Health education improves referral compliance of persons with probable Diabetic Retinopathy: A randomized controlled trial.

OBJECTIVE: Lack of awareness about Diabetic Retinopathy (DR) is the most commonly cited reason why many persons with type 2 diabetes are non-compliant with referral instruction to undergo retinal screening. The purpose of this study was to evaluate the efficacy of a culturally, geographically and socially appropriate, locally adapted five-month-long health education on referral compliance of participants.
METHOD: A prospective randomized, open-label parallel group study was conducted on persons with type 2 diabetes who underwent basic eye screening at a diabetes hospital between September 2017 and August 2018. Participants who were noncompliant with referral instruction to visit a hospital for advanced DR management were randomly divided into health education intervention group (n = 143) and control group (n = 156). Both groups received information regarding DR and referral instruction at the diabetes hospital. The intervention group was provided personalized education followed by telephonic reminders. The primary endpoint was 'increase in referral compliance' and the secondary endpoint was 'increase in knowledge of DR'. Multivariate logistic regression model was used to identify significant predictors of compliance to referral.
RESULTS: A total of nine participants dropped and 290 completed the post intervention survey. The compliance rate in intervention group was found to be significantly higher than the control group (64.3% vs 28.2%; OR 4.73; 95% CI 2.87-7.79; p<0.001). Participants in the intervention group acquired better knowledge on DR (p<0.05). Apart from intervention, referral compliance rate was also found to be significantly associated with participants' self-perception of vision problem (OR 2.02; 95% CI 1.02-4.01; p = 0.045) and participants' income (OR 1.24; 95% CI 1.06-1.44; p = 0.008). DISCUSSION: Our results suggest that intensive health education on DR should be integrated with diabetes education as it may result in significantly improved referral compliance. Outcomes may be sustainable if interventions are institutionalized at referral point. TRIAL REGISTRATION: Clinical Trials.gov (Registration # NCT03658980); https://clinicaltrials.gov/ct2/show/NCT03658980.

RCT Entities:   Population Interventions Outcomes

Introduction

Type 2 Diabetes Mellitus (T2DM) is on an alarming rise in the world [1]. In 2019, approximately 463 million adults (20–79 years) were living with diabetes and by 2045 this will rise to 700 million [2]. Bangladesh has been identified as one of the top ten countries worldwide in terms of the number of people living with diabetes [3]. In 2019, there was an estimated 8.4 million people with diabetes in the age bracket 20–79 in Bangladesh, and this number is predicted to increase to 15 million by 2045 [4].

Diabetic Retinopathy (DR) is the leading cause of vision loss in adults aged 20–74 years, and the fifth leading cause of global blindness [5,6]. In Bangladesh, the national prevalence of diabetes is estimated to be 8.4% in population aged 20–79 years, with approximately 56% of them undiagnosed [4]. The estimated number of individuals with DR in Bangladesh is 1.8 million (21.6% of people with diabetes) [7]. Higher prevalence of DR has been reported in the coastal areas compared to rural population from other areas of Bangladesh, according to a study that was conducted in Barishal Division [8].

Since DR is initially asymptomatic, many people with diabetes are not aware that their eye condition, if left unmanaged, may affect their vision and lead to blindness [9]. Screening and early intervention is critical—it is more cost effective and may result in better health outcomes [10,11].

In Bangladesh, DR management is not yet integrated as part of mainstream public health systems and as such most registered persons with diabetes do not get their eyes routinely examined [12]. DR management services are largely unavailable in facilities designated for management of cases of diabetes, where other disease management is prioritized given limited resources [12]. This may be because of low demand for DR management services possibly due to lack of awareness about the disease, even among registered persons having diabetes. DR management services can be very expensive [13]; therefore diabetes hospitals may not cater to this service in limited-resource settings where demand for services is low.

A large proportion of persons with diabetes are non-compliant with referral to an Ophthalmologist [14] because of lack of awareness about eye complications of diabetes and lack of information regarding where services are available [15,16]. Published systematic reviews assessing RCTs found that providing health education regarding DR among persons with diabetes is a promising intervention that resulted in increased DR screening rate [17,18].

There does not exist any published RCT on said topic that has been conducted in a least developed country (LDC) or in a low and middle-income country (LMIC), although RCT is considered the gold standard for evaluating effectiveness of health education interventions [19]. To the best of our knowledge, this is the first ever RCT conducted on said topic in an LDC or an LMIC.

Health education interventions must be contextualized according to geographical, cultural and socioeconomic needs. For instance, home-based mail reminders may not be feasible in remote rural locations in Bangladesh given low literacy rates and because effectiveness of mail reminders has been found to be ‘quite modest and short-lived’ in previously published studies [20]. Printed education messages alone have at times failed to increase retinal screening among persons with diabetes [11]. Interventions in limited-resource settings, where literacy rates are low, perhaps require more personalized face-to-face connections and must be trialed to study its effectiveness and scalability.

To address the challenge of suboptimal referral compliance, this study used an innovative approach that comprised of comprehensive and multicomponent modalities such as interactive face-to-face education session using demonstrative flipchart, colorful pictorial leaflet and special referral card, and telephonic follow-ups. The purpose of this study was to evaluate the efficacy of a culturally, geographically and socially appropriate, locally adapted five-month-long health education intervention on participants’ referral compliance. In our research we were able to successfully evaluate the efficacy of the intervention and recommend behavior change strategies as well.

Methods

Trial design

This was a prospective open-label parallel randomized controlled trial designed for non-compliant participants from December 2018 to May 2019. This study was registered with Clinical Trials.gov (Registration # NCT03658980) and approved by Bangladesh Medical Research Council (Registration # 12512062018). Additionally, the study was undertaken in accordance with the Declaration of Helsinki. All participants had been explained in detail about the purpose, risks and potential benefits of the research prior to voluntary consenting and recruitment into the study. Participation was completely voluntary, and a written informed consent was obtained from all participants (S1 and S2 Files).

Study sites

Barishal district under Barishal division of Bangladesh was selected as the study site for this study. This was done to explore the existing referral modality between a private diabetes hospital and a public tertiary hospital. This divisional level public tertiary hospital was the first in Bangladesh (outside of the capital Dhaka) to offer advanced DR management services (including eye screening using High Resolution Fundus camera, provision of treatment such as injections and laser surgeries) to the public at a very minimal cost.

Study population

The study population consisted of participants with T2DM, registered with a private diabetes hospital and referred for advanced DR management to a public tertiary hospital.

Participants eligible for inclusion into study were adults (18 years or above) with T2DM registered with a diabetes hospital, who had undergone preliminary screening for DR using low-resolution fundus camera at the diabetes hospital between September 2017 and August 2018, were referred to a public tertiary level hospital for advanced DR management, did not undergo a Dilated Fundus Examination (DFE) in previous 12 months and had provided informed written consent to be included in this study.

