The physical examination: a survey of patient preferences and expectations during primary care visits

Abstract

Background

Little is known about patient preference regarding the physical exam in non-urgent primary care settings.

Objective

To determine the differences between a patient’s expectations of the physical exam and the actual components of the physical examination performed during a non-urgent visit.

Design

452 surveys administered in the waiting room of a VA primary care clinic in West Haven, CT.

Key Results

The response rate was 91.6% (n = 414). For 15 of 16 maneuvers on the survey, more respondents believed a reasonable provider should conduct it than received it at their annual physical exam; for 7 of them (breast, axillary, rectal, pelvic, total body skin exam, electrocardiogram, and stress test), over twice as many respondents believed they should be done than received them. There was an association between a patient’s perception of their primary care provider and the number of maneuvers recalled at their annual exam (P <0.001), and a gap in the number of maneuvers expected from a reasonable provider by non-white and white patients (P <0.001).

Limitations

Convenience sample, response bias (healthy patients are more likely to respond) and recall bias.

Conclusion

Patient perception of their primary care provider is strongly associated with the number of maneuvers recalled during an annual physical. Furthermore, the number of maneuvers expected by a patient is influenced by race, with non-white patients desiring more. This suggests the need for further research on the role of race in expectations of healthcare providers.

Introduction

Primary care visits are among the most common reasons for adults to visit a physician. In 2008, Americans made 462 million visits to primary care providers (PCPs) (accounting for approximately half of all visits), a number projected to increase to 565 million in 2025.¹ In addition, in this age of patient-centered care, greater emphasis is being placed on framing healthcare within the context of what is important to the patient and on improving the therapeutic alliance between patient and healthcare provider.

An understanding of patients’ expectations for preventative exams is thus critical to the delivery of high-quality, compassionate care, especially in an era where the value of the periodic health examination is being questioned.^2–5 PCPs use expert opinion, professional guidelines, like those provided by the United States Preventative Task Force (USPTF), and personal experience in tailoring the physical exam, recommending screening tests, and suggesting preventative care. However, patients tend to overestimate the value of, and underestimate the harms of, screening tests from PCPs.^6–10 Interestingly, a few studies have shown that patient satisfaction is related to the number of interventions done, suggesting that patients view testing as a sign of competence and/or compassion on the part of the practitioner.^11,12 These previous studies are often limited, however, by focusing on a handful of costly screening tests, such as mammograms or colonoscopies.^2,6,7

Even with rapidly advancing technology, a well-conducted physical exam is still vital to early and correct diagnosis, judicious use of expensive tests, and building a therapeutic alliance.^13–17 To our knowledge, there is little or no literature on patient expectations for specific components of the physical exam, such as fundoscopy, palpation of lymph nodes, etc., which are low-cost maneuvers that may improve patient satisfaction with few adverse effects. Furthermore, there exists very little literature on the influence of race, education, gender, or other demographic variables on patients’ willingness and expectation to undergo specific aspects of the physical exam.

In this study, we designed a survey instrument to determine what components of the physical exam a respondent recalled undergoing at their last non-urgent primary care visit, what components they believed that a reasonable provider should perform, and what components they would be willing to undergo. We also collected data about patients’ age, race, gender, education level, perception of their PCP, and perception of their own health to determine if these variables correlate to which physical exam maneuvers were conducted at the last visit, which maneuvers they believe should be performed, and which aspects the patients would be willing to undergo.

Methods

Data Collection

Surveys were administered as paper-and-pencil tests to a convenience sample of patients awaiting primary care appointments at the West Haven Veterans Affairs (VA) Primary Care Clinic waiting room from June to August 2015. Only patients scheduled for non-urgent visits, defined as visits scheduled for chronic disease management or preventative care (rather than an acute condition), were included. For most analyses, patient demographic data (age, sex, race, education level), as well as the respondents’ perception of their own health and perception of their PCP were treated as independent variables. While race and gender are not entirely standardized medical terms, we wished to ascertain the relationships between patient’s self-reported race and gender and their preferences.

Patients were asked which of 16 physical exam maneuvers were done at the last visit, which should have been done by a reasonable provider, and which they would have allowed to occur. Maneuvers were described in layman’s terms (for example, otoscopic exams as “look into my ears”). These 48 questions about exam maneuvers (3 each for 16 maneuvers) were treated as dependent variables. Male respondents were not asked about pelvic exams.

Statistical Analysis

All analysis was done in R (v3.4.1, R Foundation, Vienna, Austria) on Windows 10. For all regression analyses, independent variables were dichotomized (Age ≥65 or <65, Male vs Female, White vs Non-white, Perception of Health ≥ or < than “neither good nor poor”, Perception of PCP ≥ or < than “neither good nor poor”, Education > or ≤ than High School) to improve convergence of statistical models and to facilitate binary comparisons, unless otherwise specified.

To relate counts of survey responses to each independent variable in Table 2, we used a multivariate quasipoisson regression (to avoid issues with overdispersion) on the counts using the glm command in base R. Pelvic exams were excluded, because only women were asked about them.

Table 2.

Correlation of patient characteristics with reported number of completed and expected components of physical exam.

	Age ≥65	Age <65	Male	Female	White	Non-White	Health ≥ “neither good nor poor”	Health < “neither good nor poor”	Perception of PCP ≥ “neither good nor poor”	Perception of PCP < “neither good nor poor”	Education > High School	Education ≤ High School
Number of maneuvers expected	12.22	12.41	12.31	12.32	12.01	13.28	12.3	12.42	12.31	12.22	12.19	12.48
P-Value	0.44		0.82		5.05E-06***		0.58		0.4		0.4
Number of maneuvers performed at last visit	7.92	8.16	8.04	7.94	8.17	7.58	8.1	7.27	8.2	4.44	8.23	7.75
P-Value	0.18		0.69		0.09		0.25		2.16E-07***		0.43

To determine which independent variables influence the responses to the survey, 48 multivariate logistic regressions were performed in Figure 2, one for each question on the survey (with 1/0 coding for “yes” and “no” responses). To account for multiple testing, P-values were deflated by the Bonferroni correction (n=48). The Hosmer-Lemeshow test was used to assess model fit, and gave p > 0.01 (not adjusted for multiple correction) in all cases in which the statistic existed.

Figure 2: Logistic regressions of each survey item.

Each question in the 48-item survey was regressed against age, gender, perception of health, perception of primary care provider, education level, and race (all dichotomized per “Methods”) using a logistic model. Significance levels (corrected using the Bonferroni method with n = 48) are marked on the color scale bar.

Results

Of 452 patients approached, a total of 417 patients agreed to complete the survey, reflecting a response rate of 92.5%. Three surveys were incompletely filled out and were excluded from the final dataset, resulting in a total of 414 included responses.

The age of respondents ranged from 24 to 94 years, with a mean of 62.20, median of 65.5 and standard deviation of 15.27. The majority of respondents were male (91.79%) and white (76.33%); only two respondents were Native American, representing 0.48% of the sample (Figure 1). Only 15 respondents did not complete high school, while 101 either completed college or had some additional schooling after college. 89.86% of patients have either a “good” or “excellent” perception of their PCP, suggesting that the vast majority of patients are pleased with the primary care received at the VA – or at the very least, with the provider they have been assigned. Furthermore, the majority of patients believe themselves to be in at least “good” health, while over 90% believe their health to be at least “neither good nor poor.”

Figure 1: Summary statistics of the study population.

Participants’ age, sex, race, perception of their overall health, perception of their primary care provider, and education level are summarized above.

Respondents were willing to undergo almost any component of the exam [Insert Table 1]. The survey also found that almost all patients recalled undergoing a heart exam, lung exam, and blood pressure measurement at their last visit (95.7%, 93.7%, and 100% respectively), while very few had received either a stress test or breast exam (9.7% and 12.6% respectively). The completion rate for blood pressure measurement is likely because it is standard procedure for a nurse to take vitals during the clinic intake process.

Table 1.

Summary of survey responses.

Maneuver	Was Examined or Performed (“Done”)	Should Have Been Done by a Reasonable Healthcare Provider (“Should Do”)	What I Would Have Allowed to Be Examined (“Would Allow”)
Blood Pressure	414 (100%)	412 (99%)	414 (100%)
Eyes	165 (39%)	308 (74%)	412 (99%)
Carotid	216 (52%)	371 (89%)	414 (100%)
Ears	305 (73%)	391 (94%)	414 (100%)
Mouth	312 (75%)	384 (92%)	412 (99%)
Neck	291 (70%)	390 (94%)	414 (100%)
Breast	52 (12%)	250 (60%)	394 (95%)
Lungs	388 (93%)	409 (98%)	414 (100%)
Axilla	153 (36%)	354 (85%)	412 (99%)
Heart	396 (95%)	412 (99%)	414 (100%)
Abdomen	337 (81%)	404 (97%)	414 (100%)
Rectal	69 (16%)	234 (56%)	378 (91%)
Pelvic	6 (17%)	21 (61%)	30 (88%)
TBSE	79 (19%)	302 (72%)	410 (99%)
EKG	109 (26%)	260 (62%)	412 (99%)
Stress Test	40 (9%)	215 (51%)	392 (94%)

We also found that for every maneuver, there is a gap between the percentage of patients that believed a competent provider should perform it and those that recalled it at their last visit. In particular, there is a greater than 40% gap between expectation and receipt of total body skin exams (TBSEs) (54%), axillary exams (49%), breast exams (48%), pelvic exams (44%), stress tests (42%), and rectal exams (40%).

The total number of maneuvers a respondent received at their last primary care visit and thought should be done at their next primary care visit (excluding pelvic exams) was recorded and compared by dependent variables in Table 2. Each dependent variable listed in Table 1 was dichotomized. Means were compared by fitting the dichotomous variable to a quasipoisson regression in which number of maneuvers was treated as the response variable and age (≥65 or <65), gender (male vs female), race (white vs nonwhite), perception of health (at least “neither good nor poor” vs lower), perception of PCP (at least “neither good nor poor” vs lower), and education (completed at least some college vs not) were treated as dichotomous dependent variables. Patients who perceived their PCP as at least “neither good nor poor” received 8.20 maneuvers on average, while those with a lower opinion of their PCP reported 4.44 (P <0.001 by quasipoisson).

The survey also showed that the majority of respondents believed that a reasonable practitioner should perform most of the maneuvers listed at a non-urgent primary care visit. The least expected was the stress test (51.9%), which, while more properly considered a diagnostic test rather than a physical exam maneuver, was included because we wished to ascertain whether stress testing was an expected portion of a patient’s yearly physical. The most expected was blood pressure screening (99.5%). Non-white patients, on average, expected 13.28 of 15 maneuvers, while white patients expected only 12.01 (P <0.001 by quasipoisson).

We next sought to determine the independent variables influencing responses to each of the 48 survey items about maneuvers. In order to do this, responses to each item were treated as dependent variables in a logistic regression. The results of this analysis, revealing several significant correlations even when corrected for multiple testing, are shown in Figure 2. As expected, we found that women are more likely to have received a breast exam at their last visit than men. Consistent with the quasipoisson regression in Table 2, we find that a patient’s perception of their PCP is strongly positively correlated to if they recalled having a heart, neck, mouth, or ear exam done. Furthermore, a patient’s expectation for a neck exam at the next visit is positively correlated with perception of PCP quality.

Finally, the analysis revealed education is negatively correlated with the belief that a competent PCP should conduct a stress test, and that non-white patients were more likely to expect a stress test than white patients. Given our previous finding that non-white patients expected, on average, 1.27 more maneuvers than their white counterparts, we sought to determine if the expectation of stress testing explains most of this difference. To that end, we repeated our quasipossion regression and excluded stress tests from the total count maneuvers the respondents believed should be conducted. Similar to the previous analysis, race was the only significant independent variable, with white patients expecting an average of 11.55 of the considered maneuvers, while non-white patients expected 12.57 (P <0.001 by quasipoisson), for a difference of 1.02.

Discussion

We conducted a comprehensive study of patient preferences at a VA primary care clinic in Connecticut. Patients were generally willing to undergo a variety of physical exam maneuvers, even those that are invasive (e.g. digital rectal exam) or involved (e.g. cardiac stress test). Furthermore, patients, in general, believed more maneuvers should be done than are generally recommended as high-value care by bodies like the USPSTF.^12,18 This is not unusual, as patients tend to overestimate the effectiveness of preventative interventions and commonly request more tests and examinations than necessary.^11,12,19,20

Explaining the need for judicious and effective use of expensive testing (e.g. colonoscopy, blood tests) is vital to decreasing medical costs and ensuring efficient and quality care. The cost per unique patient in the Connecticut VA system is within a standard deviation of the mean in the national VA system. While we do not believe a routine stress test or several other tests in asymptomatic individuals are warranted, the knowledge of patient expectations may allow a provider to discuss why certain tests or examination maneuvers are not done may improve the patient experience.

Our study reveals that patient’s perception of the quality of their PCP is correlated with whether they could recall undergoing a handful of inexpensive maneuvers (namely, cardiac auscultation, ear exam, mouth exam, and cervical lymph node exam). Furthermore, patients with a high opinion of their PCP recalled undergoing an average of 3.76 more components of the physical at their previous visit than patients with lower. This may be due to a recall bias–a patient who already has a favorable view of their PCP may state that they had a more thorough physical exam than a respondent with an unfavorable view. However, it may also be because practitioners who perform a lengthier physical exam are seen as more thorough, compassionate, and/or competent.

