Article Text
Abstract
Background Hospital readmissions are an important quality of care indicator and are tied to hospitals’ financial reimbursements. Safety-net hospitals, which serve a high proportion of patients of low socioeconomic status (SES), face unique challenges to reduce or maintain low readmission rates.
Objective We sought to understand strategies high-performing safety-net hospitals used to achieve low 30-day risk-standardised readmission rates (RSRRs) using qualitative methodology.
Methods Safety-net hospital status was defined by public ownership or a Medicaid population that is greater than 1 SD higher than the state proportion of Medicaid patients and the hospital payer source is composed of at least 15% Medicaid patients. Safety-net hospitals were selected based on their ranking among the lowest 20% of heart failure RSRRs, the best-performing quintile. We purposefully sampled hospitals to ensure variation in characteristics and conducted on-site interviews with key hospital staff. A multidisciplinary team analysed the data using thematic analysis.
Results We performed site visits at 9 safety-net hospitals (RSRR range: 18.1%–21.6%) in 9 states and conducted in-depth interviews with 108 hospital staff. Several thematic attributes and organisational strategies were evident in high-performing safety-net hospitals: (1) strong hospital support for quality improvement at all levels; (2) tailoring resources to meet patient needs; (3) facilitating collaboration and communication among and between providers and patients; (4) creating strong relationships with postacute care facilities and communities and (5) proactive approach to healthcare policy changes and other external factors.
Conclusions The provision of high-quality and equitable care in hospitals serving a high proportion of low-SES populations is influenced by several modifiable factors. These findings may serve to inform lower-performing safety-net hospitals on how to optimise patient care and improve readmission outcomes.
- QUALITATIVE RESEARCH
- Quality in health care
- Hospitals
Data availability statement
No data are available.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
STRENGTHS AND LIMITATIONS OF THIS STUDY
The study methodology allows for the identification of successful strategies from high-performing hospitals, potentially offering actionable insights for other safety-net hospitals.
The research team followed the Standards for Reporting Qualitative Research, ensuring transparency and reproducibility.
The code structure for analysing transcripts was developed collaboratively by experienced researchers, with conflicts resolved by consensus, enhancing reliability and depth of the thematic analysis.
The study recruited 9 hospitals and 108 healthcare personnel, which, while useful for qualitative insights, limits generalisability in comparing hospital strategies.
Introduction
The movement to prioritise reductions in hospital readmissions has gained considerable attention through financial accountability and public reporting programmes in the USA. Key stakeholders, including physicians, policy-makers, scholars and patients, are interested in how patient and hospital factors play a role in reducing unplanned readmissions, improving patient outcomes and closing the equity gap. Some stakeholders remain concerned that safety-net hospitals caring for patients of low socioeconomic status (SES) may put hospitals at a disadvantage regarding readmissions. Safety-net hospitals have a common mission to provide care for Medicaid beneficiaries and those who are uninsured.1
Avoidable hospital readmissions are important indicators of quality of care as they put an unnecessary physical, social and emotional burden on patients and families, incurring risk of iatrogenic disease.2 There are several reasons that caring for low SES patients could lead to challenges with achieving low readmission rates. Patients often present to safety-net hospitals with complex and chronic conditions, substance abuse issues and mental health concerns and have fragmented financial and support resources available.3 4 Safety-net hospitals may also struggle with recruitment and retention of staff, declining reimbursement and scalability issues.5 6 However, prior research demonstrates that safety-net and non-safety-net hospitals have similar 30-day risk-standardised readmission rates (RSRRs) and many hospitals serving a high proportion of low SES patients perform better than the national readmission rate for Medicare fee-for-service beneficiaries with clinical conditions such as heart failure.7–11 Despite the evidence that some safety-net hospitals attain high performance in reducing readmission rates, the methods for achieving this outcome have received insufficient attention to date and little is known about the strategies that are most critical for safety-net hospitals specifically.
Given this gap, we sought to examine the characteristics of high-performing hospitals serving low SES patients, with low 30-day RSRR for heart failure patients by gathering insights from hospital personnel involved in the care of patients. This study characterises specific aspects pertinent to safety-net hospitals regarding organisational culture, quality improvement initiatives and outpatient services with a greater focus than described in prior literature. Our findings regarding the strategies implemented by these hospitals may serve to inform lower-performing safety-net hospitals on how to optimise patient care, close the equity gap and improve readmission outcomes for patients.
