OBJECTIVE: The reliability of the diagnoses given on forms authorizing Hospital Admittance (AIH), financed by the official Health System, was analysed during the period January through December,1992, in Maringá-PR, Brazil. METHOD: A systematic method was used to sample 1,595 medical records of the admittances to 8 general hospitals. The diagnoses registered in the medical records were compared to those registered in the corresponding AIH forms. The level of agreement on the main diagnoses was analysed by using the Kappa test. RESULTS: The results of the study show that the agreement for the five leading chapters of International Classification Disease-9 (ICD-9) has ranged from k=0.79 to k=0.98. The agreement for the five leading groupings of diagnoses has ranged from k=0.79 to k=0.97. CONCLUSION: Some reasons for disagreement, such as miscoding and the hospital reimbursement system are discussed. It is also concluded that it may be possible to use the hospital admittance data of the Official Health System, according to the diagnostic chapters and groupings of the ICD-9, with some degree of confidence.
Medical records; Diagnosis; Hospital statistics; Reproducibility of results
