Presbyopia Screening by Community Health Worker in Bangladesh: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Presbyopia Screening by Community Health Worker in Bangladesh (NCT03001258), a registered phase not stated study. The primary outcome is Screening Accuracy.

2. Clinical Question

The study addresses whether community-based workers can identify presbyopia cases during eye-camp screening. Its stated purpose was to assess BRAC service-providers’ skill in identifying cases and to support more frequent services in communities with limited eye-care presence. This is an implementation question about screening performance in the setting described, rather than a treatment comparison.

The registry summary calls the study a non-inferiority equivalence randomized intervention trial. It also says that workers were assigned to three arms, with two sub-districts per arm, through permuted-block randomization. However, allocation is not stated in the registry record’s detailed trial information, and the arms are not a like-for-like control comparison. The record does not report a non-inferiority or equivalence margin, or the statistical criteria for reaching either conclusion.

The primary endpoint focuses on agreement with refractionists, not on subsequent care, visual outcomes, or the effect of glasses. Those outcomes are not reported in the registry record. The central question is therefore limited: what screening accuracy was observed for each worker group, and how precisely was it estimated? The registry record does not provide the estimates needed to answer that question quantitatively.

3. Trial Design

The registry identifies this as an implementation research study of presbyopia screening by community health workers in Bangladesh. It reports a study start of 2015-02 and a primary completion of 2015-06, with the study status recorded as completed. The summary describes the work as taking place during February-June 2015. It says data were collected by organizing eye camps over 2 weeks, although the study duration was 5 months.

Six randomly selected sub-districts were involved. According to the summary, workers were assigned to three arms, each consisting of two sub-districts, using permuted-block randomization. The summary calls the study a non-inferiority equivalence randomized intervention trial. At the same time, the registry record does not state the allocation details in its trial-design information, and the arms are not a like-for-like control comparison. The record does not explain the randomization sequence, concealment, or the practical unit at which outcomes were analyzed.

The PO arm involved program organizers employed by BRAC as field-level organizers. Eight POs organized eye camps in two sub-districts, with SS assisting in organizing and screening. The USS arm involved 20 upgraded community health workers assigned in two upazillas; they screened patients and sold glasses during 40 eye camps. In the SS arm, 27 SS organized 25 eye camps over eight days in two upazilas. Each worker group received a two day presbyopia screening training. The record describes 105 eye-camps overall and says there was a minimum 500 sample per arm.

The registry lists an enrollment of 2001, while the summary states that 2059 camp-patients were screened. These figures are not reconciled in the registry record, so their relationship should not be assumed. The summary also says refractionists were recruited to provide the gold-standard assessment against which screening reports were matched.

The record does not report a protocol, allocation ratio, sample-size justification, or detailed analysis plan. It also does not describe whether every screened patient received the refractionist assessment or how missing or inconclusive assessments were handled.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: The discrepancy in the sample size (2001 vs 1090) was because we did two separate sample size calculation in this study. Firstly, in the intervention trial for 3 arms, our sample size was 2001. Then, we took sub-sample (1090) from the intervention sample (2001) to capture patients' socio-demographics.

6. Primary Endpoint

The registry lists the primary outcome as Screening Accuracy, measured over The study duration was for 5 months, however, data was collected by organizing the eye camps over 2 weeks period.

The primary endpoint is screening accuracy, expressed as the percentage of cases. The registry defines an accurate result as one in which the screening result agrees with the gold standard assessment by a refractionist. A positive screen counted as accurate when the refractionist also reported presbyopia; a negative screen counted as accurate when the refractionist also reported a negative result.

This endpoint captures agreement across positive and negative classifications. It is useful for assessing overall classification performance in the screening process, but it does not by itself show how well a screen detects people with presbyopia specifically. Sensitivity and specificity address those two aspects separately, and the study summary says both were calculated. Their numerical results are not reported in the registry record.

Accuracy depends on both types of classification being represented in the population screened. As a general statistical point, a single overall accuracy percentage can conceal different performance for positive and negative cases. It also does not describe whether screening changes follow-up, access to services, or health outcomes. Those outcomes are not reported in the registry record.

The endpoint is accompanied in the record by a 95% confidence interval as its stated dispersion measure. A confidence interval communicates uncertainty around an estimate; its width depends on the amount and variability of the underlying information and on the analysis used. The registry record does not report the observed accuracy percentages or confidence interval limits, so the magnitude and precision of screening accuracy cannot be assessed from that record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Screening accuracy compares a screening classification with a reference classification. In this study, the registry identifies the refractionist’s assessment as the gold standard. A screen was counted as accurate when both classifications were positive or both were negative. This creates a binary agreement outcome for each assessed case, which can then be expressed as a percentage across cases.

