Do Hospitals Led by Physician CEOs Have Better HCAHPS Scores?

The chief executive officer (CEO) represents a position in healthcare that has oversight for the organization’s strategic focus, culture, and day-to-day operations. The Upper Echelon Theory (UET) suggests that organizations’ priorities are informed by the leadership of the CEOs, including their education, background, prior experiences, personal characteristics, and traits.(1) As more and more healthcare organizations seek the leadership of a physician CEO, it becomes important to understand the independent contribution that a physician CEO may have on the organization’s outcomes, including patient experience and satisfaction.

While many physician leaders accept the notion of clinical quality, the evolution of patient experience as a component of healthcare quality has required a cultural shift among physicians and physician leaders. In 2002, the Centers for Medicare & Medicaid Services (CMS) sought to increase quality transparency through public reporting.

Within the hospital environment, the formal evaluation of patient experience originated with the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS)(2) in 12 pre-defined ratings categories. The HCAHPS process remains central to value-based payment models today, with a 25 percent performance weight attributed to domains from the HCAHPS survey.(3) In addition to the financial benefits, health systems with best-practice patient experience demonstrate lower rates of hospital-acquired conditions, shorter lengths of stay, and fewer readmissions when compared with those that perform in lower percentiles.(4-7)

A hospital’s focus on patient experience seems legitimate, given the important contribution of patient experience scores to the clinical and financial well-being of the organization. While a few empirical studies and meta-analyses investigate the operative role of physicians on outcomes, these studies are equivocal in terms of their results and are dated.(8-13) As a result, we sought to compare whether hospitals led by physician CEOs had higher HCAHPS scores than those led by non-physician CEOs to determine if there were differences in patient experience based on the CEO’s personal experience of being a physician.

Method and Study Design

The data used in this study were derived from a link between two publicly available databases consisting of cross-sectional cohort data from the year 2016: the American Hospital Association (AHA) Annual Survey Database and the CMS HCAHPS survey data. The AHA Annual Survey Database is a comprehensive hospital database produced primarily from the AHA Annual Survey of Hospitals and reports information on hospital organizational structure, service lines, utilization, finances and staffing.(14) HCAHPS is a national standardized, publicly reported survey of patients’ perspectives of hospital care provided by CMS.(15)

Ninety-eight percent of the final sample consisted of acute care hospitals, all of which voluntarily participated in the 2016 AHA Annual Survey of Hospitals and HCAHPS survey. Figure 1 shows the breakdown of the linkage to create the final sample of hospital CEO data. We combined the 2016 AHA dataset with the 2016 HCAHPS dataset by using exact matches on hospital name, the state in which the hospital was located, and the hospital zip code.

Figure 1. Breakdown of 2016 hospital administrator data

Exclusions

Hospitals designated as Critical Access Hospitals, Rural Referral Centers, and those having 24 or fewer beds were excluded because they are not required to participate in CMS’ HCAHPS survey. In addition, any hospitals that did not submit data for HCAHPS were excluded.

Variables

Independent Variables from the AHA Database

Categorical and continuous variables were reported in the AHA Annual Survey of Hospitals, as well as the name of the hospital CEO for the year 2016. Categorical variables included in this study were the state and a predefined, eight-level variable for bed size. Continuous variables included in this study were total facility admissions, adjusted admissions, total facility inpatient days, adjusted patient days, total facility Medicare discharges, total facility Medicaid discharges, total births (excluding fetal deaths), total surgical operations, emergency department visits, total facility personnel full time equivalent (FTE), and adjusted average daily census.

Coded Independent Categorical Variables

Region: Hospitals included in this analysis were coded into four statistical regions defined by the U.S. Census Bureau as being the most commonly used classification system.(16) Each hospital was assigned to West, Midwest, South, or Northeast regions based on the state that was provided in the AHA database.

Gender: The CEO’s gender was identified using the CEO’s name and assigning a gender, dichotomously ascribed as male or female. For those names that were ambiguous with respect to gender, for example, Randy or Chris, a search of LinkedIn and Google was used to identify the gender of the hospital CEO.

Physician: The CEO’s medical education was identified using a specific search in LinkedIn and Google. The CEO’s name was searched and dichotomously ascribed as physician or non-physician, based on the person being identified as holding a medical or osteopathic doctoral degree.

Dependent Variables from the HCAHPS Survey

Twelve HCAHPS Star Ratings were examined as dependent variables: care transition, cleanliness of hospital environment, communication about medicines, discharge information, communication with doctors, communication with nurses, overall hospital rating, pain management, quietness of hospital environment, recommended hospital rating, responsiveness of hospital staff, and summary star rating.

Eleven of the star ratings are publicly reported measures from the HACHPS survey to capture specific aspects of the patient’s care experience. The twelfth measure combines all information from the previous measures to create the HCAHPS Summary Star Rating.(17)

Statistical Analysis

Using Statistical Analysis Software (SAS) version 9.4, we first created frequency tables on the categorical independent variables including gender, bed size, and region stratified by physician status, followed by Pearson’s chi-square test of independence in order to examine any relationship between potential confounders and physician status.

Next, we obtained the mean values for continuous independent variables, including total facility admissions, adjusted admissions, total facility inpatient days, adjusted patient days, total facility Medicare discharges, total births, total surgical operations, emergency department visits, total facility full time equivalent employees, and adjusted average daily census.

The continuous independent variables were found to be normally distributed, so a t-test was used to obtain the pooled p-value. All findings and p-values in this study were assessed for a significance level less than 0.05.

For each HCAHPS dependent variable, we obtained the frequency counts for each category of scores 1–5, stratified by physician status. We obtained the chi-square p-values to test for associations between each dependent variable and physician status.

We then examined the association between physician status and each HCAHPS outcome alone using an ordinal logistic regression and performed a sequential approach to model development by using a forward selection process. Bed size, gender, and region were added into individual models with physician status, and changes in odds ratios (ORs) and p-values were assessed.

Next, a combination of two of the independent variables, bed size and gender, followed by bed size and region, were added into each model with physician status, and changes to the ORs were assessed.

Finally, all three of the independent variables of bed size, gender, and region were added into the model with physician status, and the ORs were assessed for changes. Only independent variables showing a 10 percent or more change in the OR were included in the final model as controlling variables.

To better understand the directionality of the association between the controlling variable (bed size) and the HCAHPS outcome, we calculated the ORs, 95 percent CIs and chi- square p-values using an ordinal logistic regression model, which also included the main exposure of interest of physician status.

Results

The 2016 AHA dataset contained 6,239 hospitals, and 3,509 hospitals were included in the 2016 HCAHPS dataset (see Figure 1). A total of 2,064 exact matches were found between the two datasets on hospital name, state and ZIP code (see Figure 1). After applying the exclusion criteria, 1,588 hospitals were eligible for analysis (see Figure 1).

Table 1 compares hospital characteristics for the 1,588 hospitals, their bed size, geographic region, and gender classified by CEO physician status. Of the 1,588 available hospitals, 113 (7.1%) were led by a physician CEO and 1,475 (92.9%) were led by a non-physician CEO. Among physician CEOs, 37 (33%) were located in the Northeast, 23 (20%) had between 200 and 299 hospital beds, and 23 (20%) had 500 or more hospital beds. Additionally, 96 (85%) physician CEOs were male, and 17 (15%) were female physician CEOs.

Among the 1,475 hospitals led by non-physician CEOs, 658 (45%) were located in the South and 426 (29%) had between 100 and 199 hospital beds. Additionally, 1,119 (75.9%) were led by male CEOs and 356 (24.1%) were female CEOs.

All chi-square p-values were significant, showing that physician status is significantly associated with region, bed size and gender in our sample of U.S. hospitals.