Participants who were excluded from this study were below the age of 18, persons with T2DM registered with a diabetes hospital but not referred to the public tertiary level hospital for advanced DR management between September 2017 and August 2018, persons who had undergone a DFE in previous 12 months and those who did not provide informed written consent to be included in this study.

Statistical basis of sample size

To calculate the sample size for a statistical superiority design suitable for RCT, the formula {[(p1q1 + p2q2)/(p2—p1)2] X factor for alpha and beta} was used, where p1 is the percentage of the existing referral compliance, q1 is (1 –p1), p2 is the percentage of the expected referral compliance from intervention, q2 is (1 –p2), alpha is type 1 error and beta is type 2 error, and factor for alpha and beta with 90% power is 10.5.

It was assumed that the health education intervention would result in referral improvement from 35% to 55% (i.e. 20 percentage point improvement. The baseline referral rate of 35% (within control group) was assumed based on a previous published RCT conducted on persons with T2DM who were provided with personalized follow-up with the aim to increase screening for DR [21]. In this study, which involved mailing a brochure and videotape followed by telephonic reminder a week later, the referral was 31% higher in the intervention group.

During sample size calculation, a more conservative 20% increase in referral rate improvement was assumed keeping in mind the health education intervention modality (home-based health education session followed by telephonic reminders) and challenging sociodemographic and geographical context (lower education level compared to participants of aforementioned study [21], and typical barriers of accessing health care in Barishal which include lack of awareness about disease, challenging road communication, distance to facility, cost, lack of accompanying person, household commitments of women particularly living in rural areas, among others).

To detect a difference of 20% in the referral compliance using 2-tailed test with 90% power and type 1 error 0.05, the sample would be 125 in each group. The dropout rate was assumed to be 20% (more than standard 10%) because participants had already gone back home months ago following referral, and it was thought to be difficult to convince them to make the journey (possibly with an accompanying person) to the tertiary care hospital given the access barriers in Barishal and costs involved in undertaking the journey with multiple transport modalities. The total sample size was calculated to be 300 for the study.

Randomization

To maintain research quality and reduce investigators’ bias, randomization was conducted by a third person (an experienced researcher who was not related to this study). This person generated the random allocation sequence and distributed these in serially numbered sealed opaque envelopes. Envelopes were provided with either of two interventions, ‘Standard Care’ or ‘Health Education Intervention’. The sealed envelopes were directly couriered to the tertiary hospital and were maintained in a locked cabinet under the supervision of an ophthalmic personnel. Twenty sealed envelopes were provided to field research team at a time, which was carried to participants’ home and opened sequentially once a field enumerator completed the baseline in-depth interview. Envelopes were opened in a sequential manner (serial number in participant list and serial number mentioned on top of envelope was the same in all cases) in front of study participant and a witness (usually family member of the participant), and intervention allocation was implemented accordingly.

Outcomes

Primary endpoint

The primary endpoint was ‘increase in referral compliance’, i.e. increase in referral completion. Referral compliance was defined as participants’ visit to the tertiary hospital during a six-month period (five months intervention period and one month additional window period).

Secondary endpoint

The secondary endpoint was ‘increase in knowledge about DR’. Increase in knowledge about DR was measured using pre and post intervention questionnaires (S3 and S4 Files) based on a previously published study [22]. A total of nine questions were investigated to assess participants' knowledge on DR. Knowledge questionnaire was administered by experienced Community Health Workers (CHWs) at participants’ homes. Study participants were asked the same questions during post-intervention survey.

Health education intervention package for the intervention group

The duration of the total health education package was 5 months and included one face-to-face session and telephonic reminders. In case of participant being randomized to intervention group, health education session was delivered by the CHWs who completed tertiary level education and had more than 5 years’ experience in disseminating eye health awareness in communities. The CHWs received orientation to deliver the intervention from the Principle Investigator (first author) who is also a doctoral candidate of health sciences major. Interventions were delivered under the supervision of Principle Investigator. The health education contents are outlined in S5 File. The contents were examined and approved by Ophthalmologists prior to field study. Telephonic reminders were delivered by the CHW who delivered the face-to-face intervention, so that participants could recognize and trust the caller as a result of rapport built during past interaction.

The face-to-face health education session was delivered in local (Bengali) language in about 30–40 minutes following an in-depth interview. It consisted of key information about Diabetes, DR, DR management options, and information about available services at the government tertiary hospital. Communication materials used were a pictorial colorful portable flipchart used for demonstration purpose, colorful pictorial leaflet in native language (Bengali) and a waterproof referral card with important information which would help to measure primary endpoint (referral compliance rate) accurately.

Telephone reminders followed this on Day 7, 30 and 90 where each reminder call lasted about 15 minutes. Multiple reminders have previously been reported to be more effective than single reminder in improving DFE [23]. A similar past study had delivered face-to-face health education session followed by monthly reminders [24]. Our study restricted telephonic reminders to three because evidence suggests that after a third patient reminder, there is no incremental improvement in screening rates [23]. Telephonic reminders were discontinued for participants who completed referral compliance prior to the next scheduled call. The contents of the face-to-face education session and telephonic sessions have been outlined in detail (S5 File).

Control group

Participants from the control group also received information regarding DR and referral instruction from the eye health service providers at the diabetes hospital, as same as the intervention group. Participants of control group were not provided with any form of personalized health education (face-to-face home-based education session followed by telephonic reminders).

Statistical methods

Analysis was conducted by Intention-to-treat (ITT) method. We considered dropout participants to be those who did not receive our phone call or respond during the follow-up calls. All participants including those who dropped-out were included in analysis.

Pearsons chi-square test was used to compare the referral compliance and changes in knowledge measures of the two groups. Univariate analysis was conducted to test association of different variables with referral compliance as the primary endpoint. Subsequently, multivariate analysis with backward Likelihood Ratio (LR) binary logistic regression modeling was performed to identify significant predictors of referral compliance after adjusting for potential confounders. The significance level of association was set at p<0.05 for the univariate analysis and regression analysis. The strength of association was evaluated by calculating odds ratio (OR) and their 95% confidence intervals (CI).

For the secondary outcome, we compared the proportion of participants who provided correct (or positive) responses during pre and post intervention surveys. This comparison was done between the two study arms. In addition, chi-square test was performed for each knowledge related question. There were no participants who initially provided correct (or positive) response at baseline and later incorrect (or negative) response during the post intervention survey. Therefore, we conducted chi-square test only on those participants who initially provided incorrect (or negative) responses. The significance level of association was set at p<0.05 for analysis of the secondary outcome.

Statistical Software for Social Science (SPSS) Version 21.0 was used for data analysis (IBM Corp, 2012) [25].

