Physician-hospital integration and hospital efficiency investigation

Introduction

In the healthcare landscape, efficiency, an indicator of how healthcare resources are being used to provide the best value for patient care, is essential for hospitals (1). It calls for hospitals to provide high-quality care while controlling costs, and it measures the relationship between resource inputs (such as healthcare labor resources) and healthcare outputs (such as care quality). To tie the quality of care delivered to payments for care delivery, multiple value-based care (VBC) programs have been implemented nationwide (2,3), such as the Hospital Value-Based Purchasing Program, Hospital Readmission Reduction Program, and the Hospital Acquired Conditions Reduction Program. Studies have shown that VBC adoption was associated with improved quality of care. For instance, a study examining 3,387 United States (U.S.) hospitals from 2007 to 2015 revealed that the Hospital Readmission Reduction Program decreased readmission rates as a result of program incentives (4). Furthermore, Pandey and colleagues conducted a systematic review and concluded that higher-intensity Value-Based Purchasing Programs were more consistently associated with improvements in desired quality processes, utilization measures, and reductions in spending, compared to lower-intensity programs (5).

Increasingly, U.S. hospitals have prioritized efficiency promotion through strategies, such as improving patient safety care, shaping patient-centeredness care culture (6-8), and managing waste reduction (9,10). However, efficacious adoption of these strategies requires a mechanism that aligns staff interests and behaviors with the goals of the hospitals. The need to become more efficient has led hospital executives to implement physician-hospital integration (PHI), which encompasses a closer collaboration between physicians and hospitals through various contractual forms, including employment contracts, joint ventures, and other partnerships (6-8).

Based on a study that analyzed 2014–2018 American Hospital Association Annual Surveys, more than 60% of U.S. hospitals were involved in various formats of PHI (11). Research has suggested that PHI could play an important role in reshaping and harmonizing physician-hospital relationships and navigating physician behaviors toward hospital goals (6,12,13); thereby improving hospital performance. Hospital management is also interested in building mechanisms to facilitate physician-hospital relationships so hospitals can better navigate physicians’ inputs and achieve desirable outputs. However, studies have shown that PHI has exerted mixed impacts on hospital performance, especially on hospital efficiency (14-16).

In the literature, the most popular hierarchical categories use high, medium, and low controlled levels, or little to full controlled PHI levels (11,12,17-20). The high-controlled level PHI, usually through employment-based arrangements, is characterized by a stable, close relationship between physicians and hospitals to foster care coordination and financial sustainability (12,21). By contrast, the low-level controlled PHI, is usually based on collaborative contracts, emphasizes physician autonomy and loose financial risk sharing (12). However, these broad categories fail to clearly define the boundary of each category or the level of integration for certain PHI models (11,12,20,22); setting clear boundaries can improve the effectiveness of explaining predictors for examining the effects of PHI on hospital performance (15).

Therefore, unlike previous studies, this study proposes using an objective-based framework to classify PHI types, including (I) financial PHI, which mainly aims to improve financial performance through economies of scale (23-25), (II) clinical PHI, which mainly aims to improve the quality of care through care collaboration (24,26), and (III) financial and clinical hybrid PHI, which has dual goals of improving care collaboration and financial performance.

Studies have suggested that clear-cut goals often significantly impact an organization’s performance (27,28), we predicted that this objective-based classification of PHI would better evaluate the impact of PHI on hospital efficiency. Additionally, Li (29) found hospital-level PHI and system-level PHI demonstrated distinctly different hospital performance outcomes. These differences underscore the importance of examining the relationships between PHI and hospital efficiency in different organizational contexts. We thus examine hospital efficiency with both hospital-level PHI and system-level PHI.

The aim of this study was to examine the association between PHI and hospital efficiency with this objective-based classification of PHI. To better understand the contributors to hospital efficiency, we employed the data envelopment analysis (DEA) and performed a cross-sectional analysis, using the 2020 American Hospital Association (AHA) Annual Survey, the 2020 Centers for Medicare and Medicaid Services (CMS) Hospital Compare datasets, and the 2023 updated Health Service Areas Files (HSAF) from the National Center for Health Statistics. Our objectives were to (I) measure hospital efficiency of both hospital-level and system-level with three PHI types: financial PHI [represented by Integrated Salary Model (ISM)], clinical PHI [represented by Independent Practice Association (IPA)], and financial and clinical hybrid PHI [represented by Open and Closed Physician-Hospital Organization (Open-PHO and Closed-PHO)] (the definitions of each PHI model are shown in Table 1), respectively; and (II) identify other drivers for hospital efficiency, including organizational characteristics and market competition. The findings of this study were expected to assist hospital management in exploring pathways and developing strategies to better integrate with physicians in its PHI adoption process to improve hospital efficiency.

