BACKGROUND: The Primary Care Information Project at the New York City (NYC) Department of Health and Mental Hygiene assists providers in the adoption and use of electronic health records (EHRs), and queries health data from these EHRs using the Hub Population Health System (Hub). Covering up to two million patients seen by over 700 outpatient practices, the Hub provides a means to evaluate the completeness and representativeness of provider reporting to the NYC Bureau of Tuberculosis Control (BTBC). The NYC Health Code mandates reporting of children <5 years who have a positive test for TB infection to the BTBC.
METHODS: All children <5 years who visited their provider in 2013 and had an International Classification of Diseases (ICD-9) code for latent TB infection (LTBI) ever recorded in their EHR were queried from the Hub. Similarly, all children <5 years who were reported to BTBC with LTBI in 2013 were abstracted from the TB Registry, excluding those with known contact to a TB patient. Providers reporting ≥1 child with LTBI to the BTBC were matched by their practices’ name and address to facilities listed in Hub data.
RESULTS: Within Hub data, 1866 children with a patient visit in 2013 had an ICD-9 code for LTBI documented in their EHR in 2013 or prior years; 73% of these were at 4 practices which may be associating a LTBI diagnosis code with LTBI testing. The BTBC received reports for 131 children with LTBI in 2013. Ninety-two Hub practices had ≥1 child with LTBI; the majority of these were independent primary care practices (86%). Conversely, among 38 practices that reported ≥1 child with LTBI to BTBC, 16 (42%) were hospitals or hospital-affiliated clinics. When comparing Hub and BTBC practices, 4 healthcare facilities were in both datasets.
CONCLUSIONS: While Hub data captured children with a patient visit in 2013 who ever had an ICD-9 code for LTBI documented and BTBC data included children reported with LTBI in 2013, the substantial difference between children with LTBI identified in these datasets indicates that LTBI in this population is likely under-reported. However, non-specific use of ICD-9 codes may also contribute to this difference. Nevertheless, many providers from the Hub did not report to the BTBC. These results identified challenges to using EHRs for surveillance evaluation, and will help inform BTBC provider outreach by identifying practice types where reporting may be especially low.