A glaucoma patient arrives for a routine follow-up and spends a good share of the visit sitting in the waiting room. Clinics have long treated that time as lost, and a pilot at the Moorfields Eye Centre at Bedford Hospital asked whether some of it could go to waiting-room contrast sensitivity testing that patients complete themselves.
The answer is narrow and useful. Patients could do the tests, quickly and willingly, but the pilot was built to show that the task is feasible, not that the results improve care. This article walks through what the pilot did, separates feasibility from clinical validity and clinical utility, and sets out how a clinic team can evaluate any self-administered test before adding it to a visit. For the basics, see our explainer on what contrast sensitivity is, and for when the measure belongs in an exam, our guide to contrast sensitivity testing in clinical practice.
What the Bedford Hospital Pilot Did
Rathore and colleagues collected data between May and August 2024 in the virtual glaucoma clinic at Moorfields Eye Centre at Bedford Hospital, a technician-led service that runs three times a week and sees about 87 patients weekly. Research nurses screened patients the day before their visit. Eligible patients had suspected or confirmed glaucoma, were attending a routine follow-up, met a visual acuity criterion of 6/18 in at least one eye, and could give informed consent.
Of 173 patients approached, 155 agreed to take part. Each was randomly assigned to one of two contrast sensitivity tests and completed it in the waiting-room chair they had chosen on arrival. SpotChecks, from Precision Vision, is a single sheet of paper on which patients mark dots of decreasing contrast with a pen, while PopCSF, from Irida Health, is a tablet game in which patients tap grating patterns on screen. Only the better eye was tested, with the other eye patched, and patients were encouraged to wear their usual near-vision glasses.
To judge whether the waiting-room scores meant anything, every participant also completed a Pelli-Robson chart at 1 meter and a seven-question functional vision questionnaire (VF-7). Those reference measures took place in a small annex nearby rather than in the waiting room, because of the viewing distance and the need for privacy. The team also logged illuminance, background noise and the number of other people in the room, and each participant rated the experience on a nine-item acceptability survey.
How Waiting-Room Contrast Sensitivity Testing Performed
Completion, Time and Acceptability
Of the 155 patients who agreed to take part, 151 (97.4%) completed their assigned test before the appointment: 77 of 78 on paper and 74 of 74 on the tablet. Three participants were called in for their appointment before a test could be assigned, and one could not follow the spoken instructions because of limited English. Another 18 of the 173 patients approached declined, with reasons that included difficulty hearing, limited English, mobility concerns and personal preference.
Median completion time, including the time spent explaining the task, was 3.0 minutes for the paper test and 2.5 minutes for the tablet. No statistically significant association was detected between age and paper test time. For the tablet test, older age was associated with longer completion time, equivalent to about one second per year of age, or roughly 17 seconds longer for an 80-year-old than for a 60-year-old. Patients rated both tests as enjoyable (82% paper, 81% tablet) and said they understood the task well (99% and 96%), and 97% and 95% described the waiting-room session as a beneficial use of their time. Paper was rated easier to perform than the tablet (96% versus 84%), and no adverse events occurred.
Environment and Correlation With Reference Tests
The study found no statistically significant association between the measured lighting, noise or crowding conditions and scores for either test. The waiting room was divided into three illuminance zones (under 200 lux, 200 to 400 lux and over 400 lux), background noise averaged about 60 decibels on the side of the room where most patients were tested, and the room counted as crowded above nine people. No statistically significant difference in scores was detected between morning and afternoon sessions. These findings apply to the conditions sampled in this waiting room, rather than establishing that environmental conditions cannot affect testing.
Both tests tracked the reference measures in the expected direction. Correlation with Pelli-Robson was 0.81 for the paper test and 0.62 for the tablet, and both scores were also associated with age, distance acuity and visual field mean deviation. The gap between the two correlations with Pelli-Robson lost statistical significance after correction for multiple comparisons, so this comparison does not establish that either test correlates more strongly with the reference after that correction. Correlation alone also does not establish agreement or interchangeability.