Methods
Study design
We used the positive deviance approach12–17 to improve healthcare quality as it assumes that solutions to problems are already available within the community and identifying and sharing those solutions can help others to resolve existing issues. Accordingly, we conducted in-depth, open-ended interviews with key informants comprising hospital, clinical and administrative personnel involved in care of patients with heart failure at high-performing hospitals serving low SES patients to investigate hospital organisational strategies that may be associated with low 30-day readmission rates. We focused on heart failure readmission due to the widespread efforts to reduce readmission rates in this population, the frequent interaction of heart failure patients with multiple hospital departments and the broad impact of care improvements for these patients. The Standards for Reporting Qualitative Research was used in this study.
Patient and public involvement
The results of this publication will be shared with personnel of the participating hospitals who informed our sample.
Development of interview guide
Our team adapted an interview guide based on previous positive deviance investigations12–17 to explore dimensions of enabling structures, processes and hospital internal environments that may influence hospital quality and outcomes. The interview guide consisted of open-ended questions to understand how high-achieving hospitals attained their outcomes while caring for low SES heart failure patients and to identify specific strategies (including tools, supports and specific processes) that may be reproducible in other safety-net hospitals (see online supplemental appendix 2).
Supplemental material
Inclusion criteria and recruitment
We used data from the American Hospital Association’s annual survey (n=6317 hospitals). We applied the following inclusion criteria, which aided in the selection of hospitals: (1) short-stay acute care or critical access hospital; (2) located outside the state of Connecticut to avoid conflicts of interest; (3) safety-net as defined by public ownership or a Medicaid population that is greater than 1 SD higher than the state proportion of Medicaid patients and the hospital payer source is composed of at least 15% Medicaid patients; (4) top 20% heart failure RSRR in the 2014 publicly reported data (based on nationwide admissions from 2010 to 2013) among safety-net hospitals and top 50% heart failure RSRR; (5) top 50% heart failure risk-standardised mortality rates (RSMR) to ensure overall high quality at the hospital and (6) at least 100 heart failure admissions in the reporting period (figure 1). This yielded 43 hospitals from which we purposefully sampled to ensure diversity in the number of heart failure and proportion of Medicaid patients treated, as well as geographic, demographic and community characteristics (eg, proportion of unoccupied homes and proportion of uninsured residents).
Flow chart of hospital-level inclusion criteria for qualitative study. AHA, American Hospital Association; HF, heart failure; RSMR, risk-standardised mortality rate; RSRR, risk-standardised readmission rate.
To recruit hospitals, the project leads (KEM and SMB) first sent a recruitment letter via email to key informants (eg, cardiology, quality improvement or research departments) at selected hospitals. Hospital personnel were then contacted by the study lead to explain the study in detail, address questions, obtain consent to participate and decide on a date for the in-person site visit. The site lead organised the visit and identified the key personnel who would engage in interviews. No incentives were provided for participation in the study. In order to reach theoretical saturation, we anticipated the need to recruit 7–9 high-performing safety-net hospitals.
Data collection and analysis
We conducted in-person site visits between April 2014 and June 2017 at nine hospitals. Each site visit included a 2–3 member research team with diverse backgrounds in internal medicine, public health and health services research. All members had backgrounds in quality improvement and were trained in qualitative interviewing methods.
We conducted interviews with hospital, clinical and administrative personnel involved in the care and/or outcomes of patients with heart failure. Whenever possible, we sought to interview personnel individually and not in groups. All interviews were audio recorded (average length 30 min) and professionally transcribed, with identifying features redacted manually from the transcripts prior to analysis to preserve respondent confidentiality. Verbal consent was obtained from all participants prior to the start of the interviews.
Three health services researchers (KEM, PC and SMB) used thematic analysis to identify, analyse and report themes within data.18 Initially, we immersed ourselves in the data, reading and re-reading it to gain a deep understanding while noting initial observations. We then developed a code structure to guide the coding of transcripts. The code structure was developed using a collaborative process. Two researchers (KEM and PC) then independently read and serially coded transcripts from all interviews. Conflicts in interpretation of the transcripts were resolved through consensus and consistency in coding was maintained through thematic analysis. Next, related codes were collated to form potential themes. We reviewed and refined these themes, ensuring they accurately represented the data and were distinct from one another, revisiting the data for validation. Once themes were well defined, each was analysed in-depth and given a clear, concise name and definition. Our analysis identified several recurrent and unifying themes related to the study’s aims. Qualitative analysis software (ATLAS.ti, V.8) was used to organise the data and facilitate retrieval.
