|
Use of a stochastic simulation model to identify an efficient protocol for ovarian cancer screening |
Urban N, Drescher C, Etzioni R, Colby C |
|
|
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 Various single modality or multi-modal screening strategies for ovarian cancer involving transvaginal sonography (TVS) and the tumour antigen CA 125.
Economic study type Cost-effectiveness analysis.
Study population The hypothetical population to be screened was defined as a cohort of one million women aged 50 in 1990.
Setting Tertiary care. The economic study was carried out in the USA.
Dates to which data relate Screening was simulated over a 30-year period in a hypothetical cohort of one million women aged 50 at the beginning of the period. Costs were measured in 1990 US dollars.
Source of effectiveness data The evidence for final outcomes was based on opinion and various assumptions.
Modelling A stochastic simulation model was used to evaluate the relative costs and effectiveness of the alternative screening protocols in order to eliminate from consideration clearly inferior strategies. The model simulated the natural progression of disease, then superimposed a screening programme and evaluated how the screening strategy used alters longevity and costs.
Methods used to derive estimates of effectiveness Author's assumptions and the opinions of clinicians were used to derive estimates of effectiveness.
Estimates of effectiveness and key assumptions The key assumptions used to produce the estimates of the effectiveness of the various screening modalities were:
TVS sensitivity following delay, 1.0;
probability of false-positive TVS at first screen, 0.019;
probability of false-positive TVS at second screen, 0.010;
probability of false-positive TVS at third and subsequent screens, 0.006.
It was assumed that 5% of cases do not emit CA 125 and that CA 125 specificity in women with false positive TVS was 0.85. Of the single-modal ovarian screening strategies, strategy IB (Annual CA 125 (Elevated)) resulted in 6,347 years of life saved (YLS); strategy IC (Annual CA 125 (Elevated or rising)) resulted in 11,229 YLS; and strategy IA (Annual TVS) resulted in 15,168 YLS. Of the multi-modal ovarian cancer screening strategies, strategy IIC (two-year TVS conditional on rising or elevated CA 125) resulted in 6,222 YLS; strategy IIA (annual TVS conditional on rising or elevated CA 125) resulted in 11,556 YLS; strategy IA (annual TVS in all women) resulted in 15,168 YLS; and strategy IIB (6-month TVS conditional on rising or elevated CA 125) resulted in 17,596 YLS.
Measure of benefits used in the economic analysis Years of life saved (YLS) was used as the outcome measure in the economic analysis.
Direct costs Within-stage expectations of treatment costs attributable to ovarian cancer over the 15 years post-diagnosis were estimated from available data. Savings in treatment costs were calculated for each case with a shift from distant to local or regional stage, or from regional to local stage, as the difference in the net present value of the costs over the patient's remaining lifetime, with and without the screen. For each woman, screening and diagnosis costs were calculated by multiplying the number of procedures by their unit costs, which were obtained by means of a telephone survey of relevant laboratories and medical facilities. A sample of physician offices, hospital clinics and laboratories was contacted to gather charges for CA 125 blood tests and for TVS procedures. In the case of true positives, only half of the cost of laparoscopy was counted as a screening cost, in recognition of the clinical need to follow laparoscopy immediately by laparotomy in confirmed malignancy. Cost/quantities were not reported separately.
Statistical analysis of costs The authors treated the costs in a stochastic way.
Sensitivity analysis Sensitivity analyses were conducted in order to investigate the effects on results of alternative assumptions that:
(1) the average duration of Stage 1 disease is either 5 months or 13 months;
(2)TVS cost is reduced by one-third;
(3)the within- and between-woman variances in CA 125 are 49 and 36 rather than 5 and 80 respectively;
(4)TVS is 100% sensitive immediately in all women or 85% sensitive following the previously assumed delay;
(5)10% rather than 5% of ovarian cancers fail to emit CA 125;
(6)survival post-diagnosis depends on the rate of disease progression pre-diagnosis;
(7)when lead-time exceeds the gain in survival attributable to screening, survival gains are negative;
(8)health effects and costs are not discounted;
and (9) TVS-favouring conditions prevail.
