Directions for Future Research
Health research has at times focused on topics that, while scientifically intriguing, have not always taken population health consequences into account when shaping specific research questions (Editorial 2001; Gross, Anserson, and Powe 1999; Horton 2003). The collation of evidence on exposure and hazard for different risks and the existing data gaps revealed the areas where data and monitoring need to be improved for better quantification of important risks and for more effective intervention. This includes the need for more detailed and higher quality data on exposure to most risks using exposure variables that capture the full distribution of hazards in the population. Important examples include detailed data on alcohol consumption volumes and patterns, dietary and biological markers for micronutrients, physical activity, and indoor smoke from household use of solid fuels, all of which were quantified using indirect measures with limited resolution. Furthermore, assumptions and extrapolations were needed in quantifying the relationships between risk factors and disease because of research gaps for some important global risk factors. Examples include limited quantitative assessments of the hazards of specific sexual behaviors for HIV/AIDS and other sexually transmitted diseases (UNAIDS 2001), of alcohol drinking patterns (Puddey and others 1999), or of exposure to indoor smoke from household use of solid fuels (Ezzati and Kammen 2002). Equally important are detailed exposure data for risks that have traditionally been studied in developed countries, but have global importance and require more detailed data and hazard quantification in developing regions, including smoking, body mass index, blood pressure, and cholesterol (Yusuf and others 2004).
The limited evidence on the effects of multiple risk factors and risk factor interactions also points to important gaps in research on multirisk and stratified hazards. Including multiple layers of causality in epidemiological research and risk assessment would allow investigators to estimate the benefits of reducing combinations of distal and proximal exposures using multiple interventions. Examples of such integrated strategies include using education and economic tools to promote physical activity and a healthier diet coupled with screening and lowering cholesterol, and addressing the overall childhood nutrition and physical environment instead of focusing on individual components. In such research, risk factor groups should be selected based on both biological relationships and socioeconomic factors that affect multiple diseases. Examples include those risk factors that are affected by the same policies and distal socioeconomic factors, such as malnutrition; unsafe water, sanitation, and hygiene; indoor smoke from household use of solid fuels; and rural development policies, or affect the same group of diseases, for instance, the previous example for childhood infectious diseases and smoking, diet, physical activity, and blood pressure for vascular diseases. Once risk factors are selected, the emphasis on reducing confounding should be matched by equally important inquiry into independent and mediated hazard sizes that are stratified based on the levels of other risks.
Finally, to inform interventions and policies, similar analyses should take place at smaller scales than global or regional levels, for example, rural and urban areas or different geographical regions of individual countries, and should include micro-level data and possibly a more comprehensive list of both distal and proximal risk factors, such as adverse life events and stress, risk factors for injuries, salt and fat intake, and blood glucose.