Quality conceptualization in the era of measurement: A qualitative study of primary care physicians in Medicare’s Merit-Based Incentive Payment System

Background: While governmental programs seeking to improve the quality and value of healthcare through pay-for-performance initiatives have good intentions, participating physicians may be reluctant to participate for various reasons, including poor program alignment with considerations relevant to daily clinical practice. In this study, we sought to characterize how primary care physicians (PCPs) participating in Medicare’s Merit-Based Incentive Payment System (MIPS) conceptualize the quality of healthcare to help inform future measurement strategies that physicians would understand and appreciate. Methods: We performed semi-structured qualitative interviews with a nationwide sample of 20 PCPs participating in MIPS who were trained in internal medicine or family medicine. We asked PCPs how they would characterize quality in healthcare and what distinguished exceptional, good, and poor quality from one another. Interviews were transcribed and two coders independently read transcripts, allowing data to emerge from the interviews and developing theories about the data. The coders met intermittently to discuss ndings, harmonize the coding scheme, develop a nal list of themes and sub-themes, and aggregate a list of representative quotations. Results: Participants described quality in healthcare as consisting of two components: (1) evidence-based care that is safe, which included appropriate health maintenance and chronic disease control, accurate diagnoses, and adherence to guidelines and (2) patient-centered care, which included spending enough time with patients, responding to patient concerns, and establishing long-term patient-physician relationships that were founded upon trust. Conclusions: PCPs consider patient-centered care to be necessary for the provision of exceptional quality in healthcare. Program administrators for quality measurement and pay-for-performance programs should explore new ways to reward PCPs for providing outstanding patient-centered care. Future research should be undertaken to determine whether patient-centered activities such as forging long-term, favorable patient-physician relationships, are associated with improved health outcomes. Overview A conceptual framework describing PCPs’ characterization of quality in healthcare emerged via inductive analysis of the interviews. When asked to dene healthcare quality, many respondents described the concept as consisting of two components: 1) evidenced-based care that is safe, and 2) patient-centered care that is responsive to individual patients’ needs. Most PCPs described that their rst responsibility was to deliver evidence-based care, which consisted of routine health maintenance, chronic disease management, care consistent with current recommendations, diagnoses that were accurate, and high-value care. Providing evidence-based care was viewed as necessary, but not sucient, for the provision of exceptional healthcare. Delivering evidence-based care along with patient-centered care was viewed as exceptional. Physicians often described examples of outstanding care as offering care that was particularly timely, responding to patients’ individual needs, or advocating for a patient under special circumstances. See Fig. 1 for a graphical depiction of this conceptual framework, Table 2 for a complete list of themes and sub-themes, and Appendix A2 for a list of sample quotes exemplifying all sub-themes. The following paragraphs explore the components of evidence-based and patient-centered care in further detail.


Background
The primary care physician (PCP) practice environment has been changing rapidly over the last several decades due to many factors, (1) including the development of quality measurement and pay-forperformance programs. In 2015, for example, Congress passed the Medicare Access and CHIP Reauthorization Act (MACRA), thereby rmly establishing a payment system that alters physician payment based on performance on measures of healthcare quality and value. (2) Under MACRA, physicians participate in the Quality Payment Program either through the default track-the Merit-Based Incentive Payment System (MIPS)-or via Advanced Alternative Payment Models, such as risk-bearing accountable care organizations. In the rst year of participation, approximately 700,000 physicians participated in MIPS by submitting data related to (1) quality, (2) cost, (3) improvement activities, and/or (4) use of certi ed electronic health record technology to the Centers for Medicare and Medicaid Services (CMS). (3)(4)(5) CMS reported that 94% of all eligible physicians participated, and 74% of clinicians reported quality data for a full twelve months. (6) Meanwhile, a counter-movement against quality measurement and pay-for-performance programs has been occurring. In a 2015 survey of PCP experiences and reactions to changes in health care payment, 67% of physicians reported believing that quality measures were having negative or no impact on their ability to provide high-quality healthcare to their patients. (7) Additionally, some evidence indicates that increased burden of administrative rules and regulations has negatively impacted physician satisfaction, (8) and that diversion of physician attention from patient-centered concerns may worsen health outcomes. (9)(10)(11) Prominent stakeholders now advocate for measurement parsimony, (12) a decrease in the amount of resources dedicated toward quality measurement, (13)(14)(15)) and a reduction of administrative burdens for physicians. (16) In particular, the MIPS program has fostered controversy and criticism from physician groups because of the burdens of data collection and reporting. (17,18) Thus in a contemporary clinical practice environment characterized by both physician ambivalence to measurement and an expansion of quality measurement activities, there is a need to examine physician conceptualizations of the quality of healthcare so that quality measurement program administrators can align future measurement programs to optimize physician engagement. (19,20) Thus, our objective was to describe how PCPs characterize the nature of quality in healthcare by performing interviews with PCPs reporting under MIPS.

