Health Reports
Tuberculosis-related hospital use among recent immigrants to Canada

View the most recent version.

Archived Content

Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.

Archived

This page has been archived on the Web.

by Edward Ng, Dominique Elien Massenat, George Giovinazzo, David Ponka and Claudia Sanmartin

Globally, tuberculosis (TB) is responsible for the largest number of deaths of any infectious disease, surpassing HIV and malaria combined.^{Note 1} In 2015, there were an estimated 10.4 million new cases of active TB worldwide.^{Note 2} Global targets and milestones for reducing the burden of TB have been established as part of the United Nations Sustainable Development Goals and the World Health Organization (WHO) End TB Strategy.^{Note 3Note 4} Recommendations for countries with low TB incidence, such as Canada, as well as the Government of Canada’s federal framework for action on TB, are directed at decreasing rates of TB in Canada.^{Note 5Note 6Note 7}

There is an overall shift of TB epidemiology in countries with low incidence of TB, in that the local-born population has a lower TB incidence than immigrants.^{Note 8Note 9} This is also the case in Canada, with the notable exception of Indigenous people. Among the 1,737 cases of active TB reported in 2016, the incidence among the non-Indigenous Canadian-born population was 0.6 per 100,000 population (8% of total cases), compared with 15.2 among the foreign-born population (70% of total cases) and 23.5 among Canadian-born Indigenous people (19% of total cases).^{Note 8} TB cases among immigrants originated primarily from countries with relatively high TB incidence rates.^{Note 2Note 8}

TB is monitored in Canada by the Public Health Agency of Canada (PHAC), in collaboration with provincial and territorial public health authorities through the Canadian Tuberculosis Reporting System (CTBRS) and the Canadian Tuberculosis Laboratory Surveillance System (CTLSS). While these systems can be used to report on TB rates by country of birth, reporting by immigrant class (e.g., economic, family, refugee) is not possible.^{Note 10} Information about TB exists in health administrative data such as hospital records from the Canadian Institute for Health Information (CIHI), available at Statistics Canada. Although active TB can usually be treated on an outpatient basis, some patients require in-hospital management. Active TB includes patients with pulmonary and non-pulmonary TB (e.g., meningeal TB). As such, the number of TB-related hospitalizations is not identical to the reporting of active TB done for the CTBRS. Instead, the reported number can indicate the severity of TB disease in Canada and reflect the management of TB cases as part of the continuum of TB-related care.^{Note 11} However, immigration-related variables are not routinely collected in hospital data.^{Note 12} Linking hospital data to immigrant landing records in Canada enables immigrant-related analyses.^{Note 13Note 14}

This study provides new evidence on TB-related hospitalizations among new immigrants to Canada. It uses a unique linked data file that brings together information from immigrant landing records and hospital data for a maximum of 13 years to identify a TB-related hospital event after landing. Specifically, this paper provides a profile of the timing of TB-related acute care hospitalization, starting from the time of landing among immigrants who officially landed in Canada from 2000 to 2013, as well as an estimation of the burden of TB hospital care in Canada incurred by these recent immigrants relative to the total Canadian population.

Methods and data

Data linkage

The Immigrant Landing File (ILF) data were linked to the CIHI Discharge Abstract Database (DAD) using the Social Data Linkage Environment (SDLE) at Statistics Canada. SDLE is a highly secure linkage environment that helps create linked population data files for social analysis by creating a central depository called the Derived Record Depository (DRD).^{Note 15} The DRD is a national dynamic relational database containing only basic personal identifiers. It was created by linking selected Statistics Canada source index files, including tax, birth and death data, to produce a list of unique individuals. The DRD is the core of SDLE, to which all other databases are linked. The linkage was approved by Statistics Canada’s senior management,^{Note 16} and use of the linked data is governed by the Directive on Microdata Linkage.^{Note 17}

The ILF contains administrative information for all individuals who have landed in Canada since 1980, which is provided to Statistics Canada on a monthly basis by Immigration, Refugees and Citizenship Canada (IRCC). The ILF includes information on year of birth, sex, time of entry (month and year), intended province of destination, source country and admission category (e.g., economic, refugee).^{Note 18} Landing records from 1980 to 2013 were eligible for linkage (8,450,469 records for a total of 6,896,592 immigrants). The linkage of the ILF to the DRD was conducted using probabilistic methods based on the following variables: date of birth, names (surnames and given names), postal code, city, sex, marital status, entry date, and dates of birth of the immediately younger and older person in the family. Overall, 85% (n=5,854,949) of immigrant records were linked. This is consistent with previous linkages conducted based on the ILF.^{Note 19} Since the DRD is primarily based on tax data, analysis of linked records versus non-linked records revealed that younger immigrants were underrepresented because they do not file taxes.

