Mapping cancer risk in southwestern Ontario with changing census boundaries

Our plain-English summary of Mapping cancer risk in southwestern Ontario with changing census boundaries, by P. Nguyen, P.E. Brown, J. Stafford (Biometrics, 2012).

The study

• Authors: P. Nguyen, P.E. Brown, J. Stafford
• Journal: Biometrics
• Published: 2012;68(4):1228-37
• Cited by: 3 articles, as counted by NIH iCite
• Read it: PubMed · Publisher

Abstract

Mapping disease risk often involves working with data that have been spatially aggregated to census regions or postal regions, either for administrative reasons or confidentiality. When studying rare diseases, data must be collected over a long time period in order to accumulate a meaningful number of cases. These long time periods can result in spatial boundaries of the census regions changing over time, as is the case with the motivating example of exploring the spatial structure of mesothelioma lung cancer risk in Lambton County and Middlesex County of southwestern Ontario, Canada. This article presents a local-EM kernel smoothing algorithm that allows for the combining of data from different spatial maps, being capable of modeling risk for spatially aggregated data with time-varying boundaries. Inference and uncertainty estimates are carried out with parametric bootstrap procedures, and cross-validation is used for bandwidth selection. Results for the lung cancer study are shown and discussed.

© 2012, The International Biometric Society.

In plain English

In this methods paper, the authors built a mapping method (a local-EM kernel smoothing algorithm) for cases located only by postal code, or by coarser areas for nine cases lacking one, when census boundaries change; they tested it on simulated data. They applied it to mesothelioma (called "mesothelioma lung cancer"; pleural or peritoneal not stated). The data were 163 cases in Lambton County and 84 in Middlesex County, Ontario, from the Ontario Cancer Registry, 1985 to 2007. Cases were located by where they lived; no one's asbestos exposure was measured. Lambton's incidence was roughly 5.5 times the rest of Ontario's; Middlesex was near the provincial average. Estimated risk was six to seven times the provincial rate in Sarnia and over seven times in Petrolia. Most of Middlesex County was likely at or near the Ontario average. An exploratory analysis did not detect changes over time in Sarnia or the surrounding area, including Petrolia, relative to the provincial trend. The authors suggested Lambton's high risk could possibly relate to local industrial facilities and occupational exposures. They concluded the method was useful as a quick preliminary analysis.

We wrote this plain-English summary from the full article with the help of AI, and a second AI model checked each sentence against the article. It is not the authors' wording; the abstract and the article are their own account.

What the study did and found

• Design: Statistical methods paper: development of a local-EM kernel smoothing method for mapping disease risk, tested in a simulation study and applied to cancer registry data on mesothelioma (a geographic, area-level analysis)
• Who was studied: Mesothelioma cases from the Ontario Cancer Registry, 1985 to 2007: 163 in Lambton County and 84 in Middlesex County, southwestern Ontario, Canada. Nine of these cases had missing postal codes. Population data came from the five Canadian censuses from 1986 to 2006.
• Exposure: No individual exposure was measured. Most cases were located by residential six-digit postal code. The nine cases with missing postal codes were placed in the coarser areas of the Standard Geographical Classification codes, which are made up of census subdivisions. The authors suggested the high Lambton County risk could possibly be related to local industrial facilities and their occupational exposures. No fiber types are named.
• Main results:
  • Mesothelioma incidence in Lambton County was roughly 5.5 times higher than in the rest of Ontario. Middlesex County was close to the provincial average (average risks 5.5 and 1.1, respectively).
  • Estimated risk in the city of Sarnia was between six and seven times the provincial rate. Risk in the town of Petrolia was more than seven times the Ontario average.
  • Sarnia's risk was well above typical Ontario levels but not necessarily atypical for Lambton County.
  • The authors judged that Petrolia likely had mesothelioma risk more than 10% above the Lambton County average.
  • Most of Middlesex County was judged quite likely to be at or near the Ontario average.
  • An exploratory analysis did not detect changes in risk over time in Sarnia and Petrolia relative to the provincial time trend. It did not address whether absolute incidence changed.
  • In that exploratory analysis, there was no evidence of a time trend in zone 1 (Sarnia) or zone 2 (the area surrounding Sarnia), with p-values greater than 0.2. The authors described this as consistent with risk relative to the rest of Ontario being constant. There was marginal evidence of a trend in zone 3 (east and south of the study region, near the Ontario average), with a p-value of about 0.03.
  • On simulated data, an alternative approach was tried: assigning each case to a single location and applying the Besag, York and Mollié model. It produced much worse ROC curves than the local-EM method.
• The authors' conclusion: The authors concluded that the local-EM method is a useful way to make the most of aggregated data. They described it as quick to apply to data from existing databases, possibly as a preliminary analysis to justify more costly data collection.
• Limitations the authors state:
  • The authors assumed that population is spread evenly within each census region, an assumption they call unavoidable.
  • The authors describe the boundaries of the three zones used for the time-trend analysis as arbitrary.
  • The authors note that a more thorough analysis might need more precise street-address geocoding, residential and occupational histories of cases with a suitable control group, or information on smoking and other individual risk factors.
  • The authors state that the simulation study in Section 3 does not compare the method with alternative methods.

Topics

The National Library of Medicine indexes this study under subjects our site covers: Lung cancer.

Medical terms in this study, explained in our Lexicon: Mesothelioma, Lung Cancer.

Related studies

• Survival analysis under imperfect record linkage using historic census data — A.K. Marks-Anglin, F.K. Barg, et al. (2024)
• Modeling Variables With a Spike at Zero: Examples and Practical Recommendations — E. Lorenz, C. Jenkner, et al. (2017)
• Estimating the potential disease rate reduction under varying conditions of combined effects — H. Checkoway, J.L. Hickey (1988)
• Analysis of the effect of occupational exposure to asbestos based on threshold regression modeling of case-control data — A. Chambaz, D. Choudat, et al. (2014)
• Predictive accuracy of the Liverpool Lung Project risk model for stratifying patients for computed tomography screening for lung cancer: a case-control and cohort validation study — O.Y. Raji, S.W. Duffy, et al. (2012)
• Development and validation of a lung cancer risk prediction model for African-Americans — C.J. Etzel, S. Kachroo, et al. (2008)

Browse return to medical research.