All Patient Refined Diagnosis Related Groups: A New Administrative Tool for Identifying Elderly Patients at Risk of High Resource Consumption

To the Editor: The Health Care Financing Administration—Diagnostic Related Groups (HCFA-DRG) focused on the importance of the discharge abstracts from the hospital as a useful database for estimating the care reimbursement. Unfortunately, HCFA-DRG classification system does not give an accurate estimate of resource consumption of elderly patients.1 Recently the All Patient Refined-DRG (APR-DRG) have been validated as a useful tool for identifying the effect of clinical severity on resource consumption.2 APR-DRG is a patient classification system that is an enhancement of HCFA-DRG because it takes into account age and comorbidity risk adjustment. APR-DRG adds four subclasses to each DRG. These four subclasses are expressed as APR-Severity Index (SI): mild, moderate, severe, and extreme.3 Geriatric patients' overall care consumption is related not only to the severity of clinical impairment, but also to the functional, cognitive, and socioeconomic impairment of an individual3 as properly assessed using the Comprehensive Geriatric Assessment (CGA).4 Until now no studies have compared the APR-DRG data with the CGA in geriatric patients. With the aim of evaluating whether the APR-DRG system is a useful tool for grading clinical and functional impairment, as assessed using the CGA, in elderly inpatients, all elderly patients who were consecutively discharged from a geriatric ward from March 1 through April 30, 2003, were enrolled. Inclusion criteria were acute inpatient admission, aged 65 and older, and a CGA completed at admission. Age, sex, principal and secondary diagnoses, principal and secondary procedures, and length of stay (LOS, in days) were recorded and data included in the discharge abstracts. The discharge abstracts data were grouped using APR-DRG system version 12. At admission, a CGA was performed to evaluate patients' medical, functional, and cognitive aspects.5 Medical assessment included medication history and current pathologies. Functional status was evaluated using activity of daily living (ADL) and instrumental activity of daily living (IADL) indexes. Cognitive status was assessed using the Mini-Mental State Examination. Comorbidity was examined by using the Cumulative Illness Rating Scale (CIRS) comorbidity index, and the number of drugs prescribed to the patients were also recorded. Statistical analysis was performed using SPSS version 15 (SPSS Inc., Chicago, IL). The Kolmogorov-Smirnov test, the Kruskal-Wallis test, and the Pearson chi-square test were used. P=.05 was considered to be statistically significant. Two hundred eleven patients (107 men, 104 women, mean age ± standard deviation 76.3±7.3, range 65–96) were included in the study. Using the APR-SI, patients were divided into three groups: Group 1 (APR-SI=mild, n=65), Group 2 (APR-SI=moderate, n=111), Group 3 (APR-SI=severe or extreme, n=35). No differences were found between the three groups of patients in terms of age or sex. Significantly greater functional impairment (ADL, P<.01; IADL, P<.05), cognitive impairment (MMSE P<.05), comorbidity index (CIRS=P<.05), number of prescribed drugs (P<.05), and in-hospital length of stay (P<.05) was observed with higher APR-SI ratings (Table 1). A recent study6 reported that hospital costs were higher in older medical patients dependent in some basic ADLs on admission than in older patients independent in ADLs and that HCFA-DRG–based reimbursement did not account for this difference in costs. The HCFA-DRG classification is mainly based on resource intensity, which is predicted by the diagnoses and procedures recorded in the discharge abstract without taking into account age and comorbidity risk adjustment. The multifunctional impairment of elderly inpatients were undetectable using HCFA-DRG, as reported in two previous studies performed in frail elderly patients.7, 8 This study indicated that the APR-DRG system was capable of identifying three groups of patients with significantly different multidimensional impairments, as assessed using the CGA criteria. Dividing the patients according to the APR-SI, significantly different functional, cognitive, and comorbidity status were found. Furthermore, the number of prescribed drugs as assessed at discharge, a risk factor for disability,9 was significantly higher in patients with moderate and severe/extreme APR-SI than mild APR-SI. Similarly, LOS, an indirect index of costs caused by hospitalization, was progressively higher from mild to moderate to severe/extreme APR-SI groups. In conclusion, APR-DRG is an administrative tool useful in evaluating the clinical and functional impairment in elderly inpatients. This finding suggests a possible role of APR-DRG in identifying elderly inpatients at high risk of large health-resource consumption.