Are health systems redesigning how health care is delivered?

Abstract

Objective

To explore why and how health systems are engaging in care delivery redesign (CDR)—defined as the variety of tools and organizational change processes health systems use to pursue the Triple Aim.

Study Setting

A purposive sample of 24 health systems across 4 states as part of the Agency for Healthcare Research and Quality's Comparative Health System Performance Initiative.

Study Design

An exploratory qualitative study design to gain an “on the ground” understanding of health systems’ motivations for, and approaches to, CDR, with the goals of identifying key dimensions of CDR, and gauging the depth of change that is possible based on the particular approaches to redesign care being adopted by the health systems.

Data Collection

Semi‐structured telephone interviews with health system executives and physician organization leaders from 24 health systems (n = 162).

Principal Findings

We identify and define 13 CDR activities and find that the health systems’ efforts are varied in terms of both the combination of activities they are engaging in and the depth of innovation within each activity. Health system executives who report strong internal motivation for their CDR efforts describe more confidence in their approach to CDR than those who report strong external motivation. Health system leaders face uncertainty when implementing CDR due to a limited evidence base and because of the slower than expected pace of payment change.

Conclusions

The ability to validly and reliably measure CDR activities—particularly across varying organizational contexts and markets—is currently limited but is key to better understanding CDR’s impact on intended outcomes, which is important for guiding both health system decision making and policy making.

1.

What we already know on this topic

• In the new paradigm of health care transformation, successful health care organizations are expected to re‐engineer the way patients receive care.

• Providers join together in various organizational configurations in part, to improve the care they provide, and existing empirical evidence about the impact of redesigning care largely excludes efforts in health systems (organizations with at least one acute care hospital and at least one physician organization).

• Attempts by health care organizations to improve value cannot rely exclusively on established quality improvement and management tools, because the needed changes would likely entail new innovations or disruptive changes that go beyond modifying existing workflows.

What this study adds

• Health systems are implementing a variety of CDR activities that range from improvement‐focused to innovative to disruptive.

• Health systems’ unique contexts and motivations for implementing CDR influence whether they engage in what would be considered more innovative or disruptive CDR versus more standard improvement‐focused activities.

• This look at how CDR activities are being implemented on the ground offers insight into how a multidimensional construct of CDR might be defined and measured, including the need for more robust data to inform future efforts.

2. INTRODUCTION

It is hypothesized that successful health care organizations in the new paradigm of transformation will re‐engineer the way health care has traditionally served patients. The goals of these expected reforms have been described in various ways, ¹ , ² , ³ although the most commonly endorsed has been the “Triple Aim” which focuses on “better health care, better patient experience, and lower costs.” ⁴ The path to the Triple Aim appears to include two inextricably linked mechanisms: instituting new forms of provider organization and reforming the payment system. ⁵ While some now refer to the “Quadruple Aim,” the definition of the fourth aim varies (eg, health care worker job satisfaction, health equity, readiness) and for purposes of study we focus on the Triple Aim. ⁶

Providers join together in various organizational configurations (eg, medical groups, independent practice associations, integrated delivery systems, clinically integrated networks, accountable care organizations), in part, to improve the care they provide. ⁷ These entities are often created either through horizontal integration—bringing multiple like providers under one organization, or vertical integration—combining different types of providers across the care continuum. Relationships under these arrangements can involve direct ownership, contractual and other forms of affiliation, or complex combinations of ownership and affiliation. ⁸

As part of the broad investigation conducted by the RAND Center of Excellence on Health System Performance, our inquiry is focused on the redesign of care delivery in health systems, defined as organizations with at least one acute care hospital and at least one physician organization (PO), where the affiliation between hospitals and physicians may occur through shared ownership or a contracting relationship for payment and service delivery. ⁹ Prior research shows an increase in mergers, acquisitions, affiliations, and other organizational changes occurring in health care ¹⁰ , ¹¹ , ¹² , ¹³ and various payment changes. ¹⁴ , ¹⁵ , ¹⁶ To date, little systematic research has been conducted to understand whether and how health systems are actually redesigning their care delivery.

As a starting point, it is important to distinguish between what is traditionally considered quality improvement (QI)—relatively narrow, incremental changes—and efforts to systemically transform care—holistic, comprehensive changes. ¹⁷ Many QI tools in health care, often targeting efficiency and focusing on a limited scope, originated from models used in highly standardized manufacturing and engineering settings, such as the total quality management (TQM), continuous quality improvement (CQI), and the lean production models. ¹⁷ Implementing these tools in silos does not necessarily lead to systemic transformation of care. A recent article by Bhattacharyya et al argues that attempts by health care organizations to improve value cannot rely exclusively on established QI and management tools, because the needed changes may entail completely new innovations and therefore must go well beyond modifying existing workflows. ¹⁸ This view is echoed in the position paper by the American College of Physicians (ACP). ¹⁹

Empirical evidence about the impact of redesigning care is mainly based on studies of physician group practices, hospitals, and Accountable Care Organizations (ACO). Previous research on care transformation among physician practices examined the adoption of Care Management Practices (CMP), ²⁰ , ²¹ Health Information Technology (HIT), ²² , ²³ the Patient‐Centered Medical Home, ²⁴ , ²⁵ and QI processes. ²⁶ The growth of physician practice size has been found to correlate with the implementation of QI, but not with the adoption of CMP or HIT. ²⁷ Meanwhile, studies on ACOs showed that these organizations vary significantly in structure, focus, motivations, ²⁸ and activities, with care transformation being either centralized or frontline oriented. ²⁹ , ³⁰ While some studies suggest that ACOs might reduce hospitalizations and emergency room visits, their overall impact on patient care has not been sufficiently studied and needs further investigation. ³¹

Much less is understood about how care delivery is changing in health systems, partly due to the lack of systematic and reliable measurement of redesign efforts in such multifaceted, and often large, settings. Thus, far the field has provided little insight into three important and connected aspects of health systems’ efforts to redesign care delivery: the motivations (why), the strategies and activities (what), and the level of engagement (how much). Given the increasing scale and reach of health systems and the call for more fundamental and holistic changes to improve health care, ³² , ³³ this is an important gap in the literature.