Conceptual framework & hypotheses

Conceptual framework

The conceptual framework of this study is built upon the X-inefficiency Theory and Agency theory. X-inefficiency is concerned with situations when an organization fails to fully utilize its resources to achieve the maximum possible output level, that is, the efficiency frontier (31). These inefficiencies can arise from numerous sources, such as outdated technology, inefficient production processes, suboptimal management practices, and insufficient competitive pressures, among others. They also manifest when employee behavior deviates from optimal performance, driven by a diminished motivation to pursue efficiency. However, when appropriately motivated, the employees can behave more optimally (32). X-inefficiency is a commonly used theoretical framework for describing hospital efficiency (33-35). The root cause of X-inefficiency is market imperfection, including information asymmetries and uncompetitive pressure, which may lead to irrational decisions of an organization and its employees (36).

In hospital settings, the efficiency of hospitals is greatly influenced by physicians since they are the primary decision-makers regarding how resources should be allocated to treat patients (37). A study has found that physician practice patterns were associated with higher levels of inefficiency, and physician costs were responsible for about four-fifths of total healthcare costs (38). Thus, it is expected that PHI adoption can motivate physicians to improve hospital efficiency.

On the other hand, Agency theory holds that principals, represented by hospitals in the PHI context, can delegate duties or decision-making authority to agents (39,40), represented by physicians. Physicians and hospitals are traditionally viewed by economists as buyers and sellers of services, while organizational theorists typically view the relationship as loosely coupled bureaucracies (41). Consequently, each of these two parties allocates resources to pursue its own objectives, which may lead to conflicting outcomes (42). Regarding the relationship between PHIs types (financial PHI, clinical PHI, and financial and clinical hybrid PHI), physicians, and hospital efficiency, it is expected that PHI can help tighten the relationship between physicians and hospitals economically and psychologically, enabling the two parties to align their goals more effectively (11,20,23).

Hypotheses

Among the three PHI types (the specific PHI models are defined in Table 1), since financial PHI (represented by ISM) is typically based on employment-based integration to build a more transparent information system to monitor and navigate physician behaviors, it is more likely that financial PHI could align physician behaviors with hospital interests and increase hospital efficiency. However, prior research on the association between hospital efficiency and financial PHI was mixed (13,16,24).

For clinical PHI (represented by IPA), it is generally built through contractual collaboration. For example, through setting care managers and building a sharable medical record, physicians can easily access patient records, build registrars, and develop a variety of care coordination (24). However, hospitals that implement clinical PHI may not have sufficient control over physician behaviors, and physicians may be more likely to focus on care quality improvement without paying much attention to care inputs. Under such circumstances, clinical PHI may be negatively linked to hospital efficiency.

Additionally, for financial and clinical hybrid PHI (represented by Open & Closed PHO), although its dual goal is to improve both quality of care and financial performance, and hospitals may be able to manage and monitor physician behaviors to some degree, it may divert hospital efforts to both goals, resulting in lower hospital efficiency than financial PHI. It is important to acknowledge the distinctions between Open PHOs and Closed PHOs, especially in terms of management and control. However, as noted by Alexander et al., the differences between Open-PHOs and Closed-PHOs might predominantly be semantic rather than substantive (43). We thus categorize both Open-PHOs and Closed-PHOs into financial and clinical hybrid PHI.

These perspectives have informed our following hypotheses to this study:

• H1: hospital efficiency is positively associated with both hospital-level and system-level financial PHI (represented by ISM).
• H2: hospital efficiency is negatively associated with both hospital-level and system-level clinical PHI (represented by IPA).
• H3: hospital efficiency is positively associated with both hospital-level and system-level financial and clinical hybrid PHI (represented by Open & Closed PHO).

Methods

Data source

We conducted a cross-sectional analysis using the 2020 AHA Annual Survey data, the 2020 CMS Hospital Compare dataset, and the 2023 updated HSAF. These datasets provide information on hospital characteristics, market characteristics, and hospital inputs and outputs. The Medicare provider number was used for merging the datasets.

Study samples

This study focused on short-term general acute care hospitals in the U.S. There were 6,156 hospitals that participated in the 2020 AHA Annual Survey. From this initial group, 257 hospitals were removed due to missing Medicare provider numbers. Additionally, 1,887 hospitals were excluded for lacking data on either IPA, ISM, Open-PHO, or Closed-PHO at the hospital or system level. Furthermore, 1,998 hospitals were omitted because they had not implemented any of the aforementioned models at either level. After merging data from the AHA Annual Survey with the 2020 CMS Hospital Compare dataset (specifically, the Hospital General Information dataset), several other exclusions were made: 455 government-owned hospitals, 74 hospitals not included in the merged dataset, and 389 non-acute care hospitals (comprising 31 children’s hospitals, 298 critical access hospitals, and 60 psychiatric hospitals). Additionally, 422 hospitals were excluded for not being classified as “short-term hospitals” in the Hospital Cost Report of CMS Hospital Compare dataset. Finally, exclusions were made for hospitals missing specific outcome data: 165 for acute myocardial infarction (AMI) mortality, 38 for AMI readmission, and 37 for central line-associated bloodstream infection (CLABSI) data. Consequently, the final sample comprised 434 non-government-owned short-term general acute care hospitals.

