May 2016

Measure Care Processes to Improve Outcomes
By Dan Rubin, MD, MHI
For The Record
Vol. 28 No. 5 P. 6

Transforming health care requires more than just measuring outcomes, which by definition are an end result. To truly transform care, it's necessary to start much earlier during the provision of care. This transformation also requires a different type of measurement: clinical process measurement.

The problem for health care organizations struggling with quality performance, unnecessary care variation, and cost containment is that measurement comes after the fact. By this time, patients who developed unnecessary complications have already been harmed, inpatient stays have been lengthened, readmissions have occurred, performance has been publicly reported, and extra costs have been incurred.

To remedy this situation, clinical care processes must be managed and measured earlier in a sustainable, repeatable manner—preferably a method that doesn't require an army of additional personnel carrying patient lists with potential gaps in care to be checked one by one. While that might sound easy in theory, in practice it may be nearly impossible.

Health delivery systems have struggled to build consistency around the process of care for as long as medicine has been practiced. Around 400 BCE, Hippocrates wrote a compendium of more than 50 medical texts that included comprehensive guidelines on how to treat patients for conditions as varied as war wounds and infectious disease. For as long as there have been guidelines, there has been variation in care. Today, when care strays from accepted guidelines without sufficient reason, it's dubbed an "unnecessary variation," a situation that occurs daily at most health systems.

Real-Life Scenarios
For example, hemoglobin A1c (HbA1c) tests are used to understand long-term glucose control in diabetic patients. If the measurement is too high, it typically indicates that tighter glucose control is required either through changes to medication or enhanced compliance. If the result is too low, the patient may be too tightly controlling their glucose and may be at risk for hypoglycemic events. Because either scenario is dangerous, standards bodies such as the American Diabetes Association (ADA) have recommended that physicians administer an HbA1c test every six months to their diabetic patients. When the test is performed within the recommended interval, patient care follows the accepted guideline. When there is variation from the guideline, it typically is unnecessary.

How can the variation be fixed before the outcome is gathered and publicly reported? The solution starts with gaining access to accurate data on the care process. However, it is important to understand the difference between measuring outcomes and measuring the process of care. For one diabetic patient, the outcome measure would be whether the HbA1c test was ordered within a given six-month interval. For a group of diabetic patients cared for by a provider, a clinic, or all clinics within a health system, outcomes may be measured by the percentage of patients who had the test performed as recommended by the ADA. Either way, these are considered to be outcome measures.

Clinically Speaking
Measuring clinical process is fundamentally different. How did the providers who met the recommendations for administering the HbA1c test order the test? Were they prompted within the EHR? Or did they remember on their own? Did they use the order set? Did a documentation template remind them to place the order?

All of these questions representing clinical process metrics focus not only on the results but also on how the outcome was achieved, better known as clinical process measurement. Just as outcomes can be aggregated by a provider, a clinic, a hospital, or a health system, it's possible to examine clinical process data in the same manner to understand where variability occurs and how it can be improved.

The opportunity to eliminate unnecessary variation is where clinical process measurement shines. Once it's understood how providers are or are not being guided to the right actions, health care organizations can take steps to improve that guidance. For example, if an alert to remind providers to order an HbA1c in appropriate cases is being widely ignored, clinical process measurement can delve into the reasons. Is it being displayed on the correct patients? Is it being displayed in care settings and to providers who can take appropriate action at the time? Without measuring this aspect of the clinical process, health care organizations would be unable to identify problems and miss opportunities to improve the process and patient outcomes.

Hurdles to Clear
However, there's a rub (isn't there always?). Clinical process measurement is challenging, often requiring a high degree of technical skill to carry out. It involves crunching vast amounts of data in ways that require a significant amount of specific knowledge about clinical processes and the role of EHRs. In addition, it's necessary to understand how that process is evidenced in data deep within the EHR.

Putting all this knowledge together and displaying the results in a way that is meaningful to the people who work to standardize clinical process is a major challenge. However, when done correctly, the results can be impressive. For example, every year there are 560,000 cases of catheter-associated urinary tract infections (CAUTIs), leading to an estimated $500 million increase in health care costs caused by prolonged hospital stays. More importantly, the infections result in a higher risk of mortality.

Approximately 75% of hospitals are not monitoring catheters for duration and/or discontinuation, which presents additional challenges. But what happens when clinical process measurement is applied? Recently, a large health system introduced clinical process improvement to combat its CAUTI problem. According to HIT vendor LogicStream Health, preliminary results over a six-month period indicate a 33% reduction in the CAUTI rate. All of the system's eight hospitals used clinical process improvement to determine where unnecessary variation was occurring and suggest possible solutions.

The findings were intriguing and somewhat unexpected. The health system initially focused on having fewer catheters placed, the idea being that fewer catheters would equate to fewer CAUTIs. Makes sense, right? However, after measuring the clinical process involving catheters, an even bigger opportunity for improvement was uncovered: standard care of catheters once they were in place. As a result, the focus shifted to ensuring every patient with a catheter had an order placed so that the tube could be removed as soon as was feasible based on the patient's clinical condition. The clinical process was modified and measured iteratively to ensure it was easy for clinical staff to follow, easier in fact than the old method.

Clinical process measurement is a powerful tool in health care's arsenal to improve quality, reduce costs, and compete in a value-based environment. While it's a challenging undertaking, there are solutions available that can help organizations achieve their goals faster.

— Dan Rubin, MD, MHI, is board certified in pediatrics and clinical informatics and is chief medical officer and cofounder of LogicStream Health.