JMIR Publications

ABSTRACT

Background:

Electronic health records (EHR) provide valuable insights to address clinical and epidemiological research concerning HIV, including the disproportionate impact of the COVID-19 pandemic on this population. To identify people living with HIV (PLWH), most studies using EHR or claims databases start with diagnostic codes, which can result in misclassification without further refinement using drug or laboratory data. Furthermore, given that antiretrovirals now have indications for both HIV and COVID-19 (i.e., ritonavir in nirmatrelvir/ritonavir), new phenotyping methods are needed to better capture PLWH. Therefore, we created a generalizable and innovative method to robustly identify PLWH, pre-exposure prophylaxis (PrEP) users, post-exposure prophylaxis (PEP) users, and people not living with HIV (PNLWH) using granular clinical data after the emergence of COVID-19.

Objective:

The primary aim of this work was to use computational phenotyping in EHR data to identify PLWH (cohort 1), people prescribed PrEP (cohort 2), PEP (cohort 3) or none of the above (PNLWH, cohort 4), and describe COVID-19 related characteristics among these cohorts.

Methods:

We used diagnostic, laboratory measurement, and drug concepts within the National Clinical Cohort Collaborative (N3C) to create a computational phenotype for four cohorts with confidence levels. For robustness, we conducted a randomly sampled, blinded clinician annotation to assess precision. We calculated the distribution of demographics, comorbidities, and COVID-19 variables among our four cohorts.

Results:

We identified 132,664 PLWH with a high level of confidence, 36,088 PrEP users, 4,120 PEP users and 20,639,675 PNLWH. Most PLWH were identified by a combination of conditions, laboratory measurements, and drug exposures (74,809, 56.4%), followed by labs and drugs (15,241, 11.5%), then conditions and drugs (14,595, 11.0%). A higher proportion of PLWH experienced COVID-19-related hospitalization 4,650 (3.51%) or mortality 828 (0.62%), and all-cause mortality 2,083 (1.57%) compared to other cohorts.

Conclusions:

Using an extensive phenotyping algorithm leveraging granular data in an EHR repository, we have identified PLWH, PNLWH, PrEP and PEP users, and offer transferable lessons to optimize future EHR phenotyping for these cohorts.