Feasibility Does Not Establish Clinical Utility
Feasibility answers practical questions: will patients agree, can they finish, how long does it take, and does the setting get in the way. Clinical validity asks whether a score measures what it claims to measure in the people being tested, and clinical utility asks whether having the score changes a decision or an outcome for the better. The Bedford pilot addressed the first group of questions and only touched the second. Its authors state that it was not designed to evaluate the clinical utility of contrast sensitivity testing or its effect on glaucoma management.
A strong correlation is easy to over-read. Two measures can move together without agreeing closely enough for one to replace the other, and in a 1986 Lancet paper Bland and Altman argued that correlation coefficients are misleading for that purpose and described a graphical method based on the differences between two measurement techniques (Bland and Altman, 1986). The distinction matters here because the reference, Pelli-Robson, is a letter chart, while neither waiting-room test uses letters.
The validation history matters as well. SpotChecks has been validated in children and healthy adults, but formal validation in a glaucoma clinic population has not yet been done, and neither test was used in the pilot to classify disease or guide management. What the pilot shows is that patients with glaucoma or suspected glaucoma can complete these tests in a waiting room. It does not show that a given score should change how a patient is managed.
Costs, Staffing and Clinic Flow
Clinic flow is the claim to handle most carefully. Participants spent a median of 109 minutes in the clinic against roughly 90 minutes for non-participants (the paper gives 88 minutes in its abstract and 93 in its results text). The authors attribute the gap to consent, the questionnaire, the reference test and the acceptability survey rather than to the waiting-room tests, and a retrospective analysis excluding those activities left medians of 90 and 93 minutes with no significant difference. That is reassuring about added time, but it is a post hoc adjustment in a single clinic, not evidence that waiting-room testing shortens or speeds up visits.
Staff support is central to the cost question. In the pilot, a health care professional distributed each test, explained it and remained available for questions, so self-administration did not mean unattended testing. Supplementary Table S5 models costs over one year, assuming 15,080 patients and assessment of both eyes. This is a hypothetical service model, not the observed 155-person, one-eye pilot or its measured weekly clinic throughput.
| Annual cost model | Digital test | Paper test |
|---|---|---|
| Staff salary and national insurance | £27,514.00 | £27,514.00 |
| Equipment or test supplies | £1,998.00 for two tablets | £31,168.20 after an assumed 30% wholesale discount |
| Total per clinic per year | £29,512.00 | £58,682.20 |
| Modelled cost per patient | £1.96 | £3.89 |
The main text rounds the annual totals to approximately £29,512 for digital testing and £58,682 for paper testing. Staff costs account for most of the digital total, while discounted paper supplies are the larger component of the paper total. Under a hypothetical fully autonomous arrangement without a designated staff member, the authors estimate reductions of 93% for digital testing (£0.13 per patient) and 53% for paper (£2.07 per patient). These reductions were modelled, not demonstrated by an unattended service in the pilot.
Those figures describe this model, in UK pounds, for one pair of tests, and the authors caution that the analysis is not a general cost model for waiting-room interventions. Patients also said that having someone to talk to while they waited was part of what they enjoyed, which is a reason not to treat removing staff as a pure gain.
How to Evaluate a Self-Administered Contrast Sensitivity Test
Before adding any self-administered test to a visit, a clinic team can work through a short set of questions. The two lists below are editorial guidance drawn from the limits of this pilot and from general measurement principles. They are not a validated protocol, a standard of care or an endorsement of either test.
Questions About the Evidence
- Population fit. Was the test studied in patients like yours in age, diagnosis, acuity range and language? The Bedford sample was mostly older adults with glaucoma or suspected glaucoma from one clinic.
- Reference and agreement. What was the score compared against, and did the analysis assess agreement with a method such as a Bland-Altman plot, or only correlation? A reference that uses a different design, as Pelli-Robson does here, makes the comparison harder to read.
- Repeatability and cut-offs. Does the same patient get a similar score on a second attempt, and are there defined cut-off or normative values for your population? The published article does not report repeat testing or cut-off scores for either test.
- Reporting quality. If a study claims diagnostic accuracy, does it report the items in the STARD 2015 checklist, which lists 30 essential items for diagnostic accuracy studies?
