|Mammography screening in Norway: results from the first screening round in four counties and cost-effectiveness of a modelled nationwide screening
|Wang H, Karesen R, Hervik A, Thoresen S O
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.
Biannual screening (mammography using two views) for breast cancer was compared to no screening.
Economic study type
The study population comprised Norwegian women between the age of 50 and 69.
The setting was the community. The economic study was conducted in Norway.
Dates to which data relate
Effectiveness evidence, resource use estimates and prices were from 1996 to 1997.
Source of effectiveness data
The effectiveness data were based on a single study.
Link between effectiveness and cost data
Resource use was collected prospectively alongside a screening programme that also recorded surrogate measures for the success of the screening.
All women between 50 and 69 in four Norwegian counties were invited to participate: a total of 159,887 women, or about 40% of Norwegian women in this age range. Power calculations would not have been relevant as the whole population was considered in the analysis. The invitations were sent by letter. A total of 79.5% of the women attended the screening.
The study was an observational cohort study that recorded information from several screening units. Results from the first phase of the screening programme were recorded, and no specific duration of follow-up was reported.
Analysis of effectiveness
The analysis was based on intention to treat. Mortality reduction from breast cancer in screened populations of women was assumed based on a previous Swedish clinical study. These data were used to model survival and life years gained over a 10 year period. Life years were discounted at a rate of 4.5%. Clinical outcomes were not valued in the analysis.
Mammography was assumed to reduce mortality from breast cancer by 30%. Positive predictive value (PPV) was 16.2%, but this was assumed to be 18.0% in the economic analysis.
The authors concluded that results achieved on surrogate measures of effectiveness indicated that a mortality reduction rate of at least 30% should be obtainable in the population of women screened in the programme.
A model was used to synthesize costs and effectiveness estimates as well as to explore the sensitivity of key parameters of the screening programme.
Measure of benefits used in the economic analysis
Life years gained was the measure of benefits used in the economic analysis, but these were not valued.
Cost per screening was calculated by including the cost of invitation, screening and diagnostic work-up until a final diagnosis. The authors reported that they conducted a time-consumption study of recall examinations in one of the counties included in the screening programme. Cost of personnel was calculated by multiplying the number of procedures by time per procedure and wages of the personnel involved. The authors used accountancy data from an established breast diagnosis centre to assess the cost of equipment and material used during screening. Future costs were discounted at a rate of 4.5%. Costs and quantities were not reported separately. The study reported average costs only. The price year was 1996-1997.
Statistical analysis of costs
No statistical analysis was reported.
No indirect costs were included in the analysis of costs.
The sensitivity of the estimate of mortality reduction, cost per screen and change in positive predictive value (PPV) was explored. One-way sensitivity analysis was used to estimate cost-effectiveness when mortality reduction was reduced to 20% or increased to 40%; when cost per screen was increased by 20% and 40%, and when PPV was reduced to 13.0% and increased to 20.0%.
Estimated benefits used in the economic analysis
In a population of about 160,000 women with an attendance rate of nearly 80% the benefits of the screening programme were estimated to be 41,279 life years saved, and 1,799 lives saved.
The authors made their original cost estimates in Norwegian Kroner (Nok), but converted this to US dollars at an exchange rate of Nok 7.86 Nok per $1.00. The paper does not report costs in Nok. The average cost of one screening was reported to be $50. The cost of a recall examination was estimated at $153, and the cost of a false positive was estimated at $107. The discount rate was 4.5%. Screening was compared to no screening, and the cost of the comparator was therefore assumed to be zero. The cost of a biannual screening programme over 10 years was estimated to be $154,796,590.
Synthesis of costs and benefits
In the base case, the cost per saved life for the screening programme was $86,045 and the cost per life-year saved was $3,750 as compared to no screening. The cost per life-year saved was $5,622 assuming a 20% mortality reduction from screening, and $2,813 assuming a 40% mortality reduction. A 40% increase in cost per screening would result in a cost per life year saved of $5,250. The cost per life year saved was $2,043 when only costs were discounted. The authors commented that the cost-effectiveness ratios were most sensitive to the change in mortality reduction.
The authors concluded that the cost-effectiveness estimates in this study were low but compared well with results from other European studies. The cost-effectiveness of mammography compared favourably with other life saving initiatives such as the prevention of road accidents.
CRD COMMENTARY - Selection of comparators
The authors compared biannual screening to no screening. It may have been more relevant to have made comparison with a screening programme with different degree of intensity, as this is relevant practice in other countries.
Validity of estimate of measure of effectiveness
The authors reported that the estimate of 30% mortality reduction from breast cancer in screened women was based on the results from clinical studies on Swedish women (effectiveness measures were then applied to a Norwegian setting). Decision-makers using the results of this analysis should judge whether this risk reduction is a realistic measure of effectiveness in their own setting. The authors did, however, test their results in the sensitivity analysis, which addressed the issue of variability in their estimates.
Validity of estimate of measure of benefit
Benefits were estimated through the modelling process, which was appropriate for the approach, adopted.
Validity of estimate of costs
The cost calculations were based on one centre in the study. This was a specialised centre in a central location and the authors pointed out that the costs may therefore not have been representative of all centres in the screening programme. The authors did not report the resource use separately, and it is therefore difficult to assess both the validity of the estimates and the relevance of the costs to other settings.
The authors pointed out the potential clinical advantage of using two views (for mammography) in the Norwegian setting. The authors also discussed the potential problems of reproducing results in population-based programmes in comparison with those from clinical trials, but argued that this study had achieved comparable or better results. The importance of compliance is also identified as a pre-requisite for cost-effective screening programmes. Good comparisons with other studies were made and issues of generalisability were addressed by the nature of the study (extrapolating results from one country to another).
Implications of the study
The authors noted that the Norwegian Parliament's decision to adopt the biannual screening programme nation-wide was a well founded decision given the results of this study. Well-functioning registers are essential in population-based screening programmes.
Wang H, Karesen R, Hervik A, Thoresen S O. Mammography screening in Norway: results from the first screening round in four counties and cost-effectiveness of a modelled nationwide screening. Cancer Causes and Control 2001; 12(1): 39-45
Subject indexing assigned by NLM
Aged; Breast Neoplasms /mortality /prevention & Cost Savings; Cost-Benefit Analysis; Female; Forecasting; Health Services Research; Humans; Mammography /economics /standards; Mass Screening /economics /standards; Middle Aged; Norway /epidemiology; Program Evaluation; Value of Life; control /radiography
Date bibliographic record published
Date abstract record published