UBC Faculty Research and Publications

Predictors of hospitalization of tuberculosis patients in Montreal, Canada: a retrospective cohort study Ronald, Lisa A; FitzGerald, J. M; Benedetti, Andrea; Boivin, Jean-François; Schwartzman, Kevin; Bartlett-Esquilant, Gillian; Menzies, Dick Nov 15, 2016

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-12879_2016_Article_1997.pdf [ 486.96kB ]
JSON: 52383-1.0366833.json
JSON-LD: 52383-1.0366833-ld.json
RDF/XML (Pretty): 52383-1.0366833-rdf.xml
RDF/JSON: 52383-1.0366833-rdf.json
Turtle: 52383-1.0366833-turtle.txt
N-Triples: 52383-1.0366833-rdf-ntriples.txt
Original Record: 52383-1.0366833-source.json
Full Text

Full Text

RESEARCH ARTICLE Open AccessPredictors of hospitalization of tuberculosispatients in Montreal, Canada: a retrospectivecohort studyLisa A. Ronald1,2,3, J. Mark FitzGerald2,3,4, Andrea Benedetti1,5,7, Jean-François Boivin1, Kevin Schwartzman1,5,Gillian Bartlett-Esquilant6 and Dick Menzies1,5*AbstractBackground: Hospitalization is the most costly health system component of tuberculosis (TB) control programs.Our objectives were to identify how frequently patients are hospitalized, and the factors associated with hospitalizationsand length-of-stay (LOS) of TB patients in a large Canadian city.Methods: We extracted data from the Montreal TB Resource database, a retrospective cohort of all active TB casesreported to the Montreal Public Health Department between January 1996 and May 2007. Data included patientdemographics, clinical characteristics, and dates of treatment and hospitalization. Predictors of hospitalization and LOSwere estimated using logistic regression and Cox proportional hazards regression, respectively.Results: There were 1852 active TB patients. Of these, 51% were hospitalized initially during the period of diagnosis and/or treatment initiation (median LOS 17.5 days), and 9.0% hospitalized later during treatment (median LOS 13 days). Inadjusted models, patients were more likely to be hospitalized initially if they were children, had co-morbidities,smear-positive symptomatic pulmonary TB, cavitary or miliary TB, and multi- or poly-TB drug resistance. Factorspredictive of longer initial LOS included having HIV, renal disease, symptomatic pulmonary smear-positive TB,multi- or poly-TB drug resistance, and being in a teaching hospital.Conclusions: We found a high hospitalization rate during diagnosis and treatment of patients with TB. Diagnosticdelay due to low index of suspicion may result in patients presenting with more severe disease at the time ofdiagnosis. Earlier identification and treatment, through interventions to increase TB awareness and more targetedprevention programs, might reduce costly TB-related hospital use.Keywords: Active tuberculosis, Health service utilization, Length of hospital stay, Survival analysis, EpidemiologyBackgroundTuberculosis (TB) is generally considered an ambulatorydisease [1], however, hospitalization remains an import-ant part of management [2, 3]. Benefits of hospitalizationinclude isolation of infectious patients, and medicalstabilization of patients with or at-risk of complicationsof the disease or therapy [1]. For patients who are homelessor living in group settings, have psychiatric or substanceabuse issues, or have difficulties with activities of daily liv-ing, extended inpatient treatment may be used to ensurecompletion of treatment [4, 5].However, hospitalization increases risk of nosocomialtransmission to health care workers and other vulnerablepatients [6]. Inpatient diagnostic delays can increasetransmission risk in this vulnerable population [7, 8]. TBhospitalizations can lead to psychological difficulties forpatients due to isolation and detention [1]. Lastly, thereare cost implications [2, 9, 10]—hospitalizations are gen-erally the largest health system cost component of TBcase management, often accounting for more than 50%of total treatment costs [11–15].* Correspondence: dick.menzies@mcgill.ca1Department of Epidemiology, Biostatistics and Occupational Health, McGillUniversity, Montreal, QC, Canada5Respiratory Epidemiology and Clinical Research Unit (RECRU)/ MontrealChest Institute, McGill University Health Centre, Room 419, 2155 Guy St,Montreal, QC H3H 2R9, CanadaFull list of author information is available at the end of the article© The Author(s). 2016 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Ronald et al. BMC Infectious Diseases  (2016) 16:679 DOI 10.1186/s12879-016-1997-xBetter understanding of which patients are hospitalizedduring the diagnostic and management phases of TB canhelp policy-makers and clinicians more efficiently plantreatment programs and resource needs. Therefore, theobjectives of the current study were to identify how fre-quently patients were hospitalized, and what factorswere associated with hospitalization and length of hos-pital stay during diagnosis and treatment of active TBpatients in a large Canadian city, using patient-levelclinical and demographic data from a clinical researchdatabase.MethodsData source and study populationA retrospective cohort was identified of all individualswith