Reliability and validity

Due to the nature of the study design (open label), it was not possible to withhold knowledge about allocated intervention to participants. Therefore, to reduce bias, a data manager stationed at the tertiary hospital assessed outcome (compliance in both intervention and control groups). In addition, the accuracy of the outcome assessment was checked every 2 weeks by data manager and then by ophthalmic personnel at the tertiary hospital. This record was then shared with Principle Investigator.

Results

Baseline characteristics of the participants

The CONSORT flowchart of this study is detailed in Fig 1. Three hundred and ninety-seven persons who were referred from the diabetes hospital to tertiary hospital and met inclusion criteria were checked for eligibility. Among them, 98 persons dropped out due to reasons shown in Fig 1, and the remaining 299 persons were registered and randomly allocated to the intervention group (health education) (N = 143) or the control group (standard care) (N = 156). During study period, four and five persons dropped out from intervention and control groups respectively as we were unable to follow-up. Therefore, 139 and 151 participants were analyzed, and completion rates were 97.2% and 96.8% in intervention and control groups respectively. The socio demographic and other characteristics of study participants, obtained during baseline survey, are shown in Table 1. There were no statistically significant differences in the baseline characteristics between the two groups.

Fig 1

CONSORT flowchart.

Table 1

Baseline differences between intervention and control group.

Variables	Intervention Group N (%)(N = 143)	Control Group N (%) (N = 156)
Demographic
Mean age in years ± SD	51.5±11.4	51.0±10.8
Women	72 (50.3)	82 (52.6)
Education ≤ Grade 10	102 (71.3)	110 (70.5)
Income ≤ BDT 9999	87 (60.8)	99 (63.5)
Diabetes
Mean years of registration with diabetes hospital ± SD	4.8±5.6	5.4±6.5
Eye Care
Vision problem (self-perceived)	120 (83.9)	126 (80.8)
Duration of self-perceived eye problem < 5 years	105 (73.4)	105 (67.3)
Referral
Clear about referral instruction	135 (94.4)	150 (96.2)

SD, Standard Deviation.

Efficacy of health education

Health education was found to be a statistically significant determinant of compliance rate (Table 2). When compared between two groups, it was observed that 64.3% participants who attended the health education session visited the referred facility, compared to 28.2% participants from control group. The difference of referral compliance rate between these two groups was 36.1% (p<0.001), which was higher than the anticipated 20% (OR 4.73; 95% CI 2.87–7.79).

Table 2

Comparison of primary outcome between intervention and control group.

Variables	Intervention N (%) (N = 143)	Control Group N (%) (N = 156)	P-value
Primary Outcome (Increased referral compliance, i.e. return to tertiary hospital)
Referral compliance	92 (64.3%)	44 (28.2%)	<0.001a
Referral non-compliance	51 (35.7%)	112 (71.8%)

a Pearson chi-square test.

The number of participants in each intervention group who responded positively to knowledge measures (secondary outcome) during pre and post intervention has been reported in Table 3. The post-intervention difference in knowledge levels between both groups was also found to be statistically significant as reported in Table 4. Participants who belonged to the intervention arm and who responded negatively during pre-intervention (n = 127), were significantly more likely to respond positively to the same knowledge measures during post-intervention, when compared to the control group (n = 125).

Table 3

Number of participants who responded positively to knowledge measures (secondary outcome) during pre and post intervention.

Knowledge related measures	Study arm	Intervention Group N (%) N = 139	Control Group N (%) N = 151
Knows that long-term uncontrolled diabetes might be a cause for a vision problem known as DR	Pre-intervention	12 (8.6)	20 (13.2)
	Post-intervention	139 (100)	59 (39.1)
Knows the symptoms of DR	Pre-intervention	12 (8.6)	17 (11.3)
	Post-intervention	139 (100)	28 (18.5)
Knows that the onset of DR can be delayed	Pre-intervention	22 (15.8)	26 (17.2)
	Post-intervention	131 (94.2)	21 (13.9)
Knows that DR can be treated	Pre-intervention	111 (79.9)	136 (90.1)
	Post-intervention	136 (97.8)	136 (90.1)
Correctly indicated one of the symptoms of DR	Pre-intervention	11 (7.9)	14 (9.3)
	Post-intervention	134 (96.4)	50 (33.1)
Correctly indicated one of the treatment option for DR	Pre-intervention	18 (12.9)	33 (21.9)
	Post-intervention	131 (94.2)	35 (23.2)
Understands impact of non-compliance on vision	Pre-intervention	132 (95.0)	128 (84.8)
	Post-intervention	136 (97.8)	129 (85.4)
Knows that a trained Eye Consultant at tertiary hospital provides eye care services	Pre-intervention	30 (21.6)	32 (21.2)
	Post-intervention	139 (100)	68 (45.0)
Knows the days and times when Eye Consultant at tertiary hospital provides eye care services	Pre-intervention	31 (22.3)	30 (19.9)
	Post-intervention	133 (95.7)	30 (19.9)

DR stands for Diabetic Retinopathy.

Table 4

Comparison of number of participants who responded negatively to knowledge measures during pre-intervention.

Knowledge related measures	Responses (pre and post intervention)	Intervention Group N (%) N = 127	Control Group N (%) N = 125	P-valuea
Knows that long-term uncontrolled diabetes might be a cause for a vision problem known as DR	Negative to Positive	127 (100)	39 (29.8)	<0.001
	Negative to Negative	0 (0)	92 (70.2)
Knows the symptoms of DR	Negative to Positive	127 (100)	11 (8.2)	<0.001
	Negative to Negative	0 (0)	123 (91.8)
Knows that the onset of DR can be delayed	Negative to Positive	109 (97.2)	0 (0)	<0.001
	Negative to Negative	8 (6.8)	125 (100)
Knows that DR can be treated	Negative to Positive	25 (89.3)	0 (0)	<0.001
	Negative to Negative	3 (10.9)	15 (100)
Correctly indicated one of the symptoms of DR	Negative to Positive	123 (96.1)	36 (26.3)	<0.001
	Negative to Negative	5 (3.9)	101 (73.7)
Correctly indicated one of the treatment option for DR	Negative to Positive	113 (93.4)	2 (1.7)	<0.001
	Negative to Negative	8 (6.6)	116 (98.3)
Understands impact of non-compliance on vision	Negative to Positive	5 (71.4)	1 (4.3)	0.001
	Negative to Negative	2 (28.6)	22 (95.7)
Knows that a trained Eye Consultant at tertiary hospital provides eye care services	Negative to Positive	109 (100)	36 (30.3)	<0.001
	Negative to Negative	0 (0)	83 (69.7)
Knows the days and times when Eye Consultant at tertiary hospital provides eye care services	Negative to Positive	102 (94.4)	0 (0)	<0.001
	Negative to Negative	6 (5.6)	121 (100)

aPearson chi-square test.

