Hospitals’ Strategies to Reduce Costs and Improve Quality: Survey of Hospital Leaders

ABSTRACT

Objectives: Hospitals in the US operate under various value-based payment programs, but little is known regarding the strategies they use in this context to improve quality and reduce costs, overall or in voluntary programs including Bundled Payments for Care Improvement Advanced (BPCI-A).

Study Design: A survey was administered to hospital leaders at 588 randomly selected acute care hospitals, with oversampling of BPCI-A participants, from November 2020 to June 2021. Twenty strategies and 20 barriers were queried in 4 domains: inpatient, postacute, outpatient, and community resources for vulnerable patients.

Methods: Summary statistics were tabulated, and responses were adjusted for sampling strategy and nonresponse.

Results: There were 203 respondents (35%), of which 159 (78%) were BPCI-A participants and 44 (22%) were nonparticipants. On average, respondents reported implementing 89% of queried strategies in the inpatient domain, such as care pathways or predictive analytics; 65% of postacute strategies, such as forming partnerships with skilled nursing facilities; 84% of outpatient strategies, such as scheduling close follow-up to prevent emergency department visits/hospitalizations; and 82% of strategies aimed at high-risk populations, such as building connections with community resources. There were no differences between BPCI-A and non–BPCI-A hospitals in 19 of 20 care redesign strategies queried. However, 78.3% of BPCI-A–participating hospitals reported programs aimed at reducing utilization of skilled nursing and inpatient rehabilitation facilities compared with 37.6% of non–BPCI-A hospitals (P < .0001).

Conclusions: Hospitals pursue a broad range of efforts to improve quality. BPCI-A hospitals have attempted to reduce use of postacute care, but otherwise the strategies they pursue are similar to other hospitals.

Am J Manag Care. 2024;30(8):e240-e246. https://doi.org/10.37765/ajmc.2024.89593

Takeaway Points

• A decade after the implementation of value-based payment models, little is known about the strategies hospitals currently use to improve outcomes and reduce costs.
• On average, across 20 strategies in 4 domains—inpatient, postacute, outpatient, and community resources for vulnerable patients—hospitals reported having implemented between 65% and 89% of the strategies queried.
• A higher proportion of hospitals participating in bundled payments implemented interventions aimed at reducing postacute care compared with other hospitals (78.3% vs 37.6%; P < .0001), but patterns were otherwise similar.

Over the past 2 decades, CMS has increasingly used public reporting, value-based payment, and alternative payment models to encourage hospitals and other health care providers to deliver less-expensive and higher-quality care. Initially there was substantial emphasis on improving quality performance as defined by process measures, but more recently the emphasis has changed to focus to a greater degree on patient outcomes, patient safety, and efficiency of care delivery.

Today, the vast majority of US general acute care hospitals operate under value-based payment programs such as the Hospital Value-Based Purchasing Program or the Hospital Readmissions Reduction Program.^1,2 In addition, CMS administers a number of voluntary, less-subscribed alternative payment models, such as accountable care organizations and episode-based payment models. The Bundled Payments for Care Improvement (BPCI) model, introduced by CMS in 2013 and rereleased as BPCI Advanced (BPCI-A) in 2018, is an ongoing episode-based alternative payment model that was designed to incent care redesign and improve care coordination for patients hospitalized with an acute medical or surgical condition.^3,4 The voluntary program holds participating hospitals accountable for costs and quality over a 90-day episode of care. If cost targets for the episode are met, then the hospital retains a portion of the savings; if cost targets are not met, the hospital is penalized.

As CMS and other payers seek to drive quality improvement, understanding the clinical strategies used in response to alternative payment models as well as barriers to successful implementation of these strategies is crucial to optimize program design and ensure that programs have their intended effects. However, little is known about the clinical approaches used by hospitals to deliver less-expensive and higher-quality care or the barriers and challenges they face in that effort. To address this gap, we conducted a national survey of hospital leaders to understand the approaches they have adopted to deliver higher-quality, less-costly care and the barriers that hospitals face to making their efforts successful. We designed our sample to provide national estimates and to compare the differences between hospitals that participated in BPCI-A and those that did not.