confirmed TB, and notified to Montreal Public Healthbetween January 1996 and May 2007. Given the mandatorylegal requirement to report all TB cases, this is virtually acomplete capture of all persons diagnosed with microbio-logically or clinically confirmed TB in Montreal (covering acatchment of approximately two million people). Further,approximately two-thirds of TB cases in Quebec are re-ported in Montreal [16], thus the database includes the ma-jority of TB cases in the province for those years. Detailedinformation regarding patients’ clinical status, treatment,and hospitalization history for Montreal Island hospitalswas extracted from public health charts and from medicalcharts at treating hospitals and clinics for each case. Dataextraction was performed by trained research assistantsusing standardized data extraction forms. Ethical approvalfor this study was obtained through the McGill UniversityHealth Centre Research Ethics Board.Outcome and predictor variablesThe main outcomes were whether or not the individualhad a TB-related hospitalization and the length of stay(LOS) for the hospitalizations. Hospitalizations were de-fined as admissions to an acute care hospital on the Is-land of Montreal, and excluded emergency departmentvisits of less than 24 h, hospital day clinics, and day surger-ies. Hospital transfers were defined as when a dischargeand a new admission date were on the same or next day.When a transfer was identified, these were counted as onecontinuous hospitalization. LOS was calculated for eachcontinuous hospitalization as the difference between dateof discharge (or in-hospital death) and admission date. Weexcluded hospitalizations missing admission or dischargedates (n = 19).The TB diagnosis date was identified as the treatmentstart date, or when not available, the date the mandatorynotification was made. Hospitalizations were stratifiedbased on whether the timing was at the time of diagnosisor during treatment—defined as: 1) an initial TB-relatedhospitalization (admission within 1 month before or up to1 month after the date of TB diagnosis, or where TB wasdiagnosed in-hospital); and 2) a hospitalization later dur-ing the treatment period (admission more than 1 monthafter the TB diagnosis date and while the patient was stillundergoing TB treatment) [14]. It was hypothesized thatfactors associated with hospitalizations at the time ofdiagnosis and treatment initiation would differ from thoseassociated with hospitalizations occurring later duringtreatment.Demographic variables included age in years, sex, pro-vincial health insurance registration, self-reported Abori-ginal status, immigration details, and homeless status.We stratified foreign-born patients into categories ofimmigration from high or low-moderate TB incidencecountries based on WHO definitions [17], and timesince immigration based on published studies indicatingthat TB rates are highest within the first 2 years of arrival[18]. Additional variables included smoking, drug and al-cohol abuse history, HIV co-infection, and clinical charac-teristics (smear and culture results, disease site, TB drugresistance, cavitary and miliary TB, pulmonary and/orsystemic symptoms), and method of therapy supervision(directly observed therapy (DOT) or self-administeredtherapy (SAT)). Patient co-morbidities were extractedfrom notations in hospital and public health charts; theseincluded diabetes, cancer, renal disease, HIV, substanceabuse, liver disease or other conditions. Among hospital-ized patients, we identified if the hospital was a teaching(university) or non-teaching hospital, and if they diedwhile in-hospital.Statistical analysesWe calculated the proportion of patients with TB-relatedhospitalizations, overall, and stratified by hospitalizationtiming, and the mean (standard deviation) and median(interquartile range, IQR) LOS (in days). We furtherestimated the proportion of patients starting treatment in-hospital, and for these patients, estimated the length of in-hospital diagnostic delay as the median (IQR) number ofdays between date of hospital admission and date whenTB treatment was started.We identified factors associated with having anyhospitalization by calculating univariable and multivari-able odds ratios using logistic regression (SAS 9.3, SASInstitute Inc, Cary, USA), stratified on hospitalizationtiming. We considered variables for inclusion in adjustedmodels using a backward selection procedure.We identified factors predictive of hospital LOS usingCox proportional hazards regression [19], whereby sur-vival time was duration of time spent in-hospital and theevent was hospital discharge. Patients were censoredwhen they died in-hospital. We included hospital type inthe model as a time-dependent covariate to account fortransfers. Since some patients had multiple separateRonald et al. BMC Infectious Diseases  (2016) 16:679 Page 2 of 11hospitalizations, we used a conditional counting processmodel whereby each hospitalization was assigned to aseparate stratum and the time scale was time since studyentry [20]. This allowed the baseline hazard to differ de-pending on whether it was the