|
Cost-effectiveness of aspirin chemoprevention for Barrett's esophagus |
Hur C, Nishioka N S, Gazelle G S |
|
|
Record Status 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. Health technology Four strategies for the management of Barrett's oesophagus (BE) to prevent oesophageal adenocarcinoma were studied. The strategies were aspirin therapy, endoscopic surveillance with biopsies, both, or neither.
Study population The hypothetical population was a cohort of 50,000 55-year-old men who had BE proven by endoscopic biopsy. The reason for choosing this base-case was that BE is more common in men, and the average age at diagnosis is 55 years.
Setting The setting was secondary care. The economic study was carried out in Boston (MA), USA.
Dates to which data relate The effectiveness evidence was obtained from studies dating from 1976 to 2003. For the cost data, the studies dated from 1996 to 2001. The price year was 2000.
Source of effectiveness data The evidence was derived from a review or synthesis of completed studies and estimates based on experts' opinions.
Modelling A Markov Monte Carlo (state transition) decision model was used. The health states in the model included:
BE (no dysplasia),
low-grade dysplasia (LGD),
high-grade dysplasia (HGD),
cancer without symptoms,
cancer with symptoms,
post-successful oesophagectomy for either HGD or cancer,
ineligible for resection for either HGD or cancer,
incomplete resection for cancer, and
death.
The Markov cycle length was 1 month. To simulate clinical reality, the model allowed for each patient to have a true biologic state (for the model to be followed internally) and a perceived state (for patient management). The time horizon was from starting age (base-case 55 years) until death. The transition probabilities remained constant over time.
Outcomes assessed in the review The parameters used in the model included:
aspirin effects (reduction in cancer and complications);
BE transition probabilities per year;
procedure characteristics (oesophagectomy mortality, eligibility for resection, surgical cure, standard endoscopy, complication rate, surgery rate with complications, mortality rate in those with complications); and
endoscopic biopsy characteristics (false-negative and false-positive rates).
Study designs and other criteria for inclusion in the review No inclusion criteria for a review were reported. However, the authors included meta-analyses of clinical trials and of primary studies of varying design.
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 The effectiveness evidence was derived from 21 primary studies.
Methods of combining primary studies A narrative method was used to combine the primary studies..
Investigation of differences between primary studies Results of the review The parameters used in the model were as follows.
The parameters for aspirin characteristics were:
a 50% reduction in cancer; and
complication rates (per 100,000 individuals) of 220 noncerebral bleeds, 4.4 deaths from noncerebral bleed, 20 total haemorrhagic cerebrovascular accidents, and 6 deaths from cerebrovascular accidents.
The BE transition probabilities per year were:
0.5% BE to cancer, 6.5% BE to LGD, 34.8% LGD to BE, 16.5% LGD to HGD, 14.8% HGD to LGD, and 15.5% HGD to cancer.
The parameters for procedure characteristics were:
for oesophagectomy mortality, 2.0% for HGD, 4.0% for adenocarcinoma with surveillance, and 8.1% for adenocarcinoma without surveillance;
for eligibility for resection, 100% at age 45 and 91.9% at age 80;
for surgical cure, 80% with surveillance and 33% without surveillance; and
for standard endoscopy, a 13/100,000 complication rate, 12.3% surgery rate in those with complications, and 0.16% mortality rate in those with complications.
The parameters for endoscopic biopsy characteristics were:
a false-negative rate of 17.5% for cancer interpreted as HGD, 11.5% for HGD interpreted as LGD, and 17.5% for LGD interpreted as BE; and
a false-positive rate of 0.5% for HGD interpreted as cancer, 8.3% for LGD interpreted as HGD, and 14.5% for BE interpreted as LGD.
For most of the model parameters included, the authors adequately referenced their base case value, their range used in the sensitivity analysis, and the source.
Methods used to derive estimates of effectiveness The study was based on published data, experts' opinion, and authors' assumptions.
Estimates of effectiveness and key assumptions The authors stated that when published estimates were not available, an expert in the field was consulted to provide an expert estimate of the parameter. The false-positive rate, expressed as a percentage where HGD is interpreted as cancer for endoscopic biopsy, was assumed and the risk of cancer was reduced as the published value was considered to be overstated.
