Chris Sampson’s journal round-up for 23rd July 2018

Every Monday our authors provide a round-up of some of the most recently published peer reviewed articles from the field. We don’t cover everything, or even what’s most important – just a few papers that have interested the author. Visit our Resources page for links to more journals or follow the HealthEconBot. If you’d like to write one of our weekly journal round-ups, get in touch.

Quantifying life: understanding the history of quality-adjusted life-years (QALYs). Social Science & Medicine [PubMed] Published 3rd July 2018

We’ve had some fun talking about the history of the QALY here on this blog. The story of how the QALY came to be important in health policy has been obscured. This paper seeks to address that. The research adopts a method called ‘multiple streams analysis’ (MSA) in order to explain how QALYs caught on. The MSA framework identifies three streams – policy, politics, and problems – and considers the ‘policy entrepreneurs’ involved. For this study, archival material was collected from the National Archives, Department of Health files, and the University of York. The researchers also conducted 44 semi-structured interviews with academics and civil servants.

The problem stream highlights shocks to the UK economy in the late 1960s, coupled with growth in health care costs due to innovations and changing expectations. Cost-effectiveness began to be studied and, increasingly, policymaking was meant to be research-based and accountable. By the 80s, the likes of Williams and Maynard were drawing attention to apparent inequities and inefficiencies in the health service. The policy stream gets going in the 40s and 50s when health researchers started measuring quality of life. By the early 60s, the idea of standardising these measures to try and rank health states was on the table. Through the late 60s and early 70s, government economists proliferated and proved themselves useful in health policy. The meeting of Rachel Rosser and Alan Williams in the mid-70s led to the creation of QALYs as we know them, combining quantity and quality of life on a 0-1 scale. Having acknowledged inefficiencies and inequities in the health service, UK politicians and medics were open to new ideas, but remained unconvinced by the QALY. Yet it was a willingness to consider the need for rationing that put the wheels in motion for NICE, and the politics stream – like the problem and policy stream – characterises favourable conditions for the use of the QALY.

The MSA framework also considers ‘policy entrepreneurs’ who broker the transition from idea to implementation. The authors focus on the role of Alan Williams and of the Economic Advisers’ Office. Williams was key in translating economic ideas into forms that policymakers could understand. Meanwhile, the Economic Advisers’ Office encouraged government economists to engage with academics at HESG and later the QoL Measurement Group (which led to the creation of EuroQol).

The main takeaway from the paper is that good ideas only prevail in the right conditions and with the right people. It’s important to maintain multi-disciplinary and multi-stakeholder networks. In the case of the QALY, the two-way movement of economists between government and academia was crucial.

I don’t completely understand or appreciate the MSA framework, but this paper is an enjoyable read. My only reservation is with the way the authors describe the QALY as being a dominant aspect of health policy in the UK. I don’t think that’s right. It’s dominant within a niche of a niche of a niche – that is, health technology assessment for new pharmaceuticals. An alternative view is that the QALY has in fact languished in a quiet corner of British policymaking, and been completely excluded in some other countries.

Accuracy of patient recall for self‐reported doctor visits: is shorter recall better? Health Economics [PubMed] Published 2nd July 2018

In designing observational studies, such as clinical trials, I have always recommended that self-reported resource use be collected no less frequently than every 3 months. This is partly based on something I once read somewhere that I can’t remember, but partly also on some logic that the accuracy of people’s recall decays over time. This paper has come to tell me how wrong I’ve been.

The authors start by highlighting that recall can be subject to omission, whereby respondents forget relevant information, or commission, whereby respondents include events that did not occur. A key manifestation of the latter is ‘telescoping’, whereby events are included from outside the recall period. We might expect commission to be more likely in short recalls and omission to be more common for long recalls. But there’s very little research on this regarding health service use.

This study uses data from a large trial in diabetes care in Australia, in which 5,305 participants were randomised to receive either 2-week, 3-month, or 12-month recall for how many times they had seen a doctor. Then, the trial data were matched with Medicare data to identify the true levels of resource use.

Over 92% of 12-month recall participants made an error, 76% of the 3-month recall, and 46% of the 2-week recall. The patterns of errors were different. There was very little under-reporting in the 2-week recall sample, with 3-month giving the most over-reporting and 12-month giving the most under-reporting. 12-month recall was associated with the largest number of days reported in error. However, when the authors account for the longer period being considered, and estimate a relative error, the impact of misreporting is smallest for the 12-month recall and greatest for the 2-week recall. This translates into a smaller overall bias for the longest recall period. The authors also find that older, less educated, unemployed, and low‐income patients exhibit higher measurement errors.

Health surveys and comparative studies that estimate resource use over a long period of time should use 12-month recall unless they can find a reason to do otherwise. The authors provide some examples from economic evaluations to demonstrate how selecting shorter recall periods could result in recommending the wrong decisions. It’s worth trying to understand the reasons why people can more accurately recall service use over 12 months. That way, data collection methods could be designed to optimise recall accuracy.