patient’s first or secondhospitalization [20]. We used a robust sandwich varianceestimator for estimating standard errors. When therewas strong evidence that proportional hazards assump-tions did not hold, we included an interaction term inthe model of the variable multiplied by time. We consid-ered variables for inclusion in adjusted models using abackward selection procedure.ResultsCharacteristics of TB patientsThere were 1852 patients with confirmed TB reported toMontreal Public Health between January 1996 and May2007 (Table 1). Roughly half were male, with a median ageof 44 years, and 80.5% were foreign-born, with slightlymore than one-quarter having arrived to Canada within 2years before the TB notification date. Approximately one-quarter were current or past smokers, almost one-sixthreported past or current substance abuse, but few werehomeless. Six patients were reported as Canadian-bornAboriginal (data not shown). Concomitant HIV infectionwas documented in 7.9% of patients (Table 1).Most patients had pulmonary TB disease, of which one-third were smear-positive (Table 2). Few strains were TBdrug resistant (6.6% mono-resistant and 1.8% multi/poly-Table 2 Clinical characteristics of persons with TB notified toMontreal Public Health between January 1996 – May 2007,N = 1852Number (%) ofactive TB patientsaN %Mycobacteriology and disease sitePulmonary, Smear+ 619 33.4Pulmonary, Smear- 625 33.7Extra-pulmonary 563 30.4Clinical and radiographic features of TB diseaseMiliary TB 35 1.9Cavitary TB 336 18.1Any pulmonary TB-related symptomsb 1530 82.6Any systemic TB-related symptomsc 1018 55.0Drug resistanceMulti- and poly-TB drug resistant 33 1.8Mono-TB drug resistant 122 6.6Pan-sensitive 1664 89.8Therapy supervision, ever directly-observed (DOT) 807 43.5Treatment duration in days, median (min-max) 208 (183–295)Treatment outcomeCured 1518 82.0Died 127 6.9Defaulted/failed 34 1.8Moved 69 3.7aMissing data: TB diagnosis (n = 45), drug resistance (n = 33), treatmentoutcome (n = 104)bPulmonary TB-related symptoms: cough, hemoptysis, and abnormalchest x-raycSystemic TB-related symptoms: weight loss, fever, fatigue, night sweatsTable 1 Characteristics of persons with TB, notified to Montrealpublic health between January 1996 – May 2007, N = 1852Number (%) ofactive TB patientsaN %Patient demographicsSex- male 992 53.6Age group0–19 years 143 7.720–34 years 602 32.535–64 years 711 38.465+ years 390 21.1Had provincial health insurance coverage 1499 80.9Country of birthbCanada 317 17.1Foreign-born, low/moderate TB incidence country 1112 60.0Foreign-born, high TB incidence country 380 20.5Immigration, years since arrival≤ 2 years 485 26.23–14 years 541 29.215 or more years 338 18.3Foreign-born, unknown year of arrival 128 6.9Canadian-born 317 17.1Homeless 9 0.5Patient co-morbiditiesCancer 41 2.2Diabetes 132 7.1HIV positive 146 7.9Renal disease 105 5.7Liver disease 153 8.3Ever substance abuse 249 13.4Number of co-morbiditiesc2 or more 412 22.31 446 24.1Ever smoker 489 26.4aMissing data: sex (n = 19), age (n = 6), country of birth (n = 43)bHigh TB incidence countries (WHO estimated sputum smear positive TB rateof ≥15 per 100,000, 3 year average, 2005–2007) [28]cCo-morbidities include renal disease, liver disease, diabetes, cancer, substanceabuse, HIV infection, or otherRonald et al. BMC Infectious Diseases  (2016) 16:679 Page 3 of 11drug resistant). Most patients were cured (82%), while6.9% died and the remainder moved out of province,defaulted or failed treatment (Table 2). The mediantreatment duration was 208 days (IQR = 183–295) andwas longest for multi- and poly-drug resistant cases(median 517 days, data not shown). Further stratificationof covariates by years since immigration suggested that re-cent immigrants (within the past 2 years) were on averageyounger and had less co-morbidity than Canadian-bornactive TB patients and foreign-born patients in Canada for15 or more years, were more likely to have pulmonary TBthat was smear-negative and less likely to be symptomaticat the time of presentation (Table 3).Predictors of hospitalizations and LOSIn total, 1001 patients were hospitalized (54.1%); 942patients (50.9%) had a hospitalization initially, and 167(9.0%) later during treatment (Table 4). Of 942 hospi-talized initially, 114 were hospitalized more than onceduring this period 30 days before or after the TB diagnosisdate (Table 4). The median LOS of initial hospitalizationswas 17.5 days (IQR = 9–31) and 13 days (IQR = 6–22) laterduring treatment. Treatment was started in-hospital for731 patients (39.5%), with a median time to starting treat-ment after admission of 4 days (IQR = 2–8). The propor-tion of patients hospitalized and median LOS did notchange over the study period (Fig. 1).Table 3 Characteristics and treatment outcomes of persons notified to Montreal public health between January 1996 – May 2007,stratified by years since arrival to CanadaCharacteristics of Active TBPatientsaForeign-born, years since arrival to Canada Canadian-born p-value0–2 (N = 485) 3–14 (N = 541) 15+ (N = 338) (N = 317)n % N % N % n %Sex, male 281 (57.9) 275 (50.8) 172 (51.0) 172 (54.2) 