The paper by Maynard, Street & Hunter on ‘payment by results’ in this issue of Addiction[1] offers some provocative and useful suggestions to the British (and other) government payers who are contemplating performance contracting to enhance the effectiveness of substance use treatment. These are necessary suggestions, because behind the simple, direct appeal of ‘paying by result’ to policy makers are many important logical and procedural issues that must be understood if the procedure is to achieve its substantial potential. Consider the basic premises involved: an administrative act (purchasing) by a public purchaser will affect the intrinsic behaviors (services) of a treatment provider, which will in turn affect immediate and perhaps longer-term changes in patient behavior (outcome). In concept it is like a two-ball combination shot in billiards. Where the second ball ultimately lands is determined by the laws of physics resulting from how the first ball hits it, how the cue strikes the first ball and how skillful the player is in choosing the proper angle and strike on the cue ball. The good news is that the laws underlying reward-contingent behaviors are almost as consistent, powerful and useful as the laws of physics. However, like the billiards shot, performance contracting will only work if all the fundamental elements are aligned and properly executed. In what follows I discuss two essential—but often ignored—requirements for effective performance contracting.

The first requirement is that the ultimately desired (purchased) result should be observable (verifiable), truly valuable to the purchaser, and reasonably likely as a result of behaviors (services) by the treatment program. The most valuable performance criteria are patient outcomes with intrinsic public health value (e.g. abstinence, employment), but many of these desirable outcomes are quite distal to the programmatic behaviors that are expected to influence them; and this creates problems (see below). Next most valuable are interim patient behaviors that occur early during care and that are logically linked to more distal outcomes. These could include active participation and provision of drug-free urines during out-patient treatment. In contrast, performance criteria based on program behaviors such as completing a diagnostic assessment or timely filing of a treatment plan are not as predictably related to more distal patient outcomes—they rarely provide true value [2].

A second requirement for effective performance contracting is that the desired (purchased) result should be linked closely in time to a reward that is large enough to change the contingent behavior. This may be the most important requirement for effective performance contracting, yet this is often ignored, with invaryingly bad results. A purchaser might want 6-month post-treatment abstinence, and thus offer a performance contract that provides an extra $500 to a treatment program for every discharged patient who provides a drug-free urine at 6-month follow-up. This is an outcome with obvious public health value, and the performance measure (urine) is a reasonably valid indicator. However, there are three critical problems in the execution. First, any behaviors or actions that the program might undertake to meet the performance goal are at least 6 months removed from the performance criterion (post-treatment abstinence) and thus not likely to become engrained. Secondly, while treatment should be expected to reduce patient substance use while the patient is participating in out-patient care, it is debatable whether or for how long patient improvements will sustain after care is finished [3]. Reliably affecting that type of sustained outcome will probably involve post-treatment services such as telephone monitoring, attendance at Alcoholics Anonymous and/or periodic home visits [4]; and this raises the third problem. Most addiction treatment programs do not have significant reserve resources, so an opportunity for larger, performance contingent rewards such as this will almost always require the program to invest capital in new procedures, personnel and training. In turn, the earned rewards will have to be sufficiently large and timely to pay off the investment outlay by the program—$500, 6 months post-discharge will probably not do it.

There is much more that could be said about ‘payment by results’ or ‘performance contracting’. In my view, it is one of the most conceptually and politically appealing concepts for improving treatment effectiveness; but as briefly illustrated here, I agree with Maynard et al. that although purchasers should proceed, they should ‘proceed with care’.

Declaration of interests