This article is more than 5 years old. Information may no longer be current.
Codes for RA, GCA often not recorded at hospitalization; research based on codes may be flawed
After a patient is hospitalized, diagnosis codes for rheumatoid arthritis and giant cell arteritis are often omitted, which could yield inaccurate results when research is conducted based on the use of medical codes, according to research from the Mayo Clinic.
Researchers identified 499 patients with rheumatoid arthritis (RA) who had 2,407 hospitalizations and 119 patients with giant cell arteritis (GCA) who had 502 hospitalizations from the Rochester Epidemiology Project and conducted a retrospective, population-based study.
Clinical Classification Software (CCS) was used to define 18 chapters, which included infections and parasitic diseases; neoplasms; endocrine, nutritional, and metabolic diseases and immunity disorders; mental illness; diseases of the circulatory system; diseases of the digestive system; complications of pregnancy, childbirth, and puerperium; diseases of the skin and subcutaneous tissue; diseases of the musculoskeletal system and connective tissue; congenital anomalies; symptoms, signs, and ill-defined conditions; and residual codes, unclassified and other categories.
A diagnosis code for RA occurred more frequently in younger patients 67.2 years, in patients with longer disease duration and in more recent hospitalizations. In three CCS chapters, a significant difference was seen in primary discharge diagnoses. When the primary discharge diagnoses were related to musculoskeletal and connective disuse diseases, RA was more likely to be cited in the discharge diagnosis. The least likely diagnosis to include RA was in the presence of a primary discharge diagnosis involving neoplasms, or signs, and ill-defined conditions.
Patients who were seropositive for rheumatoid factor or anti-cyclic citrullinated peptide antibodies had a higher chance of being mentioned. Joint erosions, elevated erythrocyte sedimentation rate and rheumatoid nodules were not associated with the odds for having a diagnosis with RA.
A diagnosis of GCA was more likely to occur with younger patients, shorter disease duration and the use of corticosteroids. In the case of a primary diagnosis of a circulatory system disease, GCA was slightly more likely to be included in the diagnosis but the effect did not reach statistical significance. – by Shirley Pulawski
Disclosure: The researchers report the work is supported by a grant from the National Institutes of Health, NIAMS (R01 AR46849), and made possible by the Rochester Epidemiology Project (R01 AG034676 from the National Institute on Aging.
Perspective
Back to Top
Research into multiple domains of what is often referred to as health systems research deals with everything from cost-effectiveness research to diagnostic and therapeutic efficiency to safety and epidemiologic surveillance. It is predominantly gleaned from large-scale study of medical records. Previous studies have demonstrated the high predicative value of using ICD-coding diagnoses, which is reassuring.
What appears not so reassuring are data from inpatient databases which rely on clinicians to clearly identify rheumatic disease patients hospitalized for any reason. In other words, in a hypothetical scenario of 100 patients with rheumatoid arthritis (RA) who are hospitalized for community-acquired pneumonia and then discharged, how often are the patients accurately identified as having RA as a comorbidity? According to investigators from the Mayo Clinic, it is not very often. In their study of the unique population of Olmstead County, which has a very high presumed standard of medical care, RA or giant cell arteritis were accurately identified on hospital records 55% and 31% of the time in the study which covered the years 1996 to 2012. While there are some caveats in interpreting the findings, I find the data alarming and disappointing.
Several years ago, Eamonn Molloy, MD, MS, MRCPI, and I reported on the frequency of progressive multifocal leukoencephalopathy (PML) in hospitalized patients with RA and systemic lupus erythematosus (SLE) using the same inpatient database investigated in the current study. On the basis of hospital codings for RA and SLE, we developed estimates of prevalence PML in patients hospitalized with the diagnosis and RA and SLE. If the data from the Michet study can be generalized, then we may have missed 45% of RA patients alone. The implications for studies using hospitalized databases is that such data need to be looked at with scientific skepticism. Furthermore, new data analysis systems need to be developed and refined for health services research.
Reference:
Molloy ES, et al. Arthritis Rheum. 2009;doi:10.1002/art.24966.