A Comparison of Nonobese Versus Obese Emergency Department Patient Satisfaction Scores Utilizing Standard U.S. Hospital Survey Query Methodology.

BACKGROUND: Prior research reveals that overweight patients have higher emergency department (ED) utilization rates, longer length of stay, and face increased misdiagnosis risk.
OBJECTIVE: The objective of this study was to evaluate the association between obesity and ED patient satisfaction.
METHODS: This study was a cross-sectional study. A convenience sample of inner-city ED patients completed a written survey, then rated overall satisfaction with ED care (10-point scale), and rated components of satisfaction (4-point scale; never to always). Body mass index (BMI) was calculated using triage records (obesity = BMI >30).
RESULTS: Five hundred and sixty-four patients were included in the study group (50.5%: obese, 55.4%: female, mean age: 43.2 ± 25.4 years). With respect to overall visit satisfaction (rating 8 or greater on 10-point scale), bivariate analysis revealed no differences between nonobese versus obese patients (74.6% vs. 73.9%; P = 0.85). There were no significant differences for score of 4 (always) for components of ED satisfaction: physician courtesy (87.9% vs. 90.4%; P = 0.34), nurse courtesy/respect (89.2% vs. 88.7%; P = 0.87), doctor listened (85.4% vs. 87.1%; P = 0.5), doctor explained (80.2% vs. 85.0%; P = 0.14), and recommend to friend (72.5% vs. 81.1%; P = 0.02). Within our multivariate model, obesity was not associated with overall satisfaction (scores of 8 or greater) (P = 0.97; odds ratio = 0.99 [95% confidence interval = 0.65-1.5]).
CONCLUSIONS: Despite research that suggests that overweight patients have characteristics of their ED visit that might increase dissatisfaction risk, we found no difference in satisfaction scores between nonobese and obese patients.

INTRODUCTION

Hospital administrators and regulators are ever increasing their focus on patient satisfaction. In recent years, the Centers for Medicare and Medicaid Services have introduced the Hospital Inpatient Value-Based Purchasing Program, which links reimbursements to patient satisfaction surveys.[1] Within this program, the federal government has allocated $850 million in reduced Medicare reimbursement for facilities with lower patient satisfaction scores.[2] When faced with the potential for lost revenues, providers and hospital systems increasingly in need of data that provides a profile of patient subgroups that are at highest risk to provide poor scores for the types of standardized questions that are posed on these survey instruments.

For a variety of reasons, one such subgroup seemingly at risk for poor satisfaction scores is patients with obesity (body mass index [BMI] >30). Currently, more than one-third of the US adults and 17% of the youth are obese.[3] Further, a total of 68.8% of adults are considered to be overweight, which means a BMI of >25.[4] We commonly know that obesity is associated with increased risks for health complications (e.g. heart disease and hypertension), many of which are addressed in the emergency department (ED).

Previously, investigators have reported that obese patients have elements of their ED visit care that would likely contribute to dissatisfaction, including increased difficulty with procedures such as repeat intravenous (IV) cannulation attempts, more frequent testing, and longer ED length of stay.[56789] Surprisingly, despite the vast amount of research on factors related to patient satisfaction, our Medline search revealed a paucity of data that specifically address the potential impact of obesity on these scores. The purpose of our study was to compare satisfaction scores for obese versus nonobese patients in the ED utilizing the standard US hospital survey query methodology.

METHODS

Study design

This was a prospective, cross-sectional study.

Setting

The study was conducted at CHRISTUS Spohn Memorial Hospital in Corpus Christi, TX. The facility is a major teaching affiliate of Texas A and M medical school, a level-two trauma center, and serves an inner-city population. The annual ED census is 45,000 patients. The CHRISTUS Health Institutional Review Board approved the study before the initiation of data collection.

Population

We enrolled a convenience sample of medically stable, consenting adult patients aged >18 years who presented to the ED. Patients were excluded for any of the following reasons: refusal to provide consent, pregnancy, and inability to complete the questionnaire due to clinical instability, severe pain, or disorientation as determined by a study physician. Patients that were not English speaking were also excluded from the study.