Measures

Outputs

The outputs of quality of care were composed of desirable outputs and undesirable outputs. Based on literature review (11,18,20,44), desirable output indicators included the total facility admissions and inpatient and outpatient surgical operations, aligning with previous PHI efficiency research (18). The data came from the 2020 AHA Annual Survey.

Three components included in undesirable output indicators were: (I) CLABSI score in Intensive care units (ICUs), (II) the AMI 30-day mortality rate, and (III) the AMI 30-day readmission rate. A higher CLABSI score indicates a worse central line-associated blood infection performance in a hospital. The CLABSI score came from the 2020 CMS Healthcare-Associated Infections data of the 2020 CMS Hospital Compare datasets; the AMI 30-day mortality was calculated as the death rate for patients discharged from hospitals with a principal diagnosis of AMI within 30-day post-discharge divided by admissions for patients discharged from hospitals with a principal discharge diagnosis of AMI who had a complete claims history for the previous 12 months before discharge. A higher AMI 30-day mortality score indicates a higher AMI 30-day mortality in a hospital. The AMI 30-day mortality data were obtained from the Complications and Deaths data of the 2020 CMS Hospital Compare datasets (45). The AMI 30-day readmission rate, which was calculated by dividing the number of unplanned all-cause readmissions within 30 days after a hospital discharge by the number of admissions for AMI inpatients with the previous 12 months before discharge (45) was obtained from the Unplanned Hospital Visits data of the 2020 CMS Hospital Compare datasets. The selection of CLABSI score, along with AMI mortality and readmission rates, as indicators of care quality was made to maintain consistency with prior research on PHI (11,20,46,47). This alignment ensures that our study builds on the established body of knowledge and facilitates comparability with existing findings in the field.

To prepare the data for DEA analysis, we addressed the issue of zero values in the dataset. Because DEA requires all output and input values to be positive, we substituted zero values in CLABSI, AMI mortality, and AMI readmission rates with a small positive number (0.0001) in order to avoid distortions in the analysis. Furthermore, the negative indicators were inverted in accordance with the DEA model’s requirement that higher values should always indicate better performance. Therefore, the CLABSI, AMI mortality, and AMI readmission rates were transformed by taking their reciprocals. As a result of this transformation, a higher transformed value within the DEA model indicates a decrease in the undesirable outcome, thus maintaining compatibility with the DEA’s principle of optimizing performance indicators.

Inputs

Input measures in this study included (I) case mix index, (II) number of beds, and (III) the full-time equivalent (FTEs) total personnel. Case mix index was associated with hospital efficiency, although empirical results were mixed (32,48). To keep consistency with prior studies on PHI and its impact on efficiency, we have incorporated the case mix index into the input measures (18). Case mix index represents a hospital’s average diagnosis-related group (DRG) relative weight. It was calculated by adding up the DRG weights across all Medicare discharges, and then the number was divided by the number of discharges (49). The data came from the CMS Inpatient Prospective Payment Systems (IPPS) Final Rule and Correction Notice 2019 File; Number of beds refers to the facility beds set up and staffed at the end of the reporting period [2020]. Hospital bed size represents hospital size and has been commonly used as an input for examining hospital efficiency differences (18,44,50); FTEs, excluding medical and dental residents, interns, and other trainees, has been considered an important measure for hospital efficiency (44,51,52). Both hospital bed size and FTE data came from the 2020 AHA Annual Survey.

Key organizational factors—PHI types

The key organizational factors in this study were the three PHI types: financial, clinical, and financial and clinical hybrid PHI, which are represented by ISM, IPA, and Open-PHO and Closed-PHO, respectively. The data came from the 2020 AHA Annual Survey.

Organizational characteristics

Since many studies have suggested that for-profit hospitals’ efficiency outperformed nonprofit ones, we classified hospital ownership into two categories (a dichotomous variable): (I) nonprofit hospitals and (II) for-profit hospitals (investors-owned for-profit hospitals) derived from the 2020 AHA Annual Survey. Moreover, a study found that teaching hospitals performed better than non-teaching hospitals (18), although teaching hospitals typically yielded lower efficiency than nonteaching hospitals (53,54). We thus categorized the hospitals into two categories (dichotomous variables): (I) teaching hospitals (recognized for Accreditation Council for Graduate Medical Education accredited programs) and (II) nonteaching hospitals. In addition, since divergent results were observed regarding efficiency in different hospital locations (rural vs. urban) (55,56), two hospital locations from the 2020 AHA Annual Survey were used in this study: (I) metro, and (II) micro and rural to further investigate the difference in hospital location regarding hospital efficiency.

Market characteristic

The Herfindahl-Hirschman Index (HHI) is widely used to measure market concentration (57). Typically, when the HHI of a hospital rises, market competition decreases, and market power increases (58), hence being associated with a decrease in hospital efficiency. In this study, information regarding each hospital’s admission was obtained from the 2020 AHA Annual Survey, which was summed up at the health service area (HSA) that was from the National Center for Health Statistic. Then, we obtained the value of the market share for each hospital by dividing the individual hospital’s admission by its total admissions from the same HSA data. The value of HHI of a hospital then was calculated by squaring the market share of each competing hospital and then summed up at the HSA level. HHI calculations can be explained as, where represents the market share (percentage) of hospital n, and represents the total number of the hospitals in a given HSA.

Statistical analyses

We performed descriptive statistical analyses for hospital inputs and outputs, and organizational and market characteristics. We used frequency and percentage to characterize categorical variables and mean and standard deviation to characterize continuous variables.

We employed a DEA, a nonparametric method that is often applied to assess Decision-Making Unit’s (DMU’s) performance (59) and evaluate efficiency in operation management (60). In essence, DEA compares each DMU with a set of peers. A measure of efficiency in DEA is the ratio of the weighted sum of outputs to the weighted sum of inputs. We employed an “rDEA” package that is a conservative DEA-based method, to handle data uncertainties in DMUs, ensuring stable and reliable performance evaluation (61).

Our analysis categorized outputs into two categories: “undesirable” outputs, including CLABSI score, AMI 30-day mortality rate, and AMI 30-day readmission rate, which should be minimized, and “desirable” outputs, including total facility admissions and inpatient and outpatient surgical operations, which should be maximized. A DEA with an output-oriented approach and variable returns to scale were used to evaluate hospital performance. This approach was chosen to maximize outputs within a given input, which is particularly relevant to healthcare settings where maximizing patient care outcomes is often the objective. Based on this output-oriented DEA model, technical efficiency scores reflect how well hospitals utilize their healthcare inputs to maximize outputs. A score of “1” indicates optimal efficiency, indicating the hospital is operating at the forefront of best practice. On the other hand, a score greater than “1” indicates inefficiency, indicating that a hospital has the potential to improve its output levels without increasing its inputs.

To mitigate the potential limitations of DEA, which may not always distinguish between inefficiencies caused by operational factors and external factors beyond the control of the DMUs (62), we also performed multiple linear regression analyses to examine the associations between PHI types and efficiency scores. Before conducting the regression analyses, we checked that all necessary linear regression assumptions had been met. After removing outliers using Cook’s distance method, we determined that a log transformation of the dependent variable (efficiency scores) was more appropriate to meet the assumptions of linear regression. As a result of this transformation, the variance of our model was stabilized, and its interpretability was improved. Also, we selected the best model for our study after performing model selection. Additionally, to address the limitations of DEA for making statistical inferences, we applied a two-stage semi-parametric modeling approach (63). The detailed results of this analysis are available in Appendix 1 (Figures S1-S2, and Tables S1-S3). All analyses were performed using R Programming software (R Foundation, Boston, MA, USA, Version 4.2.1). The significance level was set at P value =0.05.

Results

Descriptive statistics of DEA

Table 2 shows the descriptive statistics of the hospitals’ undesirable outputs, desirable outputs, and inputs. Regarding undesirable outputs, both AMI 30-day mortality and AMI 30-day readmission have small variabilities, and the standard deviation values were 1.09 and 0.93, respectively. In contrast, the CLABSI score had relatively wide variability, and the value of standard deviation were 0.75 with a mean score of 0.88. Desirable outputs, both admissions and inpatient and outpatient surgical operations, demonstrated wide variabilities among the hospital samples. In terms of inputs, the variability of the case mix index of the hospital samples was relatively small, with a standard deviation of 0.27. In contrast, the variabilities of bed numbers and staffing/FTEs were significant.

DEA results

A score of “1” indicates an efficient subject in the DEA (64), and the DEA results are shown in Table 3 and Figure 1. Among the 434 hospital samples, 95 (22%) hospitals were considered efficient, while 339 (78%) hospitals exhibited inefficiency. The mean of DEA scores of inefficient hospitals was 1.093.

Figure 1 DEA efficiency plot. DEA, data envelopment analysis.

Furthermore, we summarized hospital efficiency based on our hospital-level PHI and system-level PHI in Table 4. Among hospital-level PHI, financial and clinical hybrid PHI (Open-PHI) had the highest proportion (27%) of efficient hospitals, followed by clinical PHI (IPA) of 25%, financial and clinical hybrid PHI (Closed-PHI) of 24%, while financial PHI (ISM) had the lowest proportion (19%) of efficient hospital. Meanwhile, among the system-level PHI, financial and clinical hybrid PHI (Closed-PHI) had the highest proportion (34%) of efficient hospitals, followed by financial and clinical hybrid PHI (Open-PHI) of 26%, clinical PHI (IPA) of 21%, while financial PHI (ISM) was ranked last (18%).

Descriptive statistics of hospital organization and market characteristics

A description of the hospital organization and market characteristics of the sample is shown in Table 5. Altogether, 41% financial PHI (ISM), 16% financial and clinical hybrid PHI (Open-PHO), 5% financial and clinical hybrid PHI (Closed-PHO), and 11% of the hospital sample reported a presence of clinical PHI (IPA). Ninety-four percent of the sample was non-profit, while 6% was for-profit. Over three-quarters (77%) of the sample were teaching hospitals. Additionally, the majority of the hospitals (93%) were located in metropolitan areas, and 7% were found in micro and rural areas.

