The Diabetes Population Risk Tool (DPoRT) calculates the future risk of diabetes, for diabetes-free individuals, without the requirement of clinical information. The tool was specifically designed to allow users to estimate future population risk using population survey data.
The tool is based on the publicly available national population health surveys administered by Statistics Canada (Canadian Community Health Survey). The cross-sectional nature of these surveys allows DPoRT to be applied to the newer releases—providing estimates that more accurately reflect Canada's current population.
DPoRT is a population-based risk prediction tool, designed to estimate diabetes incidence and quantify the effect of interventions using routinely collected survey data. Full details on the development and validation of DPoRT can be found in: A population-based risk algorithm for the development of diabetes: development and validation of the Diabetes Population Risk Tool (DPoRT).
Link to the tool
Uses of the Tool include:
- Projecting the number of new cases of diabetes.
- Describing the contribution of a specific risk factor, such as obesity, to population risk of diabetes.
- Assessing the potential benefit of different preventive strategies for reducing the future diabetes cases.
The tool calculates the future likelihood of developing diabetes, in a specified time period, given the current level of risk. For each individual in the population, their characteristics are input into the predictive algorithm. This calculation is repeated for all individuals in the population and summed overall or by important sub-populations (e.g., health units, risk group, etc.) to generate the number of expected diabetes cases. The tool includes time in the equation, allowing users to predict diabetes probability for a range of 1 to 10 years.
An example of the application of DPoRT can be seen in an ICES report by Manuel et al. (How many Canadians will be diagnosed with diabetes between 2007 and 2017? Assessing population risk). The baseline risk of developing diabetes in the Canadian population was estimated using information about diabetes risk factors collected in the 2007 Canadian Community Health Survey. The algorithm was applied to all respondents of the 2007 CCHS and the resulting average baseline risk was then used to estimate how many people in Canada will be newly diagnosed with diabetes in the following 10 years. 1.9 million Canadians were predicted to develop physician-diagnosed diabetes between 2007 and 2017.
In addition to the projection of new cases, population health planners can use DPoRT to describe the contribution of specific risk factors (included in the algorithm) to population risk or to evaluate different prevention strategies. For the former application, a risk exposure (such as obesity) is theoretically reduced by a small amount in the entire population. Baseline risk is calculated before and after the reduction in risk factor exposure—the difference in the two estimates represents the potential reduction in population risk. An alternative approach, to assess the contribution of a risk factor to population risk, is to multiply the baseline risk by the relative risk of exposure. This method is similar to the method for evaluating the potential benefit of prevention strategies. In population health planning, the population health benefit of a diabetes prevention strategy can be assessed by multiplying the target population size by the average population baseline risk and the relative benefit of the intervention or policy. This can be adjusted to account for factors that reduce intervention coverage (i.e., access, level of awareness, compliance, adherence, etc.).
The algorithm can be used by health planners to estimate diabetes incidence, to stratify the population risk, and quantify the effect of interventions using routinely collected survey data.
DPoRT is a valid and useful tool for estimated baseline risk in a population setting. Several interpretive cautions should be considered when reporting results from the tool:
- DPoRT estimates physician-diagnosed diabetes, as opposed to true diabetes status (diagnosed plus undiagnosed)—meaning, the prediction may not detect those whose diabetes is not yet identified. The results from DPORT, therefore, are an underestimate of diabetes cases in the population. Furthermore, the estimates may be biased: those with undetected diabetes could be patients with less severe disease and/or poorer access to medical care. Any report should emphasize that the predicted numbers are physician-diagnosed cases.
- DPoRT was developed in Canada and is most appropriate in the Canadian setting. However, like other risk tools, it may be transportable once validated and calibrated in other populations. The simplicity of the model and the fact that it was validated in two different external populations make its generalizability a genuine possibility. When applying this model outside of Canada, we recommend that it be validated to ensure accuracy.
- DPoRT has been validated to have accurate discrimination in overall population data (i.e., by province). It has been further validated in some subgroups—including ethnicity (see: Rosella et al., in press); however, it has not been validated for individual use or in small sub-populations. While we feel that overall validity would be maintained in sub-groups of the Canadian population, the validity in individuals or small subpopulations cannot be guaranteed.
- We recommend predicting cases for a maximum of ten years into the future— validity beyond this timeframe has not been assessed by the investigators.
- Given changes in testing and detection of diabetes that may occur over time, the algorithm will likely be updated and/or re-calibrated to maintain validity. Any updates to the algorithm should be incorporated.
For additional interpretive cautions and guidelines see: How many Canadians will be diagnosed with diabetes between 2007 and 2017? Assessing population risk.
The study was funded by the Canadian Institutes of Health Research and the Population Health Improvement Research Network.
Limitation of Liability
In no event shall the Canadian Institutes of Health Research or its partners, directors, employees, agents, or licensors be liable for damages of any kind arising from the use of information in the tool.
Disclaimer of Warranties
Information and content are provided "as is". By using the Diabetes Population Risk Tool (DPoRT), the users acknowledge and agree that they will use DPoRT at their own risk and liability.
1. Rosella LC, Manuel D, Burchill C, Stukel TA. A population-based risk algorithm for the development of diabetes: development and validation of the Diabetes Population Risk Tool (DPoRT). J Epidemiology Community Health 2011;65(11):613-20.
2. Manuel DG, Rosella LC, Tuna M, Bennett C. How many Canadians will be diagnosed with diabetes between 2007 and 2017? Assessing population risk. ICES Investigative Report. Toronto: Institute for Clinical Evaluative Sciences; 2010.
3. Rosella LC, Corey P, Stukel TA, Mustard C, Hux J, Roos L, Manuel DG. The role of ethnicity in predicting diabetes risk at the population level. Ethnicity and Health. In press.