Study protocol

Hemodynamically, stable, oriented, consenting patients at an inner-city, academic ED were consecutively enrolled at hours in which trained research associates were available to assist with data collection (convenience sample) from August 2015 through February 2016. The enrolled patients completed a structured, written survey providing demographic, chief complaint information. Subsequently, the patients provided structured answers to questions regarding components of patient satisfaction as well as overall satisfaction with the visit [Appendix 1].

Statistical analysis

Data were entered into the Excel for Windows (Microsoft Corporation, Redmond, WA) and transported into STATA software (STATA, College Station, Texas). Components of satisfaction were assessed with a 4-point scale (never to always). Overall satisfaction with ED care received was rated on a 10-point scale. BMI was calculated by the research assistant using height and weight obtained in triage, and obesity was defined as a BMI >30. Categorical data were presented as frequency of occurrence and analyzed by Chi-square test; continuous data were presented as mean ± SD and analyzed by t-tests. Multivariate logistic regression analysis was performed to control for confounding. The primary outcome parameter was to compare satisfaction scores between obese and nonobese patients. Secondary outcome parameters were to compare the relationship between other patient characteristics and satisfaction scores.

RESULTS

We considered 770 patients for participation; 564 completed the survey (study group), 104 refused to participate, and the remaining patients met exclusion criteria. The number of patients for which we had completed data for a given characteristic included: weight (564), gender (560), race (564), income (549), education (561), insurance (542), and chief complaint (557). Table 1 summarizes the demographics, which reveals 50.5%: obese, 55.4%: female, mean age: 43.2 ± 25.4 years. Within the study group, 70.9% had an annual income <$20,000, 29.6% had less than high school graduation/education, and 11% had private insurance. Most common chief complaints, also shown in Table 1, included chest pain (8.8%), back pain (10.4%), abdominal pain (11.5%), and injury (15.2%).

Table 1

Demographics

Total enrolled	564
Percentage obese	50.5 (285/564)
Percentage females	55.4 (310/560)
Percentage annual income <$20,000	70.9 (389/549)
Mean age	43.2 ± 25.4 (564)
Private insurance	11% (60/542)
Chest pain	8.8% (49/557)
Back pain	10.4% (58/557)
Abdominal pain	11.5% (64/557)
Injury	15.2% (85/557)

With respect to the primary outcome measure of overall satisfaction with the ED visit (rating 8 or greater), bivariate analysis revealed no differences between nonobese and obese patient scores (74.6% vs. 73.9%; P = 0.85), which is summarized in Table 2. Likewise, Table 3 shows that there were no significant differences for nonobese versus obese patients for % score of 4 (always) for the following components of ED satisfaction: physician courtesy and respect (87.9% vs. 90.4%; P = 0.34), nurse staff courtesy and respect (89.2% vs. 88.7%; P = 0.87), doctor listened (85.4% vs. 87.1%; P = 0.5), doctor took time to explain (80.2% vs. 85.0%; P = 0.14), and % who would recommend to friends (72.5% vs. 81.1%; P = 0.02). Table 4 summarizes our multivariate model, which revealed that obesity was not associated with differences in overall satisfaction (scores of 8 or greater) (P = 0.97; odds ratio [OR] = 0.99 [0.65–1.5]), whereas increasing age was the only variable associated with high satisfaction (P = 0.010; OR = 1.02 [1.00–1.03]). Multivariate analyses for each component of satisfaction, respectively, did not reveal significant associations between obesity and good scores. These findings are summarized in Tables 5-10.