Questions About Your Clinic
- Non-completers. Who cannot or will not take the test because of language, hearing, mobility or time, and what is the plan for them? In Bedford, 18 of 173 patients approached declined and four of the 155 who agreed did not finish.
- Setup. What are the lighting, viewing distance, correction and eye-testing rules, and are they written down? The pilot tested one eye, with the other patched.
- Staffing and interruptions. Who explains the test, who checks the result and what happens when a patient is called mid-test? Three Bedford participants were called in before a test could be assigned.
- Use of the result. What will a clinician do with a high or low score, where is it recorded and which patients still get the standard exam-lane test?
- Outcome evidence. Is there any evidence that acting on the score improves decisions or outcomes? Without it, a waiting-room test is best treated as additional data collection, not as a basis for changing management.
Workflow Considerations for a Real Clinic
The pilot also kept its reference testing out of the waiting room. Pelli-Robson was administered at 1 meter in a separate annex, which is a reminder that any clinic wanting a reference measure alongside a screening score needs space and time for it. Clinicians who want a clinician-administered contrast sensitivity measure in the exam lane can browse our contrast sensitivity test collection or read the CSV-1000 case study, which summarizes a clinical study involving the grating-based CSV-1000.
Clinician-Administered Contrast Sensitivity Tests
Tabla (Contraste): Juego de pruebas de bajo contraste con letras Mars Perceptrix (16"/40 cm)-(20"/50 cm)
El juego de pruebas de contraste de letras Mars Perceptrix es diferente a muchas otras pruebas de contraste que simplemente evalúan la agudeza visual me...
$628.00
Gabinete iluminado: Visor de instrumentos de prueba CSV-1000®
Presentamos el visor iluminado CSV-1000® Mejore la precisión de sus pruebas con el visor iluminado CSV-1000®, diseñado para pruebas estandarizadas de se...
$4,650.00
Limitations of the Pilot
The pilot was a proof-of-concept study at one site, and its sample size was not formally calculated. The authors describe the waiting room as relatively quiet and orderly and say a multicenter trial would be needed before generalizing to other ophthalmology clinics, and that generalizing to settings such as general emergency clinics would not be reasonable. Only the better eye was tested, and the analysis did not adjust for age, type of glaucoma, severity or visual acuity.
Staff support also limits what the pilot can say about unattended testing. The tests were self-administered but not unsupervised, the reference measures were collected away from the waiting room, and no outcome beyond completion, time, acceptability and association with other measures was assessed. Those gaps define the next studies: a multicenter design with both eyes tested, adequate power to examine age and disease severity, and outcomes that show whether acting on the score helps patients.
What This Means for Practice
The pilot is best read as a design template and a caution at the same time. It shows that two brief contrast sensitivity tests can be completed by most patients in a busy glaucoma clinic's waiting room, with high acceptance and no adverse events. It does not show that either test improves glaucoma care, saves staff time or should replace testing in the exam lane.
The unresolved questions are whether a score changes a clinical decision, whether that decision benefits the patient, and what staffing and resources are needed to obtain it reliably. For educators, this makes the pilot a useful case study in separating successful test completion from evidence of clinical benefit.
Sources
- Rathore M, Founti P, Rethi B, Wulandari R, Raja A, Scott A, Jayaram H, Crabb DP, Jones PR. Making better use of patients' time: feasibility of performing self-administered vision assessments in waiting rooms. Eye. Published November 2, 2026. DOI: 10.1038/s41433-026-04874-7. PubMed: PMID 42972052. Open access (CC BY 4.0).
- Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;1(8476):307-310. PMID 2868172.
- Bossuyt PM, Reitsma JB, Bruns DE, et al. STARD 2015: an updated list of essential items for reporting diagnostic accuracy studies. BMJ. 2015;351:h5527. DOI: 10.1136/bmj.h5527.
This article is original editorial commentary on the research and does not reproduce the study's protocol. SpotChecks and PopCSF are named as the study's tools only; Good-Lite does not endorse or compare them.