The DAD contains demographic, administrative and clinical data for all acute care hospital discharges and some psychiatric, chronic rehabilitation and day-surgery hospital discharges for all provinces and territories, excluding Manitoba before April 2004 and Quebec.^{Note 20} The data are provided to Statistics Canada annually by CIHI for each hospital fiscal year (from April to March of the following year). Hospital discharges occurring from April 1, 1994, to March 31, 2015, were eligible for linkage (n=77,925,269). The linkage was conducted using a deterministic approach using the following variables: date of birth, postal code, sex and health insurance number. Overall, 85% (n=66,246,909) of hospital records were linked to the DRD. Analysis of linked records versus non-linked records revealed similar distributions of individual-level characteristics (e.g., age, sex, year of entry), indicating no evidence of bias.^{Note 21}

Study cohort

The study cohort comprises new immigrants who landed from January 2000 to December 2013 and were linked to the DRD (n=2,730,390): immigrants in the Longitudinal Immigration Database (IMDB), which included only taxfiling immigrants (n=2,200,420), and immigrants who landed in the same period but never filed taxes, and, therefore, were not included in the IMDB (n=529,970). The IMDB is derived from the ILF and is a research database representing unduplicated immigrant records that have also been linked to tax data.^{Note 18} The post-2000 cohort of immigrants was chosen to ensure continuous follow-up after arrival and to minimize the possibility of immigrants who entered the country having already gone back to their home country or moved on to another country. To correct for the underrepresentation of younger immigrants noted above, the study cohort also includes the unique immigrants who landed during the same period from 2000 to 2013 who were not originally included in the IMDB, most of whom were young immigrant non-taxfilers (nearly 50% of immigrants younger than 20 years of age). These non-taxfilers also include some seniors (about 15% of immigrants older than 60 years of age). Adding these non-taxfilers back to the linkage with the IMDB ensures that the analysis is not confined only to taxfilers. For simplicity, the rest of this paper will use the IMDB cohort to describe the study cohort. Approximately 840,000 hospital discharges occurring from April 1, 2001, to March 31, 2014, were linked to an immigrant record.

Variables of interest

Age was derived from the IMDB as the date of birth to 2013. Age at hospitalization among those hospitalized for TB was derived based on the difference between birth year, from the IMDB, and the year of hospitalization, from the DAD. Age was grouped as follows: 0 to 17 years, 18 to 24 years, 25 to 34 years, 35 to 44 years, 45 to 64 years, and 65 years and older. Landing year was defined as the year immigrants became legal permanent residents (this is not necessarily the same as the year of arrival in Canada), as recorded in the IMDB. The categories for period of immigration were used as a proxy for the duration since landing and were categorized as follows: 2000 to 2003, 2004 to 2008, and 2009 to 2013. Because immigrants could be temporary residents in Canada before immigration, individuals in this group could be hospitalized before landing. The time to the first TB hospitalization during the follow-up time (a maximum of 13 years) was classified from the landing month as follows: less than 0 months after landing (i.e., occurred before landing), 0 to 12 months after landing, 13 to 60 months after landing, and 60 or more months after landing. Any hospitalization that occurred in the same month as landing is classified as post-landing in this study.

Immigrants were identified as having come from a country of birth with high risk of TB based on the designated country list for TB from IRCC. The IRCC list includes all countries already on the WHO list of countries with high TB burden,^{Note 22} as well as any other IRCC-specific countries (see Appendix 1). The latter group (IRCC-specific) was identified as any country not in the WHO list but with a TB incidence level of 30 per 100,000 population or more. This level is consistent with mathematical models estimating that this TB incidence level is cost-effective for interventions that aims at limiting the reactivation of post-landing latent TB infection.^{Note 23Note 24} Results among the immigrants from countries with high risk of TB were presented for all designated countries overall and by source: the WHO list, and countries added by IRCC (i.e., IRCC-specific). All remaining countries are categorized as low risk. The 12 countries among recent immigrants with the highest number of TB-related hospital events were also highlighted.

