Patient Experience and Physician Productivity: Debunking the Mythical Divide at HealthPartners Clinics
Troy J Boffeli, DPM, FACFAS; Kerri L Thongvanh, MBA; Sarah J Horst Evans, MA; Clay R Ahrens, MHA
Perm J 2012 Fall;16(4):19-25
Introduction: Physicians are continually encouraged to be more productive while providing higher levels of patient satisfaction. It is a common presumption that the two goals are somewhat exclusive—that higher productivity must entail a sacrifice in patient satisfaction or vice versa. Moreover, physicians seeking tested, evidence-based approaches to improving satisfaction have had relatively little to go on, and they commonly have justifiable concerns about how ineffective changes may hurt their productivity for no benefit.
Clinicians, whether operating their own practice or employed within another, have always been motivated to keep productivity strong. In recent years, however, especially as concepts such as pursuit of the "Triple Aim" of best care, best experience, and lowest cost1 have been embraced by leading health care organizations, many physicians are facing simultaneous expectations for delivering stronger patient satisfaction and a better patient experience.
Many published studies have examined the links between different aspects of physician behavior and patient satisfaction.2-5 Although most physicians would like to improve their productivity and patient satisfaction ratings simultaneously, several interrelated obstacles have often frustrated their attempts to achieve this.
Many physicians, for example, tend to view the balance between productivity and patient satisfaction as something of a zero-sum game: that improvement in one area can occur only at the expense of the other. Physicians with weak satisfaction ratings may believe that strong satisfaction ratings result from having a natural knack for dealing with patients that they do not happen to have. They may believe that patient satisfaction survey instruments are flawed, or that only dissatisfied patients complete them.
In our multispecialty practice, we know that strong productivity and strong patient satisfaction ratings are not necessarily exclusive. We have physicians across many specialties who achieve both, as well as those who achieve neither, despite working with similar patient populations, working conditions and facilities, and staff support.
Recognizing this, in 2008 we began work leading to an observational pilot study conducted in our organization with three purposes. We first wished to identify behaviors and characteristics of high-performance care teams and understand the differences, if they existed, in behaviors and characteristics of physicians with strong productivity and satisfaction ratings and those of physicians with weaker ratings. Second, we wished to develop and validate a set of physician and care team observation tools. Third, we wished to develop an encounter structure detailing best practices for optimal workflow and a care model process to help guide division of tasks among the care team. In doing so, we wished to give clinicians a useful tool for assessing their performance that would also suggest potential best practices for weak-performing clinicians to emulate. Although our organization has made advancements in all of these areas since beginning this work, in this article we will focus on the foundational work behind the first and second goals.
In our organization, as with many health care organizations, physicians face increasingly high expectations for achieving strong patient satisfaction ratings. In our experience, addressing weak patient satisfaction ratings has long been problematic for a number of reasons. These reasons include the physician's skepticism or reluctance to change as well as misapprehensions about patient satisfaction and productivity.
Overcoming Skepticism or Personal Reluctance to Change
Not surprisingly, most physicians do not welcome the news that their satisfaction ratings are weak when they first hear it. In our experience, they often progress through four similar stages (Table 1): 1) resistance/denial, 2) initial acceptance, 3) initial frustration, and 4) seeking meaningful help. Reflecting on these stages, we identified two obstacles contributing to difficulty in helping move physicians through these stages. First, we found that from the physician's perspective, consideration of patient satisfaction experience as a stand-alone goal could seem unfair because it did not take productivity and clinical effectiveness—also important measures—into account at the same time. Second, we recognized that physicians, once persuaded that their satisfaction scores could and should be strengthened, had few practical tools available to help them accomplish it, especially those grounded in evidence relevant to their own practice.
Misapprehensions about Patient Satisfaction and Productivity
Besides skepticism or reluctance to change, another obstacle to improvement in patient satisfaction is the presence of misconceptions. Among physicians, there is a common presumption that strong productivity and strong patient satisfaction are an "either-or" choice. Physicians who accomplish one well may presume that improvements in the other area must come at the cost of the part of their practice that is going well.
If asked to improve patient satisfaction scores, physicians may presume that changes to their workflow would be needed to achieve the improvement. They may presume that they need to see fewer patients, have more time with each patient, have more staff support, or other changes. They may also question the validity of the instruments used to assess patient satisfaction. Table 2 summarizes seven common beliefs drawn from our experience.