METHODS

Survey Development and Administration

To identify survey domains, we conducted a literature review focused on use of various clinical strategies to improve quality and decrease costs and on barriers to successful implementation. The survey was also based in part on our prior survey work with the Hospital Readmissions Reduction Program.^5,6 From these domains, we created and tested an initial survey instrument using cognitive interviews with 6 members of an advisory panel representing a range of expertise in hospital leadership and alternative payment models. Following the cognitive interviews, we adjusted the survey based on the advisory panel’s feedback and obtained another round of feedback from the panel. In examining strategies to improve quality and reduce costs as well as barriers to doing so, we focused on 4 domains: inpatient care, postacute care, outpatient care, and community resources for vulnerable patients. The finalized survey was scheduled to enter the field in April 2020. However, due to the COVID-19 pandemic, field work was delayed until November 2020. Survey questions relevant to this manuscript are available in the eAppendix (available at ajmc.com).

Based on sample size calculations, we aimed to survey 600 hospitals. We oversampled hospitals participating in BPCI-A at a 3:1 ratio based on power calculations conducted prior to survey development by selecting 450 hospitals at random from a public list of BPCI-A participants (roughly 825 hospitals at the time of survey administration) and 150 hospitals at random from the remaining general acute care US hospitals paid under the Inpatient Prospective Payment System (roughly 2200 hospitals at the time of survey administration).

To identify hospital leaders, we obtained the hospital leadership list of chief medical officers (CMOs) from the American Hospital Association. A leader’s hospital was moved into the active field stage after study staff verified each leader’s contact information by telephone. Based on this screening step, our final sample included 588 hospitals (442 BPCI-A participants and 146 nonparticipants) and excluded 12 hospitals that had closed, merged with other hospitals, or become critical access hospitals or long-term care facilities. SSRS (Glen Mills, PA) conducted the survey, which was in the field in pilot phase beginning September 22, 2020, and in 2 waves of the main release between November 1, 2020, and June 27, 2021. An initial premailing was followed a week later by a second mailing that included an incentive check for $100 along with a hard copy of the survey with a cover letter explaining the intent of the survey and the consent process. It also contained a personalized link should the recipient wish to complete the survey online. A third mailing followed 3 weeks later, and phone reminders commenced thereafter. A fourth mailing, along with a $10 Amazon gift card, was sent 1 week after the phone reminders began. Despite the CMO being our initial point of contact, we encouraged respondents to reach out to other leaders within the hospital best equipped to provide assistance or complete the survey.

Hospital Characteristics

Further information on hospital characteristics was derived from the American Hospital Association Annual Survey for 2019. These characteristics included profit status, hospital size, teaching status, urban/rural location, and region. Following CMS’ approach for the Hospital Readmissions Reduction Program, we defined safety-net hospitals as those with the highest quintile of dual Medicare/Medicaid enrollment among Medicare admissions.^7-10 Using a similar strategy, we defined disproportionately high-minority hospitals as those with the highest quintile of their proportion of identified Black or Hispanic Medicare admissions based on Medicare enrollment files.^9,11,12

Analysis

We asked each hospital whether it “currently used” each of
20 quality improvement strategies and whether each of 20 barriers to implementation was “significant” at their hospital. To better reflect a national representation of US hospitals, survey responses were adjusted for both sampling strategy and nonresponse. We assigned sample weights to hospitals to adjust for sampling strategy based on BPCI-A status, with weights representing the inverse probability of each hospital’s selection. To adjust for nonresponse, we constructed a logistic regression model in which returning the survey was the primary outcome and hospital characteristics—including size, teaching status, ownership, urban location, and region—were predictors, as has been done previously.^5,6 Each hospital received a likelihood of response based on this model; responses were then weighted with the inverse of this likelihood. Using generalized estimating equations with both sample and nonresponse weights, we conducted analyses to estimate the national prevalence of utilization of clinical approaches and the perception of barriers to care. We also compared hospitals participating in BPCI-A with nonparticipating hospitals. Item nonresponse ranged from 2 of 203 respondents to 17 of 203 for the questions reported in this manuscript, with a mean of 7.5; we omitted nonrespondents from the numerator and denominator for each question.

We considered a P value less than .05 as statistically significant. Because of the multiple comparisons, our findings should be regarded as hypothesis generating. All responses were deidentified before analysis. Informed consent was obtained as part of the survey process; the introductory page to the survey included detailed information about privacy and data deidentification and stated, “Completion of this survey implies informed consent.” The study was approved by the Human Research Protection Office at Washington University in St Louis.

RESULTS

Hospital Characteristics

Of the 588 hospitals to which surveys were sent, 203 responded (response rate, 34.5%). Of the 203 respondents, two-thirds were chief medical/clinical/quality officers (67.0%) and more than a quarter were chief executive/operating officers (27.6%). The remaining respondents were either vice president or head of population health or equivalent (3.9%) or director of case management, care coordination, or equivalent (2.5%) (Table 1).

Of the responding hospitals, 159 (78%) were BPCI-A participants and 44 (22%) were nonparticipants. Respondent institutions tended to be not-for-profit (72.9%), medium-sized (63.5%), and minor teaching (41.9%) hospitals, and the vast majority were located in urban areas (95.1%). In unweighted comparisons, survey nonrespondent hospitals were generally similar in their characteristics, without statistically significant differences (Table 1); weighted comparisons were similar and are shown in eAppendix Table 1. Characteristics by BPCI-A status, unweighted and weighted, are shown in eAppendix Tables 2 and 3.

Strategies to Improve Quality and Reduce Cost

The survey inquired about 20 specific approaches to improve quality and reduce cost (Figure 1 and eAppendix Table 4), spread across 4 domains. In the inpatient domain, on average, hospitals reported using 89.4% of strategies queried. The most commonly reported strategy in this domain was team-based discharge planning (98.9%), followed by scheduling a follow-up within a short period of time after discharge (90.5%). In the postacute domain, on average, hospitals reported using 64.9% of strategies queried. The most commonly reported strategy in this domain was monitoring patients’ length of stay (79.3%), followed by selectively referring patients to skilled nursing facilities (SNFs) based on quality (74.7%). In the outpatient domain, on average, hospitals reported using 84.3% of strategies queried. Scheduling follow-up prior to discharge (90.1%) and developing programs to deliver innovative outpatient care (77.5%) were both commonly used. In the community resources for vulnerable patients domain, on average, hospitals reported using 81.7% of strategies queried. Screening patients for frailty (93.4%), screening for mental or behavioral health needs (89.0%) or social determinants of health (84.6%), and referring to community resources for social needs (84.0%) were common.

Barriers to Improving Quality or Reducing Cost

The survey inquired about the presence of 20 specific barriers to improving quality or reducing cost across 4 domains (Figure 2 and eAppendix Table 5). In the inpatient domain, on average, hospitals reported facing 56.9% of barriers queried. The most commonly reported barrier in this domain was competing interests (67.4%), followed by a lack of financial resources to implement new programs (59.6%). In the postacute domain, on average, hospitals reported facing 50.3% of barriers queried. The most commonly reported barrier in this domain was the lack of control over clinical care delivered in the postacute setting (73.4%), followed by a lack of follow-up after the postacute care episode (68.0%). In the outpatient domain, on average, hospitals reported facing 64.2% of barriers queried. Patients’ inability to follow treatment recommendations (88.2%) and inability to attend follow-up care visits (63.2%) were both commonly reported. In the community resources for vulnerable patients domain, on average, hospitals reported facing 73.5% of barriers queried, the highest of the 4 domains. Difficulties related to patients’ lack of social support (93.6%), a lack of access to mental health care (89.3%), and a lack of access to substance use disorder treatment (84.9%) were common.

Strategies and Barriers by BPCI-A Status

We compared use of clinical approaches designed to improve care or lower utilization between BPCI-A participants and nonparticipants, controlling for sampling and nonresponse (Table 2). In 19 of the 20 clinical strategies we examined, there was no statistical difference between participants in BPCI-A and nonparticipants. The one area with a difference was in programs to reduce institutional postacute care at SNFs and inpatient rehabilitation facilities: 78% of BPCI-A hospitals employed such programs, whereas only 38% of nonparticipants employed such an approach (P < .0001).

