| This study contributes to ongoing research on validity and reliability of hospital patient safety measures derived from administrative data. Its main focus is on the coding of the complication of postoperative ileus. Initially, patient safety indicators were intended to be used as screening tools to identify adverse events and to improve quality of care. However, they are now used to assess and to publicly report performance by hospitals and physicians. A literature review revealed potential reasons for inaccurate coding of postoperative complications, including ambiguity in coding guidelines, interpretation of the guidelines and advice, coder bias, incomplete or misleading physician documentation, differences in definitions, lack of definitions, and coding unsupported by the clinical record. With the inception of the 2012 Coding Clinic regarding the coding of postoperative ileus (POI), the coding of POI increased by 24% between 2010 and 2012 at a mid-size non-profit hospital located in the northwest region of the United States where this study was conducted. Sources of variability in the coding of complications were analyzed to reveal how coding professionals arrive at decisions about coding of complications. Other aims were to examine any inconsistencies in the coding of the same cases by different coding professionals and to draw generalizations about the sources of coding variations, such as coder interpretation of the guidelines and advice, lack of uniform definitions, incomplete physician documentation, lack of communication with the physicians, and/or differences in the directions provided by ICD-9-CM Coding Conventions, Centers for Medicare & Medicaid Services (CMS) Official Guidelines for Coding and Reporting, and the American Hospital Association's (AHA) Coding Clinic Guidelines and Advice. Another aim was to verify other researchers' conclusions that current coding practices lack uniform and standardized application of regulatory rules and guidelines. Two hypothetica lmedical records (Case 1 and Case 2) with the documentation of POI were provided to 15 coders employed by the hospital that served as a study site. They coded the two cases and then completed a questionnaire on education, credentials, certification, experience, and application of CMS Coding Guidelines and AHA Coding Clinic advice. Only five out of 15 coders (33%) coded Case1 correctly by indicating, based on the documentation in the medical record, no POI complication. The case had documentation of POI but no physician linkage to support the coding of the complication of POI. Case 2 had strong physician linkage between the complication of POI and the surgery; 13 out of 15 coders(87%) coded Case 2 correctly. Coder education, credentials, certifications, and experience had no effect on the correct coding of these cases. The coders inconsistently prioritized rules and advice governing the coding of complications: ICD-9-CM Coding Conventions, CMS Coding Guidelines and AHA Coding Clinic. Next, the coders participated in a focus group led by a Health Information Management(HIM) expert in the field of HIM regulatory compliance, auditing and reimbursemen tmethodologies. The analysis of discussion on the sources of coding variability indicated that the coders valued experience more than education, credentials or professional affiliations. The application of the ICD-9-CM Coding Conventions, CMS Coding Guidelines and AHA Coding Clinic advice during the coding process was subject to the understanding and interpretation by the coder. This raises the possibility that coders perceive the rules and advice governing the coding of complications as conflicting guidance. Coders' reliance on subjective experience in dealing with this conflicting guidance may limit the accuracy of reporting outcomes of hospital and physician performance. It may also account, at least in part, for mostly incorrect coding of the complication of Case 1 POI, as well as for overall inconsistencies in general in the coding of ICD-9-CM Code 997.49, POI. |
