Table 2.

Studies evaluating feasibility of smartphone-based screening program in LMICs, staff training, and on/off-site supervision.

Studies	Population	Intervention	Results, Comment
Asgary et al., 2020 [29]	Eswatini 25–49 y n = 247 HPV status = NR HIV-positive = 128 *	Smartphone-based VIA screening program, standard VIA training, refresher course, and 6-month mHealth mentorship.	Results: agreement 100% for positive cases and 95.7% for negative; kappa 0.74, then 0.64 after 3 months and 0.79 after 6 months,
Yeates et al., 2020 [30]	Tanzania >24 y n = 10,545 HPV status = NR HIV-positive = 2561 **	Smartphone-enhanced VIA platform (SEVIA) for “real-time secure sharing of cervical images”. Follow-up of the mean VIA+ rates after implementation of SEVIA. Evaluation of VIA images by providers and reviewers.	Results: VIA+ rates increased from 4 to 6.2% after implementation of SEVIA. Provider-Reviewer concordance rate = 90% over the 1-year period. Comment: SEVIA allows enhanced quality of visual inspection, training, real-time data acquisition, monitoring, and evaluation.
Asgary et al., 2019 [25]	Ghana mean age = 33.8 y n = 21 HPV status = NR	Providers’ perceptions and experiences: 15 nurses, 1 nurse supervisor, 1 expert reviewer.	Comment: cervical images provided peer-to-peer learning opportunities, better trust of patients, targeted education, and improvement of adherence, as well as implementation of quality control.
Quercia et al., 2018 [31]	Madagascar 30–65 y n= 151 HPV status = NR	Registration of cervical cancer screening program data onto a secure web-based platform, for monitoring purposes. Quality of data evaluation.	Results: less than 0.02% of key data missing. Comment: small group. Helps for real-time monitoring, but impact on women follow-up not assessed.
Sharma et al., 2018 [32]	India Mean age = 38.79 y n = 180 HPV status = NR	Assessment of nurses’ judgment for diagnosis of cervical pre-cancerous lesions using smartphone images.	Results: moderate nurse-expert agreement, kappa 0.45. Comment: appropriately trained nurses can reliably conduct screening. Real-time expert feedback might improve reporting.
Asgary et al., 2016 [33]	Ghana 25–45 y n = 169 HPV status = NR	Providers completed a 2-week on-site training in VIA, followed by a 3-month VIA training supported by text messaging by an expert reviewer (real-time feedback). Comparison of agreement rates for VIA+.	Results: total agreement rate, 95%, average agreement rate between each provider and expert reviewer 89.6%. Kappa 0.67
Peterson et al., 2016 [34]	Kenya Age = NR n = 824 HPV status = NR	Training of providers using pictures taken. Decision support “Job Aid tool” included in the mobile application (MobileODT system) for diagnosis and treatment.	Results: 12.6% pre-cancerous lesions, 0.7% suspected cancer. Comment: deployment of the “EVA System” allows monitoring of clinical decisions made by nurses. Help of “Job Aid” decision support for treatment and gives more confidence to providers.
Yeates et al., 2016 [26]	Tanzania 25–49 y n = 1072 HPV status = NR	Training providers to perform D-VIA with real-time support from regional experts, images sent through a smartphone application.	Feasibility of smartphone camera to perform “Enhanced VIA” and level of agreement between trainee and expert over time (agreement 96.8%), Response timing (real-time), 1–5 min 48.4% and <10 min 60% of the time.

Abbreviations: D-VIA (smartphone-based visual inspection with acetic acid), EVA (enhanced visual assessment), HIV (human immunodeficiency virus), HPV (human papilloma virus), mHealth (mobile health), NR (not reported), SEVIA (smartphone enhanced visual assessment), VIA (visual inspection with acetic acid), y (years old). * 128/247 women were HIV-positive. ** 2561/10,545 women were HIV-positive.