Active screening for tuberculosis in high-incidence Inuit communities: a cost-effectiveness analysis
Additional File 1
TABLE OF CONTENTS
Clinical Pathways & TB Cascade.
Table S2. Calculating TB cascade parameters using 2019 program data
Secondary Infections & Active Cases.
Table S3. Number of secondary active TB cases per index case by smear status in both villages
Incorporation of Additional Strategies.
Results: Additional Strategies
Table S4. Outcomes over 20 years in Village 1 and Village 2, given a single outbreak in 2019
Figure S1. Cost outcomes over 20 years for all scenarios in Village 1 and Village 2.
Figure S2. Number of active TB cases over 20 years for all scenarios in Village 1 and Village 2.
Table S8. All scenario analysis results
Results: Univariate Sensitivity Analyses
Results: Probabilistic Sensitivity Analyses
Figure S7. Probabilistic sensitivity analysis for Village 1 given a single outbreak
Figure S8. Probabilistic sensitivity analysis for Village 2 given a single outbreak
Figure S9. Probabilistic sensitivity analysis for Village 1 given repeated outbreaks
Figure S10. Probabilistic sensitivity analysis for Village 2 given repeated outbreaks
The following sections describe model parameters, methods for calculating secondary infections and secondary active TB cases, methods for simulating repeated outbreaks, costs considered, all strategies that were simulated using data from both villages, and all scenario analyses.
Screening Campaigns. In 2019, the Nunavik Regional Board of Health and Social Services (NRBHSS) led active screening campaigns in Village 1 (population approximately 1,000) and Village 2 (population approximately 1,500). These campaigns were community-wide: anyone not already known and evaluated for active TB or LTBI was eligible, without age restrictions. This meant that in Village 1, approximately 60% of the population was eligible for screening, and in Village 2, approximately 70% were eligible. The NRBHSS worked with local staff as well as staff flown into the villages to organize these screening campaigns. Several types of staff were required to operationalize screening activities, such as nurses, translators, HR staff, physicians, public health practitioners, administrative staff, and communications officers. Community members were engaged by local public health authorities using educational materials and encouraged to participate through incentives such as prize draws. Screening included on-site tuberculin skin testing and chest radiographs. In both villages, over 90% of individuals who were eligible for screening in 2019 participated in the screening campaigns.
Model Parameters. Table S1 lists all model parameters for both villages, separated into key parameter categories. Table S2 illustrates how TB cascade parameters (for both active TB and LTBI) were impacted by the presence of active screening, in both villages.
Table S1. Model parameters
Parameter |
Type of Distribution for Probabilistic Sensitivity Analysis |
Value in Village 1 (Range for Probabilistic Sensitivity Analysis) |
Value in Village 2 (Range for Probabilistic Sensitivity Analysis) |
Source |
TB PATHOGENESIS |
||||
Probability of progression to active TB† |
Beta |
0.05-0.265 (range 25%) |
0.05-0.265 (range 25%) |
[1] |
Probability of reactivation to active TB† |
Beta |
0.0005-0.075 (range 25%) |
0.0005-0.075 (range 25%) |
[2,3] |
Annual risk of infection |
Beta |
0.0095 (0.0071 - 0.0119) |
0.0095 (0.0071 - 0.0119) |
[4] |
Average secondary infections per index TB case |
Uniform |
0.67 (0.50 – 0.84) |
0.67 (0.50 – 0.84) |
[1,5] |
Average secondary active TB cases per index TB case |
Uniform |
1.82 (1.37 – 2.28) |
1.23 (0.92 - 1.54) |
[6] |
Immunity conferred by previous infection |
Beta |
0.55 (0.4125 - 0.6875) |
0.55 (0.4125 - 0.6875) |
[3] |
Probability of recovery following active TB treatment |
Beta |
0.928 (0.9 - 1) |
0.928 (0.9 - 1) |
[4] |
Probability of recovery following complete LTBI treatment |
Beta |
0.875 (0.9 – 0.925) |
0.875 (0.9 – 0.925) |
[7] |
Probability of recovery following incomplete LTBI treatment |
Beta |
0.21 (0 – 0.3) |
0.21 (0 – 0.3) |
[3] |
Probability of relapse following recovery from active TB treatment |
Beta |
0.015 (0.0075 - 0.025) |
0.015 (0.0075 - 0.025) |
[8] |
Probability of spontaneous recovery from active TB |
Beta |
0.25 (0.2 - 0.3) |
0.25 (0.2 - 0.3) |
[9] |
Probability of dying from active TB during treatment |
Beta |
0 (0 - 0) |
0 (0 - 0) |
[7] |
Probability of dying of untreated TB if smear negative |
Beta |
0.02 (0.015 - 0.025) |
0.02 (0.015 - 0.025) |
[8] |
Probability of dying of untreated TB if smear positive |
Beta |
0.07 (0.053 - 0.086) |
0.07 (0.053 - 0.086) |
[8] |
CASCADE PROBABILITIES†† |
||||
Probability of diagnosing active TB (with active screening) |
NA |
1 |
1 |
[10] |
Probability of diagnosing active TB (without active screening)** |
NA |
0.82 |
0.56 |
[10] |
Probability of diagnosing new LTBI (with active screening) |
NA |
1 |
0.99 |
[10] |
Probability of diagnosing new LTBI (without active screening)** |
NA |
0.83 |
0.71 |
[10] |
Probability of diagnosing longstanding LTBI (with active screening) |
NA |
1 |
1 |
[10] |
Probability of diagnosing longstanding LTBI (without active screening)** |
NA |
0.983 |
0.998 |
[10] |
Probability of coming for a follow up visit to read TST |
NA |
0.86 |
0.86 |
[22] |
Probability of starting active TB treatment when diagnosed (with or without active screening) |
NA |
1 |
1 |
Assumption |
Probability of starting latent TB treatment (without active screening) |
NA |
0.7 |
0.7 |
[6] |
Probability of starting treatment for new infection (with active screening) |
NA |
0.72 |
0.71 |
[10] |
Probability of starting treatment for longstanding infection (with active screening) |
NA |
0.69 |
0.7 |
[10] |
Probability of completing active TB treatment (with or without active screening) |
NA |
0.997 |
0.997 |
[6] |
Probability of completing LTBI treatment (with or without active screening)* |
NA |
0.75 |
0.6 |
[10] |
Median proportion of total doses taken among those who don't complete treatment for LTBI |
Beta |
0.25 (0.18 – 0.5) |
0.25 (0.18 – 0.5) |
[21] |
OTHER PROBABILITIES |
||||
Proportion of population at baseline in LTBI states |
NA |
48% |
33% |
[6] |
Proportion of population at baseline in susceptible states |
NA |
49% |
66% |
[6] |
Proportion of Population at baseline in active TB states |
NA |
3% |
1% |
[6] |
Probability of adverse event during active TB treatment |
Beta |
0.051 (0.01 - 0.1) |
