Sight Through a Visor or Through the Eye: Visual Task Performance, Device-Related Pain and Surgical Risk at 12 Months in 138 Federation Patients Born Blind, Fitted with External VISORs or Ocular Implants from 2372 to 2374

1. Introduction

A VISOR is an external device worn across the eyes. It receives signals over a wider band of the electromagnetic spectrum than the unaided human eye, and it passes them to the brain through connectors at the temples, so that a person born blind can perceive the surroundings as a form of vision (Starfleet Medical Prosthetics Board, 2349). The connectors are implanted, but the visor itself can be taken off and replaced. Federation medicine has used it for congenital blindness for decades.

By 2373 ocular implants had begun to replace the VISOR in some wearers. As the first clinical series describe it, an implant places the sensor within the orbit and connects it to the optic pathway, so that the external device and its temple connectors are no longer needed, and the operation is not easily reversed (Adeyinka-Rowe & Lindholt, 2373). Guidance on choosing between the two has rested on engineering arguments about signal bandwidth (Daystrom Institute Cybernetics Division, 2368) and on that series and others like it, which were small and followed patients briefly.

We compare the two devices in a register that follows patients fitted between 2350 and 2374. We ask whether patients with implants perform better on functional visual tasks at 12 months, whether they report less device-related pain, and how the surgical risk of an implant weighs against any gain. We write in 2376.

2. Methods

The Federation Prosthetic Outcomes Register enrols patients aged 16 or older at fitting who receive a sensory prosthesis for congenital blindness at a Federation centre (Federation Prosthetic Outcomes Register, 2375). It records the fitting date, the device, age at fitting and standard assessments at 12 months. The Register held 312 eligible patients fitted between 2350 and 2374: 231 with a VISOR and 81 with an implant. Implants first appear in 2372. We excluded 52 patients: 34 without a 12-month assessment (23 VISOR, 11 implant) and 18 with a co-occurring neurological condition that prevented standard task testing (12 VISOR, 6 implant). Loss to assessment was 10.0% of VISOR patients and 13.6% of implant patients. We analysed 260 patients, 196 with a VISOR and 64 with an implant.

Functional performance was measured by the Visual Task Score, a 0 to 100 composite of four standard tasks scored by a trained occupational assessor: reading printed text, identifying objects across a room, walking an unfamiliar corridor and recognising faces (Pellegrin-Ashdown, 2358). Device-related pain was defined as headache or pain at the temples or orbits on at least half of the days in the month before assessment, by patient report (Ravensdale & Iwuchukwu, 2366). For implant patients we also recorded a return to theatre for a surgical complication within 12 months. Devices are visible, so assessors could not be blinded to the device.

Implants were available only from 2372, so the primary comparison was restricted to patients fitted from 2372 to 2374, 74 with a VISOR and 64 with an implant. Patients chose their device after counselling and clinical guidance on candidacy (Starfleet Medical Prosthetics Board, 2372), so allocation was not random. Because implant patients tended to be younger, we also compared devices within two age strata, under 30 and 30 or over, and combined the stratum differences by inverse-variance weighting (Wyndham-Bell, 2369). The 122 VISOR patients fitted before 2372 were used to check whether VISOR performance had changed over time. The score is continuous and was compared by a difference in means with a Welch interval. Pain is binary and was compared by a risk difference, a risk ratio and an odds ratio, with log-scale intervals for the ratios and a chi-square test. The proportion returning to theatre is given with a score interval.

3. Results

Among the 138 patients fitted from 2372 to 2374, implant patients were younger: 35 of 64 (54.7%) were under 30, against 30 of 74 (40.5%) of VISOR patients (chi-square 2.8, p = .10).

At 12 months the mean Visual Task Score was 70.4 (SD 13.9) with implants and 61.2 (SD 14.8) with a VISOR. The difference was 9.2 points (95% CI 4.4 to 14.0; Welch test, p < .001). Within age strata the difference was 8.1 points under 30 (95% CI 1.4 to 14.8) and 9.3 points at 30 or over (95% CI 2.3 to 16.3). The age-stratified difference was 8.7 points (95% CI 3.9 to 13.5; p < .001).

Among the 122 VISOR patients fitted before 2372 the mean score was 60.1 (SD 15.6). Patients fitted from 2372 to 2374 scored 1.1 points higher on average (95% CI −3.3 to 5.5; p = .62), so we found no sign that VISOR performance had changed over the period, although a change of up to about 5 points cannot be excluded.

Device-related pain was reported by 31 of 74 VISOR patients (41.9%) and 12 of 64 implant patients (18.8%) (Table 1). The risk difference was 23.1 percentage points (95% CI 8.4 to 37.9), the risk ratio 2.23 (95% CI 1.26 to 3.98) and the odds ratio 3.12 (95% CI 1.43 to 6.81), with chi-square 8.6 and p = .003.

Five of the 64 implant patients (7.8%; 95% CI 3.4 to 17.0) returned to theatre within 12 months for a surgical complication. Return to theatre was recorded only for implant patients, so we cannot compare it with VISOR patients.

4. Discussion

Patients with implants scored higher on functional visual tasks and reported less pain than patients with a VISOR, and the score difference was similar in younger and older patients. The size of the gain, about nine points on a 100-point scale, is the fair summary. We have no external standard for the smallest change a patient would notice, and we do not claim that nine points is large.

Selection probably accounts for part of it. Guidance issued when implants became available directed the operation toward patients judged likely to tolerate surgery and to use the device well (Starfleet Medical Prosthetics Board, 2372). Patients selected in that way would plausibly be expected to do well with either device. The VISOR patients fitted in the same years may not have been selected by that guidance. The age-stratified estimate is close to the crude one, but age is only one of the ways the two groups may differ, and motivation, health and the fitting centre are unmeasured.

The gain has to be set against the cost of the operation. About one implant patient in thirteen returned to theatre in the first year, and the interval is wide, from about 3% to 17%. An implant also cannot easily be undone, whereas a visor can be removed or replaced, although its temple connectors remain. Return to theatre was not recorded for VISOR patients, so the surgical burden of the two devices cannot be compared here. A patient who values freedom from pain and an unobstructed face may reasonably accept the surgical risk, and a patient who values reversibility may not, bearing in mind that the pain difference may itself partly reflect selection. The figures show what there is to weigh and do not indicate which choice is correct.

Engineers have argued that a direct optic connection carries more signal than a temple connector (Daystrom Institute Cybernetics Division, 2368). That argument is compatible with a higher score with an implant, and so is the result, although a difference that may be confounded cannot confirm a mechanism. The argument does not predict the difference in pain, and our data cannot say which component of the VISOR produced it. That question needs a study that records the timing and site of pain.

5. Limitations

Allocation was not random, and the comparison rests on patients who chose or were advised to have each device. Follow-up is 12 months, which is short for a device intended for life, and implants have been fitted for no more than three years. Assessors could not be blinded, and pain is self-reported. The Register gives no inter-rater figure for the Visual Task Score. Patients fitted at the same centre are not independent, and our intervals treat them as if they were. The Register enrols patients aged 16 or older at fitting, so children fitted at younger ages are outside it and the findings do not apply to them. The strata are small, with between 29 and 44 patients in each cell, so the stratum-specific intervals are wide. Loss to assessment was slightly higher among implant patients, and we do not know whether patients who were lost differed from those who stayed. The Register enrols patients at participating centres and may not represent centres that do not report to it. Serious harm beyond a return to theatre, such as loss of residual function, was not recorded, and a longer study that recorded it would change the balance of risks reported here.