Health promotion research literature in Europe 1995–2005
A. Clarke1, M. Gatineau2, M. Thorogood3 and N. Wyn-Roberts4
1 Health Services Research Institute, University of Warwick, UK
2 Public Health Resource Unit, Oxford, UK
3 University of Warwick
4 Healthcare Libraries, University of Oxford, UK
Correspondence: Aileen Clarke, FFPH, Health Services Research Institute, University of Warwick, Health Sciences Research Institute, Medical School Building, Gibbet Hill Campus, University of Warwick, Coventry, CV4 7AL, UK, tel: +44 (0) 24 76 51809, fax: +44 (0) 24 765 28375, e-mail: aileen.clarke{at}warwick.ac.uk
Received May 26, 2007, accepted May 28, 2007
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionConclusionsAcknowledgementsReferences |
|---|
Background: To undertake an overview of health promotion research in the EEA to inform the collaborative study—SPHERE (Strengthening Public Health Research in Europe). Methods: A ‘filter’ (search strategy) was used to search Medline and Embase for a 10-year period from 1995 to 2005. A 32% (6000) sample of the filter output was assessed for proportion constituting health promotion. Output was analysed by country, population, gross domestic product (GDP) and health need (disability-adjusted life years, DALYs). Disease prevention (screening and immunization) and health improvement papers were separately identified. The latter were classified by methodology, level of intervention and topic area. Results: 18 862 papers were identified. One-third was identified as health promotion (2206/6000, 36.7%) equivalent to 6935 (CI 6651–7230). Production varied: Nordic countries were highest producers per million population; the UK the largest net producer. There was a weak relationship between health promotion publication and population size (r2 = 0.38); a weak inverse relationship with relative health (DALYs per million population) (r2 = 0.07) and a slightly stronger relationship with GDP (r2 = 0.45). Twenty-eight percent (626/2206) of the papers identified were disease prevention (screening and immunization). The largest topic areas of the remainder (1580) were diet and exercise, smoking and tobacco, and cardiovascular disease reduction. Accidents and violence, alcohol and mental health each accounted for <5% of total output. Intervention studies were a minority; with less aimed at the regional/national or policy or legal and fiscal levels. Conclusion: Health promotion research production varies across Europe. Research commissioning should stress interventional and policy level research.
Keywords: bibliometrics, Europe, health promotion, literature review, public health research
This overview of health promotion literature in Europe for the 10 years from mid-1995 to mid-2005 was undertaken as part of the collaborative study SPHERE (Strengthening Public Health Research in Europe).1 The aim was to produce a description of European health promotion literature and to compare published research outputs by country in order to provide advice on future research policy.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionConclusionsAcknowledgementsReferences |
|---|
Filter (search strategy) design
We used the Ottawa charter2 for our definition of health promotion. It is the most commonly used definition for health promotion, and has a broad and respected approach. After discussion with SPHERE partners, a decision was taken to include disease prevention (i.e. screening and immunization) as well as the more traditional health improvement focus of health promotion, but to identify these activities separately within the results.
The search strategy, or ‘filter’, was adapted from previous similar literature reviews3,4 and used MeSH thesaurus headings to capture papers with health promotion terms such as ‘health-education’ and ‘primary prevention’. Where MeSH terms were not available, free text terms were searched for within the title and abstracts of the articles. We interrogated Medline and Embase for health promotion research with address of author from within the EEA [Our European countries included the 25 EU Member States at the time of the study in 2006, and Iceland, Norway and Switzerland. (Since this study was completed, the EU has expanded to include Bulgaria and Romania)] between 1 July 1995 and 30 June 2005. Searches were not restricted by language, but reviews were excluded, as also were papers where there was no abstract.
Table 1 describes steps in the search. Each row indicates the filter commands and the number of papers retrieved by the command at each stage. These numbers reduce in size as the search is limited by date, country of corresponding author, publication type and research pertaining to humans only.
|
Analysis framework
Screening and immunization (disease prevention) papers were separately identified. Health improvement papers were coded according to apparent methodology used intervention, observational or methodological studies.5 (Box 1). Topic areas were assigned based on known modifiable risk factors for diseases or health risks which contributed to >1% of the DALYs (disability-adjusted life years) in estimates of the Global Burden of Disease (2002 data) for the European Area A6,7 (chosen as the nearest proxy for countries within the EEA). After initial pilot categorizations of the filter output, two further disease-based topics, coronary heart disease (CHD) prevention and cancer prevention were added. Intervention studies were then further categorized for level of intervention as: Individual or Family; Community or Group; Regional or National; and Policy or Legal and Fiscal. Piloting demonstrated that interventions could pertain to more than one topic area or intervention level and papers were coded multiply for these categories where appropriate. Table 2 shows the categorization framework.