Multiple regression analysis

Table 6 shows the results of regression analysis for each of the key independent variables (hospital and system-level PHI types). Among hospital-level PHI, no statistical significances were found for all the PHI types. For hospital organizational characteristic covariates, holding all other variables constant, for-profit hospitals had a 3.3% higher DEA efficiency score (high efficiency scores indicates inefficiency) (P<0.01), compared to non-profit hospitals; for every one-point increase in case mix index, there was a 4.7% increase in DEA efficiency score (P<0.001); for a hospital in the metro area, there was a 2.1% increase in DEA efficiency score (P<0.05).

Among system-level PHI, holding all other variables constant, the presence of a financial and clinical hybrid PHI (Open-PHO) was associated with a 1.8% lower DEA efficiency score (indicating higher efficiency) (P<0.05). We did not find statistically significant associations between the presence of a financial PHI (ISM) or clinical PHI (IPA) and DEA efficiency score. We also observed no significant associations with the presence of a Closed-PHO. Notably, the P value for this association was 0.051, which is marginally above the threshold for significance but may suggest practical significance. For hospital organizational characteristic covariates, holding all other variables constant, for-profit hospitals had a 3.6% higher DEA efficiency score (indicating lower efficiency) (P<0.01), compared to non-profit hospitals; for every one-point increase in case mix index, there was a 3.8% increase in DEA efficiency score (P<0.01).

In summary, DEA analysis results have shown that hospitals with a financial and clinical hybrid PHI, encompassing both Open-PHO and Closed-PHO, had a higher proportion of efficient hospitals. Furthermore, multiple linear regression analyses have revealed that at the hospital-level PHI types did not demonstrate statistically significant associations with logged DEA efficiency scores. In contrast, the system-level financial and clinical hybrid PHI (Open-PHO) exhibited a negative association with DEA efficiency scores, which suggests that Open-PHO was positively linked to higher efficiency, albeit in a relatively modest way.

Discussion

With the new objective-based framework for PHI categorization proposed by Li (29), we anticipate this framework will be more effective in examining the association between PHI and hospital performance compared to previous PHI categorizations. A robust output-oriented DEA model was applied to set targets on the best practice frontier in this study to compare hospital efficiency among financial, clinical, and financial and clinical hybrid PHI. In addition, multiple linear regression analyses were conducted to further examine the underlying logic about efficiency scores, hospital organizational characteristics, and market characteristic, in addition to the relationship between efficiency and these three PHI categorizations at both the hospital and system-level.

One of our major findings indicates that hospitals adopting either hospital-level or system-level financial and clinical hybrid PHI were more efficient, based on the robust DEA results. This finding is inconsistent with previous research: some of which showed that a high level of controlled PHI or employment-based PHI was more efficient (18), while others suggested there was no difference or lower efficiency with PHI adoption (65). Highly controlled levels of PHI, or employment-based PHI usually have more clear-cut incentive mechanisms, such as risk sharing or salary incentives. Hence, it would be strongly tied to the relationship between physicians and hospitals and may increase the motivation of physicians to align their working goals with hospital goals (12,21). However, as we discussed previously, high levels of controlled PHI or employment-based PHI that share certain commonalities with financial PHI focusing on pursuing hospital financial performance improvement (29,66) may lead to negative side effects in clinical outcomes. As a result, the potential financial performance advantage, such as high volume of admissions, high level of controlled PHI or employment-based PHI, may be offset by the potential clinical performance drawbacks when measuring the efficiency of a hospital. In contrast, hospitals with financial and clinical hybrid PHI, which seeks to improve both financial and clinical performance, may more effectively balance hospital inputs and outputs, thus demonstrating greater efficiency, based on our results.

Our study indicates that for-profit hospitals exhibit lower efficiency compared to non-profit hospitals, a finding that challenges the expectations set by theories such as agency theory, property-rights theory, and public choice theory (67). Despite the assumption that for-profit hospitals would prioritize financial stability through various incentives (68), the relationship between hospital ownership and efficiency is complex, as indicated by previous mixed empirical findings (69,70). For instance, for-profit hospitals have shown responsiveness to external financial incentives (71), yet this does not always translate into higher efficiency (67). Our study adds to this discourse by suggesting that despite the growing market share of for-profit hospitals in the U.S. (72), non-profit hospitals may still hold an edge in employing more effective resource allocation strategies, thereby optimizing the balance between their inputs and outputs.

Furthermore, our findings show that the case mix index was positively associated with hospital inefficiency, which is consistent with existing studies (33,73). Studies have shown that a higher case mix index indicating a higher level of clinical complexity was more likely to have a lower care quality outcome (15), and care quality outcomes were predominant in the hospital efficiency model (74). Thus, hospitals with a high case mix index should devote more resources to improve the quality of care when striving to increase hospital efficiency.