0.22Age group, in years <0.010–19 51 (10.5) 35 (6.5) 1 (0.3) 50 (15.8)20–34 257 (53.0) 233 (43.0) 40 (11.8) 37 (11.7)35–64 147 (30.3) 222 (41.0) 169 (50.0) 105 (33.1)65+ 29 (6.0) 50 (9.2) 128 (37.9) 122 (38.5)Had provincial health insurance # 213 (43.9) 499 (92.2) 328 (97.0) 308 (97.2)Homeless 1 (0.2) 0 (0) 1 (0.3) 7 (2.2) <0.01Ever smoker 122 25.2 99 18.3 98 29.0 159 50.2 <0.01Ever substance abuse 27 (5.6) 56 (10.4) 51 (15.1) 98 (30.9) <0.01HIV 35 (7.2) 51 (9.4) 21 (6.2) 28 (8.8) 0.38Number of co-morbiditiesb <0.012 or more 53 (10.9) 78 (14.4) 104 (30.8) 130 (41.0)1 93 (19.2) 116 (21.4) 102 (30.2) 92 (29.0)TB diagnosis <0.01Pulmonary, Smear+ 126 (26.0) 188 (34.8) 123 (36.4) 144 (45.4)Pulmonary, other 257 (53.0) 124 (22.9) 77 (22.8) 108 (34.0)Extra-pulmonary 88 (18.1) 219 (40.5) 131 (38.8) 55 (17.4)Any pulmonary symptomsc 435 (89.7) 403 (74.5) 273 (80.8) 291 (91.8) <0.01Any systemic symptomsd 201 (41.4) 321 (59.3) 212 (62.7) 203 (64.0) <0.01Multi- or poly-drug resistant 13 (2.7) 12 (2.2) 2 (0.6) 4 (1.3) 0.20Treatment outcome <0.01Cured 421 (86.8) 465 (86.0) 263 (77.8) 246 (77.6)Died 3 (0.6) 16 (3.0) 41 (12.1) 45 (14.2)Defaulted/failed 7 (1.5) 10 (1.8) 4 (1.2) 9 (2.8)Moved 34 (7.0) 20 (3.7) 5 (1.5) 3 (1.0)aNote: table subset to people with known birth country and timing of arrival to Canada (N = 1681). Missing data: sex (n = 18), age (n = 5), TB diagnosis (n = 41),treatment outcome (n = 89)bCo-morbidities include renal disease, liver disease, diabetes, cancer, substance abuse, HIV infection, or other (as reported in public health or hospital records)cPulmonary TB-related symptoms include cough, hemoptysis, and/or abnormal chest x-raydSystemic TB-related symptoms include weight loss, fever, fatigue, and/or night sweatsRonald et al. BMC Infectious Diseases  (2016) 16:679 Page 4 of 11In adjusted models (Table 5), patients were more likelyto be hospitalized initially if they were children, had HIVco-infection, renal disease, co-morbidities, smear-positivepulmonary TB, cavitary TB, miliary TB, pulmonary or sys-temic symptoms, and multi- or poly-TB drug resistance.Factors associated with hospitalization later during treat-ment were similar to those associated with initial hospital-izations (Table 5).Factors predictive of longer initial LOS in adjustedmodels included older age, HIV infection, renal disease,pulmonary smear-positive TB, pulmonary symptoms,multi- or poly-TB drug resistance, and admission to ateaching hospital (Table 6). Factors associated with lon-ger LOS when hospitalized later during treatment weresimilar to initial hospitalizations, with the exception thathaving extra-pulmonary TB was associated with longerLOS (Table 6).DiscussionIn a setting with universal health care access, over half ofall patients were hospitalized at the time of TB diagnosisand treatment initiation. Hospitalization frequencies andLOS in our study were comparable to other US andCanadian studies [12, 14, 21]. In the one US study todifferentiate initial and during-treatment hospitalizations,Taylor found that 45% of TB patients had an initialhospitalization for TB and 8% were hospitalized duringtreatment, with a median LOS of 11 days [14].In Montreal Health region, there are six hospitalstreating TB patients which have adequate respiratory isola-tion facilities. The Canadian TB Standards recommend thatpatients with confirmed smear-positive, culture-positivedrug susceptible respiratory TB be kept under airborne pre-cautions until there is clinical evidence of improvement,evidence of adherence to at least 2 weeks of effective treat-ment, and three consecutive negative sputum smears [22].Thus some of our findings likely reflect recommendedhospitalization guidelines for TB patients. However, pa-tients may be discharged to home isolation for the periodrequiring airborne precautions provided there is clinical im-provement, drug-resistant TB is not suspected and there isno contraindication for home isolation [22].Our study also identified several population groups athigher hospitalization risk and longer LOS. HospitalizationTable 4 Number of hospitalizations and lengths of hospital stay of active TB cases, notified to Montreal public health betweenJanuary 1996 – May 2007, stratified by timing of hospitalizationTiming of hospitalization All hospitalizationsInitiala DuringbPatients with any hospitalization, n (%) 942 (50.9) 167 (9.0) 1001 (54.1)Patients with >1 hospitalization, n (%) 114 (6.1) 35 (1.9) 124 (6.7)Total number of hospitalizations 1006 206 1212Mean length of hospital stay (LOS), in days (se) 28.2 (39.0) 21.6 (36.1) 27.2 (38.6)Median length of hospital stay (LOS), in days (IQR) 17.5 (9–31) 13 (6–22) 16.5 (8–29)Patients starting TB treatment in-hospital, n (%) 731 (39.5%) - 731 (39.5%)Median # days (IQR) from admission to treat start 4 (2–8) - 4 (2–8)LOS length of stay, se standard error, IQR interquartile rangeaInitial = hospital admission within 1 month before start of TB treatment or up to 1 month after treatment start, or any admission where TB treatmentstarted in-hospitalbDuring = hospital admission more than 1 month after start of TB treatment and during TB treatment periodFig. 