Abstract
Purpose :
In low vision rehabilitation (LVR), information regarding a patient’s inability to perform activities of daily living is documented as their history of present illness. Much of that information is in narrative format, making it difficult to quantitatively review and analyze. The purpose of this study is to develop methodology for the extraction, interpretation, and analysis of a patient’s rehabilitation goals.
Methods :
We developed operational definitions for what it means to address a goal and what it means to do so successfully or unsuccessfully (Table 1). Using these definitions, co-authors (DB, BW, JG) analyzed the electronic health records of 86 consecutive low vision patients without prior LVR at the Wilmer Eye Institute LVR Service since 2015. Study inclusion required at least one follow-up visit with the same clinician. Patients were excluded if they did not have documented low vision goals when pre-screened by DB. We used multilinear regression to estimate the effects of various patient traits (visual acuity, contrast sensitivity, age, race, and sex) on the number of patient goals and on the proportion of goals successfully addressed.
Results :
The average patient age was 61.8 (SD = 20.3), and 60% of the sample was male. Two variables were statistically significant with respect to the number of patient goals: age (p = 1.59e-05) and the interaction term between logMAR best-corrected visual acuity (BCVA) and log contrast sensitivity (CS) (p = 0.0266) — we used the interaction term because the correlation between them was -0.71 (p = 3.594e-14). As age increased, and as the interaction term BCVA/CS increased (i.e., BCVA or CS gets worse), the number of patient goals decreased. No variables were statistically significant for predicting the proportion of goals successfully addressed.
Conclusions :
This study demonstrates the feasibility of determining how effectively clinicians address goals of patients through a retrospective review of patient health records when extracting such data may be difficult or impractical via other means. Increasing age or worsening vision in patients may lead to fewer LVR goals to address. As the current data set is relatively small, non-significant trends may become significant with a larger sample size.
This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.