Impact of personal protective equipment on the clarity of vision among trachoma survey graders and trichiasis surgeons in the context of COVID-19

Background

Trachoma is the leading infectious cause of blindness and remains a public health problem in 44 countries, primarily in Africa.1 2 Prior to the pandemic, in 2019, 92 622 trachomatous trichiasis (TT) surgeries were conducted to realign the eyelid to prevent eyelashes from touching the eyeball.3–5 With the onset of the COVID-19 pandemic, mitigation efforts to protect healthcare workers and patients became a priority.6 Population-based surveys are used to determine if trachoma is a public health problem and if implementation of trachoma elimination activities is required to reach WHO-defined targets for elimination.7 8 In 2019, there were at least 624 trachoma surveys undertaken globally with Tropical Data support (a global initiative that supports health ministries to conduct high-quality prevalence surveys).9 10 To complete them, at least 330 trachoma graders were either trained or retrained.

The causative agent of COVID-19, SARS-CoV-2, spreads from infected people via respiratory droplets and aerosols. There is no direct evidence that eye protection equipment (a face shield or the equivalent) prevents transmission of SARS-CoV-2.11 12 However, during eye examination, such as that required during surveys to estimate trachoma prevalence, or surgeries for treatment of TT, it is impossible to maintain physical distancing, and some form of eye protection has been routinely used by ophthalmic personnel since the onset of COVID-19.

To participate in trachoma surveys, graders must undergo a multistage training in which the rationale for and optimal conduct of each component of the survey participant encounter is broken down and practised, including the use of 2.5× binocular loupes for magnification and prevention of carry-over contamination between successive participants.13 The goal of training is to ensure a high degree of accuracy and reproducibility in the diagnosis of trachomatous inflammation—follicular (TF), the primary sign to determine if active trachoma is a public health problem in the district.

Though individual TT operations take a relatively short time, surgery is typically undertaken on an outreach basis, in sessions completed over a 6–8-hour day for 1–5 successive days. This entails extended face-to-face exposure between surgeons and a sequence of patients. To prepare for service, trichiasis surgeon trainees participate in theoretical teaching, simulator-based training3 and supervised practice on real patients.

It is unknown if face shields, a form of personal protective equipment (PPE), when used with 2.5× magnifying loupes, change the visual acuity or dexterity of health workers during trachoma grading or trichiasis surgery. In this study, we aimed to test the use of a face shield during trachoma survey grader training and TT surgery training. The focus of our investigation was which types and combinations provide the clearest view (no distortion to visualising the eye for grading or surgery) and which are the most practical (such as comfort and cleaning). After a preliminary assessment of user requirements, participants either performed simulated eye surgery on Hazardous Environment Adapted for Development of Surgical Training And Readiness Techniques (HEAD START) mannequins or TF grading on photos of everted eyelids with or without TF using two different face shield configurations as well as no face shield (control). HEAD START uses mannequins to provide a safe and controlled environment for surgical training in a hazardous area of the body, such as the eye.3 14

Methods

There were three research phases. Phase 1 was a preliminary assessment of user requirements. In phase 2, we evaluated the accuracy of trachoma survey grading and the quality of TT surgery, while in phase 3, we investigated comfort, wearability, decontamination and reuse of face shields by trichiasis surgeons during their routine work.

In phase 1, a scoping review of four different configurations of face shields with medical masks (online supplemental figure) was conducted with 16 grader trainers and trichiasis surgeon trainers, selected from a convenience sample of trainers known through Tropical Data (service that is primarily used to support health ministries with trachoma prevalence survey design, planning, training, data collection, management and analysis).7 ,10 The four different configurations were based on recommendations from WHO and the Centers for Disease Control and Prevention for healthcare workers to wear eye protection, including goggles or face shields in addition to face masks if in close contact with patients with COVID-19 or when performing aerosol-generating procedures. Most participants assessed the face shields inside their offices in controlled temperatures or outside in the shade. Assessment involved rating seven different parameters: visibility, misting, convenience, comfort, cleaning, size and functionality, and robustness. To narrow down the configurations on which to conduct more extensive testing in phase 2, responses were collected using a three-point Likert scale: good, moderate or poor. For the purposes of analysis, ‘good’ included both good and moderate, while ‘poor’ was stand-alone. Participants were encouraged to test all four configurations; however, it was not a requirement in phase 1 to do so; some participants were unable to procure goggles or unable to modify the face shield by cutting a hole into the shield in order to mount the loupes.

As a result of the exploratory assessment, two face shield configurations (options 1 and 2) were identified as the most desirable PPE candidates for further testing. Option 1 involved the loupes being mounted on a clear face shield, with a rectangular cut-out providing space for the loupe housing to sit in the correct position in front of the eyes, while in option 2, loupes were worn in front and mounted onto the face shield without a cut-out. Both options in phase 2 also required the use of a medical grade mask.