Patient and public involvement
The results of this publication will be shared with the hospital personnel of the participating hospitals who informed our sample and served as the lead contact.
Results
We interviewed 108 key staff in a variety of roles (see figure 2) from nine high-performing safety-net hospitals, each varying in geography, volume, proportion of Medicaid patients, teaching hospital status and ownership (table 1).
Number of interviews conducted by respondent role. CMO, chief medical officer.
Participating hospital readmission rates, volume and characteristics of sample by US census region and performance
We identified five recurrent and unifying themes to illustrate the organisational practices of high-performing hospitals serving predominantly low SES populations. We grouped these themes into topic areas and described each in more detail, with illustrative quotes shown in table 2 (further detail in online supplemental appendix table 1).
Supplemental material
Themes, subthemes and illustrative quotes
Strong hospital support for quality improvement at all levels
Hospitals demonstrated a commitment to investing in staff development across occupational groups (ie, hospital administrators, physicians, and medical and support staff) with a clear quality improvement objective. This was demonstrated through providing staff with professional development support—both financial support and dedicated time—to attend conferences, trainings and case review meetings. Some hospitals provided resources for leadership and management to attend large-scale external quality improvement trainings including Lean Six Sigma and Virginia Mason training. Several hospitals created new, full-time positions or new roles within existing positions specifically focused on improving the quality of care and meeting patient needs. For example, one hospital created a new position for a nurse to coordinate all postdischarge care for heart failure patients from rehabilitation to follow-up appointments and group information sessions. Another hospital incentivised physician participation in quality improvement by allocating a portion of physicians’ weekly effort to quality improvement.
Respondents described high-performing hospitals as having a ‘low-bar’ for getting involved in quality improvement and catalysing staff engagement at all levels. They further described organisational hierarchies conducive to quality improvement projects and responsive to internal staff proposals. To achieve this, hospital leadership met with staff to identify quality improvement projects and provided financial resources, staffing and effort to support the initiative. Importantly, staff reported having autonomy to exercise discretion with the development and ‘testing’ of quality improvement projects, while also not feeling personally blamed for any perceived shortcomings.
Hospitals also made significant investments in data and often used data to drive organisational change. Many respondents described efforts to consistently measure outcomes—even when mandates to collect data had subsided. Several hospitals were early adopters and members of external quality improvement networks, such as national registries, the Premier Alliance and the Institute for Healthcare Improvement, which use data to facilitate change. Hospitals became ‘data-centric’ by investing in tools and technology to analyse data. Respondents worked with IT to design custom clinical workflow tools, invested in external software or partnered with data analytics companies to run reports on trends in patient outcomes for review with care teams to analyse data systematically and efficiently.
Tailoring resources to patient needs
Staff reported hospital leadership communicated a focus on patient-centred care across institutions in our study. Respondents described working at organisations where care was both holistic and tailored to individual patient needs. Often respondents perceived the driving force behind patient-centred care to be the mission of the hospital as well as the patient population and community that the hospital served.
Identification of, and tailored responses to, individual patient readmission risks, whether due to clinical or socioeconomic factors, was an approach used across hospitals to identify patients at risk for readmission. For example, one hospital used medical informatics to develop a home-grown risk assessment tool that included social determinants of health when considering readmission risk.
