- Email: [email protected]
To Sit or Not to Sit?
THE PRACTICE OF EMERGENCY MEDICINE/BRIEF RESEARCH REPORT
To Sit or Not to Sit? Rebecca L. Johnson, MD Annie T. Sadosty, MD Amy L. Weaver, MS Deepi G. Goyal, MD
From the Mayo Clinic College of Medicine (Johnson), Department of Emergency Medicine (Sadosty, Goyal), and Division of Biostatistics (Weaver), Mayo Clinic, Rochester, MN.
Study objective: We prospectively examine whether provider posture (seated versus standing) influences patient and provider estimates of time spent at the bedside relative to actual time and patient perceptions of the provider-patient interaction. Methods: A convenience sample of consenting adult patients presenting to an academic tertiary care emergency department between September 7, 2005, and September 25, 2005, were eligible for inclusion in this randomized, controlled trial. Providers (emergency medicine attending physicians, residents, physician assistants, and medical students) were randomly assigned to sit or stand during the initial encounter, after which, participants completed questionnaires about their perceptions of provider-patient interactions and time spent therein. Actual encounter length was measured. Data were analyzed to determine whether patient and provider perception differences existed, using a multilevel regression model that was adjusted for patient-level and provider-level covariates. Results: Two hundred twenty-four consenting patients met inclusion criteria (239 approached; 15 excluded). Data from 36 providers were collected. The mean length of encounters in both study arms was 8.6 minutes (SD 4.8; range 1.5 to 34.1). Patients involved in seated interactions overestimated time providers spent performing initial encounters by an average of 1.3 minutes (SD 4.3 minutes), whereas patients involved in the standing interactions underestimated time by an average of 0.6 minutes (SD 4.3 minutes) (P⫽.001). Conversely, providers overestimated time spent with patients in both study arms (P⫽.85; mean [SD] 0.5 [3.6] versus 0.3 [3.2] minutes). Patient perceptions of the quality of patient-provider interactions were not affected by provider posture. Conclusion: Although provider posture during the initial interaction affects patient perceptions of time spent at the bedside, it does not influence patient perception of the provider’s bedside manner, sense of caring, or understanding of the patient’s problem. [Ann Emerg Med. 2008;51:188-193.] 0196-0644/$-see front matter Copyright © 2008 by the American College of Emergency Physicians. doi:10.1016/j.annemergmed.2007.04.024
SEE EDITORIAL, P. 194. INTRODUCTION Background Rapport-building is essential to the physician-patient relationship.1-3 Classic textbooks dedicate chapters to physicianpatient interactions and medical schools dedicate courses to the same.2,3 The classic text DeGowin’s Diagnostic Examination2 posits, “To obtain the patient’s confidence and rapport, present yourself as unhurried, interested, and sympathetic. Sit at eye level without a desk between you.” A textbook on the medical interview states, “. . . if you stand by the door rather than sit by the bed, patients may assume you are in a hurry . . . .”3 Indeed, it is often promulgated throughout medical education that 188 Annals of Emergency Medicine
patients perceive that providers who sit during the encounter are at the bedside longer than providers who stand for the same amount of time, yet we could find no study in the medical literature that examined the effects of provider posture on provider-actual, patient-perceived, and providerperceived time. Importance Patient perceptions of time spent with the provider have been shown to be determinants of patient satisfaction.4-6 Patient satisfaction, in turn, has been linked with increased referrals, improved compliance, decreased cost, enhanced clinical outcomes, and decreased litigation.4,6-8 Spending more time at the bedside is not always possible. On busy days, Volume , . : February
Johnson et al
To Sit or Not to Sit Editor’s Capsule Summary
What is already known on this topic It is commonly taught that physicians should sit rather than stand during the encounter to impart a sense of unhurried caring. This teaching is unsubstantiated. What question this study addressed Does provider posture—sitting or standing— during the initial emergency department evaluation affect the patient’s perception of the duration of the encounter or the provider’s bedside manner? What this study adds to our knowledge In this 224-patient randomized trial, patients overestimated the time spent by seated providers and underestimated the time spent by standing providers, both by small amounts. How this might change clinical practice Sitting may modestly improve the patient’s perception of physician care. Given the low cost of this simple act, it may be prudent to provide a place for the physician to sit in each patient care area.
altering patient perceptions about the amount of time spent may be the only alternative. Goals of This Investigation In this article, we examine provider posture (seated versus standing) during the initial emergency department (ED) encounter. We hypothesize that having the provider sit during the initial encounter will increase patient and provider estimates of the time spent at the bedside relative to actual time and improve patient perceptions of the provider-patient interaction.