Who should receive treatment? An empirical enquiry into the relationship between societal views and preferences concerning healthcare priority setting. PLoS One [PubMed] Published 27th June 2018

Part of the reason the QALY faces opposition is that it has been used in a way that might not reflect societal preferences for resource allocation. In particular, the idea that ‘a QALY is a QALY is a QALY’ may conflict with notions of desert, severity, or process. We’re starting to see more evidence for groups of people holding different views, which makes it difficult to come up with decision rules to maximise welfare. This study considers some of the perspectives that people adopt, which have been identified in previous research – ‘equal right to healthcare’, ‘limits to healthcare’, and ‘effective and efficient healthcare’ – and looks at how they are distributed in the Netherlands. Using four willingness to trade-off (WTT) exercises, the authors explore the relationship between these views and people’s preferences about resource allocation. Trade-offs are between quality vs quantity of life, health maximisation vs equality, children vs the elderly, and lifestyle-related risk vs adversity. The authors sought to test several hypotheses: i) that ‘equal right’ respondents have a lower WTT; ii) ‘limits to healthcare’ people express a preference for health gains, health maximisation, and treating people with adversity; and iii) ‘effective and efficient’ people support health maximisation, treating children, and treating people with adversity.

A representative online sample of adults in the Netherlands (n=261) was recruited. The first part of the questionnaire collected socio-demographic information. The second part asked questions necessary to allocate people to one of the three perspectives using Likert scales based on a previous study. The third part of the questionnaire consisted of the four reimbursement scenarios. Participants were asked to identify the point (in terms of the relevant quantities) at which they would be indifferent between two options.

The distribution of the viewpoints was 65% ‘equal right’, 23% ‘limits to healthcare’, and 7% ‘effective and efficient’. 6% couldn’t be matched to one of the three viewpoints. In each scenario, people had the option to opt out of trading. 24% of respondents were non-traders for all scenarios and, of these, 78% were of the ‘equal right’ viewpoint. Unfortunately, a lot of people opted out of at least one of the trades, and for a wide variety of reasons. Decisionmakers can’t opt out, so I’m not sure how useful this is.

The authors describe many associations between individual characteristics, viewpoints, and WTT results. But the tested hypotheses were broadly supported. While the findings showed that different groups were more or less willing to trade, the points of indifference for traders within the groups did not vary. So while you can’t please everyone in health care priority setting, this study shows how policies might be designed to satisfy the preferences of people with different perspectives.


The role of healthcare in promoting health equity

Why should we consider health equity? The rationale for treating health as a special good rests on the idea that a certain level of health is a precondition for achieving any of the other outcomes in life that we value. Sen identifies that health contributes to a person’s ability to choose the life she has reason to value. Our functioning is impaired by disability and particularly by death; Andrew Marvel writes in a 1681 poem, ‘The grave’s a fine and private place,/ But none, I think, do there embrace.’ As such we value health equity, but what then is the role of healthcare in promoting health equity?

In the UK we take pride in the idea of a health service that provides healthcare only on the basis of need. It is important then that there is equitable access to healthcare; everyone should have an equal capability of benefiting from that health care. It would be a poor state of affairs if socioeconomic variables, such as income, determined a person’s access to healthcare rather than need. But socioeconomic factors do determine need. As an example (there are many examples to have, consider much of the work by Sir Michael Marmot), women from areas of high deprivation are far more likely to have premature, sick babies (see here) but given the characteristics that determine need (such as gestational age, congenital anomalies etc.) the provision of care and rate of mortality is the same (see here). So deprivation goes some way to determining the characteristics that affect mortality but not the access to care or clinical outcomes conditional on need. We might therefore judge healthcare to have played its role in health equity; the social issue of health inequalities is not a concern for healthcare since there is nothing it can do beyond treating patients fairly.

It is quite possible to argue that we should favour poor people since this would reduce overall health inequalities in the whole population. Alan Williams’s idea of a ‘fair innings’ would seem to support this. The fair innings argument says that everybody has a right to a certain quality adjusted life expectancy and we should favour interventions to support this. But, as Williams points out, there is a gender difference in health outcomes; women live longer than men. As Williams says ‘We males are not getting a fair innings!’ But we certainly would object to a system where we systematically favoured men over women; health equity cannot be judged in isolation of ideas of fairness.

One of the most pervasive arguments against cost-effectiveness analysis (CEA), and utilitarianism in general, is that it is distribution blind. Williams’s argument was an attempt to find a solution to that. But in terms of equity in relation to socioeconomic factors, the poorest are normally the sickest (whichever way the arrow of causality may lie). The current system does not favour the sickest; a gain of 0.2 QALYs is treated the same for a person at 0.1 or 0.7 QALYs. Derek Parfit suggests that one way of reducing inequality would be to reduce the standard of those at the top. However, while this may increase equality we would not say that this contributes to equity in any meaningful way, and would soon reject this as a solution. This would imply that we would rather help those at the bottom of the distribution as a means of reducing inequality. So why should we not favour the sickest? It not only appeals to our sense of justice but would reduce socioeconomic health gaps too.

But could the argument above be extended to systematically favouring the poor in healthcare settings? No, because this would be unfair, not only at the individual level but also for healthcare practitioners. It would place an unfair moral burden on a doctor to treat those on the right side of an income threshold.

So I would say then that the role of healthcare should be to provide equitable access to healthcare resources, but that this should entail providing resources that favour treatment for the sickest since this would be a policy that would favour the poorest without placing any unfair restraints on practitioners or patients.

*This post takes a lot from this book, and in particular Amartya Sen’s contribution to it which is also available here.