Table 2 compares hospitals with physician and non-physician CEOs on volume, personnel, and occupancy. Hospitals led by physicians showed significantly greater total adjusted admissions, patient days, and Medicare and Medicaid discharges than those led by non-physicians (all p < .001). In addition, hospitals led by physicians had significantly more births (1,609.8 vs. 1,140.7, p = .0018); more operative procedures (1,3221.0 vs. 7,481.8, p < .001); more emergency department visits (5,4783 vs. 4,1418, p = .0003); and a higher average daily census (469.6 vs. 273.8, p < .001). Hospitals led by physicians also had significantly more full-time equivalent personnel (3,299.2 vs. 1,411.7, p < .001).

Table 3 shows frequency counts and percentages for each score (1–5) of all the 1,588 hospitals’ HCAHPS ratings categorized by whether the CEO was a physician. Additionally, the chi-square p-value is reported to determine if there is a significant relationship between physician status and each HCAHPS rating. Doctor communication, quietness, and recommended hospital rating were found to have a significant association with physician status.

Table 4 displays the ORs, 95 percent CIs, and chi-square p-values for the ordinal logistic regression models between physician status (main exposure of interest), HCAHPS rating (outcome), and controlling for a range of potential confounders (bed size, CEO gender, and region) independently and in combination.

When examining physician status and HCAHPS Rating alone, three outcome categories showed significant associations: recommended hospital rating, overall hospital rating, and quietness. After controlling for potential confounders, only bed size was found to change the OR more than 10 percent and was included in the final model with CEO physician status and the HCAHPS Rating category (see Table 4.b.1). Gender and region did not have an effect (greater than 10 percent) on the association between HCAHPS rating and physician status, hence they were excluded from the final model.

Five HCAHPS ratings showed significant associations (p < 0.05) with physician-led hospitals when controlling for bed size: recommended hospital rating, overall hospital rating, care transition, discharge information, and pain management rating (see Table 4.b.1). When controlling for bed size, the quietness rating was no longer significant (OR: 0.81; p = 0.2521) and care transition, as well as discharge information rating, became significantly associated with physician CEO status (OR: 1.55; p = 0.0143 &, OR: 1.52; p = 0.0221) (see Tables 4.a and 4.b.1). There were no associations found between physician CEO status and cleanliness, medication communication, doctor communication, nurse communication, quietness, staff responsiveness, or summary star rating (see Table 4.b.1).

Table 5 displays the ORs, 95 percent CIs, and chi-square p-values between CEO physician status, bed size, and each HCAHPS rating category. Having a higher HCAHPS score in care transition, discharge information, overall hospital rating, and pain management were found to be significantly associated with physician-led hospitals with a lesser bed size; whereas, scoring higher in recommended hospital rating was found to be significantly associated with physician-led hospitals with a larger bed size.

Discussion

Hospital CEOs have an important responsibility for the strategic priorities and operational effectiveness of the organizations they lead. Both theory and practice suggest the background, education, and experience of CEOs influence the strategic priorities they establish for their organizations and the performance outcomes they achieve.

For the past 20 years, patient experience has been prioritized as a major component of quality and now represents a fundamental component of value-based reimbursement. Hence, one would expect that patient experience would be prioritized by hospital CEOs and even perhaps more so by physician CEO; therefore, we sought to understand if having a physician CEO leading a hospital was associated with higher patient experience scores, as measured by HCAHPS, than having a non-physician CEOs.

Using the AHA and HCAHPS data for hospitals across the nation, we found that physicians represented a relatively small percentage of CEOs; however, when present, these physicians tended more often to be male and work at larger, busy hospitals. We also found that the presence of a physician CEO, when controlling for bed size, was associated with higher HCAHPS scores for care transitions, discharge information rating, overall hospital rating, pain management, and recommended hospital rating.

