
Methods: We examined a retrospective cohort of patients aged 18-64 years admitted to Duke University Health System (which includes a tertiary care center and two regional community hospitals) from January 1, 1996 through December 31, 2014 with a primary diagnosis of pneumonia. Relevant outcomes included: documented HIV testing in the health system, HIV testing during the pneumonia admission, and new HIV diagnosis.
Results: During the study period 6,858 persons were admitted to the health system with a primary diagnosis of pneumonia. Median age was 50 years (interquartile range [IQR] 40-58) and patients were primarily Caucasian (53%) or African American (41%). Of the cohort, 345 (5.0%) patients were known to have HIV prior to admission and 19 patients (0.3%) had a new positive HIV test result during the pneumonia admission. Furthermore, another 46 patients (0.7%) who were not tested during admission were subsequently diagnosed with HIV a median of 807 (IQR 244-1557) days after pneumonia admission. Of a simple random sample of 207 patients without a prior diagnosis of HIV, 34 (16%, 95% confidence interval [CI] 11-22%) had HIV testing during the pneumonia admission and 83 (40%, 95% CI 33-47%) had an HIV test result ever documented in the health system. HIV testing rates (ever or during pneumonia admission) did not change significantly after 2006 (p > 0.05).
Conclusion: Despite recommendations for universal HIV screening in all healthcare settings since 2006, inpatient HIV testing rates remain low in a population at high risk for HIV infection. Furthermore, a number of persons were subsequently diagnosed with HIV after a prolonged delay, representing a potential missed opportunity for earlier HIV diagnosis. Opt-out HIV testing among pneumonia patients should be implemented to improve testing rates in this high-risk population.

D. Clifton,
None
T. Holland, None
G. Cox, None
K. V. Dicks, None
J. Stout, None