This study was made possible by the cotlaboration of a number of persons. Firstly, Dr. Monika Naus at ...... manual search was perfomed using FOSS. Cases found in ..... Unexpectedly, the standard error of the risk estimate for PLS ...... Epidemiological and dinical study of tuberculosis in the district of Kolin,. Czechoslovakia.
INFORMATION TO USERS This manuscript has been reproduœd from the microfilm master. UMI films the text directly from the original or copy submitted.
Thus, some thesis and
dissertation copies are in typewriter face, while others may be from any type of cornputer pfinter.
The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard rnargins, and improper alignrnent can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, &arts) are reproduoed by sectioning the original, beginning at the upper lefi-hand corner and continuing from left to right in equal sections with srnall overlaps. Each original is also photographed in
one exposure and is inciuded in reduced form at the back of the book. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appeanng in this copy for
an additional charge. Contact UMI directly to order.
Bell & Howell Information and Leaming 300 North Zeeb Road, Ann Arbor, MI 48106-1346 USA
8oO-52l-06OO
RlSK FACTORS FOR ACTIVE TUBERCULOSIS AMONG FOREIGN-BORN PERSONS ARRlVlNG IN ONTARIO.
Wendy LeeWobeser
A thesis submitted in conforrnity with the requirements
for the degree of Master of Science Graduate Department of Community HeaIth University of Toronto
OWendy Lee Wobeser 1998
1+1
ofcanada
National Library
Bibliothèque nationale du Canada
Acquisitions and Bibliographic Services
Acquisitions et setvices bibliographiques
395 Wellington Street Ottawa ON K IA O N 4
395. rue Wellington Ottawa ON KI A ON4
Canada
canada
The author has granted a nonexclusive licence allowing the National Library of Canada to reproduce, loan, distribute or sell copies of this thesis in microform, paper or electronic formats.
L'auteur a accordé une licence non exclusive permettant à la Bibliothèque nationale du Canada de reproduire, prêter, distribuer ou vendre des copies de cette thèse sous la forme de mi~rofichelfilm~ de reproduction sur papier ou sur format électronique.
The author retains ownership of the copyright in this thesis. Neither the thesis nor substântial extracts fiom it may be printed or othenvise reproduced without the author's permission.
L'auteur conserve la propriété du droit d'auteur qui protège cette thèse. Ni la thèse ni des extraits substantiels de celle-ci ne doivent être imprimés ou autrement reproduits sans son autorisation.
Risk factors for active TB among foreign-bom persons aniving in Ontario. Wendy Lee Wobeser Master of Science Graduate Department of Community Health University of Toronto, 1998
ABSTRACT Migration has resulted in the resurgence of tuberculosis (TB) as a public health concem in developed wuntnes. In Ontano over 80% of cases of TB occur among foreign-bom persons. The study deterrnined nsk factors for TB among this group. This was a population based case-control study. Cases were enrolled from the Ontario Reportable Disease Information Service.
Data was obtained from this source and databases
rnaintained by Citizenship and Immigration Canada. Risk factors for the developrnent of active TB included age at time of immigration, referral for post-landing surveillance
(PLS) at time of immigration, world region of origin, time since landing and educational attainment. Persons referred for PLS, despite being at increased nsk, acounted for only 13% of al1 the cases of TB.
The provincial TB control program must develop
alternative strategies to control TB among foreign-born persons in Ontario. This study provides key information for the development of such strategies.
ACKNOWLEDGEMENTS Dr. Lilian Yuan acted as a mentor and provider of helpful and much appreciated direction throughout the course of this thesis project. This study was made possible by the cotlaboration of a number of persons. Firstly, Dr. Monika Naus at the Ontario Ministry of Health, Public Health Branch provided key input into the development of the study questions. Dr. Naus and Diane Kerbel facilitated the release of provincial data. Persons a i Citizenship and Immigration Canada who provided essential assistance included Dr. Neil Heywood. Dr. Heywood provided input into the development of the study questions and facilitated the release of data from C C . Claude Langlois provided the linkage of the provincial database and the CIC database and guidance regarding the contents of the CIC database. Dr. Paul Corey provided essential support and advice for the design and analysis of the study. Dr. Jirn Bninton acted as a much needed mentor. Dr. Dick Menzies and Dr. Jeff Edelson read the manuscript and provided very helpful comments and criticisms. Dr. Linn Holness was instrumental in btinging about the cornpletion of this project. The National Sanitarium Association and Citizenship and Immigration Canada provided financiat support for the completion of this project. My husband Brïan and our son Julius were supportive and joy provoking throughout. Our daughter Lillian arrived on the scene to provide cheers as I went over the finish line.
iii
TABLE OF CONTENTS List of tables
List of figures List of appendices Chapter 1
Background
1.1
Impact of migration on the epidemiology of TB and screening for TB among immigrants
1.2
Program d-ption
1.3
The provincial perspective
1.4
Study deçcription
Chapter 2
- Canadian immigration medicd screening program
Literatum Review
2.1
Natural history of TB
2.2
Screening for active and inactive TB four basic strategies
2.3
The chest Xray - its utility in i d e n t . g a high risk group
2.4
interventions to reduce risk among persons mith inactive tubercuiosis
2.5
The impact of migration on the epidemiology of TB in low prevalence countries
2.6
Immigration screening programs the example of the United States
2.7
Immigration screening prograxns - the Canadian experience
2.8
summary
Chapter 3
-
-
Study methods
3.1
Study objectives
3.2
Study design
3.3
Study approval
3 -4
Data sources
3.5
Data acquisition
3.6
Methods of data linkage and preparation
3.7
Method of control selection
3-8
Sample size considerations
3.9
Variables
3.10
Data anaiysis
3.11
Mode1 construction
Chapter 3
Results
-
Results of linkage Reportable Disease Information System (RDIS) and the Jhdeci immigrant Data System (LIDS) 39
Cornpison
of foreign-born cases of TB enroiied and not enrolled in the sîudy 40
Demographic characteristics of cases and controls
42
World region of origin, burden of TB in country of birth and country of last permanent residence 44
The effect of a classification for pst-landing surveiilancc (PLS)on risk of active TB Cornparison of controlling for factors in the study design or in the analysis
50
The effect of age and gender on the risk of active TB
51
The effect of year of landing on risk of active TB
53
The effect of world region of origin on risk of active TB and predictive power of a referral for PLS
53
Analysis of other factors 4.10a
The effect of incidence of TB in country of birth on the risk of active TB 56
1.lob
The effect of immigration classiEicaiïon on the risk of active TB
58
4 . 1 0 ~ The &ect of educational attainment at the time of immigration on the risk of
active TB 4.1Od
59
The effect of maritid status at the t h e of unmigration on the risk of active TB 59
4.10e
The effect of intended occupation at the time of immigration on the ri& of active TB 60
Final mode1 Chapter 5
Discussion
Chapter 6
Conclusions
References
61
List of Tables Cornparison of foreign-boni cases of TB enrolled and not enrolied in the study
Table f
Il
Table 2
Demographic characteristics of cases and unmatched controls
43
Table 3
World region of origin of cases and controls
45
Table 4
Post-landing surveillance classification of cases and controls
47
Table 5
Characteristics of cases referred versus those not referred for PLS
49
Table 6
Comparing the effect of controlling for factors in the design versus the anatysis 51
Table 7
Age and risk of active TB
52
Table 8
World region of ongin and risk of active TB
55
Table 9
World region of origin and the predïctive p v e r of a referral for PLS 56
Table 10
Immigration c a t e g o ~and risk of active TB
Table I l
Educationai attainment at time of immigration and risk of active TB 59
Table 12
Marital status and ri& of active TB
60
Table 13
Occupation and risk of active TB
60
Table 14
Risk factors for active TB among foreign-born persons - Ontario
58
62
1994-95
List of Figures Figure 1
Annual immigration to Canada - 1950 to present
Figure 2
Estimateci burden of TB in ten top source countries to Canada - 1993 3
Figure 3
Historical summary of TB by place of birth Canada, 1980-94
4
Figure 4
Estirnateci TB incidence in country of birth by year of landing
45
Figure 5
Eüsk of active TB by age at Iaading - stratified by gender
52
Figure 6
Risk of active TB by year of landing
54
Figure 7
Risk of active TB by world region of ongin
55
Figure 8
Rlsk of active TB by burden of TB in country of birth
57
-
2
List of Appendices Appendis A
Canadian Classification and Dictionasy of Occupations 1971 - Major IFOUPS 94
Chapter 1 Background 1.1 impact o f migration on the epidemiology of TB and screening for TB amoog immigrants Woridwide, tuberculosis (TB) is a disease of iremendous importance resulting in an estimated I O million cases and 3 million deaths annually population is infected with the tubercle bacillus
2.
1 3
. One third of the world's
Immigration of persons from
countnes which have a high burden of TB has infiuenced the epidemiology of TB in many developed (low burden) countries. In these countries, the proportion of cases of TB occumng among foreign-bom individuals has increased and, in some, the absolute number of cases have also increased 412. ln Ontario, over 80% of new active cases of TB occur among persons who were bom outside of Canada 13. It would appear that cases of TB among foreign-bom persons in this province will increase. The first reason that an increase in TB case numbers can be expected is that the absolute number of persons immigrating to Canada increased substantially in
the early 1990's. Currently more than 200,000 persons immigrate to Canada each year (figure 1).
Figure 1 Annual immigration to Canada 1950 to present
Year
-
Source: Citizenship and Immigration Canada Facts and Figures A995
A second reason to expect an increased number of TB arnong foreign-born persons is that recent trends indicate an increasing number of immigrants originate
from countries with high burdens of TB 14. Presently the majority of immigrants originate in a country with a high burden of TB (incidence of greater than 100 per 100,000 per year). Nine of eleven of the top source countnes of immigrants to Canada
in 1994 were countries with a high burden of TB (figure 2).
Figure 2 Estimated burden of TB in ten top source countries to Canada 1994
-
aHong Kong lndia Phillipines Taiwan
China
-ri
Lanka Vietnam Pakistan USA
Source Country Source: World Bank
- World Health Report 1993
8, Citizenship and Immigration Canada
-
Potand
Facts & Figures 1995
The risk of developing TB is similar for recent immigrants and perçons remaining in the country of origin 10.1416 . Although this risk decreases over time, the relative risk of active disease compared to that of the non-foreign-born population remains elevated throughout the life of the foreign-bom penon ''? As long as immigration to Canada continues at curent levels. numbers of cases of TB in Canada are projected to remain at the same level or potentially rise. Figure 3 illustrates the increasing contribution of TB among foreign-bom to the epidemiology of TB in Canada.
The number of cases among Canadian-born penons has trended downward over the last 15 years while numbers among the foreign-bom are trending upward.
Figure 3 Historical summary of TB by place of birth, 1980-94
Year
Source: Statistics Canada - Tuberculosis Statistics, 1994
Transmission of TB frorn the foreign-bom to the native-bom population has been demonstrated in outbreaks, but has not been shown to have a rneasurable impact at the population level. In Great
ri tain'^ and in ~ontreal", ecological studies were
unable to dernonstrate evidence of transmission of TB between these two cornmunities. In the Netherlands, however, rnolecular epidemiological evidence suggests that 25% of cases of TB among native-bom persons are Iikely a result of transmission from the
foreign-born population1g. In San Francisco, few secondary cases of TB in the local cornmunity could be attributed to immigrants and refugees 20. Weil described outbreaks have emphasized the importance of localized transmission 21"4. Canadian evidence also supports transmission in prisons and hospitals2526.
5
One of the tools used to reduce the importation and burden of TB amongst immigrants is a screening program administered at the tirne of immigration. Most developed countries screen immigrants and refugees 5"".
The pn'rnary goal is to
detect aa-ve disease. As well these programs identify persons at high risk of developing TB. Characteristics used to identiw high risk persons include chest Xray changes consistent with inactive TB or 2 history of TB. Such characteristics are used to select a group which is targeted for surveillance after they have arrived in their country of destination, a practice known as post-Ianding surveillance (PIS)_ Among persons with chest Xray changes consistent with inactive TB, age and world region of origin have been shown to influence the risk of active TB
. Persons with
changes on chest Xray consistent with inactive TB are candidates for preventive therapy to reduce their future risk of TB 32. The two key determinants of the effÏciency of a program which refers individuals
for medical evaluation after landing in the country of destination are the size and relative risk of the referred group. The size of the referred group has increased in the last decade. This increase in size is prirnarily attributable to increased immigration. As
a result of greater nurnbers of persons being referred to public health officiais for PLS an increased dernand has been made on Ontario public heaith resources. Risk
estirnates have been made for persons referred for PLS in Manitoba and British ~ o l u m b i a ~however * ~ . it is not known if these estimates are generalizable to the Ontario population of immigrants. The purpose of this study is to estimate the relative risk of TB among persons referred for PLS in Ontario and to identify other factors
associated with increased risk of developing TB.
