Background: Complimentary subscriptions to UpToDate, a decision support tool, were provided to community health workers (CHWs) in rural and remote primary care sites as part of a government-funded health system research program. A feasibility evaluation conducted after the first year of implementation showed that UpToDate was acceptable among CHWs despite infrastructural barriers. Objective: This follow-up study evaluated the longitudinal adoption of UpToDate among CHWs and examined how sociocultural, political, and environmental factors influenced its use. Drawing on the nonadoption, abandonment, scale-up, spread, and sustainability framework, this study aimed to understand not only use patterns but also broader challenges to scale-up, spread, and sustainability in a complex health system. Methods: An explanatory mixed methods design was used combining analysis of use and program activity logs; program reports; and focus groups with CHWs, health care professionals, and program implementers. Quantitative analysis of use logs (March 2021 to September 2023) compared adoption over time by using descriptive statistics, CIs, and chi-square tests. Qualitative data came from the reanalysis of previous focus group transcripts and program reports and from a new focus group with program implementers. Reflexive thematic analysis was used to interpret how CHWs and implementers perceived and used the tool, and findings were integrated to explain quantitative trends. Results: Use of UpToDate was modest and declined over time. Monthly active use among CHWs and midwives fell substantially from 3.57% (97/2720 person-months) in 2021 to 1.07% (37/3456) in 2022 and remained low at 1.50% (39/2592) up to 2023, with markedly higher engagement in the rural site than in the remote site. Peaks in use coincided with program activities, whereas prolonged troughs followed typhoons, power outages, and other disruptions. Log data showed that users primarily consulted patient education articles rather than clinician-oriented decision tools. Qualitative analyses revealed that CHWs appropriated UpToDate as a learning aid and source of professional credibility. Structural shocks, heavy workloads, language barriers, and limited device access constrained individual use, and communal practices (shared devices and learning activities) meant that meaningful engagement often went unrecorded in vendor metrics. Conclusions: Our findings show that acceptability does not guarantee sustained use and that adoption cannot be captured fully by log-in counts. UpToDate’s value for CHWs lay in how it was domesticated as a tool for building capacity and professional credibility, not in its intended function as a decision aid used at the point of care. Therefore, evaluations of digital health tools should incorporate indicators of learning and social capital alongside use metrics. Policymakers should recognize that infrastructural fragility and communal adaptation shape digital health uptake. Embedding tools into ongoing training and peer learning structures, providing offline and multilingual support, and investing in resilience planning will be crucial for meaningful scale-up and sustainability.