This study describes a bibliometric analysis of mental health research publications between 1980 and 2008. Over 350,000 papers on mental health research from 20 countries, accounting for over 95% of research output, were identified in the Thomson Reuters Web of Science. The bibliometric characteristics of these papers were analysed, which led to a set of ten key observations:
- The volume of mental health research publication output is growing faster than that in biomedicine or science as a whole.
- There is a convergence in the volume of mental health research publication output between the US and Europe, with the recent emergence of mental health research publication output in the BRIC countries.
- Smaller countries such as Sweden, Israel, The Netherlands and Finland are the most productive in terms of number of publications per capita.
- Countries with greater specialisation in mental health research, such as the US, the UK, Canada and The Netherlands, generally have higher scientific impact.
- Spain, France, South Korea, Japan and Turkey obtain fewer citations than the world average for mental health publications and publish fewer mental health publications than one would expect, given the volume of their overall publication output.
- Over the last 25 years, we observe a tendency towards publishing more basic research outputs in mental health. The only exception to this is The Netherlands.
- Mental health research publications are increasingly the result of international collaboration. While 3% of all mental health papers published in 1980 was the result of an international collaboration, that percentage was 20% in 2008.
- Countries/regions with a relatively large volume of research output—such as the US, the EU and Japan—tend to obtain relatively low international collaboration rates. This is not surprising as it is more difficult for researchers from a bigger country to find collaborators outside their borders.
- The strongest bilateral relationship in mental health research is between Canada and the US, followed by the UK and the US, Germany and the US, and Italy and the US.
- We observe a strong growth of research on autism, anxiety, bipolar disorders, hyperactivity, memory, schizophrenia, sleep and stress since the early 1980s.
The analysis provides trend and benchmark data on mental health research that will be updated on a regular basis. By their very nature these conclusions are descriptive, and do not (and cannot) explain why trends have occurred. That said, we may infer a number of emergent policy observations that warrant further and subsequent investigation. These include the following:
- The performance of national mental health research systems varies. There is an apparent correlation with the size of the country; it seems that countries that are small, specialised or both have a relatively high citation impact. Large countries that do not specialise have a lower impact.
- The rise of mental health research seems to be due to funding. There is a disproportionate increase in mental health research publications compared to all biomedical science. Though it is notoriously difficult to estimate the amount of funding going into different fields of biomedical science, research productivity (e.g. cost per paper) has not radically changed in either the US or the UK over the periods assessed, suggesting that funding is the major driver of publication increase.
- Is mental health research lost in translation? One of the frequent observations made about mental health research is that it has failed to be effectively translated from bench to bedside. Over the period analysed the output of basic research increased at a faster rate than that of clinical research. The question that remains unanswered is whether the mental health translation gap is due to funding policies or lack of scientific tractability.
The research described in this article was prepared for The Science of Science for Mental Health Research Network (SOS for Mental Health) and the Graham Boeckh Foundation and conducted by RAND Europe.