June 2022
Volume 63, Issue 7
Open Access
ARVO Annual Meeting Abstract  |   June 2022
A framework for automating psychosocial distress screening in ophthalmology clinics using an EHR-derived AI algorithm
Author Affiliations & Notes
  • Samuel Berchuck
    Duke University, Durham, North Carolina, United States
  • Alessandro A Jammal
    Duke University, Durham, North Carolina, United States
  • Kevin Weinfurt
    Duke University, Durham, North Carolina, United States
  • David Page
    Duke University, Durham, North Carolina, United States
  • Tamara Somers
    Duke University, Durham, North Carolina, United States
  • Felipe A Medeiros
    Duke University, Durham, North Carolina, United States
  • Footnotes
    Commercial Relationships   Samuel Berchuck None; Alessandro Jammal None; Kevin Weinfurt None; David Page None; Tamara Somers None; Felipe Medeiros Aerie Pharmaceuticals, Allergan, Annexon, Biogen, Carl Zeiss Meditec, Galimedix, IDx, Stealth Biotherapeutics, Reichert, Code C (Consultant/Contractor), Allergan, Carl Zeiss Meditec, Google Inc., Heidelberg Engineering, Novartis, Reichert, Code F (Financial Support), nGoggle Inc., Code P (Patent)
  • Footnotes
    Support  Supported by National Institute of Health/National Eye Institute grant K99EY033027 (SIB). The funding organizations had no role in the design or conduct of this research.
Investigative Ophthalmology & Visual Science June 2022, Vol.63, 2145 – A0173. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Samuel Berchuck, Alessandro A Jammal, Kevin Weinfurt, David Page, Tamara Somers, Felipe A Medeiros; A framework for automating psychosocial distress screening in ophthalmology clinics using an EHR-derived AI algorithm. Invest. Ophthalmol. Vis. Sci. 2022;63(7):2145 – A0173.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose : In patients with ophthalmic disorders, psychosocial risk factors play an important role in morbidity, mortality, and overall disease outcomes. Proper and early screening of psychosocial distress (i.e., anxiety and depression) can result in prompt intervention, and mitigate many of the factors listed above. Since screening is resource intensive, we developed a framework for automating distress screening using an electronic health record (EHR) derived artificial intelligence (AI) algorithm.

Methods : We performed a retrospective longitudinal study using the Duke Ophthalmic Registry (DOR). DOR consists of EHR data of over 100,000 patients seen at the Duke Eye Center (DEC) from 2009 to 2018. Our cohort included encounters for patients with at least two DEC encounters and a year of follow-up. At each encounter, distress was defined using a validated computable phenotype that relied on diagnostic and procedure codes, and medical history. Risk factors included available EHR history, including diagnostic and procedure codes, medical and encounter history, demographics (age, race, sex, ethnicity, marital status, income, education, and alcohol, smoking and illicit drug use), and problem list items. At each encounter, risk factors were used to discriminate patient distress status using the elastic-net classifier. Performance of the model was evaluated using the area under the receiver operating characteristic curve (AUC). Odds ratios (OR) for the top-25 predictors are presented, along with all non-zero demographic variables.

Results : Our cohort consisted of 358,135 encounters from 40,326 patients with an average of 9 encounters per patient over 4 years. The AUC was 0.91, with sensitivity of 0.97, and 0.90 at specificity levels of 0.75, and 0.9, respectively. Top predictors are listed in Figure and contained mostly predictors associated with psychiatric conditions. Also, present are conditions associated with high levels of distress, including suffocation, esophageal disorders, and headaches.

Conclusions : Using EHR data, we automatically identified psychosocial distress in ophthalmology patients upon encounter to a tertiary eye clinic. Our screening framework will be most effective on improving ophthalmic outcomes when paired with an effective referral and treatment program.

This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.

 

Odds ratios (OR) for the top 25 predictors of distress and non-zero demographics.

Odds ratios (OR) for the top 25 predictors of distress and non-zero demographics.

×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×