The correlation coefficient (R2) equals 0.23 for the number of Cochrane systematic reviews and protocols compared with the global percentage of total 2010 disability-adjusted life-years for 8 ophthalmologic conditions.
eTable 1. Conditions That Comprise the Other Vision Loss Category With Their ICD-10 Code Definitions and Terms Entered Into the Search Query
eTable 2. Reviews and Protocols Included for Seven GBD Ophthalmologic Conditions With Their Associated Cochrane Group and Publication Year
eTable 3. Reviews and Protocols Included for GBD Other Vision Loss Category With Their Associated Cochrane Group and Publication Year
eTable 4. Excluded Titles Generated From Search of Seven Ophthalmologic Conditions Studied by GBD 2010 in the Cochrane Database of Systematic Reviews
eTable 5. Excluded Titles Generated From Search of GBD 2010 Other Vision Loss Category in the Cochrane Database of Systematic Reviews
Boyers LN, Karimkhani C, Hilton J, Richheimer W, Dellavalle RP. Global Burden of Eye and Vision Disease as Reflected in the Cochrane Database of Systematic Reviews. JAMA Ophthalmol. 2015;133(1):25-31. doi:10.1001/jamaophthalmol.2014.3527
Eye and vision disease burden should help guide ophthalmologic research prioritization. The Global Burden of Disease (GBD) Study 2010 compiled data from 1990 to 2010 on 291 diseases and injuries, 1160 disease and injury sequelae, and 67 risk factors in 187 countries. The Cochrane Database of Systematic Reviews (CDSR) is a resource for systematic reviews in health care, with peer-reviewed systematic reviews that are published by Cochrane Review Groups.
To determine whether systematic review and protocol topics in the CDSR reflect disease burden, measured by disability-adjusted life-years (DALYs), from the GBD 2010 project. This is one of a series of projects mapping GBD 2010 medical field disease burdens to corresponding systematic reviews in the CDSR.
Design and Setting
Two investigators independently assessed 8 ophthalmologic conditions in the CDSR for systematic review and protocol representation according to subject content. The 8 diseases were matched to their respective DALYs from the GBD 2010 project.
Main Outcomes and Measures
Cochrane Database of Systematic Reviews systematic review and protocol representation and percentage of total 2010 DALYs.
All 8 ophthalmologic conditions were represented by at least 1 systematic review in the CDSR. A total of 91.4% of systematic reviews and protocols focused on these conditions were from the Cochrane Eyes and Vision Group. Comparing the number of reviews and protocols with disability, only cataract was well matched; glaucoma, macular degeneration, and other vision loss were overrepresented. In comparison, trachoma, onchocerciasis, vitamin A deficiency, and refraction and accommodation disorders were underrepresented.
Conclusions and Relevance
These results prompt further investigation into why certain diseases are overrepresented or underrepresented in the CDSR relative to their DALY. With regard to ophthalmologic conditions, this study encourages that certain conditions get more focus to create a better representation of what is causing the most disability and mortality within this research database. These results provide high-quality and transparent data to inform future prioritization decisions.
The process of setting research agendas and priorities is complex and multifactorial, with the aim to provide the greatest public health benefit.1 While a uniform approach to establishing health research priorities is not appropriate for the unique situation and context in which agendas are set, 9 common themes of good practice have been identified including, but not limited to, determination of criteria that will guide prioritization decisions, determination of decision makers, and transparency.1 The Cochrane Collaboration serves as an example of a large network of researchers that has established a diversity of priority-setting methods by its various review groups.2 The 10 guiding principles of Cochrane’s work are collaboration, building on the enthusiasm of individuals, avoiding duplication of effort, minimizing bias, keeping up to date, striving for relevance, promoting access, ensuring quality, continuity, and enabling wide participation.3
Within the Cochrane Library, the Cochrane Database of Systematic Reviews (CDSR) contains reviews produced by 53 Cochrane Review Groups. The CDSR serves as a source of peer-reviewed systematic reviews in health care. It aims to provide evidence-based information to inform decision making for individuals, health care professionals, and policy makers alike.4,5 Until now, Cochrane reviews have mainly focused on treatment effectiveness assessed in randomized clinical trials, although guidance on incorporating evidence from studies with other designs is being developed.
