Author + information
- Poornima Prabhakaran, MBBS⁎,
- Vamadevan S. Ajay, MPH⁎,
- Dorairaj Prabhakaran, MD, DM, MSc†,
- Arun Kumar Gottumukkala, PhD⁎,
- J.S. Shrihari, MPH⁎,
- Uma Snehi, MBBS⁎,
- Bijoy Joseph, PhD⁎ and
- Kolli Srinath Reddy, MD, DM, MSc⁎
Cardiovascular diseases (CVDs) and diabetes are major public health problems worldwide. They accounted for 32% of the total deaths globally and contributed to 11% of the global burden of disability adjusted life years (DALYs) in the year 2005 (1). The fast-changing demographic profile in the developing world has resulted in the emergence of cardiovascular disease and diabetes as major contributors to disease burden, leading to substantial increases in economic and health care costs.
Developing countries comprising low-income countries (LIC) and low-middle-income countries (LMIC), because of their larger population, account for a substantially greater DALY loss as compared with developed countries (2,3). The Global Burden of Disease Study projects a 55% increase in DALY loss attributable to CVD between 1990 and 2020 in developing countries (4). In contrast, developed nations will witness a 14.3% reduction in the proportion of DALY loss attributable to CVD during the same period. Thus, the increasing burden of CVD would be borne mostly by the developing countries in the next 2 decades. To address this increasing challenge, these countries need to develop effective national policies, strategies, and programs aimed at the prevention of CVDs. For this to be achieved, evidence needs to be generated from locally conducted research. Applied research is also needed to support evidence-based practice, to develop context specific guidelines, and to understand the complex dynamics of the CVD epidemic in these country settings. It has been demonstrated that research translates to better technologies and diagnostic methods, improved application of cost-effective interventions, more informed health care users, and the promotion of healthy lifestyles. However, the commitment of various governments to invest in building research capacity varies greatly. In 1990, the Commission on Health Research for Development (CHRD) estimated that <10% of the global health research resources (totaling U.S. $30 billion/year in 1986) were accessed by the developing countries, which accounted for >90% of the world’s health problems (5). Despite an increase in the global health research expenditure to U.S. $55.8 billion by 1996, the “10/90 gap” has persisted (6). Similar patterns were observed when the Global Forum for Health Research compared research and development during the years 2001 and 2003 (7). This study demonstrated a U.S. $20 billion increase in research and development for health expenditures within 2 years. However, the spending among high-income countries accounted for most of the increase globally (7). On the other hand, in developing countries, allocation to the health sector is low and to health research specifically is almost negligible (7). In 2001, an assessment of global resources for health research noted that none of the developing countries studied had met the 1990 CHRD recommendation to spend as much as 2% of the national health budget on health research: only India, Brazil, and Cuba had spent notably closer to the target of 1% to 2% (8). Likewise, most countries fall short of the recommended allocation of 5% of their health development budget for health research and research capacity building (5). Such inadequate investments could potentially translate to low research capacity, poor-quality research, and low research outputs, especially in CVD research, in developing countries, which are additionally burdened by pre-transitional diseases such as communicable disease and undernutrition.
Research publications in peer-reviewed journals, although not the best indicator, give a reasonable indication of the type and impact of research conducted by health professionals. Quantification and comparison of research publications from developed and developing countries would be potentially useful to funding agencies and governments in informed policy-making for addressing inequalities in health, setting priorities, health resource allocation, and financing health research. However, data on the nature of research publications are limited both from developed and developing countries. Mendis et al. (9) attempted to quantify the CVD research output from different categories of nations. They searched the MEDLINE database and reported that developed market economies contribute 78% of the research publications, highlighting the research gaps in the developing nations. Earlier studies on the geography of biomedical publications in general (10), and of mental health (11) and HIV/AIDS-associated publications (12) have also reported similar findings. Even within developed country settings, differentials exist in CVD research outputs. De Jong et al. (13) showed that the U.S. leads research in clinical cardiology among G-7 nations. Further, in most G-7 nations, the quality and quantity of cardiological publications lag behind those of the smaller Western European countries (13).