Finally, we found that when controlling for the CEO’s gender or the region of the hospital, there were no changes to the OR greater than 10 percent; hence, they were not included as controlling variables in the final model.

Consistent with our findings, previous research has found hospital size to affect patient experience and satisfaction. McFarland, et al., noted that larger hospitals were associated with lower HACHPS ratings in categories of doctor communication, cleanliness, and staff responsiveness.(12) While our findings suggest physician status has no association with categories of doctor communication, cleanliness, and staff responsiveness, we did detect changes in the OR after controlling for bed size.

Care transition, discharge information rating, and pain management did not show significant results until bed size was controlled for in the model, for which all three showed to be more strongly associated with smaller hospitals (OR: 0.87, p < .0001; OR:0.80, p < .0001; OR:0.78, p < .0001, respectively). Additionally, the magnitude of the relationship between physician status and overall hospital rating increased (OR: 1.51 to OR: 1.69) while the magnitude with recommended hospital rating decreased (OR: 1.95 to OR: 1.87).

Previous research suggests there is no difference in hospital performance between medical and non-medical leadership(8-11); however, differing opinions persist about which professions should manage hospitals. A modest body of evidence supports the inclusion of doctors in the composition of hospital governing boards. For example, Xirasager, et al., found increased effectiveness of doctors who completed graduate management training in Master of Business Administration, Master of Health Administration, or Master of Public Health.(8,12-13)

Minimal research provides evidence as to why physician leadership might make a difference in hospital performance.(8, 18) In our study, we tried to account for CEO gender and/or region as a possible confounder, but no significant findings were detected when controlling for gender or region in the ordinal logistic model. This largely could be due to an insufficient sample size, given there were only 17 female physician CEOs in 2016. We found gender to have a significant association with physician CEO status, with 85 percent of physician CEOs being male (see Table 1). More research is needed to establish whether gender plays a role in hospitals’ outcomes for physician leaders.

This research has some important limitations. First, it is a cross-sectional cohort study of a single year; therefore, it represents a single moment in time and may not be representative of trends in performance. Second, the designation of a CEO’s gender and status as a physician were performed using publicly available information and is subject to misclassification bias. We believe that these biases, while possible, are limited by the use of multiple data sources to interrogate the variables of gender and physician status. Third, physician CEOs are rare and represent 7.1 percent of the final sample, which can decrease the statistical power of this analysis. Fourth, there may be unknown confounders at play which were not captured in the AHA dataset.

Nonetheless, despite these limitations, we believe that this study has significant strengths. First, it provides a national perspective of physician CEOs and their HCAHPS’ performance as compared to their non-physician colleagues in non-rural hospitals greater than 25 beds. Second, we used a series of models to appropriately adjust for confounders and established that a parsimonious model that excluded gender and region provided just as much information as more resource-intensive models. Third, this is both an empiric and more contemporary study than what currently exists in the literature. Finally, the focus on specific and well-validated outcomes, namely HCAHPS scores that are important for a hospital’s value based performance, are essential to understanding the role of personal traits, like physician status, on outcome.

We believe this is one of the first evaluations of physician CEOs to link AHA data to HCAHPS ratings in order to study objective performance measures and consider gender, bed size and region as possible confounders. Further empiric studies, using contemporary datasets at scale and investigating other important hospital outcomes are warranted.

References

1. Hambrick DC and Mason PA. Upper Echelons: The Organization as a Reflection of Its Top Managers. Acad of Manage Review. 1984;9(2). https://doi.org/10.2307/258434 .

2. Centers for Medicare and Medicaid Services. HCAHPS: Patient’s Perspectives of Care Survey. www.cms.gov/Medicare/QualityInitiatives-Patient-Assessment-Instruments/HospitalQualityInits/HospitalHCAHPS . Accessed December 7, 2020.