One of the Cochrane Review Groups, the Eyes and Vision Group, registered with Cochrane in 1997 and has 200 members from 30 countries.6 This group seeks to create, sustain, and encourage access to systematic reviews that synthesize information regarding all aspects of prevention and treatment of eye disease and visual impairment.7 Its scope extends to aid those affected by visual impairment or blindness in the adjustment to their disability. Prioritization strategies for the Cochrane Eyes and Vision Group specifically emphasize chief causes of blindness worldwide and regions with a large range of clinical practices and outcomes.7 The compiled systematic reviews rely on data from randomized and quasirandomized clinical trials and create descriptive compilations where such information is scarce. Primary outcome measurements are visual function such as visual acuity, assessment of visual field, and assessment of vision-related quality of life.7
There are various efforts underway to help drive agendas and priorities within the field of ophthalmology.8 For instance, the Sight Loss and Vision Priority Setting project, initiated in 2011, is supported and guided by the James Lind Alliance.9 The priorities set by this group are within the context of the UK National Health Service. This organization uses the input of patients, caregivers, and health care professionals to identify and prioritize unanswered questions or overlooked issues regarding ophthalmology treatment, diagnosis, and prevention.9 The results are publicized to research commissioning bodies to be considered for funding.9 Li et al10 conducted an eye and vision prioritization exercise on the topic of open-angle glaucoma. This study focused on creating and evaluating a framework to determine evidence gaps and prioritize clinical questions by using clinical practice guidelines and systematic reviews. Unlike the Sight Loss and Vision Priority Setting project, this work aimed to respond to clinicians’ needs and was thus limited to the professionals’ perspective.10
When it comes to research prioritization, burden of disease is one of many factors that may be considered. In particular, the Global Burden of Diseases (GBD), Injuries, and Risk Factors Study 2010 is the result of a worldwide collaboration of 500 researchers that scientifically and systematically quantified various health metrics of loss to diseases, injuries, and risk factors by age and sex.11 The GBD 2010 included 291 diseases and injuries, 67 risk factors, 1160 sequelae (nonfatal health consequences) for 21 regions, 187 countries, and 20 age groups. The GBD 2010 uses improved methods for the estimation of disability weights, which quantify the severity of health loss associated with a particular disease state. Disability-adjusted life-years (DALYs) are a metric established by the GBD to analyze and compare years of healthy life lost. The DALY is a combination of years lost owing to premature death and years lived with disability. These data metrics enable comparison of health data across boundaries of geography and time in a way that is more comprehensive, internally consistent, and comparable than previous data sources.12- 14
The aim of the current study was to assess representation of 8 ophthalmologic categories studied by the GBD 2010 within the CDSR and whether representation corresponds to GBD 2010 disability estimates for each category. The current study is part of a larger series intending to map all 291 diseases studied by the GBD 2010 to representation in major research databases.15,16
The following 8 ophthalmologic categories were studied by the GBD 2010: refraction and accommodation disorders, cataracts, macular degeneration, glaucoma, trachoma, onchocerciasis, vitamin A deficiency, and other vision loss. The other vision loss category included a total of 57 eye conditions including choroidal degeneration, central retinal artery occlusion and diplopia (see eTable 1 in the Supplement for a complete list of conditions in the other vision loss category). Each ophthalmologic category was explored in the CDSR by searching the condition name under title, abstract, keywords (see Table 1 and eTable 1 in the Supplement for International Statistical Classification of Diseases, 10th Revision [ICD-10] definitions and search terms for each ophthalmologic category).18 Both systematic reviews and protocols were considered to assess ophthalmologic disease representation in the database.
Certain GBD categories were ultimately excluded because ophthalmologic conditions were not the focus and GBD metrics could not be attributed solely to the ophthalmologic conditions. These categories included sense organ diseases, other sense organ diseases, and other noncommunicable diseases. The category of vitamin A deficiency was predominantly eye conditions (ICD codes of nonophthalmologic conditions were E50.8. E50.9, and E64.1); thus, it was included.
