[Quality of documented diagnosis in primary care - An analysis using the example of thyroid disorders]

Background: Routine data of statutory health insurance (SHI) provide a great potential for evaluating the healthcare situation in Germany by providing, e. g., morbidity estimates. In the context of secondary data, analyses based on insurance data were conducted using the outpatient medical billing diagnosis. If, however, medical billing data, which are collected for this primary purpose, are the only source of these morbidity estimates, a limited validity of outpatient medical diagnostic data will have to be assumed or validity will have to be assured. This investigation aimed to analyse the quality of family practitioners' documentation regarding diagnostic data in patients with thyroid disease based on medical records for billing purposes.

Methods: As part of the pilot study General practitioners' Views on Polypharmacy and its Consequences for Patient Health Care, the medical records of 548 multi-morbid patients with thyroid disease from six general practices in Dresden, Saxony, were analysed with respect to the congruence of prescriptions of thyroid medication and associated thyroid diagnosis. Logistic regression was used to investigate predictors of limited diagnostic data of thyroid disease.

Results: There was insufficient documentation of thyroid diagnoses in 26.8% (n=147) of the included patients diagnosed with or treated for thyroid disease. The proportion of undocumented (1.1 to 35.8%), imprecise (4.6 to 22.3%) and non-specific documented (14.9 to 73.8%) thyroid diagnoses varied in all general practices. Due to undocumented thyroid disease, the corrected prevalence of thyroid diseases summarised for all practices was 5.5 percentage points higher than the originally documented prevalence (29.7 instead of 24.2%). An increasing number of prescribed drugs was a significant predictor for inadequate documentation of thyroid disease (for 5 to 8 routinely taken medications: OR=2.4/p<0.001; for 9 to 12: OR=4.0/p<0.001; for 13 to 20: OR=7.4/p<0.001).

Conclusions: Due to the limited data quality of outpatient medical billing diagnoses as a basis for morbidity estimates, GPs' diagnostic data should be subjected to regular internal and external diagnostic validation in SHI routine data. Additional case-related interviews with documenting GPs would significantly increase data validity. Also, intelligent e-tools supporting electronic patient documentation could be helpful to improve the quality of primary care documentation.