Hospitals also tailored resources that were initiated in the hospital and extended to outpatient settings to meet patient needs. Such programmes focused on identifying community resources to support access to food and housing while also addressing behaviours that affected health status. Many hospitals had innovative programmes that helped patients to access or afford medications, such as those that sought to bridge drug benefits, use connections with pharmaceutical companies and fill prescriptions at the bedside. Study hospitals also coordinated patient discharge with community-based programmes, such as facilitating Meals on Wheels services, and by supporting patients and providers in the transition of care to home or postacute care settings via patient navigators, care coordinators or follow-up call centres
Rather than a priori deciding which services patients needed, hospital staff described ‘meeting patients where they are’ to minimise the burden patients faced in follow-up care. This required an understanding of the specifics of patients’ lives. All respondents noted that a high proportion of their hospital census comprised vulnerable patient populations and noted the importance of identifying and understanding social determinants of health to improve care delivery. Other respondents emphasised using teach-back techniques, providing transportation to follow-up appointments and preparing documents in multiple languages as ways to minimise burden on patients. Additionally, respondents pointed to systems with acute care facilities in the community, rather than just in the hospital, as helpful to improve patient care and outcomes. Respondents also described ways to involve patient families or caretakers who could advocate for patients’ needs and share the responsibility of caring for patients postdischarge.
Hospitals also tailored education to patient SES. Many educational programmes had been in existence for years and often there were key hospital staff with whom patients established relationships. In more remote regions, where it may be difficult for patients to travel to the hospital, hospitals relied on in-home as well as phone-based follow-up protocols to engage patients in care. Some hospitals had specific postacute heart failure programmes geared towards providing resources, as well as teaching patients about their condition and how to manage symptoms appropriately. Integrating patients into the self-care process helped them to feel responsible for their own health trajectory.
Facilitates collaboration and communication among and between providers and patients
Organisational culture of hospitals focused on building and maintaining effective communication, collaboration and teamwork skills among and between providers and patients. To accomplish this, hospitals used structural factors that facilitated communication, as well as supportive leadership—including both physicians and nurses—to establish the tone and culture of the hospital.
Communication was fostered by structural factors, including modes of communication. Among larger hospitals, respondents used electronic health record software or encrypted texting to facilitate communication. In contrast, among smaller community hospitals, staff described the ease of in-person communication due to close proximity.
Hospitals also used several mechanisms to support an organisational culture that fostered teamwork. Some hospitals had multidisciplinary patient care teams that enabled and encouraged physicians, nurses, case managers and other clinical and non-clinical staff to work together. In other hospitals, communication was facilitated during interdisciplinary rounds, or during rounds focused specifically on social issues and discharge planning. Other structural means of communication and collaboration included integrating pharmacists on hospital floors for ease of access to both patients and providers. Regardless of the mechanism, collaboration was a central tenet of the hospital and an expectation for providers.
Supportive leadership at hospitals, facilitated by both physicians and nurses, helped to establish the tone and culture of the organisation. Tone was set by enthusiastic leaders who championed an inclusive vision and mission throughout the hospital. This improved team dynamics and effectiveness. A commitment to teamwork was emphasised in hiring individuals to leadership roles. Culture was also established through frequent communications in which staff all ‘spoke the same language’.
Creating strong relationships with postacute care facilities and communities in a variety of ways
Hospitals used a variety of strategies to create strong relationships with postacute care facilities and communities, such as providing education and assistance with transitions of care. This support was facilitated through building robust relationships with postacute care facilities and communities, including frequent communications, promoting trust and establishing rapport and relying on partnerships.
Many hospitals provided clinical education to postacute care facility staff, for whom such opportunities may not otherwise exist. For instance, one hospital realised that heart failure patients discharged to a specific nursing home were frequently readmitted to the hospital. To address this concern, the hospital sent staff to train nursing home aides on when a patient’s condition could be managed at the nursing home, as opposed to requiring emergent care.
Hospitals also used both written and verbal communication with outside providers to facilitate communication efforts, such as holding in-person meetings with external facilities and encouraging open lines of communication at all times. Among hospitals affiliated with a health system, inclusive ecosystems enabled hospitals to benefit from coordinated care and ease of access between facilities. For example, having shared electronic health records across a hospital system enhanced interoperability and provided for smooth communication between sister hospitals and system-owned skilled nursing facilities and long-term care facilities.
In addition to supporting postacute facilities, hospitals provided significant support to community organisations by allowing space for meetings, providing financial contributions, health vans and localised clinics, among others.
While hospitals often had a robust relationship with local communities, that support was also often bidirectional, and many respondents described ways in which their communities supported the hospital. For instance, community members volunteered at the hospital or providing financial support for expansion, renovation or growth. One study community voted for a tax hike to support the construction of a new public hospital.
