[Geriatric multimorbidity in claims data: part 3: prevalence and predictive power of geriatric conditions in an age-specific systematic sample].

BACKGROUND: In various contexts, the identification of insurants with geriatric conditions (GC) can offer new approaches for specific medical services. GC can be determined from diagnoses data of insurants retrieved from different care sectors, and supplemented with other relevant claims data, e.g., long-term care levels and pharmaceutical data.

METHODS: Part 3 of this study is based on a systematic sample of 957,447 AOK insurants (age ≥ 60 years). Prevalence of 15 GC was investigated using anonymous claims data of diagnoses from physicians in the ambulant care setting and diagnoses from hospital settings in 2008. In addition the potential relationships of GC with mortality, nursing home admission, need for long-term care and hospital utilization in the following year were examined. All results were standardized by gender and age based on the general population aged ≥ 60 years in Germany.

RESULTS: Pain and impairment of vision or hearing was the most common GC (> 25%) followed by high risk of complications, fall risk/dizziness, and cognitive deficit (8-14%). Delayed convalescence, frailty, medication problems, immobility and malnutrition occurred in < 1% of the insurants. Almost all GC occurred more often with increasing age. Only 37% insurants in the sample showed no GC, while for 31% exactly one, for 17% two, and for 15% three or more GC were observed. With the exception of pain and impairment of vision or hearing all of the GC had a significant positive association with mortality, nursing home admission, increasing need of care, and hospital utilization in the following year.

CONCLUSIONS: The applied operational approach proved to be generally practicable and successful with few adaptations. The GC pain and impairment of vision or hearing, however, do not contribute sufficiently to the identification of geriatric multimorbidity based on claims data. These GC should be therefore disregarded from such identification processes. To enhance the reliability of an identified geriatric multimorbidity, the requirements on the specificity and number of individual GC (two, three, or more) can be adapted.