MATERIALS AND METHODS Study Design and Setting We conducted this prospective nonblinded randomized study in an academic tertiary-care ED with 60,000 annual adult visits. Selection of Participants Adult patients presenting between September 7, 2005, and September 25, 2005, were eligible for inclusion in the study. Patients were enrolled by a sole study investigator (R.L.J.) by convenience sampling, with data collection generally between 8 AM and midnight, on weekdays and weekends. Exclusion criteria included patients ineligible for inclusion according to nursing assessment if they were younger than 18 years, were non-English speakers, and had a health status and cognitive function that affected their ability to communicate answers to R.L.J. Patients triaged to the critical resuscitation area were excluded according to acuity potential. Volume , . : February
Providers present during the study period included 23 emergency medicine attending physicians, 9 emergency medicine residents, 7 off-service residents, 4 physician assistants, and 6 rotating medical students. Encounters by the other authors of the study were not included. The study was designed to enroll 200 participants (100 in each study arm), given the time constraints of R.L.J. The Mayo Clinic institutional review board approved this study, and all participants provided written informed consent. Methods of Measurement and Data Collection and Processing R.L.J. approached patients deemed eligible by nursing, after they were assigned to a room, to obtain informed consent. R.L.J. confirmed eligibility and excluded patients who failed to answer a series of orientation questions relating to current month, current year, and current president of the United States correctly. Blinded to the nature of the study, eligible patients were enrolled and demographic data were collected. Before the beginning of the study period a block randomization schedule was created. Using this block randomization schedule, R.L.J. randomized the provider to sit or stand, notified the provider of their assigned posture just before entrance into the examination room, and accompanied the provider into the room. At a teaching hospital, patients are often treated by multiple providers during their visit. The first provider present for the initial encounter was selected for randomization. Those providers randomized to sit during the initial encounter (defined as introduction, history taking, physical examination, and rendering of initial impressions and plan) were instructed to perform as much of the initial encounter as possible in the seated position. Because portions of the physical examination require providers to stand, posture during the physical examination was left to the seated provider’s discretion. Providers randomized to stand were instructed to stand during the entirety of the initial encounter. R.L.J. used 3 stopwatches to capture time data. The stopwatches were concealed from provider and patient view within R.L.J.’s white-coat pocket. The entire length of the initial encounter was timed from provider entrance into the room until the provider exited the room. In addition to the running initial encounter time, R.L.J. clocked the amount of time spent seated with a second concealed stopwatch for both sit-randomized and stand-randomized providers. A third concealed stopwatch was present to collect any time spent that could interrupt the running initial encounter time (eg, answering pager, telephone interaction, requirement for room exit). Patient and provider demographic data, actual encounter time data (total encounter time minus interruption time), and time spent sitting were recorded. We developed a 5-item questionnaire with a Likert scale for this study (Appendix E1, available online at http://www.annemergmed.com). Immediately after initial encounter services were delivered and strictly following the Annals of Emergency Medicine 189
To Sit or Not to Sit language of the scripted questionnaire, R.L.J. administered the questionnaire to the patient who was blinded to time. On exit from the room, the provider was asked to complete a brief questionnaire (Appendix E2, available online at http:// www.annemergmed.com). Two sham questions, designed to address provider discomforts, were inserted into the questionnaire to distract the providers from the true nature of the study. Sham data were not reported for the purposes of this study. After administrating the questionnaires, R.L.J. would proceed to the next patient identified by nursing to be eligible for enrollment. Primary Data Analysis Data were analyzed according to the intent-to-treat principle. The responses to the questionnaire items, actual initial encounter time, estimated time spent together in the initial encounter, and the difference in the actual and estimated times were summarized using standard descriptive statistics. The outcomes variables of interest were the difference in the actual and estimated time by the patient and difference in the actual and estimated time by the provider. Multilevel or randomeffects models were fit with a generalized mixed linear randomeffects model with the SAS (SAS Institute, Inc., Cary, NC) MIXED procedure to evaluate the relationship between provider posture (sitting or standing) and each of the 2 outcome variables. The model allowed for the adjustment of both patient-level (age and sex) and provider-level (sex and provider type) covariates. All calculated P values were 2-sided, and P values less than .05 were considered statistically significant.
RESULTS R.L.J. approached 239 patients for inclusion in the study, of whom 15 were not enrolled. Nursing excluded 1 patient because of severe sore throat pain, precluding ability to communicate; R.L.J. excluded 4 patients for failing to correctly answer the orientation questions, and 10 did not provide informed consent. Of the 10 denying consent, 5 (50%) were women, with a mean age of 55.9 years (SD 22.7 years), median age of 55 years, over a range of 22 to 85 years. Results herein are based on the 224 patients and 36 providers who gave informed consent to the study. Patient and provider characteristics are summarized in Table 1. The patient and provider perceptions and estimates of time spent are listed in Table 2. On average, each encounter lasted 8.6 minutes, and the average length of the encounters was similar in the 2 study arms. None of the providers randomized to standing (N⫽109) sat down during the encounter. The providers randomized to sitting (N⫽115) sat an average of 67.4% of the total time. The patients involved in the sit-randomized interaction tended to report a longer duration for the amount of time the provider spent with them compared with the patients involved in the stand-randomized interaction (mean 9.6 versus 8.2 190 Annals of Emergency Medicine
Johnson et al Table 1. Summary of patient and provider characteristics, by study arm. Characteristics Patient sex, female, No. (%) Patient age, y Mean (SD) Range Patients treated by a female provider, No. (%) Patients treated by each type of provider, No. (%) Emergency medicine faculty Emergency medicine resident Resident off Physician assistant Medical student No. of participating providers (median no. of patients treated by each provider)* Emergency medicine faculty Emergency medicine resident Resident off Physician assistant Medical student
Sit (Nⴝ115) 72 (62.6) 48.7 (19.9) (18.0–90.0) 73 (63.5)
11 (9.6) 23 (20.0) 41 (35.6) 29 (25.2) 11 (9.6)
4 of 6 (2) 5 of 7 (4) 10 of 11 (2.5) 5 of 6 (4) 6 of 6 (1.5)
Stand (Nⴝ109) 55 (50.5) 48.0 (19.9) (18.0–94.0) 62 (56.9)
8 (7.3) 22 (20.2) 36 (33.0) 31 (28.4) 12 (11.0)
4 of 6 (2) 7 of 7 (1) 11 of 11 (2) 6 of 6 (4) 5 of 6 (2)
*A total of 36 providers participated in this study; 6 faculty, 7 emergency medicine residents, 11 off-service residents, 6 physician assistants, and 6 medical students. The phrase “5 of 7” compared with “7 of 7” means that of the 7 emergency medicine residents, 5 were assigned to sit and all 7 were assigned to stand during the encounter.