1 Number of initial hospitalizations and median length of hospital stay, notified to Montreal Public Health between January 1996 – May 2007,stratified by year of notification (N = 1852)Ronald et al. BMC Infectious Diseases  (2016) 16:679 Page 5 of 11Table 5 Predictors of hospitalization during diagnosis and treatment of active TB cases, notified to Montreal public health betweenJanuary 1996 – May 2007, stratified by timing of hospitalization (N = 1852)Characteristics Any initial hospitalization (yes/no) Any hospitalization during treatment (yes/no)Number ofpatients (%)Crude OR(95% CI)Adjusted ORa(95% CI)Number ofpatients (%)Crude OR(95% CI)Adjusted ORa(95% CI)Male 536 (54.0) 1.30 (1.07–1.57) - 91 (9.2) 1.08 (0.78–1.49) -Female 399 (47.4) 1.00 72 (8.6) 1.00Age group0–19 years 85 (59.4) 2.04 (1.38–3.00) 2.52 (1.63–3.92) 15 (10.5) 2.09 (1.10–3.97) 2.82 (1.38–5.76)20–34 years 279 (46.4) 1.00 1.00 32 (5.3) 1.00 1.0035–64 years 249 (54.7) 1.39 (1.11–1.73) 1.09 (0.83–1.43) 60 (8.7) 1.64 (1.05–2.56) 1.11 (0.68–1.81)≥ 65 years 258 (66.2) 1.92 (1.47–2.51) 1.13 (0.80–1.59) 58 (14.9) 3.11 (1.98–4.89) 2.19 (1.25–3.85)Had provincial health insurance #Yes 804 (53.6) 1.75 (1.37–2.24) - 149 (9.9) 2.05 (1.24–3.40) -No 138 (39.1) 1.00 18 (5.1) 1.00Country of birth - -Canada 196 (61.8) 2.02 (1.48–2.75) 51 (16.1) 2.61 (1.59–4.30)Foreign, low-moderate TB incidence 551 (49.6) 1.16 (0.92–1.47) 85 (7.6) 1.13 (0.72–1.78)Foreign, high TB incidence country 174 (45.8) 1.00 26 (6.8) 1.00Immigration, years since arrivalForeign-born, unknown arrival date 53 (41.4) 0.41 (0.27–0.63) - 13 (10.2) 0.59 (0.31–1.13) -15 or more years 204 (60.4) 0.90 (0.65–1.25) 37 (11.0) 0.64 (0.41–1.01)3–14 years 269 (49.7) 0.58 (0.43–0.77) 39 (7.2) 0.41 (0.26–0.63)0–2 years 199 (41.0) 0.41 (0.30–0.54) 22 (4.5) 0.25 (0.15–0.42)Canadian-born 196 (61.8) 1.00 51 (16.1) 1.00Homeless - -Yes 8 (88.9) 7.46 (0.93–59.66) 2 (22.2) 2.91 (0.60–14.11)No 934 (50.7) 1.00 165 (9.0) 1.00Cancer -Yes 30 (73.2) 2.23 (1.09–4.53) 13 (31.7) 5.00 (2.54–9.84) 3.43 (1.63–7.23)No 912 (50.4) 1.00 154 (8.5) 1.00 1.00Diabetes - -Yes 98 (74.2) 2.83 (1.88–4.26) 21 (15.9) 2.04 (1.24–3.35)No 844 (49.1) 1.00 146 (8.5) 1.00HIVPositive 111 (76.0) 3.24 (2.17–4.85) 1.46 (0.90–2.36) 36 (24.7) 3.84 (2.60–5.97) 3.31 (1.96–5.60)Negative or unknown 831 (48.7) 1.00 1.00 131 (7.7) 1.00 1.00Renal disease -Yes 83 (79.1) 3.92 (2.38–6.47) 1.99 (1.10–3.59) 22 (21.0) 2.93 (1.78–4.83)No 859 (49.2) 1.00 1.00 145 (8.3) 1.00Liver disease -Yes 105 (68.6) 2.02 (1.40–2.92) - 25 (16.3) 2.14 (1.35–3.40)No 837 (49.3) 1.00 142 (8.4) 1.00Ever substance abuse -Yes 176 (70.7) 2.69 (1.99–3.63) - 32 (12.9) 1.60 (1.06–2.42)No 766 (47.8) 1.00 135 (8.4) 1.00Ronald et al. BMC Infectious Diseases  (2016) 16:679 Page 6 of 11risk was lowest among immigrants who arrived in Canadawithin 2 years of diagnosis, and increased with time sinceimmigration, approaching rates of Canadian-born patientsby 15 years post-arrival. Recent immigrant patients mayhave less advanced disease due to earlier diagnosis throughactive screening [23]. This was suggested by our study,whereby recently immigrated TB patients were less likelyto have smear-positive, symptomatic pulmonary TB. Thesefindings may also partly reflect a “healthy immigrant ef-fect”, whereby immigrants’ health at the time of arrival isbetter than the Canadian-born population, an effect thattends to diminish with longer time spent in Canada [24].Having drug resistance was associated with a higher risk ofhospitalization and longer LOS—similar to the Taylorstudy, which found that MDR-TB cases were almost 6times more likely to be hospitalized than non-MDR pa-tients [14]. This has implications for resource planning,given increasing drug-resistant TB globally. There was alsoevidence for an increased hospitalization rate and longerLOS of homeless patients in our study, similar to previousstudies [14, 25].Strengths of this study include the virtually completecapture of confirmed TB cases in a large health region.The extraction of detailed clinical data from hospital,clinic and public health records allowed investigationinto clinical and social determinants of hospitalization inthe general population of TB patients. Limitations in-clude that these data were collected retrospectively andrecording of some details, such as smoking, alcohol, andillicit drug use may not be complete. Additionally, if pa-tients were hospitalized outside Montreal, this may nothave been recorded in the study database, thus under-estimating hospitalization rates. It is possible that somehospitalizations were due to co-morbidities, rather thanpresentation of active TB symptoms (thus overestimatingthe TB-related hospitalization rate). These results also mayTable 5 Predictors of hospitalization during diagnosis and treatment of active TB cases, notified to Montreal public health betweenJanuary 1996 – May 2007, stratified by timing of hospitalization (N = 1852) (Continued)Number of co-morbidities2 or more 316 (77.3) 5.31 (4.04–6.98) 3.56 (2.50–5.07) 76 (18.6) 5.44 (3.64–8.14) 2.90 (1.69–4.96)1 242 (53.9) 1.84 (1.46–2.32) 1.43 (1.08–1.89) 51 (11.4) 3.06 (1.99–4.70) 2.17 (1.31–3.60)0 384 (38.6) 1.00 1.00 40 (4.0) 1.00 1.00Ever smoker -Yes 306 (62.6) 1.92 (1.54–2.39) - 61 (12.5) 1.69 (1.21–2.36)No 636 (46.7) 1.00 106 (7.8) 1.00TB diagnosis -Pulmonary, Smear+ 447 (72.2) 3.87 (3.02–4.96) 1.48 (1.04–2.09) 70 (11.3) 1.43 (0.97–2.12)Pulmonary, Smear- 257 (41.2) 1.05 (0.83–1.32) 0.65 (0.46–0.90) 47 (7.5) 0.91 (0.60–1.40)Extra-pulmonary 228 (40.5) 1.00 1.00 46 (8.2) 1.00Cavitary TB -Yes 230 (68.5) 2.49 (1.92–3.23) 1.75 (1.28–2.40) 25 (7.4) 0.76 (0.48–1.19)No 712 (47.0) 1.00 1.00 142 (9.4) 1.00Miliary TBYes 29 (82.9) 5.33 (2.05–13.87) 3.53 (1.26–9.88) 8 (22.9) 3.09 (1.38–6.92) 2.09 (0.88–4.96)No 913 (50.3) 1.00 1.00 159 (8.8) 1.00 1.00Any pulmonary TB-related symptomsYes 853 (55.8) 3.45 (2.63–4.53) 2.12 (1.44–3.11) 149 (9.7) 1.77 (1.07–2.93) -No 89 (27.6) 1.00 1.00 18 (5.6) 1.00Any systemic TB-related symptomsYes 697 (68.5) 5.18 (4.22–6.36) 3.54 (2.82–4.46) 123 (12.1) 2.44 (1.70–3.50) 1.72 (1.16–2.57)No 245 (28.4) 1.00 1.00 44 (5.3) 1.00 1.00Multidrug or poly-drug resistanceYes 26 (78.8) 3.95 (1.61–9.67) 6.98 (2.65–18.30) 8 (24.2) 3.48 (1.54–7.89) 5.55 (2.18–14.14)No 914 (51.2) 1.00 1.00 157 (8.8) 1.00 1.00OR odds ratio, CI confidence intervalaModels adjusted for all reported variablesRonald et al. BMC Infectious Diseases  (2016) 16:679 Page 7 of 11Table 6 Predictors of hospital length of stay during diagnosis and treatment of active TB, stratified by timing of hospitalization,Montreal, January 1996 – May 2007Initial hospitalizations(n = 1006)Hospitalizations during treatment(n = 206)Characteristics Crude HR(95% CI)Adjusted HRa,b(95% CI)Crude HR(95% CI)Adjusted HRa(95% CI)SexMale 1.01 (0.89–1.14) - 1.29 (0.78–2.11) -Female 1.00 1.00Age in yearsc 0.99 (0.98–1.00)log(LOS)≤ 2.5 1.02 (1.02–1.03) 1.02 (1.02–1.03) - -log(LOS) > 2.5 0.96 (0.96–0.96) 0.96 (0.96–0.96) - -Had provincial health insurance #Yes 0.96 (0.81–1.14) - 1.15 (0.63–2.10) -No 1.00 1.00Country of birthCanada 0.68 (0.55–0.83) - 0.80 (0.47–1.37) -Foreign, low-moderate incidence 0.88 (0.74–1.04) 0.57 (0.33–0.98)Foreign, high incidence country 1.00 1.00Immigration, years since arrivalForeign-born, unknown arrival date 1.47 (1.13–1.91) - 0.57 (0.28–1.15) -15 or more years 1.11 (0.92–1.35) 0.70 (0.34–1.44)3–14 years 1.50 (1.26–1.79) 0.85 (0.50–1.42)0–2 years 1.42 (1.17–1.71) 1.15 (0.58–2.28)Canadian-born 1.00 1.00Homeless - -Yes 0.89 (0.56–1.40) 0.26 (0.11–0.63)No 1.00 1.00Cancer -Yes 0.82 (0.55–1.23) 1.42 (0.89–2.26) -No 1.00 1.00Diabetes -Yes 0.92 (0.76–1.11) 1.05 (0.63–1.75) -No 1.00 1.00HIVPositive 0.83 (0.70–0.99) 0.77 (0.66–0.90) 0.93 (0.59–1.47) 0.61 (0.40–0.91)Negative or unknown 1.00 1.00 1.00 1.00Renal diseaseYes 0.55 (0.43–0.70) 0.59 (0.46–0.75) 0.45 (0.21–0.95) 0.58 (0.31–1.10)No 1.00 1.00 1.00 1.00Liver diseaseYes 0.79 (0.66–0.94) - 1.62 (0.99–2.66) -No 1.00 1.00Ever substance abuseYes 0.91 (0.79–1.05) - 0.88 (0.58–1.33) -No 1.00 1.00Ronald et al. BMC Infectious Diseases  (2016) 16:679 Page 8 of 11not be generalizable to other jurisdictions, for example, TBhospitalizations have been shown to vary based on urbanor rural residence [26].This study adds to our knowledge about hospitalizationduring TB diagnosis and treatment in a large urban Can-adian population. These data can be used by planners tobetter capture TB resource implications especially with re-gard to the cost effectiveness of treatment of LTBI. Ourfinding that homeless patients were more likely to beadmitted with longer LOS, suggests a need for earlierdiagnosis and treatment for these patients, and is sup-ported by similar findings by Marks et al. in the US [25].Our finding that Canadian-born patients tended to havemore severe disease at the time of diagnosis suggests aneed for interventions to reduce diagnostic delay, such asincreased training for health care professionals, patients,and community members to increase their TB awareness.Some hospitalizations may also be preventable; in a USstudy, for example, close to 40% of hospitalizations of TBpatients were considered likely to have been avoidable [1].We found that smear-positive patients were more likely tobe hospitalized with longer LOS; it is possible that in someof these cases, physicians may have been reluctant todischarge patients if they are unsure how to manage aTable 6 Predictors of hospital length of stay during diagnosis and treatment of active TB, stratified by timing of hospitalization,Montreal, January 1996 – May 2007 (Continued)Number of co-morbidities2 or more 0.65 (0.56–0.75) - 0.74 (0.44–1.25) -1 0.85 (0.73–1.00) 0.69 (0.36–1.31)0 1.00 1.00Ever smokerYes 0.84 (0.74–0.96) - 0.91 (0.59–1.40) -No 1.00 1.00TB diagnosisPulmonary, Smear+ 0.61 (0.51–0.72) 0.64 (0.54–0.77) 1.51 (0.84–2.69) 1.26 (0.78–1.99)Pulmonary, Smear- 089 (0.73–1.09) 1.02 (0.84–1.26) 1.50 (0.80–2.83) 1.81(1.07–3.04)Extra-pulmonary 1.00 1.00 1.00 1.00Cavitary TBYes 0.92 (0.90–1.05) - 1.91 (1.13–3.25) -No 1.00 1.00Miliary TBYes 0.83 (0.63–1.10) - 1.13 (0.70–1.83) -No 1.00 1.00Any pulmonary TB-related symptomsYes 0.65 (0.50–0.85) 0.70 (0.55–0.90) 0.61 (0.42–0.90) 0.46 (0.27–0.80)No 1.00 1.00 1.00 1.00Any systemic TB-related symptomsYes 0.80 (0.69–0.93) - 1.16 (0.60–2.26) -No 