Presbyopia Screening by Community Health Worker in Bangladesh

Study context:

In aligning with the significant unmet need of presbyopia and prevalent eye care scenario of Bangladesh, BRAC- a multinational non-government organization (NGO), had been implementing a project titled 'Reading glass for improved livelihood (RGIL)' to provide eye care service to the people suffering from eye problems with special emphasis given to the near vision loss or presbyopia since 2006. The objective of this project was to provide access to vision screening and affordable eyeglasses to the people who combat diminished quality of life due to blurry up-close vision and to extend referral for people with other eye problems.

BRAC's Health Nutrition and Population Program (HNPP) Division is implementing this project in partnership with Vision Spring, USA. It has been an issue that whether program could use its existing Community Health Workers (CHW), (BRAC designated them as shasthya shebikas, here the investigators refer them as SS) instead of the assigned program organizers (BRAC designated, the investigators refer them as PO) who had been doing the job, for conducting screening and taking more leading role in selling glasses. In BRAC health program usually these SSs are expected to work six days a week, spending two hours each day for the household visits, covering around 250 Households (HHs) each month, and promote behaviour change communication on various health components. SSs are non-salaried staff with lower educational level, usually with eight years of schooling. During the initial stage on their job, they receive 16 days of basic training on basic preventive, promotive and curative health care, backed up by regular monthly refreshers training. POs usually had master level degree and were regular salaried staff of BRAC as field level worker. When hired by BRAC, they receive a six days basic health and management training to be able to work as a health worker in the community. This change of role shifting, if adopted means that program can easily ensure wider coverage within a short period as CHWs are more in numbers as well as have greater access in the community. Moreover presbyopia can easily and effectively be corrected by reading glasses and certified ophthalmologists are not necessarily required for the visual correction; a lay person with low literacy and appropriate training can prescribe reading glasses for the same. It was well-understood that organizing camp has its own limitation in regards to its coverage of whole population on a regular basis. Moreover, in order to sustain the reading glass program along with the camp, a system was warranted where trained workforce could screen and suggest eye glass during their routine community visit. It was therefore, an experimental study to assess the performance of different level of workforce to screen presbyopia cases, which was necessary to bring in any changes in the service delivery model to scale up throughout Bangladesh.

Eye camps were organized in each of the study sub-district with the aim to eliminate the vision problem faced by the presbyopic patients. Potential patients were encouraged to join camps and information on significance of the camp day, time and place of the camp, procedure of patients' registration, brief of diagnostic techniques, benefits of estimating of refractive error and offer to prescribe glasses only for presbyopia cases were largely disseminated throughout the whole community. Refraction were undertaken using E charts by the respective camp provider. Glasses were available to buy for the patients who required spectacles with a cost of around 2 USD on the spot and providers were instructed to counsel on glass use. These glasses were imported from the USA through Vision Spring, USA.

During the time of the intervention, USS were only employed in 20 sub districts where RGIL was being implemented by BRAC HNPP. To ensure USS recruitment in the study the investigators had to randomly select 6 sub-districts from these 20. Two sub districts were randomly assigned to one of the three arms through permuted-block randomization. Community people aged 35 or more, both male or female and having difficulty with near vision were invited for screening. Individuals having diagnosed eye diseases that require regular treatment and was not willing to participate in the study were excluded.

Sample size:

The sample sizes for this three-arm study were estimated using non-inferiority design. It was assumed that true detection rate in the reference arm (refractionist) will be 80% and that in the experimental arms would be identical. The investigators estimated that enrolment of 288 participants in each of the three arms with 2.5% one-sided significance level would yield 85% power to demonstrate a similarity margin of -10%, if detection rates are identical. Assuming a design effect of 1.5 for the cluster design, the investigators aimed to enrol 435 participants in each of the arms. For comparing accuracy rates between arms, required sample size was 470 per arm to detect a 10% difference with 80% power, 5% significance level and 1.5 design effect. the investigators recruited a minimum of 500 per arm.

Data collection tools and techniques:

Both quantitative and qualitative data were collected for 2 weeks simultaneously face-to-face from the respondents during the camp hours. For the quantitative data, two separate closed-ended structured questionnaires were developed. Information regarding patient's demographics and socio-economic status were collected. From the service providers, information on their age, years of working with BRAC, and duration of involvement in BRAC's RGIL project were also collected. Their level of knowledge was assessed and scored on several issues for instance, common foreign body and infection related eye problems, and its immediate measures, vision (either near or distance) related eye issues, presbyopia, it's aetiology and certain eye conditions which would require to refer to registered eye specialist. Furthermore, refractionists were present in each camp run by all three providers in the respective arms. All the patients screened under any of the above three arms were then re-examined by a refractionist to confirm the diagnosis of presbyopia and it was recorded in a defined format. The investigators kept blinded each of the refractionists to the screening outcome i.e. the refractionists were not aware of the screening report made by our service providers prior to examining a patient. Background data were collected from presbyopia positive participants only identified by BRAC workers. Data collection was performed by skilled interviewers (consisting of science graduates having field experience). A five-day intensive training was organized for them consisting of lectures, mock interviews, role play and field practice in an eye camp. Before the actual survey, the teams were deployed in camps and the participants were informed about the purpose of the study and with their informed verbal consent, the interviewers took information on the camp day.

For the qualitative observation, the investigators recruited two anthropologists as research assistant (RA) who had completed their undergraduate and masters on anthropology. who were present during the entire camp hours run by three different service providers in three study upazillas and observed the activities (from patients counselling to prescribing eye glass as well as marketing aspects). A total of eight camps were observed in each provider arm in three upazillas by two RA. There was one RA per camp who was assigned to observe two cases and note the steps of the screening process for presbyopia by the BRAC providers and some key features of the camp as per the camp protocol for example, counselling and marketing. The qualitative data were collected using a semi structured observation checklist developed following the presbyopia screening protocol and pretested prior to data collection. Observers were given four days training including a one day practical observation exercise in an eye screening camp on the relevant themes to be observed. A total of 24 cases were observed. The investigators blinded the service providers from the direct observation by the research assistants and it was possible because there were quantitative data collectors present in the camp. Blinding was done to understand the unbiased performance by the intervention groups. Content analysis was adopted to analyse the data according to predefined themes.