2.1. Research questions and conceptual framework

We adopt care delivery redesign (CDR) to capture the variety of tools and organizational change processes used by health systems to pursue the Triple Aim. In practice, health systems may include multiple hospitals and POs along with other affiliated providers such as home health agencies, diagnostic laboratory services, long‐term care providers, and care management organizations. Because health systems are heterogeneous in their structure, governance, and culture; serve different populations; and are subject to regional market conditions and state regulations, we hypothesize that these factors can significantly impact the extent to which a particular system engages in CDR and its effectiveness in doing so.

Since innovation in any organization can range from a novel approach to the application of commonly used changes that are new to that particular organization, ³⁴ we adopt Bhattacharyya et al’s three conceptualized levels of innovation to guide our exploratory analysis: improving standardization and efficiency of an existing process, innovating within an established delivery model to provide care that is not currently deployed, and disrupting the current care paradigm by employing novel approaches to address situations where customers (patients) have unmet needs. ¹⁸

As there is no agreed‐upon path for a health system to transform from volume‐based to an increasing focus on quality and efficient care delivery, we sought to understand why and how the 24 systems in our purposive sample are engaging in CDR. Our study focused on these foundational questions:

1. What appears to motivate the health systems to engage in CDR?

2. What strategies and activities are the health systems using to redesign care?

3. To what degree do the health systems, as a group, appear to be engaging in improvement activities? In innovative activities? In activities that disrupt the existing model of care?

3. METHODS

Given the dearth of information about health systems’ efforts in CDR, we took an intentional exploratory approach using qualitative methods. Our analysis is part of a broader research effort led by the RAND Center of Excellence on Health System Performance as one of the Agency for Healthcare Research and Quality (AHRQ) Comparative Health System Performance Initiative (CHSP) centers. ³² We conducted 45‐90 minute semi‐structured telephone interviews between August 2017 and March 2019 with a total of 162 executives in a sample of 24 health systems (5‐8 executives per system) that varied in size and performance based on publicly available measures. Health systems were selected from across a convenience sample of 4 states that each host a multistakeholder nonprofit health improvement collaborative. Respondents included system‐level c‐suite executives and leaders from one multispecialty PO associated with each system. Interview instruments specific to each type of respondent role were created. Interview topics included the origins, organization, and governance of the health system; market context; care structures and strategies; and relationships among units, affiliates, and providers in the health system.

Project investigators developed and tested a first‐level codebook based on the major topic areas covered across the interviews. After an extensive period of training and codebook testing, two experienced qualitative researchers independently coded and reviewed each transcript.

We developed and analyzed system‐specific memos based on the data from relevant CDR codes. We then employed a multistep, team‐based coding process to categorize the full set of CDR activities described by respondents across the 24 systems. Next, we created groupings of systems with a similar level of self‐reported approaches in each category and studied the groupings for patterns. We extended our exploration of related literatures (eg, integrated care, disruptive innovation) as we worked through each step of our process to develop working hypotheses and interpret or test observations. We also analyzed data from additional codes related to systems’ motivations and strategies that influenced their CDR work.

Given space limitations, a detailed description of sample selection, data collection, and the analysis methods used to develop the working definition of CDR and address the three research questions are provided in the Appendix S1.

4. RESULTS

While the selection of our sample was purposive, we did observe variation in system type along a number of dimensions (Table 1). For example, of the 24 systems, 63% (15) were not‐for‐profit and 29% (7) were academic or had an academic affiliation. Most systems served multiple counties within or across states, and the vast majority included more than one acute care hospital. Nearly three‐quarters of the systems had a system‐wide electronic health record (EHR) from a single vendor. The systems vary widely in terms of structure, organizational complexity, governance, market size, and have varying influence with providers (both across and within systems) due to the myriad relationship structures in place with POs. ⁸ , ³⁵ For example, nearly half of the systems affiliate with POs (and in some cases, other systems) via one or more clinically integrated networks ³⁶ and most affiliate with providers in additional ways. While the majority (19) were participating in one or more ACOs at the time of data collection, three‐quarters of the systems had only a small percentage of their book of business at any level of financial risk.

TABLE 1.

