Increasingly, hospitals, health systems and the federal government are pushing for “pay for quality” or “pay per performance” over the more traditional “pay for volume” as the underlying structure for physician compensation.

For instance
Recently, several medical groups and hospitals have shifted the payment emphasis to the quality model. For instance, Henry Ford Medical Group, located in the Detroit, Mich., area, is reportedly basing compensation for its 1,200 physicians more on how many patients choose them — a sort of popularity contest based on patient referrals and online reviews. Previously, 66% of physician pay was based on the volume of relative value units (RVUs). Now it has become a 50/50 split.

In addition, Geisinger Health System, a 12-hospital system headquartered in Danville, Pa., reportedly eliminated physician bonuses completely. In addition, it placed all 1,600 employed doctors on a straight salary at, or above, the national average in their areas of expertise. Previously, Geisinger paid doctors 80% straight salary, with possible performance bonuses of 20%.

Another physician group in Middletown, N.Y., links 15% of physician compensation to quality, cost and satisfaction measures, with that percentage expected to increase to 30% over the next three years.

Some issues
Quality and patient satisfaction are important and need to be weighed in physician compensation models. But decreasing the number of diagnostic tests performed, reducing the number of procedures performed and shrinking hospital stay periods doesn’t necessarily equate to improved quality or improved patient care. Such measures clearly decrease costs, but whether they amount to better patient care is less clear.

One of the basic tenets of productivity management is a reliance on RVUs, a system Medicare uses to determine how to reimburse physicians for the 9,000 or more services and procedures covered under its Physician Fee Schedule. The dollar amounts are linked to the Current Procedural Terminology (CPT) codes, and the three components of the dollar amount are calculated based on a physician’s work, practice expenses and malpractice insurance. A physician’s work is broken into four subcomponents:

1.      The time needed to perform the service,

2.      The technical skill and/or physical effort to perform the service,

3.      The amount of mental effort and judgment involved, and

4.       The stress related to potential risk to the patient.

But quality is a far more ephemeral issue. Some hospitals conduct patient surveys after every visit, with a link to the survey emailed to the patient. Like Henry Ford Medical Group, some pay is based on online reviews and patient referrals. Many of what are being called alternative payment models (APMs) are based on various metrics, such as the Medicare Access and CHIP Reauthorization Act (MACRA), which often take data from population data sets and then compare that data to individual physician and patient interactions.

Problems with that approach are fairly obvious — it’s hard to compare large patient populations to small patient populations or individuals, data sources aren’t comprehensive, and information systems aren’t standardized. These are technical issues. Political issues can involve physician stakeholder wariness and a lack of consensus over the appropriate quality metrics and how to report them to and from consumers.

At heart, the issue is that, while methods for objectively evaluating the amount of care are straightforward and well documented, benchmarks for value aren’t well documented — and are often fuzzy at best.

Data matters
The risk, of course, is that the patient will get lost in all the data-crunching, and/or that insurers and government agencies will be less interested in quality patient care than in driving down costs and reimbursements.

As health care systems experiment — and this is clearly an ongoing, long-term process — more data will become available about how these different models work, at least in terms of how much money they save and how they affect productivity. Until everyone agrees on what “value” is and how to measure it, the results are likely to be a constant topic of debate. Although patient surveys or referral rates don’t seem to be entirely objective metrics for value, they probably indicate a portion of value.

Going forward
Experiments in physician compensation based on value are likely to always have critics. They may be meaningful for some types of care and not for others. Patient noncompliance and other factors beyond the control of a physician, such as emergencies that throw off scheduling, can have a negative impact on patient surveys. Perhaps the search for “value metrics” is really a Holy Grail — and the experiments will only continue.