August 2023: Use of Mobile Phone Survey Methodology for Collecting Health Data

Gibson et al, Mobile Phone Surveys for Collecting Population-Level Estimates in Low- and Middle-Income Countries: A Literature Review, JMIR, 2017

Summary & Takeaways: 

MPS are increasingly used as a supplement to traditional population health household surveying. This literature review examined 6,625 articles for the use of MPS in LMICs and their effectiveness across three modalities: Computer-Assisted Telephone Interviews (CATI), Interactive Voice Response (IVR), and Short Message Service (SMS). The authors identified 11 articles containing 19 MPS to collect population health data across these modalities, with each having implications for downstream costs, metrics, and data quality. A majority (53%) of the MPS were CATI-based; fewer MPS used either IVR (32%) or SMS (16%). Due to the limited number of surveys using IVR or SMS, comparing the relative advantages and disadvantages across the modalities cannot be adequately assessed at this time. 

Comment from the Center for Global Digital Health Innovation:

Using MPS to collect population health data is in its nascent stage. To establish these modalities as robust, viable options, parameters for comparing costs, key metrics, and demographic representativeness should be included. Once MPS produces these valid and reliable data, their use can further benefit surveillance through lower average costs, more frequent survey implementation, and improved information collection leading to more responsive healthcare delivery. These benefits are key goals of digital health approaches and tools.

Assefa et al, COVID-19 Preventive Practices, Psychological Distress, and Reported Barriers to Healthcare Access during the Pandemic among Adult Community Members in Sub-Saharan Africa: A Phone Survey, The American Journal of Tropical Medicine and Hygiene, 2023 

Summary & Takeaways: 

This MPS survey, conducted by the Africa Research, Implementation Science, and Education (ARISE) Network included 9 sites across 5 countries. The authors examined practices, distress, and barriers to access during the COVID-19 pandemic. An estimated 12% of adults reported difficulty accessing health services, with 20% reporting psychological distress due to COVID-19. The results indicate a continued need to break barriers to access and reduce the economic burdens of healthcare throughout the sub-Saharan African region. While the authors did not have pre-pandemic data, the diversity of sites and locations aimed to reduce biases and aid reporting on the study parameters through the use of CATI, the gold standard for phone-based surveys. This attention to standardized tools and methods allowed timely and important assessments and comparisons which can help policymakers and managers build resilient health systems, inform monitoring, and mitigate negative impacts of pandemics across the African continent.

Comment from the Center for Global Digital Health Innovation:

On the whole, the study found CATI to be both useful and informative as a survey modality, particularly when used with localization language techniques and incorporating local cultural knowledge. Further, CATI enables remote and rapid data collection. While this study relied on CATI exclusively, the authors discovered limitations to this modality when it is used for limited duration surveys. Due to this, the survey was therefore unable to capture fully the types of services and factors influencing access to healthcare and psychological distress.

Song et al, Using Mobile Phone Data Collection Tool, Surveda, for Noncommunicable Disease Surveillance in Five Low- and Middle-income Countries, Online Journal of Public Health Informatics, 2020

Summary & Takeaways: 

This non-communicable disease (NCD) MPS used an open source, multi-modal software (Surveda) specifically developed for this study. It reviewed the capabilities of the software for data collection and improving public health surveillance in 5 LMICs – Zambia, Philippines, Morocco, Malawi, and Sri Lanka, collecting 23,682 interviews over 3 years. Surveda collects data via IVR, SMS, and/or a mobile web portal, and aims to be a powerful supplement to face-to-face data collection methods. The study found that automated MPS through Surveda provided features such as randomized questions as well as accurate and timely data collection for NCD surveillance. It demonstrated the viability of MPS use in this capacity and as a reliable source of information for policymaking.

Comment from the Center for Global Digital Health Innovation:

The use of MPS modalities is making striking inroads for data collection and disease surveillance. As the mobile phone industry grows globally, the opportunities both for capturing greater volumes of data and disseminating timely, information rich data grows with it. A flexible platform such as Surveda can deliver targeted, mixed modal surveys through a secure portal, and circulate the resulting data effectively to decisionmakers and researchers. With the goal of implementing digital health in the right way, in the right place, at the right time, Surveda was created to optimize advances in global technology with advances in public health surveillance and disease-limiting goals.

Nagpal et al, Who do phone surveys miss, and how to reduce exclusion: recommendations from phone surveys in nine Indian states, BMJ Global Health, 2021

Summary & Takeaways: 

With the goal of estimating population statistics with lower bias, this practice paper focuses on the prevalence and effects of exclusion when using CATI MPS methodology and implementation. Using original data from five surveys across nine states in India, the authors examined sources of bias due to non-observation errors across three MPS phases: design, implementation and analysis. The aim is to provide a framework for extending the utility and effectiveness of CATI MPS use. The authors found non-coverage and non-response errors correlated with household income, religion, gender, and other relevant socioeconomic and demographic characteristics. Suggestions to improve the framework across two of the phases include coverage and yield rate (design) and engagement strategies around best dates/times (implementation). Specific opportunities in error reduction include adjusting sample estimates using post-stratification weights, adopting structured callback protocols, and offering airtime incentives.

Comment from the Center for Global Digital Health Innovation:

As the use of MPS gains ground for population health data collection in LMICs, examining the shortcomings and potential exclusion of population subsets is important. This paper puts together a framework of suggestions to improve data collection outcomes. It speaks to the importance of local knowledge and cultural sensitivities when designing and implementing a data collection strategy that includes MPS digital tools.

Menon et al, Five questions to consider when conducting COVID-19 phone research, BMJ Global Health, 2021

Summary & Takeaways: 

Drawing on prior research, participation in the COVID-19 Research Network (CORE Net), and the range of phone research conducted by CORE Net from April to November 2020, the authors offer five questions to reflect on and consider before using MPS for data collection and dissemination, particularly during a crisis such as the COVID-19 pandemic:

1. Is your research needed?
2. Can you reach your target population?
3. Can your data needs be met through a phone?
4. Can you minimize risk for your respondents?
5. Does your research meet the needs of policy makers?

By posing these questions, the researchers aim to strengthen the use of MPS for much-needed real-time data collection. Moreover, the researchers emphasize the importance of collecting data that are high-quality and generalizable enough to inform healthcare providers and policymakers.

Comment from the Center for Global Digital Health Innovation:

We endorse these questions, and add that the gain of MPS must be in alignment with both best practices in research methodology and cultural awareness / sensitivity to at-risk populations. Determining the efficacy and appropriate use for digital tools comes with the responsibility of understanding the effects of crises on individual behavior change, economic situations,  and potential sensitivities of intra-household dynamics across these factors.