Morris L. Barer - An Undisciplined Economist: Robert G. Evans on Health Economics, Health Care Policy, and Population Health

An Undisciplined Economist

Robert G. Evans on Health Economics, Health Care Policy, and Population Health

Edited by
MORRIS L. BARER,
GREG L. STODDART,
KIMBERLYN M. M C GRAIL,
AND CHRIS B. M C LEOD

Carleton Library Series 237

McGill-Queens University Press
Montreal & Kingston London Chicago

2
Toward a Healthier Economics : Reflections on Ken Bassetts Problem ( 1998 )

ROBERT G. EVANS

From 1978 to 1990 Dr Ken Bassett practised as a family physician in Invermere, a small town in southeastern British Columbia. As part of that practice, he and his colleagues cared for a number of obstetrical patients. And in managing the delivery process, they would commonly use an electronic foetal monitor ( EFM ) to check on the condition of the foetus. In this their behaviour was no different from that of most other physicians, in North America at least. But they may have been somewhat unusual in reflecting upon and being troubled by their own behaviour. As they were well aware, there is an extensive research literature on the effectiveness of electronic foetal monitoring that stretches back about twenty years. Over time, the early enthusiasm for this diagnostic tool has given way to the now predominant view that, on balance, EFM provides no benefits for normal deliveries. Yet it continues to be used routinely. And it continued to be used routinely in Invermere, by family practitioners who knew that their practice had no scientific basis. Unlike most practitioners (and most other people) they found this discrepancy troubling and tried to understand it.

Dr Bassett went further, however, and in 1985 began to study for a doctorate in medical anthropology during winter sessions at McGill University. Not surprisingly, intensive further analysis of EFM from an alternative disciplinary perspective has shown that the application of a particular medical intervention is the result of a wide and quite complex array of interrelated factors (Bassett, 1996). The balance of findings from randomized trials and other forms of scientific evaluation of effectiveness is only one of these factors, and is rarely decisive in itself.

Of particular interest was the role of EFM in supporting first an increase and then a decrease in rates of use of augmentation (acceleration) of labour through the administration of oxytocin to the mother. This is a highly controversial procedure with some risk to both mother and child. The objective data provided by the EFM , available for interpretation by anyone, supported a corresponding sense of a normal birth process that could be defined objectively and represented by mathematical models.

Departures from the model, usually taking the form of prolonged labour, were then abnormal, and the intervention was indicated. The EFM could then be applied to ensure that the foetus was not getting into trouble, becoming hypoxic, during the procedure. The technology thus enabled clinicians to act with greater confidence in augmenting more patients.

But when, for a variety of reasons, including the sheer unpleasantness of the procedure and its dubious value, augmentation fell from favour, the same objective EFM data (which were, in any case, usually rather ambiguous) could be read to show that the foetus was not in trouble during an extended birth, and that no augmentation was necessary. Rates fell. Only years later did ... local doctors come to see augmentations as almost irresistible opportunities to behave badly, to be impatient, to project their fears onto patients, or to take out their general job frustrations on a particular individual (Bassett, 1996, 291).

Dr Bassett now works with the BC Office of Health Technology Assessment at the University of British Columbia, where a range of disciplinary perspectives are brought to bear to try to understand why particular technologies are adopted, and how to influence the process. Improving these decisions requires far more than simply presenting the findings of research, however relevant or competent.

The processes by which particular technologies come to be employed (or not) in medical practice are of obvious interest to economists, simply because of the impact these decisions have on patterns and levels of resource utilization costs and outcomes. But that is not the focus of this paper. Much more fundamental, as the Invermere physicians well understood, is the issue raised by the observation of informed practitioners knowingly (and in this case unhappily) applying an intervention that they knew to be inappropriate. I believe that this phenomenon generalizes, and not just among physicians; and I suggest the following label.

KEN BASSETTS PROBLEM:

Why would intelligent and competent professionals routinely behave in ways that they know to be illogical and scientifically unsound?

My focus will of course be on economics and economists, particularly on economists studying the health care sector. This paper thus complements that by Uwe Reinhardt (this volume), in which he documents quite specifically such behaviour within our own profession. Like him, I feel best equipped to talk about the inadequacies of the field I know best.

Ken Bassetts Problem, however, refers to a more general process, even if we ourselves have a narrower focus. And while in economics (as in medicine) this behaviour is revealed in the work of particular individuals, we are not or should not be engaged simply in finger-pointing.

Also at the outset we should set to one side, or at least hold in check, a hypothesis that may come most easily to the minds of economists. Ken Bassett and his colleagues were not paid more when There is usually much more going on than simple financial gain (Giacomini et al., 1996). And indeed (and quite inconsistently) we are much less ready to adopt economic explanations for our own behaviour.

In addressing Ken Bassetts Problem in the context of health economics, I am not proposing to argue for the existence of this form of professional behaviour in our field. Such behaviour is described elsewhere in this volume, not only by Reinhardt, but also by Hurley and Rice, among others. Nor are they the first to point it out; much of what they say is old news and yet the patterns they describe persist. Observing this persistence, in economics as in medicine, I am arguing that this behaviour is
a systematic phenomenon to be described and understood, not merely
an aberration to be deplored. I will also indicate what seems to me to be a relatively effective corrective strategy.

It may, however, be helpful to start by illustrating what I regard as persistent illogical and unsound analysis by competent professionals. Consider the classic summary of welfare economics by Bator (1957). As Reinhardt reminds us, in this volume and elsewhere (Reinhardt, 1992), that paper makes crystal clear a point that every professional economist knows: to rank social outcomes, you need a ranking rule, a social welfare function, embodying a set of values, that cannot be derived from economics itself. If you want to generate recommendations, normative conclusions that state A is better than state B, using economic analysis, then you must introduce those values first, explicitly or more often implicitly.

The Pareto criterion is one such externally imposed ranking rule; it imports a particular set of values into the analysis from outside. The justification may be that these are widely held, sort of universal values surely no one could object to changes that make some better off without making anyone else worse off? Indeed, they could object the Pareto criterion actually does impose restrictions on individual utility functions (non-malevolence, for one). But maybe in practice they dont? (And anyway maybe malevolent peoples values shouldnt count?).

As it happens, however, there may be a much more fundamental flaw in the Pareto criterion as a representation of our values. And it