Table 2

Overall satisfaction (rating 8 or greater)

	Obese	Nonobese	P
Overall satisfaction (%)	74.6 (206/276)	73.9 (210/284)	0.85

Table 3

Percent receiving score of 4 (always) for components of the emergency department satisfaction

	Obese	Nonobese	P
Physician courtesy and treatment	90.4 (255/282)	87.3 (241/274)	0.34
Nurse staff courtesy and treatment	88.7 (252/284)	89.2 (247/277)	0.87
Doctor listened	87.1 (245/281)	85.4 (234/274)	0.5
Doctor explaining treatment	85.0 (238/280)	80.2 (219/273)	0.14
Percentage recommend to friend	81.1 (228/281)	72.5 (201/277)	0.02

Table 4

Multivariate analysis (n=522)

Overall satisfaction	P	OR
Obesity	0.97	0.99 (0.65–1.5)
Increasing age	0.010	1.02 (1.00–1.03)

OR: Odds ratio

Table 5

Overall satisfaction level of visit (n=521)

Characteristics	P	OR	95% CI
Obese	0.973	0.992	0.653–1.508
Female	0.997	0.999	0.649–1.538
Increasing age	0.010	1.020	1.004–1.036
Hispanic race	0.885	0.965	0.598–1.558
Income of less $20,000	0.179	0.685	0.394–1.189
Private insurance	0.404	0.695	0.295–1.633

CI: Confidence interval, OR: Odds ratio

Table 10

Would you recommend this hospital to your friends and family? (n=523)

Characteristics	P	OR	95% CI
Obese	0.080	1.470	0.955–2.263
Female	0.054	1.537	0.992–2.382
Increasing age	0.205	1.010	0.994–1.026
Hispanic race	0.429	0.820	0.501–1.340
Income of less $20,000	0.841	0.943	0.529–1.677
Private insurance	0.025	2.538	1.125–5.722

CI: Confidence interval, OR: Odds ratio

DISCUSSION

Obese patients represent a challenging patient population to evaluate from a variety of perspectives. A study by Ngui et al.[10] investigated the number of IV cannulation attempts, liver function tests, cardiac enzyme tests, and abdominal X-rays obtained for obese versus nonobese patients in the ED and found that all were increased for obese patients. Kam et al.[5] performed a similar study and demonstrated that BMI most strongly correlated with difficulty finding anatomical landmarks, venous pressure measurement, physical examination, patient positioning, and procedures including venipuncture.

When one takes all of these factors into account, it would suggest some risk of prolonged and more costly ED visits for obese patients; however, the data are somewhat contradictory. Peitz et al.[6] reported that when compared with nonobese ED patients, obese patients with chest pain and dyspnea did have significantly increased costs and length of stay. However, two other studies that evaluated the care of ED patients with abdominal found that obese patients were not at risk for longer disposition times, more frequent diagnostic testing/resource utilization, and increased length of stay.[78]

To date, Baskerville and Moore conducted the only study that has examined the relationship of body mass to length of stay for ED patients that included patients with all chief complaints.[9] This retrospective cohort included 102 morbidly obese (BMI >40) and 195 normal weight or mildly obese patients. The authors found that the mean length of stay for patients with BMI >40 was 101-min longer than that which was observed for those patients with BMI <35 ([95% CI, 55–146]; P < 0.0001).

With some evidence supporting an increased potential for higher cost and length of stay, we were surprised to observe within our study group that there were no significant differences between obese and nonobese patients for an overall satisfaction with visit score of ≥8 (74.6% vs. 73.9%; P = 0.85). Similarly, there were no significant differences in the component of satisfaction scores for obese and nonobese patients.

We believe that our study is novel in its design to evaluate the relationship between obesity and satisfaction scores in the ED setting. Previously, several studies have examined this question in non-ED settings and in some cases found that obese patients were more satisfied with their visit to health facilities.[1112131415161718192021] Bottone et al. evaluated a large cohort of patients (>50,000) by survey about their general health-care experiences and number of health-care facility encounters without specifically isolating ED visits.[11] They concluded that obese individuals were more satisfied with their care and had better experiences. Fong et al. found similar results of higher satisfaction for obese patients in a survey of 9914 adult patients.[12] Based on our observations from the current study data, we cannot recommend that physicians give any specific focus to concerns for patient satisfaction as it relates to the ED care of obese versus nonobese patients.

