A decade of patient satisfaction survey results: lessons learned in a large multispecialty group practice.
The patient satisfaction program originally was founded on a number of interrelated principles:
* Executive leadership demonstrates an ongoing commitment to improving service to our patients and to patient satisfaction. This is reflected in ongoing monitoring of performance at the executive and board levels and explicitly incorporating patient service goals in the strategic plans and vision for the future.
* Systems of accountability for performance are implemented as broadly as possible and are imbedded in the personal evaluation process of our professional staff and in our compensation programs.
* Improvements, in part, come through skill building. A series of education programs, some mandatory, were developed and are presented throughout the organization.
* Measurement of our performance from the patient's perspective is done on a continuous basis using patient satisfaction surveys. System-wide summary results are reported at least semi-annually. The results are used to guide performance improvement activities.
Given the potential to improve health outcomes through better patient satisfaction and the trend for more public reporting of these results, it is incumbent on physician leaders and other health care executives to have a sound understanding of how to interpret and use patient satisfaction data.
Initially, we used the Parkside survey to gather patients' opinions about our services. Parkside was acquired by Press Ganey and we continued with them after the acquisition. Patients are asked to respond to questions in various categories including:
* Interactions with the front desk personnel
* The facility
* Interactions with nursing
* Experience with the doctor
There are 10 questions that specifically relate to the physician-patient interaction. Typically, every six months we receive results on the performance of our physicians from over 30,000 of our patients compared to over 850,000 patients in nearly 33,000 physicians' practices in the Press Ganey database.
Among the lessons we learned are:
* Results must be benchmarked, individualized, and specialty-specific to be actionable.
* It's the experience with the doctor that drives the patient's overall impression of service.
* It's all about the 4s and 5s, not about service recovery.
* Productivity and patient satisfaction are not mutually exclusive.
* Reporting results and systems of accountability help capture the physician's attention.
* Hiring for success is better that trying to change a poor performer.
Despite overall deterioration in satisfaction with the health care nationally, most patients are generally well-satisfied with the physicians who provide their own care. This is reflected in the fact that the raw scores of patient satisfaction surveys are relatively tightly grouped at the high end of the range.
Patients usually do not confront their provider when dissatisfied; they often "vote with their feet" by finding a different doctor, a process that may be unnoticed by the physician. For these reasons, physicians rarely have an accurate sense of their own performance in patient satisfaction without comparative data with other providers.
Since almost all physician-patient interactions in ambulatory care are not observed by others, variation in performance in physician-patient interaction is usually unknown unless measured by patient satisfaction surveys. Since these interactions are one-on-one encounters, it is critical for the physician to receive feedback related to his or her individual performance.
Certainly, aggregate results that paint a picture of a clinic operation can provide important insights for process improvement, but only individualized data are actionable by each practitioner.
Early in our use of satisfaction data we drew some erroneous conclusions about individual performance based on the fact that the data for our physicians were being reported in only two groupings--specialists and primary care providers.
We were concluding that our obstetricians and cardiologists were almost all top performers while our orthopedists, neurosurgeons and pain therapy physicians were usually at the low end of the curve.
In looking into the details, we found that the mean scores for obstetricians and cardiologists across the country are generally at the high end of the performance curve, while orthopedists, neurosurgeons, and pain therapists generally received lower mean scores from their patients.
Clearly, patient expectation and outcome influence satisfaction. Therefore, it is necessary to make specialty-specific comparisons. This reflects the reality of the market as well. Cardiologists compete with other cardiologists, not all specialists, just as orthopedists do.
Using specialty-specific data does pose a problem for certain highly specialized physicians. There are many cardiologists in the database, so comparison is both valuable and fair.
But, there are not a lot of pediatric nephrologists in the database, for example, so specialty comparison becomes difficult. In these instances, we have found it helpful to aggregate physicians somewhat (such as pediatric subspecialists) and to look at the individual's rank within our entire specialty practice.
One interesting observation is that, in some of our smaller practice sites where there may be only two physicians and one nurse, that nurse will be rated very differently by the patients of the two doctors.
The patients of a physician who achieves high patient satisfaction will also give the nurse high scores, while the patients of a low performer will give the same nurse lower scores. Simply put, patients come to see the doctor and that experience drives the patient's overall impression of the visit.
We have also experienced the frustration of failing to see an improvement in low patient satisfaction results when a practice moves from an outdated, inconvenient facility to a new one. If the physicians maintain their same style of relating with patients, scores do not change. We've seen phenomenally high performance coming out of some of our most physically challenging practice sites.
Providers generally fall into three groups of performers--a few at the top, a few at the bottom, and a huge middle group that are very similar.
Let's address the huge middle group first. Given the highly competitive process to become a physician and the relatively similar personal characteristics required to successfully complete medical training, it is not surprising that the majority of physicians make up a tightly clustered group of providers that do an admirable job in satisfying their patients.
Within many specialties, a change in a decimal point or two in a raw score can lead to a large change in percentile ranking. For this large majority we have successfully provided educational programs aimed at moving the curve to the right, but we do not react when a provider experiences a drop from the 70th percentile to the 40th.
There are a few top performers whom we arbitrarily define as those ranked above the 90th percentile for their specialty. And, there is a group of low performers that we define as performing below the 15th percentile. We have found these groupings to generally make sense.
The groupings are usually confirmed by the answers to the following questions:
* "Is this a provider that you would have your spouse or child see?"
* "Does the staff in the clinic recommend this doctor to their friends?"