We also compared BPCI-A hospitals with nonparticipants on the 20 barriers to implementation that we identified. Again in 19 of 20 instances, there was no difference statistically. However, respondents from BPCI-A hospitals were more likely to say they had difficulty implementing change because of resistance from clinicians (52.7% vs 35.2%; P = .03) (Table 3).

DISCUSSION

In this national survey of hospital leaders, we found that the 20 strategies we identified to improve quality and reduce cost are commonly used. More than 80% of hospitals reported that they currently use most of the strategies we queried, and the proportion of hospitals using each approach ranged from 39% to 99%. Although a lower overall proportion of hospitals reported barriers, ranging from 26% to 94% by item, many barriers in the vulnerable populations’ domain were reported by more than 80% of hospitals. Finally, there was little difference between hospitals participating in BPCI-A and those that were not participating. The only major exception was the large difference in the effort by BPCI-A hospitals to reduce postacute care at SNFs and inpatient rehabilitation facilities.

It was not entirely unexpected that the differences related to BPCI-A participation would be most prominent in the domain of postacute care. Under episode payments, there are obvious financial incentives to reduce the proportion of patients receiving postacute care and their length of stay if such care is initiated. In fact, prior studies have shown that the greatest reduction in cost among BPCI-A–participating hospitals was due to the reduction of institutional postacute care.¹³ Supporting this notion, work by the National Academy of Medicine has shown that postacute care also accounts for the largest portion of regional variation in health expenditures.¹⁴ There has also been widespread concern that postacute care is overused for Medicare patients, making it an appealing target among policy makers and perhaps also among clinical leaders. We also found that BPCI-A hospitals were more likely to use predictive analytics. Although we did not inquire as to what type of predictive tools hospitals used, we suspect that financial incentives in BPCI-A related to readmissions could have prompted increased use of tools such as readmission prediction scores that could facilitate interventions to reduce utilization.

More broadly, however, it seems likely that hospitals working alone will have difficulty addressing some of the barriers that may impede the delivery of high-quality care. We found that some of the greatest barriers to pursuing clinical strategies were difficulties related to inadequate availability of mental/behavioral health services and inadequate availability of substance use treatment services. Partnerships with mental health providers to provide support to patients outside the hospital may be necessary to improve quality and outcomes in health care, but these providers are in very short supply nationwide. Another related frequently cited barrier was a lack of inpatient personnel to support care innovations. Broad workforce-related challenges in hospital care, particularly in the inpatient nursing workforce, are likely to remain significant barriers to the pursuit of some important clinical strategies in the coming years. Lack of personnel has been highlighted recently as an important and commonly reported barrier to telemedicine utilization,¹⁵ and this problem was likely significantly worsened by the COVID-19 pandemic.

One area of particular concern is care for vulnerable and historically marginalized populations that face challenges specifically related to social determinants of health and health-related social needs. It is well demonstrated that issues such as patients’ lack of access to care, financial challenges, and lack of social support have a significant impact on health care outcomes and costs. There is evidence that supports the use of strategies pointed at social determinants of health, which has been associated with improvements in health outcomes.^16-18 Approximately 85% of hospitals reported that they have programs for screening for social determinants of health, but a much smaller proportion actually provided resources related to these needs. Thus, hospitals appear to be aware of these barriers that are particularly difficult for them to address because of their limited reach and inability to address issues outside the hospital.^6,19,20 Whether these issues can best be addressed through partnerships with community-based organizations or through direct service provision is unknown, although partnerships have the added benefit of facilitating investment in communities, which may lead to more sustained improvements in health outcomes.

Limitations

The study has several limitations. Our response rate was 34.5%. We devoted substantial resources to the survey effort and hypothesize that hospitals suffering the greatest burden from the pandemic may have been less likely to respond. We reweighted to compensate for nonresponse and cannot predict the impact of any residual bias on our findings. We did not reweight each item for item nonresponse. We recognize that others who might develop similar lists of strategies to improve care and reduce utilization and barriers that could impede those efforts might not match our lists precisely. We are unaware of any prior efforts to develop similar lists and suggest that these might serve as a good starting point for others in the future. Some of the individuals completing the survey may not have had the knowledge regarding the questions asked despite directions and communication with the office of the CMO. Given the broad proliferation of value-based payment arrangements in Medicare and among private payers, some of the strategies reported by BPCI-A participants may reflect participation in other programs rather than solely participation in BPCI-A. Although we attempted to reduce social acceptability bias on each of the data elements, this source of bias may also influence answers. We used hospitals’ fee-for-service Medicare patients to define high-minority and safety-net hospitals, which may not precisely correlate with the racial or economic breakdown of their overall patient population.

