Abstract
Background:
Sitting at the bedside may improve patient-clinician communication; however, many clinicians do not regularly sit during inpatient encounters.
Objective:
To determine the impact of adding wall-mounted folding chairs inside patient rooms, beyond any impact from a resident education campaign, on the patient-reported frequency of sitting at the bedside by internal medicine resident physicians.
Design, Setting, and Participants:
Prospective, controlled pre-post trial between 2019–2022 (data collection paused 2020–2021 due to COVID-19) at an academic hospital in Baltimore, Maryland. Folding chairs were installed in two of four internal medicine units and educational activities were delivered equally across all units.
Main Outcome and Measures:
Patient-reported frequency of sitting at bedside, assessed as means on Likert-type items with 1 being “never” and 5 being “every single time.” We also examined the frequency of other patient-reported communication behaviors.
Results:
256 and 206 patients enrolled in the pre and post-intervention periods, respectively. The mean frequency of patient-reported sitting by resident physicians increased from 1.8 (SD 1.2) to 2.3 (1.2) on education-only units (absolute difference 0.48 [95% CI 0.21–0.75]) and from 2.0 (1.3) to 3.2 (1.4) on units receiving chairs (1.16, [0.87–1.45]). Comparing differences between groups using ordered logistic regression adjusting for clustering within residents, units with added chairs had greater increases in sitting (odds ratio 2.05 [1.10–3.82]), spending enough time at the bedside (2.43 [1.32–4.49]), and checking for understanding (3.04 [1.44–6.39]). Improvements in sitting and other behaviors were sustained on both types of units.
Conclusions:
Adding wall-mounted folding chairs may help promote effective patient-clinician communication.
Keywords: Communication, Patient-Clinician Relationship, Hospital Medicine, Patient-clinician communication, Graduate Medical
Background
Effective patient-clinician communication is associated with improved patient comprehension, satisfaction, and health outcomes.1–6 However, building strong patient-clinician rapport in the hospital setting may be challenging. Patients typically interact with many clinicians. In turn, some hospital clinicians report that they are not always able to fully answer questions or discuss treatment options with patients and their families due to high workloads.7
Simple, modifiable physician behaviors, such as sitting at the hospital bedside, may improve patient-clinician communication.8–13 Sitting has been associated with improvements in patients’ perceived understanding of their condition and with patients perceiving that clinicians spend more time at the bedside, without actually making clinical encounters take longer.8,11 Despite these potential benefits, hospital clinicians do not regularly sit during patient interactions.8,9,14 In a recent study at our institution, the majority of hospitalized patients (56%) reported that internal medicine residents never sat at the bedside.8 Surveyed residents cited inadequate access to chairs as the biggest barrier to sitting.8 The chairs that were part of the standard room furniture were often occupied by visitors, patient belongings, or supplies.
The primary aim of this study was to assess the impact of adding wall-mounted folding chairs, over and above any impact from an educational campaign, on the frequency of sitting by internal medicine residents. We hypothesized that the folding chairs placed inside the entrances of patient rooms would increase sitting. We also evaluated the impact of the chairs on the frequency of other positive resident physician communication behaviors (e.g., checking for understanding, spending enough time at bedside), which we hypothesized would improve as resident physician sitting increased.
Methods
Study Design and Setting
We performed a prospective, controlled, pre-post evaluation of a 2-component intervention aimed to increase sitting at the bedside on four general internal medicine units at The Johns Hopkins Hospital.
Our intervention consisted of (1) conducting educational activities on the importance of sitting for the residents based on all 4 units, and (2) installing folding chairs in patient rooms on 2 units. Evaluation was conducted through patient surveys, with baseline surveys conducted August 2019 – January 2020. We planned to conduct post-intervention surveys beginning in February 2020 but deferred starting these until October 2021 due to the COVID-19 pandemic. Chairs were left in place throughout this period.
Each of the four units (23 private patient rooms each, comprising different floors of the same building) houses one of four chief-resident led teaching services (“firms”). Residents are randomly distributed to firms when they matriculate. Patients are generally assigned among the four units according to the first available bed.
Firm teams are comprised of 1 attending physician, one upper-year (PGY2–3) resident physician, four first-year (PGY1) resident physicians, and 2–3 medical students. On the first full hospital day, the team typically rounds together at the patient’s bedside, and the primary PGY1 resident physician sits while all others stand. On subsequent days, PGY1 resident physicians conduct brief visits alone in the morning, sometimes followed by a visit together with the PGY2–3 resident physician. If they visit together, the PGY1 resident physician may sit. Later in the day, the PGY2–3 resident physician and attending visit the patient, and the PGY2–3 resident physician may sit. Resident physicians (either PGY1 or PGY2–3s) may also provide periodic updates, typically alone, to the patient and their families throughout the day.
Study intervention
Component 1: Educational activities
In January 2020, we conducted short educational activities (15-minute noon conference presentation, flyers in work rooms, discussion with each chief resident) focusing on benefits of sitting at the bedside, including residents from all firms equally. Because two-thirds of the residents working in January 2020 graduated prior to our COVID-19-delayed post-intervention round of surveys, we provided a refresher for all firms in October 2021 consisting of the same components.
Component 2: Chair installation
In January 2020, we installed a folding chair in every patient room on 2 of 4 study units. Chairs were hung from specifically-designed wall hooks (between the room entrance and the bed), were easily visible, and labeled to indicate they were intended for clinician use (Figure 1). We chose the two units to receive chairs prior to any assessment of residents’ baseline sitting behavior.
Figure 1:
Example clinician chair
Intervention evaluation
Four different surveyors - a mixture of patient experience professionals, hospital administrators and graduate students - worked during each survey period (8 surveyors total). Using the electronic health record, surveyors identified potentially eligible patients who were admitted to a firm service and seen by the same residents for at least 2 days.15 Prior to approaching patients, surveyors asked bedside nurses about mental status, limited English proficiency, or impairments (e.g., vision impairment) that would preclude patients from recognizing pictures of residents and completing a survey (Figure 2). Because communicating with care partners is essential when patients are non-communicative, we surveyed family members of patients who were unable to complete a survey if nurses reported that family members were involved in medical decision-making, able to complete a survey in English, and present at the bedside.
