Association Between Vision Impairment And Trac Safety Outcomes In Low And Middle-Income Countries: Protocol For A Systematic Review

Background: Road trac injuries are a major public health concern. The burden and road trac fatality rate are especially high in low-and middle-income countries and the socioeconomic impact is profound. Although many authors have studied the correlation between vision and trac safety, there is no robust evidence base that could be used in advocacy. This systematic review will test the hypothesis that interventions to improve vision function are associated with good trac safety outcomes. Methods: Cochrane guidance on conducting a systematic review and Preferred Reporting Items for Systematic Review and Meta-Analysis - Protocols (PRISMA-P) were used to inform the preparation of this protocol and the Cochrane guidance and the main PRISMA guidance will inform the conduct and reporting of the review. We will search MEDLINE (Ovid), EMBASE (Ovid), PsycINFO (Ovid), CINAHL (EBSCO host), Web of Science, Cochrane Database of Systematic Reviews (CDSR) and The Cochrane Central Register of Controlled Trials (CENTRAL) in the Cochrane Library. We will include studies of any design which either attempt to access trac safety outcomes of any kind among persons with any vision decit or examine vision among persons who use roads. The primary outcome for this review is any measure of trac safety or surrogate outcomes. Study selection, data extraction and evaluation of risk of bias will be done by two reviewers. A meta-analysis or narrative data synthesis will be conducted, depending on study quality and homogeneity. Discussion: The results of this review will include summary estimates of vision and the effects of interventions to improve vision function, that are associated with trac safety outcomes in LMICs. This systematic review will ll a gap in the evidence base with policy implications that will be useful for a wide audience and may improve vision of the drivers in LMICs, leading to better trac safety outcomes.


Background
Road tra c injuries are a major public health concern, which require concerted efforts for prevention. They are the leading global cause of death among people aged 5 to 29 years, and without sustained action could become the seventh-leading global cause of death for all ages by 2030. (

1)(2) The World
Health Organization reported that road tra c injuries caused 1.25 million deaths world-wide in 2013. (3) The burden and road tra c fatality rate are especially high in low-and middle-income countries (LMICs) with annual fatality rates per 100,000 population of 24.1 in low-income, 18.4 in middle-income and 9.2 in high-income countries. (3) In LMICs, studies have found that between 30% and 86% of trauma hospital admissions are due to road crashes.(4) Furthermore, the continuous expansion of cities and rapid urban migrations, will mean that urban tra c safety problems will be even more prominent, adding to the challenges faced in LMICs. (5) The socioeconomic impact of road tra c injuries is profound. According to a survey by the Transport Research Laboratory, between 40% and 75% of motor vehicle crash victims in LMICs were the principal earners in a family group.(6) Road tra c crashes cost LMICs 1-2% of their gross national product (GNP), more than the total development aid received by these countries. (7) The main domains of visual functions that are necessary for safe driving are visual acuity, static acuity, dynamic acuity, visual elds, depth perception and contrast sensitivity. Evidence suggests that vision problems such as glare and visual eld loss are associated with increased risk of crashes in high-income countries among older drivers.(8)(9) Furthermore, any association between visual acuity and tra c safety cannot be adequately considered without evaluating other aspects of visual functioning.(9) However, the impact of poor vision on the safety of road users in LMICs, where many drivers do not undergo vision testing, is poorly understood.(10) Although it is known that uncorrected vision problems are common among drivers in LMICs, and that reduced visual acuity appears to be associated with crash risk, the di culty in establishing reliable measures of crash outcomes has been a major barrier to inference of cause and effect. (11) The currently available published literature shows that the association between vision and tra c safety is weakly described for LMICs.
Although many authors have studied the correlation between vision and tra c safety, there is no robust evidence base that could be used in advocacy. (12) The scienti c community has yet to gather the necessary evidence to prove to policy makers that modest investments to improve vision would save lives on the road. Consequently, there is a lack of evidence to inform policies with regard to the vision requirements for drivers that might reduce the burden of mortality and morbidity due to road tra c injuries and information on licensure requirements and rates of compliance in LMICs remains scarce.
This systematic review will help to ll this gap in the evidence base. It will test the hypothesis that interventions to improve vision function are associated with good tra c safety outcomes. We will seek to answer the research question: how does vision impairment impact driver and tra c safety in LMIC settings?

Methods
Cochrane guidance on conducting a systematic review (13) and the Preferred Reporting Items for Systematic Reviews and Meta-Analysis -Protocols (PRISMA-P 2015) (14) were used to inform the preparation of this protocol and the Cochrane guidance and main PRISMA statement (15) will inform the conduct and reporting of the review (PRISMA-P checklist -Additional le 1). This systematic review protocol will be registered in the international prospective register of systematic reviews, PROSPERO (Registration number -submitted).

Objective
Our main objective is to review published studies to determine whether any aspect of vision, and particularly interventions to improve vision function, are associated with tra c safety outcomes in LMICs.
Criteria for considering studies for this review Type of studies We will include any study (interventional or observational) which aimed to assess whether vision is associated with tra c safety outcomes in LMICs. Motor vehicle collisions or motor vehicle crash or road tra c injuries are rarely reported in LMICs and reliable data on crashes are rare. Therefore, we aim to include studies that report poor vision among drivers and any measure of tra c safety or surrogate outcomes.

