Dataset Information

A method for managing re-identification risk from small geographic areas in Canada.

ABSTRACT:

Background

A common disclosure control practice for health datasets is to identify small geographic areas and either suppress records from these small areas or aggregate them into larger ones. A recent study provided a method for deciding when an area is too small based on the uniqueness criterion. The uniqueness criterion stipulates that an the area is no longer too small when the proportion of unique individuals on the relevant variables (the quasi-identifiers) approaches zero. However, using a uniqueness value of zero is quite a stringent threshold, and is only suitable when the risks from data disclosure are quite high. Other uniqueness thresholds that have been proposed for health data are 5% and 20%.

Methods

We estimated uniqueness for urban Forward Sortation Areas (FSAs) by using the 2001 long form Canadian census data representing 20% of the population. We then constructed two logistic regression models to predict when the uniqueness is greater than the 5% and 20% thresholds, and validated their predictive accuracy using 10-fold cross-validation. Predictor variables included the population size of the FSA and the maximum number of possible values on the quasi-identifiers (the number of equivalence classes).

Results

All model parameters were significant and the models had very high prediction accuracy, with specificity above 0.9, and sensitivity at 0.87 and 0.74 for the 5% and 20% threshold models respectively. The application of the models was illustrated with an analysis of the Ontario newborn registry and an emergency department dataset. At the higher thresholds considerably fewer records compared to the 0% threshold would be considered to be in small areas and therefore undergo disclosure control actions. We have also included concrete guidance for data custodians in deciding which one of the three uniqueness thresholds to use (0%, 5%, 20%), depending on the mitigating controls that the data recipients have in place, the potential invasion of privacy if the data is disclosed, and the motives and capacity of the data recipient to re-identify the data.

Conclusion

The models we developed can be used to manage the re-identification risk from small geographic areas. Being able to choose among three possible thresholds, a data custodian can adjust the definition of "small geographic area" to the nature of the data and recipient.

SUBMITTER: El Emam K

PROVIDER: S-EPMC2858714 | biostudies-literature | 2010 Apr

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

A method for managing re-identification risk from small geographic areas in Canada.

El Emam Khaled K Brown Ann A AbdelMalik Philip P Neisa Angelica A Walker Mark M Bottomley Jim J Roffey Tyson T

BMC medical informatics and decision making 20100402

<h4>Background</h4>A common disclosure control practice for health datasets is to identify small geographic areas and either suppress records from these small areas or aggregate them into larger ones. A recent study provided a method for deciding when an area is too small based on the uniqueness criterion. The uniqueness criterion stipulates that an the area is no longer too small when the proportion of unique individuals on the relevant variables (the quasi-identifiers) approaches zero. However ...[more]

PMID: 20361870

Similar Datasets

Project description:BackgroundTemporal trends in risk factors for cardiovascular disease and the impact of socio-economic status on these risk factors remain unclear.MethodsUsing data from the National Population Health Survey and the Canadian Community Health Survey, we examined national trends in heart disease, hypertension, diabetes mellitus, obesity and smoking prevalence from 1994 to 2005, adjusting for age and sex. We stratified data by income adequacy category, body mass index and region of residence.ResultsAn estimated 1.29 million Canadians reported having heart disease in 2005, representing increases of 19% for men and 2% for women, relative to 1994. Heart disease increased significantly in the lowest income category (by 27%), in the lower middle income category (by 37%) and in the upper middle income category (by 12%); however, it increased by only 6% in the highest income group. Diabetes increased in all but the highest income group: by 56% in the lowest income group, by 93% in the lower middle income group and by 59% in the upper middle income group. Hypertension increased in all income groups: by 85% in the lowest income group, by 80% in the lower middle income group, by 91% in the upper middle income group and by 117% in the highest income group. Obesity also increased in all income groups: by 20% in the lowest income group, by 25% in the lower middle income group, by 33% in the upper middle income group and by 37% in the highest income group. In addition to socio-economic status, obesity and overweight also modified the trends in risk factors. Diabetes increased to a greater extent among obese participants (61% increase) and overweight participants (25% increase), as did hypertension, which increased by 80% among obese individuals and by 74% among overweight individuals. Trends in diabetes, hypertension and obesity were consistent for all provinces.InterpretationDuring the study period, heart disease, hypertension, diabetes and obesity increased for all or most income groups in Canada. Further interventions supporting modification of lifestyle and risk factors are needed to prevent future cardiovascular disease.

Project description:BackgroundMore than half of all heart disease and stroke are attributable to hypertension, which is associated with approximately 10% of direct medical costs globally. Clinical trial evidence has demonstrated that the benefits of pharmacist intervention, including education, consultation and/or prescribing, can help to reduce blood pressure; a recent Canadian trial found an 18.3 mmHg reduction in systolic blood pressure associated with pharmacist care and prescribing. The objective of this study was to evaluate the economic impact of such an intervention in a Canadian setting.MethodsA Markov cost-effectiveness model was developed to extrapolate potential differences in long-term cardiovascular and renal disease outcomes, using Framingham risk equations and other published risk equations. A range of values for systolic blood pressure reduction was considered (7.6-18.3 mmHg) to reflect the range of potential interventions and available evidence. The model incorporated health outcomes, costs and quality of life to estimate an overall incremental cost-effectiveness ratio. Costs considered included direct medical costs as well as the costs associated with implementing the pharmacist intervention strategy.ResultsFor a systolic blood pressure reduction of 18.3 mmHg, the estimated impact is 0.21 fewer cardiovascular events per person and, discounted at 5% per year, 0.3 additional life-years, 0.4 additional quality-adjusted life-years and $6,364 cost savings over a lifetime. Thus, the intervention is economically dominant, being both more effective and cost-saving relative to usual care.DiscussionAcross a range of one-way and probabilistic sensitivity analyses of key parameters and assumptions, pharmacist intervention remained both effective and cost-saving.ConclusionComprehensive pharmacist care of hypertension, including patient education and prescribing, has the potential to offer both health benefits and cost savings to Canadians and, as such, has important public health implications.

Dataset Information

A method for managing re-identification risk from small geographic areas in Canada.

Background

Methods

Results

Conclusion

Publications

A method for managing re-identification risk from small geographic areas in Canada.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets