Patient satisfaction with physician assistant/associate care: an international scoping review

Background As the role of the physician assistant/associate grows globally, one question is: what is the level of patient satisfaction with PAs? Driven by legislative enactments to improve access to care, the PA has emerged as a ready and able medical professional to address workforce shortages. The aim of this study was to review the literature on patient satisfaction of PAs. Objectives The basis for this review was to clarify working definitions, synthesize the evidence, and establish conceptual boundaries around the topic of patient satisfaction with PAs. The intent was to identify gaps in the literature and offer suggested undertakings for more clarification on the subject. Methods A scoping review was undertaken. Literature from 1968 to 2019 was searched and filtered for eligibility. Those that met criteria were categorized by date, method, geography, themes, and design. Results In total, there were 987 papers or reports that were identified through bibliography database searching. Additional articles found through snowball methodology-reviewing references (n = 11). Only English language articles emerged for analysis. From this effort, 25 articles surfaced from the filtering process for final inclusion. Most (72%) of the articles came from the United States of America, three from the United Kingdom, and one each from Ireland, the Netherlands, and New Zealand. Most articles were descriptive in nature. Some variations in methods emerged. Conclusion PAs are operational in 15 nations; their acceptance appears successful and satisfaction with their care largely indistinguishable from physicians. Findings from this analysis highlight one theory that when patient’s needs are met, satisfaction is high regardless of the medical provider. Areas for further research are identified.


Introduction
We undertook a systematic, scoping review of the literature on patient satisfaction with physician assistants and physician associates (PAs). The topic is germane to the utilization of PAs as they are employed in some fashion on four continents [1]. The rationale for this activity is their growing presence internationally. As more countries adopt and deploy PAs, the question as to what patients want emerges. A fundamental theory is that no health policy or marketing will overcome prevailing attitudes if patients are unwilling to accept care or are not satisfied with a PA's care [2]. Furthermore, patient satisfaction is an important component of the healthcare experience because patient agreement is considered a desired outcome of care [3].
The education of PAs across the globe is more similar than different, and outcomes of care diverge little from physicians even when the populations are identical [4][5][6]. The basis for this review was to clarify working definitions, synthesize the evidence, and establish conceptual boundaries around the topic. A second rationale was to summarize and disseminate the research findings, identify gaps in the literature, and offer suggested undertakings for more clarification on the subject.
The research question is: Are patients satisfied with care provided by PAs? The objective was to assess the impact on the patient experience with a PA in a medical care encounter.
A PA, for the purpose of this review, is defined as "a healthcare professional trained in medicine and works as part of a medical team in partnership with doctors to provide healthcare to patients" [7]. The PA movement is global, and their presence is noted across 15 countries [1]. However, the literature is diverse and remains to be synthesized.
The concept of "patient satisfaction" is a regularly used indicator of quality in marketing, as a patient retention measure, and a measure of healthcare quality [3]. Because satisfaction of an encounter can affect clinical outcomes, it can mean a great deal for a wide range of health providers. Where it occurs, it is noted both on inpatient hospital services as well as in medical clinics and physician's offices. Patient satisfaction is also a concept that is evolving. Observations in the twentieth century focused on whether the patient was "pleased to be cured" [8]. As early as 1971, when the development of the PA was still underway, Rousselot et al. called "public acceptance [of the PA] must be studied fully and evaluated fairly." [9]. A more contemporary view has emerged. This view is that the quality of the healthcare system as well as the health professional is needed when examining patient satisfaction [10]. Satisfied patients are more likely than unsatisfied ones to continue using healthcare services, maintaining their relationships with specific health care providers, and complying with care regimens [11].
If the contemporary PA movement began in the mid-1960s, a view spanning a half-century of patient satisfaction in regard to PA encounters is timely to understand what the shortcomings are and how experience can be improved. The Donabedian model of quality healthcare stresses that the interface of practitioner performance and patient acceptance is where "maximally effective or optimally effective care is sought." Further investigation explored whether individual or social preferences define the optimum encounter [3].

Method
Our analysis follows the scoping review methodology as outlined by Arksey and O'Malley and further refined by Peters et al. [12,13]. This review was registered at the University of York in 2017 (irss505@york.ac.uk).
An information specialist (MM) developed the strategy for MEDLINE, the primary database. After the method was peer reviewed by library colleagues, the strategy was translated into the other pre-selected databases. Database subject headings and keyword searching were utilized for increased sensitivity. Date limits were from 1968 to 2020. No methodological filters were used. Databases searched include MEDLINE (Ovid), Embase (embase.com), CINAHL Plus with Full Text (Ebscohost), PsycINFO (Ebscohost), CENTRAL (Cochrane Library), and Web of Science (Clarivate Analytics). See supplemental files for search strategies. Citation management, including removal of duplicates, was accomplished with EndNote (Clarivate Analytics). The study was considered exempt from our Institutional Review Board scrutiny.
The question "are patients satisfied with physician assistant/associates" ensured that a broad range of literature was included in this scoping review project. Comprehensive inclusion of the scale and scope of available literature was consistent with contemporary scoping review strategies. Exclusion criteria ensured that the PA alone was evaluated by the patient and not included with other health providers such as nurse practitioners (NPs) or midwives where each type of provider was not distinguishable. The methodology is cataloged in Table 1 and identifies inclusion and exclusion criteria.
Covidence (Covidence.org), an online systematic reviewing platform, was the software used to screen, review, and select studies. Two reviewers independently screened the title and abstract and reviewed the full text. Each was blind to the other's decision. A consensus strategy for inclusion was created a priori with a third party if one could not be reached. When 11 differences were found, these were Exclude if PAs were grouped together with another health professional such as an NP and could not be separately analyzed.

Study focus
Patients were asked or surveyed as to how they regarded their care or experience with a PA. The definition of a PA includes those formally trained as a PA.
Population and sample Patients that had an encounter with a PA. Other health professionals if they were merged together with PAs, e.g., NPs, nurses, and technicians.
resolved without an arbitrator. Data charting was performed with Excel (Microsoft). During the article selection, over 95% of articles reviewed were excluded when the criteria in Table 1 were applied. Examples excluded were poster abstracts or papers where specifics on the data-gathering strategies were not detailed. Another small but important percentage of articles reported on patient satisfaction with a PA and NP in the aggregate but did not separate out the two providers. We corresponded with five authors requesting more granular data, but none were forthcoming. After final eligibility filtering, 25 studies were included in this scoping review (Fig. 1).

Results
In total, there were 987 papers or reports that were identified through database searching. Additional records were identified using "snowball" methodology-reviewing references (n = 11). From this effort, 25 articles emerged from the filtering process for final inclusion. These were grouped by national origins for the first analysis ( Table 2).
Articles included in the final review were identified by country of origin to enable a comparative analysis of physician assistant/associates. The majority was American (n = 18); four were from the United Kingdom; and one each from Ireland, the Netherlands, and New Zealand. No patient satisfaction studies were reported on PAs in other countries, or on clinical associates of South Africa (Table 2). Only one study was a national survey of patient satisfaction of the elderly-and compared PAs, NPs, and doctors alongside each other [14]. No differences were found between the three types of providers, and at the same time, patient satisfaction with each type of provider was found to be consistently high.
The method of assessment was examined (Table 3). Eight studies reported anonymous surveys administered by mail and six used standardized and validated survey instruments (e.g., Meijer and Drennan) [15,16]. Older studies, many made at an early stage of PA development, were done by phone or the patient was interviewed in person (e.g., Litman and Farmer) [17,18]. Anonymous surveys began appearing in the late 1980s.
We analyzed the setting of the patient being seen by a PA. While a quarter were in some type of physician's office (N = 6) or clinic (N = 4), six were in specialty clinics such as diabetes or orthopedics ( Table 4). Two of the early studies took place in a rural America clinic site. Because many of the early studies investigating PAs were efforts to Fig. 1 Overview or schematic of the review process enhance healthcare delivery as such PAs were viewed more as a complement to physicians than replacements.
A summary of the included studies is listed in Table 5. All studies were undertaken between 1970 and 2017. Spanning 23 studies, the number of patients interviewed about PAs was at least 8062 (Mean = 350; Max = 1159; Min = 20). However, two studies did not identify the total number of patients surveyed. The total number of PAs assessed spanning all studies was more than 2234, although the accumulative number of PAs is not known because in eight studies the exact number of PAs assessed was not mentioned.
The majority of studies (N = 23) was descriptive in reporting. In two articles, a mixed-methods approach was incorporated. In most studies, the research question probed the patient's satisfaction with the PA. In the remainder, it was explored as a secondary outcome. Older articles focused more on patient acceptance instead of satisfaction, likely due to the emerging concept of the PA as a new profession. Early studies describing patient acceptance were included in this review because they were interpreted by the authors as satisfaction and willingness to be seen (e.g., Strunk 1973).
Overall, patient satisfaction compares favorably with physicians in this scoping review. This finding was consistent through all 25 studies and, where a comparison was made, patient satisfaction ranged from 94 to 100% regardless of the instrument used. Exit interviews (patients interviewed in person) did not differ significantly from anonymous paper interviews. No assessment via electronic communication was done in the 25 studies we assessed.

Discussion
The majority of studies on patient satisfaction came from the United States (18/25), reflecting over 50 years of experience with PAs. It should be noted that analysis across five countries finds remarkably consistent results-patients were generally satisfied with PAs regardless how the encounter or experience was assessed.
Early development of US PAs (including the MEDEX model and the Child Health Associate model) was aided by federal and foundation grants intended to examine if these new health professionals were going to meet the needs of society. Patient awareness and acceptance with PAs was a cornerstone of health policy research. In the US, the acceptance with PAs by patients in theory or in practice consistently grew as reflected in patient satisfaction studies. This concept mirrors our findings in the articles we analyzed, as early studies focused on acceptance, which evolved into a focus on satisfaction in more recent studies. Similar findings, using more refined survey instruments, found that the contemporary PA in the UK, IR, NL, or NZ, was widely accepted. More specifically, patients were as satisfied with PAs as the doctor [15,35]. None of the studies included in this review found patient dissatisfaction beyond single digit percentage.
Techniques in how patients were assessed for their satisfaction varied widely. This included exit interviews (as the patient left the clinic), chart reviews, telephone surveys, anonymous paper surveys, and household surveys in rural areas. Only a few of the surveys were standardized and validated (e.g., Medicare survey of the elderly, European Union Consumer Quality Index), and only one involved a pre-visit survey followed by a postvisit survey. A few of the studies matched the PA and the patient for age, gender, race, ethnicity, or other characteristics. The US Medicare study was unique as it was national in scope. Some studies looked exclusively at patient satisfaction with care provided by a PA, though some made direct comparisons with other health professionals. Of the 25 studies analyzed, none showed that patient satisfaction with the PA was significantly less than the doctor. In the one study where PAs and NPs    Patients were randomly assigned to "health assistant" functioning as a PA. In total 441 patient visits were studied for protocol adherence by the HA. Of 286 patients assigned, 6% declined to be routed into the protocol system. Of 53 patients who were told they did not have to see the doctor, only 1 asked to be seen; only 2.2% of total visits did patients seek medical attention in between visits.   were compared, the satisfaction results were found comparable. There are a number of studies on patient satisfaction that include PAs folded in with NPs. Such studies are important in their own right as they inform consumers and policy makers but were excluded from this study because they failed to discriminate between the two providers where some differences might exist. In addition to differing assessment techniques, there were other reasons why direct comparisons between studies were difficult. These factors included variability in study types, size of studies, patient populations, settings, specialties, assessing individual providers vs. teams, utilization/role with PAs, and scope of practice. Furthermore, some studies asked about satisfaction directly whereas others made inferences. Due to the wide date ranges between studies, PAs in earlier studies were less recognized as a profession than in later reports. Early demonstration projects viewed PAs more as doctors' assistants while contemporary PAs are seen as medical professionals similar to doctors. This is reflected in the studies done by Litman [18], Nelson [20], and Hla [27]. The European studies assumed the PA was filling an otherwise deemed physician role [15,35]. Spanning the 50+ years of PA utilization, the adherence to protocols to treat patients has been replaced by "best practice" or community standards of care. Notably, in this study, it is difficult to directly compare several studies as the scope of the PA in early studies was different than in later studies. In some earlier studies, patients were either satisfied or accepted care by a PA but in a very limited scope of practice. Implications were made that they may not have accepted care if a PA had increased responsibility in which they did not feel a PA had adequate training. This is also due, in part, to the lack of knowledge about the PA profession at the time the studies were conducted. In later studies, as the profession became more well known, satisfaction encompassed the current scope of PAs.
Measuring patients' perception of care is important for a number of reasons. The first is that no amount of observation science could overcome a negative perception if patients refused to accept the role of the PA. In addition, satisfaction correlates with compliance, health outcomes, and patients returning to see the same provider. Patient satisfaction surveys are essentially assessing service delivery based on the patients' viewpoints of the organization as well as the provider of that service. Outcomes are improved when the patient and the medical clinician correlate with patient satisfaction, quality of life, compliance with instructions, and most importantly, health outcomes [40,41]. Patient satisfaction also correlates with continuity of care and likelihood of returning to the provider for longitudinal care [42,43].