Figure 2.
Flow of participants through study during pre- and post-intervention periods
Surveyors approached patients on the first weekday of their eligibility. If a patient was occupied, surveyors returned on subsequent weekdays until the patient completed the survey, declined, or was discharged. Surveyors collected some surveys each week (3 days/week on average), but were not available every weekday, and some patients were discharged without being approached.
Surveyors informed participants that we wanted to understand resident physician communication behaviors through a brief, anonymous survey that would not impact their care. In order to ensure accurate attribution of survey results to resident physicians who went through the education and were based on the study floors (as opposed to visitors such as consultants), surveyors administered the survey only to patients who identified their assigned first year (PGY1) and/or upper year (PGY2–3) resident physician from a photo line-up, similar to a prior study of hospitalists.15
Surveys included a question about the frequency of sitting in general and a separate question about sitting for “important updates and family meetings,” as sitting behaviors may vary across different types of encounters. Surveys also included questions about patients’ perceptions of 4 other physician communication behaviors (spending enough time, checking for understanding, not interrupting, and being in a rush), which were selected based on prior literature.7,10,15 Answers were based on a 5-point Likert-type scale (“Never, Rarely, Sometimes, Most of the time, Every single time”).8 Surveyors entered responses into Qualtrics survey software (Qualtrics, Provo, UT).
To protect patient confidentiality, we did not collect specific patient demographics. We compared aggregate patient demographics between baseline and post-intervention periods and between education-only units and units receiving chairs using anonymized administrative data.
Statistical Analysis
Our sample size of at least 200 patients per time period was chosen to afford at least 80% power to detect a difference of 0.75 points between baseline and post-intervention sitting frequency for general bedside visits on the units with added chairs. Few data exist to inform a clinical relevance threshold for changes in Likert scale-scored clinician behaviors.9,16 In a prior study, clinician performance on a composite score of 6 observed behaviors (including sitting) correlated with improvements in patient experience ratings; the mean composite score was 22%.9 We felt a mean increase of 1 on a Likert scale of 1 – 5 has face validity as a substantive increase and conservatively powered the study to detect 0.75.
Given ordinal (Likert) response variables, our pre-specified primary analysis compared differences between groups using repeated-measures, ordered logistic regression. We adjusted for resident identity (given multiple patient surveys for individual residents). Our primary comparisons were the frequency of patient-reported sitting for education-only units compared to their baseline and for units with added chairs compared to their baseline. Secondly, we compared the changes from baseline for units with added chairs to the corresponding changes from baseline for education-only units (difference in differences analysis). Thirdly, to assess for decay in any effect and to account for the progression of resident physicians throughout training, we compared survey responses in the first half of the post-intervention survey period (October-December 2021, immediately following the refresher education in October 2021) to responses in the latter months of the post-intervention survey period (February-April 2022). Next, we performed the same analyses for other pre-specified communication behaviors (e.g., checking for understanding, spending enough time at bedside).
To make changes more explicit, as a secondary analysis, we calculated absolute mean differences in Likert responses. We generated 95% confidence intervals (CIs) for these results using standard errors that assumed a normal distribution and did not account for clustering within residents. Given the different methodologies, we anticipated some discrepancies in statistical significance between comparisons in the regressions and corresponding assessments of mean differences using 95% CIs generated by assuming a normal distribution.
Because experience and roles differ between PGY1 and PGY2–3 resident physicians, we repeated the above analyses stratified by level of training. Finally, we performed a sensitivity analysis in which we excluded surveys from family members.
We assumed an a priori alpha of 0.05 (two-sided) for all analyses. Analyses were performed using Stata (College Station, TX).
The Johns Hopkins University Institutional Review Board considered this work a quality improvement activity (IRB00184753).
Results
Among screened patients, 334/528 (63%) and 217/343 (63%) were eligible in the pre- and post-intervention periods, respectively. Reasons for exclusion are summarized in Figure 2. Among eligible patients, 256 (77%) and 206 (95%) patients completed surveys in the baseline and post-intervention periods, respectively, compiling a total of 364 and 292 surveys at baseline and post-intervention. Only 6% (15/256) and 2% (5/206) of surveys were completed by someone other than the patient in the pre- and post-intervention periods, respectively.
Approximately half of all residents (76/157, 48%) had at least one survey completed about their performance in the post-intervention period; the median surveys per resident was 3 (IQR 2–5).
While we did not collect specific patient-level data, we expect surveyed patients’ demographics to resemble those of the overall patient population within study units. Patient demographics from the 4 firm services were similar between the baseline and post-intervention survey periods (Supplementary). During the academic year including the post-intervention period, 51% of patients were female, with a median age of 59 years (IQR 44–69). The racial distribution was 56% Black, 34% White, 2% Asian, and 7% other racial identities. Most patients’ (94%) preferred language was English. There were more discharges in the post-intervention period (3663 vs. 2691), although median length of stay remained similar (4.0 days). In the post-intervention period, characteristics were similar between education-only units and units receiving chairs.
Patient perceptions of resident physician sitting at the bedside and of other communication behaviors during the baseline and post-intervention periods are shown in Table 1. At baseline, few residents usually or always sat on either education-only or units with added chairs (10% and 16%, respectively). Post-intervention, the percentage of residents who usually or always sat increased to 15% on education-only units and to 45% on units with added chairs.