minutes). Furthermore, the patients involved in the sitrandomized interaction overestimated the amount of time the providers spent with them by an average of 1.3 minutes, whereas the patients involved in the stand-randomized interaction underestimated time by 0.6 minutes. This difference between study arms was statistically significant after adjustment for patient-level and provider-level covariates in a multilevel regression model (P⬍.001). Conversely, providers overestimated the amount of time spent with the patients in both study arms (P⫽.85). A scatterplot of the difference between the estimated and actual length of the initial encounter time (minutes) versus the actual time (minutes) is shown in the Figure. As shown in Table 3, the patients’ qualitative perceptions of the provider-patient interaction did not differ between the study arms.
LIMITATIONS There are a number of limitations with studying perceived quality of the provider-patient interaction. Admittedly, there are multiple nonverbal and verbal cues that affect patient’s perception of empathy, many of which are immeasurable. It was our hope that through randomization, the risk of confounding would be reduced. Although randomization can limit some confounding, given the large number of providers and the many variables that affect the interactions, confounding may still be Volume , . : February
Johnson et al
To Sit or Not to Sit
Table 2. Summary of patient and provider perceptions of time, by study arm. Sit (Nⴝ115)
Perceptions of Time Interruptions, No. (%) None 1 2 Actual length of initial evaluation time, min Mean (SD) Median Range Actual time in seated position, min Mean (SD) Median Range Patient: How many minutes did your provider spend with you? Mean (SD) Median Range Provider: How many minutes did you spend with the patient? Mean (SD) Median Range Patient: Difference between actual and estimate of time, min Mean (SD), estimate–actual Median Range Provider: Difference between actual and estimate of time, min Mean (SD), estimate–actual Median Range Patient: Rate the amount of time your provider spent with you from first entry until his/ her exit, No. (%). Way too little Somewhat too little Appropriate Somewhat too much Way too much Provider: Rate the amount of time you spent with the patient from the time of first entry until exit, No. (%). Way too little Somewhat too little Appropriate Somewhat too much Way too much
108 (93.9) 6 (5.2) 1 (0.9)
8.3 (4.2) 7.6 2.0–20.7
5.4 (2.8)
Stand (Nⴝ109) 102 (93.6) 6 (5.5) 1 (0.9)
8.9 (5.4) 7.6 1.5–34.1
None of the providers sat down
5.0 (0.5–14.7)
9.6 (6.1) 8.0 (1.0–30.0)
8.2 (4.9) 7.0 (1.0–20.0)
8.8 (4.6) 8.0 (2.0–25.0)
9.2 (4.8) 8.0 (2.0–30.0)
1.3 (4.3) 0.7 –7.5, 13.2
–0.6 (4.3) –0.3 –14.1, 13.9
0.5 (3.6) 0.3 –6.3, 10.7
0 5 (4.3) 109 (94.8) 1 (0.9) 0
Volume , . : February
0 10 (8.7) 99 (86.1) 6 (5.2) 0
0.3 (3.2) 0.4 –12.1, 9.5
1 (0.9) 2 (1.8) 106 (97.2) 0 0
0 9 (8.3) 98 (89.9) 2 (1.8) 0
Figure. Scatterplot of the difference between the estimated and actual length of the initial evaluation time (minutes) versus the actual time (minutes). The top panel is based on the patient’s estimate of time spent, whereas the bottom panel is based on the providers’ estimate. The results are presented for the 2 different randomization groups, sit randomized and stand randomized, with different symbols. The box-and-whisker plots on the right display the distribution of estimate–actual. “Sit” is indicated by a circle; “stand,” by a triangle.