A high disenrollment rate from a physician's practice is often associated with low scores, but the individual provider is usually unaware of these rates unless they are shared with him or her.
We have seen that an individual physician's score generally tends to be within the same general range over time. An unexpected drop in an individual's scores may reflect a problem unrecognized by the supervisor, such as personal difficulties or even the onset of clinical depression. For this reason, it is important to tailor leadership's response to low scores to the individual in order to assess the situation correctly.
4s and 5s
It is a misconception that implementing effective service recovery programs to deal with dissatisfied patients will lead to substantial improvements in survey results.
This is not to say that service recovery is not important or a good thing to do. Fixing a process that went wrong is a huge opportunity to create a loyal customer. But the fact remains that there is only a small proportion of patients who are dissatisfied.
Even our worst performing providers get only several percent of survey responses in the very poor (1) or poor (2) range. What distinguishes an outstanding provider from the rest is the proportion of 4s (good ratings) and 5s (very good ratings) in patient responses.
As shown in Table 1, even the lowest scoring physicians have nearly 90 percent of responses in the good or very good range and only a few percent of responses in the very poor or poor range. For this reason, most poor performers are not aware through patient interactions that they rank so poorly.
Most of their patients are "satisfied," they are not faced with a lot of direct complaints, and they have the majority of their patients who are very satisfied. It is more common than not that when a supervisor confronts a poor performer with the poor score results, the physician responds with a number of thank you notes from grateful patients as evidence that the data are flawed.
Physicians and practice managers are challenged to increase productivity while improving patient satisfaction. But it is often argued that productivity must be sacrificed to achieve high levels of patient satisfaction. We have not found this to be true.
A review of clinical relative value unit (RVU) productivity of our physicians in our multispecialty group practice showed that there is no correlation between high productivity and poor patient satisfaction.
In fact, as the average RVU per physician increased significantly since 2001, the overall patient satisfaction scores improved as well. We have also seen a trend that our most productive physicians are generally at the higher end of the patient satisfaction rankings.
Physicians are a competitive lot. Most clearly want to be "the best" and to provide great care for their patients. Routinely providing each physician with his or her patient satisfaction survey results that are benchmarked to the specialty and ranked within the practice is a powerful motivator for most.
Within our practice, the compensation plan provides for very modest incentives and withholds of compensation (generally much less than 1 percent of total compensation) for outstanding or poor results respectively. The message is clearly more important than the dollar amount.
Simply "putting the money where our mouth is" as far as patient satisfaction has sent an important message to our physicians, even if it is just a little bit of money. Our plan allows for poor performers to have compensation withholds "forgiven" (not implemented) if they enter "remediation" programs in physician-patient communication skills. Most do so. It is a small investment on the part of the group to assist a group of our physicians to improve their performance, and potentially their own professional satisfaction.
The programs for poor performers include small group sessions with standardized patients, individualized coaching during clinic visits with real patients, or specialized support such as diction coaching.
About a third of those who have received this support will show sustained significant improvement in their scores. Given the cost of recruiting and orienting a new physician, we have judged these efforts to be cost effective. They also help identify for clinical leadership those poor performers who cannot or will not improve.
But defining the potential of a candidate to successfully satisfy patients during the interviewing process is more desirable than trying to remediate a poor performer after hire.
Reviewing patient satisfaction performance data from previous positions, if available, requesting that referees comment on performance in this arena, and behavioral event interviewing techniques may assist in better hiring practices.
In the end, patient satisfaction is an important social outcome in health care, a key determinate of professional satisfaction, and it influences other health outcomes as well. In the coming era of increased accountability for and public reporting of patient satisfaction results, it is important for physician leaders to understand the nuances of these data. Accurate interpretation provides the foundation for management actions that can lead to improvement.
Robert C. Spahr, MD, is an associate in neonatology and the senior vice president for service quality at Geisinger Health System in Danville, PA. He can be reached at firstname.lastname@example.org.
John Gerdes, PhD, is an associate in psychiatry in the department of service quality at Geisinger Health System in Danville, Pa. He can be reached at email@example.com.
Randall L Hutchison, MBA, is the director of customer service and performance enhancement for the Geisinger Health System in Danville, Pa. He can be reached at firstname.lastname@example.org.
By Robert Spahr, MD, John Gerdes PhD, Randall Hutchison MBA
RELATED ARTICLE: About GHS
The Geisinger Health System (GHS) is an integrated health care system that provides services to 2.8 million people living in 37 counties in northeastern and central Pennsylvania. The system comprises four hospitals, a health plan, an alcohol and drug dependency treatment facility, two ambulatory surgical centers, and a multispecialty group practice. Our group practice, the Geisinger Clinic, employs over 600 specialists and primary care physicians who practice in clinics at the Geisinger hospitals and at over 35 community practice sites located across our service area.
Table 1 Selected Physician Percentile Rankings Related to Patient Survey Responses Provider percentile rank Very poor Poor Fair Good Very good > 90 0.0% 0.1% 1.0% 8.2% 90.7% 62 0.0% 0.3% 1.8% 20.1% 77.8% < 5 0.6% 1.4% 6.1% 37.8% 64.2%
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Searching for Satisfaction|
|Author:||Spahr, Robert; Gerdes, John; Hutchison, Randall|
|Date:||Sep 1, 2007|
|Previous Article:||Chart your ACPE course.|
|Next Article:||'Can't get no (physician) satisfaction?'.|