Themes: The value of healthcare safety, in relation to risk attitudes and extra-consequentialist factors.
Description: The NHS is committed to safe patient care, but it is often unclear how best to value safety, or to determine what price a health service should be willing to pay for it. Work to date suggests that safety is mainly valued using approaches which estimate risk of injury as an expected mean decrease in Quality Adjusted Life Years (QALY). However, behaviour and policy are likely to be influenced more by the perceived value of avoiding accidents or errors. Literature about people’s perceptions of risk and safety reveals a broad, qualitative and complex conception. Few studies have investigated people’s value judgements about reducing healthcare incidents, or their willingness to pay for devices or systems that increase safety from this "contextualist" point of view. Bringing probabilities and consequences into the same equation as other contextual parameters assumes that risk and safety can be characterised by some combination of attributes.
Building on earlier work to value information, define the cost-effectiveness gap and determine optional pricing strategies, we will study the use of full Bayesian models to support "supply side" investment decisions, encompassed by a Real Options framework in which prospective evaluation of devices early in the development process allows for the impact on current value of later decisions. Results from evaluations of devices are considered more ephemeral than those of pharmaceuticals. We will examine these ideas, using the method of panoramic meta-analysis. Health economic analysis is built around probabilities and utilities, with more emphasis placed on ensuring the validity of the former over the latter. Building on our development of split choice theory, we will conduct studies into the sensitivity of utilities to the method by which they are obtained.
Objectives: Understanding how the NHS values safety and developing new methodologies, which incorporate extra-consequentialist attributes and reflect risk attitudes:
Strategy: To develop a validated methodology for the economic evaluation of safety, which takes account of extra-consequentialist factors and risk attitudes. We will combine a literature review to gain an overview of current approaches to valuing safety, with interviews to inform the definition of attributes. A questionnaire will be completed by health care professionals involved with budgeting at various levels, to test the hypothesis that different stakeholder groups place different values on safety attributes.
Contingent evaluation will be applied to two exemplars (Sharps Injuries and MRSA infections) in order to gain insight of the traits in different contexts and among different stakeholders. We will reveal how this approach may be used and how the results compare with the standard consequentialist approach, using one or more exemplar devices as case studies. The results of the contingent evaluation will be incorporated into standard economic evaluation frameworks. A "user tool" will be developed to measure the value of safety improvements from an extra-consequentialist, non-risk neutral perspective, as part of a broader, clear-cut economic evaluation package for SMEs. We will produce a simple work book describing, in accessible ways, the advantages and disadvantages of different methods of evaluating effectiveness and safety.
To develop methodologies for supply side economics and to complement our mathematical modelling of optimum pricing strategy under uncertainty, we will build on the headroom method to produce an elaborated model for conducting health economic analyses of technologies at the design stage. We will show how a value of information analysis can be adapted to yield a value of investment analysis. A model will be created for the expected gain in benefit under the proposed new technology. The result will be an estimate of expected profit per procedure, which can inform a calculation of total payback on investment. Investors frequently have a different attitude to losses and gains of equal magnitude. We will investigate attitudes to gain loss and factor the results into the above model. Because investment decisions, unlike purchasing decisions, are not dichotomous as they include the option to develop further and then re-evaluate, we will factor a real option approach into our models. The work on utilities will continue, especially our split choice theory. In particular we will investigate how different methods of eliciting utilities affect results obtained.