Health system characteristics, n = 24

State ID, System ID	System characteristics					System affiliations		Risk assumption
	Org. type	Region served	HS is a single legal entity	# Hospitals	Enterprise‐wide EHR	CIN	Health plan	ACO participation	At‐risk contracting
A1	Nonprofit	Part of 1 county	No	3 owned	Single instance ‐ one vendor	No	No	No	Yes (half)
A2	Academic	Single county	No	2 owned	Single instance ‐ one vendor	No	No	Yes	Yes (small %)
A3	Nonprofit	Single county	No	3 owned; 2 affiliated	Inpatient and ambulatory use different vendors	No	Yes	Yes	Yes (large %)
A4	Nonprofit	Multiple counties within a state	Yes (with subsidiaries)	2 owned; 1 JV	Single instance ‐ one vendor	Yes w/partners	No	Yes	Yes (small %)
A5	Academic	Multiple counties within a state	No	5 owned	Single instance ‐ one vendor	Yes w/partners	Employees only	Yes	Yes (small %)
A6	Nonprofit	Multiple counties within a state	No	5 owned	Single instance ‐ one vendor	No	Yes	Yes	Yes (half)
A7	Nonprofit	Multiple counties within a state	No	2 owned	Inpatient and ambulatory use different vendors	No	Employees only	No	Yes
A8	Academic	Multiple counties within a state	No	1 owned	Single instance ‐ one vendor	No	Yes, JV (but plans to drop nonemployees)	No	Yes (small %)
A9	Academic	Multiple counties within a state	No	3 owned; 1 JV; 7 affiliated	Single instance ‐ one vendor	Yes	No	Yes	Yes (small %)
A10	Nonprofit	Multiple counties across 3 states	Yes (with subsidiaries)	49 owned; 2 JV	Single instance ‐ one vendor	Yes w/partners	Yes, JV	Yes	Yes
B1	Academic (affiliation agreement)	Multiple counties within a state	No	3 owned; 2 affiliated	Single instance ‐ one vendor	No	Yes, JV	Yes	Yes (small %)
B2	Nonprofit	Multiple counties within a state	Yes (with subsidiaries)	4 owned; 1 JV; 3 affiliated	Inpatient and ambulatory use different vendors	No	Yes	Yes	Yes (small %)
B3	Nonprofit	Multiple counties within a state	Yes	7 owned	Single instance ‐ one vendor	Yes (all employed physicians are in the CIN)	No	Yes	Yes (small %)
B4	Nonprofit	Multiple counties across 2 states	Yes	8 owned	Single instance ‐ one vendor	Yes (all employed physicians are in the CIN)	Yes, JV	No	Yes (half)
B5	Nonprofit	Multiple counties across 2 states	No	2 owned	Single instance ‐ one vendor	Yes w/partners (all employed physicians are in the CIN)	Yes, JV (launching)	Yes	Yes (small %)
B6	Academic	Multiple counties across several states	Yes (HS is a subsidiary of academic enterprise)	17 owned	More than one instance (one vendor)	No	No	No	Yes (small %)
C1	Quasi‐public	Single county	Yes	1 owned	Single instance ‐ one vendor	No	No	Yes	Yes (small %)
C2	Academic (affiliation agreement)	Multiple counties within a state	No	13 owned	More than one instance (one vendor)	Yes	Yes (involves only one of the partners)	Yes	Yes (large %)
C3	Nonprofit	Multiple counties across 2 states	Yes (with subsidiaries)	12 owned	Single instance ‐ one vendor	Yes (all employed physicians are in the CIN)	Yes, JV (launching)	Yes	Yes (small %)
C4	Nonprofit	Multiple counties across 2 states	Yes	5 owned; 1 affiliated; 1 JV	Single instance ‐ one vendor	No	Yes	Yes	Yes (large %)
C5	Nonprofit	Multiple counties across 3 states	Yes	2 owned; 1 affiliated	Inpatient and ambulatory use different vendors	No	No	Yes	Yes (small %)
D1	Quasi‐public	2 counties within a state	Yes	2 owned	Multiple EHR vendors within a single setting (inpatient. or ambulatory	Yes w/partners	No	Yes	Yes (small %)
D2	Nonprofit	Multiple counties within a state	No	8 owned	Single instance ‐ one vendor	Yes	No	Yes	Yes (small %)
D3	Nonprofit	Multiple counties within a state	No	2 owned	Single instance ‐ one vendor	No	No	Yes	Yes (small %)

HS refers to health system. A quasi‐public corporation is a private company that is supported by the government with a public mandate (and public funding) to provide a given service. Many quasi‐public corporations began as government entities (eg, safety net providers) but have since become nonprofit entities. Academic affiliation agreement refers to a health system that incorporates parts of a university health system and a nonprofit health system operated under a joint governance arrangement. Hospitals are a count of acute, general hospitals and excludes behavioral health facilities. (For multistate health systems, some of the counted hospitals are located outside of the four focal states of this study). Owned refers to hospitals owned and operated by the HS. Affiliated refers to hospitals managed by the HS under affiliation agreements (such as memorandums of understanding, management contracts, joint provision of service agreements). JV refers to jointly owned and operated or managed. EHR refers to electronic health record. Enterprise‐wide refers to an EHR that all (or almost all) affiliated providers use. Single instance means that there is a single copy of content that the multiple users or computers share. More than one instance refers to a situation in which the health system uses a single vendor EHR but not a single instance of the content. CIN refers to whether the health system offers a clinically integrated network to payers or participates with another HS or hospital to offer a CIN to payers. CIN w/partners refer to participation in a CIN sponsored by two or more HS. Health plan is a health insurance company either owned by a health system or through a JV. Employees only refer to a self‐insured health plan. ACO refers to whether the health system offers an ACO product to payers or participates in an ACO network offered by another entity. At‐risk contracting is defined as percent of book of business at global (professional, hospital and pharmacy), full (professional and hospital), or partial (professional services only) risk.