However, all of these studies did not attempt to identify the type of research publications in the field of cardiovascular diseases. Specific areas of research that would be important for policy-makers and funding agencies are research addressing issues of equity, such as health systems, cost-effectiveness, and quality of life. To quantify the amount of research and to delineate the type of research in CVD, we undertook a review on CVD research-related publications among several countries belonging to the various income groups as defined by the World Bank (14). We also sought to compare the proportion of CVD research related to health system, quality of care, costing/cost-effectiveness among selected developed, and developing countries.
We used the World Bank classification of countries into high-income (HIC), high-middle-income (HMIC), LMIC, and LIC based on their gross national product estimates of 2004 (14). A total of 90 countries were selected for the study. We initially selected the top 21 countries from each of the 4 income groups. To include large/most populous countries, we also included 6 large countries that did not feature in the original selection. These countries were Russia, South Africa, Argentina, Egypt, Indonesia, and Bangladesh. The complete list of countries is given in Table 1.
Our initial search strategy was to identify the most sensitive database yielding the maximum number of search results with appropriate key words. We made multiple searches using different key words, such as “cardiovascular”, “heart”, “coronary”, “heart failure ”, “rheumatic”, “disease”, and “country name” in MEDLINE, PubMed, and EMBASE for randomly selected countries from the aforementioned list. We found that, for any particular country, PubMed retrieved the maximum number of publications with the aforementioned search terms. We therefore used the PubMed database for the purpose of this study.
A PubMed search was performed for all 90 countries using the previously mentioned key words. The search was limited to publications from human studies, in all languages, for the years 1994, 1995, 2004, and 2005. We chose these years to evaluate changes in trend of publications among the selected countries. We excluded publications not related to CVDs or diabetes. The details of the search strategy are provided in the schematic diagram (Fig. 1).
Country Group-Wise Distribution of CVD Publications
During the reference period of 4 years (1994, 1995, 2004, and 2005), a total of 253,681 articles were published from 90 countries. In the initial search we obtained 17,531 articles pertaining to CVD. Of these 17,531 published articles, we excluded 1,321 articles as they were not related to CVD. Of the remaining 16,210 (6.4% of the total) articles, 9,823 were original publications and the remaining were ancillary publications (see Fig. 1for the classification). The U.S. had the greatest number of CVD-related publications (6,161), whereas 10 countries (5 from LIC) had no such publications during this period. The majority of the CVD-related publications were from HIC (n = 13,354; 82.4%). Contributions from other country groups were considerably small, that is, HMIC (n = 1,145; 7.1%), LMIC (n = 1,077; 6.6%), and LIC (n = 634; 3.9%). The details are shown in Figure 2.Of the original publications, most (n = 7,107; 72.3%) were observational studies. Further, two-fifths of the observational studies were either case reports/case series (n = 2,940; 41.3%). The next largest group of publications were randomized control trials (n = 1,519; 15.5%) followed by basic science research and genetic studies (n = 618; 6.3%). The publications related to health system, quality of care, cost, and cost-effectiveness accounted for a minuscule 5.9% (n = 579) of original publications.
Profile of Original Publications by Country Groups
The profile of the original publications is shown in Figure 3.In brief, when analyzed country wise, the observational studies formed the major publication group among the 4 country groups: HIC (n = 5,550; 71.1%), HMIC (n = 637; 78.4%), LMIC (n = 592; 73.5%) and LIC (n = 328; 82.0%). Proportion of publications from clinical trials ranged from 10% to 17%, with the lowest number (n = 59; 10.0%) from LIC. The proportion of publications from basic science and genetics research formed 13.2% in LMIC, 9.2% in HMIC, 5.4% in HIC, and 3.3% in LIC. The largest proportion of publications related to health system, quality of care, and cost/cost-effectiveness was observed from HIC (n = 543, 94% of publications in this category). The HMIC and LMIC contributed to a smaller proportion (3% each) in this category. None of the LIC had any publication conforming to this category.