DR stands for Diabetic Retinopathy.

Univariate analyses were conducted to identify sociodemographic and other factors, obtained during baseline survey, which may have had important relationship with the primary endpoint. The dependent variable was referral compliance (yes/no) including all registered participants (n = 299). The independent variables or possible predictors of compliance are shown in Table 5. Results show that in addition to the health education intervention itself, ‘education’ and ‘income’ were found to be statistically significant determining factors of participant’s referral compliance. No other independent variables were found to be significantly associated with the primary endpoint.

Table 5

Association of primary outcome with other variables.

Variables	Compliant N (%) N = 136	Non-compliant N (%) N = 163	P-value
Intervention Group
Health Education Intervention	92 (67.6)	51 (31.3)	<0.001a
Standard Care	44 (32.4)	112 (68.7)
Sex
Women	62 (45.6)	92 (56.4)	0.061a
Men	74 (54.4)	71 (43.6)
Age Group
≤29	2 (1.5)	5 (3.1)	0.476b
30–39	19 (14.0)	19 (11.7)
40–49	40 (29.4)	43 (26.4)
50–59	36 (26.5)	42 (25.8)
60+	39 (28.7)	54 (33.1)
Education
Grade 10 or below	65 (47.8)	88 (54.0)	0.029b
Passed Grade 10 or has higher qualification	68 (50.0)	74 (45.4)
Vocational or others	3 (2.2)	1 (0.6)
Income Range (BDT)
0–4999	71 (52.2)	97 (59.5)	0.044b
5000–9999	7 (5.1)	11 (6.7)
10,000–14,999	12 (8.8)	27 (16.6)
15,000–19,999	15 (11.0)	5 (3.1)
20,000–49,999	26 (19.1)	23 (14.1)
50,000 +	5 (3.7)	0 (0)
Travel time in minutes (home to tertiary hospital)
≤ 60 mins	68 (50.7)	71 (43.6)	0.296b
61–90 mins	23 (16.9)	30 (18.4)
91–120 mins	21 (15.4)	35 (21.5)
≥ 121 mins	23 (16.9)	27 (16.6)
Vision problem (self perception)
Yes	118 (86.8)	128 (78.5)	0.063a
No	18 (13.2)	35 (21.5)
Experiences disability as a result of vision problem
Strongly agree	2 (1.5)	3 (1.8)	0.625b
Agree	11 (8.1)	15 (9.2)
Neutral	2 (1.5)	3 (1.8)
Disagree	121 (89.0)	142 (87.1)
Strongly disagree	0 (0.0)	0 (0.0)
Requires someone else’s assistance to perform day to day activities as a result of vision problem
Strongly agree	0 (0.0)	1 (0.6)	0.282b
Agree	9 (6.6)	14 (8.6)
Neutral	1 (0.7)	3 (1.8)
Disagree	125 (91.9)	144 (88.3)
Strongly disagree	1 (0.7)	1 (0.6)
Service received at referral point (diabetes hospital)
Waiting time			0.207b
<30 minutes	31 (22.8)	46 (28.2)
30–60 minutes	96 (70.6)	110 (67.5)
>60 minutes	9 (6.6)	7 (4.3)
Eye screening duration at diabetes hospital			0.623b
<20 minutes	60 (44.1)	72 (44.2)
20–40 minutes	58 (42.6)	78 (47.9)
>40 minutes	18 (13.2)	13 (8.0)
Counsel time			0.234b
<10 minutes	81 (59.6)	108 (66.3)
10–20 minutes	53 (39.0)	53 (32.5)
>20 minutes	2 (1.5)	2 (1.2)
Referral
Overall referral clarity			0.840a
Yes	130 (95.6)	155 (95.1)
No	6 (4.4)	8 (4.9)

aPearson chi-square test.

bMann-Whitney U test.

All thirteen variables shown in Table 5 were included in the adjusted model during backward Likelihood ratio (LR) binary logistic regression analysis (cut-off value was 0.5). This revealed the most significant predictors of participants’ compliance with referral instruction to the tertiary hospital. Results from the final backward logistic regression model are shown in Table 6. The most important predictor of compliance was health education intervention (OR 4.73; 95% CI 2.87–7.79; p<0.001). Next, compliance rate was positively and significantly affected by participants’ self-perception of vision problem (OR 2.02; 95% CI 1.02–4.01; p = 0.045) and participants’ income (OR 1.24; 95% CI 1.06–1.44; p = 0.008). The Hosmer and Lemeshow test result was non-significant which indicated that the model is a good fit (p = 0.151) and can significantly predict outcome.

Table 6

Multivariate binary logistic regression of the predictors of participants’ referral compliance.

Variable	B	OR	Results at final step 95% CI for OR Lower Upper		P-value	Nagelkerke R2
Intervention type	1.553	4.73	2.89	7.79	<0.001	0.21*
Self-perception (vision problem)	0.703	2.02	1.02	4.01	0.045
Monthly income	0.211	1.24	1.06	1.44	0.008

Hosmer and Lemeshow Test result indicates that model is non-significant (p = 0.151).

Prediction percentage correct for above model = 68.2%.

*The Nagelkerke R2 is 0.21 for this model.

Post-hoc sub-group analyses

Although more women than men presented in the baseline survey of this study (shown in S1 Table), more men were likely to visit referred facility (shown in Table 5). Of the total participants, 10.4% (N = 31) experienced physical disability as a result of their existing vision problems (‘Strongly agree’ and ‘Agree’ categories taken into account as shown in S1 Table). Of the total study participants, 8% (N = 24) said that they depended on someone’s assistance to perform day-to-day tasks as a result of vision problem (‘Strongly agree’ and ‘Agree’ categories taken into account as shown in S2 Table). Out of the 24 participants who depended on someone else’s assistance, only 37.5% (N = 9) visited the referred facility (as shown in S2 Table). However, it was found that among these compliant participants (N = 9), referral compliance was better in participants who received the intervention (N = 6) compared to those who did not (N = 3).

Discussion

Analysis of results

In our study, personalized health education intervention was found to be the most important statistically significant predictor of referral compliance rate. In this study the intervention improved knowledge of DR among participants, consistent with other studies that had reported that health education interventions have been successful in increasing screening rates for retinopathy [26-28]. The face-to-face interaction with participants using easy local language and pictorial tools apparently helped to promote understanding among participants. Participants most probably felt empowered to make informed decision to access DFE after receiving specific information about where, when, how and from whom to seek advanced DR management services at a very small cost. This apparently resulted in reduced perceived barriers and increased perceived benefits among participants. When coupled with cues to action (reminder and follow-up), this resulted in the uptake of health services, and this concept is consistent with the Health Belief Model [29].