Limitations

This study has limitations. First, as a cross-sectional study, it did not attempt to explain causality but the associations between efficiency and PHI categorizations. Second, the generalizability of this study could have been weakened by exclusion criteria, listwise deletion, and datasets sample inclusion. The focus on short-term performance of non-governmental, acute care hospitals engaged with ISM, IPA, and PHO (both Open-PHO and Closed-PHO) further narrows the scope, resulting in a modest sample size of 434. This sample also disproportionately represents nonprofit, teaching, and metropolitan-area hospitals, which may diminish the generalizability of our findings. Third, our dataset corresponds to 2020, a year marked by the COVID-19 pandemic, leading to significant disruptions in the healthcare industry, particularly within hospitals. This disruption could have influenced both the gathering and the integrity of the data that might have impacted the generalizability of our findings. In addition, due to the availability limitation of the CMS Hospital Compare dataset, not all variable data collection duration was exactly within the year 2020, which may have affected the accuracy of our findings. Furthermore, consistent with prior studies, we selected facility admissions, along with inpatient and outpatient surgical procedures, as the desired outputs for our analysis. However, it is important to note a potential limitation: in hospitals that serve as safety nets or where a significant portion of revenue comes from capitation arrangements or value-based plans, these outputs might not be as indicative of desired outcomes. Lastly, in line with prior research, we did not apply weights to output or input variables, which may introduce some unobserved biases.

Conclusions

The findings of this study suggest that a sole objective-based PHI, either financial PHI or clinical PHI, may not be an effective approach. Instead, a two-pronged PHI objective, namely a clinical and financial hybrid PHI model, may be a more effective strategy. It is important that hospitals commit to improving care quality as well as prioritizing resources to improve their operational efficiency to ensure their sustainability through dual goal setting and culture nurturing. In addition, it is also necessary for hospital management to consider hospital characteristics, such as hospital ownership and case mix index, when implementing strategies that aim to strengthen the physician-and-hospital relationship and improve hospital efficiency.

Acknowledgments

We would like to extend our sincere gratitude to the Research Information Technology Department and Data Core at the Ohio State University for their invaluable support to this study.

Funding: The development of this publication was partially supported through the CEDAR core funded by the MUSC Office of the Provost.

Footnote

Peer Review File: Available at https://jhmhp.amegroups.com/article/view/10.21037/jhmhp-24-10/prf