0.051 (0.01 - 0.1) |
[11] |
Probability of adverse event during LTBI treatment |
Beta |
0.003 (0.001 - 0.0045) |
0.003 (0.001 - 0.0045) |
[23] |
Probability of non-TB-related death (background mortality)†† |
NA |
0.014-0.021 |
0.014-0.021 |
[24] |
Probability of being smear positive (with active screening)†† |
NA |
0.12 |
0.13 |
[6] |
Probability of being smear positive (without active screening)†† |
NA |
0.15 |
0.11 |
[6] |
Probability that someone being evaluated for LTBI undergoes sputum examination†† |
NA |
0.23 |
0 |
[6] |
Probability of producing adequate sputum sample |
Beta |
0.82 (0.747 – 0.896) |
0.82 (0.747 – 0.896) |
[20] |
Number of days hospitalized if smear negative |
Uniform |
14 (11 - 18) |
14 (11 - 18) |
[6] |
Number of days hospitalized if smear positive |
Uniform |
60 (45 - 75) |
60 (45 - 75) |
[6] |
Number of days hospitalized for suspicion of active TB |
Uniform |
1 (0 - 1) |
1 (0 - 1) |
[6] |
Probability of being hospitalized for suspicion of active TB |
Beta |
0.05 (0.0375 - 0.0625) |
0.05 (0.0375 - 0.0625) |
Assumption |
COSTS RELATED TO ACTIVE SCREENING |
|
|
|
|
Cost of active screening: amount spent on local amenities |
Triangular |
$10,821 ($8,116 - $13,526) |
$41,118 ($30,839 - $51,398) |
[10] |
Cost of active screening: average spent on car rental |
Triangular |
$2,391 ($1,793 - $2,989) |
$28,140 ($21,105 - $35,175) |
[10] |
Cost of active screening: average spent on charter flight |
Triangular |
$3,563 ($2,672 - $4,454) |
$23,703 ($17,777 - $29,629) |
[10] |
Cost of active screening: average spent per cleaner |
Triangular |
$338 ($254 - $423) |
$0 ($0 - $0) |
[10] |
Cost of active screening: amount spent on communication and mobilization |
Triangular |
$2,865 ($2,149 - $3,581) |
$30,228 ($22,671 - $37,785) |
[10] |
Cost of active screening: average spent on construction for lodging |
Triangular |
$32,153 ($24,115 - $40,191) |
$0 ($0 - $0) |
[10] |
Cost of active screening: average spent per consultant |
Triangular |
$12,095 ($9,071 - $15,119) |
$10,212 ($7,659 - $12,765) |
[10] |
Cost of active screening: average spent per driver |
Triangular |
$1,434 ($1,076 - $1,793) |
$0 ($0 - $0) |
[10] |
Cost of active screening: amount spent on equipment and materials related to lodging and transport |
Triangular |
$4,755 ($3,566 - $5,944) |
$1,753 ($1,315 - $2,191) |
[10] |
Cost of active screening: average spent on additional hotel stay per staff |
Triangular |
$0 ($0 - $0) |
$3,983 ($2,987 - $4,979) |
[10] |
Cost of active screening: average spent per human resources and logistics staff |
Triangular |
$225,000 ($168,750 - $281,250) |
$0 ($0 - $0) |
[10] |
Cost of active screening: average spent per lab technician |
Triangular |
$9,535 ($7,151 - $11,919) |
$0 ($0 - $0) |
[10] |
Cost of active screening: average spent per local staff |
Triangular |
$0 ($0 - $0) |
$15,346 ($11,510 - $19,183) |
[10] |
Cost of active screening: average spent to lodge each staff member |
Triangular |
$2,984 ($2,238 - $3,730) |
$5,452 ($4,089 - $6,815) |
[10] |
Cost of active screening: average spent per nurse |
Triangular |
$19,539 ($14,654 - $24,424) |
$21,148 ($15,861 - $26,435) |
[10] |
Cost of active screening: average spent on other staff |
Triangular |
$0 ($0 - $0) |
$100,649 ($75,487 - $125,811) |
[10] |
Cost of active screening: average spent per pharmacy technician |
Triangular |
$11,695 ($8,771 - $14,619) |
$19,505 ($14,629 - $24,381) |
[10] |
Cost of active screening: amount spent on supplies |
Triangular |
$29,405 ($22,054 - $36,756) |
$31,902 ($23,927 - $39,878) |
[10] |
Cost of active screening: amount spent on training and workshops |
Triangular |
$1,256 ($942 - $1,570) |
$29,692 ($22,269 - $37,115) |
[10] |
Cost of active screening: average spent per translator |
Triangular |
$1,336 ($1,002 - $1,670) |
$4,200 ($3,150 - $5,250) |
[10] |
Cost of active screening: average for other lodging costs |
Triangular |
$0 ($0 - $0) |
$3,870 ($2,903 - $4,838) |
[10] |
COSTS RELATED TO MANAGEMENT OF TB AND LTBI |
||||
Cost of adverse event due to active TB treatment |
Triangular |
$16,364 ($12,273 - $20,455) |
$16,364 ($12,273 - $20,455) |
[11] |
Cost of adverse event during LTBI treatment |
Triangular |
$782 ($587 - $978) |
$782 ($587 - $978) |
[12] |
Cost of chest x-ray |
Triangular |
$31 ($23 - $39) |
$31 ($23 - $39) |
[13] |
Cost of DOT for active TB |
Triangular |
$197 ($148 - $246) |
$197 ($148 - $246) |
[4,14] |
Cost of medication for active TB |
Triangular |
$674 ($506 - $843) |
$674 ($506 - $843) |
[15] |
Cost of medication for latent TB |
Triangular |
$114 ($86 - $143) |
$114 ($86 - $143) |
[15] |
Cost of flight (non-medical evacuation charter) |
Triangular |
$305 ($229 - $381) |
$305 ($229 - $381) |
[10] |
Cost of medical evacuation charter to regional hospital |
Triangular |
$6,713 ($5,035 - $8,391) |
$6,713 ($5,035 - $8,391) |
[16] |
Cost of follow up visit |
Triangular |
$9 ($7 - $11) |
$9 ($7 - $11) |
[17, 18] |
Cost of hospitalization per day |
Triangular |
$2,050 ($1,538 - $2,563) |
$2,050 ($1,538 - $2,563) |
[19] |
Cost of induced sputum collection |
Triangular |
$76 ($57 - $95) |
$76 ($57 - $95) |
[13] |
Cost of spontaneous sputum collection |
Triangular |
$65 ($49 - $81) |
$65 ($49 - $81) |
[13] |
Cost of sending sputum samples to regional hospital |
Triangular |
$4 ($3 - $5) |
$4 ($3 - $5) |
[20] |
Cost of physical exam |
Triangular |
$167 ($125 - $209) |
$167 ($125 - $209) |
[13,17] |
Cost of TST |
Triangular |
$34 ($26 - $43) |
$34 ($26 - $43) |
[10,13,17] |
Cost of visits to manage active TB treatment |
Triangular |
$436 ($327 - $545) |
$436 ($327 - $545) |
[6,17] |
Cost of visits to manage LTBI treatment |
Triangular |
$41 ($31 - $51) |
$41 ($31 - $51) |
[6,17,18] |
Cost of GeneXpert analysis |
Triangular |
$69 ($52 - $86) |
$69 ($52 - $86) |
[20] |
NA = not applicable
Assumptions were vetted by regional experts
† The probabilities of progression and reactivation change over time. At each time point, a 25% range was evaluated in probabilistic sensitivity analysis
†† Cascade probabilities (with the exception of median proportion of total LTBI treatment doses taken) reflect the realities of the two communities. As such, we did not consider fluctuations in these parameters in probabilistic sensitivity analysis. We did, however, consider changes in certain cascade probabilities in scenario analyses. For similar reasons, parameters such as the background mortality and probability of being smear positive were not considered in probabilistic sensitivity analysis.
