|
Cost-effectiveness of screening for colorectal cancer: a simulation model |
Neilson A R, Whynes D K |
|
|
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 Screening for colorectal cancer using faecal occult blood (FOB) test and/or flexible sigmoidoscopy.
Study population A cohort of hypothetical subjects between 40 and 90 years of age.
Setting Hospital. The economic study was carried out in the United Kingdom.
Dates to which data relate The data related to the effectiveness analysis were extracted from clinical trials conducted between 1988 and 1994. The dates of resource and price data were not specified.
Source of effectiveness data The estimates for the final outcomes were derived from a semi-Markovian model. Clinical and transition probabilities were obtained from several sources including one UK randomised trial (Nottingham), European trials, and literature value averages.
Modelling A semi-Markovian model was constructed by the authors from a USA accounting matrix utilized to assess colorectal screening within Medicare. The model was used to simulate the process of screening, diagnosis, and treatment for the study cohort utilizing the clinical and transition probabilities obtained from other sources in order to estimate the net gain in expected life-years and the cost implications of each strategy for the study cohort.
Outcomes assessed in the review Since the semi-Markovian model consisted of three quasi-independent blocks the data requirement for the model was large and the values of a large number of clinical and transition probabilities were obtained from several sources including the ongoing Nottingham randomised trial, European trials, and literature value averages. Some of the clinical and transition probabilities for which values were reported were as follows: the probability of a true progressive adenoma being detected by FOB, the percentage of all cases of cancers originating in polyps (adenomas), the compliance rate with diagnostic follow-up of a positive result, the probability of a true progressive polyp being detected by colonoscopy,the number of years (on average) taken for a progressive polyp to develop into early-stage cancer and thereafter to become symptomatic, the compliance rate with screening, and the overall probability of detection of a cancer in its early stage.
Study designs and other criteria for inclusion in the review The main sources of clinical and transition probabilities were one ongoing UK randomised controlled trial (Nottingham) and three European randomised controlled trials. Adequate information was not given with regard to the literature review as the third source of data for the study.
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 Four randomized controlled trials from Denmark, England, France and Sweden were the main sources of the data for the study. The number of primary studies included in the literature review (the third source of the effectiveness data) was not specified.
Methods of combining primary studies The authors used the average values of the parameters whose values obtained from the literature review. No other method was mentioned.
Investigation of differences between primary studies Results of the review The values of the clinical and transition probabilities obtained from the clinical trials or the literature review were not fully reported in the paper. The adopted values of the parameters whose values were chosen to be reported were as follows:
probability of a true progressive adenoma being detected by FOB:20%,
percentage of all cases of cancers originating in polyps (adenomas): 83%,
compliance rate with diagnostic follow-up of a positive result: 95%,
probability of a true progressive polyp being detected by colonoscopy: 84%,
number of years, on average, taken for a progressive polyp to develop into early-stage cancer: 15 years,
number of years to become symptomatic after that: 2 years,
compliance rate with screening: 53%,
overall probability of detection of a cancer in its early stage: 59%.
Measure of benefits used in the economic analysis Quality adjusted life years (QALYs) were used as the outcome measure in the economic analysis. A semi-Markovian model was used to estimate cancer incidence and stage at detection, as the intermediate outcomes, with QALYs, as the final outcome. Adequate information was not given about the methods employed to evaluate the health states. QALYs were not discounted.
Direct costs Costs were discounted. Quantities and cost items were not reported separately. Total net (over and above the no screening option) costs of screening for different age ranges and different screening intervals were reported in relative, rather than absolute, values. A comprehensive list of cost components was not given, however, some of the cost components were: cost of surgical treatment of cancer, by stage, cost of colonoscopic polypectomy, and cost of treating operative complications. The perspective of the cost calculations was not reported. The main source of the cost data was the accumulated findings of the Nottingham trial. The dates to which the price data referred were not specified since the economic outcomes were reported in relative, rather than absolute, values.
Currency No currency was reported: the economic outcomes produced by the model were presented as relative, rather than absolute, values.
Estimated benefits used in the economic analysis Not presented in the paper.
Cost results The value adopted for the discount rate was not reported. The total net discounted costs were reported using the cost of screening of 40-74 years olds on an annual basis as the base case (index 100). Indices of total net discounted costs for screening at 1,2,3,4 and 5 year intervals were:
40-70 age group: 100,53, 38, 31 and 26,
45-74 age group: 92, 49, 35, 28 and 24,
50-74 age group: 81, 44, 32, 27 and 22,
55-74 age group: 69, 36, 27, 22 and 18,
60-74 age group: 57, 32, 23, 19 and 15,
65-74 age group: 46, 25, 21, 17 and 13.
The duration of the intervention and comparator costs were not specified.
Synthesis of costs and benefits Taking screening of 40-74 years olds on an annual basis as the base case (index 100), indices of net costs per undiscountedQALY gained for screening at 1,2,3,4 and 5 year intervals would be:
40-74 age group: 100, 81, 79, 81 and 83,
45-74 age group: 101, 82, 81, 82 and 83,
50-74 age group: 103, 85, 85, 87 and 87,
55-74 age group: 107, 89, 90, 90 and 91,
60-74 age group: 123, 104, 105, 106 and 105,
65-74 age group: 157, 135, 150, 147 and 135.
Authors' conclusions Optimal screening intervals, defined in terms of minimum costs per QALY, appear to exist, although these are different for different age ranges. It appears that three year screening intervals, commencing at the age of 40 years, represent the global optimum modality. However, if it was decided to delay screening until later ages, a two-year interval for the cohorts with starting ages of around 55 years and above would appear the most appropriate.
CRD COMMENTARY - Selection of comparators The reason for the choice of comparator is clear.
Validity of estimate of measure of benefit Although the authors utilizedrandomised controlled trials as the main sources for the model data, it is not possible to ascertain whether the estimates of measure of benefit used in the economic analysis are likely to be internally valid since the authors did not to provide adequate information about the effectiveness data, and presented only a limited description of their methodology and the assumptions used in their model.
Validity of estimate of costs Again, the authors have not presented a complete description of their methodology in this particular paper and therefore it is not possible to ascertain whether important cost items were omitted or to comment on the methods of quantity and cost estimation.
Other issues This paper only presented a brief overview of the complex methods used in a semi-Markovian model and contained few details of the data used to generate the cost-effectiveness indices of screening for colorectal cancer.
Source of funding Supported by a grant from PPP Medical Trust.
Bibliographic details Neilson A R, Whynes D K. Cost-effectiveness of screening for colorectal cancer: a simulation model. IMA Journal of Mathematics Applied in Medicine and Biology 1995; 12(3-4): 355-367 Indexing Status Subject indexing assigned by NLM MeSH Adult; Aged; Aged, 80 and over; Cohort Studies; Colorectal Neoplasms /economics /epidemiology /prevention & Computer Simulation; Cost-Benefit Analysis; Female; Great Britain /epidemiology; Humans; Male; Mass Screening /economics; Mathematics; Middle Aged; Models, Economic; Randomized Controlled Trials as Topic /statistics & control; numerical data AccessionNumber 21997006380 Date bibliographic record published 31/12/1998 Date abstract record published 31/12/1998 |
|
|
|