Identifying Populations with Complex Needs: Variation in Approaches Used to Select Complex Patient Populations

Abstract

Interventions to support patients with complex needs are proliferating. However, little attention has been paid to methods for identifying complex patients. This study aims to summarize approaches used to define populations with complex needs in practice, by cataloging specific population criteria and organizing them into a taxonomy. The authors conducted a pragmatic review of literature published January 2000-December 2018 using PubMed. Search results were limited to English-language studies of adults that specified a set of objective criteria to identify a population with complex needs. The authors abstracted data from each article on population parameters, and conducted thematic analysis guided by deductive coding. The review identified 70 studies reflecting 90 unique complex population definitions. Complex populations criteria reflected 3 approaches: stratification, segmentation, and targeting. Six domains of population criteria were found within, including age-based criteria (59 populations); income (12); health care costs (45); health care utilization (39); health conditions (35); and subjective criteria (15). Criteria from multiple domains were frequently used in combination, and exact specifications were highly variable within each domain. Overall, 83% of the 90 population definitions included at least 1 cost- or utilization-based criterion. Nearly every study in the review presented a unique approach to identifying patients with complex needs but a limited number of “schools of thought” were found. Variability in definitions and inconsistent terminology are potential sources of ambiguity between stakeholders. Greater specificity and transparency in complex population definition would be a substantial contribution to the emerging field of complex care.

Introduction

Interventions to support patients with complex needs are proliferating, and selection of a target population is central to intervention planning. In practice, complex care programs generally use a set of objective and/or subjective criteria to select eligible patients, such as focusing on patients with high costs, frequent utilization, or multiple chronic conditions.¹ Given the heterogeneity of patients with complex needs, a single care model or program cannot meet the needs of all subpopulations.² This has led to much experimentation with regard to the ideal design of interventions for maximum impact, titrating program design elements including mix of services, program staffing model, and timing of intervention delivery.

The optimal approach to defining “patients with complex needs” is a matter of ongoing discussion in the field.³ Often, theoretical descriptions of patients with complex needs are framed around the biopsychosocial model of disease, which describes the interconnectedness of biological and psychosocial factors in determining health and wellness.⁴ Although taxonomies exist to describe the major groups of complex patients, these are often high level. Most prominent is the taxonomy put forward by the National Academy of Medicine (NAM),¹ which categorizes complex patients into 6 clinical and functional groups: children with complex needs, non-elderly disabled, multiple chronic, major complex chronic, frail elderly, and advancing illness. Social risk factors and behavioral health factors crosscut through these groups.¹ The NAM taxonomy represents a theoretical framework for groups of complex patients and highlights the interplay between medical, behavioral, social, and functional status factors. However, it does not reflect the diversity of current operational approaches to identify patients for complex care programs.

While evidence about the design of effective interventions for populations with complex needs is slowly accumulating,^1,3,5 less attention has been paid to exploring the implications of patient selection decisions in the complex care field. Often, discussions of evidence surrounding complex care programs treat population selection as a sidebar and focus on the design of services rather than the specification of population criteria. As a result, little evidence is available about the advantages and disadvantages of the many ways to conceptualize complex care populations.

This study aims to catalogue and summarize the range of approaches being used to define populations with complex needs to provide insights that support informed decisions about population selection for complex care interventions. This pragmatic review of the literature presents a holistic survey of the variety of ways in which complex populations are being defined and organizes these approaches into a framework. This review can inform efforts to compose populations of complex patients for intervention planning, post hoc descriptive analyses, or predictive modeling. This study highlights some of the most common approaches and discusses implications of the variability in methods for identifying complex patients.

Methods

This study employed a pragmatic review of the literature. A pragmatic review aims to summarize a body of literature most efficiently by avoiding duplication of effort while maintaining the essential rigor and comprehensiveness of a systematic review.^6,7 A pragmatic approach is appropriate when the question of interest is broad and not intended to guide treatment for specific types of patients. Nevertheless, the methods used were largely consistent with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines.⁸ Specifically, the pragmatic compromises made included searching of a single computerized database by one reviewer, development of a custom search term because of a lack of relevant MESH terms, and use of a team-based data abstraction process.

Search strategy

The research team searched PubMed in March 2019 for studies published January 2000-December 2018 (including advanced and online-first publications) using the following search profile: (“population health management” OR “utilization management” or “risk management”) and (“high-need” or “complex needs” or “high-cost”) and (identif*). Reference lists of studies retrieved were reviewed to identify any relevant papers that were not captured by the search terms.