6
1.2
Program description - Canadian immigration medicd screening program The general goals of the Canadian immigration program are fourfold: to enrich
Canada's social fabric; to bnng a body of skiiled workers into the work force; to promote economic growth, and to realize Canada's humanitarian mission". Components of the Canadian immigration progrârn include a security clearance, medical evaluation and skiII assessment. Canadian immigration law requires that "every immigrant and every visitor of a prescribed dass shall undergo a medical examination by a medical officef 36
. The general goal of this screening policy is to exclude persons wtio are suffen'ng
from a disease, disorder, or disability which would likely be a danger to public health or
to public safety or which might reasonably be expected to cause excessive demands on health or sociaf setvices. A TB screening program cunstitutes one of the major components of the medical evaluation of al1 persons immigrating to Canada. This program is delivered by Citizenship and Immigration Canada (CIC). The two major objectives
of the program can be divided into short terni (identification of persons with
active TB) and long terni (identification of persons at high nsk for the developrnent of
active TB). The screening program is based on chest Xray scieening ûi appiic;aiïi~wiio are at least 14 years of age. If changes consistent with TB are found on the chest Xray the individual is screened with a sputum smear and culture or a repeat chest Xray at three rnonths. Persons identified with active TB are required to undergo treatrnent pnor to entry into Canada.
7
Persons felt to be at an increased risk for the development of active TB (those with abnormalities consistent with TB on chest Xray but with a negative smear and culture or stable follow up chest Xray at three months) are referred for PLS. These individuals enter Canada conditional upon their agreement to present to provincial public health authorities after arrival. There are, however, no consequences associated with failing to adhere with PLS. Notification of persons referred for PLS are sent to public heatth officials in the province of destination at the time of entry into the country. Guidelines have been published for the management of persons referred for
PLS 37. PLS consists of a medical evaluation in Canada, utilisation of therapy to prevent TB, and regular follow up where appropriate. The potential benefits of the inCanada evaluation include two possible interventions: early diagnosis of active TB and the reduction of future risk of TB. Both these interventions would be expected to decrease the amount of transmission of TB in Canada.
8
1.3
The provincial perspective The cost-effectiveness of a PLS prograrn on a provincial leveI has been
questioned. Idealiy, this program wouid prevent cases of future TB among a high n'sk group and, as a result, prevent future transmission of TB in Canada. Currently there are a number of problems with PLS in Ontario. Prominent amongst these is inadequate data regarding its effectiveness. Because PLS has not been futly operationalized in Ontario data are inwrnplete. Between 12,000 and 78,000 persons are referred for PLS in Canada each yeaP. Approximately half of these individuals amve in Ontario. In Ontario, it is estirnated that less than U3 of those referred for PLS adhere and receive a rnedical evaluation in canada".
As well, amongst those who are assessed, few are
offered an intemention which would reduce their future risk of TB (Le. isoniazid preventive therapy). lncreasing numbers of persons have been referred for PLS over
the last five years. However, persons referred for PLS rnay be unaware of the reason for such a referai. Medical care providers in Canada rnay also be unaware that their patients have been referred for PLS. In addition, active cases of TB may be treated anyway. It is not clear that the PLS program results in more cornprehensive or earlier treatment. National guidelines recommend that perçons referred for PLS should be considered for preventive therapy and be followed for 3-5 years 37. The cost to Ontario of providing a comprehensive PLS program would be substantial. Given the difficulties
in operationalizing a prograrn and concems that those referred for PLS are not tnrly at high risk for TB, this study was designed to delineate the contribution of a referral for PLS to risk of active TB. Other factors contributing to risk of active TB were also analysed.
1.4
Study description The study was population-based and case-control in design. Cases and
controls were drawn from the population of al1 persons landing in Canada between 1986 and 'i995, destined for Ontario, and greater than eleven years of age at the time of immigration. Cases were persons in this population who were diagnosed with active
TB in 1994-95. Controls were persons not known to have active TB in 1994 or 1995. A logistic regression model was constwcted using active TB as the outcome. The pnrnary exposure of interest was referrat for PLS. Additional factors of potential importance which were analyzed included age at time of immigration, gender, year of obtaining landed status in Canada, immigration classification, world region of origin, prevalence of TB in country of origin, intended occupation and level of education at time of immigration. Cases of active TB were identified from the TB registry of the province of Ontario. Cases were included in this study if the person was born outside of Canada, obtained legal landed status in Canada between 1986-95, and their landing record was available from the database ai CIC. Exclusion criteria included diagnosis with active TB in Ontario prior to the immigration medical evaluation and age less than IIyearç of age at time of visa issue (only those at least 11 years are screened with a
chest X-ray). Linkage with a CIC database (Landed Immigrant Data System (LIDS)) was perfonned to obtain PLS classification, tanding date, country of birth, immigration category, occupation and education for the cases. Controls were also enrolled from
LIDS. This database inciudes the entire cohort of persons landing in Canada for each year. The controls were selected from those destined for Ontario and at least 11 years old at the tirne of medical evaluation. The case:control ratio was i:4. Three sets of controls were used in the analysis: 1) matched on year of Ianding; 2) matched on year of landing, age at landing and gender and 3) unmatched. Year of entry was chosen as a matching factor to atternpt to control for changes in the pattern of immigration over
10
the ten pars of the çtudy. Such changes may be associated both with referral for PLS and risk of active TB- Age and gender are also potentiaily associated with the development of active TB and a PLS refenal and were therefore matched in one of the control groups.
Chapter 2 Literature review 2.1
Natural history of TB Tuberculosis is caused by the bacillus Mycobacterium tube~~uIosis.In most
cases this bacillus attacks the lungs. Upon reaching lung tissue one or more bacilli can induce a nonspecific inflamrnatory response. Activation of normal immune defenses leads to spontaneous resolution of ihis local inflammation, leaving an area of parenchymal fibrosis and calcification in a draining lymph node. The components of the primary complex include a lesion in the lung and in a corresponding lymph node. These lesions contain a srnall foci of bacilli which are "donnant". These dormant bacilli can be reactivated and cause clinical disease at any time dunng an individual's Iifetime. Factors which favor reactivation inchde 1) high bacillary load and 2) compromised host defenses. Two to six weeks following primary infection, the body's immune system develops a certain level of cell-mediated immunity to proteins of the organism. As a result a granulorna foms around the bacillary focus. These foci, in time, become calcified. The calcified lesions can be detected on chest X-ray if large enough ". The risk of developing tuberculosis is detemined by Wo distinct processes: the n'sk of being infected with the tubercle bacillus and the risk of progressing to active disease once infected. The risk of being infected is primarily determined by the annual n'sk of infection of the community in which a person is living and the duration living in this community. The effect of age on the risk of TB is influenced both by the risk of infection with TB and the rÎsk of progressing (the latter risk is not constant through Iife). In communities with a high risk of infection (such as the developing world) the nsk of
TB is concentrated among children and young adults. In communities charactenzed by
12
a low burden of TB (and subsequenffy by a fow risk of infection), TB cases tend to be more concentrated among older adults. Other factors wtiich affect risk of being infected with TB inciude gender and socioeconomic factors 41". exists for mortality from TB ".
A social class gradient
Unemployrnent has been associated with incidence of
TB l6and prevalence of tuberculous infection is inversely associated with income ". The association of prevalence of infection with incorne has also been obsenred among foreign-bom persons in Montreal 47. As well, education has been associated with morbidity and mortality ". Among persons infected with the tubercle bacillus, the Iifetime nsk of developing active TB is estimated to be 540% ". This risk depends on a number of other factors such as tirne since infection, age, imrnunosuppression, other underlying medical illness
and malnutrition 4'. On a population levef, penons who are infected with the tubercle bacilfus (based on a positive tuberculin skin test) are at 4-5 times greater risk of developing TB compared to uninfected persons
This relative risk estirnate will be
affected by the background risk of infection with TB. The most potent risk factor for TB in individuals remotely or recently infected with the tubercle bacillus is HlV infection. Persons with extensive chest X-ray changes consistent with TB have a higher rate of wnfinned TB than those with less extensive changes, who in turn have a higher rate
than those with a normal chest Xray 5*a150. Among persons immigrating from a country with a high burden of TB to one with a low burden of TB, many of the cases of active TB are assumed to occur as a result of reactivation of "donnant" foci of infectious organisrns acquired in their country of ongin 1O
- This results in these persons having a risk of TB simitar to that observed arnongst
their compatriots who did not emigrate ' 4 .
2.2
13
-
Screening for active and inactive TB four basic stratees One of the objectives of medical screening is to discover conditions suitable for
eariy preventive or curative intervention. These conditions may not be sufficiently symptomatic to induce the patient to seek medical help on his or her own. The condition being screened for must be sufficiently prevafent for the screening procedure to be effective; rnust have agreed upon diagnostic criterion; must have a known natural history; and must be amenable to a definitive intervention
'' . lt rnust be ensured that
the persons with conditions discovered through screening get prompt and definitive
medical attention, including patient counseling, which in itself needs to result in restoration of health or in a dearly targeted intervention- In the case of TB, transmission of M. tuberr=ulosisshould be intempted or prevented frorn occuming.
Each screening tool has its predefined operaüng characteristics, i.e. sensitivity and specificity. The predictive value of a positive finding is dependent on these two characteristics, but crucially on a third, the prevalence of the condition. The value of screening is improved if the condition has a high prevalence in the population which is to be screened. The implernentation of a screening program for TB must be rneasured in ternis of rewards and predictabte darnage that might be done by the screening itself 8
- The Worid Health Organization (WHO) considers a high risk population as one in
whom active casefinding (screening) rnay be effective. The definition accepted by the WHO is a population with an annual incidence of TB of greater than 100/100,000 ". Four important tools are used to detect tuberculous infection and clinical disease. These tools include the tuberculin skin test (TST), sputum smear, sputum culture, and chest Xray. The TST measures the cell-rnediated immune response to M. fuberculosis antigens. The TST is characten'zed by a number of limitations as a screening tool. The sensitivity of Me TST is less than 90% ? lnterpretation of the TST
14
is reader dependenf and is characterized by high inter-observer variabiiitya. The TST does not distinguish between recent and rernote infection. Nor does the TST reliably distinguish between an infection caused by M. tubercuiosis and one caused by other mycobacteria including BCG". BCG vaccination creates false-positive TST reactions. The interpretation of a positive TST in persons immigrating from high TB burden countries is complicated as the majority of these populations are vaccinated with BCG. A further limitation of the TST is that it does not differentiate between infection and
active disease. Ji is an inefficient tool for detecting active cases of TB. Overall the TST could be expected to identify a large pool of individuak (35-53% of foreign-bon? panons from high burden countries 6847 ) who are at a moderately increased (4-5 fold) fisk of developing disease
. In spite of its limitations the TST is a potentially strong
predictor of active TB among immigrants with an abnormal chest Xray
31
. However, it
would appear difficult to deliver a high quaiity TST screening program worldwide. TST screening may play an important role in TB control activities among foreign-bom persons, however, it is Iikely best applied af%eramval in Canada. Detection of acid-fast bacilti (smear) in sputurn is the most important tool to detect highly infecüous cases of TB. Although sputum smears perfonn welt in this setting, they have limited sensitivity (50-70%)for the diagnosis of ail 'forms of pulmonary TB. As well sputum smear may be difficult to collect under field conditions 54
Culture of specirnens for mywbacten'a detects tubercle bacilli in sputum containing insufficient organisms to be detected by sputurn smear. Culture requires four to six weeks. This is a costly and time consuming rnethodology which is not availabfe in many countries in the developing worid.
15
The chest Xray identifies a much a smaller group of individuals (2-4% of individuals screened ) than the TST. However, this group is potentially at much higher relative nsk of developing active disease (in the range of 1046) wrnpared to the general population "? The degree of the increased risk is dependent upon what is accepted as a positive finding on chest Xray. In addition to identifying a group at high nsk for the future development of TB, the chest Xray should identify rnost individuals with active pulmonary disease who are infectious and thus represent a public health
nsk 56.57.
2.3
The chest Xray - its utüity in identiqing a high risk gmup The chest Xray suffers limitations as a diagnostic and prognostic tool. This is a
result of variabifity in the interpretation of chest Xray findings. Inter-observer variability in chest X-ray interpretation is estirnated to be in the range of 20-30% "-", whereas
intra-obsewer variability is in the range of 10-20%. Published reports emphasize that the chest Xray will fail to identify clinically venfied cases of pulmonary TB at least 5% of the time 56".
Attempts made to develop a standardized classification schema for
tuberculosis based on chest X-ray findings have not been successful 57. Extensive experience was gained in the 1950s and 1960s when mass chest Xray surveys were used as a case-finding tool for active TB throughout the developed
and developing wodd. Rates of chest Xray changes consistent with TB were shown to correlate with incidence rates of bacteriologically proven TB in a number of large population-based surveys conducted by the WHO in Africa 59 . However, chest Xray changes which are consistent with TB are not specific to this infectious agent 49 . Published mass radiological surveys wnducted world-wide were summarized by Kannisto in 1 97260. The nurnber of pnmary findings (findings consistent with TB)
varied from 14.9-68/1000 subjects sutveyed. The rate of prirnary findings has been
16
noted ta Vary with sex, age BD, race 28.33.59 and geography however, is of uncertain epidemiological significance
. This variation,
""' .
The chest Xray identifies less than 70% of persons who are infected with M. tuben:uIosr's. In WHO sponsored African surveys in the 1960's 032.9% of ail those
infected with TB (TST 2 i O mm) demonstrated chest Xray evidence of active or inactive tuberculosis".
Horwitz found 7% of tuberculin reactors to have an abnomal chest Xray
consistent with TB (active or inactive) in a large, population-based study in Oenrnarkm. More cases of TB will anse from those with a normal screening chest Xray than those with an abnonnal screening chest Xray.
