Validation of the FIMA Questionnaire for Health-Related Resource Use Against Medical Claims Data: The Role Played by Length of Recall Period,Validierung des FIMA-Fragebogens zur Inanspruchnahme von Versorgungsleistungen anhand von Routinedaten der Krankenversicherung: Welchen Einfluss hat der Erinnerungszeitraum?

AIM TO VALIDATE: the questionnaire on health-related resource use in an elderly population (FIMA).

METHODS: Self-reported health care use of 1,552 participants was validated against medical claims data. Reliability was measured by intraclass correlation coefficient (ICC), sensitivity, specificity, and Cohen's Kappa. Linear regression models were used to investigate the association between validity and individual characteristics, health state, recall period (3, 6, or 12 months), or frequency of resource use.

RESULTS: On average, participants were 74 years old; 95% rated the questionnaire as easy. The number of physician contacts was underestimated depending on recall period by 9 to 28% and the ICC was moderate (3/6/12 months, ICC 0.46/0.48/0.55), whereas contacts with physiotherapists were remembered quite well (ICC>0.75). Remembering the number of days in rehabilitation and hospital differed by recall periods (3/6/12 months); rehabilitation ICC=0.88/0.51/0.87; hospital ICC=0.69/0.88/0.66. Very good reliability of self-reported long-term care insurance benefits was found for all recall periods (Kappa>0.90) while agreement in self-reported medical aid was poor (Kappa<0.30); agreement in intake of medication was good (Kappa>0.40). The chance of agreement between self-reports and claims data significantly decreased with the number of contacts. Individuals with better health had a significantly higher chance of reporting contacts with physiotherapists accurately.

CONCLUSION: The FIMA largely demonstrated good reliability. The FIMA is a coherent and valid instrument to collect health-related resource use in health economic studies in an elderly population.