Eligibility screening

A 2-phased process was used for eligibility screening with explicit criteria. First, studies were potentially eligible if they were published in English between January 1, 2000-December 31, 2018, included adults aged 18 years and older, and involved the use of objective criteria to identify a complex needs population for any purpose (studies must have specified a set of concrete criteria for defining the “complex needs” study cohort). Studies were then excluded from the sample if they focused on a single health condition, presented individual case studies, or were student dissertations and theses, trial protocols, or newsletters and commentaries without presentation of quantitative data.

One screener (TAO) reviewed the titles and abstracts of publications identified using the search strategy. Full-text articles were pulled for all publications that were not eliminated during title and abstract screening. Two screeners (TAO and LJL) then reviewed the full text of articles and applied all inclusion and exclusion criteria to arrive at the final sample of included studies.

Data abstraction

Two members of the research team (TAO and LJL) systematically abstracted relevant data from each article using a Microsoft Excel-based data abstraction form. The data elements abstracted included: full citation, article intent (program/intervention, descriptive analysis, or predictive modeling), setting/data source, objective criteria used to define the study population, notes regarding any element of subjective criteria or processes in the study population inclusion and exclusion logic, and a general category of reviewer notes. Published appendices and other supplemental article materials were used as needed to obtain all data elements.

Each team member independently abstracted data for half of the studies. The coders then exchanged the data for secondary review and confirmation. Finally, together with a third team member (ACD), the coders reviewed and discussed any studies for which data abstraction was unclear or for which there was disagreement. Articles were discussed until consensus could be reached.

Synthesis

To synthesize insights about strategies for identifying and selecting complex populations, the research team then reviewed the abstracted data related to objective and subjective criteria or processes for complex population identification, as well as study setting and data source, using a thematic analysis approach guided by a deductive coding process and data displays. Study setting and data source were included in this synthesis because some study populations were narrowed as a byproduct of these study features even though they were not presented as explicit population criteria in the original studies (eg, studies using Medicare data were focused on seniors and persons with disabilities).

Two team members (TAO and ACD) jointly reviewed the abstracted data to identify and organize the population criteria into domains and subdomains. Then, one team member (TAO) applied the domains and subdomains to the abstracted data for all studies included. Finally, for quality assurance, 20% of the articles were reviewed by a third team member who did not participate in the synthesis process (LJL) to confirm the accuracy of the domains as applied to the abstracted data. This quality assurance review did not identify any errors.

Finally, the team applied a framework presented by NAM in its report on effective care for high-need patients,^{1(pp 36–37)} to further characterize the domains of population criteria identified. The NAM framework presented 3 different approaches to identifying patients with complex needs: stratification methods, segmentation methods, and patient targeting.

Stratification methods separate a population according to risk level (eg, spending or utilization; may be past or future, persistent or acute). Stratification methods do not yield homogenous clusters of patients with shared needs or characteristics. For example, the top 1% of patients with high costs are heterogeneous in their characteristics and needs.⁹

Segmentation methods aim to drive toward homogenous groups of people with shared needs and characteristics (eg, clinical, behavioral, or social characteristics or needs). Patient segments may be heterogenous in risk level depending on the granularity of segmentation. For example, segmentation based on presence of conditions such as diabetes and kidney disease would yield a population that is more heterogeneous in terms of risk level, while segmentation based on severity of diabetes and kidney disease would homogenize risk level to a greater extent.

Patient targeting is used to identify clusters of people who are suited for a given intervention, based on presence of “actionable” needs (eg, transportation needs, polypharmacy). Although this approach might be thought of as an action taken to match patients to services after they have already been identified, some programs may use highly targeting criteria to select patients, such as programs that enroll patients based on an assessment/screening for specific needs.

Descriptive summaries of study data according to the domains and subdomains identified are presented.

Results

A total of 3411 citations met the search criteria; after screening at the title/abstract stage, 81 full-text articles were assessed for eligibility. Of these, 11 were excluded after full-text review, and a total of 70 papers were included in the final sample for qualitative synthesis (Figure 1).^{9

–78}

FIG. 1.

Diagram of study selection process for qualitative synthesis.

