General practice-based cancer research publications: a bibliometric analysis 2013–2019

Summary

It is believed that this is the first article to explore the breadth and depth of cancer in general practice research using bibliometric analysis. The results suggest cancer in general practice is a stable field where research is predominantly descriptive or observational, and where there is less evidence outlining implementation or translation into clinical practice. This could suggest not only that there remains a paucity of clinical trial evidence, but also that further emphasis on interventional and/or implementation research is still needed. A wide geographical variation was also found in the focus of cancer in general practice research, and certain tumour types that are over-represented compared with their population disease burden.

A fundamental goal with general practice research is to develop evidence to inform clinical practice and improve outcomes for patients. The results suggest that most of the published cancer in general practice literature sits early in the research translation pipeline (that is, research that is still working towards understanding a problem or clinical need rather than testing an intervention to address this problem).¹⁶ The smaller number of RCTs and implementation studies suggests greater support is needed for so-called T2 and T3 translational studies where evidence is applied in practice.¹⁶ It further highlights the long trajectory to integrate evidence into clinical recommendations¹⁷ and the need to provide support to improve research translation. In Australia, only 18% of cancer research funding is awarded to studies that have moved beyond early translation.¹⁸

Cancer in general practice research is published in a wide range of journals, with only three of the top 10 journals being primary care specific, suggesting researchers target general medical and cancer-specific journals in which to share their findings. This may reflect a pattern of researchers chasing higher impact factor journals first, which are more likely to be in these broader areas than primary care- specific ones.¹⁹ More than two-thirds of publications came from journals with an impact factor of <5, suggesting that leading primary care journals have lower impact factors and/or that it may be difficult to publish cancer in general practice research in higher impact factor general medical and cancer-specific journals. A 2007 bibliometric analysis of cancer research in medical literature found that 25% of publications were published in the top 20 medical journals with the highest impact factors.¹⁰ The top 10 medical journals had an average impact factor of 30.3, which is nearly 10 times higher than that found in the present study.

Strengths and limitations

The search strategy was designed to retrieve many publications relevant to cancer in primary care. Given the complexity of the structure of primary care and general practice between different countries, it was chosen to focus the bibliometric review on general practice only. This limits the applicability of results to this setting rather than primary care more broadly. This complexity is also a limitation to identification and interpretation of international collaborations in cancer in general practice research. Additionally, there is a dearth of relevant literature to compare the search strategies employed by other studies.

The authors tried to mitigate the variation in general practice nomenclature between countries by being inclusive around language used to describe both primary care and primary healthcare professionals. Not all general practice research is accurately labelled or signposted in indexing applied by the databases used for this search. This is particularly important when considering prevention research where the link to cancer may be less explicitly stated and result in an artificially lowered publication count.

This analysis represents research over a 7-year period recorded in two key databases. The analysis only applied to indexed journals and consequently does not cover unpublished research, works in non-indexed journals, and non-journal outputs such as books, theses, reports, or government documents. Given the English language restriction, this study does not capture all cancer in general practice publications internationally. Differences in healthcare systems between countries created an additional layer of complexity in identifying whether research was conducted in, or significantly related to, general practice. Overall, the results of this study illustrate the difficulty in accurately assessing research publications across areas of the cancer continuum based on the complexity and variation in the structure of general practice in different countries.

Some journals publish authors with first name initials only. The research team tried to identify the correct researcher based on authorship, but this was difficult in some cases. Additionally, information about impact factors and the appropriate figure at time of publication was not always available.

Lastly, impact factors remain a contentious measure of impact and have known limitations such as the length of citation windows, differences between disciplines, and the impact of commentaries and editorials artificially inflating calculations.²⁰ Alternative metrics are increasingly being used to assess impact. Given the varied approaches to journals incorporating alternative metrics, in this context, impact factor was a more widely available variable to collect.