Source: Health system interview and survey data collected by authors. Information is current as of the date of the telephone interview (2017‐2019).


4.1. Why systems are engaging in CDR

We found that systems engage in CDR for complex and multidimensional reasons. We identified seven categories of CDR motivating factors, including internal and external motivations, and listed examples for each category (Table 2). Working with data from each system, we plotted the systems to visualize the relationship between internal and external motivations and identified several clusters of systems to examine in more detail in order to better understand the diversity of systems’ motivations (Figure 1).

TABLE 2.

Health Systems’ motivation for care delivery redesign

Motivating factor	Example motivators
Internal
Organizational priorities	Organizational mission and vision Board and leadership priorities Standardization across units
Organizational culture	Quality and safety Patient centered Value driven Employee/provider priorities
Fiscal health	Market differentiation Financial performance Cost efficiency
External
Community and population needs	Patient experience Stakeholder relationships Community needs assessment
Current scientific evidence	Clinical guidelines Best practices Technological innovation
Policy and transparency	Public reporting Regulatory compliance
Alternative payment models	Value‐based reimbursement Payment incentives Risk assumption

FIGURE 1.

Four clusters of health system motivation for care delivery redesign [Color figure can be viewed at wileyonlinelibrary.com]

We made a number of observations by viewing the data in this way. Seven systems (top right) described themselves as strongly motivated by both external and internal factors. Respondents from these systems often described efforts to purposefully link motivations with strategy, especially in an era of changing payment models. One respondent noted that the Triple Aim catalyzed the system's population health movement by encouraging system leaders to ask, “How can we get from kind of the dark corners of ‘you will come to the hospital and we will treat you,’ to this amazing promise of the Triple Aim?”

Respondents from 11 systems (middle) discussed internal and external motivating factors, but did not emphasize either as strongly as the systems described above. For example, respondents from one system frequently mention how the organizational culture values patient outcomes, which in turn motivates work in QI and cost efficiency while also noting that the system's strategies were influenced by external factors, such as the Centers for Medicare and Medicaid Services (CMS) star ratings and how those ratings can impact the hospital's reputation and potentially result in financial penalties.

Five health systems described themselves as being strongly internally motivated while reporting less emphasis on external motivators (top left). Respondents at these systems frequently cited factors such as the importance of patient experience, a strong organizational culture focused on QI, or a leader setting CDR priorities. For example, a system respondent noted having “a CEO who has incredible vision in terms of where she wants to take us as a health system and is very proactive and innovative.”

Respondents from a single system (bottom right) reported being primarily driven by external factors, describing a multiyear system reorganization motivated by market and fiscal pressures to “markedly reduce, by hundreds of millions of dollars across the system what our expenses are in delivering care.”

Overall, respondents describing their system as having strong external motivations (those furthest to the right in Figure 1) report experiencing pressure to change care processes because of expected increased competition from new care models regardless of whether or not they reported strong internal motivations. Respondents noted that lack of a strong evidence base on the return‐on‐investment associated with CDR causes them to grapple with how to invest time and resources. Respondents from these systems expressed a simultaneous sense of urgency and ambiguity about how to best adapt to an unknown future environment driven by much talk about, but less actual movement toward, use of alternative payment models.

Other respondents, such as this one, describe their systems as being “stuck in the middle” between external pressures to change and not knowing how best to do it and whether it will pay off:

As an organization I think we're still trying to assess where do we think the future is going… our academic side of our institution is very much an episodic care delivery organization. To what degree will that stay in some sort of a fee‐for‐service environment?

Overall, respondents from systems with strong internal motivations (those closest to the top in Figure 1) report more certainty related to the importance and direction of CDR than do those largely driven by strong external motivation or mixed motivations. Stated simply, internal motivations steer system leaders to view their work toward quality and efficiency as the right thing to do or because those efforts clearly align with the organization's mission, as illustrated by this respondent:

Our journey on quality has been about, ‘I want this to be the best place for patients to get care.’ Period. How do we measure ourselves? Well, there are quality scores, and there are outcomes… not because the government says we need to or because we're pursuing an award.

Another example is a system that has a strong “cultural heritage” of practicing knowledge‐based medicine. Respondents from this organization described culture as “the secret sauce” and were credited with creating an environment where clinicians and staff are willing to work within and across specialties to solve problems and come to consensus to benefit their patients.

4.2. How systems are engaging in CDR

One of our research questions centered on developing an understanding of what systems were doing to redesign care. Consistent with an exploratory approach that would build our understanding from the “ground‐up,” we did not impose a constrained definition of CDR, but rather asked respondents to describe the work their system was engaged in as it related to redesigning care. Based on these responses, as well as our working definition of CDR and the three areas of innovation identified by Bhattcharyya et al, ¹⁸ we took the information we heard during the interviews and sought to identify common types of CDR activities, definitions and descriptions of these activities, and patterns of focal areas across our sample.

4.3. Defining and categorizing CDR activities

We identified 13 distinct CDR activities, established definitions, inductively developed ratings, and then categorized each system on their self‐reported efforts (Table 3). We identified several areas of care redesign a priori and explicitly asked each of our focal respondents to discuss them: primary care and specialty care redesign, use of evidence, use of technology, efforts to reduce variation in care, and efforts to reduce the cost of care. The other CDR activities emerged from the interviews.

TABLE 3.