1.00 1.00Multidrug or poly-drug resistanceYes 0.65 (0.50–0.85) 0.44 (0.32–0.60) 0.62 (0.28–1.39) 0.28 (0.12–0.66)No 1.00 1.00 1.00 1.00Teaching hospitalYes 0.86 (0.75–0.99) 0.81 (0.71–0.92) 0.99 (0.62–1.59) -No 1.00 1.00 1.00IQR interquartile range, LOS length of stay, HR hazard ratio, CI confidence intervalaModels adjusted for all reported variablesbCox Proportional Hazard model is modeling the time to discharge; a hazard ratio of less than one indicates a longer LOS (ie. A HR for multidrug or poly-drugresistance = 0.45 can be interpreted as: the rate of discharge from hospital is 55% slower for patients with drug resistance compared to those withoutdrug resistance)cGiven strong evidence that proportional hazards assumptions did not hold for age (lines crossed in the log-log plot), we included an interaction term in themodel of age (continuous, in years) multiplied by the LOS. This interaction term indicates that the effect of age was not constant throughout the length of stayRonald et al. BMC Infectious Diseases  (2016) 16:679 Page 9 of 11potentially infectious patient in their home, as previouslyreported [1]. Professional development programs for phy-sicians on evidence-based strategies for TB managementcould help to reduce time spent in-hospital. Improvingcommunity-based DOT programs could also lead to re-ductions in hospitalization rates of TB patients [27].ConclusionWe found a high hospitalization rate at the time of diag-nosis and during treatment of patients with TB. Diag-nostic delay due to low index of suspicion may result inpatients presenting with more severe disease at the timeof diagnosis. Earlier identification and treatment of pa-tients, through interventions such as enhanced trainingto increase TB awareness and more targeted preventionprograms, might reduce costly TB-related hospital use.AbbreviationsCI: Confidence interval; DOT: Directly observed therapy; HIV: Humanimmunodeficiency virus; HR: Hazard ratio; IQR: Interquartile range;LOS: Length of stay; LTBI: Latent tuberculosis infection; RAMQ: Régiede l’assurance maladie du Québec; SAT: Self-administered therapy;TB: Tuberculosis; WHO: World Health OrganizationFundingLisa Ronald received funding support through the Canadian Institutes ofHealth Research (Banting and Best Doctoral Award), the BC Lung Association,and the Quebec Respiratory Health Network Training Program. The establishmentof the database used in this study was supported through a CIHR funded grantentitled “Understanding the Keys to Tuberculosis: From Exposure to Infection andfrom Infection to Disease”).Availability of data and materialsThe establishment of the database used in this study was supported througha CIHR funded grant entitled “Understanding the Keys to Tuberculosis: FromExposure to Infection and from Infection to Disease”). These are confidentialpatient-level data and are not publicly available.Authors’ contributionsLR was responsible for the original ideas behind the manuscript, analysis andwriting. DM and JMF provided guidance on all aspects of the studies andare key editors. AB, GBE, JFB, and KS provided methodological guidance andeditorial assistance. All authors read and approved the final manuscript.Competing interestsThe authors declare that they have no competing interests.Consent for publicationNot applicable: manuscript does not contain any individual person’s data.Ethics approval and consent to participateEthical approval for this study protocol was obtained through the McGillUniversity Health Centre Research Ethics Board (BMC06-023/BMC06-006t).The Institutional Review Board for each institution and the administrators ofeach hospital gave approval for access to medical records; patient consentwas not required by the MUHC Ethics Board review as this was a retrospectivestudy.Author details1Department of Epidemiology, Biostatistics and Occupational Health, McGillUniversity, Montreal, QC, Canada. 2Division of Respiratory Medicine, Faculty ofMedicine, University of British Columbia, Vancouver, BC, Canada. 3Centre forClinical Epidemiology and Evaluation, Vancouver Coastal Health ResearchInstitute, Vancouver, BC, Canada. 4Institute for Heart and Lung Health,University of British Columbia, Vancouver, Canada. 5Respiratory Epidemiologyand Clinical Research Unit (RECRU)/ Montreal Chest Institute, McGillUniversity Health Centre, Room 419, 2155 Guy St, Montreal, QC H3H 2R9,Canada. 6Department of Family Medicine, McGill University, Montreal, QC,Canada. 