Utilization of SOGI in the Confirmation of Transgender and Gender Diverse Patient Populations by Inter-observer Agreement in the Kaiser Permanente System Open Access

Moriarty, Hannah (Spring 2021)

Permanent URL: https://etd.library.emory.edu/concern/etds/w3763812m?locale=en


Background: Despite ongoing efforts to systematically capture gender identity information, accurate identification of transgender and gender diverse (TGD) people based on data from electronic health records (EHR) is still a challenge. One way of improving identification of TGD individuals in the EHR is to use a combination of free text keywords (KWD) and diagnostic codes (DX). In recent years, health systems have also begun to capture Sexual Orientation and Gender Identity (SOGI) data. The aim for the current analysis is to assess the utility of SOGI data in identifying TGD people enrolled in two large health systems.

Methods: The Study of Transitions, Outcomes and Gender (STRONG) data in conjunction with newly available SOGI information was used to assess the identification of TGD patients in Kaiser Permanente health systems in Northern and Southern California. For each patient, two reviewers examined free text to determine TGD status. In the event of disagreement between reviewers, a third reviewer served as adjudicator. Extent of agreement between reviewers was assessed by calculating a kappa statistic and the corresponding 95% confidence interval (CI), both overall and across subgoups of study participants. Factors associated with final confirmation of TGD status were analyzed by multiple logistic regression with results expressed as odds ratios (OR) and 95% CI.

Results: Inter-observer agreement for TGD status was higher among persons with SOGI data [kappa=0.29; 95% CI: 0.24, 0.34] compared to patients without SOGI data [kappa=0.18; 95% CI: 0.14, 0.22]. The odds of inter-reviewer disagreement were significantly lower [OR=0.32; 95% CI 0.27, 0.37] amongst persons whose records contained all three data elements (DX, KWD, and SOGI) compared to health plan members with only two of the three data elements available. Patient’s age also predicted final TGD status confirmation, with those in age groups 19-44 years having reduced odds of disagreement compared to patients ≤18 years of age.

Conclusions: Availability of SOGI data in addition to other data elements improved inter-reviewer agreement and increased the likelihood of TGD status confirmation. Complete and accurate SOGI data collection is necessary for informing evidence-based equitable health care delivery among sexual and gender minority populations.

Table of Contents

Distribution Agreement___page 1

Approval Sheet___page 2

Abstract Cover Page___page 3

Abstract___page 4

Thesis Cover Page___page 5


a. Introduction___page 6

b. Methods___page 9

c. Results___page 12

d. Discussion___page 14

e. References___page 17

f. List of Tables___page 19

About this Master's Thesis

Rights statement
  • Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.
Subfield / Discipline
  • English
Research Field
Committee Chair / Thesis Advisor
Last modified

Primary PDF

Supplemental Files