CDR activity definitions, category definitions, and number of health systems within each category

CDR activity definitions	CDR activity rating definitions	#HS
Use of Evidence: scientific evidence is used to guide care practices	Systemic use: HS reports system‐wide approaches for monitoring and using evidence	13
	Fragmented use: HS reports monitoring/using evidence in specific areas of care	11
Data and analytics: use of data and metrics in designing care approaches and/or evaluating care quality and outcomes	Mature: HS reports system‐wide use of data/analytics for care design/evaluation	13
	Developing: HS reports systemic data and analytic capabilities in development, but currently used only in specific areas for care design/evaluation	8
	Immature: HS reports limited use of data/analytics	3
Telehealth: ability to provide visits and consults where patient and provider are not in the same physical location	Yes: HS reports employing telehealth activities	12
	No efforts/unknown: Telehealth activities not reported by HS	12
QI/care quality: efforts to improve care; use of care improvement practices/models	Area of emphasis: HS reports QI/care quality as an area of emphasis	18
	Some efforts: HS reports limited QI/care quality efforts	6
Patient safety: efforts to reduce harm or increase patient safety	Area of emphasis: HS reports significant activity around patient safety	13
	Some efforts: HS reports small‐scale or limited patient safety efforts	10
	Little or no efforts/unknown: HS has few or no reported patient safety efforts	1
Care standardization: efforts to reduce or eliminate variation in care	Yes: HS reports minimizing care variation as an explicit area of emphasis	16
	No efforts/unknown: HS does not report minimizing care variation as an explicit area of activity	8
Cost focus: efforts to manage cost of care (eg, reducing low value and/or medically necessary care, waste, reducing unnecessary ED utilization and/or hospital (re)admissions)	Area of emphasis: HS reports that cost reduction is a major focus, or reports that there are well‐established, systemic cost, and efficiency initiatives	12
	Some efforts: HS reports multiple initiatives around cost and efficiency, sometimes with limited reach, but does not emphasize it as a major area of focus	11
	Low focus/unknown: HS does not report explicit initiatives focused on cost and efficiency	1
Team‐based approaches: care that is delivered by organized teams of three or more disciplines	Wide reach: HS reports team‐based care is widely implemented	11
	Limited reach: HS reports team‐based care is in development or implemented on a limited basis	11
	No emphasis/unknown: HS does not report an emphasis on team‐based care or team‐based care not reported	2
Population health management: efforts to provide preventive, therapeutic, and diagnostic care for a defined group in order to improve and maintain the health of the focal population	Mature: HS reports population health management is a systemic focus and/or reports multiple well‐established population health management initiatives	6
	Developing: HS reports building population health capabilities and/or reports some population health management initiatives in place, but narrow in scope/not utilized system‐wide	13
	Immature: HS does not report a formal population health management program	5
Care coordination: efforts to manage/coordinate care across providers and/or care settings	Established: HS reports established care coordination efforts in place	19
	Developing/Limited: HS reports developing care coordination efforts and/or care coordination is in place for limited patient population(s)	5
Specialty care: efforts to redesign specialty care to improve care/outcomes and integrate with primary care	Area of emphasis: HS reports a significant emphasis on specialty care redesign	4
	Some efforts: HS reports limited efforts to redesign specialty care	17
	Little or no efforts/unknown: HS does not report formal efforts to redesign specialty care	3
Social determinants of health: efforts to address one or more nonmedical needs of patients	Area of emphasis: HS reports a strong emphasis on efforts targeting patients' SDH	5
	Some efforts: HS reports limited efforts targeting patients' SDH	5
	Little or no efforts/unknown: HS reports little or no efforts targeting patients' SDH	14
Patient‐centric care: efforts to center care on the patient perspective	Established: HS reports pathways in place for patients to influence how HS approaches care	9
	Developing: HS reports transitioning to a patient‐centric approach, but early in implementation	2
	No data: Patient‐centric approaches not reported by HS	13

Abbreviations: CDR, care delivery redesign; HS, health system; QI, quality improvement; SDH, social determinants of health.

Source: Health system interview data collected by authors. Information is current as of the date of the telephone interview (2017‐2019).


Since our data focus on each system's areas of emphasis, we limit our analysis on those areas of reported strength rather than contrasting against areas of little reported activity since respondents may not have reported all lower‐level efforts (ie, those described as being in planning or small‐scale pilot stages or as highly fragmented ongoing efforts).

4.4. Patterns in health systems’ areas of CDR emphasis

While our analysis indicates that collectively our purposive sample of systems are engaged in many of the 13 activities, overall, there are limited discernable patterns in the CDR activities reported by respondents across the systems. All 24 systems reported at least some level of activity in 4 areas: care coordination (19 reported established efforts); QI/care quality (18 reported as an area of emphasis); use of evidence (13 reported systemic approaches); and cost focus (12 reported as an area of emphasis). The activity that the fewest systems reported as an area of emphasis was addressing social determinants of health (5 systems).

Of the 13 activities, one system reported emphasizing work in 11 areas (the exceptions were specialty care and social determinants of health) while, conversely, another system reported emphasizing work in only 2 activities (ie, data and analytics and care coordination). Most systems reported emphasizing approximately 5 to 8 of the 13 activities. Because of the lack of clear patterns from our exploration across our sample and the lack of existing research on whether to expect wide implementation across activities or innovation in a more focused set of areas in “leading” health systems, we also viewed the CDR activity data by each of the health system characteristic variables reported in Table 1. The discernable patterns were very limited and were often based on reasonable, but nonscientific judgments about how to divide categories (eg, using number of hospitals as a proxy for health system size in order to compare sub‐groupings). The most salient pattern we identified was based on geography: the majority of systems (8 of 10) in state A reported a cost‐focus emphasis compared to half of systems (3 of 6) in state B, only 1 (of 5) in state C and none (of 3) in state D; all systems in state D reported a systematic use of evidence compared to two‐thirds of systems (4 of 6) in state B, half of systems in state A (5 of 10), and one‐fifth of systems in state C (1 of 5).