The second most important predictor of compliance was participant’ self-perception of vision problem. Where participants felt that their vision was already affected, there seemed to be a sense of urgency and possibly fear that vision may get worse without timely medical intervention. The perceived susceptibility and severity after the health education as per the Health Belief Model [29], have probably led to increased compliance.

Another important predictor of compliance was participant’s income or ability to pay for transportation and logistics required to attend the referred facility (possibly along with an accompanied person). The better financially resourced participants were significantly more likely to visit the referred hospital for further check-up.

Field experience

Health education, for it to be effective, must be focused, personalized and suitably adapted to local context. It must be appropriate from cultural geographical and social perspectives [30]. In our experience, participants were found to be very pleased to participate in an interactive personalized learning session, as they felt ‘cared-for’.

Telephonic reminder system was the most preferred mode of reminder since all study participants owned or had access to a personal cell phone. Participants were more likely to respond eagerly and provide verbal commitment (by confirming a date) to visit the tertiary hospital during the first telephonic reminder on Day 7 compared with reminders provided on Day 30 and 90. Timely intervention is particularly vital for priority patients such as those suffering from a disability. In our study, the health education resulted in increased DFE among those who faced physical disability as a result of vision problem and required someone else’s assistance to perform day-to-day tasks.

In countries with limited resources and high DR prevalence rates, ensuring customized home-based personalized health education even for people with advanced DR maybe costly and challenging. A more cost-effective solution may be to utilize the existing network of CHWs to provide information about DR, and where when and how to access nearby DR management services.

Similarity with other global studies

Our findings are consistent with the findings of similar studies trialed in developed countries [17, 18] where increase in DFE was reported as a result of health education intervention. Our multicomponent and locally adapted intensive health education intervention succeeded to attain a referral compliance rate that was 36.1% higher (in other words, this is equivalent to 128% increase in interventions group compared to control group). Several RCTs based in developed countries that used varied health education modalities to increase DFE reported a lower compliance rate [21, 31–37].

Apart from the fact that a multicomponent intervention was used in our study, there may be other reasons why it was possible to attain higher compliance rate compared to other global studies. It may be easier to achieve a higher change in intervention group where baseline referral compliance rates are relatively low [17] such as in countries with limited health resources. On the other hand, despite lower education level among participants, and social and infrastructural barriers to accessing health care generally expected in limited resource settings, the compliance rate among intervention group was more than what was initially expected during study design.

Future implications

To sustain increased DFEs among persons with T2DM, personalized multicomponent behavior change strategies used in this trial may be tested and provided at the institutions where DR management services are either available within the hospital or in the community. Additional demonstrative strategies such as using audio-visual modalities within hospital premises may be explored and adopted. Further studies may be conducted in LMICs to understand social infrastructural and other systemic barriers to accessing DR management services, so as to devise and trial other effective interventions to increased referral compliance.

Study limitations

This study may not be generalized for the entire population who are suffering from diabetes. Participants of this study were registered with a diabetes hospital and therefore generally aware about diabetes mellitus.

Furthermore, in this study, the referred facility was located in the same district as participant’s residence as well as the diabetes hospital. In areas where participants live further away, or where referred facility is further away, the referral compliance rate may not improve significantly among those who have been provided with health education.

Baseline characteristics of study participants according to subgroups.

(DOCX)

Click here for additional data file.

Effect of intervention on referral compliance for patients who require someone else’s assistance.

(DOCX)

Click here for additional data file.

Informed consent form in English.

(PDF)

Click here for additional data file.

Informed consent form in Bengali.

(PDF)

Click here for additional data file.

Questionnaire in English.

(PDF)

Click here for additional data file.

Questionnaire in Bengali.

(PDF)

Click here for additional data file.

Health education content.

(DOCX)

Click here for additional data file.

CONSORT checklist.

(DOC)

Click here for additional data file.

Trial protocol.

(PDF)

Click here for additional data file.

Approved protocol of ethics committee.

(PDF)

Click here for additional data file.

20 Aug 2020

PONE-D-20-18885

Health education improves referral compliance of persons with probable Diabetic Retinopathy: A randomized controlled trial

PLOS ONE

Dear Dr. Khair,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Oct 04 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Antonio Palazón-Bru, PhD

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. We note that you have stated that you will provide repository information for your data at acceptance. Should your manuscript be accepted for publication, we will hold it until you provide the relevant accession numbers or DOIs necessary to access your data. If you wish to make changes to your Data Availability statement, please describe these changes in your cover letter and we will update your Data Availability statement to reflect the information you provide.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: No

Reviewer #3: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: No

Reviewer #3: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This is an interesting article, with findings that could impact care of patients with diabetic retinopathy in developing countries. Overall, the manuscript is too long, and much of the introduction and discussion can likely be shortened. There are other issues to consider:

1. The introduction is very long. It can likely be cut significantly. There is considerable mention of previous studies that are very similar in scope and purpose, which makes the reader question why this study is necessary. The authors eventually state that this study is unique in that it is conducted in a low/middle income country, but that statement is not made until line 130. It should be made earlier in the text.

2. Line 70--the authors state that there is no treatment to restore vision that has already been lost, but that is not technically true. Anti-VEGF medications can improve vision in the case of DME, though it may be true that those treatments are not readily available in certain communities.

3. Lines 114-118 refer to a "study" that observed a large number of patients who were being referred to for dilated exams were non-compliant, though there is no reference provided. Does that mean you are referring to the current study? If so, then this is not background for the paper (or hypothesis driving). It would, however, be interesting and relevant to know what baseline compliance rates were prior to performing the intervention/conducting the study.

4. It appears that as long as the patients presented for their fundus exam within the study time frame, they were considered compliant, which means that patients who presented up to a year after referral were still considered compliant. There are many scenarios in which that would not really be considered "compliance" given the significant progression of disease that may occur in that time frame, so it would be important to note what the average time between referral to presentation/appointment scheduling was.

5. Table 1 and 2 are included in the methods section, but would be more appropriate in a results section.

6. Re: Table 3; It is interesting that there was no difference between the education and the control group on the understanding of the impact of non-compliance on vision, or the understanding that DR can be treated. Why, then, was compliance improved when there was no difference in the patients' risk perception? This is something that should be further discussed. What part of the education do the authors think was most impactful? Many would assume that a better understanding of potential risk of non-compliance and potential benefit of compliance would encourage patients to be more compliant, however the data demonstrates no difference in understanding between those in the control vs treatment group.

7. The association of eye screening duration with referral compliance is interesting. Is it possible that just spending longer with the patient fostered a sense of urgency in the patient? This should be considered in the manuscript and for further research.

8. The authors should consider a survey of the patients who were compliant to better understand the factors that encouraged them to present.

