Legitimacy and fairness in priority setting in Tanzania.

Every health system has to make decisions about how to use limited resources to meet competing claims about diverse health needs. In all systems, national priority setting, including budgeting, generally imposes constraints on other levels of decision-making – be it in states or provinces, districts or cities, or local health authorities, hospitals, or health insurance plans. Decisions at any of these levels often are contested because they create winners and losers, sometimes on matters of life and death. Winners and losers have conflicting interests and claims. What is worse, we lack consensus on the distributive principles capable of resolving disputes about who should get what1. Reasonable ethical disagreement thus surrounds these conflicts of claims and interests. In addition, our economic tools for resource allocation, such as cost-effectiveness analysis have limited ethical acceptability, for they may controversially push us to maximize aggregate health benefits without adequate consideration of the fairness of the distribution that results.

In the absence of agreement on principles or methods that yield fair outcomes, we may need to develop a fair process for decision-making and accept the outcomes of such a process as legitimate and fair. This appeal to procedural justice rather than a more substantive view was the idea behind developing the fair, deliberative process called ‘accountability for reasonableness’ (2) that Steven Maluka writes about in his PhD thesis summarized in a PhD review (3). The argument for relying on such a process, however, is largely theoretical and is not based on evidence that it will actually produce more legitimate and fairer decisions (4). Although the theoretical arguments about why such a process may enhance legitimacy and yield fair outcomes may be plausible, it would be useful to see evidence that the process works and makes things better. Indeed, such evidence is something health ministers usually ask for. Gathering such evidence requires instances in which the process being evaluated is implemented appropriately; only then can we begin to assess whether it works. Indeed, having good explanations of what difficulties of implementation face the process is also an important area of investigation. And that is what Maluka's paper is about.

Maluka's paper is one of few that focuses on decision-making about health priorities in a low-income developing country (5, 6). It throws light on the difficulties facing implementation of such a process, given the structural and political realities of a country like Tanzania. Maluka's work, and the work of the multicountry EU-sponsored REACT project he describes (7), examines the acceptability of accountability for reasonableness at the district level. He finds that district-level decision-makers generally approve of the idea of a more transparent, deliberative, district-level process that includes a range of stakeholders seeking grounds for the priorities they set.

Despite this acceptability of the ideas involved in the fair process, the study finds significant obstacles at the implementation level. Two key difficulties stand out, namely the limited decision-space created by decentralization in Tanzania and the difficulties getting adequate participation by relevant stakeholders in the process. These difficulties are not unrelated. Tom Bossert has studied decentralization in various countries and found it to be a complex concept: the action or choice space that devolves to lower levels of a system has multiple dimensions, and some forms of decentralization give true authority to lower levels while others do not (8, 9). Maluka is clearly finding that Tanzanian decentralization does not devolve full authority to set priorities at district levels. In addition, lateness in delivering funds included in budgets means further constraints on carrying out any priorities that are set at the district level. It would be interesting to use Bossert's methodology to find in an independent way the features of incomplete decentralization that may be affecting the implementation of accountability for reasonableness.

The other central difficulty derives from the many ways there are to undercut a broad range of stakeholders and keep them from making a meaningful contribution to the deliberation. Maluka points to the need to provide adequate information to the stakeholders, and this is part of a general problem that community-level stakeholders often defer to those they see as having more expertise. In addition, there are other problems: power imbalances among the different participants may mean that some stakeholders are intimidated out of making the contribution they are capable of making. Maluka suggests that empowerment be added as a further condition on the process, as some others have proposed; for the suggestion to address the problem, it would have to aim for some form of equality in empowerment, otherwise those who are more empowered than others will use their power to distort the deliberation, perhaps converting it into an unwanted form of lobbying.

The need to gather evidence about whether accountability for reasonableness improves the legitimacy or fairness of outcomes has as a first step making sure cases that purport to involve the process actually do. That is still some distance from developing ways of measuring legitimacy or of determining when decisions are fairer than others, given the lack of a prior consensus on what counts as a fair outcome. Maluka's thesis is an important early step that may help us accumulate cases in which we can examine the impact of conformance more systematically and not rely on theoretical plausibility alone.