We also analyzed the systems’ CDR activities by the 4 motivations groupings discussed above. While we identified some possible patterns, we determined that the exploratory nature of the findings was not systematic enough to ensure any identified patterns were not more than coincidental. Building on work from the broader project, we also examined the data for patterns related to the systems’ CDR activities compared to available scores on publicly reported quality measures. This analysis also did not yield any patterns, which makes some intuitive sense as it is difficult to discern if a system's emphasis on CDR activities may be due to the need to improve quality measures or the reason why its quality measures may have improved.

4.5. Analysis of CDR activity areas—improvement, innovation, or disruption

As a final step in our analysis, we compiled summaries of systems’ efforts in each of the 13 CDR activities and linked to the Bhattacharyya et al framework, ¹⁸ to assess the degree to which each activity appeared to be focused on improvement, innovation, or disruption of care. (See Table 4, eg, of CDR activities.) Based on our analysis of these summaries, several activity areas (eg, QI/care quality, patient safety, care standardization) appeared to be improvement activities, while some (eg, population health management, team‐based approaches, telehealth, data, and analytics) appeared to be generally more innovative, and only one (ie, patient‐centric care) appeared to be employed by some systems in ways that are truly disruptive to care delivery. However, we observe that it is not the selection of a particular activity, rather how systems operationalize work in that area, that dictates whether systems aim to improve, innovate, or disrupt their current care paradigm.

TABLE 4.

Examples of health system activities by CDR area

CDR activity	Examples from HS that reported an emphasis
Use of evidence	System utilizes clinical decision support mechanisms in EHR to hardwire clinical evidence into practice; clinical informatics assists with prioritization and implementation. Specialists closely monitor practice guidelines.
	System uses quality and analytic teams, IT, and providers to develop care strategies using evidence‐based guidelines and standard order sets (in EHR); builds tools for measurement and assessment.
Data & analytics	System utilizes an enterprise data warehouse (including EHR, claims, and workflow data) to glean unique insights, strategic opportunities, and to guide actionable high‐value workflows. Patient information is loaded into a population health platform and acts as a centralized care management solution. System monitors physicians’ out of network utilization, shift to in‐network, and the value of that shift.
	System uses an analytics‐driven performance improvement platform to help reduce care variation. System has three arms of analytics: medical informatics, EHR, and enterprise data warehouse (includes financial and clinical data outside EHR). System leveraging data and EHR to support best practice implementation. System connecting primary care, specialty, and other sites in network through HIE.
Telehealth	System uses multiple efforts to increase access and make care more consumer centric (eg, video visits, asynchronous e‐consultations, open‐access scheduling, building out a patient portal, automatic texting for appointment times, algorithms for primary care to sub‐specialty referrals). System aims to grow telemedicine from 5% to 25%‐40% of visits.
	Online care for nearly 100 common conditions; protocols developed by clinician leaders and staffed by nurse practitioners. Developing electronic capabilities for both e‐visits and phone visits.
QI/care quality	QI is a system‐based, centrally managed program; system employs numerous QI approaches (eg, Lean, Six‐Sigma, PDSA cycles) that follow the data and are provider‐led. Large amount of training on care improvement practices and pathway production. QI initiatives triaged to ensure they align with institutional goals; metrics reported through the board.
	System CEO prioritizes quality and safety. Quality department has multiple staff that work on projects across the system; provides training on QI (PDSA); and formally assesses QI efforts. System uses a prioritization grid for QI that includes regulatory components, capacity and risk, and compares its performance against state metrics. Different components of the system may prioritize different areas of work.
Patient safety	System has a strong culture of safety, quality, transparency, teamwork, excellence, and respect for people that is championed by long‐term leaders.
	System CEO is driving cultural shift and strategy of being a “high reliability organization” with “zero harm.” Strategy puts emphasis on safety and QI initiatives.
Care standardization	System's CDR strategy focuses mainly on organizational growth and standardization. Evidence‐based guidelines used in standardizing care, cost reduction initiatives, and QI.
	System's P&T committee drives standardization with an eye on cost. Multiple centrally coordinated QI projects focused on standardizing practices and screenings. Relies heavily on EHR to increase standardization.
Cost focus	System's quality organization oversees value‐based care strategy and related initiatives. One arm is focused on developing evidence‐based, standardized care pathways to reduce unnecessary costs. Guided by supply chain management principles, the system is prioritizing cost reduction with an initial goal of cutting $100 million in expenses.
	System reorganized so that leaders have both location‐specific and cross‐system responsibilities to reduce duplication/silos and reduce costs. Currently, in the midst of a major 3‐year effort to markedly reduce expenses in care delivery. Cost reduction is the main charge of the system's new Transformation Office.
Team‐based approaches	System has been working for several years to build a team care model in primary care clinics. The approach utilizes “pods” of physicians, other clinicians, and front office staff. Pharmacists embedded in one primary clinic with the intent to roll out to other clinics if financially viable.
	Team‐based care is intertwined with population health and viewed as a way to achieve better health care. Effort is physician‐led; motivated by reimbursement challenges and a desire for clinicians to practice at top of their license; goal of maximizing patient care while minimizing cost. Modeled after work at other institutions.
Population health management	System invested in multiyear comprehensive population health initiative with goal to move toward a care continuum model. Works with community teams to target SDH in order to improve success in population health management. Efforts include assigning care navigators to Medicare and Medicaid patients that come through the ER; offering programs and classes to patients around housing, health conditions, food, or exercise; and building social determinant networks and referral systems. Uses centralized care management platform. Evaluates population health programs using literature and cost/quality measures.
	System utilizes panel managers/care coordinators (~60 across outpatient settings), viewed as extensions of system's team‐based care model, work with patients with common chronic conditions. Remote monitoring used for complex chronic disease patients at high risk for readmission. Once stabilized, patients transition back to own care team.
Care coordination	System and PO coordinate patient care and establish goals for patient initiatives that cross inpatient and outpatient settings.
	Care coordinators at system work from central location but are typically assigned to 3‐4 clinics; efforts focused on Medicare Advantage patients, at‐risk employee population, and high cost utilizers. “Early stages” of population health management; focus on preventative (versus disease‐specific) care; driven by publicly reported measures. Early adopters of previsit planning.
Specialty care	System has implemented over 500 standardized best‐practice protocols/order sets; based on payment bundles and specialty societies’ recommendations; embedded in EHR. Leaders examine best practices, monitor order set compliance, and feed data back to physicians.
	System is incorporating specialists who practice at ambulatory clinics into standardized rooming. Specialists engaged in and held accountable for certain population health initiatives and primary care quality metrics. Has orthopedic center of excellence.
Social determinants of health	System is heavily focused on team‐based population health care to serve communities (beyond patients directly attributed to the system). Works with various communities (eg, Native American community) to proactively build relationships and address barriers to care (eg, transportation).
	System integrates claims data, EHR data, and local social service needs data to stratify by risk and assess program need for at‐risk patients. System utilizes community paramedicine to take primary care and public health services out to underserved populations, such as the homeless, rather than exclusively providing services in health care facilities.
Patient‐centric care	System has innovation division to enhance the health care experience by listening to consumers and delivering innovation solutions; looks for patient‐facing technology; aggressively trying to understand patients and how they want to partner with system. Utilizing video visits; patient portal; e‐communications; automatic refills; and centralized scheduling.
	Patient‐centric/quality of care culture; engaging patients beyond face‐to‐face interaction (eg, nurse call line, online chats with nurse practitioners, social media, using technology to overcome the rural care barrier).

