% CI −0.02 to 1.63; n=36) for caregiver burden, −2.38 (95% CI −5.42 to 0.66; n=34) for depressive symptoms, and −0.39 (95% CI −0.90 to 0.12; n=36) for loneliness. Open-ended feedback highlighted needs for more flexible delivery, greater tailoring, and easier navigation. Conclusions PflegeBot could be delivered through WhatsApp, and participants providing follow-up data reported high usability and generally favorable satisfaction. Controlled evaluations with objective engagement data and assessment of differential reach and outcomes are warranted.
EUVIMED