|
| Box 1 Definitions of study types: intervention, observation, methodology An intervention study was defined as a study in which conditions are under the control of the investigator, i.e. a study in which the investigator allocates people, (individuals or communities or populations) to an intervention (which receives a health promotion activity or programme e.g. an exercise programme) or to a comparison group which either receives a different activity or programme, or acts as a control. The investigator then measures and compares the outcomes in the two groups. (Obviously, a randomized controlled trial is a specific form of intervention study). An observational study was defined as a study in which the investigator is an observer of what happens when populations or communities vary ‘naturally’, for example in the extent of exercise undertaken. Similarly, the investigator measures and compares the outcomes between groups. (Cross-sectional, cohort and case-control studies are examples of observational studies.) A methodological study was defined as a study in which the main focus was the methods or process of the health promotion research undertaken e.g. methods of improving randomization for intervention studies or methods of adjusting for differences in baseline characteristics for observational studies.5
|
Exclusion criteria
In order to improve inter-marker reliability, two researchers initially assessed the titles of random samples of 250 titles drawn from the filter. Discussion of the relevance of each to health promotion research led to the following criteria for excluding articles from the initial output:
- Tertiary prevention and rehabilitation activities or programmes e.g. cardiac or stroke or pulmonary rehabilitation;
- studies relating to drugs, or surgery for prophylaxis or clinical treatment or adherence or concordance with clinical or drug treatment—unless clear health promotion content (e.g. smoking cessation pharmaceuticals in the context of a smoking cessation programme or drugs for obesity control/prevention in a similar context);
- welfare, benefits or individual financial interventions;
- studies describing disease incidence/prevalence with no link to health promotion or identifiable risk factors; and
- case reports or side effects of treatments or plans for clinical care, related to an individual.
Assessment of a sample
A random sample of 32% (6000 titles) was drawn from the initial output of the filter and was assessed by two researchers (A.C. and M.T.). Each researcher assessed three sets of 1100 papers (900 unique papers with 200 marked jointly for assessment of agreement), first to identify whether publications were ‘health promotion’ publications and second to allocate them to the sub-headings and topics of the analysis framework. Both titles and abstracts were assessed. After each marking round, the 200 jointly coded papers were analysed and then discussed by the reviewers. Kappa scores were calculated and agreement reached between scorers. The final kappa score for level of agreement between assessors was 0.65 (strength of agreement considered good).
The findings from the sample of papers were extrapolated to the total output of the Medline/Embase search and 95% confidence intervals calculated. Estimates of numbers of health promotion research publications were made for each European country (using country of first authorship), and by population in millions. Denominator data for populations and GDP were derived from WHO routine data.8 Estimated publication production over the 10 years was compared for each country by GDP (billions of US dollars), and health need, as represented by each country's total burden of disease (measured in DALYs x 1000).9 The time point used for all data was mid point of the year 2000.
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionConclusionsAcknowledgementsReferences |
|---|
Overall, the Medline/Embase search yielded a total of 18 862 unique publications from the two databases (table 1). In the random sample of 6000, 2206 (36.7%) articles were identified as health promotion research publications. Thus, an estimated total of 6935 (95% CI 6651–7230) health promotion research publications was produced in the 10 years from mid-1995 to mid-2005, with a steady rise in estimated numbers per year from around 400 in 1995 to around 1400 in 2005).
Using these extrapolated figures, table 3 shows the total number of publications over the 10-year period by EEA country. The range of research publications produced by country was estimated between 3 and 1911. The UK was the largest net producer, with just over a quarter (28%) of all the publications [1911, (95%CI 1827–1999)] and nearly four times more than the next set of countries, Sweden, Netherlands and Germany, which produced between 500 and 600 publications each. Thirteen of the 28 countries including Austria, Portugal and Iceland were estimated to produce <100 publications each over the 10-year period, whilst Lithuania, Latvia and Iceland each produced <20 publications.
|
Population size, GDP and burden of disease
There was a weak relationship between estimated publication output and population in millions (r2 = 0.38). The UK had a higher than expected publication output against population size compared with other EEA countries, as also did Sweden, the Netherlands, Finland, Denmark, Norway and Switzerland. Conversely, Germany, France, Italy, Spain and Poland appear to have lower outputs than would be expected.