Table 1:
Frequency of Patient-Reported Resident Physician Communication Behaviors in Baseline and Post Intervention Periods for All Residents
| Never N (%) | Rarely N (%) | Sometimes N (%) | Usually N (%) | Always N (%) | |||
|---|---|---|---|---|---|---|---|
| Physician sits: | |||||||
| When he/she visits my bedside | Education-Only Units | Baseline (N=164) | 102 (62) | 21 (13) | 24 (15) | 7 (4) | 10 (6) |
| Post-Intervention (N=150) | 49 (32) | 46 (31) | 33 (22) | 9 (6) | 13 (9) | ||
| Units with Added Chairs | Baseline (N=200) | 101 (51) | 41 (21) | 27 (14) | 11 (6) | 20 (10) | |
| Post-Intervention (N=142) | 25 (18) | 14 (10) | 39 (27) | 35 (25) | 29 (20) | ||
| When giving me important updates or talking with me and my family | Education-Only Units | Baseline (N=164) | 103 (62.8) | 18 (10) | 15 (9) | 1 (1) | 27 (16) |
| Post-Intervention (N=150) | 49 (32.3) | 42 (28) | 38 (25) | 10 (7) | 11 (7) | ||
| Units with Added Chairs | Baseline (N=200) | 107 (54) | 33 (17) | 28 (14) | 10 (5) | 22 (11) | |
| Post-Intervention (N=142) | 22 (15) | 17 (12) | 33 (23) | 30 (21) | 40 (28) | ||
| Other communication behaviors: | |||||||
| Spends enough time at the bedside with me | Education-Only Units | Baseline (N=164) | 1 (0) | 2 (1) | 31 (19) | 37 (23) | 93 (57) |
| Post-Intervention (N=150) | 2 (1) | 4 (3) | 11 (7) | 59 (39) | 74 (49) | ||
| Units with Added Chairs | Baseline (N=200) | 1 (0) | 9 (5) | 38 (19) | 49 (25) | 103 (52) | |
| Post-Intervention (N=142) | 0 (0) | 1 (1) | 5 (4) | 33 (23) | 103 (73) | ||
| Checks to be sure I understand everything | Education-Only Units | Baseline (N=164) | 4 (2) | 6 (4) | 28 (17) | 25 (15) | 101 (62) |
| Post-Intervention (N=150) | 1 (1) | 5 (3) | 8 (5) | 36 (24) | 100 (67) | ||
| Units with Added Chairs | Baseline (N=200) | 2 (1) | 3 (2) | 39 (20) | 36 (18) | 120 (60) | |
| Post-Intervention (N=142) | 2 (1) | 1 (1) | 3 (2) | 13 (9) | 123 (87) | ||
| Lets me talk without interrupting | Education-Only Units | Baseline (N=164) | 8 (5) | 0 (0) | 7 (4) | 29 (18) | 120 (73) |
| Post-Intervention (N=150) | 0 (0) | 1 (1) | 2 (1) | 24 (16) | 123 (82) | ||
| Units with Added Chairs | Baseline (N=200) | 2 (1) | 1 (1) | 13 (7) | 29 (15) | 155 (78) | |
| Post-Intervention (N=142) | 3 (2) | 1 (1) | 3 (2) | 15 (11) | 120 (85) | ||
| Is in a rush when he/she is with me | Education-Only Units | Baseline (N=164) | 105 (64) | 19 (12) | 32 (20) | 1 (1) | 7 (4) |
| Post-Intervention (N=150) | 131 (87) | 13 (9) | 2 (1) | 4 (3) | 0 (0) | ||
| Units with Added Chairs | Baseline (N=200) | 108 (54) | 29 (15) | 48 (24) | 7 (4) | 8 (4) | |
| Post-Intervention (N=142) | 122 (86) | 9 (6) | 6 (4) | 2 (1) | 3 (2) | ||
Comparisons between resident communication behaviors in the baseline and post-intervention periods are shown in Table 2. In the post-intervention period, the mean absolute difference in Likert responses for patient-reported sitting by all residents increased by 0.48 (95% CI: 0.21–0.75) on education-only units and by 1.16 (0.87–1.45) on units receiving chairs. In the primary ordered logistic regression analysis adjusting for clustering among residents, the odds of increased sitting in the post-intervention time period were higher for both education only units (OR 2.53 [1.58–4.06]) and units with added chairs (4.76 [3.08–7.38]). In difference-in-difference analyses (Table 3), units with added chairs had a greater increase in the mean response for sitting during bedside visits (0.68 [0.29–1.08]) and a greater increase in the odds of increased sitting (2.05 [1.10–3.82]). Similar findings were observed when analyzing separately by training level, though the ORs did not reach significance (2.00 [0.94–4.26] for PGY1 resident physicians and 2.26 [0.86–5.92] for PGY2–3 resident physicians). Results for sitting during “important updates or talking with me and my family” were similar (Table 3). Our sensitivity analysis excluding responses from family members revealed no meaningful differences compared to the primary analysis (results not shown).
Table 2.
Patient-Reported Resident Physician Communication Behaviors in Baseline and Post Intervention Periods*
| Education-Only Units | Units with Added Chairs | |||||||
|---|---|---|---|---|---|---|---|---|