Persons with inactive TB based on a known past record of disease are at a relative risk of 10-50 for active TB compared to the general population. The relative rÎsk is significantly decreased by a history of effective
TB chemotherapy
49.62
. Persons
with fibrotic changes on chest Xray and no history of TB are usually identified by mass, population based chest Xray surveys. The relative nsk of active TB of those identified by such surveys has been estimated in the range of 10-164g*55". It appears that
isolated apical scaning is less "high risk" than fibrotic changes on chest Xray. Healed primary wmplex or old pleurisy on chest Xray translates into a relative n'sk of TB of four 49.
This relative risk is comparable to that found for persons infected with M.
tuberculosis (positive TST) ". Estirnates of the annual rate of reactivation
(development of active TB) among perçons with fibrutic changes on chest Xray and no history of active TB or of treatment are in the range of 0.24-0.5% par year 4955*Q. Factors which have been s h o w to affect the risk of active disease amongst perçons with an abnormal chest Xray include TST status, age and burden of TB in country of *rigin28.3O.3I 34. Persons with inactive pulrnonary TB (apical fibrocalcific changes on
chest Xray) and no history of treatment are good candidates for preventive Merapy ".
17
The effect of annual surveillance of persons with abnorrnal chest Xray consistent with TB has been studied in one randomized controlled trial 64. Annual medical evaluation versus instructions to present to the TB c h i c upon development of symptoms did not significantly increase the rate of detecüon of active disease. Half of those referred for annual surveillance presented with syrnptorns outside of the surveillance program. The authors concluded that annual follow up of persons with fibrotic changes on chest Xray is not cost-effective. The chest Xray will detect most cases of infectbus pulmonary TB in addition to detecting a subgroup of persons at high risk of developing TB. The rate of detection of prevalent active TB has been esümated for a number of immigration screening programs. Prevalent active TB was detected in O. 1% of immigrants referred for evaluation at Heathrow airport 65; those screened on arrÏval in Switzerfand 28 and those applying for landed immigrancy status in Denver ". Comparatively the rate of detection of prevalent active cases of TB was 0.6-0.8% among refugees from Haiti 66 and Southeast Asia
In the one Canadian study 1.9% of lndochinese refugees were
diagnosed with active disease although many of these had minimal pulmonary TB ". Among non-immigrant populations the rate of detedon of active TB varies from 0.060.6% and appears to be dependent on the prevalence of TB in the screened
community 60. Recent studies have indicated that an immigration screening prograrn rnay fail to detect cases of active TB which occur within one year of amval 68*.
This
may be a result of either the misinterpretation of the immigration chest Xray or of the lack of universality of the screening program.
18
2.4
Interventions to reduce the risk among persons with inactive TB Preventive therapy with isoniazid reduces the risk of TB arnong persons infected
with M. tube~uIosis.The greatest protection (93% reduction in rate of active TB) is gained with twelve months of therapy ". Overail (including nonadherent patients) the rate reducüon with twelve months of isoniazid is estimated to be 75% and for six months 65% ? lsoniazid preventive therapy is not without nsk. The most important risk is of isoniazid-associated hepatitis. This is a rare, age-dependent phenornenon with risk increasing with age. Isoniazid-associated hepatitis c m be fatal- As a result of
the flsk of hepatitis and difficulties in ensuring adherence with a long regimen, the utilization of isoniazid preventive therapy in the field is highly variable. The efficacy of isoniazid preventive therapy has been well established. The cause of the vanability in its utilization appears to be primanly due to questions regarding its efficiency among
infected but healthy perçons 8.
19
2.5
The impact of migration on the epidemiology of TB in low burden countries International migration imposes on host countries the responsibility to develop
an understanding of migrant health care needs, among which TB prevention and treatment is prominent '.
In the last thirty yean a shift in immigration patterns has
been evident, with a growing proportion and number of immigrants coming from areas which have a high burden of TB
8.14.70
. The risk of active disease in foreign-bom
penons refiects the disease incidence rates in their country of origin 6v14. Persans immigrating from high burden countries may be at a severai fold increased relative risk of developing active TB when compared to individuals bom in a low burden country
.
14,15,27.30,33.34
Persons from high-prevalence areas are most Iikely to develop TB within
a few years of amval in a low prevalence country remains elevated for many years
1°*16.
10.14.71
. However, the risk of TB
TB appears to be one of the most important
health issue for rnany persons migrating to Canada. Most low burden countries have provisions for screening certain groups of
In Western immigrants and other foreigners seeking entrance to the country '"? Europe screening programs wnsist pnmarily of a chest Xray examination and, occasionally, a tuberculin skin test '. Foreign workers and asylum seekers are the two most cornmon target groups for screening. Switzerland screens only persons coming
'
from countries with a high burden of TB . The United States requires clinical examination and chest Xray for al1 individuals applying for immigration and refugee status . The Canadian screening program is targeted at al1 individuals applying for landed immigrancy or refugee status and visitors applying for a stay of >=6 months if
they originate from a country of high TB burden. The Canadian program has defined a high TB burden country as one with an annual notification rate of TB of greater than 5O/lOO,OOO based on WHO notification rates.
20
The overall population impact of immigration of persons from high burden countries to low burden countries is unknown, Ecological studies have been unable to dernonstrate evidence of transmission of TB between foreign-bom persons and those native to Me countries of England and Canada '733 - However, molecular epidemiological studies have demonstrated that such transmission does occur ". Well described outbreaks have emphasized the importance of localized transmission 2''24.
Canadian evidence supports transmission in prisons and hospibls2536. Certainly, the potential for transmission of active TB frorn persons who immigrate to Canada and those born in this country exists. This is particularly worrisome drug resistant TB, an often fatal disease.
in the setting of multiple
21
-
Immigration screening programs the exampIe of the United States
2.6
Binkin et al. recently reviewed overseas screening practkes in the United States 5. The American program is less stnngent than the Canadian program. The former excludes from entry only those individuals who are sputum smear positive and, therefore, highly infectious. Other high nsk individuals are referred for evaluation after landing. Adherence with post-landing evaluation ranged from 63-99% among five different American programs, Median time to follow up after immigration ranged from 6 days to 9 weeks. Confinned cases of TB were detected among 3.3-14% of individuals classified as B I (clinicatly active, not infectÏous)and 0.4-3,8% classified as 62 (clinically inactive, not infectious). Persons classified as B i or B2 were at increased risk of active
TB if they had a positive tuberculin skin test (odds ratio 10.7) or were young (odds ratio 4.8).
31
. In Hawaii and Seattle 23 and 32% respectively of persons developing TB
within one year of entry into the state were not referred for evaluation by the screening program 5. This grmp either had a chest X-ray interpreted as not consistent with active or inactive TB or were not screened. In San Francisco 40% of immigrants and refugees who sought further medical evaluation and follow-up a r e after amival were found to be candidates for preventive therapy?
Arnong a cohort of Southeast Asian
refugees screened for TB and followed in Seattle 2% were found to have stable chest Xray findings consistent with inactive TB. This group contributed 24% (6/25)of al1 active cases and 50% (6/12) of pulmonary cases of TB diagnosed in the following five
yean
2.7
22
-
Immigration screening programs the Canadian experience Two Canadian studies have shown mat persons referred for PLS are at
increased risk of TB a". In Manitoba, a cohort analysis was conducted among foreign-
born individuals entering the province from 1981-85? 21,586 immigrants were referred for PLS- Acv tie
During that period 2.4% of
TB was diagnosed at the initial
medical evaluation in Manitoba in 2.8%. Over the five years of the study, the PLS group contributed 21.4% of al1 cases of TB among the foreign-bom. More than half of the cases of TB which developed in this group were diagnosed at the initial evaluation. The annual TB incidence rate was 0.32% (excluding those diagnosed with TB at the initial evaluation) for those referred for PLS and 0.07% for immigrants entering at the same time and not referred for PLS (relative risk = 4.5). Non-adherence was identified as a problem, with 19.5% not attending a single visit for medical evaluation in Canada. In British Columbia from 1982-85, 5.3% of ail Asian immigrants were referred for PLS and contributed 41% of al1 cases of TB among Asian immigrants. Active TB was
diagnosed in 14 persons at the tirne of their initial assessrnent - 1.5% of ail those assessed
The mean annual incidence of TB at follow up (excluding those
diagnosed at the initial evaluation) was significantly greater for immigrants referred for PLS than both the generai population with inactive TB (0.33% and 0.17% respectively) and immigrants who were not referred for PLS (0.08%
- relative risk=4) . The
immigrants with a normal baseline chest X-ray had an 8 fold higher incidence of active TB than the general population (0.08% vs.O.Ol% annual incidence). The annual rate of active TB among those with inactive disease and without previous treatment (1-64%) was significantly reduced by a history of adequate treatment (0.18%). Adherence was again identified as a potential problem as 13.5% of those referred for PLS did not present for medical evaluation.
23
In 1994, only 6.4% of cases of foreign-bom TB in Canada were known to have been placed on PLS m. However, the PLS classification was unknown for the rnajority
(77%).As well, the rnajority (79%) of foreign-bom persons with TB were detected as a result of syrnptorns. The effect of immigration TB screening programs has not been studied in a randomized control trial. Enarson suggested that the lower rates in persons who emigrate as compared to their compatriots who do not is evidence that chest Xray screening programs are successful l4. Although this may be tue, persons who immigrate rnay be in better generai health than their compatriots wtio remain in their country of birth.
2.8
Summary In summary, the nsk of active TB is much higher among persons originating
frorn countries with a high burden of TB compared with those already resident in Canada. This elevated nsk decreases with time but remains higher than the risk of persons bom in Canada 16. Within the immigrant population the burden is not evenly distn'buted and is dependent upon the burden of TB in the country of origin, age at immigration and time since immigration. Although the chest Xray should detect most cases of active TB which are of public health significance, it may miss a nurnber of active cases and rnay not be a cost-effective tool in populations where TB is not endemic? Population-based studies have demonstrated that the chest Xray can identify a group of individuals who are at I O -16 fold increased risk for the development of active TB. However, there is no standardization of chest Xray interpretation that can
be used to select such a high risk group. A high proportion of persons referred for PLS in Manitoba and British Columbia were diagnosed with active TB at the initial rnedical evaluation in Canada. lndividuals referred for PLS were shown to be at increased risk of developing active TB over five years of follow up.
24
A nurnber of concems regarding the efficiency of providing PLS for new amvals to Ontario have ansen. The cust to Ontario of delivering a program to provide PLS would be substantial. Currently, it is estimated that less than 2/3 of those referred for
PLS in Ontario adhere with the medical evaluation. An intervention which would reduce their future risk of TB is offered to few arnongst those who are evaluated. Few cases of TB among foreign-bom persons in Canada had been referred for PLS and most presented with symptoms. These findings suggest that PLS may be having a iimited
impact on reducing the incidence of TB in Canada. The question remains whether immigration screening programs should remain or possibly be expanded. Some
Amencan experts have rewmmended the expansion
" , whereas others suggest that
screening of individuals prior to immigration will have a minimal effect unless much
more restrictive and probably unacceptable travel and immigration policies are adopted as well 74. TWOCanadian studies which defined the risk associated with a referral for
PLS were whort in design and based on relatively few cases of TB=".
Other
contn'buting factors to risk of active TB were not analyzed, The curent study not only analyzed the relative risk of a referral for PLS in Ontario but also the contribution of other important factors such as age, time since ianding, world region of ongin and socioeconornic factors.
Chapter 3 Study Methods 3.1
Study objectives
To estimate the risk associated with a referral for PLS among persons bom outside of Canada. To analyze other factors which contribute to the nsk of active TB in this group.
3.2
Study design This was a population-based case control study. Al1 cases of active TB
diagnosed in the years 1994 and 1995 in the province of Ontario among persons born outside of Canada were included. Cases were enrolled from the Ontario Provincial TB Registry (Reportable Disease Information System - RDIS).
Eligibility criteria included
the availability of a landing record from the database at CIC and Ianding in Canada between 1986 and 1995. Exclusion criteria were: diagnosis with active TB before immigration rnedical evaluation; and age < 11 years at time of rnedical evaluation (no chest Xray screening is done for those e 11 year). Three independent sets of controls were used in the analysis. The controis were enrolled from a database at Citkenship and immigration Canada (CC). The case:control ratio was 1:4.
26
3.3
Study spproval The study was a collaborative effort between the Ontan'o Ministry of Health -
Public Health Branch and CIC. Databases from these two organizations were linked. Nominal infonnation was used for the Iinkage. The Ontario TB database RDlS (Reportable Disease tnformation System) is non-nominal. Therefore, nominal
information was obtained frorn the Health Units in the province of Ontario. Access to the data from RDlS was obtained from the Freedorn of Information and Protection of
Privacy Office, Ministry of Health, Ontario. No special agreement was required from CIC. CIC holds an authonzation from ail applicants undergoing an immigration medical examination regarding release of infonnation 'Yo a public health agency or a physician in Canada, if indicated" (Section C of the IMM 1017). it was considered that this pmvided the necessary authorization regarding the interface with the Ontario Ministry of Health data. Intemal data collection and review is considered to corne within federal jurisdiction. Nominal information collected from Ontario, was destroyed once linkage
with the CIC databases had been performed.
3.4
Data sources Two electronic databases provided the data for the case-control study.