A slightly stronger relationship was found between estimated health promotion research output and GDP (r2 = 0.45) (figure 1) in contrast to other bibliometric findings of a strong relationship between research output and GDP.10
|
There appears to be a very weak inverse relationship between the production of health promotion research in relation to relative health need as assessed using DALYs per million population (r2 = 0.07). Sweden and Finland were high producers compared with health need in contrast with the relatively low output of eastern European countries such as Estonia, Latvia, Lithuania and Hungary which have greater health needs. France, Germany and Spain were all relatively low producers of health promotion research in relation to health need.
Topic area, methodology and intervention level
2206 papers from the sample of 6000 were considered to be health promotion research papers. Just over a quarter of these (626, 28.4%) was identified as disease prevention (relating to screening or immunization). Most of these were interventional and most were aimed at the regional/national level (Table 2).
For the remaining 1580 publications on health improvement, research on diet and physical activity was the commonest topic area, considered by 385 (24.4%). There were 217 (13.7%) publications concerned with research on tobacco and 132 (8.5%) with cardiovascular disease reduction. Research into interventions for accidents and violence, alcohol use or promotion of mental health all accounted for <5% of the total output. A relatively high proportion (684, 43.3%) of the health improvement studies fell into the ‘other’ category, many of which were concerned with diseases contributing <1% of the burden of disease in the European area. Examples of research areas in this ‘other’ category include sexual and dental health promotion.
Of the 1580 health improvement papers, 941 (60%) were assessed as observational research, 566 (35.8%) as interventional research and only a small proportion (64, 4%) as methodological research.
Most of the 566 studies classified as intervention studies were aimed at individuals and families (186, 32.9%) and at local communities and groups (175, 30.9%). A further third was aimed at the regional/national level (191, 33.7%). Research at the policy or legal and fiscal level was the least frequent with only 23 (4.1%) publications.
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionConclusionsAcknowledgementsReferences |
|---|
Strengths and weaknesses of methods
We were aiming to assess the quantity, type and distribution of published health promotion research in Europe. A basic assumption of the methods we used is that health promotion research is adequately reflected in the standard electronic databases we used—MedLine and Embase. We found the definition of health promotion difficult to translate into search strategies for health promotion research because of its breadth, the difficulty in separately identifying disease prevention and health promotion and of determining borders with other related areas such as occupational health. Our initial yield of health promotion research papers published in Europe during the 10-year period 1995–2005 was 18 862. However, after excluding papers which did not fit our inclusion criteria such as those which were not health promotion or were not research, we estimated that 6395 (CI 6651–7230) health promotion research papers were published in the 10-year period.
We found a substantial variation in publication of health promotion research by country as measured using these standard bibliometric databases. Although we were careful not to exclude any publications on the basis of language of publication, research in languages other than English may be underrepresented both in MedLine and Embase. This may partly explain the finding of an extremely low output of public health research publications from some smaller countries. A further problem may relate to the very small numbers of health promotion publications produced by some countries, in that random variation may have some effect on exact rankings.
Our analysis framework, particularly the section on topic area, was developed using the lists of diseases and conditions contributing >1% of DALYS in Europe. However, using this criterion for inclusion on the topic list, it was noticeable that a relatively high proportion of the health promotion research we identified fell into the ‘other’ category with certain prominent areas of research e.g. sexual and dental health promotion, were not included.
With the bibliometric overview method we used, we were not able to look in detail at the content and quality of the research publications we identified. We decided to exclude reviews, since much published review material will be discursive and would not fall within our definition of ‘research’. However, this exclusion will also have had the effect of excluding systematic reviews.
Geographic and linguistic differences
The Nordic countries are high producers of health promotion research on almost every measure. Many European countries are surprisingly low producers of health promotion research given their population size and GDP (e.g. Germany, Spain, France, Greece and Italy). Eastern European countries are mostly noticeably low in production of health promotion research, especially given their levels of health need measured using DALYS.