Limitations

Our study has several limitations that warrant discussion. Patients were surveyed in a nonconsecutive fashion over several months. We attempted to limit sampling bias by surveying patients on all shifts including nights and weekends. A second limitation may be the generalizability of our study. Our patient population is predominantly Hispanic and of lower educational and socioeconomic status. Although there might be a concern for language barriers impacting the bias of our sample, nearly all of the patients who visit our department speak English fluently, so few were likely excluded on that basis though we did not track such. Further, we attempted to control for such confounding through our multivariate regression models and generally found that few patient characteristics correlated with higher scores. We note that we did not control for other potential co-founders such as severity of illness, so there is the potential that the groups were not similar in this respect.

Another limitation of our investigation was the survey instrument itself and the time of administration. Our study instrument was significantly shorter than the actual Press Ganey questionnaire and most other surveys currently in use commercially. For proprietary/intellectual property reasons, our survey instrument cannot precisely match the one that patients receive from these firms. We have found in our setting that our patients of typically lower educational status have higher survey completion rates when the instruments are kept brief. That being noted, our survey questions are essentially paraphrasing of the majority of key components of the major surveys currently utilized in the US to evaluated EDs and their staff (e.g., physician taking the time to listen and experience with pain control). As scores for each component assessed on US hospital-type surveys are typically one question per component (i.e., components are not aggregate scores of several questions), our editing of survey length reflects a reduction in components reported rather than reducing the number of questions asked per component. While we did not perform sensitivity analysis, we believe that our questions well reflect in terms of the components utilized and scoring method the actual instruments upon which administrators and regulators assess for several elements of customer service. There is also a risk that our need to have a brief survey may led to the omission of potential cofounding data points such as severity of illness, though we excluded patients who were critically ill.

We also note that there are potential response differences based on the time of survey administration compared with standard commercial surveying. In our protocol, the survey was administered at the point of care late in the ED encounter. Surveying companies typically mail/e-mail questionnaires to patients multiple days to weeks after their ED visit. We are unsure how these differences in survey technique and timing may have impacted our results.

CONCLUSIONS

Despite research that suggests that overweight patients have characteristics of their ED visit that might increase risk for dissatisfaction, we found no difference between nonobese and obese patients for overall and component measures of satisfaction. More research should be performed to assess for patient groups at risk for decreased satisfaction.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

Table 6

Did the doctors treat you with courtesy and respect? (n=524)

Characteristics	P	OR	95% CI
Obese	0.396	1.291	0.716–2.329
Female	0.168	1.526	0.836–2.785
Increasing age	0.743	1.004	0.982–1.025
Hispanic race	0.078	0.507	0.237–1.080
Income of less $20,000	0.112	0.549	0.262–1.149
Private insurance	0.464	1.501	0.506–4.453

CI: Confidence interval, OR: Odds ratio

Table 7

Did the nurses treat you with courtesy and respect? (n=519)

Characteristics	P	OR	95% CI
Obese	0.947	0.980	0.552–1.739
Female	0.299	1.364	0.759–2.453
Increasing age	0.425	1.008	0.987–1.030
Hispanic race	0.393	0.741	0.373–1.472
Income of less $20,000	0.304	0.686	0.334–1.407
Private insurance	0.289	1.832	0.599–5.599

CI: Confidence interval, OR: Odds ratio

Table 8

How often did the doctors listen carefully to you? (n=520)

Characteristics	P	OR	95% CI
Obese	0.500	1.196	0.711–2.011
Female	0.690	0.896	0.523–1.535
Increasing age	0.796	0.997	0.979–1.015
Hispanic race	0.721	0.896	0.491–1.635
Income of less $20,000	0.346	0.721	0.364–1.424
Private insurance	0.038	2.739	1.058–7.086

CI: Confidence interval, OR: Odds ratio

Table 9

How often did doctors explain things in a way you could understand? (n=520)

Characteristics	P	OR	95% CI
Obese	0.178	1.385	0.862–2.227
Female	0.082	1.532	0.947–2.477
Increasing age	0.429	0.993	0.977–1.009
Hispanic race	0.150	0.665	0.381–1.159
Income of less $20,000	0.477	0.799	0.431–1.481
Private insurance	0.384	1.486	0.608–3.628

CI: Confidence interval, OR: Odds ratio