Abbreviations: CDR, care delivery redesign; EHR, electronic health record; HIE, health information exchange; P&T committee, pharmacy and therapeutics committee; PDSA, Plan Do Study Act; QI, quality improvement; SDH, social determinants of health.

Source: Health system interview data collected by authors. Information is current as of the date of the telephone interview (2017‐2019).


As one illustration, multiple systems reported working to adopt a population health management approach as an area of emphasis. For most systems, that work involved selecting certain populations (eg, patients with diabetes, high‐utilizers) and building special pathways, facilities, care structures. Additionally, some systems are empaneling patients to identify gaps in care and direct patients to preventive care or screenings, enhancing their ability to provide the right level of care. ³⁷ Such activities are improvement‐ and innovative‐level focused. In some instances, however, systems report adopting a population health management strategy as their overall approach to primary care. This latter approach crosses into disrupting current care delivery.

The varied use of data and analytics is another example of how systems operationalize their work and whether they seek to improve, innovate, or disrupt the current care paradigm. While all systems report the use of data and analytics, in some cases it is limited to the basic measurement of QI processes. In other cases, systems extend their use of data and analytics to measure QI outcomes. Both of those applications are improvement‐level focused. Systems that appear to be using data and analytics to innovate report activities such as building dashboards and making patient outcomes data available for planning purposes. A small number of systems report employing data and analytics in ways that could potentially disrupt the way they traditionally provide care. These systems are bringing together disparate forms of data for real‐time access with the goal of providing actionable views into the system's cost of care, timeliness of care, care outcomes, etc For example, one system is utilizing its unique relationship as a subsidiary of the county to integrate county social services data, medical claims data, and clinical data from its EHR to evaluate the needs of its population, stratify by risk, and assess what programs would be most effective. Programs implemented include, for example, the deployment of mobile clinics into the community to provide care to vulnerable populations such as the homeless.

5. DISCUSSION

While it is appealing to assume that health systems are well‐positioned to use their financial and human resources as well as their administrative leverage to redesign traditional care delivery to move toward the Triple Aim, our exploratory analysis suggests there is significant heterogeneity in both the breadth and degree of innovation in health systems’ CDR approaches. We also found variation in the factors that are prompting systems to redesign care. For example, system leaders appear to be handling payment uncertainty differently: some believe they need to “get out ahead” and prepare to assume risk, while others are waiting to develop their full strategy until payment incentives change in a meaningful way. Those motivated largely by anticipated payment changes have concerns that they would end up stuck in what is often referred to as the “valley of death”—the lag period between making investments to redesign care and when those changes financially pay off. ³⁸