However, an additional three databases were searched at CIC tu obtain background infonnation on al1 cases of foreign-bom TB. The first database, RDIS, is maintained by the Ministry of Health, Ontario. RDlS is used for al1 reportable diseases in Ontario. The Health Protection and Promotion act requires that al1 active and reacüvated cases of TB be reported to the Public Health Branch, Ontario Ministry of Health. The case definition
27
.Mycobacteriim tuberculosis CO mplex (e.g . M. tuberculosis, M. bovis ( excluding
BCG strain), or M. afncanum) demonstrated on culture from sputum, body fluids, or tissues or .Cases without bacteriological proof but with clinical symptorns or signs, radiological or pathoiogical evidence of active pulrnonary or nonpulmonary disease, preferably with: O
O
O
a positive tuberculin skin test (as defined by the provincial guidelines) andor demonstration of acid-fast bacilli in smears from sputum or other body fluids or tissues andfor response to antituberculous treatrnent.
Case information is entered eiectronicaily at the health units and transrnitted directly to the central database at the Ministry of t-iealth.
28
The second database, the Landed Immigrant Data System (LIDS), is maintained by the Program Data Division of the Immigration Service. LIDS is a batch file
assembled from the main administrative database - the Field Operations Support System (FOSS) and contains information from the record of landing (Imm 1000). Information from FOSS is batched to LIDS at the time the individual enters Canada. Therefore LlOS contains only legally landed immigrants to Canada and not visitors, students and refugees who have not obtained landed resident status. Additional databases (atso maintained by the Program Data Division of the Immigration Service) were searched for persons not found in LIDS. The databases included a visitor file, a refugee file and a file of persons landing in Canada between 1968 and 1978. Linkage procedures are described below. These searches were done to detemine how many cases had files at CIC but were not in LIDS. At the end of the linkage procedures, a manual search was perfomed using FOSS. Cases found in FOSS with records in
LlDS were added to the case file.
3.5
Data acquisition Cases of active TB were enrolled fr6m RDIS. The Ministry of Health - Public
Health Branch provided an electronic file which included al1 cases of tuberculosis bom outside of Canada and enrolled in Ontatio TB case registry for the years 1994-95. The 42 health units in Ontario were requested to provide nominal infomation (with the RDIS
unique identifier) for these cases. Nominal information was manually entered into the RDIS file. Manually entered data was independentiy reviewed for errors. Linkage with
the CIC database was performed to obtain infomation available in LlDS including PLS classification, age, gender, world region of origin, landing date, immigration classification, level of educational attainment and occupation at time of landing.
Linkage was perfomed at CIC following which al1 nominal information was destroyed.
29
Controls were afso enrolled from LIDS. Controls were randomly selected from
the same population from which the cases were derived (Le. those landing between 2986-95, were destined for Ontario and were at least at least 4 1 years of age at the
time of medical evaluation),
3.6
Methods o f data linkage and preparation Data Iinkage was perfomed by CIC personnel using strategies previously
developed for other Iinkage projects. Two methods of data Iinkage were used. The first was used to Iink the RDlS file to LIDS, visitor, refugee and old landing file. The second used the probabilistic search algorithm of FOSS. The first linkage procedure utilized a linkage routine created in FoxproTM.Personal identifiers used in the search included name, date of birth, gender. Search routines included 1) al! personal identifiers; 2) date of birth and gender and 3) name and gender, AI1 likely matches were confimed rnanually using additional information (country of birth and time of amval in Canada). Following search of LIDS the same procedure was used for the visitor, refugee, and old landing database. Files from RDlS without a match in LIDS,visitor, refugee or old landing database were then searched for in FOSS using the search algorithm of this database. The algorithm is probabiiistic and uses a scoring system based on sumame, first name, date of birth, gender and country of birth. As a result of this search a number of additional files in LIDS were found and these were added to the case file. The databases were merged using the RDlS unique identifier and stored using
the relational database: Microsoft Accessm. Data was analyzed using SPSSTM for windows version 7.s7=.
30
3.7
Method of control seiection Three control groups were selected. The prirnary reason was to gain
experience in using matched and unmatched control groups. Controls were selected from the entire cohort of files in LlDS for which Ontario was Iisted as the province of destination and age at visa issue (time of medical evaluation) was at least 11 years. Controls are, therefore, a probability sample of the M o l e population from which the cases came. Previous studies at C C indicate that the majon'ty (over 95%) of individuals who immigrate to Ontario remain in this province, therefore it is unlikely thai sewndary migration would change the composition of foreign-bom perçons who had landed in Ontario. It was assumed that the controls did not develop TB. To verify that the controls did not receive a diagnosis of active TB in ttre province of Ontario it would be necessary to search for the controls in the provincial TB registry (ROIS). This is not possible because RDIS is non-nominal.
31
LlDS is stored as a number of files each containing a year of data. Al1 control groups were selected from LlDS using the SPSS random selection routine. A total of three groups of controls were selected: the first was matched for year of entry; the second on year of entry, age and gender; and the third was unmatched. Year of entry was chosen as a matching factor to atternpt to control for changes in the pattern of
immigration over the ten years of the study- Such changes may be associated both with referral for PLS and risk of active TB. Age and gender are ako potentially
associated with the development of active
TB^^^^ and a PLS r e f e r r a ~ "and ~ ~ were
ttierefore matched in one of the controt groups. For the first (rnatched) control group a random sample from each year of landing was drawn in a ratio of 4:l controls:case per year of landing. For the second (matched) control group the sample was again drawn
frorn each year, however in addition a rnatching variable was created using year of birth and gender. A randorn sample of those matched to cases was drawn in a ratio of 4:l for each case. For the third (unmatched) control group, a 5% randorn sample was taken from each landing year of LlDS for 1986 to 1995. These ten files were merged and a random sample corresponding to 4:i controls:cases of the merged file was taken. Utilizing three control groups provided experience in unmatched and rnatched
-
analysis, however, is relatively inefficient As the disease (TB) exposure (PLS) association was expected to be large one could predict that the unmatched design would be more efficient 'O. group would be preferable.
Therefore if one were to use only one an unmatched control
32
3.8
Sample size considerations A total of 602 cases were eligible for the case control analysis. Previous studies
suggest that the relative dsk of actv ie 10 3J?
TB associated with a referral for PLS is at least
Based on the unmatched analysis the 602 cases with 4:l controls give 80%
power to detect an odds ratio of 2 (assuming a=0.05,Psurv in wntmls of 3%). The power to deted an odds ratio of 4 (assuming al1 else equal) is greater than 99% ".
3.9
Variables Enrolled cases were compared to those not enrolfed using variables found in
RDIS. For the case control analysis the variables which were used were derived from LIDS. The variables of greatest interest were PLS, gender, age at landing in Canada, year of landing, and world region of origin. Additional factors which were analyzed included type of immigrant, marital status, occupation, estimated burden of TB in country of birth, and level of educational attainment at the time of immigration. Further variables from RDIS which were used were method of detection, site of TB disease and smear status.
Post-landing surveiUance (PLS) classification PLS classification is found in LlDS and is derived from the surveillance code found on the IMM 1000 fom. Diseases for which PLS is required under the curent regulations include inactive TB and syphilis. lt is estimated that greater than 90% of PLS ~Iassificationsare based on inactive TB. It is not, however, possible to distinguish the type of referral using the available data sources.
33
Age at landing This was cafculated from two variables found in LlDS: age at landing = land date - date of birth. The effect of age on the risk of active TB was analyzed using age both as continuous and categorical variable.
Gender Gender was available from LIDS.
Marital status Marital status was categorized into single, married, divorce or separated and widowed from LIDS.
World region of origin Worid region of origin was created using the country of birth field of LIDS. The classification schema divided the worid into regions. These regions included the five cuuntrÎes which were the most common countries of birth among cases of TB in Ontario in 1994-95 (Vietnam, Philippines, Somalia, lndia and China), with the remainder of the world being divided using the scheme devised by the World Bank. This scheme has been s h o w to be useful in other analyses
The regions include Sub-Saharan
Africa, India, China, Other Asia and islands, Latin Arnerica and the Canbbean, Middle
East crescent, formeriy socialist econornies of Europe (FSE) and established market
34
Year of Ianding The year of landing is the year which the immigrant received landing status in Canada and is derived from LIDS. For persons who applied for landing status from outside of Canada, the time of landing will usually be the same as the time of amval.
For persons who apply from within Canada the time of arriva1 may precede the time of landing.
Burden of TB in country of birth and country of last permanent residence Two international sets of measures of disease burden exist for TB. The burden of TB in country of birth was created using the LlDS country of birth variable. This variable was recoded into two numeric variables: 1) World Health notification rate for
1991 (midpoint of the study) World Bank for 1990
and 2) estirnates for annual incidence of TB by the
World region of origin and estimated TB burden in country of
birth were highly correlated (lambda coefficient = 0.36,p 70 Male gender Land yead
1.2 (0.9-1.7) 1.6 (1.2-2.0) reference 0.9 (0.6-1.2) 1-0(0.7-13) 1.6 (1.1-2-4) 1.7 (1.0-2.7) 1.0 (0.8-1.2)
-
7.0 (4.7-10.5)
-* -
-
-
5.2(3.6-7.5) 1.3 (1.1-1.8) 1.4 (1.1-1.8) reference 0.8 (0.6-1.2) 1.0 (0.7-1.6) 1-2(0-8-1-8) 2.4 (1.4-4.0) 1.0 (0.9-1.2)
-
2.8 (1-6-51) 2.0(1.1-3-4) 2.0 (1.1-3.5) 2.3 (1.2-3.9) 2.7 (1.6-4.6) 3.5 (2.2-5.8) 3.2 (2.0-5.3) 4.3 (2.6-6.9) 3.5 (2.2-5.8) 1994 *where adjusted odds ratios are not provided, u : factor was controlled for in the design freference category 1995
-
-
-
-
4.7 The effeet of age and gender on N k of active TB When analyzed as a wntinuous variable, no association was seen between risk of active TB and age at time of landing. However by dividing age at time of landing into decades (Table 7) or deciles (Figure 5) bimodal relationship is seen between age and
risk ofactive TB (figure 5 and table 7). The adds ratios presented in figure 5 are
adjusted for PLS,gender, worid regian of origin and for year of tanding. Two age groups at increased risk of active TB were identified, the first group being those from twenty to thirty years at landing and the second those greater than 70 ygars at landing.
No association was seen between n'sk of active TB and gender. The unadjusted and adjusted odds ratio for male gender was 1. tiowever, a stratified
analysis of the effect of age by gender reveafed that the age effect on rÏsk was
primarily seen among fernales (figure 5). The age effect among males was not
significant on rnultivariate analysis.
Table 7 Age and risk of active TB unadjusted odds ratio
adjusted odds ratio*
(95%CI)
(95% CI)
Age at time of landhg (years) 11.1 - 2 0 20.1 30 30.1 40 30.2 50 50.1 - 60 60.1 - 70
-
1.2 (0.9-1.6) 1.4 (1-1-1.8)
reference 0.9 (0.7-1.3) 1-3 (0.9-1.9) 1-6 (1.1 2.2)
-
>70 1 3.1 (L.9-5.0) *adjusteci for gender?year of landing, world region of ongis and PLS
Figure 5 Risk of active TB by age at landing stratified by gender
Age at landing (years) 'adjusted for PLS. A
d r e g i of ~ &gin and year of ianding
1-2 (0.9-1.7) 1.3 (1.0-1.7)
reference 1.0 (0.7-1.6) 1.0 (0.7-1.6) 1- 1 (0.7-1.7)
3-4 (1.9-6.0)
53
4.8 The effect of year of landing on risk of active TB The relationship of year of landing and nsk of acv tie
TB is s h o w in figure 6.
The reference year is 1995. The risk of being diagnosed with TB is not equal for al1 years. In effect each year is a cohort with different risk. Those landing in 1995 were at decreased nsk of active TB compared to other years. This is likely a result of them being in Ontario for a relatively short period of tirne, It is also possible that new immigrants may delay presentation to public health offcialsas a result of fear of immigration a u ~ i o f l t i e s or~as ~ a result of delays in obtaining health care insurance.
ln spite of the differences between the years, with respect to the potential to be diagnosed, a Iinear relationship is evident with the risk of TB- Those amving more recently were at increasing risk (except for 1995). Unexpectedly those amving in 1986 are at particulariy high risk. By excluding 1986 from the analysis the R* of the linear relationship is 0.79compared to 0.42 when 1986 is included. As the inclusion of 1986 did not have an impact on the risk estirnates for PLS, age and worfd region of on'gin in
the final mode1 it was included in the analysis.
Figure 6
Risk of active TB by year of landing
Year of ianding 'adjusted for age at landing. gender. world region ofarigin and PLS
4.9 The effect of world region of origin on risk of active TB and predictive power of a referrai for PLS World region of origin was a powerful predictor of active TB (table 8). Two very high risk regions of ongin were identified: Somalia with an adjusted (age, gender, PLS, year of landing) odds ratio of 81-5and Vietnam with an adjusted odds ratio of 31.5 using established market economies as the reference category. World regions of origin with intemediate risk (adjusted odds ratios 7-13) included Philippines, Indian, China, and other sub-Saharan Africa. Wotld regions with low risk (adjusted odds ratios 2-5) included other Asia, Latin America, Middle East and fomerly socialist economies of
Europe. The lowest dsk category was established market economies which include
Western Europe, United States, Japan, Australia and New Zealand.