Recommendations
While this review found a significant body of literature on the subject of patient satisfaction with PAs, 72% was produced in the United States. As more countries expand their observations of PA behavior, the number of patient satisfaction studies is expected to grow. We believe that all countries should have some fundamental understanding how well and to what extent their citizens need and accept their providers of medical care.
In undertaking this project, a wide assortment of studies was examined (although not necessarily included). From this body of patient satisfaction literature, we identified variables we believe need to be introduced into the research model when testing the hypothesis that patients are largely accepting and, in general, satisfied with PAs in terms of care, experience, and outcomes. These are: Gender-A match for gender between provider and patient should assess and compare where there is a difference. In the case of children, the gender of the parent or guardian should be known as well as the provider.
Race and ethnicity-A match for race and ethnicity needs to assess if the differences found in patients and physicians change with patients and PAs. Age-The four broad age groups needed for examination are children, 18-40, 40-65, and > 65. Younger versus older providers are broad areas that need to be compared in provider-patient encounters. Medical specialty-PAs in hospitals, clinics (urban and rural), physicians' offices, emergency rooms, urgent care clinics, and orthopedic clinics were assessed. Because the majority of PAs in the US, UK, NL, and Canada are not in primary care settings, the roles, specialties, and settings not assessed compile a large list. Country-All countries should undertake patient satisfaction studies and results compared across borders to see where areas of improvement can be made. This is not only for PAs but also for the wide range of health professionals.

Limitations
The greatest limitation to this study is the criteria we applied: that the included study needed to be peer reviewed and published. By this gauge, a large number of studies on patient satisfaction were not included. This is due to the observation that many health organizations undertake patient satisfaction studies for enrollment and marketing concerns but do not publish their results or even make the results publicly known. We were aware of these studies, some by marketing companies for commercial purposes, but including them would be outside the scrutiny of scientific study and scoping review criteria.

Conclusion
The contemporary physician assistant/associate emerged in the 1960s and now occupies a role across a wide set of societies. We asked the question: Are patients satisfied with care provided by a PA? Using the scoping review format, almost 1000 mentions of some aspect of PAs and patient satisfaction were screened; 25 met criteria for inclusion. Of those analyzed, 18 were US studies and three were from the UK. The settings ranged widely from small rural clinics to large urban hospitals. In almost all studies comparing PA care to physicians, the patients made little if any distinction between the two. In this scoping review, it appears that patients are satisfied with PA-led care. The next phase of patient satisfaction research should compare provider and patient race, age, and include a diverse type of setting and medical specialty.