Whatever the perceived obstacle may be, these beliefs may reinforce physicians' resistance to change or their denial that a need or opportunity for improvement exists. Furthermore, without being shown evidence to counter these beliefs, physicians may be unwilling to make changes if they sense a potential risk to productivity.
It should be noted that some of these beliefs are, at times, true. For example, it is true that some physicians are more naturally adept than others at behaviors that promote strong patient satisfaction. We surmised, however, that these beliefs alone are insufficient support for the broad premise that physician productivity and patient satisfaction are exclusive. We simply lacked a way to demonstrate that persuasively to physicians in our practice, and we believed that an approach involving some degree of rigor would be useful. Therefore, we undertook the two-phase study described herein.
Our organization, HealthPartners Inc, operates a comprehensive specialty center in St Paul, MN, in which this research project took place. The project involved clinicians in four specialty departments located in the same complex: Orthopedics, Podiatry/Foot and Ankle Surgery, Gastroenterology, and General and Vascular Surgery. The physicians work in largely similar conditions. They have the same kinds of patients, similar workloads, the same facility and support staff, and so forth.
Observation was conducted of physicians and, to lesser extent, staff. In this report, we focus on physician-related observations and findings, because we recognize them as the strongest drivers of patient satisfaction.
Phase 1 Methods
Our first goal was to assess our physicians' lack of knowledge about the true range of their productivity and satisfaction performance. To explore the baseline productivity-satisfaction equation, we plotted individual measures for physicians from our 4 departments on a scattergram, distributed into performance quadrants for both productivity and patient satisfaction (weak and strong scores for each). The productivity score was based on productivity targets established by the Medical Group Management Association (now MGMA-ACMPE, Englewood, CO), with the 63rd percentile as a nationally recognized goal for financial stability in multispecialty groups. Patient satisfaction scores were based on a sampling derived from the ongoing collection of our organization's survey data obtained from NRC Picker (Lincoln, NE) and an organizational goal of a "problem score" of no higher than 18%. This score refers to the percentage of patients having a negative response to being asked whether they would recommend the physician's office to family or friends.
The plotted baseline data were shared with physicians in our practice. By showing physicians their individual place in the scattergram and the places of others in the overall practice, we hoped that physicians would be able to understand clearly where their performance is rated alongside their peers.
Phase 2 Methods
Having plotted the baseline performance data, we constructed an observational study to examine characteristics and behaviors of physicians in each of the quadrants. We also designed a standardized observational tool for use during the assessment. A summary of the methods in this phase appears in Table 3.
Step 2 of Table 3 is detailed here. A trained, independent observer observed approximately 24 physicians as a "fly on the wall" during a half-day of patient encounters involving approximately 200 patients. We selected a nonclinically trained observer to allow for focus on behavioral aspects of the visit and to avoid potential distraction of observing care-related events. Using a set of more than 100 internally developed observational criteria, the observer took note of the presence or absence of characteristics and behaviors demonstrated by the physician during the encounter. These behaviors included whether the physician offered a handshake and a smile, whether treatment alternatives were offered, whether medication side effects were explained, and so forth. Behaviors were rated by the observer on a 5-point scale ranging from "Always" to "Never." The same observer was present at each observation and had no knowledge of which performance quadrant each physician's baseline productivity-satisfaction rating fell.
The observer noted aspects of workflow details surrounding the visit, such as the preparation of the examination room, the rooming process (nurse calling patients from the waiting area, taking vitals, clarifying reason for visit, and prepping patients for the exam or procedure), and discharge. Using a separate questionnaire, the observer also interviewed physicians and staff and performed exit interviews with patients about their experience during the visit.
The goal of using an observer was to allow assessment of characteristics and behaviors that are difficult for a physician to assess on his or her own. The observation of workflow behaviors allowed an itemization of which members of the care team were performing specific tasks.
The results of all observations were synthesized and referenced against individual physician productivity-satisfaction scores to identify associations between the noted characteristics or behaviors and the ratings plotted on the productivity-satisfaction scattergram. All physicians and staff in the study were debriefed in one-on-one meetings, during which they were shown their physician productivity scores as well as the sets of behaviors and characteristics common to those achieving stronger scores. This, we hoped, would give the physicians a set of relevant, evidence-based tactical suggestions for what they could do to affect their scores in the desired direction.