| Baseline 9/2019 – 1/2020 | Post-Intervention 10/2021 – 4/2022 | Absolute differencea (95% CI) | Odds ratiob for change between time periods (95% CI) | Baseline 9/2019 – 1/2020 | Post-Intervention 10/2021 – 4/2022 | Absolute differencea (95% CI) | Odds ratiob for change between time periods (95% CI) | |
| Mean (SD) | Mean (SD) | Mean (SD) | Mean (SD) | |||||
| All residents | (N=164) | (N=150) | (N=200) | (N=142) | ||||
| Physician sits: | ||||||||
| When he/she visits my bedside | 1.79 (1.21) | 2.27 (1.23) | 0.48 (0.21–0.75)*** | 2.53 (1.58–4.06)*** | 2.04 (1.33) | 3.20 (1.36) | 1.16 (0.87–1.45)*** | 4.76 (3.08–7.38)*** |
| When giving me important updates or talking with me and my family | 1.97 (1.50) | 2.28 (1.20) | 0.31 (0.01–0.61)* | 2.26 (1.39–3.68)** | 2.04 (1.37) | 3.35 (1.40) | 1.31 (1.01–1.61)*** | 5.35 (3.43–8.34)*** |
| Other communication behaviors: | ||||||||
| Spends enough time at the bedside with me | 4.34 (0.87) | 4.33 (0.83) | −0.01 (−0.20–0.18) | 0.92 (0.56–1.50) | 4.22 (0.94) | 4.68 (0.58) | 0.46 (0.28–0.63)*** | 2.85 (1.78–4.55)*** |
| Checks to be sure I understand everything | 4.30 (1.03) | 4.53 (0.80) | 0.23 (0.02–0.43)* | 1.45 (0.91–2.30) | 4.35 (0.91) | 4.79 (0.65) | 0.44 (0.27–0.62)*** | 4.40 (2.48–7.79)*** |
| Lets me talk without interrupting | 4.54 (0.96) | 4.79 (0.48) | 0.25 (0.08–0.43)** | 1.77 (1.02–3.10)* | 4.67 (0.71) | 4.75 (0.73) | 0.08 (−0.08–0.23) | 1.57 (0.88–2.80) |
| Is in a rush when he/she is with me | 1.70 (1.08) | 1.19 (0.59) | −0.50 (−0.70–−0.31)*** | 0.25 (0.13–0.47)*** | 1.89 (1.13) | 1.27 (0.79) | −0.62 (−0.83–−0.40)*** | 0.20 (0.11–0.36)*** |
| First-year residents (PGY1s) | (N=94) | (N=90) | (N=104) | (N=90) | ||||
| Physician sits: | ||||||||
| When he/she visits my bedside | 1.85 (1.29) | 2.32 (1.18) | 0.47 (0.11–0.83)* | 2.69 (1.41–5.11)** | 2.00 (1.25) | 3.14 (1.33) | 1.14 (0.78–1.51)*** | 5.07 (2.84–9.04)*** |
| When giving me important updates or talking with me and my family | 1.90 (1.44) | 2.29 (1.16) | 0.38 (0.00–0.77)* | 2.60 (1.30–5.18)** | 1.96 (1.32) | 3.24 (1.38) | 1.28 (0.90–1.67)*** | 5.57 (3.14–9.88)*** |
| Other communication behaviors: | 0.99 (0.52–1.90) | |||||||
| Spends enough time at the bedside with me | 4.37 (0.79) | 4.38 (0.80) | 0.01 (−0.23–0.24) | 1.58 (0.85–2.94) | 4.24 (0.93) | 4.60 (0.63) | 0.36 (0.13–0.59)** | 2.13 (1.21–3.74)** |
| Checks to be sure I understand everything | 4.26 (1.00) | 4.50 (0.84) | 0.24 (−0.03–0.51) | 1.80 (0.87–3.72) | 4.29 (0.98) | 4.78 (0.75) | 0.49 (0.24–0.74)*** | 5.12 (2.24–11.72)*** |
| Lets me talk without interrupting | 4.52 (0.94) | 4.77 (0.52) | 0.25 (0.02–0.47)* | 4.59 (0.81) | 4.77 (0.75) | 0.18 (−0.04–0.40) | 2.55 (1.14–5.71)* | |
| Is in a rush when he/she is with me | 1.73 (1.06) | 1.18 (0.57) | −0.56 (−0.81–−0.31)*** | 0.19 (0.08–0.47)*** | 1.95 (1.18) | 1.36 (0.92) | −0.60 (−0.90–−0.29)*** | 0.24 (0.12–0.48)*** |
| Upper-year residents (PGY2–3s) | (N=70) | (N=60) | (N=96) | (N=52) | ||||
| Physician sits: | ||||||||
| When he/she visits my bedside | 1.71 (1.08) | 2.20 (1.30) | 0.49 (0.07–0.90)* | 2.30 (1.15–4.60)* | 2.08 (1.42) | 3.31 (1.41) | 1.22 (0.74–1.71)*** | 4.66 (2.36–9.23)*** |
| When giving me important updates or talking with me and my family | 2.06 (1.57) | 2.27 (1.26) | 0.21 (−0.29–0.71) | 1.87 (0.92–3.80) | 2.11 (1.42) | 3.52 (1.43) | 1.40 (0.92–1.89)*** | 5.53 (2.87–10.66)*** |
| Other communication behaviors: | ||||||||
| Spends enough time at the bedside with me | 4.29 (0.97) | 4.25 (0.88) | −0.04 (−0.35–0.29) | 0.82 (0.40–1.68) | 4.20 (0.96) | 4.81 (0.44) | 0.61 (0.33–0.89)*** | 5.20 (2.26–11.95)*** |
| Checks to be sure I understand everything | 4.36 (1.08) | 4.57 (0.74) | 0.21 (−0.12–0.54) | 1.28 (0.61–2.71) | 4.41 (0.83) | 4.81 (0.44) | 0.40 (0.16–0.65)** | 3.49 (1.70–7.18)** |
| Lets me talk without interrupting | 4.57 (1.00) | 4.83 (0.42) | 0.26 (−0.01–0.54) | 1.78 (0.70–4.51) | 4.76 (0.58) | 4.71 (0.70) | −0.05 (−0.26–0.16) | 0.82 (0.40–1.71) |
| Is in a rush when he/she is with me | 1.64 (1.10) | 1.22 (0.61) | −0.43 (−0.74–−0.11)** | 0.34 (0.13–0.89)* | 1.82 (1.09) | 1.13 (0.49) | −0.69 (−1.00–−0.38)*** | 0.11 (0.03–0.38)** |
95% confidence intervals (CIs) calculated assuming normal distribution and without adjustment for multiple observations among individual residents
Odds ratios (OR) from ordered logistic regression models adjusting for multiple observations among individual residents. OR compare the odds of being in a higher category versus all lower categories combined when comparing the two groups.
