Diagnostic Coding and Medical Rehabilitation Length of Stay

Their Relationship

Published in: Archives of Physical Medicine and Rehabilitation, v. 79, no. 3, Mar. 1998, p. 529-548

Posted on RAND.org on January 01, 1998

by Margaret G. Stineman, Jose J. Escarce, Charles J. Tassoni, James E. Goin, Carl V. Granger, Sankey Williams

Read More

Access further information on this document at www.sciencedirect.com

This article was published outside of RAND. The full text of the article can be found at the link above.

The objective of this study was to determine if diagnostic information provided in the form of International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) codes improves rehabilitation length of stay (LOS) prediction when used in combination with the Functional Independence Measure-Function Related Groups (FIM-FRGs) classification system. Various models characterizing diagnostic information using ICD-9-CM codes were created that included individual ICD-9-CM codes and groupings of those codes by organ or etiology involved. Each method was evaluated using linear regression with the natural logarithm of LOS as the dependent variable. Separate validation data sets were held back to quantify the incremental effect of diagnosis when combined with the FIM-FRG classification system. The addition of ICD-9-CM diagnostic information to the FIM-FRG classification system increased the variance explained by a maximum of 1.9%, from 31.5% to 33.4%. Refinement of the FIM-FRGs to include ICD-9-CM diagnoses does not appear warranted on the basis of the small increase in the percentage of explained variance in LOS. The authors believe the lack of improved prediction with the addition of ICD-9-CM codes relates primarily to incomplete coding practices and to the effect of patients' diagnoses being absorbed in variables as already expressed by the FIM-FRG system. Although ICD-9-CM codes, overall, did not greatly improve LOS prediction, they appeared to have some impact in certain impairment categories.

This report is part of the RAND Corporation External publication series. Many RAND studies are published in peer-reviewed scholarly journals, as chapters in commercial books, or as documents published by other organizations.

The RAND Corporation is a nonprofit institution that helps improve policy and decisionmaking through research and analysis. RAND's publications do not necessarily reflect the opinions of its research clients and sponsors.