The plotted baseline data (Figure 1) countered the idea that strong productivity and strong patient satisfaction could not be achieved simultaneously. The aggregate distribution (which looked generally similar when plotted by individual departments within the aggregate) suggested that we had roughly equal numbers of physicians in each of the performance quadrants. The quadrants were as follows: physicians achieving strong scores in both categories (upper left quadrant of Figure 1), physicians with strong productivity and weak patient satisfaction (upper right quadrant), physicians with weak productivity and strong satisfaction (lower left quadrant), and physicians with weak scores in both categories (lower right quadrant). Physicians tended to remain in their given quadrants consistently, over the period of the study and over time, both before the study and in the quarters to follow.
There were essentially no variables that could explain the distribution of the data other than that the physicians and their care teams in the quadrants were doing something differently—that they had different characteristics and behaviors in their workflows and with their patients. This provided a counter for the commonly held beliefs in Table 2. The data, when shared with physicians in our practice, were arresting. The data showed, at a glance, that strong productivity and strong patient satisfaction could be achieved in tandem.
These findings suggested to us that identification of associated behaviors and characteristics for each quadrant could be a valuable step in helping to provide evidence-based context for physicians, their supervisors, and their care teams in the assessment of their own productivity and patient satisfaction ratings. It also suggested that the potential existed for development of meaningful recommendations designed to improve those ratings.
Sample Scenario: Physicians "A" and "B": In the data set shown in Figure 1 are two physicians whom we will call "Dr A" and "Dr B." At the time of the ratings, the physicians practiced in the same clinic, and the essentials of their practices were identical; they shared the same building and staff, and handled the same kinds of cases. A patient calling for an appointment would have an equal chance of being seen by either physician.
Drs A and B, however, achieved dramatically different ratings for both productivity and patient satisfaction—the highest and lowest ratings in a single practice subset. In the scattergram (Figure 1), Dr A was plotted in the "strong-strong" quadrant (strong in both productivity and patient satisfaction), whereas Dr B was in the "weak-weak" quadrant (weak in both categories). Shown in Figures 2 and 3 are their baseline individual productivity and patient satisfaction scores, illustrating, for example, that Dr B had 40% less productivity during the 2-year period of 2008 to 2009 than did Dr A.
In the observational study, the cross-referencing of observational data with productivity-satisfaction scores, although based on qualitative assessments, yielded fairly clear and discrete sets of common characteristics for physicians and staff in each of the performance quadrants. Although space prohibits a complete listing of physician characteristics found, common characteristics for strongest and weakest scoring physicians are shown in Tables 4 and 5, respectively. These characteristics incorporate both interpersonal behaviors and administrative or workflow-related aspects of the patient encounter.
Sample sizes used in deriving these characteristics were not large enough to allow statistically significant results, but we did note some interesting qualitative aspects. For example, the "strong-strong" physicians scored "always" in showing characteristics associated with warmth, such as smiling, shaking hands or initiating touch, and using a friendly, familiar greeting, whereas the "weak-weak" physicians scored "mostly" or only "sometimes" in these areas.
Interestingly, "weak-weak" physicians were actually more likely than were "strong-strong" physicians to perform a certain desired behavior: concluding a visit by asking the patient if s/he had any further questions. This may suggest that the ability of the physician to connect with the patient and listen during the visit has a stronger bearing on patient satisfaction than how the physician concludes the visit.
"Dr B"—One Year Later: Although analysis of changes in performance in productivity and patient satisfaction is still ongoing, a sample postintervention result for the weak-scoring "Dr B" referenced earlier is shown in Figure 4. Over the course of two years, we saw Dr B progress through the four stages listed in Table 1. When shown data depicting individual weak performance within the context of the overall practice, Dr B at first felt frustrated and unable to understand what could be done about the weakness of the scores. Ultimately, however, Dr B reached a point of readiness to be counseled on what was needed for improvement. Figure 4 depicts Dr B's patient satisfaction improvement after being debriefed and presented with suggestions based on the observational study.
The behaviors and characteristics in each of the performance quadrants were generally what one might have predicted, were one to surmise what drives strong and weak ratings. By associating the behaviors and characteristics with actual performance data, however, we were able to demonstrate a real, relevant, and localized connection between what our physicians were doing and how they were performing.
