Societal costs and effects of implementing population-based mammography screening in Greenland

Abstract

With a low breast cancer incidence and low population density, Greenland is geographically and organisationally challenged in implementing a cost effective breast cancer screening programme where a large proportion of the Greenlandic women will have to travel far to attend. The aim of this paper is to evaluate the cost effectiveness and cost utility of different strategies for implementing population-based breast cancer screening in Greenland. Two strategies were evaluated: Centralised screening in the capital Nuuk and decentralised screening in the five municipal regions of Greenland. A cost effectiveness and cost utility analysis were performed from a societal perspective to estimate the costs per years of life saved and per QALY gained. Two accommodation models for the women’s attendance were examined; accommodation in ordinary hotels or in patient hotels. The least costly accommodation model was the hotel model compared with the patient hotel model, regardless of screening strategy. The decentralised strategy was more cost effective compared with the centralised strategy, resulting in 0.5 million DKK per years of life saved (YLS) and 4.1 million DKK per quality-adjusted life year (QALY) gained within the hotel model. These ratios are significantly higher compared with findings from other countries. The sensitivity analysis showed a substantial gap between the most and least favourable model assumptions. The investigated strategies were all estimated to be extremely costly, mostly due to high transportation and accommodation costs and loss of productivity, and none would be accepted as cost-effective per YLS/QALY gained within a conventional threshold level. The least expensive strategy was regional screening with hotel accommodation.

Valuing the health states associated with breast cancer screening programmes: A systematic review of economic measures

“…For screening disutility, almost half of the economic evaluations (29,48,51,54,56,58,60,63,65,66,68,73,74,78,84) used expert VAS utilities derived from a second study in the Netherlands (32), but only three economic evaluations (51,54,73) considered the generalisability of the expert sample to the general population in the model to which this was applied. Other economic evaluations made their own adjustments to local population EQ-5D or SF-6D data (47,49,50,59,69,79,81) applied to reflect this uncertainty in more than half of the economic evaluations, which may bias results (QALYs) toward more frequent screening (29,49,52,54,59). This limitation was justified in five studies due to the lack of robust HSUVs for mammography screening.…”

“…This limitation was justified in five studies due to the lack of robust HSUVs for mammography screening. For the 13 studies (47,50,60,63,65,66,(78)(79)(80)(81)(82)(83)(84) which did attempt to value overdiagnosis in their analysis, an assumption was made that this was captured in the QALYs across screening strategies by including the temporary disutility of diagnosis and treatment without a corresponding gain in life years. However, the utilities applied used sources which had not highlighted that there was a risk the treatment was unnecessary during the valuation process and therefore is unlikely to fully capture the impact of the risk of overdiagnosis on quality of life.…”