What does overall accuracy capture?

In general, accuracy combines correct positive classifications and correct negative classifications. It is an intuitive summary of agreement, but it is influenced by how common positive and negative cases are in the screened population. If one classification is much more common than the other, overall accuracy can appear high even when performance for the less common classification is weaker. That is why the study summary’s mention of sensitivity and specificity is relevant: sensitivity concerns detection among reference-positive cases, while specificity concerns correct negative results among reference-negative cases. The registry record does not report the numerical values for any of these measures.

What does the confidence interval add?

A confidence interval gives a range of values compatible with the observed data under the statistical method used. In general, narrower intervals indicate greater precision and wider intervals indicate more uncertainty. The registry specifies a 95% confidence interval for the primary endpoint, but gives neither its limits nor the underlying accuracy estimate. The method used to calculate the interval is not reported in the registry record. Without those values and method details, precision cannot be evaluated.

Why does the reference assessment matter?

Accuracy depends on the quality and consistency of the reference classification. The summary says screening reports were matched with a refractionist’s assessment, but the record does not describe how refractionists were trained, whether they were unaware of the screening result, or how disagreements and incomplete assessments were handled. As a general matter, a reference standard that is applied inconsistently or with knowledge of the screening result can affect the measured agreement. These procedural details are not reported in the registry record.

How should the three arms be understood?

The record describes POs, USS, and SS as different groups of BRAC workers who received a two day screening training and worked in eye-camp settings. The summary states that workers were assigned to arms by permuted-block randomization across sub-districts, with two sub-districts per arm. Yet the detailed allocation information is not stated, and the arms are not a like-for-like control comparison. The registry record does not say whether the statistical analysis accounted for outcomes being grouped within camps or sub-districts. In general, observations within a shared setting may be more alike than observations from different settings; ignoring that dependence can make uncertainty appear smaller than it is.

The registry summary calls the trial non-inferiority equivalence, but the record does not report a margin, the primary pairwise comparison, or a prespecified decision criterion. In general, a non-inferiority analysis asks whether a difference is not worse than a specified margin, while an equivalence analysis asks whether it lies within specified bounds. The intended conclusion depends on the margin and analysis plan, not on the design label alone. No such numerical bounds or analysis details are reported in the registry record.

What can planning calculations tell readers?

Any sample-size quantities associated with the posted material are described as planning illustrations based on an observed effect, not as the trial’s original design or as observed power. The stated calculation assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation, with the effect set to a stated fraction of the observed effect. These assumptions do not establish that the trial was planned that way. The registry record does not report its own sample-size justification or the information needed to assess the planning calculations independently. A planning illustration is not evidence that a study had a particular probability of finding a result.

The record states that sensitivity and specificity were calculated, but lists no analyses in its analysis information and does not provide an analysis plan. It does not describe hypothesis tests, adjustment for multiple comparisons, handling of missing data, or which participants were included in each analysis. As a general principle, these choices affect how estimates and uncertainty should be interpreted. Their absence means the analysis cannot be reconstructed from the registry record.

10. Multiplicity, Missing Data and Analysis Populations

The registry record lists one primary outcome measure, screening accuracy, and no analyses in its analysis information. The study summary also says that sensitivity and specificity were calculated, but it does not provide their results or specify whether they were treated as additional primary outcomes, secondary outcomes, or descriptive measures. The record does not report a multiplicity plan or explain how testing across worker groups and outcome measures was handled.

Multiplicity arises when several statistical comparisons are made. In general, the more comparisons considered, the greater the chance that at least one result may appear unusual by chance if no adjustment or prespecified hierarchy is used. Whether and how this applies to this study cannot be determined because the registry record does not state the planned comparisons, testing procedure, or decision rules. The description of a non-inferiority equivalence trial does not fill in those missing details.

Missing-data handling is not reported in the registry record. The record does not say how many screened patients had a completed refractionist assessment, whether some results were inconclusive, or how incomplete screening-reference matches were treated. It also does not describe whether all eligible camp-patients were included in the accuracy calculation.

The analysis populations are not specified. In general, an analysis population defines which participants contribute to an estimate, and differing inclusion rules can affect results. Here, the reported enrollment of 2001 and the summary’s statement that 2059 camp-patients were screened are not reconciled. Neither figure can be assumed to be the denominator for the primary endpoint. These gaps limit interpretation of both the intended statistical analysis and the population represented by any estimate.

11. What the Result Does, and Does Not, Mean

The study concerns the feasibility of community health workers screening for presbyopia at eye camps, with refractionists serving as the reference assessment. The endpoint definition makes the intended measure clear: agreement between the screening result and the refractionist’s result, expressed as a percentage. The study summary also reports that sensitivity and specificity were calculated.