Comparison with existing literature

Overall, cancer screening, diagnosis, and survivorship were the most investigated areas, although coverage across the cancer continuum varied considerably by country. This may highlight the different research priorities and targeted funding within countries. For example, from 2006–2011, Australia and the UK had similar levels of investment into cancer research when measured per head of population and as a percentage of the gross domestic product.²¹ In 2014, Cancer Research UK, the world’s largest independent cancer research charity, began substantially investing in early detection research through their national Early Diagnosis Initiative. This includes the International Alliance for Cancer Early Detection, which is a £55 million partnership between Cancer Research UK and the US.²² This initiative was partly in response to poorer survival outcomes for UK patients with cancer when compared with countries such as Australia, Canada, and Sweden, which was demonstrated through the International Cancer Benchmarking Partnership.²³ While the pattern of research investment across the cancer continuum has been similar between the UK and Australia,²¹ key organisations, such as the Cancer Council Australia, have invested more in survivorship research. Cancer Council Australia is the largest non-government funder of cancer research in Australia.²⁴ In 2020, they invested AUD$54 million into cancer research, over 70% of which was awarded to support and survivorship research.²⁴

The distribution of publication by cancer type also highlights the well-known mismatch between disease burden and funding allocation.^25,26 Publication outputs can be used as a proxy measure for funding and investment.²⁶ In Australia, the cancer type with the greatest difference between incidence and mortality is liver cancer.²⁷ It is the 15th most common cancer type but represents the eighth most common cause of cancer-related deaths.²⁷ Only three Australian publications were found to be investigating liver cancer.^28–30 However, the proportion of lung cancer publications identified in this study is comparable with current broader lung cancer research, representing only 6.5% of cancer research.³¹ This small percentage does not align with the global burden of lung cancer, as the second most commonly diagnosed cancer and the leading cause of cancer death.¹ A 2020 bibliometric analysis of lung cancer research worldwide identified that almost all countries, except for China, have seen a decline in lung cancer research outputs over the past 15 years.³¹ In Australia, lung cancer is the greatest contributor of cancer disease burden but receives less research funding than colorectal, breast, prostate, melanoma, leukaemia, brain, ovarian, and liver cancer.¹⁸ The present study revealed that at least twice as many publications were about either breast or colorectal cancer in comparison with lung cancer. The higher frequency of these publications may be owing to the strong existing role of general practice in the care of patients with these cancers. For example, many countries have population- based colorectal cancer screening where general practice involvement has been widely researched.^32–35 Additionally, the prevalence of breast and colorectal cancer publications aligns with that seen in the wider cancer literature, both of which are in the top three cancer types published.³⁶ As countries consider future models of implementing lung cancer screening, it is possible that this pattern of research activity in general practice may change.

Overall, the variation in distribution of research publications by cancer type suggests that this inequity in cancer research funding may also be reflected in the support provided to primary care-based research.^25,37 The potential mismatch in research funding needs to be considered in the context of variation of disease prevalence and economic standing between countries.³⁸ These two lenses, along with the structure of the local health system, all likely play a role in the geographical variation seen in publication outputs by cancer type and area of the continuum.

Implications for research and practice

The findings suggest that a large proportion of cancer in general practice research is observational. There is a need to provide additional support to assist researchers to use this evidence to underpin an increase in the development of interventional research. Additionally, the results suggest that certain tumour types are over-represented in the cancer in general practice literature compared with their population disease burden. Future consideration should be given on how to increase research of under- represented areas including both specific tumour types and areas of the cancer continuum such as cancer prevention.

There are a number of aspects of medical research that demonstrate sex inequality, such as research grant applications,³⁹ and authorship of publications where, depending on the research field, females are significantly less likely to occupy first or last author positions.^40–42 The results suggest this inequality may not extend to authors working in cancer in general practice research, and, in fact, female authors were found to represent two-thirds of first authors. This result is in line with a bibliometric study in primary health care and general internal medicine that found 63% of first authors in primary healthcare journals were female.⁴³

International collaborations appear to be limited and consideration of how to better enable the development of large-scale, multi-country studies may be valuable. This is also affected by the known limited capacity of low- and middle-income countries to undertake primary healthcare research.⁴⁴

In conclusion, it is believed this is the first study to explore cancer in general practice research using bibliometric analysis. The study provides an interesting insight into the breadth of cancer in general practice research. It offers a concise view of the type of research conducted, where it was conducted, and the different research focus between different countries. It could also be used for future priority-setting exercises by funders and researchers to determine where best to invest limited resources for primary care cancer research.