When shown their individual place in the scattergram, and the places of others within the overall practice, physicians were able to see and understand clearly where their performance rated alongside their peers. It also suggested that movement toward the "strong-strong" category was at least theoretically possible. For weaker performers, a common reaction at this point was, "I can't help it if patients don't like me. I don't know what I'm doing wrong. I'm working hard and nothing is getting better. Tell me what I need to do because I don't want to be that dot."
We presented the performance improvement for one physician ("Dr B"). Given the limited scope of this study, we obviously cannot present this individual, qualitative finding as a generalizable finding, and we make no such suggestion. We can say, however, that the experience described here was quite representative of others in our practice and that it suggests some interesting directions for further study.
Implications for Our Organization: Physicians, even those with a sincere desire to improve patient satisfaction, are often stymied by a lack of meaningful guidelines and best practices for how to achieve it beyond a vague admonition to "do better." The literature contains many examples of expert opinion on methods for improving patient satisfaction or best practices for constructing patient satisfaction surveys, but scientific studies are far less common. Physicians seeking tested, evidence-based approaches to improving satisfaction have had relatively little to go on, and they have justifiable concerns about how ineffective changes may hurt their productivity for no benefit. Alternatively, they may presume that available guidelines lack rigor or relevance to their specific situation.
Our work on this project was founded on our increasingly evidence-driven approach to the issue throughout our Medical Group. For example, on the basis of the pilot observational tool constructed for this study, we developed a simplified observation tool for physicians and staff during patient encounters, which is now in use (see Sidebar: Physician office visit checklist).
The observational study (Phase 2) was foundational to the subsequent development of an experience framework now in use throughout our Medical Group after widespread review and feedback from our clinicians. It presents a simplified model depicting the elements of strong patient satisfaction: 1) a warm, genuine greeting; 2) patient-centered care; and 3) a warm, genuine close. The framework (Figure 5) includes an at-a-glance overview of what these attributes look like in practice.
The study also paved the way for the adoption and welcome acceptance of a shadow coaching program for physicians in 2010. All of these elements are now key elements in our overall improvement toolbox.
In sharing the findings of the observational study, we do not mean to suggest they represent a set of directions or a prescriptive "how to," but rather that they identify common characteristics of strong and weak performers that physicians and their care teams can use as an evidence-based basis for evaluation and self-reflection. We have translated this understanding into the production of educational tools that have helped our physicians broaden their understanding of the kinds of behaviors and characteristics typical of performers at different levels.
In our practice, we have found these educational tools to deliver practical assistance for physicians performing at any level to assess their own situation and to chart a path, on their own or with coaching, that leads to improvement. The results of this study also help to dispel the commonly held myths about the exclusivity of productivity and patient satisfaction, in that they show there are many physicians who excel in both areas simultaneously, and there are different characteristics associated with varying levels of performance.
Strengths and Limitations: A strength of this study was the use of an objective observer who was blinded to the physicians' individual baseline performance scores—we would even suggest that an equal, if not greater, strength was the use of a nonclinically trained observer—this ensured the focus was on behavior, not medicine. Use of an observer was helpful in that it allowed assessment of characteristics and behaviors that are difficult for a physician to assess on his or her own. As individuals, it is difficult to ascertain how others perceive us. The observer's assessment of posture, facial expression, emotional reactions, and other variables was often eye opening, and many physicians and staff were surprised by what was shared with them afterward.
This pilot study was relatively modest in scope, however, and findings were derived in part from qualitative observation. We make no assertion about the broad applicability of these findings to wide-scale, systemic improvements in practices or care systems. The findings help dispel commonly held myths about the exclusivity of productivity and patient satisfaction, suggesting that 1) there are many physicians who excel in both areas simultaneously, and 2) there are different characteristics associated with varying levels of performance. We do assert, however, that our work has helped our own physicians improve while reinforcing the need for further development of evidence-based methods for improving the patient experience while enhancing—not sacrificing—productivity.
The author(s) have no conflicts of interest to disclose.
Kathleen Louden, ELS, of Louden Health Communications provided editorial assistance.
1. Berwick DM, Nolan TW, Whittington J. The triple aim: care, health, and cost. Health Aff (Millwood) 2008 May-Jun;27(3):759-69. DOI: http://dx.doi.org/10.1377/hlthaff.27.3.759