7Department of Medicine, McGill University, Montreal, QC, Canada.Received: 7 May 2016 Accepted: 29 October 2016References1. Thomas JA, Laraque F, Munsiff S, Piatek A, Harris TG. Hospitalizations fortuberculosis in New York City: how many could be avoided? Int J TubercLung Dis. 2010;14(12):1603–12.2. MacIntyre CR, Plant AJ, Hendrie D. Shifting the balance between in-patientand out-patient care for tuberculosis results in economic savings. Int JTuberc Lung Dis. 2001;5(3):266–71.3. Taylor Z, Nolan CM, Blumberg HM. Controlling tuberculosis in the UnitedStates. Recommendations from the American Thoracic Society, CDC,and the Infectious Diseases Society of America. MMWR Recomm Rep.2005;54(RR-12):1–81.4. Oscherwitz T, Tulsky JP, Roger S, Sciortino S, Alpers A, Royce S, et al.Detention of persistently nonadherent patients with tuberculosis. JAMA.1997;278(10):843–6.5. Singleton L, Turner M, Haskal R, Etkind S, Tricarico M, Nardell E. Long-termhospitalization for tuberculosis control. Experience with a medical-psychosocial inpatient unit. JAMA. 1997;278(10):838–42.6. Edlin BR, Tokars JI, Grieco MH, Crawford JT, Williams J, Sordillo EM, et al.An outbreak of multidrug-resistant tuberculosis among hospitalizedpatients with the acquired immunodeficiency syndrome. N Engl J Med.1992;326(23):1514–21.7. Greenaway C, Menzies D, Fanning A, Grewal R, Yuan L, FitzGerald JM. Delayin diagnosis among hospitalized patients with active tuberculosis–predictorsand outcomes. Am J Respir Crit Care Med. 2002;165(7):927–33.8. Rao VK, Iademarco EP, Fraser VJ, Kollef MH. Delays in the suspicion andtreatment of tuberculosis among hospitalized patients. Ann Intern Med.1999;130(5):404–11.9. Marks SM, Taylor Z, Miller BI. Tuberculosis prevention versushospitalization: taxpayers save with prevention. J Health Care PoorUnderserved. 2002;13(3):392–401.10. Nutini S, Fiorenti F, Codecasa LR, Casali L, Besozzi G, Di Pisa G, et al. Hospitaladmission policy for tuberculosis in pulmonary centres in Italy: a nationalsurvey. AIPO Tuberculosis Study Group. Italian Association of HospitalPulmonologists. Int J Tuberc Lung Dis. 1999;3(11):985–91.11. Bocchino M, Greco S, Rosati Y, Mattioli G, Marruchella A, De Mori P, et al.Cost determinants of tuberculosis management in a low-prevalencecountry. Int J Tuberc Lung Dis. 2006;10(2):146–52.12. Brown RE, Miller B, Taylor WR, Palmer C, Bosco L, Nicola RM, et al. Health-careexpenditures for tuberculosis in the United States. Arch Intern Med.1995;155(15):1595–600.13. Menzies D, Lewis M, Oxlade O. Costs for tuberculosis care in Canada.Can J Public Health. 2008;99(5):391–6.14. Taylor Z, Marks SM, Rios Burrows NM, Weis SE, Stricof RL, Miller B. Causesand costs of hospitalization of tuberculosis patients in the United States.Int J Tuberc Lung Dis. 2000;4(10):931–9.15. Wurtz R, White WD. The cost of tuberculosis: utilization and estimatedcharges for the diagnosis and treatment of tuberculosis in a public healthsystem. Int J Tuberc Lung Dis. 1999;3(5):382–7.16. MSSS. Épidémiologie de la tuberculose au Québec de 2004 à 2007(2009-04-03): Sante and services sociaux Quebec, vol. 73. 2009.17. Clancy L, Rieder HL, Enarson DA, Spinaci S. Tuberculosis elimination in the countriesof Europe and other industrialized countries. Eur Respir J. 1991;4(10):1288–95.18. Langlois-Klassen D, Wooldrage KM, Manfreda J, Sutherland K, Ellis E, Phypers M,et al. Piecing the puzzle together: foreign-born tuberculosis in an immigrant-receiving country. Eur Respir J. 2011;38(4):895–902.19. Fenn P, Davies P. Variations in length of stay. A conditional likelihoodapproach. J Health Econ. 1990;9(2):223–34.20. Therneau T, Grambsch P. Modeling Survival Data: Extending the Cox Model.New York: Springer Science + Business Media; 2000.21. Rajbhandary SS, Marks SM, Bock NN. Costs of patients hospitalized formultidrug-resistant tuberculosis. Int J Tuberc Lung Dis. 2004;8(8):1012–6.22. Canadian Thoracic Society. Canadian Tuberculosis Standards. 7th ed. 2013.http://www.respiratoryguidelines.ca/tb-standards-2013 . (Accessed 11 Nov2016).Ronald et al. BMC Infectious Diseases  (2016) 16:679 Page 10 of 1123. Laifer G, Widmer AF, Simcock M, Bassetti S, Trampuz A, Frei R, et al. TB in alow-incidence country: differences between new immigrants, foreign-bornresidents and native residents. Am J Med. 2007;120(4):350–6.24. Ng E. The healthy immigrant effect and mortality rates. Health Rep.2011;22(4):25–9.25. Marks SM, Taylor Z, Burrows NR, Qayad MG, Miller B. Hospitalization ofhomeless persons with tuberculosis in the United States. Am J PublicHealth. 2000;90(3):435–8.26. Martens P, Brownell M, Au W, MacWilliam L, Prior H, Schultz J, et al. HealthInequities in Manitoba: Is the Socioeconomic Gap Widening or NarrowingOver Time? Winnipeg: Manitoba Centre for Health Policy; 2010.27. Terry MB, Desvarieux M, Short M. Temporal trends in tuberculosishospitalization rates before and after implementation of directlyobserved therapy: New York City, 1988-1995. Infect Control HospEpidemiol. 2002;23(4):221–3.28. Public Health Agency of Canada. Canadian Tuberculosis Reporting System.Reporting Form Completion Guidelines (version 1.9). Effective January 1,2011. http://www.phac-aspc.gc.ca/tbpc-latb/pdf/guidelinesform-eng.pdf .(Accessed 11 Nov 2016).•  We accept pre-submission inquiries •  Our selector tool helps you to find the most relevant journal•  We provide round the clock customer support •  Convenient online submission•  Thorough peer review•  Inclusion in PubMed and all major indexing services •  Maximum visibility for your researchSubmit your manuscript atwww.biomedcentral.com/submitSubmit your next manuscript to BioMed Central and we will help you at every step:Ronald et al. BMC Infectious Diseases  (2016) 16:679 Page 11 of 11


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items