Change in the Number and Pattern of Publications Over a Decade
We compared the number of publications in the years 1994 to 1995 with the years 2004 to 2005. We observed a 130% increase in the number of publications related to CVD (4,912 vs. 11,298); this increase was observed across all of the country groups. Despite this increase, we did not observe any change in the ratio of publications between the high-income (HIC and HMIC) and low-income (LMIC and LIC) groups of countries (8.6 in 1994 to 1995 vs. 8.4 in 2004 to 2005; p = 0.64). Furthermore, the proportion of the subcategories of publications remained similar between the 2-year groups under comparison.
Quality of Publications
To assess the quality of the published studies, we randomly chose 250 CVD publications from each country group and assessed their impact factor by obtaining the impact factor values listed at the website in reference 15. Of the 1,000 articles (250 from each of the country groups), <50% (n = 479; 49.7%) were published in journals with a listed impact factor at bioscience.org. When disaggregated into World Bank country groups, HIC had the highest publication in a journal with a listed impact factor (n = 148; 60%) followed by HMIC (n = 118; 47%) and LMIC (n = 16; 46%), and LIC had the least publication in journals with impact factor (n = 97; 39%). This trend was statistically significant (p < 0.001). The mean impact factor of publications was again greater in HIC (2.96) followed by LMIC (2.35), HMIC (2.01), and LIC (1.87).
Publications in Relation to the Population
We compared the number of publication from each of the country groups with their relative population size. We used the mid-2005 population for the population size (16) and found that HIC (which accounts for 14.6% of the population of these 90 countries) had the greatest number of publications per hundred thousand population (27.3/100,000). Similarly, the number of CVD-related publications per hundred thousand population (1.83/100,000) was also greatest among this group. We also found that HMIC, accounting for 9.3% of the total population from the 90 selected countries, had 7.2% and 7.1% of total and CVD publications, respectively. This number translated into 0.25/100,000 CVD-related publications per population in the HMIC. The LMIC and LIC groups, which together account for 76% of the population, had just 14.3% of total publications and 10.5% of CVD publications. In addition, LIC had the least number of CVD-related publications per hundred thousand population (0.04/100,000), whereas LMIC had a slightly greater number of publications per hundred thousand populations (0.05/100,000). Publications on health system, quality of care, and cost/cost-effectiveness related to CVD were substantially low among LMIC (0.001/100,000 population), and publications in this category were nonexistent in LIC. The corresponding figures for HIC and HMIC were 0.074/100,000 and 0.004/100,000, respectively. The details are presented in Table 2.
Publications and Health Expenditure Indicators
To ascertain the relationship of government spending and health research output, we matched the total number of CVD-related publications against the expenditure on health in that country. As indicators of health expenditure, we used both the total expenditure on health as a percentage of GDP and the per capita government expenditure in health at international dollar rates, which are available from the website of World Health Organization (17). For this comparison, the countries were grouped into 2 (above and below the mean health expenditure indicators) and were analyzed using a ttest with the SPSS software package (version 11.5, SPSS Inc., Chicago, Illinois).
Health expenditure indicators were available for 78 countries. To evaluate the relationship of health expenditure and publications, we categorized these countries into 2 groups according to their total spending on health care as a percentage of gross domestic product (GDP), that is, countries spending more or less than 6.7% of their GDP (which was the mean value for these 78 countries). Countries spending more than 6.7% of their GDP toward health had significantly greater number of CVD-related publications (p = 0.042) as compared with those with lower allocation levels. A similar comparison was made between country groups categorized according to the per capita government expenditure (at international dollar rate) on health care. Countries with more than $729 as per capita government expenditure (mean value for 78 countries) on health had a significantly greater number (p < 0.001) of CVD publications.
What Are the Reasons for the Insufficient Research From LIC and LMIC?
