Increasing access to eye care using mobile phone-based interventions : the development, validation and implementation of Peek to optimise human resources and lower barriers to access for those most in need

Background: A combination of limited access to eye services and low numbers of eye care providers in low and middle income (LMIC) populations results in high prevalence of avoidable visual impairment. Aim: To develop and evaluate a demonstration model of community volunteers and teachers using a novel mobile phone-based technology (Peek) in communities and schools, respectively, to identify and refer those with referable eye conditions and increase adherence to services so as to reduce avoidable visual impairment. Methods: This thesis comprises: (1) a three-year retrospective review of utilisation of hospital eye care services; (2) a cluster randomized trial (C-RCT) to determine the effectiveness of using of the mobile phone based, Peek School Eye Health System (Peek SEH) to increase identification and referral adherence to hospital of school pupils with visual impairment; (3) the development and validation of a smartphone based community screening decisionsupport algorithm (Peek Community Screening App) that enables Community Volunteers (CVs) to make referral decisions about patients with eye problems in the community; and (4) a second single masked C-RCT where the Peek Community Screening App was integrated into an mHealth system, Peek Community Eye Health system (Peek CEH) for integrating eye health screening in communities with primary and secondary health care facilities in Kenya to increase access and to optimise health system utilisation. Results For the first study, the retrospective analysis of records showed an average annual attendance rate increase from 60.9 to 79.2 per 10,000 population, incidence rate ratio (IRR) 1.30 (95% confidence interval (CI) 1.26–1.35) between 2013 and 2015. Also 61.0% of consultations in the three-year period were for primary eye conditions (allergic or other conjunctivitis or normal eyes) which could potentially be managed by primary eye care (PEC) and only 8.3% were for the three leading causes of vision loss in this population (cataract, glaucoma and refractive errors). In the second study, Peek SEH was validated by comparing the results of Peek Acuity test and the Snellen Tumbling-E card when performed by teachers against a reference standard backlit EDTRS LogMAR visual acuity test chart, when performed by trained ophthalmic workers. Sensitivities were similar (77% [95% CI 64·8–86·5] vs 75% [63·1–85·2]). In the CRCT, the Peek SEH intervention, comprising of Peek Acuity test, sight simulation referral cards, and short message service [SMS] reminders, was compared to standard Tumbling E card vision test and paper referral letters. The proportion of pupils identified as having visual impairment who attended their hospital referral was higher in the intervention group (285 [54%] of 531) than in the standard group (82 [22%] of 366; odds ratio 7·35 [95% CI 3·49–15·47]; p<0.0001). This result informed the design of a school project that subsequently screened a further 168,820 children, identified 6,696 (4.0%) with VI and achieved 93% treatment coverage. In the third study, the sensitivity of community volunteer referral decisions using the “Peek Community Screening App” as compared to decisions made by the reference Ophthalmic Clinical Officer was 91.0%, (95% CI 87.7% - 93.7%) and Specificity was 78.1%, (95% CI 71.6% - 83.6%). The fourth study, compared the Peek CEH (comprising Peek Community Screening app, short message service [SMS] reminders and door to door screening), delivered by CVs against current care (periodic health centre-based outreach clinics with onward referral). The intervention was associated with increased utilization to PEC services at four weeks from sensitization. The mean attendance was 14.3% vs. 5.2%, risk difference 9.1% (95%CI: Hillary Rono – PhD thesis – October 2019 6.9-11.3%); p<0.00001. Overall, 76% of participants were treated at PEC level. About 11% of hospital consultations were for primary eye conditions and 56% for cataract, glaucoma and refractive errors (compared to 63% and 8% respectively in the hospital utilisation study, in the first study). Conclusion: This study provides strong evidence that integration of Peek interventions in to the health system can increase access to eye care, whilst making more appropriate use of limited eye care resources.