|Cost-effectiveness of azithromycin for preventing Mycobacterium avium complex infection in HIV-positive patients in the era of highly active antiretroviral therapy
|Sendi P P, Craig B A, Meier G, Pfluger D, Gafni A, Opravil M, Battegay M, Bucher H C
This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn.
Mycobacterium avium complex (MAC) infection prophylaxis with azithromycin (1,200 mg per week) in HIV-infected patients receiving care in a system with access unrestricted by financial barriers and in the era of highly active antiretroviral therapy (HAART).
Type of intervention
Primary prevention of MAC infection.
Economic study type
A hypothetical population of HIV-positive patients living in a health care system unrestricted by financial barriers.
Hospitals. The economic study was carried out in Switzerland.
Dates to which data relate
The effectiveness data were collected between 1988 to 1998 (the exact dates were not reported). Most of the resource use data were collected in the period 1993-1995. The prices used refer to the same period. All costs were inflated to 1997 prices using the Swiss consumer price index for health care.
Source of effectiveness data
The main evidence for final outcomes was derived from the Swiss HIV Cohort Study (SHCS) and some evidence was derived from a randomised controlled trial by Oldfield at al (1998).
The authors developed a Markov (or state transition) model to estimate the expected survival and average costs over a 10 year period using matrix multiplication, as well as to capture the HAART effect.
Outcomes assessed in the review
The study used data from the SHCS on the clinical condition of HIV-positive patients under treatment in Swiss hospitals and data from a randomised controlled trial on the toxicity and efficacy of azithromycin.
Study designs and other criteria for inclusion in the review
The design of the SHCS was a non-randomised trial with historical controls. It was a multi-centre study and took place in 4 university hospitals in Switzerland and included 9,000 patients. Two other studies were used but it was not stated how these were selected. The evidence from the first study was used by the authors since it provided a wide range of data on a large sample of HIV-positive patients in Switzerland.
Sources searched to identify primary studies
Criteria used to ensure the validity of primary studies
Methods used to judge relevance and validity, and for extracting data
Number of primary studies included
At least 3 studies were included.
Methods of combining primary studies
The primary studies had separate inputs to the model.
Investigation of differences between primary studies
The difference between primary studies was not investigated.
Results of the review
Based on the data from the SHCS the authors calculated the monthly probabilities of transition among CD4-cell count strata in and before the era of HAART, the monthly probabilities of MAC infection and other AIDS defining diseases in and before the era of HAART, and survival probability over a period of 10 years for patients with initial CD4 count less than 50 cells/mm^3.
Before the era of HAART, the HIV-positive patients with initial CD4-cell stratum of 0-49 cells/m^3 had a 98% chance of remaining in the same stratum, a 1.22% chance of moving to a stratum with 50-74 CD4 cells/m^3 and a 0.59% chance of moving to an even higher CD4 cell stratum.
In the era of HAART, the HIV-positive patients with initial CD4-cell stratum of 0-49 cells/m^3 had a 87.85% chance of remaining in the same stratum, a 7.35% chance of moving to a stratum with 50-74 CD4 cells/m^3, and a 4.80% chance of moving to an even higher CD4 cell stratum.
Before the era of HAART, patients with 50-74 CD4 cell count had a 17.66% chance of moving to a stratum with 0-49 CD4 cell count, a 75.17% of remaining in the same CD4 cell count stratum, and a 7.17% of obtaining a CD4 cell count higher than 79 per mm^3.
In the era of HAART, patients with 50-74 CD4 cell count had a 12.26% chance of moving to a stratum with 0-49 CD4 cell count, a 66.07% of remaining in the same CD4 cell count stratum, and a 21.67% of obtaining a CD4 cell count higher than 79 per mm^3.
Before the era of HAART, the patients with an initial CD4-cell count over 75 per mm^3 had a 1.77% chance of obtaining a CD4 cell count of 0-49 per mm^3, a 9.33% chance of obtaining 50-75 CD4 cells per mm^3 and an 88.3% chance of remaining with higher than 75 CD4 cells per mm^3.
In the era of HAART, patients with an initial CD4-cell count over 75 per mm^3 had a 0.96% chance of obtaining a CD4 cell count of 0-49 per mm^3, a 2.49% chance of obtaining 50-75 CD4 cells per mm^3 and an 96.55% chance of remaining with higher than 75 CD4 cells per mm^3.
