Trends in duration of hospitalization for patients with tuberculosis in Montreal, Canada from 1993 to 2007.
Dehghani, Kianoush ; Allard, Robert ; Gratton, Jean 等
Although the incidence of tuberculosis (TB) is low in Canada, the
disease remains an important health problem, mainly among Aboriginal
Peoples and immigrants. (1) In 2008, 1,600 TB cases (4.8 per 100,000)
were reported to the Canadian Tuberculosis Reporting System, of whom 21%
were Canadian-born Aboriginals and 62% foreign-born individuals.
Pulmonary TB constituted 68% of all reported TB cases in Canada (based
on prerelease results). (1)
For many years, initial hospitalization followed by discharge with
anti-TB medications have constituted the approach of choice for the
great majority of patients with pulmonary TB in Canada. (2) Indications
for hospitalization of TB patients include investigation and/or
treatment of symptoms, establishment of an appropriate therapeutic
regimen for patients with significant side effects to anti-TB
medications and/or with known/suspected resistance to these agents, and
provision of airborne isolation if this cannot be effectively provided
as an outpatient. (3) In fact, it has been argued that an initial period
of hospitalization is important to achieve a reliable control of the
infection in the community. (3)
Hospitalizing TB patients can be challenging, especially in areas
with minimal health care resources or appropriate inpatient facilities.
Moreover, hospitalization for TB is exceptionally costly. (4-6) In
Canada, the total cost per hospitalized TB patient has been estimated at
$19,157. (6) Longer hospitalization for TB can also have detrimental
effects on the psychosocial and economic well-being of the patients and
their families.
The duration of hospitalization for patients with newly diagnosed
pulmonary tuberculosis has decreased significantly over the years, and
remains variable depending on the sputum smear, chest radiography,
presence of other medical co-morbidities, patient's response to
anti-TB medications and patient's unfavourable psychosocial status.
(2,7,8) According to the Canadian TB Standards (CTBS), for patients with
smear-negative respiratory TB, airborne isolation may be discontinued
after 2 weeks of appropriate multi-drug therapy, provided that the
patient is improving clinically. On the other hand, for patients with
smear-positive TB, the CTBS recommends that airborne isolation should
continue until three consecutive sputum smears are negative, there is
evidence of clinical improvement and the patient has shown adherence to
at least 2 weeks of appropriate multi-drug therapy. (3) TB patients
initially smear-positive may be discharged home after 2 weeks of
therapy, even if they remain-smear positive, provided that the discharge
conditions ensure continued treatment and acceptable airborne isolation,
while minimizing risk of close contact with vulnerable individuals. (3)
In Quebec, 57.3% of TB cases (annual cumulative incidence: 7.1 per
100,000 persons) were reported in Montreal region (population:
1,620,693) during the period 2004-2007. (9) Concerns have been raised
that the growing shortage of health care professionals, along with
restructuring and the reduced capacity of institutionally-based health
care services have been contributing to shortened hospital stays. (10)
Hence, we hypothesized that, as for other types of hospitalizations,
hospital stays for pulmonary TB have likely become shorter through the
years. This can be problematic because of continued infectivity of
suboptimally treated patients after premature discharge.
METHODS
We used two data sources on TB hospitalizations in the region of
Montreal from 1993-2007. One is the Quebec registry system for hospital
discharge summary information (also called the Systeme de maintenance et
d'exploitation des donnees pour l'etude de la clientele
hospitaliere or Med-Echo registry). The Med-Echo data source is
anonymous and does not include universal identity information such as
name, date of birth or health insurance number. However, the cases in
Med-Echo are distinguished by complex identifiers unique to the system.
The data in Med-Echo contain demographic and hospitalization information
including name of hospital, dates of patient's admission and
discharge, and patient's principal and secondary diagnosis on
admission, but not clinical information such as results of sputum smear
or chest x-ray (CXR). In order to include such clinical information in
our analysis, we also used the computerized reportable diseases central
registry (also called the Fichier des maladies a declaration obligatoire
or MADO registry). This registry is used to manage and process
surveillance data on all reportable diseases in Quebec. The MADO
registry contains both demographic and clinical information including
the type and results of sputum smear, culture, drug sensitivity test,
and CXR. Of all reported TB cases in this registry, we selected those
with a positive sputum culture (Figure 1).
From the Med-Echo data, we selected all registered hospitalizations
for patients with a principal diagnosis of TB. The coding of diseases in
Med-Echo is based on the 9th revision of the International
Classification of Diseases (ICD) from April 1, 1992 to March 31, 2006,
and thereafter on the 10th revision of the ICD. For patients with
multiple TB hospitalizations, we only selected the first one.
