Using SMS data to improve Ebola models and estimate the impact of Ebola on healthcare seeking behaviour in Liberia
The current Ebola crisis of West Africa has ravaged the health and social sectors of affected countries and has also had pronounced effects on the economies of these countries. Liberia has seen the most deaths, at 4,808 to date. Despite the significance of the outbreak, there has been little data available to measure the extent to which this outbreak has directly affected households in West Africa, in particular with regards to their non-Ebola health and health seeking behavior.
Using a Short Message Service (SMS) platform in order to reach a large and diverse set of people around the country, this project aims to fill in some of these information gaps in Liberia. We believe that this project will help to improve our understanding of how outbreaks of this nature can affect different demographic groups, including both urban and rural populations. Our findings will be of great interest to both international actors involved in helping to contain the epidemic but also to national actors who are interested in mitigating the impact of the outbreak on households. We also believe the data collected in this project can help inform preparation efforts for future outbreaks.
Using a commercial SMS polling platform, this project will survey approximately 10,000 Liberians from early March through early June 2015 to understand their exposure to the outbreak, as measured by the number of people they know who acquired Ebola, their healthcare seeking behavior over the course of the epidemic, as well as any economic effects of reduced health seeking behavior. The SMS strategy will enable us to reach individuals from areas all around the country, and evaluate differences between subpopulations (male vs. female, between different age groups and urban vs. rural populations), identify where these differences may arise, and examine temporal changes aligned with spatio-temporal intervention data in each of the groups. Thus within this uncertain dynamics of the outbreak, our data will provide an assessment of the how this epidemic is and has affected the health and health seeking behavior of these households, now and as the epidemic moves into the recovery phase.
To date, we have developed an appropriate set of questions, incorporating appropriate wording and translation fo Liberian pigeon english where needed. We have also designed the survey to enable comparison with data from other survey efforts in the country, and deployed the first survey resulting in the anticipated number of responses (roughly 2000 responses to date). Going forward, the project will conduct three additional rounds of data collection to increase statistical significance of our results, and examine trends over time.
We will compare the data on health seeking behavior to the nationally representative data collected in the 2013 Demographic and Health Survey in Liberia, and estimate what impact the epidemic has had on the use of health services in affected regions and whether it is worse in some areas relative to others. Finally, by repeating the survey to representative samples at a set frequency, we can also collect data on indicator changes during the epidemic and recovery process in Liberia.
The data and our analysis will be published and shared in order to allow planners to better understand the spread of outbreak in real time, which can help to target resources more effectively and minimize the negative impact of the outbreak on households in the long run, enable academics to generate more accurate models.
