Example Of Article Review On Statistics Research
Compensation of Chief Executive Officers at Nonprofit US Hospitals
Karen E. Joynt et al have carried out a research to determine the aspects of hospitals on which the CEO compensations were based in the case of nonprofit hospitals. Various aspects of the report are described below.
Data Collection Procedures and Instruments
Dependent Variables. The dependent variable was CEO compensations. The CEO compensations were a combination of compensation received from the primary organization; reportable compensation from related organizations such as foundations associated with the hospital and estimated other compensations. The compensations of all CEOs were adjusted for regional variations through the Medicare Wage Index (Joynt et al.).
Characteristics of the Independent Variables. The independent variables were structural and organizational measures, financial performance measures, technologic measures, quality-of-care measures and community benefit measures. The researchers assessed structural and organizational aspects from the number of hospital beds, number of hospitals overseen by the CEO, the membership of the hospital in any larger ecosystem and from whether the hospital was a major teaching institution. The researchers also assessed the proportion of Medicare patients in a hospital and the degree of disproportionality of its share hospital index, to corroborate whether a less favorable payer mix affected the CEO compensation adversely. The financial performance measures included the total margin calculated as the ratio of net income to net revenue plus other income, liquidity (ratio of current assets to liabilities), capitalization (ratio of fund balance to total assets) and occupancy rates of hospitals. The technologic measures included a technologic index that combined the presence of advanced technologies such as positron-emission tomography and magnetic resonance imaging. Quality-of-care measures were created out of quality metrics endorsed by the National Quality Forum. For each hospital, the researchers calculated a composite measure of performance on processes of care for acute myocardial infarction, congestive heart failure and pneumonia. They assessed patient satisfaction using the Hospital Consumer Assessment of Healthcare Providers. Community Benefit measures assessed charity care and other community benefits rendered by the nonprofit hospital, as reported in its 990 form (Joynt et al.).
Data Sources. The researchers picked up data from seven sources: the publicly available Form 990 tax returns for nonprofit hospitals, the American Hospital Association annual survey, rural-urban community codes, Hospital Care data, Medicare provider analyses and review files and the Medicare impact file. They identified 1877 CEOs of nonprofit hospitals from the 990 forms. The data was pre-existing, and not created specifically for the research (Joynt et al.).
Data Analysis Procedures
The researchers first plotted the distribution of CEO compensation. They then calculated the summary statistics for the hospitals that the CEOs managed, breaking down the results into the lowest decile, middle 8 deciles and the topmost decile. If CEOs managed multiple hospitals, the researchers averaged hospital characteristics weighted by the number of beds. They subsequently used multivariable linear regression with CEO compensation as the outcome (dependent variable) and the hospital structure characteristics as predictors (independent variables). Next, the researchers analyzed whether the hospitals’ performance on measures of financial performance, technology index, quality and patient outcomes were associated with CEO compensation. For this step, each individual characteristic was adjusted for the respective hospital’s structural characteristic and then the combined association with CEO compensation was worked out. For the quality of care measures, the researchers built patient-level hierarchical logistic regression models to calculate a 30-day risk adjusted mortality and readmission rates for patients admitted for congestive heart failure, myocardial infarction and pneumonia. To obviate the pitfall of concentrating on only three conditions to determine mortality, the researchers added a sensitivity analysis using a 19-condition composite of medial and surgical mortality rates. They then compared the relationship between CEO compensation across these 19 conditions. They accounted for the possibility of CEO compensation improving over time by conducting a sensitivity analysis using the change in financial performance and quality performance over the years 2006 to 2008 as predictors. Because the CEO pay data was right skewed owing to differential pay received in wealthy hospitals, the researchers repeated their analyses using quantile regression at the 25th, 50th and 75th percentiles. They considered a 2-tailed value of p<0.05 to be significant and used commercial software (Stata 12.1) for their analyses (Joynt et al.).