Table 8 World region of origin and risk of active TB
1 unadjusted odds ratio
adjusted odds ratio*
World Region of Ongin Vietnam Philippines Somalia India China Other Asia Latin America Other Sub-Saharan Afiïca Middle East Formerly sociaiist economies of Europe Established G k e t economies e x m t Canada *adjusteci for age. gender and year of Ianding
1
reference
niference
Figure 7 Risk of active TB by world region of ongin Vietnam PhiIlipines somalia lndia china
.osier ~ s i a Latin Amerfca
'
û?her~frica
&iddie Eart World Region of Origin 'adiusted for age, gender, year of Ianding and PES
Former socialkt
kt
The relative risk associated with a referral for PLS was not equal for al1 worid regions of ongin (table 9). A PLS referral appeared to be fairiy effective in defining a high risk group for Asia, Latin Arnerica and the formerly socialist economies of Europe. However the PLS referral did not define a high risk group for those originating from Afnca or the Middle east Few cases of active TB arnong persons bom in Afnca (4/122 - 3.3%) or the Middle East (0/36- 0%) had been referred for PLS.
Table 9 World region of origin and the predictive power of a referral for PLS Unmatched controls Number (%)
Cases
Number (?%O)
Unadjusted odds ratio
World Region of Origiri Asia
3.8 (2.6-5.5)
referred for PLS not rderred for PLS
67 (17.4%) 3 17 (82.6%)
56 (5.3%) 998 (94.7%)
referred for PLS not referred for PLS
4 (3.3%) 118 (96.7%)
3 (2.2%) 135 (97-8%)
referred for PLS not derred for PLS
2 (6.9%) 27 (93.1%)
3 (03%) 374 (99-2%)
referred for PLS
O
4 (1.7%) 227 (98.3%)
1.5 (0.3-7.0)
Afiica
9.2 (1557.6)
Latin Amenca
-
Middle East 36 (100%) not referred for PLS Fomerly socialist economies of Europe referred for PLS 3 (15.0%) 17 (85.0%) not referred for PLS Esiablished market economies except Canada referred for PLS 1 (9.10/0) 10 (90.9%) not referred for PLS 1 *odds ratio associateci with a referral for PLS by world region of origin
17.9 (3.4-95.3) 3 (1.0%)
304 (99.0%) O 301 (100%)
4.10 Analysis of other factors 4.10a The effect of incidence of TB in country of birth on the risk of active TB The estimated incidence of TB (World Bank) in the country of birth was a strong independent predictor for active TB. The unadjusted odds ratio (95%CI) for estimated for 25-49.9/100,000; 3.0 (1-0-8.6) incidence of TB in country of birth was 3.2 (1 -0-9.6) for 50-99,9/i00,000 and 22.3 (8.2-60.1) for slOO/lOO,OOO (reference category c 25/100,000). To further analyze the relationship between incidence of TB in country of birth and risk of TB the former was divided into deciles and the adjusted odds ratios were calculated (controlling for PLS, age at landing, year of landing and gender). The results of this analysis are shown in figure 8. A linear trend is suggested, however, the It was observed mat three countries were largely data is dispersed (~~=0.51). responsible for positive outliers in the adjusted odds ratio - these were Vietnam,
-
57
Ethiopia and Somalia. When these three wuntn'es were controlled for in the analysis the relationship between n'sk of TB and burden of TB in country of birth was more clearly linear (~~=0.72). The relationship between the WHO notification rates and tisk does not dernonstmte the sarne linear pattern (~'=0.16).
Figure 8
Risk of active TB by burden of TB in country of birh
Burden of TB in COB (1100,000)" 'adjusîed forage at ianding. gender. PLS and year d landing
7 B burden by deciles (World Bank esbmate - Worid Health ~eport1993)
4.10b The effeet of immigration classification on the N k of active TB The immigration classification of cases and controls is detailed in table 1. The
group of refugees, designated class and family class/assisted relatives are overrepresented among the cases whereas cases were less Iikely to be in the other independent category than controls. The adjusted odds ratio for immigration category was calculated wnlolling for age at landing, gender, PLS, world region of birth, and year of landing. The results of this analysis are given in table 10. Using the other independent group as the reference category, farnily class and refugees were at a significantly increased risk of active TB (adjusted odds ratio 1.6 and 1.6 respectively).
There was a trend toward higher risk among the designated class. The other independent group includes independent not otherwise specified, entrepreneur, selfemployed, investors and Iive-in-caregivers. The breakdown within the other independent group was not different for cases and controls. The proportion of the other independent group comprised by each of these subsets for case and wntrols respectively is: independent not otherwise specified 53/73(73%) and 453/586(77%); entrepreneur 12/73(16%)and 59/586(10%);self-employed 0/73(0%) and 23/586(6%); investors 6/73(8%)and 351586 (6%); and live-in-caregivers 2/73 (3%) and 16/586(3%).
Table 10 Immigration category and risk of active T ES unadjusted adjusted odds odds ratio ratio (9S%Cr)
(95%CI)
Type of immigrant 2.2 (1-7-2.9) 1.6 (1.2 -2.2) F d y class 2.8 (2.0-3-8) 1.6 (1.1-2.4) Convention refuge & dependent 2.2 (1.5-3.3) 1.5 (0.9-2.5) Designabxi cIass reference teference Other independent? Retiredi 1.9(0.9-3.9) 1.4 (0.6-3.O) *adjusteci for age, gender, yeai of landing, PLS and worfd region of origin Ymcludes independents, entrepreneurs, self-employed, imrestors and live-utaregivers
59
4.10~ The effect of educational attainment at the time of immigration on the risk of
active TB The educational attainment of cases and controls is s h o w in table 1. In general the cases are less highly educated than the controls. Table 11shows the nsk estimates (adjusted for age at landing, gender, PLS, world region of origin and year of landing). A dose response is not clearly dernonstrated. Persons with greater than 12 years of education were at a decreased risk of actv ie
TB compared tu persons with less
than or equal to 12 years of education. However, persons with undergraduate and postgraduate education do not appear to be at a lesser risk than those who have nonuniversity education beyond i2 years.
TabIe I I Educationai attainment at time of immigration and risk of active TB unadjusted odds ratio
adjusted odds ratio*
(9S0hCI)
(9S%CT)
Educational qualifications 0-9 10-12
reference
reference
0.8 (0.7-1.1) 0.4(0.3-0.5)
0.8 (0.6-1.0) 0.4 (0.34.6)
>13yearsnodiploma,trade certincate, nonuniversity diploma 0.7(0.6-1.0) 0.7 (0.5- 1.0) SomeuniversityBA 0.4 (0.24.7) 0.5 (0.3-0.8) Some postgrad/MA/Ph.D. * adjusted for age at Ianding, gender, PLS, world region of origin and year of Ianding.
4.10d The effect of marital status at the time of immigration on the risk of active TB Cases were more likely to be single or widowed compared to controls (table 1). The estirnated risk associated with marital status is shown in table 12. After controlling
for age, gender, PLS, worfd region of on'gin and year of landing only those wtio were widowed appeated to be at increased risk.
Table 12 Marital status and r i s k of active TB unadjusted odds d o (95OhCI)
adjusted odds ratio* (95%CI)
Maritai status
1-3(1.1-1.6) 1.1 (0-9-1-5) derence reference widowed 2 -4 (1 -6-3 -6) 2 . 3 (1.4-3.5) -ed divorced/separatedI L.O(O.5-2.1) 1.9 (0.83.3) *adjusteci for age at landing, gender. PLS, world region of origin and yeâr of hdulgsingle
1
4.10e The effect of intended occupation at the t h e of immigration on the risk of
active TB The intended occupation of the cases and controls are detailed in table 1.
There is overrepresentation of workers, new workerç and retired persons (al1 CIC categories) among cases compared to controls and under-representation of
adrninistrators/entrepreneurs and persons in the science, religion, teaching and health related fields. The odds ratio adjusted for age, gender, year of tanding and TB burden in country of birth are shown in table 13. The two CIC categories of worker and new
worker were found to be associated with an increased risk of active TB whereas students, administrative/entrepreneur and science, religion, teaching and health related fields were relatively low risk categories.
Table 13 Occupation and risk of active TB unadjusted odds ratio (95%CI)
adjusted odds ratio* (95%C9
Occupation Worker (CIC) 1.8 (1.1-2.8) 1.3 (1.0-1.7) New worker (CTC) 1.2 (0.9-1.6) Homemaker (CIC) Student (CIC) reference Retired 1.7 (1-2-2.5) 0.6 (0.3-1-1) Administrahve(cCDO)/entrepre~e~ (CIC) 0.5 (0.34.8) Science, religion, teachuig and health (CCDO) 1.0 (0.7-1.4) Clencal, sales and service (CCDO) 1.2 (0-8-1.7) Fanning, forestry, processing, equipment operators (CCDO)( *adjusteci for age at landing,gender, PLS,world region of origin and year of Ianding
2.4 (1.34.2) 1.6 (1-0-2.5)
1.4(0.8-2.3) reference 1.5 (0.8-3.0) 0.8(0.4-1.7) 0.7 (0.4-1-3) 1.2(0.7-2.1) 1.6 (0-9-2.6)
61
4.1 1 Final model
The initial final model contained PLS, age, gender, world region of origin, year of landing, education, occupation, immigration category and marital status (table 14). Using a cut-off for elimination from the model of p>0.2 occupation and gender were
removed from the final model. Using the cut-off of p0.05 immigration category and marital status were also rernoved. The three covariafes which appear to make the most significant contribution to the final model are wotld region of origin, PLS and year of landing. Age and education were also important contributors.
62
Table 14 Risk factors for active TB among foreign-boni persons - Ontario 1994-95 adjusted odd ratio (95% CI)
4.0
PLS
(2.7-6.0)
Age at landing (years)
-2LL (df) 48.0 (1) 13.9 (6)
1.3 (0-7-2.3) 1.5 (0.9-2.51 1.2(0.8-2.0) 1.2 (0.7-2.0)
reference 0.9 (0.6-1.6) Z.S(l.3-5-0) 1.0 (1)
Gender male
female
1.0
(0.8-1.2)
reference 375.9(10)
World region of ongin
Vietnam Philippines Somalia hdia China m e r Asia Latin Amerka Other Sub-Saharan Afnca Middle East Formerly socialist economies of Europe Established market economies except Canada Educational qualifications S 9 pars >9years
23.6 (11.6-48.0) 12.6 (5.8-23.3) 65.5 (29.0-140.2) 9.6 (4.9-19.0) 6.2 (3.1-12.6) 4.9 (2.5-9.5) 1.8 (0.9-3.8) 11.3 (5.6-23.0) 4 - 1 (2.0-8.4) 1.7 (0.8-3.8)
reference
1.3 (1.1-1.7)
reference
Immigration category Famiiy dass Convention refuge & dependent Designateci dass Other independent Retired MaritaI status single mamed widowed divorWseparated Land year
1.5 (1.1-2.0) 1.6 (1-1-2-3) 1.3 (0.7-2.3)
reference 1.3 (0.6-2.9) 1.0 (0.8-1.4)
reference i -7 (1 -0-3.O) 1.7 (0.73.0)
P
Chapter 5 Discussion TB is the leading infectious disease killer worldwide 87. The risk of TB among persons who immigrate parallels the risk among persons in their country of origin
14-16
.
Among persons immigrating frorn a high TB burden country to Canada the risk decreases over tirne. However, compared to the Canadian bom population, the nsk of developing active disease remains elevated throughout life 1°-16. The number of persons who immigrate to Canada each year rose substantially in the Iate 1980's and continues at very high levels. In excess of 200,000 landed immigrants arrive each year. This is one of the highest proportional rates of immigration with alrnost i% of the population immigrating to Canada each year. Over 50% of those persons immigrating to Canada chose ontano as their province of destination 88. ln the last thirty years a shift in immigration patterns has been evident with a growing proportion and number of immigrants coming from areas which have a high burden of TB 8.1 4.70 . In Ontario, more
than 80% of persons with TB are born outside of Canada 13. This paper provides an in
depth study of TB among this group including an analysis of the contribution of the system of referral for post-landing surveillance (PLS) after immigration. Two previous Canadian studies
evaluated the impact of a referral for PLS. These studies were
conducted in provinces ottier than Ontario where unlike Ontario, centralized TB control
64
programs exist. The current study adds valuable Ontario-specific data which is more conternporary than the tvvo previous studies (done in the 1980's). As well the previous
studies were cohort in design each enrolling less than fifty cases of TB. These studies were limited in their power to anafyze the contribution of other factors on botb the risk of TB and the efficiency of a PLS referral. This study was the first to use databases
maintained by CIC,thus providing population-based controls.
Presently immigrants ta Canada are screened for TB. Such persons are screened with a history, physical exam and chest Xray. The objectives of the TB screening program are: 1) to identify persons with active TB and 2) to identify persons
at high n'sk for the development of active TB. Perçons identified with active
TB are
required to complete treatment prior to entry into Canada. Persons with abnomalities consistent with TB on chest Xray but with a negative sputum smear and culture or stable follow up chest Xray at three months are felt to be at an increased risk for the developrnent of active TB. Such persons are referred for post-landing surveillance
(PLS)after amval in Canada. They may enter Canada conditional upon their agreement to present thernselves to public health authorities after amval.
It is not known how many active cases of TB are detected each year by the
Canadian immigration screening program. Other studies suggest a rate of case
65
detection using a chest Xray screening program in the range of 0.04-0.8%
2830.ôû.65.66.89
The detection rate depends primarily upon the background prevalence of disease in
the screened community. Mass screening with chest Xray has become less comrnon with the decline of TB in developed countries. This tool may, however, be useful in populations with a high burden of TB 52.72 . In contrast to other countries, Canada screens al1 applicants for landed status independent of the burden of TB in the country of origin 28.