Measure of benefits used in the economic analysis The measure of benefit was the quality-adjusted life-years (QALYs). The utility weights were determined for each health state and adjusted for age and gender. The utility weights were mainly derived from the literature and expert opinion. The model assumed that the post-procedure quality of life was 70% that of the pre-procedure quality of life. There was no discussion of the methods used in the literature to derive the utility weights. The QALYs were discounted at a real annual rate of 3%.
Direct costs The medical direct costs were included in the analysis. These were for aspirin, oesophagectomy, the hospital, endoscopy, and cancer care. The cost of 325 mg enteric-coated aspirin was obtained from the Drug Topics Red Book. The cost of an oesophagectomy was derived from the Medicare reimbursement rates corresponding to the appropriate Diagnosis-Related Group (DRG) and Current Procedural Terminology (CPT) codes. DRG codes were used to derive estimates for hospital costs and CPT codes were used to derive physician costs. A similar approach was used to calculate the cost of an endoscopic perforation requiring surgery and the cost of an endoscopy with biopsies. Estimates for the cost of cancer care were derived from published literature. The costs from prior years were converted to year 2000 dollars using the medical care component of the Consumer Price Index. All the costs were discounted at a real annual rate of 3%.
The authors reported some examples of the process followed to derive the costs for the model. However, they did not report the quantities and the costs separately. The estimations of the quantities and the total costs were derived using a modelling approach.
Statistical analysis of costs No statistical analysis of the costs was reported.
Indirect Costs Productivity costs were included in the analysis. The authors mentioned that the day's wages used to calculate the productivity costs was estimated using data from the US Bureau of Labour Statistics, but the quantities were not reported. The costs from prior years were converted to year 2000 dollars using the medical care component of the Consumer Price Index. All the costs were discounted at a real annual rate of 3%.
Sensitivity analysis A sensitivity analysis was conducted to investigate areas of uncertainty related to variability in the data. The ranges selected were based on published literature. The parameters investigated included the starting age of aspirin therapy, the benefits of aspirin, the delayed beneficial effects of aspirin, and the annual rate of BE to cancer progression. Also investigated was the quality of life after an oesophagectomy, the complication rates associated with aspirin therapy, and the oesophagectomy mortality rates. Further parameters were the discount rate, cost of aspirin, cost of oesophagectomy, and cost of endoscopy.
Estimated benefits used in the economic analysis In the base-case analysis, the no therapy strategy resulted in 12.92 QALYs, 21.59 years of life expectancy, a 12.90% rate of oesophageal cancer rate, and 7.29% of cancer deaths.
The aspirin strategy resulted in 13.11 QALYs, 22.00 years of life expectancy, a 6.91% rate of oesophageal cancer, and 4.07% of cancer deaths.
The endoscopic surveillance strategy resulted in 13.13 QALYs, 22.18 years of life expectancy, a 4.74% rate of oesophageal cancer, and 1.64% of cancer deaths.
The combined aspirin-endoscopy strategy resulted in 13.19 QALYs, 22.28 years of life expectancy, a 2.74% rate of oesophageal cancer, and 1.01% of cancer deaths.
Cost results The cost was $7,100 for the no therapy strategy, $4,200 for the aspirin strategy, $31,900 for endoscopic surveillance, and $20,500 for the combined aspirin-endoscopy strategy.
Synthesis of costs and benefits The estimated benefits and costs were combined using an incremental cost-effectiveness ratio (ICER). This was calculated as the difference in cost divided by the gain in QALYs.
In the base-case analysis, aspirin dominated no therapy, while the combined aspirin-endoscopic surveillance strategy dominated endoscopic surveillance. To be dominated means to be more costly and less effective than a comparator. The ICER of the combined aspirin-endoscopic surveillance strategy compared with aspirin alone was $203,800.
Sensitivity analyses showed that the model was sensitive to an increased starting age, and to reduced or delayed aspirin benefits. The model was robust to several other one-way sensitivity analyses that were performed.
Authors' conclusions Aspirin chemoprevention might be both an effective and cost-effective strategy. When aspirin was added to either no therapy or to the currently recommended endoscopic surveillance regimen, it resulted in strategies that dominated its respective comparator strategies. The combination strategy was the preferred strategy because it resulted in the greatest number of quality-adjusted life-years (QALYs) compared with no therapy, with an incremental cost-effectiveness ratio (ICER) within the limits of acceptability ($49,600/QALY).
