10.08.2013 Views

Pay for Quality

Pay for Quality

Pay for Quality

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

72 <strong>Pay</strong> <strong>for</strong> <strong>Quality</strong> KCE Reports 118<br />

However, on the various types of intermediate outcomes often also positive effects<br />

were found (S). This contrasts with long term outcome targets <strong>for</strong> which no significant<br />

effects were detected, whenever included (W).<br />

It is striking that most studies are focused on the correction of underuse of appropriate<br />

care, with varying P4Q results, whereas only two studies also focus on the overuse of<br />

inappropriate care (lab testing prescription, medical imaging prescription, drug<br />

prescription, etc.) 171 , 215 . The first study found positive results, the second reported<br />

mixed effects (some positive, some showed no significant difference and the effect<br />

seemed to decrease with time) (C). For the second study the behavioural independent<br />

nature of the P4Q scheme should be taken into account. GPs agreed be<strong>for</strong>ehand, by<br />

means of a democratic majority decision at a regional meeting, to improve prescribing<br />

according to a one page evidence based <strong>for</strong>mulary and to be more critical and efficient<br />

in their prescribing on antibiotics and gastric medicine in general. However, the bonus<br />

was per<strong>for</strong>mance independent, and given to all the GPs.<br />

Most P4Q studies are in compliance with SMART goals (specific, measurable, achievable,<br />

relevant, and timely), although it was often unclear why and by whom specific patient<br />

groups and targets were selected. If we look at P4Q results the reporting of a high<br />

involvement of stakeholders led more often to positive results, as compared to studies<br />

that did not report such an involvement (S). In addition, with regard to targets being<br />

relevant and timely, almost no study indicated that a previously detected quality<br />

problem (high variability or low per<strong>for</strong>mance) on a specific target was the reason to<br />

include it in a P4Q programme. Most studies referred to general lack of quality findings<br />

without assessing them in a local context as a first step, and to base indicator selection<br />

on those findings. Studies with a baseline measurement only used those results <strong>for</strong><br />

comparison reasons, not <strong>for</strong> selection purposes. Especially in acute care conditions a<br />

number of baseline reports with already a high per<strong>for</strong>mance level are presented (e.g. 80<br />

to 90% achievement on certain included measures). Looking at P4Q results this<br />

presence or absence <strong>for</strong> room <strong>for</strong> improvement seems to have been of influence on<br />

finding positive or no effects (S). A related lack in the reporting of studies concerns the<br />

dynamical aspects of quality enhancement. No studies report on a cyclical selection and<br />

refinement of indicators through time as a function of room <strong>for</strong> improvement. For the<br />

QOF a number of small scale changes to the indicator set have been made and others<br />

are planned, but their effects have not been studied yet.<br />

There is an evolution in the number of targets and indicators which are included in P4Q<br />

programmes. Programs during the nineties included often only one or a few targets.<br />

Later on, this number expanded gradually through the years with the start up of new<br />

programmes. The QOF with almost 150 indicators is one example. In terms of P4Q<br />

results a very limited target selection seems to correspond with lesser P4Q effects,<br />

especially in the domain of preventive care (S).<br />

5.1.2 <strong>Quality</strong> measurement<br />

Data validity and its acceptance are reported as sufficient in most studies. With the aid<br />

of the Healthcare Effectiveness Data and In<strong>for</strong>mation Set (HEDIS), which is standardized<br />

over USA health plans to foster comparability, and based on strict data validity criteria<br />

many studies make use of existing databases originally intended <strong>for</strong> other purposes<br />

(although mostly also financing related). Differences in data collection method<br />

(automatic extraction, secondary data use, additional registration; sampling versus<br />

continuous) have not led to substantial differences in clinical P4Q results (S). However,<br />

the effect of the choice of the data source on cost effectiveness of the programme and<br />

on provider satisfaction (due to the level of paperwork and workload effects) remains<br />

unclear after reviewing empirical results.<br />

A few studies tested <strong>for</strong> data coding quality and the level of gaming the system. Their<br />

results show that coding quality was generally good, but that there was also some<br />

gaming present, however to a limited extent (W). How detected gaming is dealt with is<br />

often not described.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!