As a secondary goal, the Canadian immigration TB screening program identifies a group felt to be at high risk of reactivation of TB. These individuals are referred for medical evaluation (PLS) in Canada. Concems have ansen that a program for PLS
rnay not be effective, first because the group refenced is not tmly at high risk ,and second as a result of inwmplete operationakation. The present study focused on the first concem and addressed the question of whether a referral for PLS predicts the development of TB in Ontario.
The current study found that only 13% of foreign-born cases of TB (A 7% of pulmonary cases of TB) arnong landed individuals in Ontario had been referred for
PLS. Persons referred for PLS were at a 4 fold greater risk of developing active TB than those not referred. ln other studies the ability of a screening chest Xray program
66
to identify persons who will go on to develop active TB has been highly variable. In Manitoba, immigrants referred for PLS contributed 22% (79/86)of al1 cases of TB among foreign-bom persons ". In British Columbia, immigrants referred for PLS wntributed 41% (21/51)". Southeast Asian refugees in Seattle identified by chest Xray contributed 24% (6/25)of al1 active cases and 50% (6112) of pulmonary cases of
TB
"?
In Europe the screening chest Xray identified 1929% of al1 future cases of TB Based on these studies mass chest Xray screening should be able to identify a
group which is at a 10-15 fold greater risk than persons with a normal chest Xray.
The risk of active TB arnongst persons referred for PLS in this study appears lower than expected based on previous studies. Potential reasons for this include: referral patterns which are relatively inefficient at identifying persons at high risk for TB; poor performance of the chest Xray in certain populations; enrollment bias; length of time since immigration; underdiagnosis of TB in the referred group and attenuation of risk by the use of isoniazid preventive therapy in the referred group.
The size of the relative risk of active TB expedenced by persons with an abnormal chest Xray compared to persons with a normal chest Xray depends on what is accepted as a positive chest Xray finding, history of active disease and history of chernotherapy. Persons with minimal findings on chest Xray such as a healed primary
cornplex or old pleurisy appear to be at lesser risk Vian those with more extensive findings such as apical fibrocalcific changes. As well, the chest Xray suffers significantly as a result of both inter and intra-reader variability in the interpretation of chest Xray changes
There is no consensus on the specific chest Xray
changes which define a high risk group. The current Canadian immigration screening program utilues chest Xray readers from around the wodd- These readers have not received forma1 training in the interpretation of chest Xrays for prediction of future risk of TB. It is likely that there is great variability in the identification of chest Xray abnomalities which will result in a referral for PLS. lt is, therefore, possible that many of the referrals are based on low risk chest Xray appearances such a healed pirnary comptex. Persons with a history of active disease and an abnormal chest Xray are at higher risk for reactivation than those with no such history who are detected solely as a result of chest Xray screening4'. Risk is decreased if the individual has received chemotherapy. Therefore, if many of those referred for PLS had previously received treatment, the observed risk would be lower ütan expected. Data regarding previous disease or treatment is not available from sources used in this study. It cannot be determined if these factors influenced the risk estimate.
68
Almost al1 referrals for PLS are for inactive TB, however, other conditions such as serological evidence of syphilis are referred for PLS. lt is estimated that less than 10% of refenals for PLS are for conditions other than TB. It is not possible to
d other conditions using currently available information. differentiate those r e f e ~ e for
Such misclassification woutd be expected to reduce the nsk estimate associated witb PLS, however, given that the proportion of those referred for non-TB conditions is
relatively srnall, the impact on misclassificationshould also be srnall.
This study provides evidence that the chest Xray rnay not be equally useful in identifying high tisk groups in all regions of the worid. Not only did world region of origin have an impact on the risk of active TB, it also impacted on how well the PLS program operated in the definition of a high risk group. Notably, the wodd region of origin which was associated with the highest relative risk of disease (Africa) was also
the region where PLS appears to be the least useful in defining a high n'sk group. Previous studies have noted a geographical variability in the rate of findings by screening chest Xray 33*59. This variation, however, has been judged to be of uncertain epidemiological significance
59,61
. Two possible explanations include high annual rates
of infection which result in a relative increase in the proportion of progressive primary
69
TB at the expense of reactivation disease (the latter being potentially identified by chest
Xray), and high rates of HIV coinfection which result in atypical chest Xray changes.
92
Enrollment bias couid result in an observed n'sk which is lower than expected. In this setting persons referred for PLS would be preferentially excluded from the case group or included in the control group. lt would not appear that cases of TB exduded from the study group were more likely to have been referred for PLS. A total of 602 cases were included in the analysis, which represents 45% of al1 cases of foreign-bom persons diagnosed with TB in the petiod of study (1994-95). The enrolled cases were generaliy representative of the foreign-born population of TB patients in Ontario in 1994-95. Cases were excluded if they were not Iinked to the C
C database, if they had
not received landed status at the time of diagnosis, or if they landed in Canada before
1986. Most cases (90%) which had anived in Canada since 1977 were Iinked to the
C C database, therefore, Iinicage limitations are unlikely to be a cause of enrollment bias. In British Columbia 31% of cases of TB among foreign-bom persons occurred in individuals who were not legally landed at the time of diagnosis 14. In this study 24% were not legally landed at the time of diagnosis and were excluded from the study. However, it is not expected that these would be more likely to have been referred for
PLS as most of them would not receive a screening evaluation. As controls were
70
randomly selected from a population based sample it is aIso unlikely that enrollment bias is a significant factor for the mntrol group.
It is recognized that risk of development of active TB decreases over time amongst persons immigrating from a high prevalence to a low prevalence country 10.16.34.71
. One study suggested that this decrease is concentrated among persons who
have abnormal chest Xray
" at the time of immigration.
Therefore if many of the cases
in this study had amved remotely one would expect the nsk associated with PLS to be less than if they had am'ved more recentiy. The study enrolled cases diagnosed in a two year period but allowed thern to have landed in Canada over a ten year pen'od.
However, over 75% had amved within five years. As detailed below tirne since immigration was independently associated with risk of active TB. Those referred for
PLS were diagnosed with TB eariier than those not referred. However, a stratified analysis did not demonstrate a difference in the risk associated with a referral for PLS between those arriving within or greater than 5 years. It is unlikely that time since amval impacted significantfy on the risk estimate for PLS.
Poor adherence with PLS may result in the underdiagnosis of actvi e
TB arnong
those referred for PLS. This was identified as a limitation of the PLS program in two
Canadian studies ? ln Ontario, it would appear that the situation is wone.
71
Prelirninary studies done by the Ministry of Health indicate that approximately 3367% of panons referred for PLS actually present to public health authorities ". This may result in a relative underdiagnosis of actv ie
TB in the referred group compared to
programs which have higher rates of follow up. This, in tum would be expected to decrease the observed risk estimate associated with PLS in this group.
Greater utilization of isoniazid preventive therapy in the group referred for PLS compared to the group not so referred would be expected to decrease the relative risk of TB experienced by the former group. Based on preliminary studies in Ontario, it is estimated that the utilization of isoniazid preventive therapy is low (less than 10% of
those referred for PLS who receive a medical evaluation in the province). As a result, the risk estimate for PLS referral is unlikely to be significantly affected by isoniazid
preventive therapy.
Other factors which were predictive of active TB (in addition to referral for PLS) were age at landing, worid region of origin, year of landing and educational attainment.
Age has been shown to influence the risk of active TB. In Montreal a birnodal pattern has been observed in incidence rates among foreign-bom individuals
11
. Two
peaks of increased Rsk were evident , the first being arnong those 15-29 years of age at landing the second among the elderiy. Among foreign-bom persons in the United
72
States a predorninance of cases among persans in the 15-29 year age group was
observed
The curent study also demonstrated that age at landing was a risk factor
for TB. Two age groups were identified which were at an increased risk. These two groups included those 20-30 years old and those greater than 70 years. The pattern rnay be a reflection an overiap of the pattern seen in high prevalence cornrnunities
wtiere cases are concentrated in the young, peaking at 25-34 years and the pattern seen in low prevalence cornrnunities where cases are concentrated in the elderiy. The
peak in the younger group may be a refiection of an increased risk of acquiBng infection f r m infancy ta early adult (ifeand time since infecüon as a high proportion of infections arnong young people are likely to have been recent
In low burden
countn'es the annual risk of infection is currentfy low, however, was usually higher in the past. As a result a whort phenornenon can be observed. In this setting, latent infection is rnuch more common at older ages than among young penons 'O. This, in
h m , results in a higher risk of active disease arnongst the elderiy. This decline in annual risk of infection may not occur in high burden wuntnes. As well life expectancy becomes a major deteminant of the Iifetime risk of developing tuberculosis arnong tuberculin reactors 41. As a result in high TB prevalence countries where, for the most
part, Iife expectancy is relatively short, the second peak in the elderiy may not be
73
obsenred. The Amencan study attributed the predominance of cases among young persons to the large number of persons in this age group who immigrate. This study suggests that there may also be a higher risk of active TB in this group.
This study use age at landing as opposed to age at diagnosis to evaluate the impact of age on risk of active TB. This is a limitation when it cornes to the
developrnent of policy designed to target high risk groups (Le.do we target them at entry or following settlernent in Canada?). Given mat the median time to the development of active TB is relatively short (3.8 yearç) it appears that a similar age relationship between active TB and tirne of diagnosis would be found with the age at diagnosis being slightiy older than age at landing. This is supported by the finding in Montreal that the highest incidence age group arnong foreign-bom persons is 15-29
yean ". The effect of gender on risk of TB should be divided into risk of infection and n'sk of active disease once infected. Generally males appear tu have a higher risk of
being infected than fernales 41.76 . This risk difference is first detectable dunng adolescence where behavioral differences may result in different rates of exposure. Once infected it appears that women of reproductive
age are at higher risk than men.
Later in life this difference is reversed 7s78. Undernotification of women may contribute
to the male predominance of TB notifications in developing countnes 76. Overall gender did not affect risk of active TB in this study. However, when the effect of age at the time of landing is analyzed by gender it is seen that the age effect is seen almost entirely among fernales, whereas there is littie variation in the age effect amongst
males. The rate of disease among immigrants is greatest within the first few years after immigration when irnmigrating from a high burden to a low burden country 6*10"'*44971. The findings of this study are consistent with previous studies. In this study risk of active TB declined with an increased duration of residence in Canada. The observed pattern is Iikely a result of the risk of active TB being highest shortly after infection 6.41 M.71
. Once persons have landed in a low burden country the risk of new infection
can be expected to decline significantiy. Alternative explanations for this finding
include 1) entry of persons with active TB into the country and 2) the act of immigration itself could be a risk factor for TB.
Persons who have recently immigrated may be less likely to present to health officiais as a result of fear of immigration difficulties
In the curent study, persons
in the five years prior. This wtio amved in 1995 were at Iower risk than those a ~ v i n g
may be explained by those persons being in Canada a short period of time and
therefore being less likely to be diagnosed with TB. Altematively it could suggest a delay in diagnosis among recent immigrants. Such a delay could be exacerbated by wrrent policy which does not permit the attainment of heaith insurance average for the first 90 days of residence in Ontario.
This study identifieci factors which were associated with time to diagnosis. Persons referred for PLS were diagnosed sooner after amval than were those not referred. This may simply be a result of the higher n'sk in this group, although early diagnosis as a result of the PLS medical evaluation is also a possible explanation. Other factors associated with time to diagnosis (on rnultivaflate analysis) were refugee status, age at landing and gender. Refugees, younger persons (at time of Ianding) and females presented earlier than non-refugees, older persons and males. Factors which were not associated ~ Ï t time h to diagnosis were TB burden in country of birth, fevel of education and site of disease, Certain groups may be more likely to be diagnosed late either as a resuit of deiays in presentation to health care providers or in delays in diagnosis by the providers. Delays in diagnosis, in tum, may result in increased transmission of TB in Canada.
An important Rsk factor both for TB infection and disease is world region of
ongin '0.1428" . This relationship was again demonstrated in the current study. World
regions of ongin which were associated with a particularly high risk of active TB included Vietnam, the Philippines, Somalia, India and other sub-Saharan African
countnes. These countn'es/regions al1 have a high burden of TB- Persons bom in established market economies were at much fesser risk compared to those originating
in other regions of the worfd. Persons from Vietnam appear to be at a similar tisk compared to their compatriots immigrating to the United States while persons immigrating to Canada from the Philippines may be at lower nsk than their compatriots immigrating to the US and those immigrating from sub-Saharan A f r h to Canada at higher risk than their compatriots imrnigrating to the US Io.For Vietnam the adjusted odds ratio in this study (adjusted for age with established market economies as the reference categories) versus the relative nsk in the American study (age-standardized rate country xlage-standardized rate established market economy) was 33.6 vs. 34.7; Philippines 13.4 vs. 18.0 ; China 7.4 vs. 9.3; sub-Saharan Africa 25.7 vs. 12.9;Middle
East 4.4 vs. 4.4; and former socialist economies 1.8 vs. 2.4. The difference between the US and Canada from the perspective of sub-Saharan Africa may be a result of higher rates of immigration from Somalia to Canada.
Afmost al1 cases arose from persons originating from a country with an incidence rate of greater than 100 per i00,000 population in their country of origin .
77
During the ten years of this study an increasing Iinear trend in the mean incidence of
TB in country of bitth was noted for controts. This suggests that ongoing changes in immigration patterns are occumng with recent immigrants coming from increasingly higher TB burden countn'es. This, in tum, would be expected to result in increasing numbers of TB cases assuming numbers of pesons imrnigrating to Canada remains the same.