*Assumption that probability of completing LTBI treatment does not change with the addition of active screening
**Assumption that persons found through active screening would not have been diagnosed otherwise
Clinical Pathways & TB Cascade. We considered clinical pathways for individuals with active TB or LTBI who were undiagnosed or diagnosed and not treated. Individuals with LTBI who fell into this category eventually moved into the longstanding LTBI branch, where they could recover without becoming reinfected, recover and become reinfected, or reactivate to active TB. In the case that community-wide screening was repeated in subsequent cycles of the model, these individuals could be diagnosed with LTBI and would follow the respective treatment pathway. On the other hand, individuals with active TB who were undiagnosed or diagnosed and not treated had a probability of dying or remaining with active TB otherwise, as well as spontaneous recovery. Notably, the probabilities of diagnosing and treating individuals with active TB were high, so most were diagnosed and treated (Table S2).
Active screening may influence probabilities of diagnosis, treatment initiation and treatment completion for both active TB and LTBI. We assumed that any persons with LTBI detected during active screening in each community would not have otherwise been found, as active screening supplemented established, ongoing contact investigation practices. Similarly, we assumed that persons with active TB found as a result of active screening in each community would have otherwise been found after symptom onset, i.e. when they became more infectious. We used data from years when there was no active screening, as well as data from 2019, when active screening took place, to inform changes in TB cascade parameters, as described in Table S2.
Table S2. Calculating TB cascade parameters using 2019 program data
TB Cascade Variable |
Value in Village 1 |
Value in Village 2 |
Reference |
Active TB |
|||
DIAGNOSIS |
|
|
|
(a) true number of individuals with active TB at the beginning of 2019 |
33 |
16 |
Calculated |
(b) number of individuals with known active TB at the beginning of 2019 |
27 |
9 |
[6] |
(c) number of additional individuals diagnosed during active screening |
6 |
7 |
[10] |
(d) diagnosis rate in absence of active screening; (b) (a) |
82% |
56% |
Calculated |
(e) diagnosis rate in presence of active screening; (b + c) (a) |
100% |
100% |
Calculated |
TREATMENT INITIATION |
|
|
|
(f) treatment initiation rate in absence of active screening |
100% |
100% |
[6] |
(g) treatment initiation rate in presence of active screening |
100% |
100% |
[6] |
TREATMENT COMPLETION |
|
|
|
(h) treatment completion rate in absence of active screening |
99.7% |
99.7% |
[4] |
(i) treatment completion rate in presence of active screening |
99.7% |
99.7% |
[4] |
LTBI* |
|
|
|
DIAGNOSIS |
|
|
|
(a) true number of individuals with LTBI at the beginning of 2019 |
138 |
104 |
Calculated |
(b) number of individuals with known LTBI at the beginning of 2019 |
114 |
74 |
[6] |
(c) number of additional individuals diagnosed during active screening |
23 |
29 |
[10] |
(d) diagnosis rate in absence of active screening; (b) (a) |
83% |
71% |
Calculated |
(e) diagnosis rate in presence of active screening; (b + c) (a) |
100% |
99% |
Calculated |
TREATMENT INITIATION |
|
|
|
(f) among those actively screened, number who start treatment |
19 |
21 |
[10] |
(g) treatment initiation rate in absence of active screening |
70% |
70% |
[6] |
(h) treatment initiation rate in presence of active screening; (b + c) (b*g + f) |
72% |
71% |
Calculated |
TREATMENT COMPLETION |
|
|
|
(i) treatment completion rate in absence of active screening |
75% |
60% |
[6] |
(j) treatment completion rate in presence of active screening |
75% |
60% |
[6] |
* LTBI refers to “new” infection, meaning infection among individuals who either had previously unknown status or had tested negative for LTBI
Secondary Infections & Active Cases. The total predicted number of incident active TB cases was used to estimate secondary infections and secondary active TB cases, using a ratio of secondary infections and active cases per index case that was informed by historical regional data [6]. The number of secondary infections was calculated by multiplying the number of incident active TB cases by the average number of household contacts in Nunavik (2.36) [5] and the average proportion of household contacts with secondary infection (26%; not specific to Nunavik) [1]. Similarly, the number of secondary active TB cases was calculated by multiplying the number of incident active TB cases by the proportion of smear positive individuals and the average number of secondary cases per index case, as listed in Table S3. Data in Table S3 comes from pooled estimates during both outbreak and non-outbreak years in the communities. We were not able to obtain the same type of data for the average number of persons with new LTBI per index active TB case, hence the other method of calculation.
Table S3. Number of secondary active TB cases per index case by smear status in both villages
Average number of secondary cases by index case in 2017-2019 |
Village 1 |
Village 2 |
Smear-positive |
4.9 |
2.8 |
Smear-negative |
1.4 |
1.0 |
The proportion of smear positive individuals (out of all individuals diagnosed with active TB) in Village 1 was 12% and 13% in Village 2, in 2019 [6].
In other words,
Number of secondary infections = number of incident active TB cases * average number of household contacts * proportion of household contacts with secondary infection
Number of secondary active TB cases = number of incident active TB cases * proportion of smear positive individuals in the village * average number of secondary active TB cases per index case dependent on smear status and village
The clinical pathway for secondary infections and secondary active TB cases resembled that of new infections and index active TB cases. We took measures not to double count these individuals; during each cycle of the model, persons with secondary LTBI and secondary active TB were taken from the pool of susceptible individuals.
Longstanding Infections. Longstanding infections were calculated in each cycle of the model’s analytic horizon as the sum of those with LTBI who were (1) never diagnosed, (2) diagnosed, but never treated, and (3) diagnosed, treated, but never completed treatment. These individuals were especially important for simulating repeated outbreaks, as described in the following section.