A systematic review or protocol was matched to 1 of the 8 ophthalmologic categories according to its subject content, and the particular Cochrane Review Group that published each review and date of publication were determined. For a review or protocol to be classified under a disease, the disease was required to be a predominant focus of the abstract objectives and main results. If the systematic review or protocol included the search term in the title, it was automatically included. Reviews or protocols with more than 1 category as a predominant focus were counted once within each respective category. For instance, if a review pertained to both cataracts and glaucoma, it was counted twice.
Methods used by the GBD 2010 to generate DALY estimates have been previously described.13,14 As previously mentioned, the DALY metric combines years of life lost owing to premature mortality and years lived with a disability.19 Disability-adjusted life-year metrics for each of the 8 ophthalmologic categories, expressed as the percentage of total DALYs of 291 conditions measured in the GBD 2010, were obtained from the GBD Compare interactive time plot.20 Using this tool, search parameters of time plot, DALYsmetric, global place, all ages, both sexes, and %units were selected for each ophthalmologic condition. The DALY rankings for the 8 ophthalmologic categories compared with 176 conditions measured in the GBD 2010 were obtained from the GBD interactive arrow diagram.21 Although the GBD project studied 291 conditions, rankings only include 176 conditions, excluding conditions for which the project made no explicit estimates such as etiologies of diarrhea, pneumonia, and more specific conditions under maternal or congenital. Search parameters of global place, all conditions, both sexes, DALY metric, and age standardized were used. The DALY percentage change from 1990 to 2010 was also obtained from the GBD interactive arrow diagram for each ophthalmologic condition. Age standardization was used in the interactive arrow diagram for DALY ranking and percentage change to eliminate effects of population growth and aging over the 20-year span.
Matching of the CDSR representation with GBD disability estimates was accomplished by creating a data plot of representation vs disability to generate a linear line of best fit with a coefficient of determination (R2) and qualitatively determining whether conditions were well matched or not well matched. Because research prioritization of a major database, such as the CDSR, was not expected to solely correlate with a single factor, qualitative rather than quantitative methods were used to match funding with disability.
Two authors (L.N.B. and C.K.) collected data independently, with consensus review during March 2014. Because the current study solely involved nonhuman subjects, institutional review board approval was not necessary.
A total of 116 reviews and protocols represented the 8 ophthalmologic categories, with 106 (91.4%) of these being from the Cochrane Eyes and Vision Group (Table 2). Glaucoma had the greatest representation in the CDSR, with 30 reviews and protocols, followed closely by the other vision loss category, with 26 reviews and protocols, and macular degeneration, with 20 reviews and protocols. Those categories with lower representation were trachoma, with 4 reviews, and onchocerciasis and vitamin A deficiency, both with only 1 review. A list of all included review and protocol titles for each condition is available in eTables 2 and 3 in the Supplement and a list of all excluded titles in eTables 4 and 5 in the Supplement. The range of review and protocol publication year was from 1999 to 2014, with 73.3% occurring from 2010 to 2014; the median and mode year was 2012.
Looking at the global GBD 2010 DALY metrics, other vision loss had the greatest percentage of total DALY (0.25%), followed by refraction and accommodation disorders (0.23%) (Table 2). Vitamin A deficiency had the lowest percentage of total DALY (0.032%) of the conditions assessed. The conditions that experienced an increase in DALY from 1990 to 2010 listed from greatest to least were macular degeneration (percentage change in DALY from 1990 to 2010 was +56%), trachoma (+48%), glaucoma (+31%), other vision loss (+2%), and refraction and accommodation disorders (+1%). The conditions that experienced a decrease in DALY from 1990 to 2010 listed from greatest to least were onchocerciasis (percentage change in DALY from 1990 to 2010 was −31%), cataracts (−30%), and vitamin A deficiency (−9%). The 2010 DALY ranks for the 8 ophthalmic conditions assessed ranged from 67 for other vision loss to 153 for vitamin A deficiency.
Comparing the CDSR representation with DALY metrics, the overall R2 was 0.23, indicating an overall poor correlation between these variables (Figure). Specifically, cataract was the only category that was well matched (Table 2). Glaucoma, macular degeneration, and other vision loss were overrepresented when matched with corresponding DALYs. Conversely, trachoma, onchocerciasis, vitamin A deficiency, and refraction and accommodation disorders were underrepresented in the CDSR relative to corresponding DALYs.