3. Centers for Medicare and Medicaid Services. Hospital Value-Based Purchasing. www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/HospitalQualityInits/Hospital-Value-Based-Purchasing- . Accessed December 7, 2020.

4. Haley RD, Hamadi H, Zhao M, Xu J, Wang Y. Hospital Value-Based Purchasing: The Association Between Patient Experience and Clinical Outcome. Health Care Manag. 2017;36(4). doi: 10.1097/HCM.0000000000000183.

5. Boulding W, Glickman SW, Manary MP, Schulman KA, Staelin R. Relationship Between Patient Satisfaction with Inpatient Care and Hospital Readmission within 30 days. Am J Manag Care, 2011;17(1). PMID: 21348567.

6. Press Ganey. Achieving Excellence: The Convergence of Safety, Quality, Experience, and Caregiver Engagement. Strategic Insights Report; 2017. http://healthcare.pressganey.com/2017-Strategic-Insights?s=White_Paper-PGPost .

7. Press Ganey. 2020 Strategic Insights: From Principles to Practice: Powering Transformation through High Reliability. White Paper; 2020. https://www.pressganey.com/resources/white-papers/2020-strategic-insights .

8. Clay-Williams R, Ludlow K, Testa L, Li Z, Braithwaite J. Medical Leadership, A Systematic Narrative Review: Do Hospitals and Healthcare Organizations Perform Better When Led by Doctors? BMJ Open. 2017;7(9):e014474. doi:10.1136/bmjopen-2016-014474.

9. Schneller ES, Greenwald HP, Richardson ML, J Ott. The Physician Executive: Role in the Adaptation of American Medicine. Health Care Manage Rev. 1997;22(6):90–6. doi:10.1097/00004010-199704000-00010.

10. Schultz FC, Pal S, Swan DA. Who Should Lead a Healthcare Organization: MDs or MBAs? J Healthc Manag. 2004;49(2):103–16. PMID: 15074119.

11. Clay-Williams R, Braithwaite J. Doctors in Executive Management: A Systematic Review of the Peer-reviewed Literature. Sydney, Australia: Centre for Clinical Governance Research, Australian Institute of Health Innovation, University of New South Wales; 2012. https://www.yumpu.com/en/document/read/35427639/doctors-in-executive-management-australian-institute-of-health- . Accessed December 1, 2020.

12. Xirasagar S, Samuels ME, Stoskopf CH. Physician Leadership Styles and Effectiveness: An Empirical Study. Med Care Res Rev. 2005; 62(6):720–40. doi:10.1177/1077558705281063.

13. Xirasagar S, Samuels ME, Curtin TF. Management Training of Physician Executives, Their Leadership Style, and Care Management Performance: An Empirical Study. Am J Manag Care. 2006;12(2):101–8. PMID: 16464139.

14. American Hospital Association. AHA Annual Survey Database. 2016. www.ahadata.com/aha-annual-survey-database .

15. Centers for Medicare and Medicaid Services. Hospital CAHPS (HCAHPS). CMS.gov . www.cms.gov/Research-Statistics-Data-and-Systems/Research/CAHPS/HCAHPS1 . Accessed November 19, 2019.

16. US Census Buraeu. Census Regions and Divisions of the United States. Census.gov . www2.census.gov/geo/pdfs/maps-data/maps/reference/us_regdiv.pdf . Accessed October 30, 2020.

17. Centers for Medicare and Medicaid Services. Hospital Consumer Assessment of Healthcare Providers and Systems. Technical Notes for HCAHPS Star Ratings. https://hcahpsonline.org/globalassets/hcahps/star-ratings/tech-notes/2016-12_star_ratings_tech_notes.pdf . Accessed December 1, 2020.

18. DC McFarland, M Johnson Shen, P Parker, S Meyerson, RF Holcombe. Does Hospital Size Affect Patient Satisfaction? Qual Manag Health Care. 2017; 26(4):205–209. doi: 10.1097/QMH.0000000000000149.