Province of residence was measured in two ways: (1) based on the province of intended residence, as reported in the IMDB; and (2) based on the province submitting the data to the DAD. Since some immigrants may move to another province after landing, the province of intended residence may not be the same as the province of hospitalization. Therefore, the province information submitted for the hospital discharge abstract was used as a proxy for the province of residence. Because of low numbers, selected provinces were grouped as follows: the Atlantic region (Newfoundland and Labrador, Nova Scotia, New Brunswick, and Prince Edward Island) and the Prairies (Manitoba, Saskatchewan and Alberta).

Immigrant category, as reported on the IMDB, is defined as immigrants admitted to Canada under the economic, family or refugee class defined in the 2001 Immigration and Refugee Protection Act.^{Note 25} The immigrant classes examined were the following: (1) economic, principal applicant; (2) economic, spouse or children (including provincial nominees, available as of 1996); (3) family (for family reunion purposes); (4) refugee (for humanitarian purposes); and (5) others, unknown or missing.

Outcomes

The primary outcome is TB-related acute care hospitalizations of immigrants discharged from April 1, 2001, to March 31, 2014, as identified using the International Classification of Diseases (ICD-9 and ICD-10) codes for the most responsible diagnosis (Table 1).^{Note 26Note 27} Hospitalizations among immigrants were derived from the linked IMDB–DAD data; hospitalizations among the remaining population were identified using the full DAD files linked in SDLE, minus records linked to an IMDB record for the same period.

Statistical methods

Descriptive statistics were generated to profile the study cohort. The distributions of immigrants with at least one TB-related hospitalization overall and of immigrants by duration of time from immigration landing to the first hospitalization are presented by selected characteristics. Person-year adjusted incidence of TB per 100,000 population is also calculated to present rates adjusted for years of follow-up. Hospitalizations occurring before landing were excluded from this calculation, since time of entry before official landing is needed to derive the exposure-adjusted incidence rate.

To measure the burden of TB-related hospitalizations among new immigrants, the mean, median and 90th percentile of all TB-related hospitalizations are derived for each of the characteristics of interest. Because the length of hospitalization for TB has a skewed distribution to the right, the results will include means and medians. The percentage of TB-related hospital discharges among recent immigrants is derived from two sources: the number of TB-related hospital discharges among recent immigrants included in the hospital file from 2001 to 2013, and the total number of discharges among all people within SDLE, regardless of immigrant status. This percentage is calculated by age at hospitalization, sex, region of submitting province, and hospital fiscal year, each available from the DAD database.

Results

Baseline characteristics of immigrant cohort

Around 2.7 million immigrants who arrived in Canada from 2000 to 2013 were in the study cohort (outside of Quebec and the territories, as well as Manitoba before April 2004). Approximately 52% of the cohort were females, and more than 45% were 25 to 44 years of age (Table 2). The majority (73%) of the cohort landed in Canada after 2003. Approximately half of the study cohort were from the following six countries: China (15%), India (15%), the Philippines (11%), Pakistan (4%), Iran (3%) and the United States (3%) (data not shown). Approximately 74% of the study cohort came from countries on the IRCC designated list for TB, composed of countries from the WHO high TB-burden list (58%) and IRCC-specific countries (16%). Most immigrants intended to reside in Ontario (60%). The majority (59%) of the cohort arrived as immigrants in the economic category, while 28% arrived in the family category and 11% arrived in the refugee category.

