2019 Volume 70 Issue 2 Pages 53-61
The diagnosis is the most important element of medical data. This study focused on diagnoses collected in the Diagnosis Procedure Combination/Per-Diem Payment System and designed a method to assess their quality. Quality was defined based on the guidelines of ISO 9000:2015. The level-of-detail of diagnoses was chosen as their characteristic, and the uniformity of level-of-detail of diagnoses among hospitals was considered as a scale of data quality. Specifically, level-of-detail was measured based on the disease classification scheme used, and a target value for level-of-detail was set to assess the quality of diagnoses. This study focused on diagnoses of primary malignant tumors. Their level-of-detail value was the sum of measurements of the following four classification axes: laterality, initial site, pathological diagnosis, and staging. Analyses using data from three general hospitals revealed that one hospital reported diagnoses with a level-of-detail below the target value. Further, non-uniformity among hospitals was indicated. The three hospitals utilized the same process for reporting diagnoses. Specifically, physicians wrote diagnoses on the medical records and administrative staff transcribed them to the medical bills. It was found that the work analysis performed by physicians and administrative staff in that process is important for improving the quality of diagnoses. The present findings suggest the possibility of developing a new hospital evaluation index.