Estimated benefits used in the economic analysis Of the single-modal ovarian screening strategies, strategy IB (Annual CA 125 (Elevated)) resulted in 6,347 YLS; strategy IC (Annual CA 125 (Elevated or rising)) resulted in 11,229 YLS; and strategy IA (Annual TVS) resulted in 15,168 YLS. Of the multi-modal ovarian cancer screening strategies, strategy IIC (two-year TVS conditional on rising or elevated CA 125) resulted in 6,222 YLS; strategy IIA (annual TVS conditional on rising or elevated CA 125) resulted in 11,556 YLS; strategy IA (annual TVS in all women) resulted in 15,168 YLS; and strategy IIB (6-month TVS conditional on rising or elevated CA 125) resulted in 17,596 YLS.
Cost results Total net costs for the three single-modal ovarian screening strategies were $719 million, $722 million and $2,217 million. Total net costs for the four multi-modal strategies were $320 million, $585 million, $2,217 million and $1,127 million.
Synthesis of costs and benefits The average cost per YLS for the three single-modal ovarian screening strategies were $113,300, $64,300 and $146,200 respectively. The average cost per YLS for the four multi-modal strategies were $51,500, $50,700, $146,200 and $64,000 respectively. Shorter average duration for the early stage of disease (5 rather than 9 months) reduces the effectiveness and cost-effectiveness of all of the three most effective strategies. A higher within-woman variance in CA 125 adversely affects the performance of the multi-modal strategies, but they still perform better than TVS used alone. A combination of four TVS-favouring conditions improves the performance of annual TVS to $80,000 per YLS, making it efficient relative to the multi-modal strategy performed annually.
Authors' conclusions The results suggest that annual screening by TVS, conditional on elevated or rising CA 125, is more efficient than annual screening by TVS in an unselected group of postmenopausal women. Using this multi-modal strategy every six months increases the chance of early detection and, therefore, the possibility of observing a significant mortality differential. The multi-modal strategy is relatively cost-effective under a wide-range of assumptions.
CRD COMMENTARY - Selection of comparators The reason for the choice of the comparators is clear and a justification is given: the authors suggest that several annual screening protocols have been proposed including transvaginal sonography (TVS), CA 125 with a criterion for positivity of elevation to 35U/ml, CA 125 followed by TVS if elevated to 30U/ml, and CA 125 (elevated to 35U/ml) in combination with TVS and pelvic examination.
Validity of estimate of measure of benefit The estimate of the measure of benefit used in the economic analysis is likely to be valid. Extensive sensitivity analyses were presented.
Validity of estimate of costs Resource quantities were not reported separately from prices but adequate details of methods of cost estimation were given.
Other issues The authors' conclusions were justified given the uncertainties in the data.
Source of funding Supported by Contract N01CN-05230 from the Division of Cancer Control of the NCI, National Institutes of Health.
Bibliographic details Urban N, Drescher C, Etzioni R, Colby C. Use of a stochastic simulation model to identify an efficient protocol for ovarian cancer screening. Controlled Clinical Trials 1997; 18: 251-270 Indexing Status Subject indexing assigned by NLM MeSH Aged; Aged, 80 and over; CA-125 Antigen /blood; Cohort Studies; Cost-Benefit Analysis /statistics & Endosonography /economics /statistics & Female; Humans; Mass Screening /economics /statistics & Middle Aged; Models, Statistical; Ovarian Neoplasms /economics /mortality /prevention & Predictive Value of Tests; Stochastic Processes; Survival Analysis; control; numerical data; numerical data; numerical data AccessionNumber 21997008215 Date bibliographic record published 28/02/1999 Date abstract record published 28/02/1999 |
|
|
|