Twenty-six systematic reviews and protocols covered 57 ophthalmologic conditions defined by the GBD 2010 other vision loss category (see eTable 1 in the Supplement for ICD codes and conditions that comprise the other category; see eTable 3 in the Supplement for included titles and eTable 5 for excluded titles). The Cochrane Eyes and Vision Group was responsible for 73.1% of the systematic reviews in the CDSR for these 57 ophthalmologic conditions.
The 8 ophthalmologic categories assessed were represented by a total of 116 reviews and protocols in the CDSR, with most (91.4%) published by the Eyes and Vision Group. While the CDSR covers a broad diversity of ophthalmologic conditions, prioritization is not solely guided by burden of disease data. This was exemplified by the poor correlation between the CDSR representation and DALYs for the ophthalmologic categories, which yielded an R2 of 0.23. Other factors and considerations play critical roles in the prioritization process by the CDSR. In addition, the fact that most reviews and protocols were published from 2010 to 2014 provided evidence of the current and frequently updated nature of the CDSR content.
Cataract was the only category that was well matched with respect to DALY and CDSR representation. Although cataracts had the third highest DALY, the condition experienced the greatest decrease in DALY from 1990 to 2010 of the conditions studied. Age-related cataracts are the leading cause of blindness, causing an estimated 51% of all blindness and affecting 19.7 million people worldwide.22 In addition, surgery for cataracts and refractive errors are among the most cost-effective interventions in health care.22
Glaucoma, macular degeneration, and other vision loss were overrepresented when compared with their DALY metrics. Glaucoma not only had the greatest number of CDSR reviews and protocols, but also experienced a 31% increase in DALY from 1990 to 2010. Similarly, macular degeneration experienced the greatest increase in DALY (+56%) of all the categories. The apparent overrepresentation of these 2 conditions in the CDSR is more accurately reflected by their 20-year disability trends.
Representing 57 eye conditions, the other vision loss category was second in terms of CDSR representation, with 29 reviews and protocols and has the greatest DALY and highest DALY ranking of all the categories. This category experienced a DALY percentage change increase of 2%, the smallest of the 3 conditions that were overmatched.
Trachoma, onchocerciasis, vitamin A deficiency, and refraction and accommodation disorders were underrepresented in the CDSR relative to their disability estimates. Of the conditions studied, refraction and accommodation disorders had the second greatest DALY. While uncorrected refractive errors are responsible for 43% of all visual impairment and affect approximately 120 million people worldwide, most of these cases, such as near sightedness, far sightedness, or astigmatism, could be corrected with the assistance of eyeglasses, contact lenses, or refractive surgery.23 Perhaps this existing knowledge of effective interventions contributed to a relatively constant DALY from this category (+1%) from 1990 to 2010. Similarly, vitamin A deficiency and onchocerciasis demonstrated the lowest 2010 DALYs of the 8 ophthalmologic conditions, in addition to large decreases in DALY from 1990 to 2010. Their low CDSR representation may thus reflect the relative decreased burden of disease over time. In contrast, trachoma was underrepresented in the CDSR but experienced the second greatest 20-year increase in DALY (+48%) of all the conditions assessed. While trachoma is the leading cause of infectious blindness in the world, the condition is irreversible and highly contagious.23 Active efforts, such as those led by the World Health Organization’s Alliance for the Global Elimination of Trachoma, predominantly focus on prevention at the population level.24
Limitations of this study were related to synthesizing ophthalmologic presence in the CDSR and the use of GBD metrics. First, quantification of reviews and protocols in the CDSR was subject to how the reviews and protocols were compiled. For instance, multiple conditions may be covered in one review with a broader focus, referred to as lumping. Alternatively, several reviews may cover the same information, each with a more narrow focus, referred to as splitting.25 Hence, consideration of a systematic review as one unit of representation may not be appropriate for certain topics. In addition, the CDSR reviews synthesize studies and the best evidence in the literature. For this reason, meta-analyses for topics that lack studies in the literature are generally not possible in the CDSR. Variables beyond Cochrane research prioritization determine database presence. Additionally, the assignment of a particular review or protocol by the study authors to a particular ophthalmologic category was not completely objective; the use of independent data collection by 2 authors with consensus review strove to minimize this source of error.
