35 - Missed and Unrecorded Drug Use Among Infective Endocarditis Cases Is Associated with Underestimated Burden of Disease and Fragmented Care: Evidence from Six States

Background: Studies using national administrative data suggest that hospitalizations for drug use-associated infective endocarditis (DUA-IE) have increased over the last ten years. However, drug use as a contributing factor to IE hospitalizations is often missed or not included in coding documentation, resulting in undercount of DUA-IE. We assessed whether missed drug use during IE hospitalizations was associated with higher levels of fragmented care and underestimation of DUA-IE burden.

Methods: We analyzed data from State Inpatient Databases and State Emergency Department Databases from six states (FL, GA, IA, NY, UT, VT) from 2011-2015. Patients older than 16 with ICD-9/10 codes for admissions with IE were included. IE was categorized as DUA using ICD-9/10 codes for drugs/conditions associated with injection drug use. We labeled IE cases as a “missed” DUA-IE case if they had no diagnosis of drug use during their index hospitalization but received a drug use diagnosis during an ED visit or inpatient stay in the calendar year of their index IE hospitalization. We compared “missed” DUA-IE cases to DUA-IE cases where drug use was identified in the index hospitalization and non-DUE-IE cases with respect to demographics, length of stay (LOS) and total charges. To assess care fragmentation, we stratified IE groups by whether the patient was admitted to 1 or >1 hospital within 90-days of the index IE admission.

Results: There were 52147 non-DUA-IE cases, 6872 DUA-IE cases, and 2676 “missed” DUA-IE cases identified by linking drug use across multiple encounters. Missed cases represented a 39% increase in total DUA-IE cases. Compared to DUA-IE cases identified at index hospitalizations, missed cases were more likely to be older, Black, insured by Medicare, and from rural areas. They also had higher 30-day readmission rate (23.2% vs 14.5%, p< 0.001) and higher charges (p< 0.001), with similar LOS. Fragmented care was most common among patients with missed DUA-IE (33.3%), followed by DUA-IE cases identified during index hospitalization (20.5%) and non-DUA-IE cases (13.7%).Table 1
Table 2


Conclusion: Missed and/or unrecorded drug use and fragmented care are common features of DUA-IE. This results in underestimation of both DUA-IE prevalence and hospital utilization due to DUA-IE.