Canada-U.S. Mortality Comparison Highlights Analytical Pitfalls

A highly publicized advocacy group report contending that some Canadian cancer patients die at rates much higher than those in the United States has raised as many questions about the availability, comparability, analysis, and reporting of cancer statistical information in Canada as it has about the country’s quality of care. Citing barriers to care that include long treatment waiting lists and inadequate numbers of specialists, the Toronto-based Cancer Advocacy Coalition of Canada’s analysis maintains that Canadian cancer patients—with the exception of those in Alberta and British Columbia—fare far worse than their American counterparts. Using data obtained from the Springfield, Ill.–based North American Association of Central Cancer Registries, the group’s “Report Card 2000” found that certain regional comparisons between the United States and Canada revealed worse outcomes for breast, prostate, colon, and other cancers. In addition, the group noted that in some provinces Canadians can wait up to 10 weeks longer than they would in the United States for radiation therapy. But despite the widespread national attention the report received in Canada, researchers there are cautioning that the report more clearly demonstrates the hazards of sophisticated data analysis than a badly flawed system of cancer care. Although the analysis raises important questions about cancer mortality rates, the lack of data standardization and comparability between Canada and the United States reveals the extent to which data can sometimes lead to inaccurate conclusions, said Don Carlow, M.D., the chief executive officer of the British Columbia–based Canadian Association of Cancer Registries. Carlow said that “the public should have knowledge of and access to reliable sources of information” so that public conclusions are authoritative and strongly evidence-based rather than based on what he called “ad hoc” cancer statistics. But on the flip side, Carlow said, as more statistical databases open their Internet portals to the public, the widespread availability of increasingly complex data provides opportunities for misinterpretation that could have disastrous results. Mary McBride, an epidemiologist and director of the British Columbia Cancer Registry in Vancouver, added, “We have easier access to information than we did more than a decade ago, but we also have the problem of understanding what is good quality data and what is most useful and relevant to what we need to know.” In addition, with the news media acting as a “go-between” for consumers and professionals, there is “heightened responsibility” among reporters and researchers to accurately and adequately convey the implications, value, and meaning of cancer data, McBride said. “Because there is so much information out there it is difficult for end users to sort through and speak to the value of having better assessment and interpretation,” she added. In fact, survey results released this month by the Vancouver-based market research company Ipsos-Reid revealed that while 70% of Canadian physicians report that their patients commonly cull health information from the Internet, 91% of the doctors said they are at least “somewhat suspicious” of the quality and accuracy of what their patients find. In addition, almost a quarter said that patient use of Internet health information has generally caused more problems that it has solved. Because Canada’s universal health care system is administered by provincial authorities, there is better integration of screening, diagnosis, treatment, and outcomes information and thus more patient-linked comprehensive data available nationally than in the United States, McBride said. Canada is also the largest contiguous geographic area completely covered by mandatory universal population-based monitoring of cancer with some provincial cancer registries dating to the 1930s. Not only is the variation in cancer incidence and mortality rates across North America generally small, but the rates can be complicated by differences in age, risk factors, socioeconomic status, and ethnic or racial patterns as well as an overall declining cancer mortality rate, McBride said. Perhaps even more importantly, the CACC analysis suggests that the growing acceptance of advocacy groups by the research community over the last decade may have come full circle—so that now researchers must likewise have a place at the advocacy table. “This analysis has upset quite a number of people across the country, and I think the criticisms are unfair. The cancer control system has an obligation and a responsibility to explain these things and not leave it to advocacy groups alone to try to interpret this data,” said Carlow. “We should help everybody understand what the truth is.” Carlow said administrators of national cancer databases should move more quickly toward internationally accepted standards of data collection, guaranteed public funding that will maintain long-term operation and data quality, and better integration of all cancer-related activities that allow for more centralized and comprehensive databases. To that end, the CACC report has made existing Canadian cancer data initiatives more urgent and also prompted the launching of new ones. For example, Carlow said that in January his Canadian registry group and a coalition of other cancer interests will be investigating ways to make Canadian and foreign data more comparable. In the meantime, the recommendations of a 1998 national initiative to improve cancer survival information, which brought together cancer database authorities from across Canada as well as the United States, Italy, and Finland, will be reviewed. “The cancer rankings were used as a way of indicating where provinces and states are in terms of actual mortality,” said Cancer Advocacy Coalition of Canada founder Pat Kelly. “The issues we are trying to raise are that we should have good benchmarks in place to not only measure how we are doing but how well we should be doing.” Kelly said the analysis represents a “quick and dirty” description of cancer mortality rates in the United States and Canada that is intended to raise questions about variations in treatment, the lack of data on cancer survival, and the impact of cancer risk reduction initiatives. In addition, the group had to raise funds to conduct the analysis as well as convince database administrators and cancer agencies not only to provide data but that the data would be used “ethically.” “We were accused of taking an overly simplistic and alarmist position,” Kelly said. “I’m not saying we did an outstanding job by any means given the barriers we had to overcome. “We’re in a time of transition here and if we’re going to have accountability we’re going to have to share information. Patients want to regain a sense of control over treatment decisions and are much less deferential to authority. We’ve reached a new phase in patient advocacy where we’re not merely cheerleaders,” she added. Mary McBride Although cancer incidence rates in Canada are similar to those in other countries, a recent report suggests that mortality rates may be much higher in Canada. The above incidence rates are per 100,000 and are standardized to the world standard population. (Courtesy of the National Cancer Institute of Canada.)

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.