Distribution of immigrants with at least one TB-related hospitalization

Approximately 1,120 immigrants were found to have at least one acute care hospital discharge for TB from the fiscal year 2001/2002 to the fiscal year 2013/2014 (Table 3), totalling 1,340 TB-related hospitalizations. Approximately half (45%) of immigrants hospitalized for TB were aged 18 to 34 at the time of hospitalization. More males than females were found to have had TB-related hospitalization (52% males compared with 48% females). Most cases occurred among immigrants who landed before 2009. A large majority of cases (97%) occurred among immigrants from the IRCC list of designated countries for TB: 77% were from the WHO list of high TB-burden countries, while an additional 20% were from IRCC-specific countries. India, the Philippines, China, Ethiopia, Pakistan and Somalia were the leading countries of birth of immigrants hospitalized for TB; immigrants from these countries together represented 61% of all recent immigrants with TB-related hospitalization. By immigrant category, most cases occurred among the family and refugee categories (37% and 29%, respectively).

Time from landing year until first TB hospital event

Pre-landing TB hospital event

Nearly 10% (n=110) of immigrants with a TB-related hospitalization experienced their first TB hospital event before “landing” as permanent residents (Table 3). This proportion differed by characteristics, varying from 3% among those aged 45 to 64 to 14% among those aged 25 to 34 (Table 4). By immigrant category, this ranged from 3% among the family category to 41% among others (including unknown and missing). Among the top source countries, Sri Lanka had the highest proportion of immigrants experiencing their first TB hospital event before landing at 14%, while Sudan and Afghanistan each had a proportion of 0%.

Among pre-landing TB hospital events (Table 3), 42% occurred among immigrants aged 25 to 34. The majority (95%) occurred among immigrants from high-risk countries. Among the top source countries, a total of 32% of TB hospital events before landing occurred among immigrants from the Philippines (16%), China (9%) and India (7%). Most of these events occurred among immigrants intending to stay in Ontario (66%). Approximately 39% occurred among refugees, followed by 30% among immigrants entering as principal applicants in the economic category.

Post-landing TB hospital event

Among the 1,120 new immigrants, 18% experienced their first TB-related hospitalization within 12 months of landing, 49% within 13 to 60 months of landing, and 24% after more than 60 months since landing (Table 3). Around 41% of post-landing TB discharges occurred among immigrants who landed in Canada from 2000 to 2003. Immigrants aged 25 to 34 had the highest proportion of post-landing TB-related discharges, at 27%. These TB-related hospital patients tended to be male (53%), be from high-risk countries (97%) such as India and the Philippines, intend to reside in Ontario (63%), and be from the family category (39%)—with some exceptions by time of landing. For example, more females than males experienced a TB-related hospitalization within 12 months of arrival, but not in other periods.

Incidence of TB hospitalization since landing

On average, incidence of TB-related hospitalization in person-years adjusted per 100,000 population was 5.2 per 100,000 person-years for the landing cohort from 2000 to 2013 (Table 5). Among immigrants in the landing cohort, rates were highest among those aged 65 and older (11.3), males (5.6), and those who arrived from 2009 to 2013 (6.8). The rate among immigrants from high-risk countries was more than 11 times that of those from low-risk countries (6.8 compared with 0.6, respectively). The incidence was highest among those from Ethiopia (51), followed by Somalia (43) and Nepal (42). In addition, the incidence was highest among refugees (14) and immigrants intended to live in the Prairies (8).

Burden of TB-related hospitalizations among new immigrants

The overall mean length of stay for all TB-related hospitalizations was 22 days, while the median was 14 days. In comparison, the average among the new immigrant cohort from 2001 to 2013 was lower, at a mean of 17 days and a median of 11 days (Table 6). Among new immigrants, those who are older, male, residing in British Columbia, and from the family category tended to have longer stays. The mean length of stay differed only slightly between high-risk and low-risk countries, at 17 days and 16 days, respectively. The corresponding median lengths of stay were 11 days and 8 days, respectively. New immigrants from Sudan had the highest mean length of stay (26 days), while immigrants from Sri Lanka and Afghanistan had the highest median stay (14 days).

Approximately 17% of the total number of TB-related hospitalizations (n=7,675) occurred to immigrants who arrived in Canada from 2000 to 2013. Immigrants arriving during this 13-year period represented 7% of the Canadian population (Table 7).^{Note 28} The percentages of TB-related hospitalizations varied by age group, from 8% among new immigrants aged 65 and older to 37% among new immigrants aged 18 to 24 and 25 to 34. Each group represented only 2% (aged 65 and older),12% (aged 18 to 24) and 8% (aged 25 to 34) of the population share. Among recent immigrants to Ontario, representing 8% of Ontario’s population, had 24% of the TB-related hospitalizations.