Within these 70 papers, the research team identified 90 distinct complex populations (some studies presented multiple distinct cohorts of complex patients). A detailed tabulation of inclusion and exclusion criteria for complex patient population definitions, as abstracted for the 90 populations from the 70 studies in the review, is provided in Supplementary Table S1 (online Supplementary Data).

Based on the thematic review and synthesis, the research team developed a hierarchy of population definition components. At the highest level, the team grouped inclusion and exclusion criteria for selecting complex patient populations into domains. In total, 6 domains of criteria for complex population selection were identified (Figure 2). Five of these were based on objective data, including health care costs (used as a criterion in 45 out of 90 complex population definitions)^{9

–12,16,19,21
–23,26,28

–31,34,35,37
–39,44,47,50,51,54,55,58,60,64,67,72,74,75,77,78}; health care utilization (39/90)^{9

–12,16,19,21
–23,26,28

–31,34,35,37
–39,44,47,50,51,54,55,58,60,64,67,72,74,75,77,78}; age (59/90)^{9,10,13,15
–17,19,20,22

–25,27,28,32

–36,40,42,43,45,47,48,51
–53,56
–58,60,61,63

–67,69,71

–75}; income (12/90)^{12,14,18,21,48,54,56,64,71,72,74,75}; and health conditions (35/90).^{12,15,17,20,21,23,25,27,28,32,42,45,48,50

–53,59,61
–63,65

–69} The sixth domain encompassed approaches based on more subjective methods of patient selection such as provider referral (15/90 complex populations).^{13,21,24,30
–32,45,52,53,57,62,64,68,70,75}

FIG. 2.

Domains and subdomains of criteria for defining populations of patients with complex needs. The cost- and utilization-based criteria in this framework can be thought of as stratification approaches, while age, health status, and income-based criteria can be thought of as segmentation approaches or targeting approaches, depending on their construction. Most population definitions reviewed combined stratification and segmentation approaches, using a mix of criteria capturing cost and/or utilization as well as health conditions, age, or income.

Within each domain of population criteria, the research team identified subdomains through iterative clustering of the criteria within each domain (Figure 2). For example, the domain of cost-based criteria was split into subdomains of approaches focusing on (a) prior costs versus those focusing on (b) predicted costs.

Building on the NAM framework for approaches to identifying high-need patients,¹ the team classified the domains of cost- and utilization-based approaches as stratification methods. In contrast, age-, income-, and health-based approaches, as well as subjective approaches were classified as segmentation and/or targeting methods (Figure 2).

In this review, most studies employed a combination of criteria across several domains to structure their complex population (Table 1 and Supplementary Table Table S1), highlighting that the 3 broad approaches (stratification, segmentation, and targeting) are not mutually exclusive. Many complex patient populations were defined using criteria that combine the approaches, such as by identifying high-risk patients based on their prior costs, and then segmenting the high-risk group into subgroups based on their functional limitations or clinical needs. The team analyzed the ways in which criteria from multiple domains were used in combination with one another. The most commonly used combination of criteria was an age-based criterion in combination with a cost-based criterion (15 out of 90 complex population definitions),^{9,10,16,19,22,28,34,35,47,51,58,60} followed by age with health conditions (12/90),^{17,20,27,42,48,51,61,63,65,69} and age with utilization (8/90).^{33,36,40,43,51,73} Criteria could be combined with “and” or “or” logic (Table S1).

Table 1.

Distinct Combinations of Domains of Complex Population Criteria and Frequency

Distinct combination of domains	Cost-based criteria	Utilization-based criteria	Age-based criteria	Income-based criteria	Health conditions-based criteria	Subjective criteria	Number of complex populations
Age & Cost	X		X				15
Age & Health Conditions			X		X		12
Age & Utilization		X	X				8
Cost	X						7
Age, Cost & Health Conditions	X		X		X		7
Cost & Utilization	X	X					5
Utilization		X					5
Age, Utilization & Health Conditions		X	X		X		4
Age, Utilization, Health Conditions & Subjectivity		X	X		X	X	3
Cost, Utilization & Health Conditions	X	X			X		3
Age, Utilization & Subjectivity		X	X			X	3
Income & Utilization		X		X			2
Utilization & Subjectivity		X				X	2
Utilization, Health Conditions & Subjectivity		X			X	X	2
Age, Income, Cost & Subjectivity	X		X	X		X	2
Age, Income & Cost	X		X	X			2
Income, Cost & Health Conditions	X			X	X		1
Income, Cost, Utilization, Health Conditions & Subjectivity	X	X		X	X	X	1
Cost & Subjectivity	X					X	1
Age, Income & Health Conditions			X	X	X		1
Income & Cost	X			X			1
Age, Income & Utilization		X	X	X			1
Health Conditions & Subjectivity					X	X	1
Age & Income			X	X			1
Number of Complex Populations	45	39	59	12	35	15	90

