Undercoding of Alzheimer's disease and related dementias in hospitalized elderly patients in Italy.

The prevalence of Alzheimer's disease (AD) and AD-related dementias (ADRD) in acute ward-hospitalized elderly patients is not well known, owing principally to misclassification and undercoding of AD and ADRD on hospital discharge abstract forms (DAFs). The aims of this study were to evaluate the prevalence of AD and ADRD, as evaluated by the DAF in elderly patients hospitalized in acute wards, and to compare clinical severity, length of stay, comorbidity, and number of diagnostic procedures in patients with AD versus ADRD to explain the different reimbursement costs of DRG12 (AD) versus DRG429 (ADRD). From the inpatient DAF database of the Casa Sollievo della Sofferenza Hospital, the DAFs of patients aged 65 years or over discharged from January 1, 2001, to March 31, 2003, with principal or secondary diagnoses of AD (ICD9-CM code 331) or ADRD (ICD9-CM codes from 290.0 to 290.43) were extracted and grouped by APR-grouper version 12. Age, gender, length of stay, principal and secondary diagnoses and procedures, and APR-DRG severity index (SI) and mortality risk (MR) were evaluated in these patients. Senile dementia was reported in 294 patients (0.58 percent, N = 50,253). In 123 patients (41.8 percent) dementia was the principal diagnosis, whereas in 171 patients (58.2 percent) dementia was reported on the DAF as a secondary diagnosis. Of the 123 patients with a principal diagnosis of dementia, 35 patients were included in the DRG-12 (AD) and 88 patients were included in the DRG-429 (ADRD). No differences were found in mean age, length of stay, comorbidity, or number of diagnostic procedures, as well as in the APR-DRG SI and APR-DRG MR between AD and ADRD patients. Conversely, reimbursement amounts were established as Euro4,033 for DRG-12 (AD) and Euro2,952 for DRG-429 (ADRD). AD and ADRD are undercoded in elderly hospitalized patients. The limits of the ICD9-CM classification system and the influence of reimbursement amounts may influence the coding reports by physicians.