Cost-comparison or cost-optimisation studies
Cost-comparison studies analyse the costs of different screening strategies but not the effectiveness of the different treatments.
The Swiss paper analysed the cost of introducing the IGRA in routine screening of contacts and was published in 2006 by Wrighton-Smith and Zellweger [
15]. For the analysis, probabilities of a positive TST (> = 10 mm) and a positive IGRA (T-SPOT.TB) were taken from contacts evaluated in 2004 and 2005 in accordance with the Swiss protocol for contact tracing. Only the costs of chemoprevention based on the probability of a positive test were compared while the effect on TB prevention was not taken into consideration. Three strategies were evaluated: IGRA-only, IGRA as a confirmation test for a positive TST, and TST-only. The two-step strategy was less expensive than the IGRA-only strategy, however only by a small margin of 5%. Compared to the TST-only strategy, the costs of the IGRA-only strategy were 44% lower. Compared to the strategy using both tests they were 49% lower.
Fox et al. [
16] analysed the costs of screening HCWs for tuberculosis with the IGRA versus the TST. Based on Israeli HCWs, the cost analysis comprised 100 HCWs who were referred for routine screening. TST (> = 10 mm) was positive twice as often as QFT-IT (17% vs. 34%). Assuming a 50% adherence to chemoprevention, the total cost of screening and treating these 100 HCWs was minimised to € 4,155 by using the QFT-IT in order to confirm a positive TST (reduction of 49%). However, the possibility that some HCWs would not come back to get their TSTs read was not included in the model. The probability of having an indeterminate IGRA result was surprisingly high, and the figure provided by the authors was more than 50% in comparison to the probability of having a positive QFT result (0.09 vs. 0.17). Under these circumstances the QFT-IT-only strategy cost 12% less than the TST-only strategy (€ 7,248 vs. € 8,217). Observed adherence to chemoprevention in the QFT-positive group was 47%, compared to 12% in the TST-positive group.
Diel et al. [
17] used a decision analytic model to simulate the costs of screening for LTBI in close contacts over a two-year time period. They analysed the costs of the different screening strategies: 1) QFT-only, 2) TST-only, 3) QFT to confirm a positive TST, and 4) positive TST followed by a QFT in BCG-vaccinated contacts. Based on the probabilities of a previous TST/QFT-comparison study among close contacts, the combined TST/QFT strategy was less costly than the TST-only strategy. The cost of the TST-only strategy was 48.6% higher than the QFT-only strategy (€ 91.06 vs. € 61.29 per close contact). The combined screening strategy for all contacts was the least expensive one (€ 52.02 per contact), followed by the strategy which used QFT to confirm a positive TST in contacts with BCG vaccination and for all others TST-only (€ 55.45)
Hardy et al. [
18] analysed the screening cost for LTBI in 280 immigrants moving from high-incidence countries to Great Britain based on NICE guidelines or when using the IGRAs first. With few exceptions (pregnancy, young age) all immigrants receive chest x-rays (CXRs) and those from high-incidence countries are also tested with the TST. If the TST is positive, an IGRA is performed. The alternative protocol provides for QFT-IT in all immigrants and stipulates CXRs in those with a positive QFT-IT. The number of chest x-rays needed decreased from 275 to 105 (38%) and the number of QFT-ITs needed increased from 153 to 280 (183%). The number of LTBI cases diagnosed increased from 83 using the NICE protocol to 105 (126%) using the QFT-first protocol. Total costs for the screening of the 280 immigrants were 27% lower for the QFT-first protocol and costs per detected LTBI case were reduced from £ 160.81 to £ 93.16 (-42%). Despite the title ("Cost effectiveness of the NICE guidelines...") no cost-effectiveness analysis was performed.
In a further paper, Diel et al. [
19] analysed the health and economic outcomes of isoniazid treatment of 1,000 close contacts followed hypothetically for two years with respect to isoniazid-related hepatotoxicity and early post-exposure TB over a two-year period using the QFT-IT, TST-only or QFT-IT as a test to confirm positive TST results (TST/QFT-IT). The model incorporated the results of a prior predictive value study [
8] assuming a higher progression rate to TB disease in QFT-positive than in TST-positive subjects.
Screening and treatment based on QFT-IT-only (€ 215.79 per close contact) was less costly compared to the TST/QFT-IT strategy (€ 227.89) and the TST-only strategy (€ 232.58) because the more targeted preventive therapy provoked fewer secondary hepatotoxic events. There were also fewer missed LTBI cases, due to, among other things, misread or false-negative TST results. This lead to a lower number of unprevented TB cases.
In summary, the IGRA-only screening strategy was less costly than the TST screening strategy in two cost analysis studies [
18,
19]. Both studies used the QFT-IT. Three other studies found the two-step strategy to be less costly [
15‐
17]. The Swiss study analysed the T-SPOT.TB. In all studies the TST-only strategy was the most expensive one (Table
1).
