The NHS can’t afford to fund every new medical treatment. Budgets are limited. Resources are finite. It’s a boring fact of life, from which people like me – economists – benefit. In recent years the UK has championed the provision of cost-effective healthcare. The efficiency of the NHS owes no small thanks to NICE, which evaluates health technologies for all kinds of conditions. To do so they need to be able to compare these technologies with each other. The general consensus amongst health economists, and the approach adopted by NICE, is to use quality-adjusted life years (QALYs). QALYs are required by the NICE reference case and we have excellent tools such as the EQ-5D to capture these. The use of QALYs as an outcome is almost ubiquitous in the evaluation of health technologies. So much so that they have come to define cost-effectiveness. The area in which their use and study appears most limited is in mental health research. Herein lies a problem: from the decision-makers perspective, not knowing whether an intervention is cost-effective isn’t all that different from knowing that it isn’t.
Generic preference-based measures
Health technologies are competing for the same pot of NHS money and are therefore pitted against each other, regardless of dissimilarities in the conditions they treat. We therefore need an outcome measure that is relevant to all conditions; from baldness to bunions; abdominal pain to Zellweger syndrome. We need an outcome measure that is ‘generic’. For economists (and to some extent decision-makers), preferences are paramount. A hypothetical change in an individual’s health only matters if they (or, in practice, the public) actually assign any value to this health change. As such, the last few decades have witnessed the rise of generic preference-based measures. The EQ-5D is the most well-know of these, but others do exist. These measures enable researchers to calculate the benefits of an intervention in terms of quality and length of life – combined in to one number. Decision-makers are then presented with an illuminating ‘cost-per-QALY’ of an intervention. Such a minimalist result is of great value in funding decisions. Unfortunately, in many cases, economic evaluations in mental health are not armed with this figure. This is no doubt detrimental to future provision and research.
Condition-specific preference-based measures
Mental health researchers’ apparent unawareness of generic preference-based measures is justifiable. The EQ-5D, for example, only includes a single question relating to mental health. There are also greater methodological problems with using generic preference-based measures in mental health; are public values representative of patients’ preferences; can severe patients understand the questions; are ‘general’ questions even relevant? There are certainly pros and cons to using a measure like the EQ-5D in mental health research1.
Fortunately there’s a happy medium that still allows for the calculation of QALYs and, therefore, the generic valuation of mental health technologies. Condition-specific preference-based measures. These measures capture changes in an individual’s quality of life based on dimensions relevant to specific conditions. The development of a preference-based measure involves two stages: development of the classification system (questionnaire) and the elicitation of values. Unfortunately most existing measures have only completed the first stage. Nonetheless, measures do exist and I implore you to research them further, get involved in their development and include them in your studies. There are measures under development that are specific to particular mental health problems, such as DEMQOL: a quality of life measure for individuals with dementia2. Some of the most promising work relates to the development of preference-based measures that are specific to mental health but general across disorders. This work includes development of a preference-based measure from the CORE-OM3,4.
Unfortunately these measures are almost solely employed and researched by economists. Researchers involved in the evaluation of interventions for mental health need to champion these measures, as economists alone cannot. If you’re a researcher, why not try to include fledgling preference-based measures (both general and condition-specific) in your studies, and aid their development. Cost-effectiveness is often an ‘unknown’ in mental health. This is no longer acceptable. If mental health research and care is to obtain the funding it needs, then researchers will have to bend to accommodate these methods and engage with economists. If you do nothing else please read this5, then read this6, and do not forget to read this7. The long-term benefits could be huge.
University of Nottingham
1 Brazier, J., 2010. Is the EQ-5D fit for purpose in mental health? The British Journal of Psychiatry, 197(5), pp.348-9.
2 Mulhern, B., Smith, S.C., Rowen, D., Brazier, J.E., Knapp, M., Lamping, D.L., Loftus, V., Young, T.A., Howard, R.J. and Banerjee, S. (2010) Improving the measurement of QALYs in dementia: Developing patient- and carer-reported health state classification systems using Rasch analysis. Discussion Paper. (Unpublished)
3 Mavranezouli, I., Brazier, J.E., Young, T.A. and Barkham, M. (2011) Using Rasch analysis to form plausible health states amenable to valuation: the development of CORE-6D from CORE-OM in order to elicit preferences for common mental health problems. Quality of Life Research, 20 (3). pp. 321-333. ISSN 1573-2649
4 Mavranezouli, I, Brazier, JE, Rowen, D and Barkham, M (2011) Estimating a preference-based index from the Clinical Outcomes in Routine Evaluation – Outcome Measure (CORE-OM): valuation of CORE-6D. Discussion Paper. (Unpublished)
5 Brazier, J., 2008. Measuring and valuing mental health for use in economic evaluation. Journal of health services research & policy, 13 Suppl 3, pp.70-5.
6 Jacobs, R., 2009. Investigating Patient Outcome Measures in Mental Health. CHE Research Paper 48, Centre For Health Economics: York
7 Chisholm, D., Healey, A. & Knapp, M., 1997. QALYs and mental health care. Social psychiatry and psychiatric epidemiology, 32(2), pp.68-75.