9. Lines 482-488: The authors discuss cost of scaling a program such as what they have presented. It would be interesting to consider the difference in the cost of scaling a program like that presented in the paper to the cost, for example, of increasing access to care (via programs like eye camps, etc.)

10. Lines 493-505 and line 531-532: It is not really accurate to say that the program presented in the study is "more successful" than previously published programs, as the two are not really comparable. Previous studies, were not only conducted in higher income countries, but also measured rates of DR screening, whereas this study refers to rates of presentation for further care after a positive screening exam. These are two different cohorts of patients, with very different risks, so should not be compared in a head-to-head fashion.

Reviewer #2: The study includes only a modest number of subject compared to other similar studies.

The Health Education is too cumbersome and time-consuming to be useful on a large scale. It involves the 30-40 initial interview. The drive to the patient's house, the flip charts, the brochure, the 15 minute phone calls, the videotape etc. The patients also often had a long drive to the tertiary hospital. Even with all this work, the follow-up rate was still only 67%. The point is, wouldn't it be easier to screen the retina for DR at the initial visit? Have a doctor present t or take a photo with an inexpensive camera for telemedicine. It seems like it would save so much time and effort for both the health care team and patient.

It's not quiet clear how the compliance rate was measured? Was it a visit to the tertiary hospital in the 6 months following the initial visit. It's also not quite clear about how the knowledge test was administered.

Reviewer #3: The manuscript entitled ‘Health education improves referral compliance of persons with probable Diabetic Retinopathy: A randomized controlled trial’ with the aim to evaluate the efficacy of a culturally, geographically and socially appropriate, locally adapted five month-long health education on referral compliance of participants.

The manuscript requires improvement based on the following comments.

Comments

2 decimal points are sufficient for OR and 95%CI figures.

Methods

Trial design

Page 7, ethics committee approval to be stated in the method section.

Sample size calculation

Page 10, 1 or 2 tailed test to be stated.

Statistical methods

Page 13 Line 291, for Independent t tests and chi-squared tests, s to be omitted.

Page 13 Line 294, proper citation including publisher name for SPSS to be stated.

Results

Page 14 Table 1, for p value readings, actual p value to be given.

However, based on the CONSORT guidelines, all statistical tests for group comparison at baseline to be avoided. Symbol <= to be replaced with symbol ≤. n to be provided apart from percentage figures. N(%) to be placed on the first row. Symbol % before the variable name to be removed.

Page 14 Table 2, the figures for non-compliance to be displayed.

Page 14 Table 1 & 2, total N to be provided.

Page 14 & 15, Table 1, 2, 3 to be placed in the results section.

Page 15, what type of chi-square test used here? Chi-square test of independence is not suitable for repeated measures.

Page 17 Line 365 typo Reaults.

Page 18 Table 4, N to be stated on top. The words ‘to 1 Decimal place)’ to be removed. Symbol % for individual figures to be omitted. Likewise with Table S6.

Page 18 Table 4, the presentation to include figures for ‘No’ apart from ‘Yes’. If not ‘n’ for each subcategory for each variable to be provided. Symbol <=, >= to be replaced with ≤, ≥ respectively. Total N to be stated. Symbol % for individual figures to be omitted. The selection criteria for variable(s) selection in the univariate analysis for the inclusion in the adjusted model to be clearly stated. If based on the referral compliance rate improvement, what was the figures chosen or if based on the p value, what was the cut off. The name Pearson chi-square test or chi-squared test to be standardized where appropriate.

Page 19 Line 378-379, the sentence ‘multivariate binary backward logistic regression modeling.’ to be revised. The exact type of backward elimination method to be stated.

Page 19 Line 382-387, p value to be placed after 95%CI.

Page 20, Annex 2?

Page 20 Line 412, Table 4 to be written as Table S7.

Page 20 Table 5, title to be revised. The word ‘Backward’ to be omitted. Exp B to be replaced with OR.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Reviewer #3: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

27 Sep 2020

Date: 25th September, 2020

To

Antonio Palazón-Bru, PhD

Academic Editor

PLOS ONE

Dear Dr. Palazón-Bru,

Greetings!

We thank you and the all the respected Reviewers for thoroughly assessing and providing constructive feedback on our manuscript titled ‘Health education improves referral compliance of persons with probable Diabetic Retinopathy: A randomized controlled trial’ [PONE-D-20-18885].

We have addressed all comments in the following pages and are pleased to resubmit our revised manuscript for your consideration for publication in PLOS ONE. We believe that the paper has improved significantly in quality, relevance and presentation, since all Reviewers’ comments have been addressed.

On behalf of my co-authors, I thank you for your consideration of this resubmission. We appreciate your highly effective engagement and look forward to your positive response.

Sincerely

Zara Khair

(on behalf of co-authors)

Response to Reviewer # 1

This is an interesting article, with findings that could impact care of patients with diabetic retinopathy in developing countries. Overall, the manuscript is too long, and much of the introduction and discussion can likely be shortened.

Thank you for your valuable comments. We have revised the article accordingly, and have now shortened the introduction and discussion sections significantly.

There are other issues to consider:

1. The introduction is very long. It can likely be cut significantly. There is considerable mention of previous studies that are very similar in scope and purpose, which makes the reader question why this study is necessary. The authors eventually state that this study is unique in that it is conducted in a low/middle income country, but that statement is not made until line 130. It should be made earlier in the text.

We agree with Reviewer’s comments and have reduced the Introduction section as well as the Discussion section significantly.

The fact that this study is unique in that it is conducted in a low/middle income country is now mentioned much earlier in the Introduction section (page 5, lines 87 – 91).

2. Line 70--the authors state that there is no treatment to restore vision that has already been lost, but that is not technically true. Anti-VEGF medications can improve vision in the case of DME, though it may be true that those treatments are not readily available in certain communities.

We agree with Reviewer’s comment and have omitted the sentence that stated that ‘there is no treatment to restore vision that has already been lost’.

3. Lines 114-118 refer to a "study" that observed a large number of patients who were being referred to for dilated exams were non-compliant, though there is no reference provided. Does that mean you are referring to the current study? If so, then this is not background for the paper (or hypothesis driving). It would, however, be interesting and relevant to know what baseline compliance rates were prior to performing the intervention/conducting the study.

That is correct - in our first manuscript submission, in lines 114 – 118, ‘study’ referred to our current study. During first submission, we included this information as we wanted readers to understand that similar high baseline non-compliance rates reported in global studies was also found in our current study. To avoid confusion, we have now removed that sentence.

4. It appears that as long as the patients presented for their fundus exam within the study time frame, they were considered compliant, which means that patients who presented up to a year after referral were still considered compliant. There are many scenarios in which that would not really be considered "compliance" given the significant progression of disease that may occur in that time frame, so it would be important to note what the average time between referral to presentation/appointment scheduling was.