Potential reasons for such insufficient research activity could be several but speculative. First, limited funding for health care has a cascading effect on health research, specifically in basic and epidemiological research, which is resource intensive. Again, resulting from a lack of resources, research outputs from clinicians in countries with poor research support are likely to be restricted to case reports and case series. Such trends are particularly evident from the LMIC and LIC publications. Second, non–English-speaking countries tend to have lower research output in the highest-ranking general medical journals. Man et al. (18) reported that Asian countries had generally low rates of publication in the five highest-ranked general medical journals when compared with English-speaking nations and certain northern European countries such as Denmark, the Netherlands, Switzerland, and Sweden (18). Third, governments and policy-makers in LIC and LMIC believe that noncommunicable diseases are not a major problem and consequently apportion lower resources for research in this area. In an earlier study conducted by our group in 3 developing countries, policy-makers did not perceive CVD as a major problem and considered pre-transitional communicable disease and nutritional diseases to be greater threats among the population (19). Such beliefs could potentially influence health research funding patterns and the resource flow in the health sector. Identifying financial flow in health research is the key to mapping various constraints in CVD health research. Hence, it would be appropriate to carry out in-depth analysis of the funding pattern for CVD research in developing countries using appropriate methodologies such as the Health Research Financial Flow Matrix developed by the Global Health Research Forum (20).
Implications for Governments and Policy-Makers
This study highlights the differences in the publications related to CVD among countries of the world. Studies on health system, cost/cost-effectiveness, and quality of care formed a small proportion of publications overall and were minuscule or absent among LIC and LMIC. We found that the number of publications per 100,000 population was almost 37 times greater in HIC as compared to that of LIC and LMIC.
Despite the 15-year gap, our results almost mirror the findings of the CHRD conducted in 1990. The CHRD finding of a significant “10/90” gap in health and health research between developed and developing countries remarkably did not change even after 10 years. Although 80% of CVDs occur in LIC and LMIC, even now, the research output from these countries is just around one-tenth of global research output on CVD. Therefore, there is an urgent need to alter this pattern and evaluate methods for channelizing resources for health research to the needier, LIC and LMIC where there is a definite upsurge in chronic diseases, affecting increasingly younger population groups. Given the low priority afforded to research related to altering health systems and evaluating cost-effectiveness and quality of care related to CVD, especially in the LIC and LMIC, governments and funding agencies need to prioritize this area of research funding.
In this study, not all databases have been included, but are confined to the PubMed database and articles published in journals indexed in PubMed alone, limited to a period of 2 years. This may underestimate the proportion of publications from developing nations, as many of their journals may not be indexed in PubMed. However, the validity of our results is suggested by another publication that also obtained a nearly similar proportion of CVD publications in the years 1996 to 2001 using the Scientific Citation Index (21). We also did not account for gray literature in this study, which could be another source for underestimation of research publications among LMIC and LIC, because most researchers from these countries may be publishing in local journals that may not be indexed. We did not evaluate the funding sources, the flow, and pattern of funding in this study. These issues need to be addressed in a separate study focusing on mapping of research flow and funding patterns in a few selected countries.
The research output (measured through PubMed citations) is low among LIC and LMIC. Specifically, research publications that could influence policy are virtually nonexistent in LIC and LMIC. On quality parameters, as assessed by the mean impact factor of publications, LIC and LMIC, although having lesser impact factor levels, did not fare worse, as compared with the quantity parameters. A major reason for the lower output may be the low government spending among LIC and LMIC.
- American College of Cardiology Foundation
- World Health Organization
- World Health Organization
- Lopez A.D.,
- Mathers C.D.,
- Ezzati M.,
- Jamison D.T.,
- Murray C.J.L.
- Murray C.J.L.,
- Lopez A.D.
- Commission on Health Research for Development
- WHO Ad-Hoc Committee on Health Research Relating to Future Intervention Options
- Global Forum for Health Research
- Global Forum for Health Research
- Yach D.,
- Kenya P.
- de Jong J.W.,
- Schaper W.
- World Bank
- ↵Frontiers in Bioscience. Impact Factor of Journals in the Order of Impact Factor. Available at: http://www.bioscience.org/services/impact15.htm. Accessed June 28, 2007
- The Population Reference Bureau
- World Health Organization
- Global Forum for Health Research
- Country Group-Wise Distribution of CVD Publications
- Profile of Original Publications by Country Groups
- Change in the Number and Pattern of Publications Over a Decade
- Quality of Publications
- Publications in Relation to the Population
- Publications and Health Expenditure Indicators
- What Are the Reasons for the Insufficient Research From LIC and LMIC?
- Implications for Governments and Policy-Makers
- Study Limitations