In the era of HAART the monthly probabilities of obtaining MAC infection and other AIDS-defined diseases by the HIV-positive patients decreased compared to the era before HAART was introduced. The survival probabilities of the HIV-positive patients were presented in a graph. The survival probabilities were highest when MAC prophylaxis took place and its effect was assumed to be continuous. The survival probability under the 5 years and 3 years effect scenarios were lower. No MAC prophylaxis led to the poorest survival probability. The survival probabilities with no AIDS initially were higher than the survival probabilities with AIDS under all scenarios.
Measure of benefits used in the economic analysis
The economic benefit was measured in terms of years (or months) of life saved with MAC prophylaxis over a period of 10 years for HIV positive patients with initial CD4 count less than 50 cells/mm^3. To calculate the years of life saved a Markov model was used to extrapolate the outcome results. The authors considered different scenarios for durability of the highly active antiretroviral therapy:
a continuous time effect scenario (CTES);
a 5 year effect scenario (5 YES); and
a 3 year effect scenario (3 YES).
The years of life saved were estimated for HIV-positive patients with no AIDS initially and for those with AIDS initially. Benefits were discounted at 4% annually.
Costs were evaluated from the point of view of the health care system. Therefore, the costs measured in the study reflected the use of health care resources. In the case of HIV-positive patients with no AIDS, the costs were based on the assumption that the patients had regular check-up visits, routine laboratory examinations, Pneumocystis carinii prophylaxis and antiretroviral drugs. In the case of HIV-positive patients with AIDS, the costing was based on the information extracted from a random sample of patient charts from four Swiss university hospitals and adding the assumed costs for protease inhibitors. Quantities and costs were not reported separately in this study, but were reported separately elsewhere (see Meier, 1992). Most of the costs were valued in the period 1993-1995. Costs were inflated to reflect 1997 prices and were discounted at 4% annually.
Statistical analysis of costs
Costs were treated stochastically. Costs were estimated under 3 different assumptions for the durability of the HAART: a continuous time effect scenario (CTES); a 5 year effect scenario (5 YES), and a 3 year effect scenario (3 YES).
Indirect costs were not measured.
Bayesian probabilistic analysis was used to estimate how the results vary given the uncertainty in the model parameters, which depended on the model's assumptions, prior opinion and the data. The joint distribution of the model parameters was used to approximate the distributions of the expected years of life saved and the average costs. The authors estimated the results for a sample of patients with a different initial CD4 cell count (50-74 cells/mm^3 instead of 0-49 cells/mm^3). Sensitivity analysis on the discount rate was performed with costs being discounted at an annual rate of 8% and undiscounted.
Estimated benefits used in the economic analysis
Under the CTES scenario the MAC infection prophylaxis resulted in 3.17 months of life saved, (95% CI: 2.63 - 6.13) for HIV-positive patients without AIDS, and 3.55 months of life saved, (95% CI: 2.31 - 5.58) for patients with AIDS.
Under the 5-YES scenario, the MAC infection prophylaxis lead to 4.81 months of life saved for patients without AIDS, (95% CI: 2.35 - 5.48) and 2.58 months of life saved, (95% CI: 1.64 - 3.90) for patients with AIDS. Under the 3-YES scenario, the MAC infection prophylaxis resulted in 3.68 months saved, (95% CI: 2.25 - 5.36) for patients without AIDS and 1.91 months, (95% CI: 1.17 - 2.92) for patients with AIDS.
Under the CTES scenario, the treatment with MAC infection prophylaxis of no AIDS patients cost SFr230,885 while the treatment with no MAC infection prophylaxis cost SFr210,155. The 95% CI of the incremental costs was SFr17,000 - SFr31,000. The treatment with MAC infection prophylaxis of AIDS patients cost SFr522,319, while the treatment with no MAC infection prophylaxis cost SFr487,554. The 95% CI of the incremental costs was between SFr25,000 and SFr54,000.
Under the 5-YES scenario, the treatment with MAC infection prophylaxis of no AIDS patients cost SFr205,629 while the treatment with no MAC infection prophylaxis cost SFr189,064. The 95% CI of the incremental costs was between SFr12,000 and SFr26,000. The treatment with MAC infection prophylaxis of AIDS patients cost SFr421,149, while the treatment with no MAC infection prophylaxis cost SFr395,756. The 95% CI of the incremental costs was between SFr17,000 and SFr49,000.
Under the 3-YES scenario, the treatment with MAC infection prophylaxis of no AIDS patients cost SFr165,242 while the treatment with no MAC infection prophylaxis cost SFr179,589. The 95% CI of the incremental costs was between SFr9,000 and SFr24,000. The treatment with MAC infection prophylaxis of AIDS patients cost SFr335,490, while the treatment with no MAC infection prophylaxis cost SFr316,766. The 95% CI incremental costs were between SFr79,000 and SFr165,000.
Costs were estimated over a period of 10 years.