present, particularly for the assessment of the patient perception of the provider. Our study was further limited by the focus of the health care setting in which the research was conducted. Though this was an ED-based study, results may likely be generalized to some outpatient settings (eg, urgent care centers), but to be certain, additional study in lower acuity illness settings is needed. Also, before results are applied to the critically ill population within the ED, further study would be required. A Likert scale was used for ease of questionnaire administration. The survey questionnaires were, however, untested and not validated. Debate exists about whether an observer’s presence within the examination room would affect provider behavior. Within our teaching hospital, it is not uncommon for observers (eg, medical students, nursing students, patients’ family members) to be present during an initial encounter, and we did not believe that this would adversely affect provider behavior. Finally, the initial encounter is only one aspect, a brief part, of the patient’s time spent in the ED. The time perceptions and even the perceptions of care the provider imparted may have been different if measured at the end of the ED visit. Annals of Emergency Medicine 191
To Sit or Not to Sit
Johnson et al
Table 3. Summary of patients’ perception of the providerpatient interaction, by study arm. Patients’ Perceptions
Sit (Nⴝ115)
Stand (Nⴝ109)
0 0 4 (3.5) 58 (50.4) 53 (46.1)
0 0 4 (3.7) 62 (56.9) 43 (39.4)
0 0 6 (5.2) 54 (47) 55 (47.8)
1 (0.9) 1 (0.9) 4 (3.7) 58 (53.2) 45 (41.3)
0 0 11 (9.6) 51 (44.3) 53 (46.1)
1 (0.9) 0 13 (11.9) 47 (43.1) 48 (44.0)
The health care provider had good bedside manner, No. (%) Strongly disagree Disagree Neutral Agree Strongly agree The health care provider cared about me, No. (%) Strongly disagree Disagree Neutral Agree Strongly agree The health care provider understood my problem, No. (%) Strongly disagree Disagree Neutral Agree Strongly agree
DISCUSSION Patients perceived that time with the emergency physician passed slower when the provider was sitting and faster when the provider stood. No previous studies have addressed this issue, but our results tie nicely into previous work by Lin et al,4 which has shown the relationship between higher overall patient satisfaction ratings when patient expectations for visit length are surpassed. Our study did not measure overall satisfaction or exact patient expectations for visit length, but this may be an area for future research. Surprisingly, a seated posture was not predictive of higher patient ratings of the provider-patient interaction. Similar results were published in an abstract evaluating overall patient satisfaction, patient perception of time spent with the physician, and patient perception of total time spent in the ED.9 Results of 73 fast-track patients concluded no difference in patient satisfaction with physician care, regardless of whether the physician sat or stood during the patient encounter. Our methods and those by Gambarota et al9 resemble each other closely; however, we compared patient perceptions of time spent to actual time spent. In an overtaxed ED, rapport-building can be challenging because providers do not have, and often cannot take, a lot of time to establish rapport. Yet, doing so, as Rosenzweig1 implored, is important. Considering that time and motion studies have shown that emergency physicians, for example, spend approximately 32% of their time doing direct patient care activities, even less time exists for rapportbuilding.10 192 Annals of Emergency Medicine
Given the time pressures inherent in medicine, altering perceptions of time may prove essential. Our study had strength of focus, high questionnaire completion rates, and hard endpoints. Our study was further strengthened by data collection by a single investigator, R.L.J. We acknowledge that many nonverbal and verbal cues by the provider influence patient perceptions. Posture is just one of many. Although posture affects patient perceptions of time spent at the bedside, it does not influence patient perceptions of the quality of that interaction. Supervising editor: J. Stephan Stapczynski, MD Author contributions: RLJ, ATS, and DGG conceived of the study, designed the trial, and obtained research funding. RLJ, ATS, ALW, and DGG participated in study design. RLJ administered the survey and collected the data. All authors analyzed and interpreted the data, took responsibility for quality control, and contributed to its reporting. All authors contributed to the drafting and editing of the article. The final article was approved by all authors. RLJ and ATS take responsibility for the paper as a whole. Funding and support: By Annals policy, all authors are required to disclose any and all commercial, financial, and other relationships in any way related to the subject of this article, that might create any potential conflict of interest. See the manuscript Submission Agreement in this issue for examples of specific conflicts covered by this statement. The Mayo Clinic Foundation supports the funding of projects conducted on site. Publication dates: Received for publication June 22, 2006. Revisions received November 22, 2006, and March 29, 2007. Accepted for publication April 30, 2007. Available online June 27, 2007. Presented as a poster at the Society of Academic Emergency Medicine annual meeting, May 2006, San Francisco, CA. Reprints not available from authors. Address for correspondence: Rebecca L. Johnson, MD, 200 First St SW, Mayo Clinic, Rochester, MN 55905; 507-2552216, fax 507-255-6592. E-mail: [email protected].
REFERENCES 1. Rosenzweig S. Emergency rapport. J Emerg Med. 1993;11:775778. 2. LeBlond RF, DeGowin RL, Brown DD. History taking and the medical record. In: Stapczynski JS, ed. DeGowin’s Diagnostic Examination. 8th ed. New York, NY: McGraw-Hill Companies, Inc; 2004:19-20. 3. Coulehan JL, Block MR. With simple, kindly words. In: The Medical Interview: Mastering Skills for Clinical Practice. 4th ed. Philadelphia, PA: F.A. Davies Company; 2001:31-33. 4. Lin CT, Albertson GA, Schilling LM, et al. Is patients’ perception of time spent with the physician a determinant of ambulatory patient satisfaction? Arch Intern Med. 2001;161:1437-1442. 5. Strasser F, Palmer JL, Willey J, et al. Impact of physician sitting versus standing during impatient oncology consultations: patients’ preference
Volume , . : February
Johnson et al and perception of compassion and duration: a randomized controlled trial. J Pain Symptom Manage. 2005;29:489-497. 6. Gross D, Zyzanski S, Borawski E, et al. Patient satisfaction with time spent with their physician. J Fam Pract. 1998;46:133-137. 7. Moore PJ. Medical malpractice: the effect of doctor-patient relations on medical patient perceptions and malpractice intentions. West J Med. 2000;173:244-250. 8. Trout A, Magnusson AR, Hedges JR. Patient satisfaction
Volume , . : February
To Sit or Not to Sit investigations and the emergency department: what does the literature say? Acad Emerg Med. 2000;7:695-709. 9. Gambarota M, Eberhardt M, Melanson S, et al. Should emergency physicians sit or stand? it does not much matter [abstract]. Ann Emerg Med. 2005;46:S56. Abstract 199. 10. Hollingsworth JC, Chisholm CD, Giles BK, et al. How do physicians and nurses spend their time in the emergency department? Ann Emerg Med. 1998;31:87-91.
Annals of Emergency Medicine 193
Figure E1. Patient questionnaire.
193.e1 Annals of Emergency Medicine
Volume , . : February
Figure E2. Provider questionnaire.