Design, Setting, and Participants
We conducted a qualitative study using semi-structured interviews with MIPS-eligible PCPs in 2017 and 2018. To facilitate investigation of a variety of clinical experiences, we recruited PCPs from diverse practice settings across the United States by using maximum variation sampling. (21) We considered PCPs who were trained in Family Medicine or Internal Medicine to be eligible for the study if they were expected to report under MIPS. We veri ed MIPS expected reporting status by using an online tool. (22) PCPs were recruited by word-of-mouth, using a mixture of referrals from study investigators and physician organizations, including the Los Angeles County Medical Association. Interviews were planned to last approximately thirty to sixty minutes, and participants were additionally asked speci c questions about the MIPS program as described in a related manuscript. (23) We provided a $50 gift card as compensation for participation.

Description of interview sessions
At the beginning of each interview session, a brief pre-session survey was administered to con rm eligibility for the study and ascertain physician and practice characteristics. One author (CTB) performed all interviews either in person, for participants local to the Los Angeles area, or by telephone, for participants in other regions. In-person interviews were completed in participants' o ces and audio recorded. Telephone interviews were performed using secure audio recording. The interviewer used a semi-structured interview guide. (See Box 1 for a brief overview of interview questions. Appendix A1 provides the pre-session survey and the full semi-structured interview guide.) A professional transcription service transcribed all interviews. It should be noted that a related publication focusing on the MIPS policy itself describes results from the same interview sessions with the same PCPs. (24) Analysis We utilized a two-step process to allow data to emerge from the semi-structured interviews, developing theories about the data and maintaining a log of ideas about the meaning of its content. Two coders, CB and ME, independently read transcripts of the rst ve interviews and agged all instances of responses related to the study questions using the text analysis software Atlas.ti Version 8 (ATLAS.ti Scienti c Software Development GmbH, Berlin, Germany). We had an initial meeting to discuss and harmonize our coding schemes based on the emerging themes. We reviewed and coded through the eighteenth interview and then met again to discuss themes and sub-themes. At this point, we decided that thematic saturation had likely been reached. We coded two nal interviews without discovering new themes, con rming that thematic saturation had been achieved. We met once more to discuss ndings and develop a nal list of themes and sub-themes, aggregating representative quotations from participants.

Ethics
The institutional review boards of both Cedars-Sinai Medical Center and the University of California, Los Angeles, approved this study. We obtained verbal consent from all study participants, which included consent for audio recording.