Simulating Outbreaks. Outbreaks were simulated by changing three parameters in the two villages: the annual risk of infection (i.e. the transmission parameter), the probability of progression, and the probability of reactivation. First, the annual risk of infection was set to 5% during outbreak years, in lieu of 0.95%. 5% is the upper bound of the estimate for the annual risk of infection that was derived from literature in Nunavut, a jurisdiction that has had a similar pattern of outbreaks to communities in Nunavik [1]. There is evidence from Village 1 that the annual risk of infection increased during outbreaks [6]. Second, we increased the rate at which individuals would progress or reactivate to active TB. Both progression and reactivation parameters started off high (26.5% and 7.5%, respectively), reflecting data from a Nunavik community in 2010, which was experiencing an outbreak at the time [1]. In the absence of outbreaks, both of these parameters would decline over the course of five years to 5% and 0.05%, respectively. However, in order to simulate outbreaks, these two parameters would increase to their baseline values (26.5% and 7.5%) during the cycle of the model that the outbreak occurred. Following an outbreak cycle, the three parameters would go back to their lower values. Because the outbreak in 2010 was large, we tested lower peaks of progression and reactivation in scenario analyses.
Costs. All costs were considered from the health system perspective, meaning that any out-of-pocket costs borne by patients were not included—although there were no direct charges made to patients. Cost inputs fell into two categories. The first category included costs related to active screening. These costs came from Nunavik program data and reflected the steps needed to conduct active screening activities in both communities in 2019. The second category included costs related to standard TB care. Wherever possible, these costs came from Nunavik, or Nunavut when necessary. Where such information was unavailable, costs came from published literature, but were confirmed with regional experts. We considered several components of active TB treatment and LTBI treatment, such as medication, hospital visits, adverse events, hospitalization, and medical evacuation (in the case of severe illness requiring treatment in larger cities). Aside from medication and hospital visits, which applied to all undergoing treatment, some components of treatment (e.g. cost of adverse events) were pro-rated according to the probability of that event arising.
Incorporation of Additional Strategies. The following strategies were considered, given a single outbreak in 2019 (Strategies A and B were considered in the main text):
- No active screening: This strategy was the counterfactual; i.e. what was predicted to occur had no active screening program been introduced to either community in 2019. We used background rates of diagnosis, treatment initiation and treatment completion for both active TB and LTBI, informed by data from each community during years where there was no active screening.
- Community wide active screening in 2019: Both Village 1 and Village 2 had active, community-based screening programs in 2019. This strategy incorporated program data to reflect increased rates of diagnosis, treatment initiation, and treatment completion compared to Strategy A.
- Community wide active screening in 2019 and active screening for new infections only in 2020, given a single outbreak in 2019: This strategy used active screening data from 2019 to inform a hypothetical repeated screening effort in 2020. The key difference between the screening program in 2019 and 2020 in this strategy was that the screening in 2020 was solely focused on diagnosing new infections. Hence only those who were previously skin test negative, or had unknown infection status were eligible to be screened in 2020; we assumed that everyone eligible for screening in 2020 would be screened.
- Community wide active screening in 2019 and 2020, given a single outbreak in 2019: This strategy also incorporated repeated screening in 2020. The 2020 screening focused on both LTBI and active TB. Hence individuals without infection, with previously negative skin tests, and those with unknown status were eligible for LTBI screening, and all others were eligible for active TB screening. As with Strategy D, we assumed that all those eligible for repeat screening in 2020 would be screened accordingly.
The following strategies were considered, given an outbreak in 2019, and every three years following that (Strategies A, B and C were also considered in the main text):
- No active screening: This strategy was the counterfactual; i.e. what was predicted to occur had no active screening program been introduced to either community in 2019. We used background rates of diagnosis, treatment initiation and treatment completion for both active TB and LTBI, informed by data from each community during years where there was no active screening.
- Community wide active screening in 2019: Both Village 1 and Village 2 had active, community-based screening programs in 2019. This strategy incorporated program data to reflect increased rates of diagnosis, treatment initiation, and treatment completion compared to Strategy A.
- Community wide active screening every two years for twenty years total: Community wide active screening was simulated every two years for the entire analytic horizon of 20 years.
- Community wide active screening annually for twenty years total, in the presence of repeated outbreaks every three years: Similar to scenario C, but with screening conducted annually.
Scenario Analyses. Model assumptions were evaluated using extensive univariate sensitivity analysis as well as scenario analyses. Univariate sensitivity analyses allowed us to identify parameters that were most influential in driving model outcomes, and to incorporate uncertainty around each parameter’s point estimate. On the other hand, scenario analyses addressed specific assumptions related to program efficiencies and community characteristics. We considered the following seven scenarios:
- Strengthened LTBI cascade: Rates of LTBI diagnosis are already high in both communities, but rates of treatment initiation and completion may be improved (with active screening, they stand at approximately 70% in both communities). Strengthening the cascade is always a goal of TB care, so we considered a scenario where treatment initiation and completion rates were increased to 80%.
- Local staff for active screening: During active screening in both communities, many staff members were flown in from the South. Village 1 and Village 2, however, have their own community health workers. To mimic building local capacity, we considered a scenario where community health workers were not flown into the communities. All other staff, such as nurses, were still flown into both communities.
- Reduced adherence to the second round of screening: Strategies C, D, E, and F rely on multiple rounds of screening. In our base case analysis, we assumed that everyone who is eligible for repeated rounds of screening undergoes active screening, however, there may be some level of fatigue associated with consecutive courses of screening. In this scenario, we reduced the level of adherence to 50% for any screening event following the one in 2019.
- Outbreak Intensity x0.75: In the base case analyses (main text analyses), the probabilities of progression and reactivation jumped up to 26.5% and 7.5%, respectively, during an outbreak. These values reflect a relatively large outbreak that occurred in a Nunavik community in 2010. This scenario reduced the high values of progression and reactivation during outbreaks by 25%, so that they would jump up to 19.9% and 5.6%, respectively, instead.
- Outbreak Intensity x0.5: This scenario was very similar to the one above, except the high values of progression and reactivation were reduced by 50%, so that they would jump up to 13.3% and 3.8%, respectively.
- Outbreak Intensity Progressively Decreasing: This scenario assumed that outbreak intensity would progressively decrease each time an outbreak would occur. As such, outbreak intensity was high during the first outbreak, then reduced by 10-15 percentage points each outbreak, so that by the last outbreak, the values of progression and reactivation were 5.3% and 1.5% (80% reduction from 26.5% and 7.5%), respectively, which approaches their values during non-outbreak years.
- Lower rates of LTBI diagnosis: As shown in Table S2, the rates of diagnosis are quite high in both villages, even without active screening. In this scenario, we reduced rates of LTBI diagnosis, both with and without active screening, by 25%.
