Elsevier

Social Science & Medicine

Volume 47, Issue 2, 16 July 1998, Pages 213-222
Social Science & Medicine

Valuing temporary and chronic health states associated with breast screening

https://doi.org/10.1016/S0277-9536(98)00065-3Get rights and content

Abstract

The aim of this study was to derive quality of life values for the four key breast screening outcomes (true negative, false positive, true positive and false negative), including the quality of life effects of the screening and treatment processes. In doing so, methodological issues in health status measurement were explored, in particular the valuation of temporary health states. The true negative and false positive descriptions were temporary health states, lasting for short term durations (12 months) and the true positive and false negative outcomes were chronic health states lasting for long term durations (rest of life). Descriptions of breast screening outcomes were valued using the time trade-off technique and the visual analogue scale. Paired comparisons between TTO values for states with the same duration found a difference between the true negative and the false positive time trade-off values but no difference for true positive and false negative descriptions. The TTO values for the false positive were low. The study highlights several important methodological issues such as the use of the two stage procedure for valuing temporary health states, the impact of duration on values, the impact of anchor points, and the importance of qualitative analysis of respondents values. Further empirical testing of all these issues is recommended.

Introduction

The benefits of breast screening programmes are often defined solely in terms of gains from detecting breast cancer earlier. Yet other benefits may be derived from breast screening, such as reassurance for those women in whom no evidence of breast cancer is found. At the same time, attendance at breast screening may generate higher than usual levels of anxiety, even if this is short lived, and it may cause false reassurance in women whose breast cancer is missed or unnecessary anxiety if results are false positive. The actual process of receiving the invitation, attending the breast screening centre, having the mammogram and waiting for results may also affect quality of life. Recent studies of breast screening, both quality of life and non quality of life studies, have confirmed the importance of breast screening outcomes beyond those involving treatment for breast cancer. It has been suggested that a true negative outcome may give women reassurance (Gerard et al., 1992) but there is evidence of short lived periods of raised anxiety and physical discomfort or pain associated with a breast screening attendance (Fallowfield et al., 1990; Cockburn et al., 1994). Women who receive false positive results have been found to be extremely anxious after receiving a letter recalling them for further tests (Sutton et al., 1995) and to experience altered moods (Gram et al., 1990).

Determining the impact of breast screening on quality of life is complex for several reasons. Firstly, the breadth of outcomes involved can include possible reassurance, anxiety, or life threatening disease if treatment for breast cancer is required. Secondly, the range of durations of outcomes, with non breast cancer treatment outcomes for temporary periods of time followed by a return to previous health levels, and breast cancer treatment outcomes for durations covering the rest of the womans life. Thus, screening involves both temporary and chronic health states. Thirdly, individuals are asymptomatic at their screening attendance, rather than presenting with disease. The multiple outcomes of screening have been recognised for other screening programmes, such as antenatal screening, where researchers have recently argued for improved measures of benefit to capture the true effects of screening (Mooney and Lange, 1996; Shackley and Cairns, 1996).

The ideal outcome measure would address all these features. Within the context of outcome measurement for economic evaluation, however, previous studies of breast screening programmes have not addressed the issue of different durations and have placed emphasis on health states associated with breast cancer, deriving values for health states associated with treatment for breast cancer only (Buxton et al., 1987; Hall et al., 1992). de Haes et al. (1991) derived values for 15 health states associated with breast screening, 13 were associated with treatment for breast cancer, one with screening attendance and one with the diagnostic phase of screening. It is clear that breast screening has an impact on quality of life both in terms of treatment for breast cancer but also in terms of the screening process itself.

This paper has two related aims. Firstly, to generate quality of life values to be incorporated, at a later time, into the findings of two clinical trials investigating policy options for breast screening, in order to estimate quality adjusted life years. Secondly, to explore the methodological issues surrounding the valuation of temporary and chronic health states associated with breast screening. Issues particular to the valuation of temporary health states have not been discussed in any comprehensive manner in the literature.