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jhmhp.amegroups.com/article/view/10.21037/jhmhp-24-10/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. There is no need for IRB approval as no human subjects were involved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Palmer S, Torgerson DJ. Economic notes: definitions of efficiency. BMJ 1999;318:1136. [Crossref] [PubMed]
2. ChangeHealthcare. Value-based care in America: State-by-state. Change Healthcare; 2019.
3. CMS. What are the value-based programs? 2023. Available online: https://www.cms.gov/medicare/quality/value-based-programs
4. Zuckerman RB, Sheingold SH, Orav EJ, et al. Readmissions, Observation, and the Hospital Readmissions Reduction Program. N Engl J Med 2016;374:1543-51. [Crossref] [PubMed]
5. Pandey A, Eastman D, Hsu H, et al. Value-Based Purchasing Design And Effect: A Systematic Review And Analysis. Health Aff (Millwood) 2023;42:813-21. [Crossref] [PubMed]
6. Rundall TG, Shortell SM, Alexander JA. A theory of physician-hospital integration: contending institutional and market logics in the health care field. J Health Soc Behav 2004;45:102-17. [PubMed]
7. Witt M, Jacobs L; The Camden Group. Physician-hospital integration in the era of health reform. California HealthCare Foundation; 2010.
8. Robinson JC. Physician-hospital integration and the economic theory of the firm. Med Care Res Rev 1997;54:3-24; discussion 25-31. [Crossref] [PubMed]
9. Bentley TG, Effros RM, Palar K, et al. Waste in the U.S. Health care system: a conceptual framework. Milbank Q 2008;86:629-59. [Crossref] [PubMed]
10. Jimmerson C, Weber D, Sobek DK 2nd. Reducing waste and errors: piloting lean principles at Intermountain Healthcare. Jt Comm J Qual Patient Saf 2005;31:249-57. [Crossref] [PubMed]
11. Upadhyay S, Opoku-Agyeman W. Physician hospital integration and its effect on hospital performance: A study of US hospitals. 2022 Academy of Management; 2022; Seattle, WA, USA.
12. Dynan L, Bazzoli GJ, Burns R. Assessing the extent of integration achieved through physician-hospital arrangements. J Healthc Manag 1998;43:242-61; discussion 261-2. [Crossref] [PubMed]
13. Burns LR, Muller RWHospital-physician collaboration. Landscape of economic integration and impact on clinical integration. Milbank Q 2008;86:375-434. [Crossref] [PubMed]
14. Amado GC, Ferreira DC, Nunes AM. Vertical integration in healthcare: What does literature say about improvements on quality, access, efficiency, and costs containment? Int J Health Plann Manage 2022;37:1252-98. [Crossref] [PubMed]
15. Cervin A, Li X, Gemeinhardt G, et al. Physician hospital integration influence on quality performance. In: Management PCHD, The University of Texas Health Science Center at Houston School of Public Health, editor. 2024.
16. Jia H. Hospital-physician integration and physician collaboration: Implications for care efficiency and outcomes. University of Tennessee, 2022.
17. Madison K. Hospital-physician affiliations and patient treatments, expenditures, and outcomes. Health Serv Res 2004;39:257-78. [Crossref] [PubMed]
18. Leleu H, Al-Amin M, Rosko M, et al. A robust analysis of hospital efficiency and factors affecting variability. Health Serv Manage Res 2018;31:33-42. [Crossref] [PubMed]
19. Satiani B, Vaccaro P. A critical appraisal of physician-hospital integration models. J Vasc Surg 2010;51:1046-53. [Crossref] [PubMed]
20. Cervin A. Physician hospital arrangement influence on nonprofit hospital quality, financial performance and community benefit expenditures. UT Health School of Public Health; 2020.
21. Boone B. Integrated healthcare delivery systems’ challenges. International Risk Management Institute; 2000.
22. Ciliberto F, Dranove D. The effect of physician-hospital affiliations on hospital prices in California. J Health Econ 2006;25:29-38. [Crossref] [PubMed]
23. Goes JB, Zhan C. The effects of hospital-physician integration strategies on hospital financial performance. Health Serv Res 1995;30:507-30. [PubMed]
24. Colla C, Yang W, Mainor AJ, et al. Organizational integration, practice capabilities, and outcomes in clinically complex medicare beneficiaries. Health Serv Res 2020;55:1085-97. [Crossref] [PubMed]
25. Neprash HT, Chernew ME, Hicks AL, et al. Association of Financial Integration Between Physicians and Hospitals With Commercial Health Care Prices. JAMA Intern Med 2015;175:1932-9. [Crossref] [PubMed]
26. Bazzoli GJ. Hospital consolidation and integration activity in the United States. In: Blank JLT, Valdmanis VG. editors. Evaluating hospital policy and performance: Contributions from hospital policy and productivity research. Leeds: Emerald Group Publishing Limited; 2007.
27. Kleingeld A, van Mierlo H, Arends L. The effect of goal setting on group performance: a meta-analysis. J Appl Psychol 2011;96:1289-304. [Crossref] [PubMed]
28. Smith KG, Locke EA, Barry D. Goal setting, planning, and organizational performance: An experimental simulation. Organ Behav Hum Decis Process 1990;46:118-34. [Crossref]
29. Li X. The association between physician-hospital integration and the performance of acute care hospitals. The University of Texas Health Science Center at Houston; 2024.
30. American Hospital Association. AHA Annual Survey American Hospital Association. 2020;2020:30.
31. Leibenstein H. Allocative efficiency vs." X-efficiency Am Econ Rev 1966;56:392-415.
32. Leibenstein H. On the basic proposition of X-efficiency theory. Am Econ Rev. 1978;68:328-32.
33. Rosko MD, Chilingerian JA. Estimating hospital inefficiency: does case mix matter? J Med Syst 1999;23:57-71. [Crossref] [PubMed]
34. Rosko MD, Proenca J. Impact of network and system use on hospital X-inefficiency. Health Care Manage Rev 2005;30:69-79. [Crossref] [PubMed]
35. Rosko MD. Performance of US teaching hospitals: a panel analysis of cost inefficiency. Health Care Manag Sci 2004;7:7-16. [Crossref] [PubMed]