Differences by level of intervention/topic area
In this report, we chose to identify research publications dealing with the two major disease prevention activities of screening and immunization separately. Once these were excluded, the majority of research was observational, with a smaller proportion of interventional research and only a tiny proportion of methodological research. The majority of the interventional research was targeted at the individual/family and community levels as opposed to at the national/regional and policy/legal/fiscal levels. In a review of public health intervention research for the government in England, Millward et al.3 suggested that targeting interventions at the latter two leveIs is of greater benefit. Nevertheless, implementation of national policies inevitably also involves action through local (and indeed individual) behavioural and cultural change.
|
Conclusions |
|---|
|
TopAbstractMethodsResultsDiscussionConclusionsAcknowledgementsReferences |
|---|
From our overview, it appears that there are variations in production of health promotion research between European countries. Research production does not appear to relate to health need, either at the country level as measured using DALYs per million population within countries, or in relation to specific topics important to the health of European populations such as alcohol, accidents and violence and mental health. Judgements need to be made by researchers and research funders as to the appropriate balance between observational, interventional and methodological research, and to topic areas and levels of intervention. Evidence-based research commissioning would be likely to favour a greater stress on interventional research for a wider range of topics than is currently favoured, and to include more emphasis on research at higher (national and policy) levels than currently occurs.
We recommend further investigation into research commissioning policy for health promotion research at the European level and by each country in relation to health need, in particular taking into account evidence on the most appropriate level for health promotion intervention and exact topic areas where targeting of research would yield most benefit.
|
Acknowledgements |
|---|
|
TopAbstractMethodsResultsDiscussionConclusionsAcknowledgementsReferences |
|---|
SPHERE (Strengthening Public Health Research in Europe), coordinator Professor Mark McCarthy, was funded by the European Commission Sixth Framework Research programme during the period 2005–07. Data from the present study were presented, in part, at the 14th European Conference on Public Health (EUPHA) Montreux, Switzerland, 16–18 November 2006.
Conflict of interest: None declared.
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionConclusionsAcknowledgementsReferences |
|---|
1 McCarthy M, Clarke A. Editorial. Eur J Pub Health (2007) 17(Suppl. 1). 1–8.
2 World Health Organisation. Ottawa Charter for Health Promotion, First International Conference on Health Promotion, Ottawa,1986. WHO/HPR/HEP/95.1. Available at: http://www.who.int/hpr/NPH/docs/ottawa_charter_hp.pdf.
3 Millward LM, Kelly MP, Nutbeam D. Public Health Intervention Research: the Evidence (2003) London: Health Development Agency.
4 Eppi Centre Databases of promoting health effectiveness reviews (DoPHER). Available at: http://www.eppi.ioe.ac.uk//webdatabases/Intro.aspx?ID=2.
5 World Health Organisation (WHO)—WHO sites. Health statistics and health information systems. In: About the Global Burden of Disease Project. Available at: http://www.who.int/healthinfo/bodabout/en/index.html.
6 Murray CJL, Lopez AD. The Global Burden of Disease (1996) Cambridge: Harvard University Press.
7 Murray CJL, Salomon JA, Mathers CD, Lopez AD, editors. Summary measures of population health: concepts, ethics, measurement and applications. WHO, Geneva. ISBN 92 4 154551 8 (NLM classification: WA 950), 2002. Available at: http://whqlibdoc.who.int/publications/2002/9241545518.pdf.
8 Murray CJL, Salomon JA, Mathers CD, Lopez AD. Summary measures of population health: conclusions and recommendations. In. In: Summary measures of population health: concepts, ethics, measurement and applications—Murray CJL, Salomon JA, Mathers CD, Lopez AD, eds. Geneva: WHO. ISBN 92 4 154551 8 (NLM classification: WA 950), 2002. Available at: http://whqlibdoc.who.int/publications/2002/9241545518.pdf.
9 World Health Organisation. (1986) Ottawa Charter for Health Promotion, First International Conference on Health Promotion, Ottawa. WHO/HPR/HEP/95.1http://www.who.int/hpr/NPH/docs/ottawa_charter_hp.pdf.
10 Rahman M, Mahbubur FT. Biomedical publication - Global Profile and Trend. Public Health (2003) 117:40274–280.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  L. Fond-Harmant Le developpement de la recherche-action europeenne en promotion de la sante mentale : Pourquoi ? Comment ? Promotion & Education, March 1, 2008; 15(1_suppl): 64 - 70. [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||