We then linked all culture-positive pulmonary TB cases from the
MADO registry with selected hospitalizations from the Med-Echo data
(Figure 1). The linking criteria included the first 3 characters of
postal code of residence, sex and age (plus or minus one year). For the
linked cases, we calculated the difference between the date of
notification of a TB case to the public health department (based on the
MADO registry) from the date of patient's admission to a hospital
for TB (based on Med-Echo registry). We accepted a difference of up to
150 days if hospitalization occurred before notification, and up to 120
days if hospitalization occurred after notification. We also eliminated
patients who were linked more than once from the MADO registry to the
Med-Echo registry, and those who stayed in long-term care hospitals or
hospitals outside Montreal (Figure 1).
The data were analyzed using the Statistical Package for the Social
Sciences, version 12.0 (SPSS Inc., Chicago, IL, USA). The median
duration of hospitalization and the interquartile range (IQR) were
calculated for the whole study period and for each separate year. Age,
sex, and country of birth (Canada or outside of Canada), results of
sputum smear, presence of radiological pulmonary cavitary lesion(s),
presence of resistance to anti-TB medication(s), and starting anti-TB
medication regimen (i.e., whether the patient is started on the
isoniazid (INH), rifampine (RMP), ethambutol (EMB) and pyrazinamide
(PZA) combination or not), period of hospitalization (1993-97,
1998-2002, 2003-07), and hospital TB volume were examined as independent
variables for patients' duration of hospitalization. For hospital
TB volume, our final linked sample of patients was used to divide
hospitals into two categories: lower volume, i.e., those with <2 TB
patients per year, and higher volume, i.e., those with [greater than or
equal to] 2 TB patients per year, on average. Pearson's [chi
square] test of significance was applied to calculate the p-value for
cross-tabulated variables. Logistic regression was used to further test
the association between the variables deemed significant in the
bivariate analysis (p<0.05). Odds ratios (OR) with 95% confidence
intervals (CI) were calculated.
[FIGURE 1 OMITTED]
RESULTS
From 1993 to 2007, there were 1,697 TB hospitalizations of all
types registered in Med-Echo for residents in the Montreal region. After
linking the data from Med-Echo and MADO, we found 633 cases who were
hospitalized (first episode) with culture-positive pulmonary TB. Upon
eliminating those with more than one link, and those who stayed in
long-term care hospitals or hospitals outside Montreal, we were left
with 563 cases (Figure 1).
The study cases had a mean age of 46.5 (standard deviation = 21.8);
44.6% were hospitalized in 3 hospital centres in Montreal. In total, 18
acute care hospital centres had at least one hospitalized patient with
culture-positive pulmonary TB. On admission, 48% of cases had positive
sputum smears, and 38.4% had radiological evidence of pulmonary cavitary
lesion(s) (Table 1); 14.2% had a strain resistant to at least one
anti-TB medication, and 2.3% had multi-drug resistant TB.
INH/RMP/EMB/PZA was the starting anti-TB medication combination for 416
(73.9%) cases.
Median duration of hospitalization was 17.0 days (IQR: 10 to 28)
for the whole study period. The minimum and maximum yearly medians for
duration of hospitalization were 13.0 days (IQR: 7 to 19) and 22 days
(IQR: 7 to 19) in 1993 and 2003, respectively. Among cases, 203 (36.1%)
remained in hospital for less than 14 days. The duration of
hospitalization was not significantly associated with the time period of
hospitalization (Table 1). The age group of cases and the presence of
smear-positive sputum were the only two independent variables associated
with the duration of hospital stay. In a logistic regression model
including these two variables, cases with a positive sputum smear were
more likely to stay in hospital 14 days or more compared to those with
no positive sputum smear (OR=1.90, 95% CI: 1.34-2.70). With respect to
age, TB cases [greater than or equal to] 50 years of age stayed in
hospital longer than those between 18-49 years of age (OR=1.66, 95% CI:
1.15-2.40).
DISCUSSION
Our results indicate that the median hospital stay for cases with
pulmonary TB in Montreal during their first episode of hospitalization
was 17 days. The duration of hospital stay has not changed significantly
over the years from 1993 to 2007. We observed that cases 50 years of age
and older had generally longer hospital stays than those in the middle
18-49 age group. Furthermore, our data suggested that cases with
positive sputum smear results were more likely to have longer stays.
Initial hospitalization of newly diagnosed TB patients, although
not mandated in Canada, is a common practice. Hospitalization is an
especially important consideration when the patient is highly
infectious, since ensuring airborne isolation is often not feasible
outside hospitals or other appropriate institutions. The challenge of
preventing outpatient secondary transmission of TB is further
complicated in densely populated and multicultural urban centres, like
Montreal, with many population subgroups at high risk for active and
latent TB infection (LTBI). (11)
With regard to duration of hospital stay, our results are
comparable to those from other studies in Canada and the United States.
