BMC Medical Research Methodology (Jun 2021)

Validity of using mobile phone surveys to evaluate community health worker program in Mali

  • Xiaomeng Chen,
  • Diwakar Mohan,
  • Abdoulaye Maïga,
  • Emily Frost,
  • Djeneba Coulibaly,
  • Luay Basil,
  • Birahim Yaguemar Gueye,
  • Mariam Traore Guindo,
  • Assa Sidibé Keita,
  • Haoua Dembelé Keita,
  • Melissa A. Marx

DOI
https://doi.org/10.1186/s12874-021-01317-7
Journal volume & issue
Vol. 21, no. 1
pp. 1 – 10

Abstract

Read online

Abstract Background The monitoring and evaluation of public health programs based on traditional face-to-face interviews in hard-to-reach and unstable regions present many challenges. Mobile phone-based methods are considered to be an effective alternative, but the validity of mobile phone-based data for assessing implementation strength has not been sufficiently studied yet. Nested within an evaluation project for an integrated community case management (iCCM) and family planning program in Mali, this study aimed to assess the validity of a mobile phone-based health provider survey to measure the implementation strength of this program. Methods From July to August 2018, a cross-sectional survey was conducted among the community health workers (ASCs) from six rural districts working with the iCCM and family planning program. ASCs were first reached to complete the mobile phone-based survey; within a week, ASCs were visited in their communities to complete the in-person survey. Both surveys used identical implementation strength tools to collect data on program activities related to iCCM and family planning. Sensitivity and specificity were calculated for each implementation strength indicator collected from the phone-based survey, with the in-person survey as the gold standard. A threshold of ≥ 80% for sensitivity and specificity was considered adequate for evaluation purposes. Results Of the 157 ASCs interviewed by mobile phone, 115 (73.2%) were reached in person. Most of the training (2/2 indicators), supervision (2/3), treatment/modern contraceptive supply (9/9), and reporting (3/3) indicators reached the 80% threshold for sensitivity, while only one supervision indicator and one supply indicator reached 80% for specificity. In contrast, most of the stock-out indicators (8/9) reached 80% for specificity, while only two indicators reached the threshold for sensitivity. Conclusions The validity of mobile phone-based data was adequate for general training, supervision, and supply indicators for iCCM and family planning. With sufficient mobile phone coverage and reliable mobile network connection, mobile phone-based surveys are useful as an alternative for data collection to assess the implementation strength of general activities in hard-to-reach areas.

Keywords