In these days of belt-tightening, Western governments are targeting healthcare as an area where public spending needs to be streamlined. One way to do so is to move from fee-for-service reimbursement (each medical act is paid for) to a system where reimbursements are tied to performance, in this case, quality of care. The UK's National Health Service is trying out this “pay for performance” at the primary care level, as is France, where efforts are in earlier stages.
In the US, Medicare introduced such financial incentives as part of the Affordable Care Act. But how is quality of healthcare measured accurately? For some medical conditions, outcomes are clearly measurable; for example, mortality rate is a good indicator of care of cardiac condition. “But it's harder to measure improvements in quality of life after a knee or hip replacement operation,” says Dimitrios Andritsos. That is why much attention has been paid to processes, i.e., the series of medical steps to be followed, as they are easier to define and track. An example of process measure would be the fraction of heart-attack patients for whom aspirin, a blood thinner, was prescribed at hospital arrival, a medically recommended action. In theory, higher process quality, defined by better adherence to such medical guidelines, should improve care and efficiency, and diminish resource usage.
Tracking steps to improve cardiac care
To put this hypothesis to the test, Dimitrios Andritsos and his co-author Christopher Tang carried out an empirical analysis of cardiac care in a sample of US hospitals. They chose to focus on heart failure and acute myocardial infarction (commonly known as a heart attack) because these are widespread and expensive to treat conditions; incidentally, this is the same reason why Medicare's pay for performance program began by focusing on heart condition and pneumonia. The researchers used total length of stay – the number of days a patient occupies a hospital bed – as a proxy for resource usage.
They counted full episodes, meaning not only initial stays but also any readmission within 30 days, which is when patients come back if there are complications down the road. “It's important to measure patients’ total length of stay, because if they are discharged early, then need to be readmitted, it doesn't help, it just shifts resource usage to a later point in time,” explains Dimitrios Andritsos. They measured process quality by creating a global indicator of adherence to evidence-based medical guidelines. Some of these are clinical in nature, including delivering appropriate medication, running diagnostic tests, and so on, while others are administrative; for example, providing advice about quitting smoking.
The researchers did find that in hospitals where process quality increased, the average total stay of cardiac patients decreased.
Adhering to proven processes improves care
The researchers did find that in hospitals where process quality increased, the average total stay of cardiac patients decreased. Specifically, a 10% increase in adherence to medical guidelines reduced stay by 2.7% for heart-attack patients and by 0.8% for heart-failure patients. “When we took a more granular look, distinguishing the effect of clinical and administrative processes, we found that clinical ones mattered more,” says Dimitrios Andritsos. Another aspect of his empirical study examined the structure of hospital environments, which he supposed affected the nature of the relation between process and impact.
Certain hospitals deal with patients distributed across wider ranges of conditions, usually large hospitals in urban areas, while others deal with narrower ranges and are more “focused.” As Dimitrios Andritsos explains, operations management theory predicts that focused environments tend to be more efficient, because less operational complexity – in this case, fewer different medical protocols – means fewer chances of error and better understanding of processes. “Indeed, what we found was that the effect of process quality was more pronounced in 'messier' environments,” says the researcher.
The cost of reducing costs
Dimitrios Andritsos warns that the impact of quality process on performance is modest. The global figures for reduced length of stay from the study’s sample translate into an approximate reduction of respectively 0.13 days and 0.04 days per average patient. Knowing that an average hospital treats 189 heart attack patients a year, the annual reduction in inpatient stay would amount to 24 days, or just over $49,000 in savings, and less than half that amount for heart-failure patients. It may sound like a lot, but from the point of view of public spending on healthcare, it's really a drop in the ocean.
Also, as Dimitrios Andritsos says, such savings come at a cost. “There is an administrative cost to running process-quality programs and keeping track of quality depending on the environment in which you operate,” he says, emphasizing the need for a cost-benefit analysis. He also adds one caveat, namely that more sensitive indicators may be needed as quality processes evolve. Medicare’s pay for performance program in the US, for example, dropped certain indicators when they showed very high conformance rates across hospitals, making the indicators next to irrelevant. “The figure of $49,000 worth of savings is based on the indicators that are there, but the savings could be more pronounced with more sensitive indicators,” the researcher explains.