In total, 75 of the 90 complex population definitions (83.3%) included a cost-based criterion, a utilization-based criterion, or both^{9

–16,18,19,21

–26,28

–41,43

–47,49

–62,64,66,67,70,72

–78}; 35 of 90 (38.9%) included a health conditions-based criterion^{12,15,17,20,21,23,25,27,28,32,42,45,48,50
–52,59,61
–63,65

–69}; and 89 of 90 (98.8%) had at least 1 criterion from the cost, utilization, or health domains (Table 1). Among the 75 populations defined using objective measures of cost or utilization, 14 of 75 (18.7%) also included a subjective component (eg, referral or screening of the candidate patient list).^{13,21,24,30
–32,45,52,53,57,59,62,64,70,75} Among the 35 populations that included a health-based criterion, 7 of 35 (20%) included a subjective component.^{21,32,45,52,59,62,68}

Finally, the team characterized the dimensions on which criteria varied, after reviewing the detailed population specifications used by the studies in the review (Table 2). Cost-based criteria varied along 5 different dimensions: the threshold set for level of spending (eg, top 5%); the time period applied to the criteria (eg, prior year); the use of explicit specifications for narrowing types of costs being counted (eg, excluding pharmacy or long-term care costs); the method used to classify costs (eg, a percentile-based approach vs. setting a specific dollar amount threshold); and lastly, whether the cost assessment was retrospective or predictive. For example, Cunningham et al defined complex patients as those whose spending was in the top 25% in the prior year,¹⁹ while Bell et al selected patients whose predicted costs in the next year were 50% higher than the average Medicaid recipient.¹² These same dimensions of variation also were seen in utilization- and health-based criteria; age- and income-based criteria were much simpler in their specifications (Table 2).

Table 2.

Dimensions of Variation in Specification of Population Criteria, by Domain

Criteria domains	Dimensions of variation
Criteria domains	Thresholds	Time period	Narrowing specifications	Method	Predictive approach
Cost-Based Criteria	Amount of spending (either past or predicted)	Variable look-back period for summing costs; time frame for predicted costs	Specific types of costs included or excluded	Specific dollar amount; percentile of costs; costs relative to mean or median; trend over time	Model for predicting future costs
Utilization-Based Criteria	Amount of utilization, past or predicted	Variable look-back period for counting utilization	Specific types of visits or medications	Count of distinct providers; count of encounters or dates of service; length of stay	Model for predicting admission or readmission risk
Age-Based Criteria	Upper and lower age bounds	NA	NA	NA	NA
Income-Based Criteria	Federal Poverty Level	NA	NA	Method of assessing income or need	NA
Health Condition-Based Criteria	Number of conditions present	Variable look-back period for assessing presence of conditions	Lists of specific conditions being counted	Number of conditions based on diagnoses or self-report; number of diagnoses related to specific condition(s); recent encounters related to specific condition(s); score on an index measure of comorbidity	NA
Subjective Criteria	NA	NA	NA	Screening or filtering of a list of patients; direct referral	NA

NA, not applicable.

Discussion

In this review of 70 papers containing 90 populations of patients with complex needs, the research team observed a multitude of approaches to defining complex populations. The team presents a framework of simple language that distinguishes broad approaches (stratification, segmentation, and targeting) and categorizes specific types of criteria into 6 domains and 15 subdomains, with itemized dimensions of variation in criteria components. In the studies reviewed, population criteria from multiple domains were used in combination with one another, with expansive variability in how individual criteria were specified. The result was that nearly every study used a unique approach to identify patients with complex needs.