However, the registry record does not report the primary accuracy percentages, their 95% confidence intervals, or the sensitivity and specificity estimates. It therefore does not show the observed magnitude or precision of screening performance. It also does not provide results that would establish whether one worker group performed differently from another.

The registry’s description of a non-inferiority equivalence randomized intervention trial should not be read as a finding of non-inferiority or equivalence. No margin or decision rule is reported, and the record does not provide the analysis results needed to evaluate such a claim. The summary reports a randomization approach, but detailed allocation information is not stated, and the arms are not a like-for-like control comparison.

Even a reported accuracy estimate would address agreement with the reference assessment, not every question about a screening service. As a general point, accuracy alone does not show whether patients received follow-up, used glasses, or experienced changes in health. Those outcomes are not reported in the registry record. Nor does the record establish how performance would vary outside the eye-camp context described.

The listed zero adverse-event values do not, on their own, describe the completeness of safety monitoring or reporting. The registry record does not provide monitoring procedures. Overall, the record establishes the research aim, the worker groups, and the endpoint definition, but it does not supply the numerical results or methodological detail required for a firm statistical conclusion about screening accuracy.

12. Why This Trial Matters Statistically

This study poses a useful statistical distinction: screening performance is not captured by a single label such as accuracy. The registry defines an overall agreement measure and reports that sensitivity and specificity were calculated. Those measures address different aspects of classification, so reporting each with its uncertainty would be important for understanding how screening results correspond to the refractionist assessment.

The implementation setting also matters to statistical interpretation. The work involved eye camps across six randomly selected sub-districts, and the three worker groups operated in different sub-districts. In general, measurements gathered within shared camps or locations may be correlated. A valid uncertainty estimate may need to reflect the way outcomes are grouped. The registry record does not explain whether this was considered in the analysis.

The study summary uses the language of non-inferiority and equivalence, while the record does not report margins or decision criteria. This illustrates a broader statistical point: design labels do not replace the quantitative details that define the question being tested. The estimate, confidence interval, comparison, and prespecified bounds are needed to assess an equivalence or non-inferiority claim.

Finally, the discrepancy between the enrollment figure and the number of camp-patients reported as screened highlights why analysis denominators should be explicit. The registry record does not reconcile those figures or identify the denominator used for the primary endpoint. Without that information, readers cannot connect the reported study size to the precision of an accuracy estimate, even if the estimate were available.

13. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The main limitation for statistical interpretation is that the registry record does not report the primary outcome results. It gives no accuracy percentages or confidence interval limits, and it provides no numerical sensitivity or specificity results despite stating that these measures were calculated. As a result, the size and precision of screening performance cannot be assessed.

The record also leaves important design and analysis details unclear. The summary describes permuted-block randomization across sub-districts, but the detailed allocation information is not stated. The arms are not a like-for-like control comparison. No non-inferiority or equivalence margins, statistical decision rules, sample-size justification, or detailed analysis plan are reported. The record does not explain whether clustering by camp or sub-district was addressed.

The enrollment figure of 2001 and the summary’s screened-patient figure of 2059 are not reconciled. The endpoint denominator and the number of completed screening-reference matches are not reported. Missing-data handling, analysis populations, and the treatment of inconclusive results are also not described.

Further limitations specific to the reported information include the absence of participant characteristics, details on refractionist procedures, and information on whether reference assessors were aware of screening results. The registry describes a two week period for organizing eye camps within a study duration of 5 months, but it does not provide further details on how the timing or setting affected data collection. These omissions restrict reproducibility and make it difficult to judge how broadly any eventual estimate would apply.

14. Related Tutorials

15. Related Statistical Calculators

16. Sources

• ClinicalTrials.gov: Presbyopia Screening by Community Health Worker in Bangladesh — NCT03001258.
• Registry record last updated 2017-04-21; status last verified 2017-03.
• Results first posted 2017-04-21.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

17. Record Summary

The registry record describes a completed implementation research study of presbyopia screening by BRAC community health workers in Bangladesh. It lists an enrollment of 2001 and a study start of 2015-02, with primary completion in 2015-06. Its summary says 2059 camp-patients were screened at 105 eye-camps in six randomly selected sub-districts.

The three worker groups were PO, USS, and SS. Each received a two day screening training. The primary outcome was screening accuracy, defined as agreement with a refractionist’s assessment and intended to be reported as a percentage with a 95% confidence interval. Sensitivity and specificity were also calculated, according to the summary.

The registry record does not report numerical outcome estimates, confidence interval limits, analysis details, or the reason the enrollment and screened-patient figures differ. It describes the study as non-inferiority equivalence randomized intervention research, but does not report margins or decision criteria. Its numerical results therefore cannot be interpreted from the registry record.