We agree with this suggestion, and have now included specific information regarding ‘time between referral to presentation’ in page 6, line 116 and in page 10, lines 205-208.

5. Table 1 and 2 are included in the methods section, but would be more appropriate in a results section.

We have now included Table 1 and 2 in the Results section (page 14-15)

6. Re: Table 3; It is interesting that there was no difference between the education and the control group on the understanding of the impact of non-compliance on vision, or the understanding that DR can be treated. Why, then, was compliance improved when there was no difference in the patients' risk perception? This is something that should be further discussed. What part of the education do the authors think was most impactful? Many would assume that a better understanding of potential risk of non-compliance and potential benefit of compliance would encourage patients to be more compliant, however the data demonstrates no difference in understanding between those in the control vs treatment group.

We thank Reviewer for this highly relevant and thought-provoking questions and suggestions. We have now provided our further insights and explanations in the Statistical Methods and Discussion section.

Please note that as per suggestion made by Reviewer#3, we have revised the statistical analysis for the secondary outcome (‘increase in knowledge’).

In the Statistical Analysis section (page 13, Lines 276-284), we have now mentioned that chi-square test was performed for each knowledge related question. For the secondary outcome, we compared the proportion of participants who provided correct (or positive) responses during pre and post intervention surveys. This comparison was done between the two study arms. In addition, chi-square test was performed for each knowledge related question. There were no participants who initially provided correct (or positive) response at baseline and later incorrect (or negative) response during the post intervention survey. Therefore, we conducted chi-square test only on those participants who initially provided incorrect (or negative) responses.

Our revised Tables 3 and 4 (pages 16-17) now highlight that there have been significant differences across all of the knowledge measures. We have updated Discussion section accordingly.

Please note that during our first manuscript submission, we did not exclude 9 drop out participants from the Pre-intervention survey result. Now, we have excluded the drop out participants from pre and post intervention survey results, and compared using chi-square test. Therefore, our results have also been updated accordingly in Table 3 and 4 (pages 16-17).

7. The association of eye screening duration with referral compliance is interesting. Is it possible that just spending longer with the patient fostered a sense of urgency in the patient? This should be considered in the manuscript and for further research.

We thank Reviewer for this valuable comment.

Please note that as per suggestion made by Reviewer#3, we have now conducted regression analysis with a cut-off value of 0.5 and included all 13 variables used to conduct univariate analysis shown Table 5 (page 18-19). As a result, the final regression model does not have the ‘eye screening duration’ as one of the three significant predictors. Please refer to Table 6 in page 20. All other predictors (‘heath education intervention’ and ‘income’) reported during first submission of manuscript were still found to be significant.

8. The authors should consider a survey of the patients who were compliant to better understand the factors that encouraged them to present.

We agree with this suggestion. Reviewer will be pleased to know that authors are currently considering a study on complaint participants to better understand the factors that encouraged them to present.

9. Lines 482-488: The authors discuss cost of scaling a program such as what they have presented. It would be interesting to consider the difference in the cost of scaling a program like that presented in the paper to the cost, for example, of increasing access to care (via programs like eye camps, etc.)

We have now indicated that a more cost-effective approach would be to utilize the existing network of Community Health Workers to disseminate information about DR and nearby available service (page 22, lines 446-448).

10. Lines 493-505 and line 531-532: It is not really accurate to say that the program presented in the study is "more successful" than previously published programs, as the two are not really comparable. Previous studies, were not only conducted in higher income countries, but also measured rates of DR screening, whereas this study refers to rates of presentation for further care after a positive screening exam. These are two different cohorts of patients, with very different risks, so should not be compared in a head-to-head fashion.

Authors agree with this review. All references to current study being ‘more successful’ have now been omitted. Statements where direct comparisons were made have been modified (Page 23, lines 451 – 457).

Response to Reviewer # 2

1. The study includes only a modest number of subject compared to other similar studies.

We thank Reviewer for this insight. We would humbly like to state that to conduct a high quality Randomized Controlled Trial, we used the superiority sample calculation formula to ensure a sufficient power of 90% with 95% CI, so that study results are considered significant (page 8, lines 154 – 167).

The sample size of our trial (N=299) is more than some of the other similar published studies as well, e.g. sample size used in similar published RCTs was reported to be 132 by Anderson et al. [21] and 280 by Basch et al [37].

2. The Health Education is too cumbersome and time-consuming to be useful on a large scale. It involves the 30-40 initial interview. The drive to the patient's house, the flip charts, the brochure, the 15 minute phone calls, the videotape etc. The patients also often had a long drive to the tertiary hospital. Even with all this work, the follow-up rate was still only 67%. The point is, wouldn't it be easier to screen the retina for DR at the initial visit? Have a doctor present t or take a photo with an inexpensive camera for telemedicine. It seems like it would save so much time and effort for both the health care team and patient.

We thank the Reviewer for discussing valuable insights on a very relevant area.

All participants in our study had already undergone a preliminary DR screening at diabetes hospital using low-resolution fundus camera (please kindly refer to page 7, line 141).

In our study, we were required to implement a multi-component health education intervention for several reasons. Not only was the education level of community generally low in our research setting, but the health system is also quite resource poor (e.g. lack of ophthalmic health personnel, low referral compliance, etc.). Moreover, the psychological aspect of our study participants had to be taken into account – persons with diabetes often forego their appointments even in developed countries because of their mental health status among other reasons.

We have now replaced with the phrase ‘behavior change strategies used in this trial may be tested and provided at the institutions’ in page 23, line 470-471. We have updated the ‘Future Implications’ section in page 23-24 as well. Institutionalization of health education intervention would mean that those who are detected with probable DR at diabetes hospitals might be right away provided with health education interventions to motivate them to undergo DFE. This would eventually save a lot of time and effort for health care team and persons with diabetes.

3. It's not quiet clear how the compliance rate was measured? Was it a visit to the tertiary hospital in the 6 months following the initial visit. It's also not quite clear about how the knowledge test was administered.

We thank the Reviewer for specifying this information gap in our manuscript. We have now included information about how compliance rate was measured (page 10, lines 205–208). We have also detailed how knowledge test was administered (page 10, lines 214-216).

Response to Reviewer # 3

1. Comments:
2 decimal points are sufficient for OR and 95%CI figures.

We have rounded OR and 95% CI figures to 2 decimal points.

2. Methods:
Trial design
Page 7, ethics committee approval to be stated in the method section.

We have now mentioned about ethic committee approval in the methods section (page 6, line 117-118).

3. Sample size calculation:
Page 10, 1 or 2 tailed test to be stated.

We have now mentioned about 2-tailed test (Page 9, line 179).

4. Statistical methods:
Page 13 Line 291, for Independent t tests and chi-squared tests, s to be omitted.

We have omitted ‘s’ from ‘tests’ (page 12, line 266).

