Reliability of Data Collected via Ecological Momentary Assessment on the Example of FeverApp Registry.

The FeverApp registry is an ambulant ecological momentary assessment (EMA) model registry focusing on research of fever in children. Verification of EMA reliability is a challenge, due to absence of other source data. To ensure the reliability of EMA data, 973 families were invited to reassess their documentation in a survey. The survey contained questions (a) regarding the number of children, (b) genuineness of entries, (c) completeness of submitted fever episodes, (d) medication, (e) usefulness and further usage of the app. Of those invited, 438 families (45% response rate) participated in the survey. Of these, 363 (83%) families have registered all their children, 208 families have one child. The majority (n = 325, 74.2%) of families stated that they only made genuine entries in the app. Agreement between survey and app for fever episodes is 90% (Cohen's κ = 0.75 [0.66, 0.82]). Medication shows 73.7% agreement, κ = 0.49 [0.42; 0.54]. The majority (n = 245, 55.9%) consider the app as an additional benefit and 87.3% would like to use it further. Email surveys are a possible approach to evaluate EMA based registry data. The possible observation units (children and fever episodes) show an adequate reliability. With this approach, surveys of further samples and variables could help to improve the quality of EMA based registries.