Accessing Evidence from Research Studies Conducted in Low-and Middle-Income Developing Countries

Member states should... collaborate with the Organization in developing and reinforcing comprehensive cancer control programmes tailored to the socioeconomic context, and aimed at reducing cancer incidence and mortality and improving the quality of life of cancer patients and their families, specifically through the systematic, stepwise and equitable implementation of evidence-based strategies for prevention, early detection, diagnosis, treatment, rehabilitation and palliative care and to evaluate the impact of implementing such programmes. World Health Assembly Resolution WHA58.22: Cancer Prevention and Control1

Resolution WHA58.22 presupposes the existence of an accessible evidence-base of relevant research that can be used to guide and inform effective cancer control plans, a sufficient cadre of scientifically trained health professionals and the necessary infrastructure and resources to evaluate their outcomes¹. Existing evidence, however, is largely derived from studies conducted in the high-income countries. Eight countries (USA, England, Germany, Netherlands, Japan, Switzerland, France and Italy) account for over 80% of the journals indexed on Medline 2007 ². While scientific research in high-income countries has provided a solid foundation upon which to build cancer control measures, its context is very different from that of developing countries. Translational research, for example, is tailored almost exclusively to the needs of high-income countries. High-, middle- and low-income countries differ widely with respect to the types and patterns of cancer, levels of cancer incidence, patient genetics, health care services, socioeconomic status, literacy rates and behavior of populations as well as with respect to the resources they can command (human, physical and financial). Hence, cutting-edge research may offer little practical benefit to clinicians who practice in a context of limited institutional resources and a poor, less-educated population; although there are exceptions to this rule e.g. the development of vaccines.

In light of the remarkably different context in which cancer control must be practiced in developing countries, WHA resolution WHA58.22 also requested that the Director-General of WHO should promote and support research that evaluates low-cost interventions that are affordable and sustainable in low-income countries¹. Reports of studies conducted in low- or middle-income countries, however, are not only few in number, but are significantly harder to access, since they are often not published in journals indexed by the major databases. Whereas research papers published by authors living in the USA or UK accounted for 32% of the cancer research indexed on PubMed in 2006, papers from the whole of Africa accounted for only 0.5% of the indexed cancer articles³. The task of gathering and assessing the value of non-indexed research evidence is, therefore, much more difficult.

A pilot study designed to assess the magnitude of these difficulties is being jointly undertaken by INCTR, the European School of Oncology (ESO) and the National Cancer Institute, Cairo. Preliminary results are presented here. The initial study entailed conducting a search of the Journal of the Egyptian National Cancer Institute (2000-2007)⁴ and the Yearbooks of the National Cancer Institute, Cairo (2000-2006)⁴ for published reports of original research in breast cancer, bladder cancer, lymphoma, pediatric cancer and palliative care conducted in Egypt in the years 2000-2007 inclusive. The results were compared with electronic searches of four databases (Medline, Embase, Scopus and PsycLit) using search terms associated with the type of cancer or palliative care in the title, abstract or medical subject heading (MeSH) combined with 'Egypt’ or ‘Egyptian’ in the title, abstract, address or affiliation details. Animal studies were excluded. Searches were conducted between January and April 2008 and the results downloaded into a bibliographic database (ProCite 5.1).

Of 1356 studies identified, 942 met the inclusion criteria. J Egypt Nat Cancer Inst was indexed on Medline (2003-2007) and Scopus (2005-2006) but not on Embase or PsycLit. Searches of the J Egypt Nat Cancer Inst 2000-2007 identified 140 studies, 70 (50%) of which were also cited in NCI Cairo Yearbooks 2000-2006. Searches of the NCI Cairo Yearbooks 2000-2006 identified 241 citations, of which 77 (32%) were reports of dissertations or unpublished studies and 164 (68%) were reports of published studies. The 77 citations, dissertations or unpublished studies were excluded from the analysis, leaving 865 citations of published studies that met the inclusion criteria. The results of the searches are presented in Table 1.