CONCLUSIONS

We identified 20 clinical strategies that are commonly used to foster higher-quality and less-costly care and 20 barriers that impede hospitals’ ability to achieve these goals. Hospitals reported implementing the majority of the strategies we queried. BPCI-A participation was associated with higher reported use of strategies to reduce postacute care costs, but otherwise, clinical approaches used by BPCI-A participants were similar to those used by nonparticipants. Broad approaches may be required to aid hospitals in achieving wider changes in quality and efficiency.

Author Affiliations: Cardiovascular Division, Department of Medicine, Washington University School of Medicine in St Louis (GH, TL, FW, KEJM), St Louis, MO; Center for Advancing Health Services, Policy and Economics Research, Washington University School of Medicine and Institute for Public Health (GH, RJW, KEJM), St Louis, MO; Division of Biostatistics, Institute for Informatics, Data Science & Biostatistics, Washington University School of Medicine in St Louis (RJW), St Louis, MO; Department of Health Policy and Management, Harvard T.H. Chan School of Public Health (JZ, AME), Boston, MA; Division of General Internal Medicine and Primary Care, Department of Medicine, Brigham and Women’s Hospital (EJO), Boston, MA.

Source of Funding: National Heart, Lung, and Blood Institute (R01HL143421).

Author Disclosures: Dr Joynt Maddox reports prior membership on the Centene Health Policy Advisory Council, attendance at meetings of the American Heart Association and Journal of the American Medical Association, and grants received from the National Heart, Lung, and Blood Institute (R01HL164561); National Institute of Nursing Research (U01NR020555); National Institute on Aging (R01AG060935, R01AG063759, and R21AG065526); National Center for Advancing Translational Sciences (UL1TR002345); and Humana. The remaining authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (GH, AME, EJO, KEJM); acquisition of data (AME, EJO, KEJM); analysis and interpretation of data (TL, FW, RJW, JZ, AME, EJO, KEJM); drafting of the manuscript (GH, TL, KEJM); critical revision of the manuscript for important intellectual content (GH, TL, FW, RJW, JZ, AME, EJO, KEJM); statistical analysis (FW, RJW, JZ, EJO); obtaining funding (AME, EJO, KEJM); administrative, technical, or logistic support (TL); and supervision (KEJM).

Address Correspondence to: Karen E. Joynt Maddox, MD, MPH, Washington University School of Medicine, 660 S Euclid Ave, St Louis, MO 63110. Email: kjoyntmaddox@wustl.edu.

REFERENCES

1. The Hospital Value-Based Purchasing (VBP) Program. CMS. Updated February 6, 2024. Accessed July 10, 2024. https://www.cms.gov/medicare/quality/value-based-programs/hospital-purchasing

2. FY 2013 IPPS final rule: Hospital Readmissions Reduction Program supplemental data file. CMS. Updated March 2013. Accessed July 10, 2024. https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/AcuteInpatientPPS/Archived-Supplemental-Data-Files/FY2013-IPPS-Final-Rule-HRRP-Supplemental-Data-File

3. Bundled Payments for Care Improvement (BPCI) initiative: general information. CMS. Accessed July 10, 2024. https://www.cms.gov/priorities/innovation/innovation-models/bundled-payments

4. BPCI Advanced. CMS. Accessed July 10, 2024. https://www.cms.gov/priorities/innovation/innovation-models/bpci-advanced

5. Joynt KE, Figueroa JE, Oray J, Jha AK. Opinions on the Hospital Readmission Reduction Program: results of a national survey of hospital leaders. Am J Manag Care. 2016;22(8):e287-e294.