Discussion

This is the first national study (excluding Quebec) of TB-related hospitalizations among new immigrants based on the linked IMDB–DAD database. It demonstrates the value of linking administrative data to understand immigrant health, even for uncommon outcomes such as TB in Canada, complementing the results from the annual TB reports issued by PHAC. This study focused on TB-related hospitalizations, which are generally the most costly component of TB control programs.^{Note 29} The results reveal that new immigrants represent a disproportionate number of people in Canada undergoing TB-related hospitalizations. This finding is supported by previous evidence suggesting that immigrants are overrepresented among TB cases in low-incidence countries such as Canada.^{Note 7} The sociodemographic factors highlighted in this study are comparable with previous studies that look at predictors of TB-related hospitalizations in Canada.^{Note 30}

While this study focused on TB-related hospitalizations, the characteristics of hospitalized patients generally reflected the characteristics of TB patients included in the case-based surveillance system maintained by PHAC.^{Note 8} For example, while males accounted for 54% of reported cases in the CTBRS, males represented around 52% of the IMDB cohort with TB-related hospitalizations. In terms of age, around 33% of the reported TB cases were from individuals aged 25 to 44, while around 45% of immigrants hospitalized for TB were among the same age group. However, direct comparison of the CTBRS with the linked IMDB–DAD database is challenging. The CTBRS did not contain information on the immigrant categories, the type of care provided (outpatient or hospital care), or hospital admissions not related to an active TB diagnosis (e.g., hospitalizations related to side effects of TB treatment). Further studies would be required to make an appropriate comparison.

Based on its linkage to the IMDB, this study revealed that immigrants in the refugee and the family categories are disproportionately represented among individuals with TB. In addition, the CTBRS reported more than 1,600 active TB cases per year across Canada during the study period, approximately 1,000 cases per year being related to foreign-born individuals (60% to 70% of cases per year). The results show that 17% of total TB hospitalizations occurred among new immigrants, despite new immigrants who landed between 2000 and 2013 accounting for only 7% of the population. The data did not include hospitalizations from Quebec and the territories, or from Manitoba before 2004. Because of this, hospital use reported in this study is an underestimate of the true national figure.

A large majority of immigrants with TB hospitalizations were from countries with high risk of TB, as identified by IRCC’s designated country list. Specifically, findings from this study identified the top source countries for TB, including India and the Philippines (in terms of cases), as well as Ethiopia and Somalia (in terms of rates).^{Note 31} The CTBRS reported that approximately 60% of active TB cases occurring among foreign-born individuals occurred among people from the Philippines, India, China, Viet Nam and Pakistan, but Ethiopia and Somalia were not identified. Known migration trends and the domestic TB rates in the home countries of migrants can inform immigration policies and in-Canada TB control programs.

An unanticipated advantage of this linkage is the ability to derive information on the timing of TB-related hospitalization in relation to landing date rather than arrival date, which can inform screening practices. Approximately 1 in 10 TB-related hospitalizations occurred prior to landing in Canada. These cases represent newcomers who acquired the legal right to “land” (permanently) as immigrants as per IRCC requirements only after their original arrival in Canada. They include in-Canada asylum seekers and some temporary residents (students, visitors and temporary workers). Hospitalization in this group likely occurred among individuals diagnosed through activities related to public health (e.g., contact tracing of active TB index cases, and review of clients referred by IRCC) after seeking medical attention for abnormal findings or symptoms, or as part of a new in-Canada immigration medical examination (IME). The IME is used by IRCC to screen for active pulmonary TB and identify clients with latent TB infection. Some individuals would not have had an IME before originally entering Canada (e.g., asylum seekers in Canada). These pre-landing results can inform immigration policy and TB screening procedures based on sociodemographic TB risk factors.^{Note 32Note 33} Recent papers have identified sociodemographic predictors of active TB in migrants (not restricted to hospitalization) and provide complementary evidence to inform immigration policy.^{Note 34}