5. Page 13 Line 294, proper citation including publisher name for SPSS to be stated.

We have now mentioned proper citation including publisher name for SPSS (page 13, line 287).

6. Results:
Page 14 Table 1, for p value readings, actual p value to be given.
However, based on the CONSORT guidelines, all statistical tests for group comparison at baseline to be avoided. Symbol <= to be replaced with symbol ≤. n to be provided apart from percentage figures. N(%) to be placed on the first row. Symbol % before the variable name to be removed.

We thank the Reviewer for this important insight and suggestion. We have removed the ‘p value’ column from Table 1 (page 14-15). We have replaced the symbol <= with symbol ≤. We have provided ‘n’ apart from percentage figures, placed N (%) on the first row, and removed the symbol % before the variable names.

7. Page 14 Table 2, the figures for non-compliance to be displayed. Page 14 Table 1 & 2, total N to be provided.

We have now displayed the figures for non-compliance in Table 2 (page 15). We have also provided total N in Table 1 and Table 2 (page 14-15).

8. Page 14 & 15, Table 1, 2, 3 to be placed in the results section.

Tables 1, 2 and 3 are now placed in the Results section.

9. Page 15, what type of chi-square test used here? Chi-square test of independence is not suitable for repeated measures.

We thank Reviewer for this important observation. In the Statistical Analysis section (page 13, line 276-284), we have now mentioned that chi-square test was performed for each knowledge related question (line 278-279).

For the secondary outcome, we compared the proportion of participants who provided correct (or positive) responses during pre and post intervention surveys. This comparison was done between the two study arms. In addition, chi-square test was performed for each knowledge related question. There were no participants who initially provided correct (or positive) response at baseline and later incorrect (or negative) response during the post intervention survey. Therefore, we conducted chi-square test only on those participants who initially provided incorrect (or negative) responses.

Please note that during our first manuscript submission, we did not exclude 9 drop out participants from the Pre-intervention survey result. Now, we have excluded the drop out participants from pre and post intervention survey results, and compared using chi-square test. Therefore, our results have also been updated accordingly (Table 3 and 4).

10. Page 17 Line 365 typo Reaults.

We have corrected this typo (Page 18, Line 356).

11. Page 18 Table 4, N to be stated on top. The words ‘to 1 Decimal place)’ to be removed. Symbol % for individual figures to be omitted. Likewise with Table S6.

In the Table, we have now stated ‘N’ on top, removed the words ‘to 1 Decimal place’, omitted the symbol % for individual figures (previously Table 4, now Table 5 in Page 18). We have done the same for S6 Table and resubmitted S6 Table along with revised manuscript.

12. Page 18 Table 4, the presentation to include figures for ‘No’ apart from ‘Yes’. If not ‘n’ for each subcategory for each variable to be provided. Symbol <=, >= to be replaced with ≤, ≥ respectively. Total N to be stated. Symbol % for individual figures to be omitted. The selection criteria for variable(s) selection in the univariate analysis for the inclusion in the adjusted model to be clearly stated. If based on the referral compliance rate improvement, what was the figures chosen or if based on the p value, what was the cut off. The name Pearson chi-square test or chi- squared test to be standardized where appropriate.

For Table 5 (previously Table 4), there are no figures for ‘No’ as it does not apply for the variables presented (we have now mentioned total N in the Table).

We have replaced symbols <= and >= with ≤ and ≥ respectively. We have stated total N, omitted % symbol for individual figures.

All the thirteen variables shown in Table 5 were included in the adjusted model during logistic regression analysis (cut-off value was 0.5). This has been clearly mentioned now (Page 19, Line 367-369).

We have now revised the data type of few of the thirteen variables from ‘nominal’ to ‘ordinal’, where appropriate, and this had led to slight changes in the result of the regression as shown in Table 6 (page 20).

We have now reworded as ‘chi square test’ throughout the manuscript.

13. Page 19 Line 378-379, the sentence ‘multivariate binary backward logistic regression modeling.’ to be revised. The exact type of backward elimination method to be stated.

We have now revised the phrase as ‘multivariate analysis with backward Likelihood Ratio (LR) binary logistic regression modeling’.

The exact type of backward elimination method was ‘Likelihood Ratio’, which has now been specified. (Page 19, Line 368 and Page 12, Line 269-270).

14. Page 19 Line 382-387, p value to be placed after 95%CI.

All p-values are now placed after 95%CI.

15. Page 20, Annex 2?

We have replaced the word ‘Annex 2’ with ‘S6 Table’. (Page 20, Line 392).

16. Page 20 Line 412, Table 4 to be written as Table S7.

We have replaced the word ‘Table 4’ with ‘S7 Table’ (Page 20, Line 395).

(We have followed PLOS ONE numbering and naming guideline for supplementary tables.)

17. Page 20 Table 5, title to be revised. The word ‘Backward’ to be omitted. Exp B to be replaced with OR.

We have omitted the word ‘Backward’ from the Regression Table title, and replaced ‘Exp B' with ‘OR’ (page 20, now Table 6).

20 Oct 2020

PONE-D-20-18885R1

Health education improves referral compliance of persons with probable Diabetic Retinopathy: A randomized controlled trial

PLOS ONE

Dear Dr. Khair,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Dec 04 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Antonio Palazón-Bru, PhD

Academic Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #2: All comments have been addressed

Reviewer #3: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #2: Yes

Reviewer #3: Partly

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #2: Yes

Reviewer #3: (No Response)

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #2: Yes

Reviewer #3: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #2: Yes

Reviewer #3: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #2: (No Response)

Reviewer #3: Minor comments

Table 6, for 95% CI for exp B, exp B to be replaced with OR.

Line 377, it was stated p=0.330 but Line 383 stated p=0.151

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #2: No

Reviewer #3: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

20 Oct 2020

Response to Reviewer # 3

1. Table 6, for 95% CI for exp B, exp B to be replaced with OR.

We have now replaced the term ‘Exp (B)’ with OR.

2. Line 377, it was stated p=0.330 but Line 383 stated p=0.151

We have fixed this typo in line 377 to read p=0.151.

Therefore both the p-values in Line 377 and Line 383 now read the same.

27 Oct 2020

Health education improves referral compliance of persons with probable Diabetic Retinopathy: A randomized controlled trial

PONE-D-20-18885R2

Dear Dr. Khair,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Antonio Palazón-Bru, PhD

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #3: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #3: (No Response)

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #3: (No Response)

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #3: (No Response)

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #3: (No Response)

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #3: (No Response)

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #3: No

4 Nov 2020

PONE-D-20-18885R2

Health education improves referral compliance of persons with probable Diabetic Retinopathy: A randomized controlled trial

Dear Dr. Khair:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Antonio Palazón-Bru

Academic Editor

PLOS ONE