We are what we do: research outputs of public health

If public health research does not progress beyond descriptive research to testing the effectiveness of interventions that can provide causal data, its contribution to evidence-based practice will not be realized. This paper examined the number and percentage of public health research publications over three time periods (1987-1988, 1997-1998, 2005-2006) for three preventive health issues: tobacco use, alcohol use, and inadequate physical activity. A computer-based literature search was conducted, using the Medline and PsycINFO databases. A random sample of 1000 abstracts for each preventive health issue was examined per time period. The abstracts were first categorized as public health or not, and then as data-based or non-data-based public health research. Data-based publications were classified according to research type as measurement-, descriptive/epidemiologic-, or intervention-oriented. Data analysis occurred in 2007. The number of data-based public health research publications examined for each topic ranged between 180 and 346 in 1987-1988, 199 and 322 in 1997-1998, and 302 and 364 in 2005-2006. Despite a significant increase over time in the percentage of intervention publications for tobacco (from 10% in 1987-1988 to 18% in 2005-2006, p=0.008), the majority of publications in all three time periods were descriptive/epidemiologic for all topics (62%-87% in 1987-1988, 64%-85% in 1997-1998, 78%-79% in 2005-2006). There were fewer measurement (3%-7% in 1987-1988, 2%-6% in 1997-1998, 4%-10% in 2005-2006) or intervention publications (9%-31% in 1987-1988, 10%-30% in 1997-1998, 12%-18% in 2005-2006). Descriptive research does not provide optimal evidence for how to reduce preventable illness. Reasons for the lack of measurement and intervention public health research are explored.