Search Category J Egypt Nat Cancer Institute 2000-2007 NCI Cairo Yearbooks 2000-2006 Dissertations, Unpublished studies  Medline   Embase   Scopus   PsycLit  Published studies not indexed Breast cancer (290) 44 64 24 84 88 172 94 33 Bladder cancer (319) 28 45 5 111 100 232 105 27 Lymphoma (134) 29 39 15 51 41 56 36 24 Pediatric malignancies (99) 25 34 10 27 11 13 43 20 Palliative care (100) 14 59 23 20 5 36 16 25 Total studies 140 241 77 289 245 509 288 129

Table 1. Studies in the indicated categories identified by searches of J Egypt Nat Cancer Inst (2000-2007), NCI Cairo yearbooks (2000-2006) and the Medline, Embase, Scopus and PsycLit databases.

Access to studies indexed on the four databases was variable. Of 865 published reports identified by the searches only 81 (9%) were indexed on all four databases (as determined by checking reports individually) and 129 (15%) were not indexed on any of the databases. Of the 164 published studies cited in the Yearbooks of the National Cancer Institute, Cairo (2000-2006) and 140 included studies published in the J Egypt Nat Cancer Inst , 95 (58%)and 77 (55%) respectively, were not indexed on any of the four databases searched. The yield rate according to the number of databases searched are presented in Table 2.

Published studies Indexed on all 4 databases Indexed on 3 databases Indexed on 2 databases Indexed on 1 database only Published studies not indexed on databases Non-indexed studies published in J Egypt Nat Cancer Inst 2000-2007 Non-indexed published studies cited in NCI Cairo yearbook 2000-2006 Breast cancer (266) 25 41 45 121 33 19 25 Bladder cancer (314) 42 49 37 159 27 16 20 Lymphoma (119) 12 11 25 47 24 17 15 Pediatric malignancies (89) 1 4 14 50 20 14 12 Palliative care (77) 1 5 11 36 25 11 23 Total studies (865) 81 110 132 413 129 77 95

Table 2. Access to studies according to the number of databases searched.

From these results we can draw three conclusions: (1) coverage by indexing on the general medical databases is, at best, partial; (2) routine searching of one or even two databases is insufficient to adequately identify studies and (3) studies conducted in developing countries may be seriously under-represented on the major databases.

Less that 50% of the studies published in the J Egypt Nat Cancer Inst and NCI Cairo Yearbooks were included in the four major databases. This is partly because Medline has not yet indexed articles published in J Egypt Nat Cancer Inst prior to 2003 and partly because many of the articles cited in the NCI Cairo Yearbooks are published by journals not indexed by the databases.

Fifty years ago Aneurin Bevan observed: many people have died and have suffered, not because knowledge was lacking, but because they did not have access to it. To all the suffering that attends illness there is the added bitterness that, if the poor could have had access to the knowledge available, they might have been saved or at least might have been helped⁵. The continuing difficulties in both the creation of and access to evidence from research conducted and published in developing countries has five negative effects.

1. Reliance must continue to be largely upon the outcomes of studies conducted in the high-income developed countries, which may not always be relevant.
2. The barriers to dissemination of information derived in an appropriate context may lead to delays in improving the care of patients.
3. Exclusion of data from developing countries in the major data bases infers that research conducted in developing countries Is inferior to that conducted in the developed countries; although this may often be true, its value may be underestimated.
4. Because research conducted in developing countries is only partly visible, the evidence-base appears smaller than it actually is. National research profiles are minimized and motivation to invest in, and conduct, research is lessened.
5. The preparation and maintenance of context-relevant systematic reviews of effects and effectiveness of health care interventions and of the quality assessment of research methodology is hindered.

1. Raise the international profile of research conducted in Egypt.
2. Provide an accessible, comprehensive and relevant evidence-base that will lead to improvements in clinical practice.
3. Identify gaps in the evidence base, thereby stimulating and guiding the conduct of additional, nationally relevant research.
4. Contribute to improving the standard of published research by facilitating comprehensive systematic reviews of methodological quality.
5. Help prevent possible duplication of research effort.
   The emphasis in Resolution WHA58.22 on evidence-based strategies is well placed, but there is a paucity of evidence created in the context of low- and middle-income countries. We hope that building the Database for Cancer Control in Developing Countries - piloted in Egypt, then expanded in a stepwise process - coupled with training in systematic review, will help lay the foundation for both improved public health interventions and patient care while simultaneously encouraging the development of a culture of scientific research in low- and middle-income countries.