Results
Descriptive characteristics of the sample Twenty PCPs participated in the study between November 2017 and June 2018. Eighteen of the twenty PCPs were board certi ed at the time of interview (seven in Family Medicine and eleven in Internal Medicine). Nine physicians worked in small practices (groups of fewer than fteen providers), and ve worked in rural areas. Most PCPs worked in practices where the majority of income arose from fee-forservice payment arrangements. Table 1 describes participants and their practice environments in further detail. Physician conceptualizations of healthcare quality: Overview A conceptual framework describing PCPs' characterization of quality in healthcare emerged via inductive analysis of the interviews. When asked to de ne healthcare quality, many respondents described the concept as consisting of two components: 1) evidenced-based care that is safe, and 2) patient-centered care that is responsive to individual patients' needs. Most PCPs described that their rst responsibility was to deliver evidence-based care, which consisted of routine health maintenance, chronic disease management, care consistent with current recommendations, diagnoses that were accurate, and highvalue care. Providing evidence-based care was viewed as necessary, but not su cient, for the provision of exceptional healthcare. Delivering evidence-based care along with patient-centered care was viewed as exceptional. Physicians often described examples of outstanding care as offering care that was particularly timely, responding to patients' individual needs, or advocating for a patient under special circumstances. See Fig. 1 for a graphical depiction of this conceptual framework, Table 2 for a complete list of themes and sub-themes, and Appendix A2 for a list of sample quotes exemplifying all sub-themes.
The following paragraphs explore the components of evidence-based and patient-centered care in further detail. Participants also expressed the belief that safety was an important dimension of healthcare quality: They strived to make accurate diagnoses and avoid low-value care to ensure that patients received high-quality, evidence-based care. One PCP explained that making an accurate diagnosis could be challenging, but also believed doing so consistently was a physician's responsibility: "Everybody can do checklist medicine and check [a hemoglobin A1C] twice this calendar year. [On the other hand,] when you have a patient with a di cult diagnosis which you establish and come up with an effective treatment for-now that's good care." There was also a sense among participants that low-value care could be harmful to patients, either because of costs or adverse consequences related to testing or treatments that were not indicated. One PCP provided an example by describing that his grandmother might demand care, but the right course of action was to withhold care in cases where potential harms outweighed potential bene ts: "My grandma was a very wealthy person and she was very demanding. Basically, she would want you to forget about standards in healthcare and spend every healthcare dollar on her and waste everything on her to get every test possible, which is absolutely not right because [it's wasteful and potentially harmful.]" See Table 2 for more examples of sub-themes that were categorized as related to evidencebased care.

Component 2: Patient-centered care
Many respondents reported believing that a good PCP provided patient-centered, personalized service. In other words, simply providing evidence-based medicine was not enough because patients needed and deserved excellent service as well. One respondent relayed, "I think broadly about right care, right place, right time for the patient, in a patient-centered way. I mean, I think in my own terms, I would think of quality care being all the care you need and none of the care that you don't need, delivered in a personalized way, in accord as much as possible with the best available evidence and done e ciently and in the ow of…working within the ow of patients' lives." Several participants mentioned that helping patients achieve personalized goals was more important than scoring well on quality measures: "I think our job as doctors is to help patients achieve their life goals, their health-related life goals. So for me, Several PCPs spoke about the importance of establishing long-term relationships with patients founded on trust. For example, one physician in a large group practice mentioned, "What people don't get is that if patients go to the person that they trust, they're more likely to be adherent to their medication or adherent to whatever regimen, if they know and trust the physician." PCPs considered demonstrating caring and compassion to be important, and they believed that this helped establish good rapport with patients: "I think the patients are looking for someone who will listen, someone who seems to be interested in the problems that they're presenting…Obviously, they want you to be knowledgeable and compassionate and timely and everything else, but I think patients want to be listened to and taken more seriously than anything else." PCPs perceived a contrast between evidence-based care, which tends to be feasible to measure, and patient-centered care, which tends to be hard to measure. Several physicians lamented that measurement priorities may be misaligned simply because measurement of patient-centered care is challenging: "The quality movement has been perverted to a kind of big data enterprise now and really, I think the focus of primary care is actually about relationships and building sort of meaningful relationships with people over time." Conceptualizing exceptional and poor healthcare quality To further re ne the PCP de nition of healthcare quality, we asked participants to describe what separated exceptional care from good care. In describing exceptional care, respondents tended to emphasize the patient-centered component of quality even more, including relationship-building and shared decision-making. One respondent explained, "I think, also, there's that human component that is hard to measure, the connection that people feel with their primary care doctor for those who feel connected." Respondents also tended to include immediate access and timely responses in their de nition of exceptional care, whether that was a patient's ability to obtain an appointment immediately, or engage in brief conversations with PCPs by phone or email: "I think it's about hearing people and really listening, engaging them, letting them know that you're there, even when they're not in the o ce. So they can run something by me either by email or call my o ce. I guess for me, exceptional care has been really more about the relationship." Some respondents included patient advocacy in their de nition of exceptional healthcare quality: "When I think of a doctor who is exceptional, I just think of someone who is an awesome patient advocate and making sure that the patients get what they need to the best of their ability." Finally, we asked PCPs to distinguish poor care from good care. Participants cited faults in evidencedbased care that resulted in missed diagnosis or inappropriate treatment: "I guess I would say low quality of care, if I meet people who have been with other physicians, I would say honestly, either under-or over-treatment…. So, I see people who are getting annual chest x-rays and cardiac stress tests for no reason.
So, I nd that as painful as people who are not receiving great care. So, I'll get people who have never been asked if they've had their screening colonoscopy and they're 60 or are overdue for screening mammograms." Another participant agreed, also adding that lack of care coordination contributes to low quality care: "So, I think low quality of care could be a couple of different things. One could beunfortunately, I have seen providers in practice that are not maybe following evidence-based recommendations or maybe don't have the skill level to address the problem they're trying to address but they're not kind of seeking out help. So, I guess I have seen misdiagnoses, mistreatments, so that's low quality of care de nitely. Then I think also the other piece of it could be that they tend to just really not follow up well with their own patients and not coordinate care back to the primary care doctor well. That can sometimes result in low quality of care too because things that they recommend never get carried out".