p<0.05,
p<0.01,
p<0.001
Table 3:
Difference in Difference Analysis of Patient-Reported Resident Physician Communication Behaviors on Education-Only Units vs Units Receiving Chairs
| Difference in Education-Only Unit (95% CI) | Difference in Units with Added Chairs (95% CI) | Difference in Differences (95% CI) | Odds ratio of a greater increase within survey responses (95% CI) | |
|---|---|---|---|---|
| All residents | ||||
| Physician sits: | ||||
| When he/she visits my bedside | 0.48 (0.21–0.75) | 1.16 (0.87–1.45) | 0.68 (0.29–1.08)** | 2.05 (1.10–3.81)* |
| When giving me important updates or talking with me and my family | 0.31 (0.01–0.61) | 1.31 (1.01–1.61) | 1.00 (0.57–1.42)*** | 2.43 (1.32–4.49)** |
| Other communication behaviors: | ||||
| Spends enough time at the bedside with me | −0.01 (−0.20–0.18) | 0.46 (0.28–0.63) | 0.46 (0.21–0.72)** | 3.18 (1.58–6.37)** |
| Checks to be sure I understand everything | 0.23 (0.02–0.43) | 0.44 (0.27–0.62) | 0.22 (−0.06–0.49) | 3.04 (1.44–6.39)** |
| Lets me talk without interrupting | 0.25 (0.08–0.43) | 0.08 (−0.08–0.23) | −0.17 (−0.41–0.06) | 0.90 (0.39–2.04) |
| Is in a rush when he/she is with me | −0.50 (−0.70–−0.31) | −0.62 (−0.83–−0.40) | −0.11 (−0.40–0.18) | 0.79 (0.32–1.94) |
| First-year residents (PGY1s) | ||||
| Physician sits: | ||||
| When he/she visits my bedside | 0.47 (0.11–0.83) | 1.14 (0.78–1.51) | 0.67 (0.16–1.19)* | 2.00 (0.94–4.26) |
| When giving me important updates or talking with me and my family | 0.38 (0.00–0.77) | 1.28 (0.90–1.67) | 0.90 (0.36–1.44)** | 2.23 (1.00–4.99) |
| Other communication behaviors: | ||||
| Spends enough time at the bedside with me | 0.01 (−0.23–0.24) | 0.36 (0.13–0.59) | 0.35 (0.03–0.68)* | 2.21 (0.93–5.25) |
| Checks to be sure I understand everything | 0.24 (−0.03–0.51) | 0.49 (0.24–0.74) | 0.24 (−0.12–0.61) | 3.47 (1.26–9.56)* |
| Lets me talk without interrupting | 0.25 (0.02–0.47) | 0.18 (−0.04–0.40) | −0.07 (−0.38–0.25) | 1.47 (0.50–4.35) |
| Is in a rush when he/she is with me | −0.56 (−0.81–−0.31) | −0.60 (−0.90–−0.29) | −0.04 (−0.43–0.35) | 1.22 (0.39–3.80) |
| Upper-year residents (PGY2–3s) | ||||
| Physician sits: | ||||
| When he/she visits my bedside | 0.49 (0.07–0.90) | 1.22 (0.74–1.71) | 0.74 (0.10–1.37)* | 2.26 (0.86–5.92) |
| When giving me important updates or talking with me and my family | 0.21 (−0.29–0.71) | 1.40 (0.92–1.89) | 1.20 (0.50–1.89)** | 2.98 (1.23–7.22)* |
| Other communication behaviors: | ||||
| Spends enough time at the bedside with me | −0.04 (−0.35–0.29) | 0.61 (0.33–0.89) | 0.65 (0.22–1.07)** | 6.35 (2.15–18.72)** |
| Checks to be sure I understand everything | 0.21 (−0.12–0.54) | 0.40 (0.16–0.65) | 0.19 (−0.22–0.60) | 2.59 (0.92–7.34) |
| Lets me talk without interrupting | 0.26 (−0.01–0.54) | −0.05 (−0.26–0.16) | −0.31 (−0.66–0.04) | 0.46 (0.14–1.52) |
| Is in a rush when he/she is with me | −0.43 (−0.74–−0.11) | −0.69 (−1.00–−0.38) | −0.26 (−0.71–0.18) | 0.31 (0.06–1.52) |
95% confidence intervals (CIs) for differences in means calculated assuming normal distribution and without adjustment for multiple observations among individual residents
Odds ratios are for the interaction term from ordered logistic regression models adjusting for multiple observations among individual residents.
p<0.05,
p<0.01,
p<0.001
Compared to baseline, units with added chairs saw improvements in perceptions of three other communication behaviors (Table 2): spending enough time (OR 2.85 [1.78–4.55]), checking for understanding (4.40 [2.48–7.79]), and being in a rush (0.20 [0.11–0.36]). Education-only units saw improvements in two behaviors: not interrupting (1.77 [1.02–3.10]) and being in a rush (0.25 [0.13–0.47]). In the difference in differences analysis (Table 3), adding chairs was associated with greater improvements in spending enough time (OR 3.18 [1.58–6.37]) and checking for understanding (3.04 [1.44–6.39]). Similar trends were apparent when stratified by level of training, but not all comparisons reached statistical significance.
Patient-reported resident behaviors in the 3 month-period immediately following the educational activities compared to the latter 3 month-period are shown in Table 4. In general, improvements in sitting and in other communication behaviors were sustained or increased in the latter 3-month period on both units receiving chairs and education-only units. On units receiving chairs, the odds of sitting increased significantly for all residents (2.56 [1.19–5.48]) and for the subgroup of PGY1 resident physicians (2.27 [1.04–4.97]), and on education-only units for PGY1 resident physicians (2.77 [1.16–6.62]).
