Reading Glass on Livelihoods and Quality of Life in Rural Bangladesh

BRAC (previously known as Bangladesh Rural Advancement Committee) has conducted a study on the Assessment of the Impact of Reading Glasses on Livelihoods and Quality of Life in the Context of Rural Bangladesh in 2017 with financial and technical assistance from VisionSpring. The goal is to assess the impact of reading glasses on economic return and social angles for people in rural areas in Bangladesh.

BRAC and VisionSpring have developed an experimental trial to test the effects of reading glasses. The study will utilize both pre and post-intervention data to identify the causal effects. A randomized control trial is considered to be the most rigorous method of determining whether a causal relationship exists between an intervention and outcome. This is a Randomized Control Trial (RCT) conducted at the household level.

First, investigators selected 59 villages in Bangladesh where the BRAC/VisionSpring distribution of reading glasses has not been present. A census survey was carried out in the 59 villages to identify those suffering from Presbyopia who were eligible for glasses.

Second, participants were recruited through a household Presbyopia screening of individuals over 35 years old and less than 65 years old. Participants with chronic eye diseases that require regular and intensive treatment were excluded from the study.

A structured questionnaire was used at the census survey where demographic information, income, and occupation were asked. Randomization was taken place after the census survey, and a study sample was selected based on the inclusion criteria of the study which are: (1) Over 35 years of age, (2) Presbyopia positive, (3) Have never used reading glass, (4) Selective occupation (tailor, barber, self-employed artisan etc), (5) Willing to provide contact information and give informed consent.

The study proposal defines two rounds of data collection to be conducted; baseline data collection was completed in August 2017 and follow-up data will be conducted in March 2018. A quantitative questionnaire was distributed at the baseline survey in order to understand the livelihoods and quality of life before the intervention began. Skilled interviewers were recruited for this study, and a three-day intensive training was held to reinforce the goals of the study. The training module was developed in Bangla to guide the interviewers in the field. Questions related to socio-demographic characteristics, economic status of the household (e.g. monthly income), the productivity of the respondent and quality of life were collected during baseline, and the same questionnaire will be used in the baseline survey.

A multi-stage simple random sampling technique has been used in this study. The sample size of the study is pre-determined through a power calculation, the inclusion criteria of being a Presbyopic patient. As our primary outcome is income, a non-normally distributed parameter, it is better to compare proportions falling in various quartiles for sample size calculations. With a two-sided significance level of p=0·05 and power of 80%, to detect a 10% between-group difference in the proportion in the highest quartile of income (35% versus 25%), corresponding to an odds ratio of 1.6, 349 participants would be needed in each group. Assuming a 15% rate of attrition based on previous studies in this population, a total sample of 822 persons (411 in each group) would be required.

As the study is randomized at the individual level, there is a possibility to have the spillover effect. To minimize the spillover effect, one participant was selected from each household. There is a risk that members of the control group may see the benefits experienced by others in their community and buy a pair of reading glasses during the study. In an attempt to mitigate this risk, members of the control group were promised free glasses at the end of the study.

Investigators will use a difference-in-difference analysis to estimate the impact of the reading glasses study based on monthly income (primary outcome), and near vision-related quality of life.

To measure the quality of life, investigators will use a standardised instrument to look at respondents' perceptions of their facility with caregiving, household work, ability to (for those who are literate) mobile phone use, personal hygiene, and dependence.

Differences in baseline characteristics by study group will be analysed using the t-test or Wilcoxon rank-sum test for continuous variables with or without normal distribution respectively, the Chi-squared or Fisher's exact test (with cell frequency<5) for binary categorical variables, and ordinal logistic regression for ordinal categorical variables. Normality will be checked using histograms and Q-Q plots.

Data for the primary outcome will be presented as median and inter quartile range due to lack of normality, The change in self-reported monthly income will be categorized by quartiles and the unadjusted and adjusted intervention effects, odds ratio and 95% CI estimated using ordinal logistic regression. Unadjusted and adjusted mean between-group differences will be estimated for the secondary outcomes using linear regression. All variables with P<0.05 in the simple regression analysis will be included in multiple regression models. All statistical analyses will be performed using a commercially available software package (Stata 15.1, StataCorp, College Station TX, USA).