Original articles
The Use of Automated Data to Identify Complications and Comorbidities of Diabetes: A Validation Study

https://doi.org/10.1016/S0895-4356(98)00161-9Get rights and content

Abstract

We evaluated the accuracy of administrative data for identifying complications and comorbidities of diabetes using International Classification of Diseases, 9th edition, Clinical Modification and Current Procedural Terminology codes. The records of 471 randomly selected diabetic patients were reviewed for complications from January 1, 1993 to December 31, 1995; chart data served to validate automated data. The complications with the highest sensitivity determined by a diagnosis in the medical records identified within ±60 days of the database date were myocardial infarction (95.2%); amputation (94.4%); ischemic heart disease (90.3%); stroke (91.2%); osteomyelitis (79.2%); and retinal detachment, vitreous hemorrhage, and vitrectomy (73.5%). With the exception of amputation (82.9%), positive predictive value was low when based on a diagnosis identified within ±60 days of the database date but increased with relaxation of the time constraints to include confirmation of the condition at any time during 1993–1995: ulcers (88.5%); amputation (85.4%); and retinal detachment, vitreous hemorrhage and vitrectomy (79.8%). Automated data are useful for ascertaining potential cases of some diabetic complications but require confirmatory evidence when they are to be used for research purposes.

Introduction

Diabetes imposes a tremendous burden on individuals and health care systems owing in part to increased morbidity and mortality from a wide range of complications [1]. Automated data hold promise as a cost-effective and efficient mechanism for identifying these complications [2]. The ability to identify complications and comorbidity using administrative data would provide an important resource for tracking disease burden related to these conditions, selecting high-risk patients for intensive intervention, and evaluating the effect of changes in clinical management strategies and medication regimens on outcomes over time.

Automated data can also be a valuable resource for diabetes research [3]. Identifying complications and comorbidities for epidemiologic and health services research often requires time-intensive and costly efforts involving medical records review and individual patient follow-up. The use of automated data to identify patients with selected complications or outcomes would be a cost-effective approach for use in case-control and other studies [4]. Although automated data have been utilized to identify complications and outcomes of diabetes [5], with the exception of coronary heart disease [6] little is known about the validity of this approach. Having established a registry of enrollees with diabetes at a large staff-model health maintenance organization (HMO) in western Washington, we sought to determine whether administrative data sources could be used to identify validly the complications of diabetes.

Section snippets

Methods

Group Health Cooperative of Puget Sound (GHC) is a staff-model HMO in western Washington with approximately 400,000 enrollees. We identified a cohort of 8905 patients aged 18 and older, with type I or type II diabetes, who were continuously enrolled from January 1, 1992, through March 31, 1996, or until their death. The algorithm for identifying enrollees as having diabetes was adapted from that used for the Diabetes Patient Outcomes Research Team (PORT) Study [7]. Patients were defined as

Results

Almost all requested medical records sampled for the validation study (97.7%) were located. The majority of enrollees in the validation sample were older than 65 years of age, white, and had a duration of diabetes greater than 10 years (Table 2). Diabetes was confirmed in the medical record for 90.4% of reviewed charts. The laboratory data of the 37 subjects identified from the algorithm, without documentation of diabetes in the medical record, were reviewed (KMN). In 22 cases (59.5%), there

Discussion

Computerized databases maintained by HMOs include an enormous variety of information including administrative data for billing, accounting and utilization of services such as office visits and hospitalizations with coded diagnoses, and data including laboratory and pathology test results and pharmacy utilization. These data are a rich resource for program and policy analyses and for health services, economic, outcomes, and epidemiologic research. The advantages of such data include ready

Acknowledgements

The funding source for this work was an unrestricted grant from Parke-Davis, Inc.

References (14)

  • D.G. Ives et al.

    Surveillance and ascertainment of cardiovascular eventsThe Cardiovascular Health study

    Ann Epidemiol

    (1995)
  • Guiss LS, Herman WH, Smith PJ. Mortality in non-insulin-dependent diabetes. In: Diabetes in America, 2nd Edition....
  • J.V. Selby

    Linking automated databases for research in managed care settings

    Ann Intern Med

    (1997)
  • T.K. Young et al.

    Estimated burden of diabetes mellitus in Manitoba according to health insurance claimsA pilot study

    CMAJ

    (1991)
  • B.M. Psaty et al.

    An approach to several problems in using large databases for population-based case-control studies of the therapeutic efficacy and safety of anti-hypertensive medicines

    Stat Med

    (1991)
  • J.V. Selby et al.

    Excess costs of medical care for patients with diabetes in a managed care population

    Diabetes Care

    (1997)
  • S. Greenfield et al.

    The uses of outcomes research for medical effectiveness, quality of care, and reimbursement in type II diabetes

    Diabetes Care

    (1994)
There are more references available in the full text version of this article.

Cited by (125)

View all citing articles on Scopus
View full text