Cost-effectiveness studies
Cost-effectiveness studies are done as part of a complete economic evaluation with the aim of comparing the costs and consequences of various measures [
20]. All cost-effectiveness studies in this review used Markov modelling for the transition to different health states. This dynamic decision analytic technique allows the progression from LTBI to active TB and the treatment outcome to be modelled over time. The follow-up period after MTB infection varied from two years to life-long in the various studies and all studies discounted the costs and health effects using a rate of 3%. Most of these studies used 'quality-adjusted life years' (QALYs) and 'life years gained' (LYGs) as effects and calculated incremental cost effectiveness ratios (ICERs). One study, Pooran et al. [
21], provided cost per avoided TB cases instead. A strategy is considered dominant if it is less expensive and at the same time more effective than the alternative strategy, which then becomes the dominated strategy (Table
2).
Pooran et al. [
21] analysed the cost-effectiveness of five different screening scenarios in contact tracing over a two-year time period in the UK using: 1) TST-only, 2) the T-SPOT.TB-only, 3) positive TST followed by T-SPOT.TB, 4) QFT-IT-only, and 5) positive TST followed by QFT-IT.
Cost-effectiveness was measured as total costs per active TB case and the ICER per active TB case prevented. Cost for performing the T-SPOT were assumed to be only ₤ 55. Figures provided for sensitivity and specificity of the T-SPOT were clearly higher than those for the QFT (95% and 100% vs. 89% and 95%); sensitivity for the TST was assumed to be 85%. There was no stratification with respect to BCG vaccination, and TST specificity was considered to be 80% for all contacts. The total cost of TST screening amounted to ₤ 199,598 per 1,000 contacts compared to T-SPOT.TB at ₤ 203,983, QFT-IT at ₤ 202,921, TST/T-SPOT at ₤ 162,387 and TST/QFT-IT at ₤ 157,048. The incremental cost per active TB case prevented, compared with no screening, was ₤ 47,840 in TST, ₤ 39.712 in T-SPOT.TB, ₤ 42,051 in QFT. The most cost-effective strategy was the two-step strategy with TST and T-SPOT (₤ 37,206), followed very closely by the TST/QFT strategy (₤ 37,699).
To assess the cost-effectiveness of QFT-G vs. the TST in diagnosing contact persons with active TB cases in Canada, Marra et al. [
22] used a decision analytic Markov model. Three different screening strategies were evaluated over a 20-year-period: TST-only, QFT-only and the two-step strategy using the QFT-G to confirm a positive TST. The model was stratified by ethnicity (foreign-born, non-aboriginal Canadian-born and aboriginal), and BCG vaccination status, as the noted groups have different rates of prior infection and BCG use.
The most cost-effective strategy was to administer QFT-G in BCG-vaccinated contacts and reserve TST for all other patients, assuming specificity for the TST of more than 99% in all BCG-unvaccinated subgroups, but of only 96% for the QFT. Driven primarily by the extremely high specificity value, which was not varied in sensitivity analysis, and in combination with an assumed low BCG vaccination rate, the TST alone-strategy is dominant (ICER) and an incremental net monetary benefit (INMB) of CA$ 3.70 per contact investigation was calculated. The least cost-effective strategy was the use of the QFT-G for all cases, which resulted in an INMB of CA$ -11.15.
Oxlade et al. [
23] used Markov modelling to compare expected TB cases and costs over 20 years following screening for TB with different strategies among hypothetical cohorts of foreign-born immigrants and close contacts in Canada. Canada. The authors compared five different strategies for immigrants: 1) no screening, 2) CXR, 3) TST-only, 4) QFT-only, and 5) QFT for confirmation of a positive TST. For screening of contact persons they compared three strategies: 1) no screening, 2) TST only and 3) QFT.
The least costly strategy for immigrants coming from intermediate and high incidence countries versus non screening was CXR screening. There, patients were considered positive for LTBI if they had an "abnormal" radiograph and subsequently a positive TST. The sensitivity of that method for identifying those infected was assumed to be 11%. The approach had an incremental cost of CA$ 825 and $ 30,680 per prevented case when applied to entry screening of immigrants from high-incidence and intermediate incidence countries, respectively. In contrast, initial screening with QFT was the most expensive one when immigrants had no BCG vaccination or had been BCG vaccinated in infancy.
With respect to contact screening, assuming a TST specificity of 98% in BCG-unvaccinated cases, of 92% among infants and 60% among older BCG-vaccinated, screening of close contacts and casual contacts (when coming from low-incidence countries) with QFT or TST would result in savings compared to non-screening. In such circumstances, TST would generally be more cost-effective than the QFT with the exception of screening older close and casual contacts who had received BCG vaccination. However, Kowada et al. [
24] evaluated the cost-effectiveness of the QFT-IT for TB screening in close contacts in Japan over the lifetime of a contact (age 20) in a nearly completely BCG vaccinated society. They compared the QFT-only strategy with the TST followed by the QFT strategy and TST-only strategy. The target population was a hypothetical cohort of 1,000 immunocompetent 20-year-old close contacts to a sputum smear positive index case. Based on a very low baseline specificity of the TST among BCG vaccines of 15% and a high prevalence of LTBI, the QFT-only strategy was dominant (US$ 471.54/28.1099 QALYs) compared to the TST/QFT strategy (US$ 500.55/28.1087 QALYs) and the TST-alone strategy (US$ 573.98/28.1079 QALYs). The incremental cost-effectiveness ratio of the QFT was a cost saving of US$23.043/QALYs. Accordingly, the QFT-only strategy is the most cost-effective for contact investigation in a medium-incidence country like Japan.