During phase 2, trachoma grader trainees who were involved in routine Tropical Data refresher trainings (Nigeria and Ethiopia) or a stand-alone grading activity (Nauru and the United Republic of Tanzania) were invited to participate in the study. As per standard trachoma grader training practice,13 photos of everted eyelids with or without TF were uploaded onto mobile hand-held devices to simulate the scale and distance that graders use when examining eyes in the field. Participants were asked to grade TF as ‘present’ (defined as the presence of five or more follicles, each at least 0.5 mm in diameter, in the central part of the upper tarsal conjunctiva)15 or ‘absent’ in a series of 150 images (3 sets of 50 images each). Participants conducted this assessment three times, once without a face shield (control), once with option 1 and once with option 2. The order of the photos was rotated, with the sequence of the options used (1, 2, control) being randomised by individuals to minimise potential bias. Graders were assessed against the unanimous reference grading previously conducted by five international trachoma grading experts.13 A kappa score was calculated to assess graders’ diagnoses against the reference. Each grader generated nine kappa scores (three options times three image sets) and for analysis, a mean of the kappa scores for each option was calculated. As Tropical Data uses a kappa score of ≥0.7 for determining if a grader qualifies as a survey grader, our analysis used the same cut-off. Graders were also asked to complete a questionnaire (phase 2 questionnaire is provided in online supplemental material) to share their feedback on the use of face shields during trachoma grading.

To assess the impact of face shields on trichiasis surgery, certified Nigerian trichiasis surgeons were recruited. Participants performed simulated eye surgery using the HEAD START mannequin with option 1, option 2 or no face shield (control). The option order was randomised to avoid potential bias. Surgeon trainers observed surgery and used a 5-point scale (ranging from 1 ‘poor’ to 5 ‘excellent’) to score each surgeon in 10 skill areas. The overall skill score for each option was calculated by summing the product of the number of surgeons who scored each question by the score.

Trichiasis surgeons were also tested for near vision acuity using a single optotype near vision test, binocularly at 40 cm, using the tumbling E near vision chart.16 This was conducted three times, once for each option (1, 2, no face shield), following the same randomised order used for the simulated eye surgery. Findings for options 1 and 2 were compared with the near vision level achieved while not wearing any face shield (control). Near vision was assessed with the N6 line as cut-off (pass/fail) considering 50% or more of the optotypes correctly identified.

Trichiasis surgeons were also asked to complete a questionnaire (phase 2 questionnaire is provided in online supplemental material) to provide feedback on various characteristics associated with using and cleaning the two face shield options.

Phase 2 results indicated that option 1 was the preferred choice. In phase 3, a second group of trichiasis surgeons in Nigeria were asked, after training, to use option 1 while they conducted routine trichiasis surgeries at their respective surgical sites and to provide feedback, via a questionnaire (phase 3 questionnaire is provided in online supplemental material), on comfort, wearability, decontamination and reuse.

The research was carried out from August 2021 to March 2022 (phase 1 from August to September 2021, phase 2 from October 2021 to January 2022 and phase 3 during March 2022).

To check if the number of people included in phase 2 was enough to identify differences among the three configurations, we conducted post hoc power analyses with power (1–β) set at 0.80 and α=0.05, two tailed for cross-over study. The power analyses were performed for a non-inferiority test with a difference threshold set to 0.05 for the mean. The results of the power analyses showed that the estimated sample sizes for our study needed to be 61 individuals. Thus, given the power analysis results, the number of people included in the study was enough to compare the three configurations.

The score obtained by people using the three configurations was compared using Wilcoxon’s rank test. The agreement of scores of each person obtained using the three configurations was evaluated using Kendall’s concordance test. We considered Kendall’s W<0.7 as low agreement, ≥0.7 and <0.8 as good agreement and ≥0.8 as high agreement when comparing grades.17

Results

During the scoping review (phase 1), four participants tested all four options, one tested three options, five tested two options and six tested one option. In total, 35 assessments were undertaken, 26 (74%) of which were of off-the-shelf face shields. While testing different face shield configurations, 26 assessments used medical masks, 7 used N95 masks and 1 used a cloth mask (all participants were given options to choose the face masks). Phase 1 of the research identified two options (option 1 and option 2) for further testing to determine the potential impact on clarity of vision of trachoma graders and trichiasis surgeons. This conclusion was drawn based on the higher percentage of respondents reporting that visibility and comfort were moderate to good for option 1 and option 2 compared with option 3 and option 4 (table 1).

For phase 2, 124 grader trainees and 28 surgeons were enrolled. Among the 124 grader trainees, 95 (77%) achieved a kappa of ≥0.7 without wearing a face shield (table 2). The majority of grader trainees had kappa scores ≥0.7 using both options 1 and 2, with option 1 associated with a slightly better response. Among grader trainees that would be deployed for survey work (kappa ≥0.70 using no face shield), sensitivity analysis of both options revealed good results for both face shield options (table 3). Among graders who achieved a kappa of ≥0.7, 24 of 94 (26%) reported that face shield option 1 negatively impacted their ability to grade TF, compared with 49 of 103 (48%) using option 2.