Results
Tables S4 and S5 illustrate the results for Strategies A, B, D and E, given a single outbreak in 2019. Tables S6 and S7 illustrate the results for Strategy A, B, C, and F, given an outbreak in 2019 and every three years following that. Figures S1 and S2 illustrate the results of all scenario analyses. Figures S3-S6 illustrate the results of univariate sensitivity analyses. Lastly, Figures S7-S10 illustrate the results of probabilistic sensitivity analyses.
Results: Additional Strategies
Table S4. Outcomes over 20 years in Village 1 and Village 2, given a single outbreak in 2019
Strategy* |
Cost |
Incident Active TB |
Incident LTBI |
Longstanding LTBI |
TB-Related Deaths |
Secondary Infections |
Secondary Active TB |
Village 1 |
|||||||
B |
$6,996,027 |
90 |
38 |
61 |
0.6 |
19 |
50 |
D |
$7,001,561 ($5,773,909 to $8,818,454) |
77 |
29 |
55 |
0.3 |
16 |
43 |
E |
$7,004,953 |
90 |
38 |
60 |
0.5 |
19 |
50 |
A |
$7,493,340 |
103 |
42 |
70 |
0.9 |
21 |
60 |
Village 2 |
|||||||
A |
$5,034,527 |
83 |
42 |
63 |
2.0 |
20 |
35 |
B |
$5,139,231 |
79 |
41 |
55 (51 to 60) |
1.6 |
19 |
34 |
D |
$5,645,676 |
68 |
33 |
50 (46 to 54) |
0.8 |
16 |
29 |
E |
$5,659,758 |
78 |
40 |
54 |
1.6 |
19 |
34 |
Values in parentheses indicate 95% uncertainty ranges.
*Strategy A: No active screening; Strategy B: Community wide active screening in 2019; Strategy D: Community wide active screening in 2019 and active screening for new infections only in 2020; Strategy E: Community wide active screening in 2019 and 2020
Table S5. Incremental cost per active TB case averted in Village 1 and Village 2, given a single outbreak in 2019
Strategy* |
Incremental cost per person compared to the preceding strategy |
Incremental cost per active TB case averted compared to preceding strategy |
Incremental cost per active TB case averted compared to Strategy A |
Village 1 |
|||
B |
-- |
-- |
Dominant** |
D |
$4 |
$442 |
Dominant** |
E |
$2 |
Dominated** |
Dominant** |
A |
$348 |
Dominated** |
-- |
Village 2 |
|||
A |
-- |
-- |
-- |
B |
$47 |
$22,134 |
$22,134 |
D |
$227 |
$48,382 |
$40,213 |
E |
$6 |
Dominated** |
$121,165 |
Values in brackets indicate 95% uncertainty ranges. Incremental cost per active TB case averted is the difference in costs divided by the difference in active TB cases (primary and secondary) between two strategies. The population of Village 1 at the end of the simulation was 1402 and the population of Village 2 was 2235.
*Strategy A: No active screening; Strategy B: Community wide active screening in 2019; Strategy D: Community wide active screening in 2019 and active screening for new infections only in 2020; Strategy E: Community wide active screening in 2019 and 2020
** Dominated means that a strategy is more costly and less effective than the one it is being compared to, while dominant means that a strategy is less costly and more effective than the one it is being compared to.
Table S6. Outcomes over 20 years in Village 1 and Village 2, given an outbreak in 2019 and every three years thereafter
Strategy* |
Cost |
Incident Active TB |
Incident LTBI |
Longstanding LTBI |
TB-Related Deaths |
Secondary Infections |
Secondary Active TB |
Village 1 |
|||||||
B |
$14,745,984 |
249 |
87 |
83 |
1.5 |
51 |
136 |
C |
$15,691,149 |
102 |
30 |
55 |
0.3 |
21 |
57 |
A |
$16,359,259 |
276 |
94 |
94 |
1.9 |
55 |
156 |
F |
$22,511,235 |
89 |
26 |
51 |
0.0 |
19 |
50 |
Village 2 |
|||||||
A |
$12,028,207 |
239 |
91 |
89 |
4.8 |
55 |
98 |
B |
$12,203,936 |
232 (211 to 248) |
88 |
81 |
4.5 |
54 |
97 |
C |
$15,008,450 |
99 |
38 |
50 |
0.6 |
24 |
43 |
F |
$22,097,123 |
83 |
32 |
46 |
0.0 |
20 |
36 |
Values in brackets indicate 95% uncertainty ranges.
* Strategy A: No active screening; Strategy B: Community wide active screening in 2019; Strategy C: Community wide active screening every two years for 20 years total; Strategy F: Community wide active screening annually for twenty years total
Table S7. Incremental cost per active TB case averted in Village 1 and Village 2, given an outbreak in 2019 and every three years thereafter
Strategy* |
Incremental cost per person compared to the preceding strategy |
Incremental cost per active TB case averted compared to preceding strategy |
Incremental cost per active TB case averted compared to Strategy A |
Village 1 |
|||
B |
-- |
-- |
Dominant |
C |
$674 (-$1,427 to $2,808) |
$6,430 |
Dominant |
A |
$477 (-$1,827 to $2,865) |
Dominated |
-- |
F |
$2,753 |
$32,797 |
$32,797 ($6,078 to $64,884) |
Village 2 |
|||
A |
-- |
-- |
-- |
B |
$79 |
$24,282 |
$24,282 |
C |
$1,255 |
$21,129 |
$21,292 |
F |
$3,172 |
$453,014 |
$64,702 |
Values in brackets indicate 95% uncertainty ranges. Incremental cost per active TB case averted is the difference in costs divided by the difference in active TB cases (primary and secondary) between two strategies. The population of Village 1 at the end of the simulation was 1402 and the population of Village 2 was 2235.
* Strategy A: No active screening; Strategy B: Community wide active screening in 2019; Strategy C: Community wide active screening every two years for 20 years total; Strategy F: Community wide active screening annually for twenty years total
** Dominated means that a strategy is more costly and less effective than the one it is being compared to, while dominant means that a strategy is less costly and more effective than the one it is being compared to.
Figure S1. Cost outcomes over 20 years for all scenarios in Village 1 and Village 2. X-axis legend: Strategy A: No active screening; Strategy B: Community wide active screening in 2019; Strategy C: Community wide active screening every two years for 20 years. Error bars represent 95% uncertainty ranges.
Figure S2. Number of active TB cases over 20 years for all scenarios in Village 1 and Village 2. X-axis legend: Strategy A: No active screening; Strategy B: Community wide active screening in 2019; Strategy C: Community wide active screening every two years for 20 years. Error bars represent 95% uncertainty ranges.