Section snippets

Temporary and chronic health states

Temporary health states are states lasting for a specified period of time until a return to good health whilst chronic health states are states lasting for the rest of life until death (Torrance, 1986). The duration of temporary health states are short term and may be specified in terms of weeks, months or years, the only requirement being that the duration is less than life expectancy. The appropriate duration selected may depend on the health care intervention being evaluated. The duration of

Description

Screening programmes can be classified into four outcomes: true negative, false positive, true positive and false negative. These four outcomes were used as the basis for descriptions and are shown in Table 1. These four breast screening outcomes clearly involve combinations of events (invitation, screening, follow up and treatment) but this approach has the advantage that the value of the screen is not separated from the invitation or the result of the screen.

A literature review was conducted

Sample and characteristics

The response rate from the letter of invitation was 43.5% and the number of interviews conducted was 440. Individual and household characteristics of respondents are shown in Table 3. The characteristics of respondents are compared to U.K. census data for the same age groups of women in the general population (OPCS, 1991). The characteristics of respondents suggest that a higher proportion of women in the sample were married, although the proportions were similar in terms of ethnic origin. The

Summary

This study has explored the values associated with the four key outcomes of breast screening and the extent to which they differ. Comparisons were limited to paired comparisons between descriptions of the same duration. The paired ranking results showed that, at an ordinal level (as indicated by the ranking), differences were observed between the true negative and false positive descriptions and between the true positive and false negative descriptions. These differences were also observed at a

Acknowledgements

We thank the staff at the four participating breast screening centres; the interviewers and the women for being interviewed. We also thank Dawn Winpenny, Karen Arnold, Nicky Gillard and Jo Holland for secretarial support; Gavin Mooney for comments on a earlier draft; Martin Buxton and members of the advisory group; and two anonymous referees for their helpful comments.

References (30)

  • J. Cook et al.

    A cost utility analysis of treatment options for gallstone disease: Methodological issues and results

    Health Economics

    (1994)
  • J.C.J.M. de Haes et al.

    The impact of a breast cancer screening programme on quality-adjusted life-years

    International Journal of Cancer

    (1991)
  • P. Dolan et al.

    The time trade off method: Results from a general population study

    Health Economics

    (1996)
  • L.J. Fallowfield et al.

    Psychological outcomes of different treatment policies in women with early breast cancer outside a clinical trial

    British Medical Journal

    (1990)
  • D.H. Feeny et al.

    Incorporating utility-based quality-of-life assessment measures in clinical trials

    Medical Care

    (1989)
  • Cited by (40)

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

      2019, Social Science and Medicine
      Citation Excerpt :

      The literature is similarly heterogeneous in the duration and methods used to apply HSUVs. A difficulty in valuing screening interventions is that the process encompasses both temporary and chronic health states (Johnston et al., 1998). The intensity and duration of the utilities associated with screening (Gerard et al., 1999) and diagnostic anxiety (Gram et al., 1990; Lindfors et al., 2001) vary significantly to the long-term sequelae associated with treatment (Parker et al., 2007), depending on whether this is classified by intervention or disease stage.

    • Challenges in Valuing Temporary Health States for Economic Evaluation: A Review of Empirical Applications of the Chained Time Trade-Off Method

      2018, Value in Health
      Citation Excerpt :

      The most common anchor state in the identified studies, “hospitalization after a serious accident,” was justified by Jansen et al. [12] on the grounds that although this is not likely to be something that has been experienced by the respondents, it is an imaginable scenario. An alternative approach is for respondents to rank all the health states before the valuation task, and the health state considered to be the worst can be used as the anchor health state, as done in several studies [11,14,19]. A potential issue with this approach is that within a single study, more than one anchor health state could be valued in stage 2 (across respondents), which may raise concerns regarding the comparability of utilities across respondents.

    View all citing articles on Scopus
    View full text