Synthesis of costs and benefits
Benefits and costs were combined using incremental cost-effectiveness ratios (ICER) which denoted the incremental costs divided by the incremental life expectancy (years of life saved) rounded to the nearest 1,000 Swiss francs.
For HIV-positive patients with no AIDS, the ICER was estimated at SFr60,000 under the CTES scenario, (95% CI: 48,000 - 88,000), SFr52,000 under the 5-YES scenario (95% CI: 37,000 - 81,000), and SFr47,000 under the 3-YES (95% CI: 29,000 - 76,000).
For patients with AIDS, the ICER was estimated at SFr118,000 under all three scenarios. The 95% CI was estimated to be between SFr91,000 - SFr151,000 under the STES scenario, SFr86,000 - SFr157,000 under the 5-YES scenario and SFr79,000 - SFr165,000 under the 3-YES scenario.
The above results were obtained for patients with 0-49 cells/mm^3 initial level of CD4 cells. When the same results were estimated for patients with 50-74 cells/mm^3 initial level of CD4 cells, the intervention was less cost-effective. All costs were inflated to 1997 prices. When alternative estimates of costs were performed, namely discounted at 8% annually and undiscounted costs, no significant change in the ICER was observed.
In the era of HAART, the MAC infection prophylaxis with azithromycin for people who are HIV-positive was more cost-effective compared to no MAC prophylaxis. The MAC prophylaxis was more effective and cost-effective on patients without AIDS initially than on patients with AIDS.
CRD COMMENTARY - Selection of comparators
The authors did not provide an explicit justification for their choice of comparator, namely a therapy for HIV-positive patients with no MAC infection prophylaxis with azithromycin. However, from the context of the paper, it can be inferred that this choice was made in order to isolate and evaluate the cost-effectiveness of treatment of HIV-positive patients which includes the MAC infection therapy with azithromycin. You, as a user of the database, should decide if this is a widely used health technology in your own setting.
Validity of estimate of measure of benefit
The study combined and analysed evidence from previously published effectiveness studies. The authors did not state that a systematic review of the literature had been undertaken nor were any details provided about the methods of selection of the studies reported. The authors used the data from the available studies in order to derive the benefits, using a Markov model, which appears to have been appropriate.
Validity of estimate of costs
All categories of costs relevant to the perspective adopted were included in the analysis. Costs and quantities were reported separately in another paper, which was referenced in the study. Prices (unit costs) were taken from published sources.
A Bayesian approach was adopted by the authors to account for the uncertainties surrounding the model parameters. The authors made appropriate comparisons of their findings with those from other studies. The authors stated that the study would be generalisable to European settings. The authors do not appear to have presented their results selectively.
Implications of the study
The study confirmed in the Swiss setting conclusions about the cost-effectiveness of MAC infection prophylaxis that were previously reached in the USA.
Source of funding
Supported in part by Abbott Switzerland, Pharmacia & Upjohn Switzerland, Pfizer Switzerland and the Swiss HIV Cohort Study. Dr P Sendi was supported by an unlimited research fellow grant from Merck Sharp & Dohme-Chibret, Switzerland.
Sendi P P, Craig B A, Meier G, Pfluger D, Gafni A, Opravil M, Battegay M, Bucher H C. Cost-effectiveness of azithromycin for preventing Mycobacterium avium complex infection in HIV-positive patients in the era of highly active antiretroviral therapy. Journal of Antimicrobial Chemotherapy 1999; 44(6): 811-817
Other publications of related interest
Ledergerber B, von Overbeck J, Egger M, Luthy R. The Swiss HIV cohort study: rationale organization and selected baseline characteristics. Sozial- und Praventivmedizin 1994;39:387-394.
Sendi P P, Craig B A, Pfluger D, Gafni A, Bucher H C. Systematic validation of disease models for pharmacoeconomic evaluations. Journal of Evaluation in Clinical Practice 1999;5:283-295.
Meier G. Kosteneffektivitat der Mycobakterium avium komplex prophylaxe bei HIV infektion in der Schweiz. Thesis, University of Basel, Switzerland.
Subject indexing assigned by NLM
AIDS-Related Opportunistic Infections /prevention & Anti-Bacterial Agents /economics /therapeutic use; Anti-HIV Agents /therapeutic use; Antibiotic Prophylaxis /economics; Azithromycin /economics /therapeutic use; Bayes Theorem; Cohort Studies; Cost-Benefit Analysis; HIV Infections /complications /drug therapy; Health Care Costs; Humans; Mycobacterium avium-intracellulare Infection /prevention & control; control
Date bibliographic record published
Date abstract record published