Geographically linking population and facility surveys: methodological considerations

Abstract

BackgroundThe relationship between health services and population outcomes is an important area of public health research that requires bringing together data on outcomes and the relevant service environment. Linking independent, existing datasets geographically is potentially an efficient approach; however, it raises a number of methodological issues which have not been extensively explored. This sensitivity analysis explores the potential misclassification error introduced when a sample rather than a census of health facilities is used and when household survey clusters are geographically displaced for confidentiality.MethodsUsing the 2007 Rwanda Service Provision Assessment (RSPA) of all public health facilities and the 2007–2008 Rwanda Interim Demographic and Health Survey (RIDHS), five health facility samples and five household cluster displacements were created to simulate typical SPA samples and household cluster datasets. Facility datasets were matched with cluster datasets to create 36 paired datasets. Four geographic techniques were employed to link clusters with facilities in each paired dataset. The links between clusters and facilities were operationalized by creating health service variables from the RSPA and attaching them to linked RIDHS clusters. Comparisons between the original facility census and undisplaced clusters dataset with the multiple samples and displaced clusters datasets enabled measurement of error due to sampling and displacement.ResultsFacility sampling produced larger misclassification errors than cluster displacement, underestimating access to services. Distance to the nearest facility was misclassified for over 50% of the clusters when directly linked, while linking to all facilities within an administrative boundary produced the lowest misclassification error. Measuring relative service environment produced equally poor results with over half of the clusters assigned to the incorrect quintile when linked with a sample of facilities and more than one-third misclassified due to displacement.ConclusionsAt low levels of geographic disaggregation, linking independent facility samples and household clusters is not recommended. Linking facility census data with population data at the cluster level is possible, but misclassification errors associated with geographic displacement of clusters will bias estimates of relationships between service environment and health outcomes. The potential need to link facility and population-based data requires consideration when designing a facility survey.

Linking household survey and health facility data for effective coverage measures: a comparison of ecological and individual linking methods using the Multiple Indicator Cluster Survey in Côte d’Ivoire

“…The average or nearest quality score can be calculated in a number of ways, including an average quality of care score for an administrative area and/or provider category, the quality score of the closest provider (straight-line distance or road distance), or an average score for all providers within a certain radius of the household or cluster [12]. We conducted five types of ecological analyses, as described below.…”

“…In middle- and high-income countries with high quality individual medical records, other approaches may be possible. Another approach, ecological linking, involves linking each care-seeking episode in a household survey to an average quality of care score of providers within certain administrative or geographical boundaries of the household survey cluster, or the quality score of the nearest provider(s) (not necessarily the provider visited by the respondent) [12]. …”

Limited Service Availability, Readiness, and Use of Facility-Based Delivery Care in Haiti: A Study Linking Health Facility Data and Population Data

“…Because DHS cluster coordinates are displaced before release of the data (to ensure respondent confidentiality is maintained), the closest facility identified based on the released geographic data may not be the nearest facility in reality. Skiles and colleagues indicated that, due to displacement of cluster coordinates, the distance to the closest facility can be misclassified for 34% to 43% of clusters 21 . The displacement is an important limitation of the data, so instead of only looking at the closest facility our study measured the service environment (all available facilities) within a reasonable distance of the displaced cluster, thus representing the likely service environment of the real cluster location.…”

The Effect of Access to Contraceptive Services on Injectable Use and Demand for Family Planning in Malawi

“…18 Studies have linked individuals and clusters of households with the geographically closest facility, 19–24 yet this implicit assumption that the closest facility is the one used likely introduces substantial misclassification error. 25 Another common method, based on Euclidean distance measurements, has been to link a woman or cluster with one or more facilities within a specific buffer or polygon. 8,26 While this approach may reduce the misclassification error, the size of the buffer or polygon has implications for analysis and interpretation.…”