Effectiveness of peer support to increase uptake of retinal examination for diabetic retinopathy: study protocol for the DURE pragmatic cluster randomized clinical trial in Kirinyaga, Kenya

The global prevalence of diabetes has escalated in recent decades, with important implications on the health system. In 2015 the International Diabetes Federation estimated that there were 415 million people with diabetes aged 20–79 years (global prevalence of 8.8%), and this is predicted to increase to 642 million by 2040 (global prevalence of 10.4%). [1] The incidence and prevalence of diabetes is increasing disproportionately faster in resource-poor regions and 75% of people living with diabetes currently reside in low- and middle-income countries. [1‐3] This dramatic increase in incidence is occurring in both rural and urban areas. [4] The regional prevalence for Africa was 3.8% in 2015, and the number of people with diabetes in this continent is expected to increase by 140% between 2015 and 2040. [1] In Kenya, the STEPwise survey for risk factors of non-communicable diseases in 2015 found a diabetes prevalence of 2% in the population 18–69 years, and 5.4% in the population 45–59 years. [5]

All patients with diabetes are at risk of developing diabetic retinopathy (DR), the most severe and progressive ocular complication of diabetes. One third of patients with Type 2 diabetes have DR while 10% of them have sight-threatening DR, which represents a significant public health concern. [6, 7]. A population-based study in Nakuru county, Kenya found that 35.9% of people living with diabetes (PLWD) have DR [8]. Visual impairment and blindness from DR is preventable mainly through early detection and timely treatment. Since DR is asymptomatic until the advanced stages, regular retinal screening is of paramount importance. DR meets the Wilson and Jungner (1968) criteria for screening, and current clinical guidelines support annual screening [9‐12]. Participation of PLWD in regular retinal screening, has been shown to be clinically effective in preventing blindness and is also cost-effective. [13, 14]

In developed countries, health systems have formal surveillance programmes for detection of DR. Kenya does not have a national population-based DR screening service where PLWD are systematically invited for screening, but opportunistic screening is available in various hospitals. Importantly, participation and re-participation rates in screening for DR are sub-optimal in Kenya and other resource poor settings. [15‐22]. The determinants of the attendance to retinal examination are complex and include both supply and demand factors. [23, 24] For instance, a Tanzanian study found that PLWD also have limited awareness on diabetic retinopathy, particularly on the need for annual eye examination [25]. This is a barrier that appropriate demand side interventions could address. A health system assessment conducted before this study has shown that 87% of PLWD in Kenya have an unmet need for annual retinal screening. [22] One of the gaps associated with this is the lack of strong links between diabetes services and eye care services. There is need for context-specific pragmatic solutions to address this gap.

A systematic review of interventions to increase diabetic retinopathy screening attendance reported that several strategies are effective, including those targeting the patient (e.g. increasing patient awareness), the health care practitioner (e.g. improving adherence to recommendations) or the organization (e.g. improving patient records) [26]. Members of diabetes support groups (DSGs) are a population subgroup that might benefit from additional support to initiate screening, and adhere to re-screening. Targeting screening interventions towards PLWD in DSGs provides a timely opportunity for three reasons. Firstly this is a community resource that is already available in the community setting. Secondly this population is likely to consider health as an important rationale for behaviour change and the health seeking behaviour of members is potentially malleable to change through peer support. Peer support refers to the provision of emotional and informational support from a created social network member who is considered an equal and who has characteristics similar to the target population. [27] Thirdly it is an economical, culturally-sensitive and flexible intervention for improving diabetes care and outcomes. [28, 29]

Self-efficacy is a direct predictor of health behaviour, according to the social cognitive theory and the self-efficacy theory [30] [31]. Self-efficacy is a predictor of uptake of screening for DR among PLWD. [32] Interventions that improve patients’ self-efficacy decrease perceived barriers and improve the likelihood of initiating the desired health behaviour. There are four main sources of self-efficacy [31]: (i) Successful performance accomplishments (e.g. having attended a previous eye examination) (ii) vicarious experience (e.g. learning that peers have successfully attended an eye examination) (iii) verbal persuasion (e.g. encouragement and recommendation to go for an eye examination by a trusted person, such as a peer or a health care worker) and (iv) psychological cues (decreased sense of isolation of PLWD interacting with a peer, or increased awareness of the risk of DR after receiving educational messages on DR). Fig. 1 shows how self-efficacy for taking a retinal examination might improve through peer support in the Uptake of Retinal Examination in Diabetes (DURE) study.

This study is pragmatic in that it tests the effectiveness of this intervention in the real-world situation of the community and the health system in Kirinyaga. There is a strong need to develop interventions that can reach PLWD populations in real world settings to ensure that any effect found is generalizable.

Public health strategies to manage the diabetes and DR in sub-Saharan Africa (SSA) are known to be inadequate or non-existent. [60, 61] Given that we are at the emergence of the epidemic, this is an appropriate time to develop contextual interventions that will enable our health system to cope with this challenge. To our knowledge, this is the first study that has targeted the DSG population in DR research. The use of peer support in DR is a relatively new field and little has yet been published on the topic.

The trial is important for a number of reasons. For the individuals with diabetes, this trial is in line with the growing global focus on patient empowerment. The PLWD will be empowered to demand for retinal examination, thus reducing demand side barriers to uptake of the examination. These actively engaged PLWD will be linked to eye care providers by the peer supporters. The Chronic Care Model, which has been proposed as a suitable model for managing diabetes, emphasises on the need to implement such links between patients and the health system using community resources. [62] [47] As all PLWD are at risk of DR, empowerment to initiate and maintain screening will be beneficial to all.

The intervention will be provided by trained peer supporters, which is a form of task-shifting. Task shifting is commonly applied in both diabetes and eye care services in our setting. It helps to address the severe shortage of human resources for health. The peer supporters will refer patients to the eye clinic, thus linking diabetes patients with eye clinics and strengthening the referral system in the county.

The explicit use of a theoretical construct (self-efficacy theory) to conceptualise the potential determinants that would influence attendance to DR screening is a key strength of the study. It enhances the understanding of the plausibility of the intervention. The intervention package combines both standardization and flexibility, which allows for scalability in diverse settings. A further strength of this study is the inclusion of process evaluation, which will assist in the interpretation of how and why the intervention did, or did not, bring about the predicted effects.

The study has potential limitations. There are only 16 support groups in the county, which limits the possibility of increasing the number of clusters to further enhance statistical power. Delayed recruitment of the required sample size and loss to follow-up during the trial may be a challenge. In mitigation, a 15% loss to follow-up contingency has been built into the sample size calculation. Sample attrition can result from any inaccuracies in the data collection, such as incorrect address and telephone number information. Other diabetes studies have documented that patients were unable or unwilling to participate due to transportation issues and lack of time or interest [63]. However this is not anticipated because: alternative contact information of participants will be documented, only one visit to the eye clinic is required of participants, and the intervention is expected to build participants’ self-efficacy. Attrition bias may occur if whole clusters drop out, however this is not anticipated as the study period is short.

Despite these limitations, DURE study illustrates the tremendous potential of implementing pragmatic cluster RCTs in the diabetes support group setting. Implementing the trial in this at-risk population will be an invaluable learning opportunity. Many of the lessons learned from this experience could be useful to other research projects.