The varied response to payment uncertainty is perhaps due to the fact that despite the amount of attention on the creation of new risk‐based payment models, the movement away from traditional fee‐for‐service has not been as dramatic as expected. For example, Burns and Pauly report that only 30% of physicians received any value‐based payments in 2016 and that 78% of 2015 payments from commercial payers were still fee‐for‐service. ⁵ Similarly, while executives in the majority of our sample reported some risk‐based contracts—for most systems, it was a small percentage of their book of business. If investment capital for CDR presumably comes from payment models that pass some risk to the provider, the pace of implementation may be too slow to underwrite CDR efforts outside of time‐limited programs that have provided up‐front funds to seed CDR‐type investments (eg, CMMI’s Comprehensive Primary Care Demonstration). ³⁹

System leaders also reported highly individualized approaches to implementing CDR activities in terms of scope, goals, depth, and breadth, but also in how connections are made between what are often interrelated CDR activities. Multiple studies on organizational change in health care have shown that while the impetus to change is necessary, in order to achieve care transformation, elements such as leadership commitment, alignment of organizational goals, and integration across organizational levels are needed. ⁴⁰ , ⁴¹ , ⁴² It is not surprising that the areas of CDR with the most uptake across the systems are those that are centered on improving standardization and efficiency of existing practices rather than establishing innovative delivery models or the adoption of disruptive methods. ¹⁸

This understanding, however, stands in contrast to policy expectations and national conversations that care is rapidly moving toward the Triple Aim. In most other industries, organizational change strategies would be expected to include a high percentage of improvement activities, a much smaller percentage of innovation activities, and a minor percentage of disruptive changes of their product or services lines. ⁴³ Is it realistic to expect health systems to be exceptional?

Studies have noted both the lack of an evidence base for transformation in health care ³⁴ , ⁴⁴ and the importance of regional and organizational context in effectively implementing innovation. ⁴⁴ , ⁴⁵ The tension between the need for models and the strategic selection and adaptation of models to fit local needs points to the importance of informed and skilled health system leaders. While it may be tempting to look for the entrepreneurial health systems that disrupted care approaches and have high care quality outcomes and expect other systems to be able to follow those models, it has been noted that those efforts may be effective in making systems look more similar, but not actually bring about improvements in efficiency or outcomes when changes are implemented without consideration of each system's unique characteristics and context. ⁴⁵ , ⁴⁶

While labels such as “redesigning” or “transforming” care have intuitive appeal and many health care stakeholders recognize there are numerous opportunities for improving processes that should lead to better outcomes, we are still a long way from being able to systematically measure these changes. The ability to validly and reliably measure CDR activities—particularly across varying organizational and market contexts—is key to better understanding CDR’s impact on intended outcomes and thus potentially better managerial decision making and policy making. Our identification of specific CDR activities from a “grounded approach,” and a first attempt to label the variation observed in these activities across our sample, as well as identifying variation in the associated motivations for CDR, may serve as a resource for future efforts to develop valid and reliable measurement approaches. Doing so would require a larger sample of health systems and significant investments in data collection of various types (eg, administrative data, document review, observations from accreditation type reviews, interviews with health system leaders and others representing various perspectives and levels of the organization across multiple points in time) with the ability to link these mixed methods data and develop an approach for weighting the contribution of each data source to the latent construct we refer to as CDR.

Our analysis has several limitations. First, our data come from a purposive sample in 4 states with a history of public reporting of performance measures, QI, and patient engagement initiatives. Our sample should not be considered generalizable to the broader universe of health systems in the United States, which as mentioned earlier is quite varied. Since the aim of this study was to systematically understand CDR—both in terms of systems’ motivations and efforts, and ultimately to develop an initial conceptualization of CDR that might be useful for systematic measurement in the future, the purposive nature of our sample is reasonable.

Second, we interviewed 5‐8 executives within each system. While the multiple respondents allowed for some triangulation, the study did not include all leaders (eg, chief nursing officers in certain systems) nor insights from people further down in the organizational structures. Interestingly, as Tables 3 and 4 illustrate, while our data are far from perfect, leaders’ responses did not always describe their organizations as advanced, suggesting an element of face validity that systems vary on motivation and implementation.

Third, we focused on data from specific types of respondents (CMOs and CQOs) whom we believed to be most familiar with their system's CDR goals and activities. Our inquiry also was fixed on a particular timeframe which does not allow us to account for organizational dynamics. Thus, it is possible that systems with strong internal motivation for CDR today, for example, were initially motivated from external factors or vice versa.

Fourth, because this analysis focused on creating an initial understanding of systems’ CDR approaches, the findings presented here are exploratory. Future research should be able to address these limitations and we believe our preliminary findings can help inform the design of more granular data collection approaches to better capture and explain the heterogeneity we observed.

In conclusion, we found that health systems are motivated and engaged in myriad CDR activities, and the magnitude and pace of change appears to be variable across systems. Furthermore, many system leaders reported uncertainty about how to approach CDR, especially considering the slow pace of payment change. Developing an evidence base by measuring health systems’ CDR activities over time and measuring associated changes in CDR capacity with changes in Triple Aim outcomes could provide a foundation for health system leaders who are developing CDR strategies, and for designing better quality measures and payment models. Our identification of CDR activities being implemented on the ground offers insight into how a multidimensional construct of CDR might be defined and measured, including the need for more robust data to inform these efforts.

Supporting information

Author Matrix

Appendix S1

Scanlon DP, Harvey JB, Wolf LJ, et al. Are health systems redesigning how health care is delivered?. Health Serv Res. 2020;55:1129–1143. 10.1111/1475-6773.13585