Innovative Real-World Data Use for Identifying Stroke Survivors and Access to Rehabilitation in Primary Care in...

As the impact of stroke remains, primary healthcare will continue to be a critical platform managing the poststroke journey. We aimed to identify stroke survivors assisted by community health worker in Brazil and how they relate to the location of rehabilitation facilities locations.

We developed a cross-sectional study using deidentified data from a real-world database generated by a free data collection app used by community health workers from May 2015 to January 2021 in Brazil to identify stroke survivors and to assess demographics and clinical characteristics. We used data from a public database, Cadastro Nacional de Estabelecimentos de Saúde, for identifying rehabilitation facilities. Locations were obtained by a geocoding application programming interface (Google Maps Platform), distances were measured in kilometers, and travel time in minutes.

Among 2 397 764 individuals assisted by community health workers, 21 785 were stroke survivors, representing a 0.9% prevalence. Among this subgroup, the majority were in the Northeast region (n=10 951; 50.3%) and 16 922 (77.7%) in urban areas. Most individuals (n=11 504; n=142; 52.8%) were women, the mean age was 66.5 (SD, 14.7), and 4313 reported physical disability. In total, 348 rehabilitation facilities were identified, mostly located in the Southeast region (40.8%). The mean distance from stroke survivor to facility was 79.13 km (SD, 97.73; median [1Q, 3Q], 47.64 km [12.19, 107.80 km]), and mean travel time was 81.18 minutes (SD, 85.85). The Southern region recorded the largest mean and median distance (mean 175.58 km; SD, 163.18; median [1Q, 3Q] 88.47 [59.38, 425.38]) to rehabilitation center and the longest mean travel time (144.48 minutes; SD, 112.57; median [1Q, 3Q] 92.34 [60.59, 305.12]).

Despite the availability of rehabilitation centers in Brazil, geographic access as represented by the distances and travel times observed access is still suboptimal. As a means of improving the clinical pathway and resource allocation, the use of large real-world databases and adequate analysis may become a key component for real needs assessments.