Table S8. All scenario analysis results
Outbreak Status |
Village |
Scenario |
Strategy |
Cost |
Active TB Cases |
Cost: Low |
Cost: High |
Active TB Cases: Low |
Active TB Cases: High |
Repeated |
1 |
Base Case |
A |
$16,359,259 |
276 |
$12,832,993 |
$20,867,686 |
251 |
295 |
Repeated |
1 |
Base Case |
B |
$14,745,984 |
249 |
$11,730,036 |
$18,714,978 |
226 |
266 |
Repeated |
1 |
Base Case |
C |
$15,691,149 |
102 |
$13,020,951 |
$18,980,447 |
90 |
116 |
Repeated |
1 |
Strengthened LTBI |
A |
$13,565,899 |
222 |
$10,648,005 |
$17,259,884 |
201 |
238 |
Repeated |
1 |
Strengthened LTBI |
B |
$12,053,920 |
194 |
$9,595,547 |
$15,201,632 |
176 |
209 |
Repeated |
1 |
Strengthened LTBI |
C |
$14,115,953 |
71 |
$11,579,326 |
$17,021,748 |
62 |
82 |
Repeated |
1 |
Outbreak Intensity x0.5 |
A |
$13,024,727 |
211 |
$10,298,825 |
$16,679,404 |
193 |
229 |
Repeated |
1 |
Outbreak Intensity x0.5 |
B |
$11,834,939 |
189 |
$9,449,061 |
$15,042,292 |
172 |
205 |
Repeated |
1 |
Outbreak Intensity x0.5 |
C |
$15,391,897 |
96 |
$12,787,179 |
$18,575,656 |
85 |
110 |
Repeated |
1 |
Outbreak Intensity x0.75 |
A |
$14,732,415 |
244 |
$11,523,026 |
$18,777,829 |
223 |
263 |
Repeated |
1 |
Outbreak Intensity x0.75 |
B |
$13,326,094 |
220 |
$10,543,138 |
$16,861,868 |
200 |
236 |
Repeated |
1 |
Outbreak Intensity x0.75 |
C |
$15,542,021 |
99 |
$12,901,575 |
$18,801,793 |
87 |
114 |
Repeated |
1 |
Outbreak Intensity Decreasing |
A |
$13,287,737 |
213 |
$10,494,208 |
$17,026,898 |
194 |
232 |
Repeated |
1 |
Outbreak Intensity Decreasing |
B |
$12,063,973 |
191 |
$9,664,294 |
$15,351,868 |
174 |
207 |
Repeated |
1 |
Outbreak Intensity Decreasing |
C |
$15,511,652 |
99 |
$12,864,214 |
$18,754,439 |
87 |
113 |
Repeated |
1 |
Local Staff |
A |
$16,359,259 |
276 |
$12,832,993 |
$20,867,686 |
251 |
295 |
Repeated |
1 |
Local Staff |
B |
$14,745,508 |
249 |
$11,574,572 |
$18,569,976 |
226 |
266 |
Repeated |
1 |
Local Staff |
C |
$15,677,288 |
102 |
$12,447,748 |
$18,342,142 |
90 |
116 |
Repeated |
1 |
Reduced Adherence |
A |
$16,359,259 |
276 |
$12,832,993 |
$20,867,686 |
251 |
295 |
Repeated |
1 |
Reduced Adherence |
B |
$14,745,984 |
249 |
$11,730,036 |
$18,714,978 |
226 |
266 |
Repeated |
1 |
Reduced Adherence |
C |
$17,411,846 |
144 |
$14,469,461 |
$21,013,495 |
125 |
164 |
Repeated |
1 |
Lower LTBI Diagnosis |
A |
$19,998,062 |
347 |
$15,261,861 |
$26,001,714 |
294 |
392 |
Repeated |
1 |
Lower LTBI Diagnosis |
B |
$18,143,036 |
321 |
$13,979,740 |
$23,601,984 |
270 |
366 |
Repeated |
1 |
Lower LTBI Diagnosis |
C |
$18,222,725 |
154 |
$14,977,013 |
$22,640,592 |
119 |
198 |
Repeated |
2 |
Base Case |
A |
$12,028,207 |
239 |
$9,462,816 |
$15,376,583 |
218 |
255 |
Repeated |
2 |
Base Case |
B |
$12,203,936 |
232 |
$9,613,500 |
$15,465,190 |
211 |
248 |
Repeated |
2 |
Base Case |
C |
$15,008,450 |
99 |
$13,665,701 |
$17,132,157 |
87 |
112 |
Repeated |
2 |
Strengthened LTBI |
A |
$8,809,318 |
168 |
$6,967,803 |
$11,260,810 |
154 |
181 |
Repeated |
2 |
Strengthened LTBI |
B |
$8,698,480 |
157 |
$6,899,541 |
$11,069,285 |
143 |
169 |
Repeated |
2 |
Strengthened LTBI |
C |
$12,700,053 |
53 |
$11,798,663 |
$14,256,772 |
46 |
61 |
Repeated |
2 |
Outbreak Intensity x0.5 |
A |
$9,430,327 |
180 |
$7,504,871 |
$12,169,181 |
165 |
196 |
Repeated |
2 |
Outbreak Intensity x0.5 |
B |
$9,579,346 |
174 |
$7,643,429 |
$12,280,009 |
159 |
189 |
Repeated |
2 |
Outbreak Intensity x0.5 |
C |
$14,617,531 |
91 |
$13,367,071 |
$16,683,944 |
80 |
104 |
Repeated |
2 |
Outbreak Intensity x0.75 |
A |
$10,760,435 |
210 |
$8,508,643 |
$13,825,367 |
192 |
226 |
Repeated |
2 |
Outbreak Intensity x0.75 |
B |
$10,923,006 |
204 |
$8,677,313 |
$13,923,560 |
185 |
220 |
Repeated |
2 |
Outbreak Intensity x0.75 |
C |
$14,813,953 |
95 |
$13,531,655 |
$16,948,370 |
84 |
108 |
Repeated |
2 |
Outbreak Intensity Decreasing |
A |
$9,661,053 |
182 |
$7,641,654 |
$12,468,109 |
166 |
197 |
Repeated |
2 |
Outbreak Intensity Decreasing |
B |
$9,812,274 |
175 |
$7,765,975 |
$12,578,619 |
160 |
191 |
Repeated |
2 |
Outbreak Intensity Decreasing |
C |
$14,762,240 |
94 |
$13,497,523 |
$16,901,232 |
82 |
107 |
Repeated |
2 |
Local Staff |
A |
$12,028,207 |
239 |
$9,462,816 |
$15,376,583 |
218 |
255 |
Repeated |
2 |
Local Staff |
B |
$12,203,936 |
232 |
$9,613,500 |
$15,465,190 |
211 |
248 |
Repeated |
2 |
Local Staff |
C |
$15,008,450 |
99 |
$13,665,701 |
$17,132,157 |
87 |
112 |
Repeated |
2 |
Reduced Adherence |
A |
$12,028,207 |
239 |
$9,462,816 |
$15,376,583 |
218 |
255 |
Repeated |
2 |
Reduced Adherence |
B |
$12,203,936 |
232 |
$9,613,500 |
$15,465,190 |
211 |
248 |
Repeated |
2 |
Reduced Adherence |
C |
$16,576,533 |
139 |
$14,881,011 |
$19,256,691 |
121 |
157 |
Repeated |
2 |
Lower LTBI Diagnosis |
A |
$14,040,450 |
284 |
$10,882,334 |
$18,186,037 |
248 |
315 |
Repeated |
2 |
Lower LTBI Diagnosis |
B |
$14,377,396 |
279 |
$11,075,105 |
$18,605,230 |
242 |
312 |
Repeated |
2 |
Lower LTBI Diagnosis |
C |
$16,978,544 |
139 |
$14,955,534 |
$20,247,870 |
110 |
174 |
Single |
1 |
Base Case |
A |
$7,493,340 |
103 |
$5,927,277 |
$9,748,954 |
90 |
118 |
Single |
1 |
Base Case |
B |
$6,996,027 |
90 |
$5,647,525 |
$8,975,360 |
79 |
103 |
Single |
1 |
Strengthened LTBI |
A |
$6,456,462 |
84 |
$5,095,052 |
$8,420,016 |
73 |
96 |
Single |
1 |
Strengthened LTBI |
B |
$5,998,226 |
70 |
$4,864,534 |
$7,606,907 |
61 |
81 |
Single |
1 |
Local Staff |
A |
$7,493,340 |
103 |
$5,927,277 |
$9,748,954 |
90 |
118 |
Single |
1 |
Local Staff |
B |
$6,995,551 |
90 |
$5,624,667 |
$8,941,536 |
79 |
103 |
Single |
1 |
Lower LTBI Diagnosis |
A |
$8,788,731 |
128 |
$6,819,381 |
$11,668,638 |
108 |
153 |
Single |
1 |
Lower LTBI Diagnosis |
B |
$8,177,016 |
116 |
$6,452,799 |
$10,725,792 |
96 |
140 |
Single |
2 |
Base Case |
A |
$5,034,527 |
83 |
$3,978,665 |
$6,536,382 |
