Development of a program to assess and treat distress in glaucoma patients using an automated EHR-derived AI algorithm

PROJECT SUMMARY/ABSTRACT
Glaucoma is a disease that results in irreversible blindness and due to its chronic, progressive nature, imposes
a psychosocial burden on patients. Appropriately, the focus of ophthalmologists is on controlling the disease to
prevent vision loss. Yet, patient’s psychosocial distress during and after therapy has not been routinely
addressed and is another important target of care. Psychosocial distress (i.e., anxiety, depression) negatively
impacts all outcomes in glaucoma and is associated with poor follow-up and medication adherence, worse
vision-related quality-of-life and disease severity, and faster rates of visual field progression. Direct
assessment and treatment of psychosocial distress is likely to improve glaucoma outcomes. While uncommon
in glaucoma clinics, psychosocial distress screening has been occurring with some consistency in other
medical settings (e.g., oncology) for more than a decade, leading to referrals for intervention and
improvements in psychosocial distress and subsequently overall health. Our overarching scientific premise is
that a screening program for psychosocial distress (i.e., anxiety, depression) in glaucoma clinics would
enhance the patient’s adherence to medical recommendations, and quality-of-life, ultimately leading to
improvements in vision-related outcomes (e.g., visual field progression). Patient-reported outcome measures
are the gold standard measures of distress, however are not routinely collected in patients with glaucoma due
to perceived time and cost burdens. To remedy this, the PI proposes an automated pre-screening framework,
motivated by preliminary analyses that demonstrate that distress can be reliably identified using predictive
modeling based on glaucoma clinical risk factors from electronic health records (EHR) data. This predictive
model will be developed in aim 1 using an existing EHR database, the Duke Glaucoma Registry, and will yield
automated risk estimates of distress that can be used to inform clinical decision making, regarding the
administration of a distress survey; therefore, limiting distress assessment to a subset of high-risk patients.
Secondary aims will focus on external validation of the automated technique, and gauging acceptability to
distress screening in a glaucoma clinic (aim 2), and the refinement of a behavioral intervention to improve
coping skills for distress in patients with glaucoma (aim 3). This research will positively impact patient well-
being in glaucoma, serving as an evidence-based assessment of a distress screening program. The proposal
also details a training plan to help the PI transition from a postdoctoral scholar to an independent researcher.
The mentored phase of the award will be supervised by the primary mentor, Dr. Felipe Medeiros, and
multidisciplinary mentoring team including Dr. Tamara Somers (Psychiatry and Behavioral Sciences), Dr.
David Page (Biostatistics & Bioinformatics), and Dr. Kevin Weinfurt (Population Health Sciences). Performing
the proposed research, formal coursework, and mentored career development will provide the PI with highly
sought-after skills and experiences to help ensure a successful transition to independence.

Grant Number: 5R00EY033027-05
NIH Institute/Center: NIH
Principal Investigator: Samuel Berchuck