Volume , . : February
Annals of Emergency Medicine 193.e2
To Sit or Not to Sit? Rebecca L. Johnson, MD Annie T. Sadosty, MD Amy L. Weaver, MS Deepi G. Goyal, MD
From the Mayo Clinic College of Medicine (Johnson), Department of Emergency Medicine (Sadosty, Goyal), and Division of Biostatistics (Weaver), Mayo Clinic, Rochester, MN.
Study objective: We prospectively examine whether provider posture (seated versus standing) influences patient and provider estimates of time spent at the bedside relative to actual time and patient perceptions of the provider-patient interaction. Methods: A convenience sample of consenting adult patients presenting to an academic tertiary care emergency department between September 7, 2005, and September 25, 2005, were eligible for inclusion in this randomized, controlled trial. Providers (emergency medicine attending physicians, residents, physician assistants, and medical students) were randomly assigned to sit or stand during the initial encounter, after which, participants completed questionnaires about their perceptions of provider-patient interactions and time spent therein. Actual encounter length was measured. Data were analyzed to determine whether patient and provider perception differences existed, using a multilevel regression model that was adjusted for patient-level and provider-level covariates. Results: Two hundred twenty-four consenting patients met inclusion criteria (239 approached; 15 excluded). Data from 36 providers were collected. The mean length of encounters in both study arms was 8.6 minutes (SD 4.8; range 1.5 to 34.1). Patients involved in seated interactions overestimated time providers spent performing initial encounters by an average of 1.3 minutes (SD 4.3 minutes), whereas patients involved in the standing interactions underestimated time by an average of 0.6 minutes (SD 4.3 minutes) (P⫽.001). Conversely, providers overestimated time spent with patients in both study arms (P⫽.85; mean [SD] 0.5 [3.6] versus 0.3 [3.2] minutes). Patient perceptions of the quality of patient-provider interactions were not affected by provider posture. Conclusion: Although provider posture during the initial interaction affects patient perceptions of time spent at the bedside, it does not influence patient perception of the provider’s bedside manner, sense of caring, or understanding of the patient’s problem. [Ann Emerg Med. 2008;51:188-193.] 0196-0644/$-see front matter Copyright © 2008 by the American College of Emergency Physicians. doi:10.1016/j.annemergmed.2007.04.024
SEE EDITORIAL, P. 194. INTRODUCTION Background Rapport-building is essential to the physician-patient relationship.1-3 Classic textbooks dedicate chapters to physicianpatient interactions and medical schools dedicate courses to the same.2,3 The classic text DeGowin’s Diagnostic Examination2 posits, “To obtain the patient’s confidence and rapport, present yourself as unhurried, interested, and sympathetic. Sit at eye level without a desk between you.” A textbook on the medical interview states, “. . . if you stand by the door rather than sit by the bed, patients may assume you are in a hurry . . . .”3 Indeed, it is often promulgated throughout medical education that 188 Annals of Emergency Medicine
patients perceive that providers who sit during the encounter are at the bedside longer than providers who stand for the same amount of time, yet we could find no study in the medical literature that examined the effects of provider posture on provider-actual, patient-perceived, and providerperceived time. Importance Patient perceptions of time spent with the provider have been shown to be determinants of patient satisfaction.4-6 Patient satisfaction, in turn, has been linked with increased referrals, improved compliance, decreased cost, enhanced clinical outcomes, and decreased litigation.4,6-8 Spending more time at the bedside is not always possible. On busy days, Volume , . : February
Johnson et al
To Sit or Not to Sit Editor’s Capsule Summary
What is already known on this topic It is commonly taught that physicians should sit rather than stand during the encounter to impart a sense of unhurried caring. This teaching is unsubstantiated. What question this study addressed Does provider posture—sitting or standing— during the initial emergency department evaluation affect the patient’s perception of the duration of the encounter or the provider’s bedside manner? What this study adds to our knowledge In this 224-patient randomized trial, patients overestimated the time spent by seated providers and underestimated the time spent by standing providers, both by small amounts. How this might change clinical practice Sitting may modestly improve the patient’s perception of physician care. Given the low cost of this simple act, it may be prudent to provide a place for the physician to sit in each patient care area.
altering patient perceptions about the amount of time spent may be the only alternative. Goals of This Investigation In this article, we examine provider posture (seated versus standing) during the initial emergency department (ED) encounter. We hypothesize that having the provider sit during the initial encounter will increase patient and provider estimates of the time spent at the bedside relative to actual time and improve patient perceptions of the provider-patient interaction.