Comfort was reported as moderate to good by 109 (89%) graders for option 1 and by 92 (74%) graders for option 2. Graders reported multiple reasons for poor comfort while wearing face shields, with option 1 reported to press on their chest whereas for option 2, graders reported a need to touch the face shield frequently. Ease of cleaning was reported by virtually all graders for both options 1 and 2.

All 28 trichiasis surgeons were assessed for near vision, and there was no difference at the N6 line using the two different face shield options. Compared with not using a face shield, 89.3% of surgeons had the same near vision using option 1 and 92.9% of surgeons had the same near vision score using option 2.

Among the 27 trichiasis surgeons who provided feedback on both face shields, 19 (70%) reported that they did not feel safe to undertake surgery without a face shield. Three surgeons reported that option 2 negatively affected their surgical ability, their dexterity handling sutures or ability to see clearly due to fogging.

Among surgeons reporting discomfort with option 1, the most frequently reported reason was due to the face shield pressing on their chest during surgery (n=13), poor ventilation (n=2) and sweating (n=3). Among surgeons reporting discomfort with option 2, the most frequently reported discomfort was due to breathing difficulties (n=4), feeling hot (n=2) and fog obscuring their view (n=2).

Independent assessment of surgeons by surgeon trainers revealed that, without a face shield, surgeons achieved an overall score of 984 out of 1500, lowest for alignment of sutures and spacing of sutures (table 4). Using option 1, the overall score was 969, while for option 2 the overall score was 941.

Twenty-three surgeons were enrolled in phase 3 of the work in Nigeria. As a group, they carried out a total of 119 surgeries, ranging from 1 to 12 per surgeon, all using the option 1 face shield. The most common reported challenges using the face shield included surgery taking longer with than without a face shield and that it hindered communication with the rest of the surgical team (table 5). Seven surgeons reported that they had to pause the procedure at some point while wearing the face shield due to breathing difficulties. Three of the nine who said they could not do so reported that the face shield was too long, and bumped on their chest; others reported fogging and difficulty breathing.

Discussion

Some trachoma elimination activities paused temporarily during the early part of the COVID-19 pandemic, but were quickly resumed in adapted form. These activities including mass drug administration of an antibiotic, azithromycin, and efforts to encourage facial cleanliness and environmental improvement. Because of close contact between graders or surgeons and members of the public, specific safeguards were needed to reduce the risk of SARS-CoV-2 transmission from patient to provider or vice versa. We set out to identify the most appropriate face shield that could be used easily with 2.5× magnifying loupes. Ideally, it would not hinder the quality of the work, would be comfortable to wear and would be acceptable by users and community members. Finally, for logistical reasons, the ideal face shield configuration would be readily available and easily adaptable, robust, easy to clean and at a cost that programmes could afford.

Our findings suggest that neither of the face shields that we tested led to inaccurate grading of TF. Further, agreement was strongest among grader trainees who passed their initial intergrader agreement test by demonstrating a kappa of ≥0.70; those not passing this initial test would not proceed to field work. We suggest that since option 1 was generally preferable to wear compared with option 2, there is a greater likelihood it would be worn in the field for grading trachoma.

The surgeon trainers’ assessments of trichiasis surgeon skills suggested that option 1 was a slightly better choice than option 2. Regardless of whether a face shield was worn or not, spacing and alignment of sutures was less than ideal, indicating a need to strengthen existing surgical training programmes and supervision. Measures related to comfort, wearability, fogging and cleaning were slightly better for option 1, with the notable exception of discomfort associated with the face shield pressing on the chest. There may be value in finding ways to reduce this discomfort. Few surgeons had difficulties wearing it over extended periods of time in the field, and routine use is indicated as a COVID-19 mitigation measure.

There are limitations to our study. We could not test the use of face shields in all environments (climatic conditions, etc) in which trachoma surveys and trichiasis surgery are undertaken. Long-term viability could not be assessed; it is unclear, for example, how long a face shield will remain intact and functional after repeated use. However, we have collected feedback of graders from different geographical conditions (Nauru, Nigeria, Ethiopia, United Republic of Tanzania), to increase the generalisability of our results.

Since the completion of this study, face shields have been adopted for use by graders and trichiasis surgeons.18 The face shields are generally inexpensive (about US$9 each), but this still represents an additional expense which needs to be included in budgets and procurement for surveys and surgery. Protecting healthcare workers and the populations they look after is of utmost concern and the likelihood of future waves of COVID-19, and potentially other infectious diseases, suggests that face shields may be indicated for years to come.