Discussion: Tuberculosis in Canada 2012

Discussion

Overall, Canada has one of the lowest TB disease rates in the world, and drug resistance in Canada remains consistently below international levels. Available surveillance data highlight the unique aspects of TB in Canada, including the disproportionate effect on Aboriginal people and immigrants to Canada from areas of the world with high rates of TB disease.

The annual number of cases reported in the Canadian-born Aboriginal population and in the foreign-born has remained relatively unchanged from 2002 to 2012, despite the reduction over time in the number of reported cases of TB in the Canadian-born non-Aboriginal population.

In 2012, among TB cases with known HIV status (positive or negative), 8% were HIV-positive. If one assumes that all HIV co-infected TB cases were reported, then the overall co-infection rate would have been approximately 4%. However, this figure should be interpreted with caution given the high percentage of cases for which HIV status was not reported (60%). The prevalence of HIV among TB cases in Canada thus remains unclear.

The prevalence of drug resistance, especially MDR-TB and XDR-TB, remains low in Canada, indicating that drug resistance has not yet emerged as a significant problem in Canada.  However, with growing global concern over the increasing rates in some parts of the world, continued monitoring of drug resistance in Canada is imperative.

Strengths and Limitations

The CTBRS is a long-standing surveillance system and is the result of successful collaboration between federal, provincial/territorial, and local governments. The system is comprehensive in that it collects detailed case information (both demographic and clinical), including drug-sensitivity results. Yet it remains flexible and information needs are reviewed regularly in consultation with provincial and territorial partners so that the system can be adapted and improved upon in response to changes in the epidemiology of TB in Canada.

The CTBRS is a passive surveillance system that relies on data retrospectively gathered from medical and laboratory records, as opposed to active surveillance involving prospective actions aimed at identifying cases. As a result, coverage of the system (i.e. whether all people with TB disease are being identified) is difficult to ascertain without the use of special evaluation studies. However, the WHO estimates that Canada’s case detection rate is between 87% and 100%.^{Footnote 4} The WHO also notes that “TB incidence has never been measured at national level because this would require long term studies among large cohorts of people (hundreds of thousands) at high cost and with challenging logistics. Notifications of TB cases provide a good proxy indication of TB incidence in countries that have both high-performance surveillance systems (for example, there is little underreporting of diagnosed cases) and where the quality of and access to health care means that few cases (or a negligible number) are not diagnosed” (p.8).^{Footnote 4} Under the reasonable assumption that the CTBRS is a high-performing surveillance system (i.e. there is little underreporting of diagnosed cases, and a negligible number of people with active TB disease are undiagnosed), notifications of TB cases in the Canadian context provide a good proxy indication of TB disease incidence.

In general, the majority of data elements for case and outcome reports submitted to the Agency are complete. For the majority of demographic and clinic data, over 95% of data elements are complete. Reporting is less complete for some of the data elements introduced in 1997, such as HIV status, however response rates have improved over time. In 1997, when the data element “HIV status at the time of diagnosis” was added to the reporting form, it was reported only for 6% of all TB cases; by 2012, reporting of this data element had increased to 40%.

The reporting of information on final treatment outcomes was added to the CTBRS in 1997.  Initially not all provinces and territories were able to report outcomes for each case, but rather only reported outcomes at an aggregate level. Case-level reporting of final treatment outcomes has increased over time and, in 2012, outcome data was reported at the case level for approximately 90% of cases reported in 2011.

The methods used to collect and analyze the data in this report have been designed to minimize error. However, surveillance data are subject to certain types of errors (e.g. measurement and processing error, completeness).

Before analysis and report preparation, all data are reviewed for errors, inconsistencies, and completeness. Follow-up is done with the reporting jurisdictions to resolve concerns or issues with the reported data. Data submitted in previous years are subject to revision in the event of late reporting or when revised information from the provinces or territories is received.  Revisions to previously reported data are reflected in subsequent surveillance reports.

The surveillance data in this report may be subject to coding, reporting, and processing errors.  As an example, not all provinces and territories use ICD 9 or ICD 10 coding systems for recording disease diagnoses, a means that the CTBRS requires to classify patients according to the main diagnostic site. Efforts are made to align alternate coding systems used by these provinces and territories to ensure that diagnostic reporting is as accurate as possible.

The accuracy of the data is partially a function of timely reporting and updates to the Agency from the provinces and territories. Some degree of lag does occur (i.e. reporting delay), but typically only affects the preliminary data published in the pre-release reports.

In some cases, variables on the reporting form are left blank or marked as “unknown”. Where a response to a question or field on the reporting form is left blank, efforts are made to obtain a valid response from the reporting jurisdiction. When no additional information is available, the data are reported as “unknown” in the data table. Therefore, values recorded as “unknown” include truly unknown information as well as unreported data.