In addition, de Perio et al. [
25] used a Markov state transition decision model to compare cost and quality-adjusted life years (QALYs) with three strategies for a hypothetical 35-year-old HCW cohort with and without BCG vaccination, also over a lifetime horizon. Costs and QALYs were discounted at 3% per year. They compared two versions of QFT and TST-only and accounted for inadequate and indeterminate outcomes of both QFTs, for failure to return for TST reading, and for 2-step TST testing. In this study, sensitivity of the QFT was assumed to be clearly higher than that of the TST (76% vs. 67%) and the drop-out rate for reading of the first TST was high at 12%. Both IGRAs were more effective and less costly than the TST. The TST strategy was thus dominated. The ICER of the QFT-G compared with the QFT-IT was US$14.092/QALY for BCG-unvaccinated and US$103.047/QALY for BCG-vaccinated HCWs. In conclusion, the authors stated that the use of IGRAs leads to superior clinical outcomes and lower costs than screening with the TST does.
In contrast to these findings, four other cost-effectiveness studies showed that the two-step strategy with TST and IGRA was the most cost-effective strategy compared to the IGRA-only strategy. The first one is a cost-effectiveness study among adult close contacts in France. Deuffic-Burban [
26] generated a decision analytic model. Lifetime costs and life expectancies for no testing, TST (basically positive at a cut off of ≥ 10 mm) and QFT-IT only and TST/QFT-IT strategies were calculated and compared using incremental cost-effectiveness ratios (ICERs) in euros per life year gained (LYG). The authors provided a sensitivity figure of only 76% for the QFT (vs. 73% for the TST) and a very high cost to cost relation between the QFT (€ 40.50) and the TST (€ 2.16). The proportion for adherence to LTBI treatment as a basic value for their calculations was assumed to be only 57%, thus reducing the cost for preventative INH treatment initially started due to false positive TST results. Given these assumptions, the discounted direct medical lifetime costs of care per patient were € 417 for no testing, € 476 for TST, € 443 for QFT and € 435 for TST/QFT. The TST/QFT strategy was associated with an ICER of € 560/LYG compared to no testing, while the QFT-only strategy was associated with an ICER of € 730/LYG. The TST-only strategy, irrespective of whether a cut off of 5 or 10 mm was used, was strongly dominated (higher costs and lower life expectancy).
Another study from Diel et al. [
27] assessed the cost-effectiveness of the QFT assay for screening and treatment of close contacts in Germany. They analysed the health and economic outcomes of isoniazid treatment for 20-year-old contacts over a 20-year time period using two different cut-off values for the TST (≥ 5 and 10 mm), the QFT-only and then QFT as a confirmatory test for a positive TST. QFT-based treatment led to cost savings of US$ 542.9 and 3.8 life days per LTBI case compared to non-treatment, TST-based treatment at 10 mm induration size saved US$ 177.4 and saved 2.0 life days per test-positive contact. Choosing a 5 mm cut-off for the TST resulted in additional expenditures and saved only 0.9 days. Although the ICER for treatment based on a TST < 5 mm was below the commonly used willingness-to-pay threshold (US$ 30,170/LYG) it resulted in unnecessary treatment of 77% due to false-positive TST results. Combining TST at a 5 mm cut-off followed by the QFT in a dual-step screening approach was only marginally less expensive (0.6%) than using the QFT solely.
In a further study, Diel et al. [
28] analysed the outcomes of INH treatment of close contacts in Switzerland using the Markov model over a 20-year period following screening with the TST-only (three different cut-off values: 5, 10 and 15 mm) and the T-SPOT.TB-only or as a two-step strategy with the TST. T-SPOT.TB-based treatment was cost-effective in both age groups at € 11,621 (20-yr-old cohort) and € 23,692 per LYG (40-yr-old cohort). Only in the younger group, and with a TST cut-off of > 15 mm, was the ICER of € 26,451/LY below the willingness-to-pay threshold of US$ 50,000 (or € 40,195; average exchange rate for 2004: US$ = € 0.8039); all other TST-only options were not cost-effective. Combination of the TST with T-SPOT.TB slightly reduced the total cost compared with the T-SPOT.TB alone, by 4.4% and 5.0% in the younger and older groups respectively.
The ultimate aim of LTBI screening is the prevention of progression to active TB via chemopreventative therapy. Whether the introduction of IGRA in the TB screening of contacts is cost-effective (in terms of producing expenditures below a predefined WTP threshold per LYG) in this respect was analysed in eight studies from six different countries (US, Canada, Japan, Switzerland, France, UK and Germany). With the exception of Japan, a country with medium TB prevalence, these countries are considered to have low TB prevalence. All studies performed TB screening on groups at high risk for developing tuberculosis: HCWs [
25], close contacts [
15,
19,
21,
22,
24,
26‐
28], and immigrants from high-incidence countries [
23].