Table 4:
Patient-Reported Resident Physician Communication Behaviors: Comparison of Immediate (October-December 21) versus Sustained (February-April 22) Periods Following Educational Activities
| Education-Only Units | Units with Added Chairs | |||||||
|---|---|---|---|---|---|---|---|---|
| Oct-Dec ‘21 Mean (SD) | Feb-April ‘22 Mean (SD) | Absolute differencea (95% CI) | Odds ratiob for change between time periods (95% CI) | Oct-Dec ‘21 Mean (SD) | Feb-April ‘22 Mean (SD) | Absolute differencea (95% CI) | Odds ratiob for change between time periods (95% CI) | |
| All residents | (N=67) | (N=83) | (N=62) | (N=80) | ||||
| Physician sits: | ||||||||
| When he/she visits my bedside | 2.16 (1.29) | 2.36 (1.17) | 0.20 (−0.20–0.60) | 1.53 (0.80–2.93) | 2.79 (1.49) | 3.53 (1.15) | 0.73 (0.30–1.17)** | 2.56 (1.19–5.48)* |
| When giving me important updates or talking with me and my family | 2.22 (1.35) | 2.33 (1.07) | 0.10 (−0.29–0.49) | 1.40 (0.73–2.68) | 2.77 (1.44) | 3.79 (1.21) | 1.01 (0.57–1.45)*** | 3.74 (1.80–7.77)*** |
| Other communication behaviors: | ||||||||
| Spends enough time at the bedside with me | 4.19 (0.91) | 4.43 (0.75) | 0.24 (−0.03–0.51) | 1.70 (0.85–3.38) | 4.55 (0.67) | 4.78 (0.48) | 0.23 (0.04–0.42)* | 2.38 (1.12–5.04)* |
| Checks to be sure I understand everything | 4.61 (0.67) | 4.46 (0.89) | −0.15 (−0.41–0.11) | 0.70 (0.32–1.50) | 4.76 (0.69) | 4.81 (0.62) | 0.05 (−0.16–0.27) | 1.47 (0.59–3.68) |
| Lets me talk without interrupting | 4.75 (0.59) | 4.83 (0.38) | 0.09 (−0.07–0.24) | 1.25 (0.55–2.82) | 4.71 (0.86) | 4.78 (0.62) | 0.07 (−0.18–0.31) | 0.92 (0.34–2.48) |
| Is in a rush when he/she is with me | 1.30 (0.67) | 1.11 (0.49) | −0.19 (−0.38–0.00) | 0.25 (0.07–0.83)* | 1.37 (0.96) | 1.20 (0.62) | −0.17 (−0.44–0.09) | 0.58 (0.20–1.66) |
| First-year residents (PGY1s) | (N=39) | (N=51) | (N=42) | (N=48) | ||||
| Physician sits: | ||||||||
| When he/she visits my bedside | 2.00 (1.15) | 2.57 (1.15) | 0.57 (0.08–1.06)* | 2.77 (1.16–6.62)* | 2.81 (1.47) | 3.44 (1.13) | 0.63 (0.08–1.17)* | 2.27 (1.04–4.97)* |
| When giving me important updates or talking with me and my family | 2.08 (1.22) | 2.45 (1.10) | 0.37 (−0.11–0.86) | 2.03 (0.90–4.61) | 2.71 (1.44) | 3.71 (1.17) | 0.99 (0.45–1.54)*** | 3.83 (1.79–8.18)** |
| Other communication behaviors: | ||||||||
| Spends enough time at the bedside with me | 4.18 (0.97) | 4.53 (0.61) | 0.35 (0.02–0.68) | 2.05 (0.92–4.57) | 4.45 (0.71) | 4.73 (0.54) | 0.28 (0.02–0.54)* | 2.70 (1.04–7.01)* |
| Checks to be sure I understand everything | 4.64 (0.67) | 4.39 (0.94) | −0.25 (−0.60–0.10) | 0.52 (0.18–1.53) | 4.76 (0.79) | 4.79 (0.71) | 0.03 (−0.29–0.34) | 1.15 (0.30–4.43) |
| Lets me talk without interrupting | 4.72 (0.65) | 4.80 (0.40) | 0.09 (−0.13–0.31) | 1.13 (0.44–2.90) | 4.81 (0.77) | 4.73 (0.74) | −0.08 (−0.40–0.24) | 0.40 (0.11–1.52) |
| Is in a rush when he/she is with me | 1.26 (0.68) | 1.12 (0.48) | −0.14 (−0.38–0.10) | 0.46 (0.12–1.75) | 1.50 (1.11) | 1.23 (0.69) | −0.27 (−0.65–0.11) | 0.45 (0.14–1.49) |
| Upper-year residents (PGY2–3s) | (N=28) | (N=32) | (N=20) | (N=32) | ||||
| Physician sits: | ||||||||
| When he/she visits my bedside | 2.39 (1.45) | 2.03 (1.15) | −0.36 (−1.03–0.31) | 0.68 (0.28–1.68) | 2.75 (1.59) | 3.66 (1.18) | 0.91 (0.13–1.68)* | 3.17 (0.59–16.89) |
| When giving me important updates or talking with me and my family | 2.43 (1.50) | 2.13 (1.01) | −0.30 (−0.96–0.35) | 0.83 (0.31–2.23) | 2.90 (1.48) | 3.91 (1.28) | 1.01 (0.23–1.79)* | 3.66 (0.81–16.43) |
| Other communication behaviors: | ||||||||
| Spends enough time at the bedside with me | 4.21 (0.83) | 4.28 (0.92) | 0.07 (–0.39 – 0.52) | 1.28 (0.39–4.23) | 4.75 (0.55) | 4.84 (0.37) | 0.09 (−0.16–0.35) | 1.43 (0.32–6.41) |
| Checks to be sure I understand everything | 4.57 (0.69) | 4.56 (0.80) | −0.01 (−0.40 – 0.38) | 1.07 (0.33–3.48) | 4.75 (0.44) | 4.84 (0.45) | 0.09 (−0.16–0.35) | 2.20 (0.68–7.08) |
| Lets me talk without interrupting | 4.79 (0.50) | 4.88 (0.34) | 0.09 (−0.13 – 0.31) | 1.57 (0.34–7.26) | 4.50 (1.00) | 4.84 (0.37) | 0.34 (−0.05–0.73) | 2.53 (0.66–9.75) |
| Is in a rush when he/she is with me | 1.36 (0.68) | 1.09 (0.53) | −0.26 (−0.58–0.05) | 0.09 (0.01–1.05) | 1.10 (0.45) | 1.16 (0.51) | 0.06 (−0.22–0.34) | 1.92 (0.17–21.49) |
95% confidence intervals (CIs) calculated assuming normal distribution and without adjustment for multiple observations among individual residents
Odds ratios (OR) from ordered logistic regression models adjusting for multiple observations among individual residents. OR compare the odds of being in a higher category versus all lower categories combined when comparing the two groups.
p<0.05,
p<0.01,
p<0.001
Discussion
Placing easily visible, wall-mounted folding chairs inside patient rooms was associated with an increase in patient-reported frequency of sitting at the bedside by internal medicine residents beyond the impact of education alone. Compared to the effects seen with education alone, adding folding chairs was associated with greater improvements in patient perceptions of spending adequate time at the bedside and of checking for understanding.
At the outset, we hypothesized that the folding chairs might increase sitting through either or both of two mechanisms. First, the folding chairs might address the previously identified barrier of a place to sit. Patient rooms have always had a single, non-folding chair; however, these are often occupied by patient visitors or are filled with room supplies. Second, folding chairs were placed in line of sight for clinicians entering patient rooms, with the intent to be a behavioral nudge. Our data do not provide information about the relative effect of either mechanism. Future studies, including resident surveys, might be designed to assess this. Some residents continued to sit infrequently despite the presence of folding chairs; future work should also examine whether this reflects other barriers to sitting or disagreement with the value of sitting.