MATERIALS AND METHODS Study Design and Setting We conducted this prospective nonblinded randomized study in an academic tertiary-care ED with 60,000 annual adult visits. Selection of Participants Adult patients presenting between September 7, 2005, and September 25, 2005, were eligible for inclusion in the study. Patients were enrolled by a sole study investigator (R.L.J.) by convenience sampling, with data collection generally between 8 AM and midnight, on weekdays and weekends. Exclusion criteria included patients ineligible for inclusion according to nursing assessment if they were younger than 18 years, were non-English speakers, and had a health status and cognitive function that affected their ability to communicate answers to R.L.J. Patients triaged to the critical resuscitation area were excluded according to acuity potential. Volume , . : February
Providers present during the study period included 23 emergency medicine attending physicians, 9 emergency medicine residents, 7 off-service residents, 4 physician assistants, and 6 rotating medical students. Encounters by the other authors of the study were not included. The study was designed to enroll 200 participants (100 in each study arm), given the time constraints of R.L.J. The Mayo Clinic institutional review board approved this study, and all participants provided written informed consent. Methods of Measurement and Data Collection and Processing R.L.J. approached patients deemed eligible by nursing, after they were assigned to a room, to obtain informed consent. R.L.J. confirmed eligibility and excluded patients who failed to answer a series of orientation questions relating to current month, current year, and current president of the United States correctly. Blinded to the nature of the study, eligible patients were enrolled and demographic data were collected. Before the beginning of the study period a block randomization schedule was created. Using this block randomization schedule, R.L.J. randomized the provider to sit or stand, notified the provider of their assigned posture just before entrance into the examination room, and accompanied the provider into the room. At a teaching hospital, patients are often treated by multiple providers during their visit. The first provider present for the initial encounter was selected for randomization. Those providers randomized to sit during the initial encounter (defined as introduction, history taking, physical examination, and rendering of initial impressions and plan) were instructed to perform as much of the initial encounter as possible in the seated position. Because portions of the physical examination require providers to stand, posture during the physical examination was left to the seated provider’s discretion. Providers randomized to stand were instructed to stand during the entirety of the initial encounter. R.L.J. used 3 stopwatches to capture time data. The stopwatches were concealed from provider and patient view within R.L.J.’s white-coat pocket. The entire length of the initial encounter was timed from provider entrance into the room until the provider exited the room. In addition to the running initial encounter time, R.L.J. clocked the amount of time spent seated with a second concealed stopwatch for both sit-randomized and stand-randomized providers. A third concealed stopwatch was present to collect any time spent that could interrupt the running initial encounter time (eg, answering pager, telephone interaction, requirement for room exit). Patient and provider demographic data, actual encounter time data (total encounter time minus interruption time), and time spent sitting were recorded. We developed a 5-item questionnaire with a Likert scale for this study (Appendix E1, available online at http://www.annemergmed.com). Immediately after initial encounter services were delivered and strictly following the Annals of Emergency Medicine 189
To Sit or Not to Sit language of the scripted questionnaire, R.L.J. administered the questionnaire to the patient who was blinded to time. On exit from the room, the provider was asked to complete a brief questionnaire (Appendix E2, available online at http:// www.annemergmed.com). Two sham questions, designed to address provider discomforts, were inserted into the questionnaire to distract the providers from the true nature of the study. Sham data were not reported for the purposes of this study. After administrating the questionnaires, R.L.J. would proceed to the next patient identified by nursing to be eligible for enrollment. Primary Data Analysis Data were analyzed according to the intent-to-treat principle. The responses to the questionnaire items, actual initial encounter time, estimated time spent together in the initial encounter, and the difference in the actual and estimated times were summarized using standard descriptive statistics. The outcomes variables of interest were the difference in the actual and estimated time by the patient and difference in the actual and estimated time by the provider. Multilevel or randomeffects models were fit with a generalized mixed linear randomeffects model with the SAS (SAS Institute, Inc., Cary, NC) MIXED procedure to evaluate the relationship between provider posture (sitting or standing) and each of the 2 outcome variables. The model allowed for the adjustment of both patient-level (age and sex) and provider-level (sex and provider type) covariates. All calculated P values were 2-sided, and P values less than .05 were considered statistically significant.
RESULTS R.L.J. approached 239 patients for inclusion in the study, of whom 15 were not enrolled. Nursing excluded 1 patient because of severe sore throat pain, precluding ability to communicate; R.L.J. excluded 4 patients for failing to correctly answer the orientation questions, and 10 did not provide informed consent. Of the 10 denying consent, 5 (50%) were women, with a mean age of 55.9 years (SD 22.7 years), median age of 55 years, over a range of 22 to 85 years. Results herein are based on the 224 patients and 36 providers who gave informed consent to the study. Patient and provider characteristics are summarized in Table 1. The patient and provider perceptions and estimates of time spent are listed in Table 2. On average, each encounter lasted 8.6 minutes, and the average length of the encounters was similar in the 2 study arms. None of the providers randomized to standing (N⫽109) sat down during the encounter. The providers randomized to sitting (N⫽115) sat an average of 67.4% of the total time. The patients involved in the sit-randomized interaction tended to report a longer duration for the amount of time the provider spent with them compared with the patients involved in the stand-randomized interaction (mean 9.6 versus 8.2 190 Annals of Emergency Medicine
Johnson et al Table 1. Summary of patient and provider characteristics, by study arm. Characteristics Patient sex, female, No. (%) Patient age, y Mean (SD) Range Patients treated by a female provider, No. (%) Patients treated by each type of provider, No. (%) Emergency medicine faculty Emergency medicine resident Resident off Physician assistant Medical student No. of participating providers (median no. of patients treated by each provider)* Emergency medicine faculty Emergency medicine resident Resident off Physician assistant Medical student
Sit (Nⴝ115) 72 (62.6) 48.7 (19.9) (18.0–90.0) 73 (63.5)
11 (9.6) 23 (20.0) 41 (35.6) 29 (25.2) 11 (9.6)
4 of 6 (2) 5 of 7 (4) 10 of 11 (2.5) 5 of 6 (4) 6 of 6 (1.5)
Stand (Nⴝ109) 55 (50.5) 48.0 (19.9) (18.0–94.0) 62 (56.9)
8 (7.3) 22 (20.2) 36 (33.0) 31 (28.4) 12 (11.0)
4 of 6 (2) 7 of 7 (1) 11 of 11 (2) 6 of 6 (4) 5 of 6 (2)
*A total of 36 providers participated in this study; 6 faculty, 7 emergency medicine residents, 11 off-service residents, 6 physician assistants, and 6 medical students. The phrase “5 of 7” compared with “7 of 7” means that of the 7 emergency medicine residents, 5 were assigned to sit and all 7 were assigned to stand during the encounter.