In Canada, data supplied by the Canadian Institute for Health
Information indicate that for the period 1996-2000, 50.2% of TB patients
were hospitalized, and that the average duration of hospital stay for
these patients was 20.6 days. (6) Holmquist et al. report an average
duration of 15 days for patients hospitalized in the US with a principal
diagnosis of TB. (5)
CXR abnormalities (including involvement of both lungs, cavitation,
pleural effusion and miliary pattern), positive sputum AFB results, body
temperature and MDR status were shown to be significantly correlated
with the duration of hospital stay and with hospitalization cost in
previous studies. (4) We did not find a significant association between
radiological findings and hospital stay. However, we only considered the
presence of any cavitary lesion, without further restricting the finding
to more specific TB radiological findings involving both lungs. The
number of MDR cases in the current study was too small to establish any
meaningful association with the duration of hospital stay.
The linking formula used to identify hospitalized pulmonary TB
patients was not perfect and may represent a limitation of the study. As
indicated, Med-Echo, the hospitalization data source, is anonymous and
does not include the case's name, health insurance number or any
other unique case identifier that is in common with MADO, the second
data source. It is therefore possible that some patients were mis-linked
between the two databases used. On the other hand, since the total
number of patients hospitalized with pulmonary TB in Montreal per year
is limited, the probability of mis-linking should be low. Hence, we
believe that the number of mis-linked cases should not significantly
affect our study conclusions. Moreover, our linking procedure left a
significant number of TB patients unaccounted for. However, we compared
the original data (n=1697) with our post-linking sample (n=563) with
respect to demographic frequencies of sex and age groups, and the
datasets were comparable. Therefore, we conclude that our post-linking
sample, although significantly smaller in number, is likely a fair
representation of all hospital cases.
Our study demonstrates that despite potential indications for
longer hospital stays, 36% of patients with pulmonary TB were kept in
hospital for less than the recommended 14-day period. However, the
implications of this finding in the community are unclear. We recommend
further research to investigate the association between premature
discharge of patients with active pulmonary TB during their
hospitalization and possible increased risk of TB transmission in the
community.
REFERENCES
(1.) Gallant V, Scholten D, Dawson K, Ellis E. Tuberculosis in
Canada 2008 (Pre-Release). Public Health Agency of Canada, 2009.
Available at: http://www.publichealth.gc.ca/tuberculosis (Accessed
January 15, 2010).
(2.) Wicks CA. Tuberculosis: Hospitalization and outpatient
treatment. CMAJ 1964;91:380-84.
(3.) Hoeppner VH, Ward H, Elwood K. Treatment of tuberculosis
disease and infection. In: Long R, Ellis E (Eds.), Canadian Tuberculosis
Standards. Ottawa, ON: Public Health Agency of Canada, 2007.
(4.) Bocchino M, Greco S, Rosati Y, Mattioli G, Marruchella A, De
Mori P, et al. Cost determinants of tuberculosis management in a
low-prevalence country. Int J Tuberc Lung Dis 2006;10(12):146-52.
(5.) Holmquist L, Russo CA, Elixhauser A. Tuberculosis Stays in
U.S. Hospitals, 2006. Healthcare Cost & Utilization Project (Agency
for Healthcare Research & Quality), 2008. Available at:
http://www.hcup-us.ahrq.gov/reports/statbriefs/sb60.pdf (Accessed
January 15, 2010).
(6.) Menzies D, Lewis M, Oxlade O. Costs for tuberculosis care in
Canada. Can J Public Health 2008;99(5):391-96.
(7.) Marks SM, Taylor Z, Burrows NR, Qayad MG, Miller B.
Hospitalization of homeless persons with tuberculosis in the United
States. Am J Public Health 2000;90(3):435-38.
(8.) Goldstein RS, Contreras M, Craig GA, Cheung OT.
Tuberculosis--A review of 498 recent admissions to hospital. CMAJ
1982;126(5):490-92.
(9.) Ministere de la Sante et des Services sociaux (MSSS).
Epidemiologie de la tuberculose au Quebec de 2004 a 2007. Quebec, QC:
MSSS, 2009. Available at:
http://msssa4.msss.gouv.qc.ca/fr/document/publication.nsf/961885cb24e4e9
fd85256b1e00641a29/bb549fb28d0a511a8525758d00412dec?OpenDocument
(Accessed January 15, 2010).
(10.) Canadian Labour & Business Centre. Physicians Workforce
in Canada: Literature Review & Gap Analysis (Final Report). A
Physician Human Resource Strategy for Canada (Task Force Two), 2003.