Another component of cultivating relationships was building trust with outpatient providers, especially among community-based physicians who may have concerns about encroachment on their patient populations. To address this, hospitals frequently communicated with outpatient providers during and following a patient’s hospital stay. Explicit connections were often made between hospital staff and providers, and ‘warm handoffs’ were commonplace to external facilities such as skilled nursing or long-term care facilities.
Importantly, respondents reported substantial and explicit efforts to build relationships between hospital and outpatient providers. Some hospitals helped achieve this through hiring people with professional experience in outpatient settings. Other hospitals noted that employee longevity also tracked with improved overall understanding of relationships inside and outside of the hospital. Retention and strategic partnerships were proxies of a good relationship between hospitals and postacute care facilities.
Proactive approach to healthcare policy changes and other external factors
While hospitals had unique internal environments and robust partnerships with outpatient providers, respondents also noted the influence of factors external to the hospital or community that could spur innovation. In this sense, hospitals were proactive in responding to recent changes in health policy, the shift to value-based payments, public reporting of quality and outcomes, and payment models such as risk-based contracts.
Numerous respondents were early to respond to the Centers for Medicare and Medicaid Services (CMS) Hospital Readmission Reduction Programme, for example, as characterised by a proactive focus on quality and the need to institute efforts to reduce readmission rates. In some hospitals, the focus on quality improvement was longstanding; in others, it was relatively new. Nevertheless, having a quality-focused culture in place primed hospitals to think about quality in a new way when government policy decisions were implemented. Furthermore, many respondents spoke openly about the potential implications of ongoing and future policy changes for quality measurement, including the Medicare Access and Reauthorization Act of 2015, the Merit-based Incentive Payment System and risk adjustment for socioeconomic disparities. Policy changes also had a perceived effect on building collaboration with those inside the hospital or health system. The reasoning was that changes in government policy can impact the entire system, requiring swift responsiveness.
In addition to policy decisions spurring quality improvement, hospitals sought to meet goals set by external institutions, other agencies (eg, professional medical or nursing societies) and hospitals in the system, including the adoption of evidence-based practice.
In states that opted not to expand Medicaid following passage of the Affordable Care Act, hospitals forged relationships with community-based providers and organisations, or applied for Section 1115 Medicaid demonstration waivers to conduct specific quality improvement projects with CMS. For example, in one hospital, the CMS waiver prompted the development of outpatient clinics in the community where patients live to obtain better follow-up care.
Hospitals described actively responding to external initiatives and policy in a positive and negative manner, such as avoiding readmission. Some also spoke about the dynamic of balancing admissions and a full census with the cost of readmission penalties.
Finally, changes in federal policy towards value-based purchasing payment paradigms encouraged postacute care facilities, such as skilled nursing and long-term care facilities, to be concerned about their readmission rates due to pending financial penalties. This often fostered a sense of synergy between hospitals and postacute facilities to collaborate and address all-cause readmissions to improve quality across the care continuum.
Discussion
Our qualitative investigation is the first to identify key attributes of hospital organisational strategies related to providing exemplary care to attain low readmission rates for heart failure patients in safety-net hospitals serving low SES populations. Our findings indicate that the provision of high-quality care is influenced by a number of factors, including strong hospital support for quality improvement at all levels, tailoring resources to patient needs, promoting collaboration and communication, hospital support to postacute care facilities and communities, and proactive approaches to healthcare policy changes and other external factors. These characteristics of high-performing hospitals may influence health equity, as well as the quality of clinical care and readmission rates of hospitals serving low SES populations. Safety-net hospitals seeking to improve readmission rates may consider some of the approaches identified in this study.
Prior research has used qualitative methodology to identify paradigms of quality among patients with cardiovascular disease15 19–21 yet we are the first to specifically explore the concept based on high-performing hospitals that serve a large proportion of patients with low SES. As detailed throughout this paper, vulnerable patients experience a number of economic, social and behavioural challenges that may impede their ability to achieve optimal health. Despite the scope of these challenges, there are a number of features we identified among high-performing hospitals that also resonated with prior similar work.15 22–25 Concepts including supportive leadership, teamwork, effective communication, investment in data, broad staff presence and the development of innovative protocols given limited resources may all be important components of providing high-quality care across institutions and levels of SES.15 24 25 Moreover, our findings can be contextualised in trauma-informed approaches to patient care such as understanding patient demographics, creating a safe environment, building trust, empowering patients in their care, demonstrating and training in cultural competence, investing in integrated support services, as well as support programmes such as peer navigators.