minutes). Furthermore, the patients involved in the sitrandomized interaction overestimated the amount of time the providers spent with them by an average of 1.3 minutes, whereas the patients involved in the stand-randomized interaction underestimated time by 0.6 minutes. This difference between study arms was statistically significant after adjustment for patient-level and provider-level covariates in a multilevel regression model (P⬍.001). Conversely, providers overestimated the amount of time spent with the patients in both study arms (P⫽.85). A scatterplot of the difference between the estimated and actual length of the initial encounter time (minutes) versus the actual time (minutes) is shown in the Figure. As shown in Table 3, the patients’ qualitative perceptions of the provider-patient interaction did not differ between the study arms.
LIMITATIONS There are a number of limitations with studying perceived quality of the provider-patient interaction. Admittedly, there are multiple nonverbal and verbal cues that affect patient’s perception of empathy, many of which are immeasurable. It was our hope that through randomization, the risk of confounding would be reduced. Although randomization can limit some confounding, given the large number of providers and the many variables that affect the interactions, confounding may still be Volume , . : February
Johnson et al
To Sit or Not to Sit
Table 2. Summary of patient and provider perceptions of time, by study arm. Sit (Nⴝ115)
Perceptions of Time Interruptions, No. (%) None 1 2 Actual length of initial evaluation time, min Mean (SD) Median Range Actual time in seated position, min Mean (SD) Median Range Patient: How many minutes did your provider spend with you? Mean (SD) Median Range Provider: How many minutes did you spend with the patient? Mean (SD) Median Range Patient: Difference between actual and estimate of time, min Mean (SD), estimate–actual Median Range Provider: Difference between actual and estimate of time, min Mean (SD), estimate–actual Median Range Patient: Rate the amount of time your provider spent with you from first entry until his/ her exit, No. (%). Way too little Somewhat too little Appropriate Somewhat too much Way too much Provider: Rate the amount of time you spent with the patient from the time of first entry until exit, No. (%). Way too little Somewhat too little Appropriate Somewhat too much Way too much
108 (93.9) 6 (5.2) 1 (0.9)
8.3 (4.2) 7.6 2.0–20.7
5.4 (2.8)
Stand (Nⴝ109) 102 (93.6) 6 (5.5) 1 (0.9)
8.9 (5.4) 7.6 1.5–34.1
None of the providers sat down
5.0 (0.5–14.7)
9.6 (6.1) 8.0 (1.0–30.0)
8.2 (4.9) 7.0 (1.0–20.0)
8.8 (4.6) 8.0 (2.0–25.0)
9.2 (4.8) 8.0 (2.0–30.0)
1.3 (4.3) 0.7 –7.5, 13.2
–0.6 (4.3) –0.3 –14.1, 13.9
0.5 (3.6) 0.3 –6.3, 10.7
0 5 (4.3) 109 (94.8) 1 (0.9) 0
Volume , . : February
0 10 (8.7) 99 (86.1) 6 (5.2) 0
0.3 (3.2) 0.4 –12.1, 9.5
1 (0.9) 2 (1.8) 106 (97.2) 0 0
0 9 (8.3) 98 (89.9) 2 (1.8) 0
Figure. Scatterplot of the difference between the estimated and actual length of the initial evaluation time (minutes) versus the actual time (minutes). The top panel is based on the patient’s estimate of time spent, whereas the bottom panel is based on the providers’ estimate. The results are presented for the 2 different randomization groups, sit randomized and stand randomized, with different symbols. The box-and-whisker plots on the right display the distribution of estimate–actual. “Sit” is indicated by a circle; “stand,” by a triangle.