Estimating risk as a function of incidence of TB in country of birth is subject to the shortcomings in the measurement of TB incidence. Such measurement requires an intact surveillance system which rnay not exist or be possible in many developing countries. The WHO notification rate suffers from serious limitations as health information systems in developing countries are too incomplete to provide rneaningful information on the incidence or mortality of tuberculosis M. A second estimate of TB incidence was used in this study
- the Wodd Bank estirnate of the incidence of TB in
the country of birth. This estirnate attempts to correct for under surveillance in a
countrysz. The current study suggests however that the Wodd Bank estimates may be low for such countries as Somalia, Vietnam and Philippines. lt is also possible that other risk factors were responsible for the unexpectedly elevated nsk in these groups
which were not contrulled for in the anaiysis.
78
Other studies have demonstrated the impact of socioeconornic status on risk of TB "".
A social class gradient exists for rnortality frorn TB
Unemployment has
been associated with incidence of TB I6and prevalence of tuberculous infection with incorne ". The inverse association of prevalence of infection with income has also been obseived among foreign-bom panons in Montreal 47. AS well, education has been associated with morbidity and mortality ". This study also provides evidence of the impact of socioeconomic factors on the risk of active TB among foreign-born
persons to Ontario. The socioeconomic indicators analyzed in this study included immigration classification, educational attainment, and occupation. A relationship between risk of active TB was obsewed for all factors on univanate analysis. Only education remained independentiy associated with risk of active TB after controlling for
other socioeconomic factors. Measunng the impact of socioeconornic factors among immigrants may be subject to a number of limitations. First the indicators were al1 measured at the time of immigration. They, therefore, reflect social standing in the country of on'gin- Other important factors such as racefethnicity, gender and acculturation are likely to impact
on the social standing in Canada. Social class was estimated only on an individual Ievef and did not take into account household and neighborhood social standing. 80th
79
of these factors rnay have an independent impact on the healffi of individuals ". To identify tnie risk factors for diseases when planning prevention prograrns, the effect of s~cioewnomicstatus rnust be differentiated from that of race or ethnicity and be considered separately as a contributor to rnorbidity. Further, different cultural heritage and social structures may affect how interventions are perceived or used and this must also be taken into account M e n planning programs 95.
TE3 is a more prominent health issue among persons who may be least able to
access health care. Persons who immigrate may rnaintain a social class parallel with their social class in their country of origin. This may have an impact on acculturation and access to health care. Control tools wiil need to be educationally appropriate. As welI, specific strategies will need to be developed to access and adequately treat persons of low educational attainrnent.
In practical ternis, the system of referral for PLS appears to play some potential role in TB wntrol activities in Ontario, however, the contribution of this program is limited. Further studies need to be done to detemine if better results could be achieved either by improving the quality of chest Xray interpretation or by changing the way immigrants are selected for PLS. This study demonstrates that world region of origin plays a very strong role in detemining future risk of TB. Therefore, world region
80
of origin could be used to target specific high risk groups for expanded suweillance for
TB. For example, two identified countries of high nsk (Sornalia and Vietnam) in this
study accounted for 29% of al1 cases of TB (27% of cases of putmonary TB), whereas only 5% ofimmigrants each year corne from these countries- Such a stiategy could
fows on such a relativeiy small group while having a potentially large impact on future
cases of Tl3 in Ontario. This study also suggests that efforts should be targeted at young persons less than 30 years of age, particularly fernales. Strategies to access such groups could include the incorporation of TB screening into pnmary health care services provided for immigrants
"or prenatal programs ".
The Canadian immigration screening program is not universal. Only persons who are applying to be legally landed and certain visitors receive an immigration
medical evaluation. As a result some of the foreign-bom persons who go on to develop TB will not have been screened. In this study 24% were not legally landed. lt is not known if these persons are at increased risk. Further studies should be done to evaluate the nsk of active TB among foreign-born persons not legally landed in
Canada. If such persons are at high risk of active TB strategies will need to be developed to idenüfy and prevent disease in mis group of persons.
81
Currently PLS is not operationalized in Ontario. Although it rnay appear that operaüonalizationof a such a program may not be justifiable, other sfudies have demonstrated that a high proportion ( 4 53%) of persons referred for PLS were diagnosed with acvtie
TB at the initial medical evaluation in Canada- Under current
conditions it is not possibfe to estirnate this proportion in Ontario, prirnarily as a result of poor follow up of those referred. It will be important to detemine the proportion of persons referred for PLS who have active disease at their initial visit prior to making a decision that this service should be discontinued.
Curent Canadian guidelines recommend isoniazid preventive therapy for tuberculin-reactive immigrants younger than 35 years of age and those of any age with
additional risk factors for reactivation such as fibrotic lesions on chest xrayn.
Persons
c 35 years of age are good candidates for the targeting of preventive strategies, firstiy
as they rnay be ai increased risk of active TB and secondly as they benefit from preventive therapy if infected with TB with a relatively low risk of isoniazid associated toxicity. A study by Nolan suggests that such a strategy will significantly reduce the number of TB cases amongst high risk immigrants to a low prevalence country 30. In this study the majority of cases of TB were less than 35 at the time of immigration. Providing isoniazid prevenüve therapy to this group rnay be an effective way of
reducing future TB in Canada. By selecting migrants from countries with a high prevalence of TB who are less than 35 at the tirne of immigration the target group would comprise approximately V 3 of al1 individuals who amve in Ontario. Assuming 100,000 persons arrive each year in Ontario, such a strategy would attempt to access 33,000 penons per year. Assuming 113 of them are infected with TB 6-47 and
candidates for preventive therapy and that 70% of those eligible complete a course of preventive therapy and that preventive therapy reduced future risk of TB by 75% that future risk of TB without a preventive intervention is 5%
and
" such a strategy could be
expected to prevent 261 future cases of TB each year. A program which enrolled only
10% for preventive therapy would be expected to prevent 41 cases of TB per year.
83
TB is an important heajth issue for rnany persons who immigrate to Canada. It is essential that provincial TB control programs develop an good understanding of the epidemiology of TB among foreign-bom persons. This understanding will be needed not only to provide adequate health care to this population but also to minimize the importation and transmission of TB in Canada. Globalization, the persistence of the TB epidemic woridwide and the devastating complications of multiple dmg resistant TB are all a reality. To combat the potential for costly and life threatening outbreaks in this province it will be necessary to optimize TB control acüvities. lmproved efforts to prevent disease and transmission of TB should be considered a public health prionty.
Chapter 6 Conclusions International migration imposes on host muntnes the responsibility to develop an understanding of migrant health issues. TB is a prominent health care need among
immigrants to Canada. This study has shown that persons who are identified as high risk by the immigration screening program are at a four fold increased risk of TB. However this group conûibuted only 13% o f future cases of active disease in Ontario. The study also demonstrated that within the group of immigrants risk of active TB is not
evenly distributed. Risk was also associated with worid region of origin; age at immigration; time since immigration; and educational attainment. Such risk factors may be useful in the design of targeted TB control activities.
Further study is required addressing the nsk of TB among foreign-bom persons who are not legally landed and, therefore, have not received an immigration medical evaluation. As well, a study should be done to estimate the proportion of persons referred for PLS who have active TB at the initial visit Certainly if the proportion is as high as it has been in other Canadian studies (1.53%) it will be difficult to justify a
decision not to provide medical evaluation of this group after amval in Canada.
85
Further studies are needed to examine current patterns of referral for PLS - with the primary question being: can we use the chest Xray to define a higher risk group. This in tum woutd increase the efficiency of a provincially-based PLS program.
The most powerful predictor of TB was worid region of origin. Targeting a screening program based on worid region of origin could substantially decrease the number of persons to be screened. However, a focused screening program is subject to significant limitations. The primary limitation is that such a strategy requires good data regarding n'sk of TB in most countries- This information may not be available, particulady for countnes which suffer from the highest burden of TB.
Strategies will necessarily concentrate on persons who have recently amved, however, those who have been in the country more than five years will continue to contribute approximately 112 of cases of TB among foreign-bom penons 'O.
Accessing
persons who have been in the country for a long time may be more difficult than accessing new amvals. Further research will be needed to identify factors which will be
useful in targeting groups who continue to be at high risk many years after immigration. The most efficient ~ontrolof TB demands the concentration of activities on
-
specific narrow targets the high risk groups. Early diagnosis of reactivation and, if possible, prevention of disease in high risk groups is an important cornponent of TB
control activities. Bamers to adherence with PLS will need to be identified and removed. As well it wili be necessary to increase rates of utilkation of preventive treatrnent, partiwlarly among persons less than 35 years of age.
Immigrants to Canada have unique health needs mmpared to persons bom in Canada. TB is an important health issue among many immigrants- However, this group is heterogeneous both frorn a n'sk and cultural perspective.
The TB control
prograrn will need to incorporate this heterogeneity into its control activities. The immigration screening prograrn may make some contn'bution to TB control in Ontario. However, the contribution appears to be seriously limited. Further study will be required to address whether a program for PLS is justifiable. The curent study suggests other ways that TB control activities can be targeted to high risk groups. 1
References 1. Kochi A. The global tuberculosis situation and the new control strategy of the World Health Organization. Tuberde i991; 72:i-6.
2. Raviglione MC, Snider DE, Jr., Kochi A. Global epiderniology of tuberculosis. Morbidity and mortality of a worldwide epidemic. JAMA 1995; 273:220-226.
3. Dolin PJ, Raviglione MC, Kochi A. Global fuberculosis incidence and mortality during 1990-2000. Bull WHO 1994; 721213-220. 4. Billo H, Personal communication1997; Screening for tuberculosis among foreign-
bom in Switzedand.
5. Binkin NJ, Zuber PLF, Wells CD, Tipple MA, Castro KG. Overseas screening for tuberculosis in immigrants and refugees to the United States: current status. CID 1996; 23:1226-1232. 6. McKenna MT, McCray E, Onorato 1. The epiderniology of tuberculosis arnong foreign-bom persons in the United States, 1986 to 1993. N Engf J Med 1995; 332: 1071-1076.
7.Raviglione MC, Sudre P, Rieder HL, Spinaci S, Kochi A. Secular trends of tuberculosis in western Europe. Bull WHO 1993; 71:297-306. 8. Rieder HL, Zellweger J, Raviglione MC, Keizer ST, Migliori GB. Tuberculosis cuntrol in Europe and international migration. Eur Respir J 1994; 7:1545-1553. 9. Stehr-Green JK. Tuberculosis in New Zealand, 1985-90. New Zealand Med J 1992; 1O5:3OI-303.
10. Zuber PLF, McKenna MT, Binkin NJ, Onorato IM, Castro KG. Long-terni risk of tuberculosis among foreign-bom persons in the United States. JAMA 1997; 278:304307.
1 'i.Rivest P, Tannenbaum T,Bedard L. Epidemiology of tuberculosis in Montreal. Can Med Assoc J 1998; 158:605-609. 12.Cowie RL, Sharpe JW. Tuberculosis among immigrants: interval from amval in Canada to diagnosis. A 5-year study in southem Alberta. Can Med Assoc J 1998; 158~599-602.
13. Kerbel D. Epidemiology of tuberculosis in Ontario 1995. PHERO 1997; 8: 14. Enarson Dl Ashfey MJ, Grzybowski S. Tuberculosis in Immigrants to Canada. A study of present-day patterns in relation to immigration trends and birthplace. Am Rev Respir Dis 1979; 119: 11-18.
88
15. Orr PH, Hershfield ES. The epiderniology of tuberculosis in the foreign-bom in Canada and the United States. In: Reichman LB, Hershfield ES, editors. Tuberculosis: a comprehensive international approach. New York: Marcel Dekker, inc. 1993:531-548. 16. Enarson DA, Sjogren t, Gnybowski S. Incidence of tuberculosis among Scandinavian immigrants in Canada. Eur J Resp Dis 1980; 61:13Q-M2.
17. Medical Research Council Tuberculosis and Chest Diseases Unit. The geographical distribution of tuberculosis notifications in a national suwey of England and Wales in 1983. Tubercle 1986; 67: 163-187, 18. Menzies D, Chan ClVissandjee B. Impact of immigration of tuberculosis infection arnong Canadian-bom schoolchitdren and young adults in Montreal- Am J Respir Crit Care Med iQ9-l.
19. Van Soolingen D. Molecular epidemiology of tuberculosis in a low incidence country: a nation wide study on transmission of tuberculosis between immigrants and native population in the Netheriands. In: Use of DNA fingerprinting in the epidemiology of tuberculosis. University of Utrecht, 1996:178-195 20. DeRiemer K, Chin DP, Schecter GF, Reingold AL- Tuberculosis among immigrants and refugees. Arch lntem Med 1998; 158:753-760.
21. Barnes Pl El-Hajj HlPreston-Martin S, Cave MD, Jones BE, Otaya Ml et al. Transmission of tuberculosis among the urban homeless. JAMA 1996; 275:305-307.
22. Edlin BR, Tokars JI, Grieco MH, Crawford JT, Williams J, SordilIo EM, et al. An outbreak of multi-drug resistant tuberculosis arnong hospitalized patients with acquired immunodeficiency syndrome. N Engl J Med 1992; 326:l 5 W l 5 2 l . 23. Coronado V, Beck-Sague C, Hutton M, Davis BJ, Nicholas Pl Villareal C,et al. Transmission of multi-drug resistant Mycobacterium tuberculosis among persons with human irnmunodeficiency virus infection in an urban hospital: epidemiologic and restriction fragment polymorphism anaIysis. J Infect Dis i993; 168:1052-2 0%. 24. Di Pem G,Cruciani M, Danzi MC, et al. Nosocornial epidemic of active tuberculosis among HlV-infected patients. Lancet 1989; 2: 502-1504.