6. Figueroa JF, Joynt KE, Zhou X, Orav EJ, Jha AK. Safety-net hospitals face more barriers yet use fewer strategies to reduce readmissions. Med Care. 2017;55(3):229-235. doi:10.1097/MLR.0000000000000687

7. 21st Century Cures Act, 42 USC (2016). Accessed July 10, 2024. https://www.govinfo.gov/content/pkg/PLAW-114publ255/pdf/PLAW-114publ255.pdf

8. Report to Congress: Social Risk Factors and Performance Under Medicare’s Value-Based Purchasing Programs. HHS Office of the Assistant Secretary for Planning and Evaluation; December 2016. Accessed July 10, 2024. https://aspe.hhs.gov/sites/default/files/migrated_legacy_files//171041/ASPESESRTCfull.pdf

9. Shashikumar SA, Waken RJ, Luke AA, Nerenz DR, Joynt Maddox KE. Association of stratification by proportion of patients dually enrolled in Medicare and Medicaid with financial penalties in the Hospital-Acquired Condition Reduction Program. JAMA Intern Med. 2021;181(3):330-338. doi:10.1001/jamainternmed.2020.7386

10. Joynt Maddox KE, Reidhead M, Qi AC, Nerenz DR. Association of stratification by dual enrollment status with financial penalties in the Hospital Readmissions Reduction Program. JAMA Intern Med. 2019;179(6):769-776. doi:10.1001/jamainternmed.2019.0117

11. Figueroa JF, Joynt KE, Zhou X, Orav EJ, Jha AK. Safety-net hospitals face more barriers yet use fewer strategies to reduce readmissions. Med Care. 2017;55(3):229-235. doi:10.1097/MLR.0000000000000687

12. Johnston KJ, Wiemken TL, Hockenberry JM, Figueroa JF, Joynt Maddox KE. Association of clinician health system affiliation with outpatient performance ratings in the Medicare Merit-based Incentive Payment System. JAMA. 2020;324(10):984-992. doi:10.1001/jama.2020.13136

13. Joynt Maddox KE, Orav EJ, Zheng J, Epstein AM. Year 1 of the Bundled Payments for Care Improvement-Advanced model. N Engl J Med. 2021;385(7):618-627. doi:10.1056/NEJMsa2033678

14. Newhouse JP, Garber AM. Geographic variation in Medicare services. N Engl J Med. 2013;368(16):1465-1468. doi:10.1056/NEJMp1302981

15. Payán DD, Frehn JL, Garcia L, Tierney AA, Rodriguez HP. Telemedicine implementation and use in community health centers during COVID-19: clinic personnel and patient perspectives. SSM Qual Res Health. 2022;2:100054. doi:10.1016/j.ssmqr.2022.100054

16. Hammond G, Johnston K, Huang K, Joynt Maddox KE. Social determinants of health improve predictive accuracy of clinical risk models for cardiovascular hospitalization, annual cost, and death. Circ Cardiovasc Qual Outcomes. 2020;13(6):e006752. doi:10.1161/CIRCOUTCOMES.120.006752

17. Parish W, Beil H, He F, et al. Health care impacts of resource navigation for health-related social needs in the Accountable Health Communities Model. Health Aff (Millwood). 2023;42(6):822-831. doi:10.1377/hlthaff.2022.01502

18. Renaud J, McClellan SR, DePriest K, et al. Addressing health-related social needs via community resources: lessons from Accountable Health Communities. Health Aff (Millwood). 2023;42(6):832-840. doi:10.1377/hlthaff.2022.01507

19. Figueroa JF, Duggan C, Toledo-Cornell C, Zheng J, Orav EJ, Tsai TC. Assessment of strategies used in US hospitals to address social needs during the COVID-19 pandemic. JAMA Health Forum. 2022;3(10):e223764. doi:10.1001/jamahealthforum.2022.3764

20. Fraze TK, Brewster AL, Lewis VA, Beidler LB, Murray GF, Colla CH. Prevalence of screening for food insecurity, housing instability, utility needs, transportation needs, and interpersonal violence by US physician practices and hospitals. JAMA Netw Open. 2019;2(9):e1911514. doi:10.1001/jamanetworkopen.2019.11514