Validation of ICD-9-CM diagnostic codes for inflammatory bowel disease among veterans

Selvi Thirumurthi, Reezwana Chowdhury, Peter Richardson, Neena S. Abraham

Research output: Contribution to journalArticlepeer-review

63 Scopus citations


Background: Inflammatory bowel disease (IBD) is well described among young whites and less so among the elderly and non-whites. Population-level data is required to assess outcomes among minority groups. Aim: To validate diagnostic codes for IBD from the Department of Veterans Affairs. Methods: National databases were used to identify local patients with Crohn's disease (CD) and ulcerative colitis (UC), the extra-intestinal manifestations and surgical procedures associated with IBD. Diagnosis was confirmed by manual chart abstraction. Multivariable logistic regression was used to derive diagnostic algorithms for CD and UC, which were then validated in an independent cohort. Results: The test cohort of 3,827 patients (1,316 potential cases, 2,511 random controls) was predominantly male (94%), white (56%), and of age of 58 (standard deviation 15). The positive predictive value (PPV) of CD codes was superior (88-100%) to UC (50-93%). The accuracy of extra-intestinal manifestations and surgeries was poor (PPV 0-29%). ICD-9-CM code 555.x without 560.9 had a PPV of 91% for CD in the validation cohort. Code 556.x with age, gender, and race factors was highly predictive of UC (c-statistic 0.9, PPV of 81%). Conclusion: VA administrative data can diagnose elderly and non-white patients with IBD.

Original languageEnglish (US)
Pages (from-to)2592-2598
Number of pages7
JournalDigestive diseases and sciences
Issue number9
StatePublished - Sep 2010


  • Crohn's disease
  • Elderly
  • ICD-9-CM code
  • Positive predictive value
  • Ulcerative colitis
  • Validation studies
  • Veterans

ASJC Scopus subject areas

  • Physiology
  • Gastroenterology


Dive into the research topics of 'Validation of ICD-9-CM diagnostic codes for inflammatory bowel disease among veterans'. Together they form a unique fingerprint.

Cite this