CRD COMMENTARY - Selection of comparators The authors justified their choice of the comparators on the grounds that recent data suggested that non-steroidal anti-inflammatory drugs, including aspirin, may prevent the progression of BE to adenocarcinoma. You should judge whether these preventive strategies are relevant in your setting, or whether other comparators from other drugs could also be relevant.
Validity of estimate of measure of effectiveness The authors did not state that a systematic review of the literature had been undertaken. Although this is a common practice with models, it does not always ensure that the best data available are used in the model. The authors used data from the available studies selectively. One cannot be sure that all the relevant literature was identified, although it is positive that meta-analyses were used to derive the effectiveness and complications of aspirin, and that few authors' assumptions were made. The estimates of effectiveness were derived credibly from the studies identified. The authors used data from published sources, experts' opinions and their own assumptions. The effectiveness evidence was derived from a meta-analysis and randomised clinical trial, which is an adequate source to estimate effectiveness. The authors justified their assumptions with reference to the medical literature. The estimates were investigated by sensitivity analyses, using ranges from the literature, but the authors did not provide any justification for the ranges selected and reported.
Validity of estimate of measure of benefit The authors used the QALYs as a measure of benefits. These were derived from a Markov model. This measure of benefit enables comparisons across different health technologies. The methods used in the literature to derive the utility weights were not reported. Sensitivity analyses over adjusted QALYs were conducted, although the method used to select the ranges was not reported.
Validity of estimate of costs The authors reported that the study had been conducted from a societal perspective. However, although the indirect costs were included, they were not reported in sufficient detail. The costs to the patient were not considered either. The unit costs were taken from published sources. No statistical analysis of the costs was undertaken. Sensitivity analyses of selected direct costs, to assess the robustness of the estimates used, were conducted and reported. Discounting was appropriately carried out since the time horizon exceeded two years. A revaluation of the costs was carried out and the price year was reported, which will aid any future reflation exercises.
Other issues The authors made appropriate comparisons of their findings with those from other studies. The authors did not explicitly address the generalisability of the results, but they did consider assessing the impact of population heterogeneity. The authors' conclusions reflected the scope of the analysis. However, the authors drew their conclusions from an inappropriate statistic. They stated that the ICER of the combined strategy compared with the no therapy strategy was within reasonable limits, but the no therapy strategy was dominated so the relevant comparator was aspirin. This ICER was $203,800, which is rather more expensive than $49,600.
The authors stated that the principal limitation of the model was related to the uncertainty of some parameter estimates. However, the possibility of conducting randomised controlled trials to show the benefits of aspirin in patients with BE is unrealistic, because the rate of progression from BE to oesophageal adenocarcinoma is low. In addition, the authors stated that the model did not specifically assess the potential impact of population heterogeneity on the results, because data on which to base the range estimates were not available for all model parameters.
Implications of the study The authors stated that aspirin used in patients with BE is effective when considered purely from the standpoint of oesophageal cancer prevention, independently of endoscopic surveillance. In a patient who has BE and concomitant cardiac risk factors, aspirin use for primary prevention may be an obvious recommendation. In patients without cardiac risk factors, aspirin use may also be a reasonable recommendation, assuming that more studies confirm that its efficacy exceeds the threshold of this study.
Bibliographic details Hur C, Nishioka N S, Gazelle G S. Cost-effectiveness of aspirin chemoprevention for Barrett's esophagus. Journal of the National Cancer Institute 2004; 96(4): 316-325 Other publications of related interest Provenzale D, Schmitt C, Wong JB. Barretts esophagus: a new look at surveillance based on emerging estimates of cancer risk. American Journal of Gastroenterology 1999;94:2043-53.
Indexing Status Subject indexing assigned by NLM MeSH Aged; Anticarcinogenic Agents /administration & Aspirin /administration & Barrett Esophagus /complications /drug therapy /pathology; Cost-Benefit Analysis; Drug Costs /statistics & Esophageal Neoplasms /economics /etiology /pathology /prevention & Esophagoscopy; Female; Humans; Male; Markov Chains; Massachusetts; Middle Aged; Monte Carlo Method; Sensitivity and Specificity; control; dosage /economics; dosage /economics; numerical data AccessionNumber 22004000335 Date bibliographic record published 31/12/2004 Date abstract record published 31/12/2004 |
|
|
|