73 |
95 |
Single |
2 |
Base Case |
B |
$5,139,231 |
79 |
$4,085,546 |
$6,660,691 |
69 |
90 |
Single |
2 |
Strengthened LTBI |
A |
$3,903,293 |
59 |
$3,093,574 |
$5,050,255 |
52 |
68 |
Single |
2 |
Strengthened LTBI |
B |
$3,866,123 |
52 |
$3,122,848 |
$4,931,604 |
46 |
60 |
Single |
2 |
Local Staff |
A |
$5,034,527 |
83 |
$3,978,665 |
$6,536,382 |
73 |
95 |
Single |
2 |
Local Staff |
B |
$5,139,231 |
79 |
$4,085,546 |
$6,660,691 |
69 |
90 |
Single |
2 |
Lower LTBI Diagnosis |
A |
$5,701,595 |
98 |
$4,496,403 |
$7,491,771 |
84 |
114 |
Single |
2 |
Lower LTBI Diagnosis |
B |
$5,864,065 |
95 |
$4,605,265 |
$7,674,281 |
80 |
111 |
Results: Univariate Sensitivity Analyses
The following two figures show how cost-effectiveness of Strategy B compared to Strategy A was affected by changes in model parameters, in both villages.
Figure S3. One-way sensitivity analysis for incremental cost per active TB case averted comparing once-off active screening to no active screening given a single outbreak, in Village 1. Red = high value of parameter; Blue = low value of parameter. Because parameters for progression and reactivation were changing over time, a “multiplier” was included in one-way sensitivity analysis (multiplier = 1 in the base case).
Figure S4. One-way sensitivity analysis for incremental cost per active TB case averted comparing once-off active screening to no active screening given a single outbreak, in Village 2. Red = high value of parameter; Blue = low value of parameter. Because parameters for progression and reactivation were changing over time, a “multiplier” was included in one-way sensitivity analysis (multiplier = 1 in the base case).
Figure S5. One-way sensitivity analysis for incremental cost per active TB case averted comparing once-off active screening to no active screening given repeated outbreaks, in Village 2. Red = high value of parameter; Blue = low value of parameter. Because parameters for progression and reactivation were changing over time, a “multiplier” was included in one-way sensitivity analysis (multiplier = 1 in the base case).
Figure S6. One-way sensitivity analysis for incremental cost per active TB case averted comparing once-off active screening to no active screening given repeated outbreaks, in Village 2. Red = high value of parameter; Blue = low value of parameter. Because parameters for progression and reactivation were changing over time, a “multiplier” was included in one-way sensitivity analysis (multiplier = 1 in the base case).
Results: Probabilistic Sensitivity Analyses
Figure S7. Probabilistic sensitivity analysis for Village 1 given a single outbreak. Each data point represents an incremental cost per active TB case averted comparing Strategy B (one round of active screening) to Strategy A (no active screening) given a single outbreak in 2019. There are 10,000 data points. The solid black lines divide the four quadrants of the cost-effectiveness plane: the upper left (Strategy B is more costly and less effective than Strategy A, i.e. B is dominated by A; points shown in orange); the upper right (Strategy B is more costly and more effective than Strategy A; points shown in red); the lower right (Strategy B is less costly and more effective than Strategy A, i.e. B dominates A; points shown in yellow); the lower left (Strategy B is less costly and less effective than Strategy A); Moving clockwise from the upper left quadrant, the proportions of simulations in each quadrant are: 5%, 8%, 86%, and 0%.
Figure S8. Probabilistic sensitivity analysis for Village 2 given a single outbreak. Each data point represents an incremental cost per active TB case averted comparing Strategy B (one round of active screening) to Strategy A (no active screening) given a single outbreak in 2019. There are 10,000 data points. The solid black lines divide the four quadrants of the cost-effectiveness plane: the upper left (Strategy B is more costly and less effective than Strategy A, i.e. B is dominated by A; points shown in orange); the upper right (Strategy B is more costly and more effective than Strategy A; points shown in red); the lower right (Strategy B is less costly and more effective than Strategy A, i.e. B dominates A; points shown in yellow); the lower left (Strategy B is less costly and less effective than Strategy A); Moving clockwise from the upper left quadrant, the proportions of simulations in each quadrant are: 26%, 36%, 38%, and 0%.
Figure S9. Probabilistic sensitivity analysis for Village 1 given repeated outbreaks. In the graphs from left to right, each data point represents an incremental cost per active TB case averted comparing a) Strategy B (one round of active screening) to Strategy A (no active screening), b) Strategy C (biennial active screening) to Strategy A, and c) Strategy C to Strategy B, all given an outbreak every three years. There are 10,000 data points in each of the three graphs. The solid black lines divide the four quadrants of the cost-effectiveness plane: the upper left (more costly and less effective; points shown in orange); the upper right (more costly and more effective; points shown in red); the lower right (less costly and more effective; points shown in yellow); and the lower left (less costly and less effective; points shown in purple). For Strategy B vs. Strategy A, moving clockwise from the upper left quadrant, the proportions of simulations in each quadrant are: 1%, 0%, 98%, and 1%. For Strategy C vs. Strategy A, moving clockwise from the upper left quadrant, the proportions of simulations in each quadrant are: 0%, 38%, 62%, and 0%. For Strategy C vs. Strategy B, moving clockwise from the upper left quadrant, the proportions of simulations in each quadrant are: 0%, 75%, 25%, and 0%.