(11.) Haase I, Olson S, Behr MA, Wanyeki I, Thibert L, Scott A, et
al. Use of geographic and genotyping tools to characterize tuberculosis
transmission in Montreal. Int J Tuberc Lung Dis 2007;11(6):632-38.
Received: May 11, 2010
Accepted: November 1, 2010
Author Affiliations
Kianoush Dehghani, MD, MPH, [1] Robert Allard, MDCM, MSc, FRCPC,
[2,3] Jean Gratton, MSc, [2] Louise Marcotte, BSc, [2] Paul Rivest, MD,
MSc [2,4]
[1] Community Medicine Residency Program, McGill University,
Montreal, QC
[2] Direction de sante publique, Agence de la sante et des services
sociaux, Montreal, QC
[3] Department of Epidemiology, Biostatistics and Occupational
Health, McGill University, Montreal, QC
[4] Departement de medecine sociale et preventive, Universite de
Montreal, Montreal, QC
Correspondence and reprint requests: Dr. Paul Rivest, Direction de
sante publique de Montreal, 1301 Sherbrooke Street East, Montreal, QC
H2L 1M3, Tel: 514-528-2400, ext. 3678, Fax: 514-528-2452, E-mail:
privest@santepub-mtl.qc.ca
Conflict of Interest: None to declare.
Table 1. Characteristics of Hospitalized Patients With Culture-positive
Pulmonary TB and Factors Possibly Associated With Duration, Montreal,
1993-2007
Characteristics
Hospital Stay
1-13 Days
([dagger])
n (%)* n (%)
Sex
Male 341 (60.6) 75 (33.8)
Female 222 (39.4) 128 (37.5)
Age groups (years)
1-17 22 (3.9) 7 (31.8)
18-49 315 (56.0) 129 (41.0)
[greater than or equal to] 50 226 (40.1) 67 (29.6)
Country of birtht
Outside Canada 437 (77.6) 158 (36.2)
Canada 124 (22.0) 45 (36.3)
Sputum smear
AFB ([section]) positive 270 (48.0) 77 (28.5)
AFB negative/Unknown ([parallel]) 293 (52.0) 126 (43.0)
Chest X-ray ([paragraph])
With cavitary lesion(s) 216 (38.4) 73 (33.8)
No cavitary lesion 340 (61.6) 126 (37.1)
Resistance to anti-TB medications
No resistance 483 (85.8) 176 (36.4)
To [greater than or equal to] 1
anti-TB medication 80 (14.2) 27 (33.8)
Initial anti-TB medication regimen
INH+RMP+EMB+PZA ** 416 (73.9) 144 (34.6)
Not INH+RMP+EMB+PZA 147 (26.1) 59 (40.1)
Year of hospitalization
1993-1997 215 (38.2) 80 (37.2)
1998-2002 192 (34.1) 74 (38.5)
2003-2007 156 (27.7) 49 (31.4)
Hospital TB volume
Lower volume 198 (35.2) 78 (39.4)
Higher volume 365 (64.8) 125 (34.2)
Total 563 (100) 203 (36.1)
Hospital Stay
[greater than or
equal to] 14
Days ([dagger])
n (%) p-value
Sex
Male 147 (66.2) 0.365
Female 213 (62.5)
Age groups (years)
1-17 15 (68.2) 0.024
18-49 186 (59.0)
[greater than or equal to] 50 159 (70.4)
Country of birtht
Outside Canada 279 (63.8) 0.978
Canada 79 (63.7)
Sputum smear
AFB ([section]) positive 193 (71.5) 0.000
AFB negative/Unknown ([parallel]) 167 (57.0)
Chest X-ray ([paragraph])
With cavitary lesion(s) 143 (66.2) 0.434
No cavitary lesion 214 (62.9)
Resistance to anti-TB medications
No resistance 307 (63.6) 0.643
To [greater than or equal to] 1
anti-TB medication 53 (66.3)
Initial anti-TB medication regimen
INH+RMP+EMB+PZA ** 272 (65.4) 0.231
Not INH+RMP+EMB+PZA 88 (59.9)
Year of hospitalization
1993-1997 135 (62.8) 0.350
1998-2002 118 (61.5)
2003-2007 107 (68.6)
Hospital TB volume
Lower volume 120 (60.6) 0.225
Higher volume 240 (65.8)
Total 360 (63.9) --
* column percent
([dagger]) row percent
([double dagger]) Data are missing for 2 patients
([section]) AFB: Acid-fast bacilli
([parallel]) Includes patients with negative sputum AFB, patients with
no sputum smears, and missing data
([paragraph]) Results not available or data missing for 7 patients
** INH: isoniazide; RMP: rifampine; EMB: ethambutol; PZA: pyrazinamide
