JMIR Publications

ABSTRACT

Background:

The FeverApp registry uses ecological momentary assessment (EMA) to collect parental data on pediatric fever for scientific research. The mobile application, FeverApp, educates parents in safe fever management and serves as a fever diary.

Objective:

The focus of this analysis is an evaluation of completeness and concordance with regards to data quality of this EMA based registry in a multi-level perspective.

Methods:

Structured descriptions of fever episodes by healthcare professionals from an office are used as reference. Number of children, their sociodemographic data and agreement of fever episodes, with maximum temperature, intake of antipyretics, antibiotics and physician visits are compared with the entries in the corresponding doctor’s reference records. The data quality indicators for completeness, meaning the extent to which the necessary data that could be included in the registry has actually been submitted, and concordance, which is the correspondence of the value of a data element with a reference source, were chosen to proof whether EMA may be a sufficient method for this kind of registry.

Results:

In both data sources, 1012 children are available for comparison over 16 months. The completeness of gender (100%) and date of birth (99.2%) is high, the mismatches are 0.7% and 1.9% between the sources. Of these 1012 children, 668 have registered fever episodes in the FeverApp. They relate to 534 families with 953 fever episodes in the reference records and 1452 episodes in the FeverApp registry. Of 534 families in both sources, 183 (34.3%) refrain from visiting the office during fever episodes but nevertheless document them in the FeverApp. 766 (52.8%) of these episodes from 183 families were recorded exclusively in the FeverApp registry. The remaining 686 episodes (47.2%) of 391 children from 351 families are comparable to the reference data source in terms of medication and physician visits. Of the attending 351 families due to acute fever episode, 37.7% do not document the fever episodes additionally in the registry. Completeness ranges, depending on the kind of variable, from 11.5% to 65% in the registry and from 7.6% to 42.6% in the office. The 953 fever episodes reported by the reference office consists of 681 acute and 272 past episodes. In the FeverApp most past (96.3%), but less acute, episodes (62.3%) have been entered. The concordance rates varied from 90.2% (antibiotics), 67.3% (antipyretics), 61.7% (physician visits) and exact highest temperature during the fever episode (16%).

Conclusions:

Both sources deliver only partial data and rate of completeness and concordance depends on the kind of variable. However, the FeverApp registry shows a higher documentation and precision rate than professional records in all considered variables. Therefore, EMA may play a unique supplement for research in ambulatory care. The FeverApp could support pediatric offices, especially during the pandemic. Clinical Trial: drks.de DRKS00016591