Hospitalizations that occurred within 12 months of landing could represent immigrants whose active TB was either missed during IME procedures or whose infection was acquired after the IME but before arrival in Canada. In contrast, cases that were identified as occurring after one year of landing more likely represent a reactivation of more distant latent TB infection. Since most cases occurred after one year, this suggests that the reactivation of latent TB infection accounts for a significant proportion of hospital admissions for TB among new arrivals to Canada.^{Note 35Note 36Note 37} WHO recommendations for countries with low incidence of TB, such as Canada, indicate that screening for latent TB infection in TB contacts and selected high-risk groups should be a priority action.^{Note 5} As a result, this analysis provides insight that can help develop strategies to decrease the reservoir of infected individuals and eliminate TB.^{Note 38Note 39}

In this study, the mean length of hospital stay was 17 days, while the median length was 11 days (Table 6) among the immigrant landing cohort. This is less than the mean length of 22 days and median length of 14 days for TB patients nationwide, and less than the 20.6 days reported previously.^{Note 29} It is possible that recent immigrants are younger or have less advanced disease at the time of admission, resulting in a shorter mean or median length of hospital stay. The burden of hospital care for TB among the landing cohort aged 65 and older (mean 22 days, median 16 days) could be explained by previous findings. Previous findings show that this age group had statistically significant factors prolonging the length of stay.^{Note 40Note 41} In addition, this age group most likely represented more complex TB cases, a higher incidence of significant co-morbidities, and more disease severity with associated delays in TB diagnosis and mortality.^{Note 42Note 43} According to immigration category (economic, family and refugee), the economic category had the shortest length of hospital stay, while the refugee category had the longest. The shorter hospital stays for the immigrant cohort might also reflect an increased index of suspicion as a result of IME for TB in this group, leading to earlier diagnosis and shorter hospitalization.

This study found that certain provinces, such as British Columbia, had longer TB hospital days. It is beyond the scope of this paper to interpret the differences of hospital stay between provinces, but appropriate interpretation would likely involve factors such as age, co-morbidities, index of suspicion, differing patterns of practice and alternate care options.^{Note 30Note 32Note 34Note 40Note 41Note 42Note 43} These interrelated factors may partially explain why some provinces have longer hospital stays.

Limitations

This study has certain limitations. The hospital use reported is an underestimate of the national figure for recent immigrants because it does not include Manitoba before April 2004 or Quebec. The study also does not include hospitalization for processes that were not identified initially as TB-related, but ultimately proven to be TB-related after discharge. Pre-landing data may also be underestimated if there were issues related to one’s ability to pay for hospitalization. This could lead to an outpatient management for conditions usually managed in hospital. It is assumed that landed immigrants are covered by provincial insurance, except for any health care waiting period that is required by new immigrants. A group of in-Canada residents that are not identified in this study includes temporary residents (e.g., students, workers, visitors) who either did not apply for permanent resident status or did apply and were unsuccessful.

When appropriate, TB can be treated in an ambulatory setting, including measures such as home isolation. TB cases treated using strictly ambulatory management or cases where the ICD codes for TB were not entered in the hospital discharge summary are not reflected in this study. Despite this, linkages in this study can provide insight into epidemiological patterns of TB-related hospitalization in Canada (e.g., immigration category and countries of concern with high risk of TB) and associated costs (related to lengths of hospital stays for TB).

Conclusion

This study provides new insights into aspects of TB-related hospitalization occurring among new immigrants to Canada using linked administrative immigration and hospital data. With the increasing number of immigrants to Canada, this linked dataset provides a timely contribution to developing immigrant health surveillance systems and monitoring the health and health service use of immigrants, by immigration category, as well as to contribute to the commitment by the Government of Canada to end TB.^{Note 44} It can also provide policy makers with information to help develop approaches for managing a disease that, in Canada, is relatively uncommon among the non-Indigenous Canadian-born population, but more prevalent among the immigrant population.

Acknowledgements

This research was supported by Immigration, Refugees and Citizenship Canada (IRCC). The authors would like to acknowledge Michael Mackinnon who provided insight and all the reviewers who provided critical reviews which greatly improved the manuscript. The authors also want to thank Maria Syoufi of IRCC who provided general administrative support and participated in the technical editing of the manuscript, as well as the Canadian Institute for Health institution for reviewing the ICD-9 and ICD-10 codes for TB.