Figure S10. Probabilistic sensitivity analysis for Village 2 given repeated outbreaks. In the graphs from left to right, each data point represents an incremental cost per active TB case averted comparing a) Strategy B (one round of active screening) to Strategy A (no active screening), b) Strategy C (biennial active screening) to Strategy A, and c) Strategy C to Strategy B, all given an outbreak every three years. There are 10,000 data points in each of the three graphs. The solid black lines divide the four quadrants of the cost-effectiveness plane: the upper left (more costly and less effective; points shown in orange); the upper right (more costly and more effective; points shown in red); the lower right (less costly and more effective; points shown in yellow); and the lower left (less costly and less effective). For Strategy B vs. Strategy A, moving clockwise from the upper left quadrant, the proportions of simulations in each quadrant are: 25%, 37%, 37%, and 0%. For Strategy C vs. Strategy A, moving clockwise from the upper left quadrant, the proportions of simulations in each quadrant are: 0%, 100%, 0%, and 0%. For Strategy C vs. Strategy B, moving clockwise from the upper left quadrant, the proportions of simulations in each quadrant are likewise: 0%, 100%, 0%, and 0%.
1. Khan FA, Fox GJ, Lee RS, Riva M, Benedetti A, Proulx JF, Jung S, Hornby K, Behr MA, Menzies D. Housing and tuberculosis in an Inuit village in northern Quebec: a case-control study. CMAJ open. 2016 Jul;4(3):E496.
2. Behr MA, Edelstein PH, Ramakrishnan L. Revisiting the timetable of tuberculosis. BMJ. 2018;362:k2738
3. N’Diaye DS, Nsengiyumva NP, Uppal A, Oxlade O, Alvarex GG, Schwartzman K. The potential impact and cost-effectiveness of tobacco reduction strategies for tuberculosis prevention in Canadian Inuit communities. BMC Medicine. 2019;17(26).
4. Gallant V, Duvvuri V, McGuire M. Tuberculosis (TB): Tuberculosis in Canada-Summary 2015. Canada Communicable Disease Report. 2017 Mar 2;43(3-4):77.
5. Inuit Statistical Profile 2018, Inuit Tapiriit Kanatami. 2018.
6. Regional Tuberculosis Data, Nunavik Regional Board of Health and Social Services. 2019.
7. Ditkowsky JB, Schwartzman K. Potential cost-effectiveness of a new infant tuberculosis vaccine in South Africa--implications for clinical trials: a decision analysis. PLoS One. 2014 Jan;9(1):e83526.
8. Tiemersma EW, van der Werf MJ, Borgdorff MW, Williams BG, Nagelkerke NJ. Natural history of tuberculosis: duration and fatality of untreated pulmonary tuberculosis in HIV negative patients: a systematic review. PloS one. 2011 Apr 4;6(4):e17601.
9. Grzybowski S, Enarson DA. The fate of cases of pulmonary tuberculosis under various treatment programmes. Bull IUAT. 1978;53(2):70-5
10. Program Data from Active Screening Initiatives. Nunavik Regional Board of Health and Social Services. 2019.
11. Tan M, Menzies D, Schwartzman K. Tuberculosis screening of travelers to higher-incidence countries: a cost-effectiveness analysis. BMC Public Health. 2008;8(1):201.
12. Campbell JR, Johnston JC, Cook VJ, Sadatsafavi M, Elwood RK, Marra F. Cost-effectiveness of latent tuberculosis infection screening before immigration to low-incidence countries. Emerging infectious diseases. 2019 Apr;25(4):661.
13. Manuel Des Medecins Specialistes, Regie de l’assurance maladie Quebec. 2020.
14. Registered nurses and registeres psychiatric nurses (NOC 3012), Emploi Quebec. 2019.
15. Liste des medicaments. Regie de l’assurance maladie Quebec. 2020.
16. Banerji A, Panzov V, Robinson J, Young M, Ng K, Mamdani M. The cost of lower respiratory tract infections hospital admissions in the Canadian Arctic. International journal of circumpolar health. 2013 Jan 31;72(1):21595.
17. Salary scales and list of job titles: Collective Agreement, FIQ Sante. 2020.
18. Alsdurf H, Oxlade O, Adjobimey M, Ahmad Khan F, Bastos M, Bedingfield N, Benedetti A, Boafo D, Buu TN, Chiang L, Cook V. Resource implications of the latent tuberculosis cascade of care: a time and motion study in five countries. BMC health services research. 2020 Dec;20:1-1.
19. Fees from Regional Hospital, Nunavik, Nunavik Regional Board of Health and Social Services. 2020.
20. Oxlade O, Sugarman J, Alvarez GG, Pai M, Schwartzman K. Xpert® MTB/RIF for the diagnosis of tuberculosis in a remote Arctic setting: Impact on cost and time to treatment initiation. PloS one. 2016;11(3):e0150119.
21. Menzies D, Adjobimey M, Ruslami R, Trajman A, Sow O, Kim H, Obeng Baah J, Marks GB, Long R, Hoeppner V, Elwood K. Four months of rifampin or nine months of isoniazid for latent tuberculosis in adults. New England Journal of Medicine. 2018 Aug 2;379(5):440-53.
22. Brassard P, Anderson KK, Schwartzman K, Macdonald ME, Menzies D. Challenges to tuberculin screening and follow-up in an urban Aboriginal sample in Montreal, Canada. J Health Care Poor Underserved. 2008 May;19(2):369-79.
23. Smith BM, Schwartzman K, Bartlett G, Menzies D. Adverse events associated with treatment of latent tuberculosis in the general population. CMAJ. 2011 Feb;183(3):E173-9.
24. Births, deaths, natural increase and marriages, administrative regions, Québec, 1986-2020 (in French only). Institut de la statistique du Québec. 2021 Mar 18. Available at: https://statistique.quebec.ca/en/produit/tableau/births-deaths-natural-increase-and-marriages-administrative-regions-quebec
25. Fox GJ, Nhung NV, Sy DN, Hoa NL, Anh LT, Anh NT, Hoa NB, Dung NH, Buu TN, Loi NT, Nhung LT. Household-contact investigation for detection of tuberculosis in Vietnam. New England Journal of Medicine. 2018 Jan 18;378(3):221-9.
26. Riva M, Fletcher C, Dufresne P, Perreault K, Muckle G, Potvin L, Bailie RS. Relocating to a new or pre-existing social housing unit: significant health improvements for Inuit adults in Nunavik and Nunavut. Canadian Journal of Public Health. 2020 Feb;111(1):21-30.
|