Type of participants
Eligible participants for these studies will be any person driving a vehicle (e.g., motorcycle, car, bus, truck or other commercial vehicle) with special attention on persons whose income derives in whole or in part from driving a vehicle in a LMIC, where this information is available.

Type of outcome measures
The primary outcome of this review is any measure of tra c safety or a surrogate outcome (e.g., hard braking, accelerometer-measured events, etc.) proven or expected to be associated with tra c safety. We will also collect data on the quality of the vision of drivers as a secondary outcome measure.

Search methods
A comprehensive search strategy has been developed in consultation with an information specialist. We will search MEDLINE (Ovid), EMBASE (Ovid), PsycINFO (Ovid), CINAHL (EBSCO host), Web of Science, Cochrane Database of Systematic Reviews (CDSR) and The Cochrane Central Register of Controlled Trials (CENTRAL) in the Cochrane Library. These sources will be searched from the year of inception to April 2020. A list of search terms for LMICs (as recommended by the Cochrane Library) will be applied in each strategy to identify studies from LMICs. (16) The search strategy is shown in Additional le 2. We will also check the reference lists of eligible studies, to identify additional studies for the review and efforts will be made to translate publications not in English. (17) Data collection and analysis

Selection of studies
We will include eligible studies of any design which either attempt to access tra c safety outcomes of any kind among persons with any vision de cit or examine vision among persons who use roads. Two reviewers (PP, VO) will use the eligibility criteria to independently check the titles and abstracts retrieved by the searches to identify potentially eligible articles. Any disagreements will be adjudicated by a third reviewer (NC). The full text of the potentially eligible articles will be retrieved, and their eligibility con rmed. Data extraction will be done for the eligible articles by the co-reviewers. The study authors will be contacted for any relevant data that are not in their article. The extracted data and data collected from authors will be used in data synthesis of the review.

Assessment of methodological quality
The risk of bias and quality of the included studies will be assessed according to the guidelines of the critical appraisal of skills program (CASP) tool and National Institute of Health (USA) quality assessment tool (NIH-QAT) for the relevant study design. Study characteristics will be tabulated according to the domains described in these tools. We will assess risk of bias of each included study using a tool appropriate to the study design.

Data items and extraction
We will extract data from each study into a specially prepared MS Excel spreadsheet. This will include country, year, study design, setting, sample size, participant characteristics, type of vehicle / driver, measure of visual acuity/other vision related domains and reported outcomes. One reviewer will extract the data and a second reviewer will verify it.

Data synthesis
We will rst describe the study characteristics, such as study design, country and setting, type of driver, category of vehicle. In the next stage, we will provide quantitative synthesis of the ndings for the reported outcomes. We will use odds ratio for binary outcomes and mean differences for continuous variables.
Heterogeneity will be assessed across the studies. If meta-analyses are appropriate, we will use a random effect model, with a xed effect model used in a sensitivity analysis. We will also conduct sensitivity analyses to assess the impact of unmeasured confounding and study assumptions, such as experience of the driver, underlying other visual pathologies other than primary cause measured, different driving environments that will affect the primary outcome, time to event data, different exposure de nitions and different outcome de nitions.
Where the necessary data are available, we will conduct subgroup analysis for persons whose income derives in whole or in part from driving a vehicle in a LMIC and other drivers who drives commercial vehicles. We will also conduct subgroup analyses based on the type of vehicle, and demographic characteristics of the driver (sex and age). We also hope to conduct a subgroup analysis of adherence to vision related licensure requirements and rates of compliance.

Results
We will present the result of our search using the PRISMA study ow diagram (Additional le 3). We will report the quality and risk of bias of each included study based on the use of appropriate tools for the study design (as described above). The systematic review will assess the association of vision and tra c safety in LMIC settings and pooled data tables and graphs will be presented for tra c crashes or near crashes or surrogate events per 10, 000 miles or 100,000 miles based on the outcomes reported in each study. Results of any meta-analyses will be presented using forest plots as appropriate. If we identify a high level of heterogeneity among the studies, we will present the results using narrative synthesis methods.

Discussion
The results of this review will provide summary estimates of vision, and of the effects of interventions to improve vision function, that are associated with tra c safety outcomes in LMICs. To our knowledge, this will be the rst review of the impact of vision on tra c safety in LMICs, lling a gap in the evidence base which has many policy implications in these countries. Our comprehensive search strategy will help to ensure the identi cation of as many articles as possible from LMICs. However, this might still yield a relatively small number of studies from LMICs with high level of heterogeneity, minimising the opportunities for meta-analyses. Nevertheless, the results should be useful for a wide audience and should help identify strategies to improve vision of drivers in LMICs, leading to improvements in tra c safety.
Strengths and limitations of this review -This systematic review protocol follows Cochrane guidance and Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) guidelines.
-This systematic review will address a gap in current evidence base and be the rst review to explore the association of vision and tra c safety outcomes in low-and middle-income countries.
-The inclusion of any study design without a time limit, will support the capture of a wide range of studies, but there may be few studies, especially controlled trials from low-and middle-income countries.
-This systematic review may nd heterogeneous studies, minimising the opportunities for meta-analyses.

Declarations Ethics and Dissemination
There will be no requirement for ethical approval for conducting this systematic review, because it will be based on published articles only. We will disseminate our ndings through publication in a high impact, global health, peer-reviewed journal and presentations at international conferences, and will use the evidence in advocacy in relevant countries.