present, particularly for the assessment of the patient perception of the provider. Our study was further limited by the focus of the health care setting in which the research was conducted. Though this was an ED-based study, results may likely be generalized to some outpatient settings (eg, urgent care centers), but to be certain, additional study in lower acuity illness settings is needed. Also, before results are applied to the critically ill population within the ED, further study would be required. A Likert scale was used for ease of questionnaire administration. The survey questionnaires were, however, untested and not validated. Debate exists about whether an observer’s presence within the examination room would affect provider behavior. Within our teaching hospital, it is not uncommon for observers (eg, medical students, nursing students, patients’ family members) to be present during an initial encounter, and we did not believe that this would adversely affect provider behavior. Finally, the initial encounter is only one aspect, a brief part, of the patient’s time spent in the ED. The time perceptions and even the perceptions of care the provider imparted may have been different if measured at the end of the ED visit. Annals of Emergency Medicine 191
To Sit or Not to Sit
Johnson et al
Table 3. Summary of patients’ perception of the providerpatient interaction, by study arm. Patients’ Perceptions
Sit (Nⴝ115)
Stand (Nⴝ109)
0 0 4 (3.5) 58 (50.4) 53 (46.1)
0 0 4 (3.7) 62 (56.9) 43 (39.4)
0 0 6 (5.2) 54 (47) 55 (47.8)
1 (0.9) 1 (0.9) 4 (3.7) 58 (53.2) 45 (41.3)
0 0 11 (9.6) 51 (44.3) 53 (46.1)
1 (0.9) 0 13 (11.9) 47 (43.1) 48 (44.0)
The health care provider had good bedside manner, No. (%) Strongly disagree Disagree Neutral Agree Strongly agree The health care provider cared about me, No. (%) Strongly disagree Disagree Neutral Agree Strongly agree The health care provider understood my problem, No. (%) Strongly disagree Disagree Neutral Agree Strongly agree
DISCUSSION Patients perceived that time with the emergency physician passed slower when the provider was sitting and faster when the provider stood. No previous studies have addressed this issue, but our results tie nicely into previous work by Lin et al,4 which has shown the relationship between higher overall patient satisfaction ratings when patient expectations for visit length are surpassed. Our study did not measure overall satisfaction or exact patient expectations for visit length, but this may be an area for future research. Surprisingly, a seated posture was not predictive of higher patient ratings of the provider-patient interaction. Similar results were published in an abstract evaluating overall patient satisfaction, patient perception of time spent with the physician, and patient perception of total time spent in the ED.9 Results of 73 fast-track patients concluded no difference in patient satisfaction with physician care, regardless of whether the physician sat or stood during the patient encounter. Our methods and those by Gambarota et al9 resemble each other closely; however, we compared patient perceptions of time spent to actual time spent. In an overtaxed ED, rapport-building can be challenging because providers do not have, and often cannot take, a lot of time to establish rapport. Yet, doing so, as Rosenzweig1 implored, is important. Considering that time and motion studies have shown that emergency physicians, for example, spend approximately 32% of their time doing direct patient care activities, even less time exists for rapportbuilding.10 192 Annals of Emergency Medicine
Given the time pressures inherent in medicine, altering perceptions of time may prove essential. Our study had strength of focus, high questionnaire completion rates, and hard endpoints. Our study was further strengthened by data collection by a single investigator, R.L.J. We acknowledge that many nonverbal and verbal cues by the provider influence patient perceptions. Posture is just one of many. Although posture affects patient perceptions of time spent at the bedside, it does not influence patient perceptions of the quality of that interaction. Supervising editor: J. Stephan Stapczynski, MD Author contributions: RLJ, ATS, and DGG conceived of the study, designed the trial, and obtained research funding. RLJ, ATS, ALW, and DGG participated in study design. RLJ administered the survey and collected the data. All authors analyzed and interpreted the data, took responsibility for quality control, and contributed to its reporting. All authors contributed to the drafting and editing of the article. The final article was approved by all authors. RLJ and ATS take responsibility for the paper as a whole. Funding and support: By Annals policy, all authors are required to disclose any and all commercial, financial, and other relationships in any way related to the subject of this article, that might create any potential conflict of interest. See the manuscript Submission Agreement in this issue for examples of specific conflicts covered by this statement. The Mayo Clinic Foundation supports the funding of projects conducted on site. Publication dates: Received for publication June 22, 2006. Revisions received November 22, 2006, and March 29, 2007. Accepted for publication April 30, 2007. Available online June 27, 2007. Presented as a poster at the Society of Academic Emergency Medicine annual meeting, May 2006, San Francisco, CA. Reprints not available from authors. Address for correspondence: Rebecca L. Johnson, MD, 200 First St SW, Mayo Clinic, Rochester, MN 55905; 507-2552216, fax 507-255-6592. E-mail: [email protected].
REFERENCES 1. Rosenzweig S. Emergency rapport. J Emerg Med. 1993;11:775778. 2. LeBlond RF, DeGowin RL, Brown DD. History taking and the medical record. In: Stapczynski JS, ed. DeGowin’s Diagnostic Examination. 8th ed. New York, NY: McGraw-Hill Companies, Inc; 2004:19-20. 3. Coulehan JL, Block MR. With simple, kindly words. In: The Medical Interview: Mastering Skills for Clinical Practice. 4th ed. Philadelphia, PA: F.A. Davies Company; 2001:31-33. 4. Lin CT, Albertson GA, Schilling LM, et al. Is patients’ perception of time spent with the physician a determinant of ambulatory patient satisfaction? Arch Intern Med. 2001;161:1437-1442. 5. Strasser F, Palmer JL, Willey J, et al. Impact of physician sitting versus standing during impatient oncology consultations: patients’ preference
Volume , . : February
Johnson et al and perception of compassion and duration: a randomized controlled trial. J Pain Symptom Manage. 2005;29:489-497. 6. Gross D, Zyzanski S, Borawski E, et al. Patient satisfaction with time spent with their physician. J Fam Pract. 1998;46:133-137. 7. Moore PJ. Medical malpractice: the effect of doctor-patient relations on medical patient perceptions and malpractice intentions. West J Med. 2000;173:244-250. 8. Trout A, Magnusson AR, Hedges JR. Patient satisfaction
Volume , . : February
To Sit or Not to Sit investigations and the emergency department: what does the literature say? Acad Emerg Med. 2000;7:695-709. 9. Gambarota M, Eberhardt M, Melanson S, et al. Should emergency physicians sit or stand? it does not much matter [abstract]. Ann Emerg Med. 2005;46:S56. Abstract 199. 10. Hollingsworth JC, Chisholm CD, Giles BK, et al. How do physicians and nurses spend their time in the emergency department? Ann Emerg Med. 1998;31:87-91.
Annals of Emergency Medicine 193
Figure E1. Patient questionnaire.
193.e1 Annals of Emergency Medicine
Volume , . : February
Figure E2. Provider questionnaire.
Volume , . : February
Annals of Emergency Medicine 193.e2