Abstract:
Background In-person health care poses risks to health workers and patients during pandemics. Remote consultations
can mitigate these risks. The REaCH intervention comprised training and mobile data allowance provision for mobile
phones to support remotely delivered primary care in Africa compared with no training and mobile data allowance.
The aim of this study was to estimate the effects of REaCH among adults with non-communicable diseases on remote
and face-to-face consultation rates, patient safety, and trustworthiness of consultations.
Methods In these two independent stepped-wedge cluster randomised controlled trials, we enrolled 20 primary care
clusters in each of two settings (Oyo State, Nigeria, and Morogoro Region, Tanzania). Eligible clusters had 100 or more
patients with diabetes, hypertension, and cardiovascular or pulmonary disease employing five health workers.
Clusters were computer-randomised to one of ten (Nigeria) or one of seven (Tanzania) sequences to receive the
REaCH intervention. Only outcome assessors were masked. Primary outcomes were consultation, prescription, and
investigation rates, and trustworthiness collected monthly for 12 months (Nigeria) and 9 months (Tanzania) from
open cohorts. Ten randomly sampled consulting patients per cluster-month completed patient reported outcome
measures. This trial was registered with ISRCTN, ISRCTN17941313.
Findings Overall, 40 clusters comprising 8776 (Nigeria) and 3246 (Tanzania) patients’ open cohort data were analysed
(6377 [72·7%] of 8776 females in Nigeria, and 2235 [68·9%] of 3246 females in Tanzania). The mean age of the
participants was 55∙3 years (SD 13∙9) in Nigeria and 59∙2 years (14∙2) in Tanzania. In Nigeria, no evidence of change
in face-to-face consulting rate was observed (rate ratio [RR] 1∙06, 95% CI 0∙98 to 1∙09; p=0∙16); however, remote
consultations increased four-fold (4∙44, 1∙34 to >10; p=0·01). In Tanzania, face-to-face (0∙94, 0∙61 to 1∙67; p=0∙99)
and remote consulting rates (1∙17, 0∙56 to 5∙57; p=0∙39) were unchanged. There was no evidence of difference in
prescribing rates (Nigeria: 1∙05, 0∙60 to 1∙14; p=0∙23; Tanzania: 0∙92, 0∙60 to 1∙67; p=0∙97), investigation rates
(Nigeria: 1∙06, 0∙23 to 2∙12; p=0·49; Tanzania: 1∙15, 0∙35 to 1∙64; 0·58) or trustworthiness scores (Nigeria: mean
difference 0∙05, 95% CI –0∙45 to 0∙42; p=0·89; Tanzania: 0∙07, –0∙15 to 0∙76; p=0∙70).