36. Leibenstein H. General X-efficiency theory and economic development. Oxford: Oxford University Press; 1978.
37. Chilingerian JA, Sherman HD. Managing physician efficiency and effectiveness in providing hospital services. Health Serv Manage Res 1990;3:3-15. [Crossref] [PubMed]
38. Chilingerian JA. Evaluating physician efficiency in hospitals: A multivariate analysis of best practices. Eur J Oper Res 1995;80:548-74. [Crossref]
39. Eisenhardt KM. Agency theory: An assessment and review. Acad Manage Rev 1989;14:57-74. [Crossref]
40. Jensen MC, Meckling WH. Theory of the firm: Managerial behavior, agency costs and ownership structure. J Finan Econ 1976;3:305-60. [Crossref]
41. Shortell MS. Effective hospital-physician relationships. Ann Arbor, Michigan: Health Administration Press; 1990.
42. Pontes MC. Agency theory: a framework for analyzing physician services. Health Care Manage Rev 1995;20:57-67. [Crossref] [PubMed]
43. Alexander JA, Morrisey MA. Hospital-physician integration and hospital costs. Inquiry 1988;25:388-401. [PubMed]
44. Nayar P, Ozcan YA. Data envelopment analysis comparison of hospital efficiency and quality. J Med Syst 2008;32:193-9. [Crossref] [PubMed]
45. Services CfMM. CMS Measures Inventory. Available online: https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/QualityMeasures/CMS-Measures-Inventory
46. Scott KW, Orav EJ, Cutler DM, et al. Changes in Hospital-Physician Affiliations in U.S. Hospitals and Their Effect on Quality of Care. Ann Intern Med 2017;166:1-8. [Crossref] [PubMed]
47. Silvera GA. The Moderating Role of Hospital Size on the Relationship Between Patient Experience and Patient Safety. Qual Manag Health Care 2017;26:210-7. [Crossref] [PubMed]
48. Kazley AS, Ozcan YA. Electronic medical record use and efficiency: A DEA and windows analysis of hospitals. Socio-Economic Planning Sciences 2009;43:209-16. [Crossref]
49. Case mix index. Available online: https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/AcuteInpatientPPS/Acute-Inpatient-Files-for-Download-Items/CMS022630
50. Chirikos TN, Sear AM. Measuring hospital efficiency: a comparison of two approaches. Health Serv Res 2000;34:1389-408. [PubMed]
51. White KR, Ozcan YA. Church ownership and hospital efficiency. Hosp Health Serv Adm 1996;41:297-310. [PubMed]
52. Harris J, Ozgen H, Ozcan Y. Do mergers enhance the performance of hospital efficiency? J Oper Res Soc 2000;51:801-11. [Crossref]
53. Mechanic R, Coleman K, Dobson A. Teaching hospital costs: implications for academic missions in a competitive market. JAMA 1998;280:1015-9. [Crossref] [PubMed]
54. Rosko M, Wong HS, Mutter R. Characteristics of High- and Low-Efficiency Hospitals. Med Care Res Rev 2018;75:454-78. [Crossref] [PubMed]
55. Garcia-Lacalle J, Martin E. Rural vs urban hospital performance in a ‘competitive’public health service. Soc Sci Med 2010;71:1131-40. [Crossref] [PubMed]
56. Pink GH, Murray MA, McKillop I. Hospital efficiency and patient satisfaction. Health Serv Manage Res 2003;16:24-38. [Crossref] [PubMed]
57. Weinstock DS. Using the Herfindahl Index to measure concentration. Antitrust Bull 1982;27:285-301. [Crossref]
58. Rhoades SA. The herfindahl-hirschman index. Fed Res Bull 1993;79:188.
59. Banker R, Charnes A, Cooper W. Some models for estimating technical and scale inefficiencies in data envelopment analysis. Manag Sci 1984;30:1078-92. [Crossref]
60. Malik M, Efendi S, Zarlis M. Data Envelopment Analysis (DEA) model in operation management. IOP Conf Ser: Mater Sci Eng 2018;300:013008.
61. SimmJBesstremyannayaG. Package ‘rDEA’. 2023.
62. Kooreman P. Data envelopment analysis and parametric frontier estimation: Complementary tools. J Health Econ 1994;13:345-6. [Crossref]
63. Simar L, Wilson PW. Estimation and inference in two-stage, semi-parametric models of production processes. J Econometrics 2007;136:31-64. [Crossref]
64. Antunes BBP, Bastos LSL, Hamacher S, et al. Using data envelopment analysis to perform benchmarking in intensive care units. PLoS One 2021;16:e0260025. [Crossref] [PubMed]
65. Machta RM, Maurer KA, Jones DJ, et al. A systematic review of vertical integration and quality of care, efficiency, and patient-centered outcomes. Health Care Manage Rev 2019;44:159-73. [Crossref] [PubMed]
66. Li X. The association between physician-hospital integration and the performance of acute care hospitals. The University of Texas Health Science Center at Houston; 2023.
67. Tiemann O, Schreyögg J, Busse R. Hospital ownership and efficiency: a review of studies with particular focus on Germany. Health Policy 2012;104:163-71. [Crossref] [PubMed]
68. Dunsire A, Hartley K, Parker D, et al. Organizational status and performance: A conceptual framework for testing public choice theories. Publ Admin 1988;66:363-88. [Crossref]
69. Chen A, Hwang Y, Shao B. Measurement and sources of overall and input inefficiencies: Evidences and implications in hospital services. Eur J Oper Res 2005;161:447-68. [Crossref]
70. Valdmanis VG, Rosko MD, Mutter RL. Hospital quality, efficiency, and input slack differentials. Health Serv Res 2008;43:1830-48. [Crossref] [PubMed]
71. Sloan FA, Valdmanis VG. Relative Productivity of For-Profit Hospitals: A Big or a Little Deal? Med Care Res Rev 2023;80:355-71. [Crossref] [PubMed]
72. Kaiser-Family-Foundation. Hospitals by ownership type 2023. Available online: https://www.kff.org/other/state-indicator/hospitals-by-ownership/?activeTab=graph&currentTimeframe=0&startTimeframe=22&sortModel=%7B%22colId%22:%22Location%22,%22sort%22:%22asc%22%7D
73. Chang J, Zhang L. Case Mix Index weighted multi-objective optimization of inpatient bed allocation in general hospital. J Comb Optim 2019;37:1-19. [Crossref]
74. Sommersguter-Reichmann M. Health care quality in nonparametric efficiency studies: A review. Central Europ. J Operations Res 2022;30:67-131. [Crossref]