Notably, we identified several unique features distinguishing high performance in hospitals serving low-SES populations. First, an explicit focus on patient-centred care and vulnerable patients was stated in the mission of the hospital and infused among the healthcare workforce and throughout the local community. Hospitals achieved patient-centred care by providing education, minimising burdens of care, instilling patient ownership and holding patients accountable, identifying and alleviating social risk factors and providing access to medications and proactive care coordination. Second, hospitals provided significant support to postacute care facilities and communities through education, financial support, transitions of care and capitalising on larger networks or systems. This was accomplished through frequent and careful communication, building trust and rapport over time, and longevity and partnerships as a proxy of a good relationship with external facilities. Third, hospitals were responsive to new policy changes, including the Affordable Care Act and Medicaid expansion, the shift to value-based payments (eg, readmission penalties), public reporting of quality and payment models such as risk-based contracts (eg, Accountable Care Organisations). Proactive approaches to changes in healthcare policy were achieved through foresight, investing in quality improvement and undergoing structural changes such as hospital mergers.
Our findings must be interpreted in the context of the limitation that our site visits occurred at a single point in time. Thus, high-performing hospitals might have been on a downward trajectory that was not captured in our sampling frame. However, we did compute outcomes over a multiyear period among selected hospitals and found little movement in ranking, suggesting overall sampling stability. Second, hospitals were selected based on HF RSRRs, thus these findings may not be generalisable to hospitals that want to improve performance on other condition-specific outcomes. Additionally, our study identified conceptual domains that we hypothesised would influence short-term patient outcomes (RSRR and RSMR). Specific measurement of these concepts is needed to test these hypotheses quantitatively in future studies with a larger, representative sample of hospitals. Finally, our findings were established without a comparison group of lower-performing safety-net hospitals, nevertheless, the exploratory nature of this study and related results may serve to inform safety-net hospitals of all levels of performance.
In conclusion, although specific approaches varied, the overall commitment to quality improvement among hospitals was consistent and clear. We noted across all five themes, there was a consistent message of reaching beyond what is traditionally done within a hospital inpatient visit. That is, respondents at high-quality hospitals serving low SES populations routinely conceived of their responsibilities for patient care as extending beyond the hospital walls. The approaches identified in this qualitative study can inform the quality of clinical care received, reduce the health equity gap and improve the outcomes of hospitals serving low-SES populations.
Data availability statement
No data are available.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants and was approved by Yale University Human Investigations Committee. Participants gave informed consent to participate in the study before taking part.
Acknowledgments
We would like to thank all interviewees for taking the time to speak with us, and all coordinators for their efforts in arranging our site visits. This work is a collaborative effort, and the authors gratefully acknowledge and thank our colleagues for their thoughtful input and support. Thanks in particular to Anita Arora, MD, MBA, Harlan Krumholz, MD, SM, Melissa Miller, MPH and Stefanie Rohde, BS. These individuals provided technical support and guidance throughout our project. Thanks also to Lein Han, our Contracting Officer Representative at the Centers for Medicare & Medicaid Services (CMS).
References
Footnotes
Contributors All authors contributed to the study and warrant authorship. KEM: data collection, data analysis, primary writer of the manuscript. KEM is the guarantor. PC conceived of study, data collection, data analysis, critical reviewer of the manuscript. KL: data analysis, project support, critical reviewer of the manuscript. LS: data analysis, project support, critical reviewer of the manuscript, critical reviewer of the manuscript. SMB obtained funding, conceived of idea, participated in data collection and analysis efforts, critical reviewer of the manuscript.
Funding Funded by the Centers for Medicare & Medicaid Services, an agency of the US Department of Health and Human Services.
Disclaimer The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services nor does the mention of trade names, commercial products, or organisations imply endorsement by the US government. The authors assume full responsibility for the accuracy and completeness of the ideas presented.
Competing interests All authors have completed this work under contract with the Centers for Medicare and Medicaid Services (CMS). The analyses on which this publication is based were performed under the Measure & Instrument Development and Support (MIDS) contract # HHSM-500-2013-13018I, Task Order: HHSM-500-T0001.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.