Understanding decision making factors influencing referrals to vision rehabilitation services at a tertiary medical center
Data files
May 08, 2026 version files 35.68 KB
-
questions_revised_3.13.26.csv
10.24 KB
-
raw_7.13.23_revised_3.13.26.csv
23.19 KB
-
README.md
2.25 KB
Abstract
Functional vision loss at any stage affects quality of life (QOL) even before clinically quantified to be moderate or severe. Eye care providers (ECP) are the gatekeepers to vision rehabilitation services (VRS) and often underestimate how small variations in functional vision loss can have a significant impact in a patient’s life. Greater patient-provider communication about vision loss is an opportunity to build trust and increase access to VRS to improve QOL. Tertiary eye care centers serve the greatest number of visually impaired (VI) patients. The purpose of this study was to find what factors influenced ECPs with VRS referral decision-making at a tertiary medical center. We also further studied whether decision making was influenced by the ECP’s specialty, volume of practice, years of practice, number of practicing locations, gender, and continuing education (CE) status. Our survey found that about 77% of respondents, early (<10 years) in their career, and 75% posterior segment (PS) specialists agreed that people with any functional vision complaint should be referred to VRS regardless of visual acuity status. Fifty-two percent of all respondents indicated the reason for low referrals to VRS was to a lack of inquiry by patients about vision services. Inquiry about driving status by patients was an important factor for referral to VRS for 35% of respondents, and especially for PS specialists. Forty-five percent of respondents reported lack of VRS providers and long wait times as a barrier to referrals; however, 68% respondents reported never referring to community services. PS specialists attended more CE courses related to low vision/ vision rehabilitation than anterior segment (AS) specialists. By understanding influencers on ECPs' decision-making, we will be able to make more targeted recommendations on workflow changes, continuing education, and quality improvement processes to increase access to vision rehabilitation services.
This repository contains the de-identified survey data used in the associated manuscript.
Two files are provided:
- raw_7.13.23_revised_3.13.26.csv – de-identified survey response data (redacted Q31 and Q11_4_TEXT)
- questions_revised_3.13.26.csv – mapping of coded variable names to the full survey questions presented to participants
Descriptions
raw_7.13.23_revised_3.13.26.csv
- Rows represent individual respondent submissions.
- Columns correspond to survey questions.
- The Progress column indicates the percentage of the survey completed by the respondent (0–100).
- Column names are coded for brevity (e.g., Q5, Q6). These codes correspond to the question identifiers listed in questions_revised_3.13.26.csv.
questions_revised_3.13.26.csv
- Rows correspond to the survey questions in the order they were presented to respondents.
- Column 1: coded question identifiers corresponding to the column names in raw_7.13.23_revised_3.13.26.csv.
- Column 2: the full survey question text shown to participants.
Missing Data (NULL Values)
Cells containing NULL values in raw_7.13.23_revised_3.13.26.csv represent missing responses and may arise for the following reasons:
- Survey non-completion: Respondents who did not complete the full survey (as indicated by the Progress column < 100) will have NULL values for questions not reached.
- Item non-response: Respondents may have skipped individual questions, resulting in NULL values for those specific items.
No imputation or data modification has been performed; all NULL values reflect the original survey responses as collected.
Human subjects data
This study was determined by the University of Wisconsin Minimal Risk Research Institutional Review Board not to constitute human subjects research under institutional definitions. No participant consent was required under this determination. Only de-identified data are included in this Dryad submission, and the authors confirm that these data are appropriate for unrestricted public sharing in the public domain (CC0).