Discussion
Quality of healthcare is notoriously challenging to de ne, and varying stakeholders have often de ned it differently. In this qualitative investigation of PCP opinions about the nature of healthcare quality, participants revealed that they conceptualize healthcare quality as having two components: (1) evidencebased care that is safe and (2) patient-centered care. Moreover, evidence-based care that is safe is necessary but not su cient in the provision exceptional care. According to our panel of physicians, patient-centered care is an important component of the care experience, especially the establishment of long-standing patient-physician relationships built upon trust. Understanding how physicians conceptualize quality is important to optimizing engagement in quality improvement efforts. (20) The Institute of Medicine identi ed six domains of healthcare quality: safety, effectiveness, timeliness, patient-centeredness, equity, and e ciency. (25) Physicians in our study identi ed all of these sub-themes except equity, and it is di cult to know whether this sub-theme might have emerged with a larger sample size or a different sample of PCPs. The framework that emerged from our data (Fig. 1) (29) The OECD reviewed these models in a 2019 publication and proposed three core dimensions of quality that align with our components: effectiveness, safety (which we refer to collectively as "evidence-based care that is safe") and responsiveness ("patient-centeredness" Finally, our ndings may have additional implications for the quality measurement and pay-forperformance movements, since our participants communicated that PCPs believe long-standing, trustbased relationships are important in the provision of exceptional healthcare. Further research is warranted to verify our ndings across a larger population of physicians and determine whether rewarding performance on other aspects of patient-centered care, such as development of long-term patient-physician relationships founded on trust, could improve health outcomes for patients.

Limitations
Our study has several limitations. First, small-sample qualitative studies are useful for developing theories and identifying questions that should undergo further inquiry. Therefore, this study may be limited in its generalizability, though we attempted to minimize this risk by recruiting PCPs in different areas across the national and different practice environments. Second, because of our decision to use maximum variation sampling, the characteristics of our sample may not re ect those of the population of physicians nationwide. Third, our interviews included questions about the MIPS program, and recruiting physicians for a study involving questions about a quality measurement program may have biased responses about the nature of quality in healthcare. Finally, while we made attempts to ensure rigor and limit biases by employing re exivity and bracketing 33 during study design, recruitment, data collection, and analysis, there remains risk that investigator and coder biases may have limited the reliability and/or validity of our ndings.
Systems/Research/CAHPS/mips.html. Figure 1 exceptional quality of healthcare