The researchers found that the unadjusted mean compensation of the 1877 CEOs chosen from 2681 hospitals was$595781for the year 2009, with a median compensation being $404938. The compensations were ordered into deciles. The researchers observed that the CEOs in the lowest decile with a median of $117933 were mostly from small, non-teaching hospitals in rural areas largely in the Midwest. The CEOs in the highest decile of compensation with a median of $1662548 oversaw large, urban hospitals that were mostly engaged in teaching as well. The researchers found that the overall median hospitals margin was 3.5% and median hospital occupancy rate was 63%. The hospitals had a median performance on process measures at 93.4% and 65% on patient satisfaction. Median charity and other community benefits figured at 1.6% and 1.4% of hospital expenditures. The researchers found that CEO compensation was higher for CEOs who managed more beds ($550 per additional bed, 95% CI, p<0.001). Managing more hospitals was also correlated with higher pay ($32609 per additional hospital, 95% CI, p=0.02). CEOs of major teaching hospitals were paid $425078 more (95% CI, p< 0.001). CEOs of hospitals with a higher proportion of poor patients and Medicare patients received lesser pay. The researchers did not find significant association of hospital financial performance with CEO compensation. They found a strong correlation between CEO pay and the level of technology in the hospitals. Hospitals higher on the technlogic index paid $135862 more to their CEOs (95% CI, p< 0.001). The researchers also did not find any significant correlation between performance on process quality, risk-adjusted mortality and readmission rates on CEO compensation. However, they found that hospitals with higher patient satisfaction scores paid their CEOs higher by $51706 (95% CI, p= 0.006). They found no association between charity care or community benefit programs and CEO compensation. The researchers also found no association of the 19-condition composite mortality rate in the hospitals to CEO compensations. When they assessed the changes in indices over the years 2006 to 2008, they found that it was the technology level alone that was significantly associated with CEO compensation. When using quantile regression, the researchers found little relationship between total margins and CEO compensation overall; however, they found that 25th percentile and median CEO salaries were higher at high margin hospitals. Similarly, they found that the 25th percentile of CEO compensation was higher among hospitals with high performance on process quality measures (Joynt et al.).
The researchers found that CEOs who oversaw larger teaching hospitals were the most highly compensated. They also found a higher level of technology to be significantly associated with higher CEO compensation. While they found no relationship with hospital performance on standard processes or outcome metrics, they found that patient satisfaction had a significant relationship with CEO compensation. They also found no relationship between community outreach programs with CEO pay, despite the fact that the hospitals being analyzed were nonprofit institutions. The researchers felt that patient satisfaction was associated with CEO packages due to the relative ease with which hospital board members could assess patient satisfaction as compared to other metrics. The fact that CEO pay did not associate with quality metrics of processes and patient outcomes confounded the researchers. They observed that the 25th percentile of CEO pay reflected better adherence to quality standards. The researchers felt that hospital boards could make the link between CEO pay and hospital quality performance more explicit. They assumed that such was not the case because of the fear of the unintended consequence of hospitals turning away complex cases. The researchers were surprised at the weak association between CEO pay and financial performance. They felt that the total margins might affect CEO pay up to a point, after which other factors such as technology level and complex care played a bigger role. The researchers assumed that high levels of technology simply reflected the hospitals’ financial health, and thus found an indirect relationship between hospital performance and CEO pay (Joynt et al.).
The researchers have relied on data from 990 forms for CEO compensation figures. This data, while audited, has not been extensively validated. The possibility of under-reporting of CEO compensations in the 990 forms could not be ruled out. Hospitals might not have been consistent in the definitions of uncompensated care, and might have inflated the costs attributed to charity. Quality measures such as patient satisfaction metrics were only from publicly available data and were not exhaustive. The possibility of internal quality metrics monitored by respective hospitals being a better measure of quality could not be ruled out. The metrics of financial performance might have the limitation of not fully reflecting the health of the hospital. The study also did not take into account the personal characteristics of the CEOs such as experience and duration of time with the organization while assessing CEO compensation. The study measured the data only over 2008 and 2009, and as such, the association of various metrics over a longer time continuum might have changed (Joynt et al.).
Karen E. Joynt et al have carried out a research using recently available data of CEO compensations at nonprofit US hospitals to determine the aspects of hospitals on which the CEO compensations were based. The researchers took into account the compensation of 1877 CEOs from 2681 private nonprofit US hospitals and carried out linear regression to determine which characteristics of the hospital were associated with CEO pay. The characteristics of the hospital that were checked for correlation with CEO pay were structural and organizational measures (such as number of hospital beds, institutional membership or teaching institution), financial performance measures, technological measures of hospital advancement and quality-of-care measures. The researchers found that CEO pay was linked to the number of beds in the hospital and the level of technology used in the hospital. The researchers felt that the linkage between high technology and CEO pay might be incidental, as the hospitals would be interested in projecting themselves as technologically advanced to their patients. They found that beyond a point, the revenues did not correlate to CEO pay. The researchers found a high correlation between declared patient satisfaction levels and CEO pay, and ascribed this linkage to the fact that Hospital Directors might have found it easier to accumulate patient satisfaction data rather than metrics related to hospital process management and performance. The researchers found little correlation between the CEO pay and patient processes, patient outcomes and community benefit. The analysis showed that considerable ground could be covered to link hospital processes and patient outcomes to CEO pay. The researchers acknowledged that linking patient outcomes to CEO pay may have unintended consequences of the hospital refusing complex cases (Joynt et al.).
Joynt, Karen E. et al. “Compensation of Chief Executive Officers at Nonprofit US Hospitals.” JAMA Internal Medicine 174/1: 61-67. January 2014. Web. April 13, 2015.