This is one of the first studies to examine the impact of placing chairs in healthcare settings. Orloski found that an emergency-department based intervention consisting of folding chairs and an educational video increased the odds of clinician sitting by 30%.17 Several studies have examined the impact of randomizing clinicians to sit versus stand but have not addressed the effect of placing folding chairs in a convenient location.10,11,18,19 Folding chairs are a relatively simple intervention, and while an upfront investment was required, our chairs have been durable, with none needing to be replaced in 4 years.
The relative improvement in perceptions of other communication behaviors on units with chairs supports the concept that sitting augments patient perceptions of clinician communication. In Orloski’s study, sitting was associated with improvements in multiple measures of patient experience. We acknowledge that the four other communication behaviors assessed in this study may be inter-related; however, even if we adjust our alpha by a factor of 4 to 0.0125, the increased odds in spending enough time at the bedside and checking for understanding would still be statistically significant. We did not observe a statistically significant improvement in the frequency of interrupting patients or resident physicians seeming to be a rush on units receiving chairs, which may reflect that these behaviors are infrequent at baseline. Moreover, our study was not powered to assess for these differences. Among hospitalists, McCaffrey found that a multi-faceted approach including encouraging physicians to sit was associated with increased patient satisfaction with communication.20 In some contrast, Donovan found that randomizing individuals to sit versus stand on teaching rounds did not impact patient experience; however, this work only examined large group teaching rounds with one team member randomized to sit.13 Further work is needed to understand how sitting may be optimized in various types of inpatient interactions (e.g. teaching rounding versus goals-of-care conversations) and the clinical implications of sitting, including quantifying how much additional sitting may be required to impact patient perception of provider communication.
We observed improvements in sitting frequency and some other communication behaviors during the post-intervention period on education-only units. Potential reasons include our educational activities, a general trend since baseline, or an effect from being later in the academic year. We are not certain why sitting improved among some groups of residents in the later 3 months of the post-intervention period compared to the initial 3 post-intervention months. Possible reasons (again) include an effect from being later in the academic year, and/or gradually growing accustomed to chairs and sitting.
Our work is limited by not being able to assess several important factors that could influence patient perceptions and/or help to generalize findings. To protect privacy, we did not collect information on individual patient or clinician demographics. Although we can only compare the overall populations from which we recruited patients to do surveys, we used the same methodology for recruitment in both instances. However, factors such as shared patient and clinician social identities (e.g., racial concordance) and culture context may influence perceptions of patient-clinician communication and should be considered in subsequent work.21–23 Furthermore, our study did not include patients with limited English proficiency, who are an important demographic given persistent patient-clinician communication inequities which undermine efforts to promote health equity.24 Future work examining sitting can and should address these factors.
Our work has other limitations. First, due to the COVID-19 pandemic, we paused data collection and conducted educational activities twice. However, this should not have affected our post-COVID comparison between units with and without chairs. Second, this is a single site study of internal medicine patients and residents, which may limit generalizability. Third, our findings may be subject to patient recall bias. Fourth, we relied on nursing assessments to inform which patients we approached and only surveyed patients who could recognize resident physicians, which may introduce bias.25 Finally, our analyses stratified by level of resident (PGY1 vs. PGY2–3) were significantly underpowered (group size range 52 – 104, Table 2). We nevertheless present the stratified analyses because of inherent interest in the different role types.
Our work also has notable strengths. First, we arbitrarily picked study units to receive chairs, and both resident and patient assignments to these units is relatively random. Second, we assessed patient-reported communication outcomes in order to minimize bias introduced by having observers present. Finally, a relatively high percentage of approached patients in both the pre-intervention period (77%) and post-intervention period (95%) agreed to surveys, which may reduce nonresponse bias. We do not know why this rate was appreciably higher in the post-intervention period. We did not ask patients why they declined. Surveyors (who were different individuals) used the same basic language in both periods. Given a different set of other work obligations, surveyors in the pre-intervention period may have felt more time pressure to return to other tasks and may have spent less time explaining surveying. We do not believe this would have had a differential effect according to the presence of a folding chair in the room.
Conclusions
Wall-mounted folding chairs placed inside the entrance of patient rooms were associated with increased patient-reported sitting by internal medicine residents and with improvements in other patient-reported communication behaviors. Placing folding chairs in patient rooms, introduced with resident physician education, may represent a simple and durable way to improve patient-clinician communication; evaluations in other clinical settings are warranted.
Supplementary Material
Acknowledgements
Contributors:
None.
Sources of funding and support:
Dr. Golden’s work is supported by the National Institute of Health (1K23AG081458-01). The findings and conclusions in this paper are solely those of the authors and do not necessarily represent the views of the National Institute of Health.
Previous presentation of the information reported in the manuscript:
A preliminary version of these findings was presented at Society of Hospital Medicine Converge 2023 (March 26–29th, Austin TX).
Conflicts of Interest Statement:
Dr. Golden’s work is supported by the National Institute of Health (1K23AG081458-01). The Johns Hopkins University Institutional Review Board considered this work a quality improvement activity (IRB00184753). The authors will make data available from this project upon reasonable request to the corresponding author.