25- Jochem K, Tannenbaurn T, Menzies D. Prevalence of tuberculin skin test reactions among prison workers. Can J Public Health 1997; 88:202-206. 26. Schwartzman K, Loo V, Menzies R, Paztor J. Tuberculosis infection among health Gare workers in Montreal. Am J Respir Crit Care Med 1996; 155.
27.Thomas RE, Gushulak B. Screening and treatrnent of immigrants and refugees to Canada for tuberculosis: Implications of the expenence of Canada and other industrialized countries. Can J Infect Dis 1995; 6:246-255. 28. Bonvin L,Zelhveger JP. Mass miniature X-ray screening for tuberculosis among immigrants entering Wtzerland. Tubercle & Lung Disease 1992; 73:322-325.
89
29. Blum RN, Polish LB, Tapy JM, Catlin BJ, Cohn DL. Results of screening for tuberculosis in foreign-bom persons applying for adjustment of immigration status. Chest 1993; 1O3:d6iO-l674. 30. Nolan CM, EfarthAM. Tuberculosis in a cohort of Southeast Asian Refugees. A five-year surveillance study. Am Rev Respir Dis 1988; 137:805-809. 31. Zuber PLF, Binkin NJ, Ignacio AC, Marshall KL, Tribble SP, Tipple MA, et al. Tuberculosis screening for immigrants and refugees: diagnostic outcornes in the state of Hawaii. Am J Respir Crit Care Med 1996;154: 151-155. 32. Standards Cornmittee (Tubercuiosis) of the Canadian Thoracic Society. Canadian tuberculosis standards. 4th ed. The Canadian Lung Association, 1996. 33. Orr PH, Manfreda J, Hershfield ES. Tuberculosis surveillance in immigrants to Manitoba. Can Med Assoc J 1990; 142453-458. 34- Wang JS, Allen €A, Enarson DA, Grzybowski S. Tuberculosis in recent Asian immigrants to British Columbia, Canada: 1982-85. Tubercle 1991; 72:277-283.
35. lnto the 21st century: A strategy for immigration and citizenship. 1995; Ottawa: Citizenship and Immigration Canada. 36. Statutes of Canada: Immigration Act. 1976; Chapter 52:
37. Guidelines for the investigation of individuats who were placed under surveillance for tuberculosis post-landing in Canada. Canada Communicable Disease Report 1992; Vol. i 8-20: 38. Heywood N, Personal communication 1998. 39. Naus M , Personal communication1998. 40. Murray CJL, Styblo K, Rouilion A. Tuberculosis. In: Jamison DT, Mosely WH, Measham AR, et al. editors. Disease control prionties in developing counb-ies.New York, NY: Oxford University Press (for the World Bank), l993:233-259.
41. Comstock GW, Cauthen GM. Epidemiology of tuberculosis. In: Reichman LB, Hershfield ES, editors. Tuberculosis: a comprehensive international approach. New York, NY: Marcel Dekker, i993:23-48. 42. Krieger N, Moss N. Accounting for the public's health: an inttoduction to selected papers from a U.S. conference on "measuring social inequalities in health". InternationalJournal of Health Semices 1996;26:383-390. 43. Mangtani P, Jolley DJ, Watson JM, Rodrigues LC. Socioeconomic deprivation and notification rates for tuberculosis in London dunng 1982-91. BMJ 1995; 310:963-966. 44. Rieder HL, Cauthen GM, Comstock GW, Snider DE. Epidemiology of tuberculosis in the United States. Epid Rev 1989; 11:79-98.
90
45. Anonymous. Tuberculosis and social ciass. Tubercle 1979; 191:194
46.Reichman LB, O'Day R. Tuberculous infection in a large urban population. Am Rev Respir Dis 1978; 705-714. 47. Menues R, Vissandjee B, Amyot D. Factors associated with tuberculin reactivity among the foreign-bom in Montreal. Am Rev Respir Dis 1992; 146:752-756. 48. Kn'eger N, Williams DR, Moss NE. Measuring social class in US public health research: concepts, methodologies, and guidelines. Ann Rev Pub Health 1997; 18:341378. 49. Grzybowski S, Fishaut H, Rowe J, Brown A. Tuberculosis arnong patients with vatious radiologie abnomalities, followed by the chest clinic service. Am Rev Respir Dis 1971; -104:605-608. 50. Nakielna EM, Cragg R, Grzybowski S. Lifelong follow-up of inactive tuberculosis: its value and limitations. Am Rev Respir Dis 1975; i 12765-772-
51.Cadman O,Chambers L, Feldman W, Sackett D. Assessing the effectiveness of community screening prograrns. JAMA 1984; 251:158O-I585.
52. Clancy L, Rieder HL, Enarson DA, Spinaci S. Tuberculosis elimination in the countnes of Europe and other industiialized countries. Eur Respir J 1991; 4:12884295. 53. Bass JB. The tubercuIin skin test. In: Reichrnan LB, Hershfield ES, editors. Tuberculosis: a comprehensive international approach. New York: Marcel Dekker, Inc. 1993: 54. Hsing CT. Chest microfilm versus direct microscopie sputum examination in tuberculosis control in the developing country. Diseases of the Chest 1967; 52648-
651.
55. Dankova D. Tuberculosis risk in pesons with 'fibrotic' lesions. Bull Int Union Tuberc Lung Dis 1973; 145-150. 56- Nyboe J. Evaluation of efficiency in interpretation of chest x-ray films. Bull WHO 1966;35535545.
57. Nyboe J. Results of the international study on x-ray classification. Bull Int Union Tuberc Lung Dis 1968; 41:115-124. 58. Birkelo CC, Chamberlain WE, Phelps PS, Schools PE, Zacks O, Yerushalmy J. Tuberculosis case finding. A cornpanson of the effectiveness of vanous roentgenographic and photofluorographic rnethods. JAMA 1947; 133:359.
59. Geser A, Thorup 1, Geographical variations in the prevalence of chest x-ray abnomalities detected in tuberculosis surveys in eight African wuntries south of the Sahara. 8uII WHO W67; 363301-820.
91
60.Kannisto M. Deteetion of pulmonary tuberculosis by mass RP suweys: a study of their effÏcacy. Scandinavian Journal of Respiratory Diseases - Supplementum 972; 8111-116. 61. Vdorld Health Organization. Expert cornmittee on tuberculosis. 1964;WHO.
62.Horwitz O, VVilbek E, Enckson PA. Epidemiological basis of tuberculosis eradication. Bull WHO 1969;41:95-113. 63. International Union Against Tuberculosis Cornmittee on Prophylaxis. Efficacy of various durations of isoniazid preventive therapy for tuberculosis: five years of follow-up in the lUAT trial, Bull WHO 1982;60:555-564. 64. Styblo K, van Geuns HA, Meijer J. The yield of actv i e case-finding in perçons with n i actv ie pulmonary tuberculosis or fibrotic lesions. A 5-year study in tubercutosis clinics in Amsterdam, Rotterdam and Utrecht. Tuberde 7 984; 65237-251. 65. Markey AC, Forster SM, Mitchell R, Larson E, Smith H. Suspected cases of pulmonary tubercuIosis referred from port of entry into Great Btitain, 1980-83.Br Med J
1986;2921378-379.
66. Pitchenik AE, Russell BW, Cleary T, Pejovic I, Cole ClSnider DJ. The prevalence of tuberculosis and d m g resistance among Haitians. N Engl J Med 1982; 307:162-i65. 67.Enarson DA. Active tuberculosis in Indochinese refugees in British Columbia. Can Med Assoc J 1984;131:39-42. 68.Centers for Disease Control and Prevention. Tuberculosis among foreign-bom persons who had recently arrived in the United States -Hawaii, 1992-1993and Los Angeles County, 7993.MMWR 1995;44703-70769.Cummings K, Mohle-Boetani J, Sciortino S, Royce S, Chin D. Idenüfying infectious tuberculosis (TB) cases among recent foreign-born (FB) entrants into California: limitation of immigration screening. Amencan Thoracic Society 1997;A222: Abstract. 70.Statistics Canada - Health Statistics Division. Tuberculosis Statistics 4994. 7 996; Ottawa, ON:Minister of Industry. 1 p. 71.British Thoracic and Tuberculosis Association. Tuberculosis among immigrants related to length of residence in England and Wales. Br Med J 1975;3:698-699.
72.Qualls N, Taylor 2,Binkin N, Saralya M, Nguyen C -A decision analysis of screening for tuberculosis among immigrants prÎur to amving in the US. International Union Against Tuberculosis and Lung Disease - North American Region Third Annual Conference 1998;Abstract. 73. lseman MD, Starke J. Immigrants and tuberculosis wntrol. N Engl J Med 1995; 332:1094-1095. 74. Hopewell PC. Immigrants and tuberculosis. N Engl J Med 1995; 333:667-668.
92
75. SPSS base 7.5 for Wndows. 1997; (7.5)
76.Holmes CB, Hausler HlNunn P. A review of sex differences in the epidemiology of tuberculosis. Int J Tuberc Lung Dis 1998; 2:-104. 77.Comstock GW, Livesay VT, Woofpert SÇ. The prognosis of a positive tuberculin reaction in childhood and adolescence. Am J Epid i974; 99: 131-137. 78. Gizybowski S, Allen EA. The challenge of tuberculosis in dec1ine:a study based on
the epidemiology of tuberculosis in Ontario, Canada. Am Rev Respir Dis 1964; 90:707720. 79, Comstock GW. Modem epidemiology of tuberculosis. Am J Epid 1975; 101:363383.
-
80.Schlesselman JJ. Case-control studies design, conduct, analysis. New York: Oxford University Press, 1982.
81.Epi lnfo 6. 1994; Centers for Disease Control and Prevention, U.S.A. and World Health Organization, Geneva, Switzerland.
82.Worid Bank. World Development Report i993. lnvesfing in Health. 1993; New York, NY: Oxford University Press. 83.Global Tuberculosis Programme. Global Tuberculosis Control - WHO Report 1997. 1997; Geneva: World Health Organization. 84. Manpower and Immigration. Canadian Classification and Dictionary 1971. Ottawa: Manpower and Immigration Canada, 1971.
of Occupations
85. Rothman KJ, Greenland S. Modem epidemiology. 2nd ed. Philadelphia: Lipincott Raven, 7998. 86. Asch SlLeake 6,Gelberg L. Does fear of immigration authonties deter tuberculosis patients from seeking care? West J Med 1994; 182373-376.
87. Murray CJL, Styblo K, Rouillon A. Tuberculosis in developing countries: burden, intervention and cost. Bull Int Union Tuberc Lung Dis 1990; 656-24. 88. Citizenship & Immigration Canada. facts and Figures: overview of immigration. 1994; Ottawa: Citizenship & Immigration Canada.
89.KrÎvinka R, Drapela J, Kubik A, Dankova DlKn'vanek Jl Ruzha Jl et al. Epidemiological and dinical study of tuberculosis in the district of Kolin, Czechoslovakia. Second report (1965-7 972). BuII WHO 1974; 5159-69. 90. Spnngett VH. Results of the international study on x-ray classification. Conclusions. Bull Int Union Tuberc Lung Dis 1968; 41A25-129.
93
91. Spflngett VH. Results of the study on x-ray readings of the AD HOC Cornmittee for the Study of Classification and Terminology in Tuberculosis. Bull Int Union Tuberc Lung Dis 1968; 41:loi-109.
92.Hopewell PC. Impact of hurnan imrnunodeficiency virus infection on the epidemiology, clinical features, management, and wntrol of tuberculosis. Clin Infect Dis 1992; 15540-547. 93-Powell KE, Meador MP, Farer LS. Foreign-bom persons with tuberculosis in the United States. Am J Public Health 1981; 71:lZ23-l227. 94. Styblo K, Rouillon A. Estimated global incidence of smear positive pulmonary tuberculosis. Unreliabifity of officially reported figures on tuberculosis. Bull lnt Union Tuberc Lung Dis 1981; 56:i 18-126.
95.Hinman AR. Disease prevention programs for racial and ethnic rninorities. Annals of Epidemiology 1993; 3: 185-192. 96. Fowler N. Providing primary health care to immigrants and refugees: the North Hamilton expenence. Can Med Assoc J 1998; l59:388-391. 97. Sent L, Ballem P, Pafuck E, Yelland L, Vogel AM. The asian wornen's health clinic: addressing cultural barriers to preventive health care. Can Med Assoc J 1998; 159:350354.
-
Appendix A Canadian Classification and Dictiooary of Occupations 1971 Major groups Managerial, administrative and related occupations Occupations in natural sciences, engineering and mathematics Occupations in social sciences and related fiefds Occupations in religion
Teaching and related occupations Occupations in medicine and health Artistic, literary, performing arts and reiated occupations Occupations in sport and recreation Clerical and related occupations Sales occupations Service occupations
Farming, horticultural, and animal-husbandry occupations Fishing, hunting, trapping and related occupations
Forestw and logging occupations Mining and quanying including oil and gas field occupations Processing occupations Machining and related occupations
Product fabricating, assembling, and repairing occupations Construction trades occupations Transport equipment operating occupations Material-handling and related occupations Other crafts and equipment operating occupations