Our data also shows that for several maneuvers (particularly TBSE, axilla, breast, and pelvic exams), there is a wide gap between the number of respondents who recalled it being performed and those that believe a reasonable provider would conduct it at a non-urgent PCP visit. In some cases, this may be due to recall bias—a PCP may have conducted a quick skin examination while completing the other parts of the exam, for example. However, because many of these procedures are invasive (especially breast and pelvic exams), recall bias is unlikely.

These findings suggest the possibility that a simple way of strengthening the therapeutic alliance is to perform a few additional components of the physical exam at every visit. Many practitioners conduct a focused exam based on symptoms, omitting those parts of the exam believed to be low-yield. For example, if a patient presents without organ-specific complaints or generalized symptoms like fevers or weight loss, a provider may feel justified in omitting several components of the physical exam. However, our data suggests that even in the absence of relevant symptoms, additional maneuvers – even maneuvers as simple as examining the patients mouth, ears, and axilla – may improve the doctor-patient relationship.

To our knowledge, our study is one of the first to determine the effect of race, gender, and education level on patient expectation for a more thorough physical exam. We found that non-white patients sought more maneuvers than their white counterparts, on average an additional 1.27 maneuvers. Further analysis showed that race did not have a significant effect on the belief that any particular maneuver should be done, except for cardiac stress testing; eliminating cardiac stress testing from the quasipoisson models did not change the conclusion that non-white respondents expected more maneuvers than white respondents. This suggests, robustly, that even if race does not influence the expectation of any one maneuver, it does influence the total amount of maneuvers expected.

There are a few potential explanations for the finding that non-white patients expected more maneuvers than white patients. One possibility is that minority patients feel underserved by the VA or by the healthcare system, in general. Previous studies have focused on the effect of race on patient satisfaction (including studies by Zickmund et al. on VA clinics), with mixed results.^21–25 We found that there was no significant difference in patients’ perception of their PCP between minority respondents and white respondents (data not shown), and minority respondents reported undergoing a similar number of physical exam maneuvers as their white counterparts (data not shown). It may also be the case that minority patients are more cautious about their health, leading them to believe a more thorough exam should be conducted. Further data gathering, possibly with more in-depth interviews, would be able to assess the reasons for this difference.

There are a few potential limitations of our study; in particular, the conclusions drawn from a VA patient population may not be generalizable either globally or to other patient populations. First, the average age of our patient population is significantly higher than the age of the VA population as a whole. This may suggest that either this VA’s patient population skews older than the national average, that primary care patients at the VA skew older than the general patient population, or some combination of the two. Additionally, the VA patient population is disproportionately white and male, compared to the general population – the preferences of this population may not necessarily reflect the American public as a whole. Finally, there are other explanatory variables (like socioeconomic status, past medical history, and percent service connection) that may provide additional insight that our study did not include.

It is also notable that the average perception of a respondent’s PCP in our dataset is quite high, as is the respondents’ self-perception of their health. This may reflect a response bias—patients willing to take the survey may have a more positive view of their own health and of their PCP, and patients regularly attending primary care visits may believe themselves to be in better-than-average health and may also have higher-than-average opinions of their PCP. However, our study is notable for our high response rate, diminishing the role of possible response bias. This response rate may be specific to this study population – a VA population may be more likely to fill out a survey than the average patient population. It may also be the case that the VA patient population is more willing to offer their opinions of their care.

Additionally, our study could have been influenced by recall bias, and what patients reported as being done at their last examination may not have actually been performed. Patients who have a favorable view of their physician might have assumed a component of the exam was performed even if it wasn’t, or simply recall the exam in more detail.

Understanding the relationship of race, gender, education level, and other demographic variables to preferences and expectations of the physical exam is vital to shaping compassionate primary care. In this study, we have demonstrated that performing a handful of exam maneuvers is correlated positively with a patient’s perception of their PCP. We have also demonstrated a correlation between a respondent’s race and the number of maneuvers they expected from their PCP at the next visit, with non-white patients desiring more than white patients. However, we found no correlation between the expectation for any one maneuver (except the cardiac stress test) and race.

Our results underscore the need for further studies on the relationship between a patient’s race and their expectations from primary care, which could potentially be used to tailor patient education. Furthermore, our data is vital to understanding the relationship between the comprehensiveness of a physical exam, a patient’s perception of their PCP, and the strength of the therapeutic alliance. This data suggestions that physicians may benefit from performing a few additional components of the physical exam in order to improve patient’s perceptions of their healthcare experience.

What is already known about this topic?

Patients tend to view practitioners that order expensive screening tests as more competent, and a physical examination is vital to the doctor-patient relationship.

What does this article add?

Patient perception of their PCP is strongly associated with the number of maneuvers recalled during an annual physical. Furthermore, the number of maneuvers expected by a patient is influenced by race, with non-white patients desiring more.