References
- 1.Ha JF, Longnecker N. Doctor-patient communication: a review. Ochsner J 2010;10(1):38–43. (https://www.ncbi.nlm.nih.gov/pubmed/21603354). [PMC free article] [PubMed] [Google Scholar]
- 2.Lie HC, Juvet LK, Street RL Jr., et al. Effects of Physicians’ Information Giving on Patient Outcomes: a Systematic Review. J Gen Intern Med 2022;37(3):651–663. DOI: 10.1007/s11606-021-07044-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Stewart MA. Effective physician-patient communication and health outcomes: a review. CMAJ 1995;152(9):1423–33. (https://www.ncbi.nlm.nih.gov/pubmed/7728691). [PMC free article] [PubMed] [Google Scholar]
- 4.Boissy A, Windover AK, Bokar D, et al. Communication Skills Training for Physicians Improves Patient Satisfaction. J Gen Intern Med 2016;31(7):755–61. DOI: 10.1007/s11606-016-3597-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ruiz-Moral R, Perez Rodriguez E, Perula de Torres LA, de la Torre J. Physician-patient communication: a study on the observed behaviours of specialty physicians and the ways their patients perceive them. Patient Educ Couns 2006;64(1–3):242–8. DOI: 10.1016/j.pec.2006.02.010. [DOI] [PubMed] [Google Scholar]
- 6.Moslehpour M, Shalehah A, Rahman FF, Lin KH. The Effect of Physician Communication on Inpatient Satisfaction. Healthcare (Basel) 2022;10(3). DOI: 10.3390/healthcare10030463. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Michtalik HJ, Yeh HC, Pronovost PJ, Brotman DJ. Impact of attending physician workload on patient care: a survey of hospitalists. JAMA Intern Med 2013;173(5):375–7. DOI: 10.1001/jamainternmed.2013.1864. [DOI] [PubMed] [Google Scholar]
- 8.Golden BP, Tackett S, Kobayashi K, et al. Sitting at the Bedside: Patient and Internal Medicine Trainee Perceptions. J Gen Intern Med 2022. DOI: 10.1007/s11606-021-07231-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Tackett S, Tad-y D, Rios R, Kisuule F, Wright S. Appraising the practice of etiquette-based medicine in the inpatient setting. J Gen Intern Med 2013;28(7):908–13. DOI: 10.1007/s11606-012-2328-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Merel SE, McKinney CM, Ufkes P, Kwan AC, White AA. Sitting at patients’ bedsides may improve patients’ perceptions of physician communication skills. J Hosp Med 2016;11(12):865–868. DOI: 10.1002/jhm.2634. [DOI] [PubMed] [Google Scholar]
- 11.Swayden KJ, Anderson KK, Connelly LM, Moran JS, McMahon JK, Arnold PM. Effect of sitting vs. standing on perception of provider time at bedside: a pilot study. Patient Educ Couns 2012;86(2):166–71. DOI: 10.1016/j.pec.2011.05.024. [DOI] [PubMed] [Google Scholar]
- 12.Strasser F, Palmer JL, Willey J, et al. Impact of physician sitting versus standing during inpatient oncology consultations: patients’ preference and perception of compassion and duration. A randomized controlled trial. J Pain Symptom Manage 2005;29(5):489–97. DOI: 10.1016/j.jpainsymman.2004.08.011. [DOI] [PubMed] [Google Scholar]
- 13.Bruera E, Palmer JL, Pace E, et al. A randomized, controlled trial of physician postures when breaking bad news to cancer patients. Palliat Med 2007;21(6):501–5. DOI: 10.1177/0269216307081184. [DOI] [PubMed] [Google Scholar]
- 14.Block L, Hutzler L, Habicht R, et al. Do internal medicine interns practice etiquette-based communication? A critical look at the inpatient encounter. J Hosp Med 2013;8(11):631–4. DOI: 10.1002/jhm.2092. [DOI] [PubMed] [Google Scholar]
- 15.Ferranti DE, Makoul G, Forth VE, Rauworth J, Lee J, Williams MV. Assessing patient perceptions of hospitalist communication skills using the Communication Assessment Tool (CAT). J Hosp Med 2010;5(9):522–7. DOI: 10.1002/jhm.787. [DOI] [PubMed] [Google Scholar]
- 16.Velez VJ, Kaw R, Hu B, et al. Do HCAHPS Doctor Communication Scores Reflect the Communication Skills of the Attending on Record? A Cautionary Tale from a Tertiary-Care Medical Service. J Hosp Med 2017;12(6):421–427. DOI: 10.12788/jhm.2743. [DOI] [PubMed] [Google Scholar]
- 17.Orloski CJ, Tabakin ER, Shofer FS, Myers JS, Mills AM. Grab a Seat! Nudging Providers to Sit Improves the Patient Experience in the Emergency Department. J Patient Exp 2019;6(2):110–116. DOI: 10.1177/2374373518778862. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Osborn R, Grossman M, Berkwitt A. The Effect of Sitting Versus Standing on Family Perceptions of Family-Centered Rounds. Hosp Pediatr 2021;11(11):e313–e316. DOI: 10.1542/hpeds.2021-005999. [DOI] [PubMed] [Google Scholar]
- 19.Donovan AK, Spagnoletti C, Rothenberger S, Corbelli J. The impact of residents sitting at the bedside on patient satisfaction during team rounds. Patient Educ Couns 2020;103(6):1252–1254. DOI: 10.1016/j.pec.2019.12.013. [DOI] [PubMed] [Google Scholar]
- 20.McCaffrey R, Hale D, Kunupakaphun S, Kaufman L, Eamranond P. A Multifaceted Approach to Improve Physician Communication Scores. J Patient Exp 2020;7(4):522–526. DOI: 10.1177/2374373519860041. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Shen MJ, Peterson EB, Costas-Muniz R, et al. The Effects of Race and Racial Concordance on Patient-Physician Communication: A Systematic Review of the Literature. J Racial Ethn Health Disparities 2018;5(1):117–140. DOI: 10.1007/s40615-017-0350-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Thornton RL, Powe NR, Roter D, Cooper LA. Patient-physician social concordance, medical visit communication and patients’ perceptions of health care quality. Patient Educ Couns 2011;85(3):e201–8. DOI: 10.1016/j.pec.2011.07.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Gupta A, Madhavapeddi S, Das A, Harris S, Naina H. Physician Posture at the Bedside: A Study of African-American and Hispanic Patient Preferences at a Teaching Hospital. J Med Pract Manage 2015;31(3):144–6. (https://www.ncbi.nlm.nih.gov/pubmed/26856020). [PubMed] [Google Scholar]
- 24.Berdahl TA, Kirby JB. Patient-Provider Communication Disparities by Limited English Proficiency (LEP): Trends from the US Medical Expenditure Panel Survey, 2006–2015. J Gen Intern Med 2019;34(8):1434–1440. DOI: 10.1007/s11606-018-4757-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Francis JJ, Pankratz VS, Huddleston JM. Patient satisfaction associated with correct identification of physician’s photographs. Mayo Clin Proc 2001;76(6):604–8. DOI: 10.4065/76.6.604. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.