Second, this study was also subject to limitations of the GBD 2010 such as inconsistent or scarce data sources.26 The ICD codes comprising disease categories were established by GBD collaborator consensus. These ICD codes were used to generate search terms and may not have generated all search results pertaining to each topic. Several GBD categories were excluded from the current study that did not focus exclusively on ophthalmologic conditions because the DALYs could not be solely attributed to the ophthalmologic condition. In addition, the GBD 2010 does not provide disability estimates for individual diseases in the other vision loss category. Further breakdown of this category would enable a more in-depth and inclusive study of disease impact. As GBD data become available on an annual basis, analysis of the relationship and trends between database representation and GBD disability over time can be explored further. Finally, this study did not explore the contributions of disease inclusion in the CDSR other than global disease burden such as the skewed financial burden of diseases that disproportionately cause vision impairment in developed populations.
Overall, the global burden of ophthalmic conditions was poorly proportioned to the reviews and protocol representation in the CDSR. There are likely a variety of factors contributing to the lack of correlation. Diseases with known cures, but with significant public health and economic barriers to treatment, such as vitamin A deficiency, were underrepresented in the CDSR. Furthermore, economic impact of diseases disproportionally affecting wealthy populations may contribute to overrepresentation in the CDSR. This article highlighted ophthalmic diseases where global burden fails to correlate with current research and reviews. We suggest that considerations be taken of such underrepresented and overrepresented ophthalmic conditions in pursuit of future research, reviews, and funding.
Overall, global burden of disease comprises one of many factors guiding research prioritization of ophthalmic conditions by Cochrane Review Groups. The process of mapping systematic reviews with disability metrics, such as DALY, provides reproducible, transparent, and valuable data for use in future research prioritization discussions. Certainly, the allocation of research funds and topic prioritization is an intricate process guided by many factors.1 Within Cochrane, prioritization processes are largely determined and maintained by individual Cochrane Review Groups.
This study is part of a series that aims to map all conditions studied by the GBD to representation in various major databases such as the CDSR.16 In the future, we hope that this initial manual mapping will become automated, using the advanced and large-scale methods possible with computer programs and technological advances. Global burden of disease is a useful metric that may guide prioritization decisions. The GBD has created unprecedented public access of electronic information. As GBD metrics become increasingly used by health care professionals, policy makers, and individuals alike, the global burden of disease may ultimately help to reduce research gaps and profoundly impact the human population.
Corresponding Author: Robert P. Dellavalle, MD, PhD, MSPH, Dermatology Service, Department of Veteran Affairs Medical Center, 1055 Clermont St, Box 165, Denver, CO 80220 (email@example.com).
Submitted for Publication: May 13, 2014; final revision received June 24, 2014; accepted July 4, 2014.
Published Online: September 18, 2014. doi:10.1001/jamaophthalmol.2014.3527.
Author Contributions: Ms Boyers and Dr Dellavalle had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Boyers, Karimkhani, Hilton, Dellavalle.
Acquisition, analysis, or interpretation of data: Boyers, Karimkhani, Richheimer.
Drafting of the manuscript: Boyers, Richheimer.
Critical revision of the manuscript for important intellectual content: All authors.
Administrative, technical, or material support: Boyers, Karimkhani, Dellavalle.
Study supervision: Richheimer, Dellavalle.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Ms Boyers and Dr Dellavalle are both employees of the US Department of Veterans Affairs. Dr Dellavalle is also chair of the Colorado Skin Cancer Task Force. Mr Hilton is an employee of The Cochrane Collaboration. No other disclosures were reported.
Funding/Support: Dr Dellavalle receives a salary from the US Department of Veterans Affairs. Dr Dellavalle is also supported by grants from the Centers for Disease Control and Prevention and National Institutes of Health.
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclaimer: The opinions expressed herein do not necessarily reflect those of the US Department of Veterans Affairs, the Centers for Disease Control and Prevention, the National Institutes of Health, or The Cochrane Collaboration.