Nevertheless, approaches tended to cluster around specific conceptual and operational groupings. Age-based approaches were most ubiquitous but were relatively simple as they varied only in lower and upper bound thresholds. More complicated (and variable) was the specification of cost, utilization, and health conditions-based approaches. The team observed population definitions in these domains that varied along multiple different dimensions at once. Some examples of approaches were more nuanced than others, such as acute care utilization criteria that focused specifically on treat and release visits to the emergency department,¹⁴ or cost-based criteria that used multiple years of data to establish persistence of high spending.⁷⁴ The team also noted several population definitions that relied on more complex (and sometimes proprietary) computed scores that were based on data from multiple domains and/or subdomains.^21,72,73

Age-based (and income-based) criteria often were presented as byproducts of data considerations, such as data source or setting of the study (eg, studies based on Medicaid data were focused on low-income patients; studies based on Medicare data usually were focused on older adults) rather than as intentional patient selection decisions rooted in conceptual understanding.

The NAM report framed 3 core approaches to identifying populations with complex needs: stratification, segmentation, and targeting. The research team applied this framing to the domains of criteria observed in the complex care population definitions found in this pragmatic review. The cost- and utilization-based criteria in the framework can be thought of as stratification approaches, while age, health conditions, and income-based criteria can be thought of as segmentation approaches or targeting approaches, depending on their construction. Most population definitions reviewed combined stratification and segmentation approaches, using a mix of criteria capturing cost and/or utilization as well as health conditions, age, or income.

The team observed the use of subjective criteria for defining complex populations; in general, these were found in studies describing programs/interventions, when there were frontline processes related to intervention delivery through which subjective eligibility determinations could be applied. The addition of subjective criteria may be used to narrow an otherwise heterogenous group of patients and can be thought of as a segmentation approach (seeking more homogenous groups of individuals) or a targeting approach (seeking out specific individuals with actionable needs).

Amid the diversity of approaches observed, many were similar in intention and demonstrate that there are “schools of thought” that likely reflect the specific priorities of individuals in the complex care community with regard to how to identify patients with complex needs. “High utilizer” approaches based on costs and/or utilization combined with an age-based criterion were the most common. In some instances these criteria were combined with a subjective determination, reflecting a desire to, for example, reduce health care utilization among older adults who might benefit from a specific intervention or program. Other schools of thought that the team posited based on this review were represented by criteria focused on predictions (future risk or health) found in both cost and utilization domains; and approaches to identifying “clinically complex” patients by specifying criteria focused on measures of health conditions or functional status.

However, within schools of thought the research team observed an array of distinct criteria specifications, given the wide variability observed across multiple dimensions of each criterion and the combinations of criteria. For example, a common construct was the “frequent acute care utilizer,” but the team observed many variants of this construct: 4+ treat and release emergency department visits in the prior year¹⁴; 1+ inpatient stay in the prior year²⁵; 3+ emergency department visits or inpatient stays in the prior year³²; 2+ emergency department visits in the prior 28 days⁴⁶; predicted inpatient readmission risk score of ≥50 out of 100 within the next year⁵⁶; and so on.

Very little information is available about how important distinctions such as these are, in which conceptually similar definitions employ distinct thresholds or look-back periods. In general, studies rarely discussed the rationale for their approach to identifying complex patients, or the process through which population criteria were selected (eg, via expert opinion, health system prioritization, exploratory analyses). Without information about the implications of these varied approaches to defining constructs of complexity, it is very difficult to draw meaning from the complex care literature. Poor match of intervention services to patient needs can be one reason for program failure, and better understanding the implications of population selection decisions is critical to interpreting and comparing results of interventions.

Limitations

Although this study is a comprehensive assessment of the current landscape of approaches to conceptualizing and identifying patients with complex needs, it has several limitations. The team conducted a pragmatic review and the scan may not have included every paper in the literature that defined a population of patients with complex needs. Studies were included that were described by the authors as focusing on populations with complex needs, high needs, or high costs, but authors varied in their intentions. The team did not observe examples of selection criteria representing measures of social needs. However, this was not an explicit term in the search strategy. Furthermore, few data sources such as claims or electronic health records contain discrete data fields on social needs, and thus their absence from the review may be a reflection of their absence in underlying data systems.⁷⁹

The variability in approaches to identifying patients with complex needs has strengths and weaknesses. Given the general agreement that patients with complex needs are a composite of many distinct subgroups, testing of many different population criteria contributes to growing understanding of these subgroups. The diversity of approaches also provides examples of methods that may work in certain contexts (eg, utilization-based methods can be used in contexts that do not have access to cost information).

It may be that the time is right for leaders in the field of complex care to consider a more focused and standardized approach to definition of complex patient populations as one means to further the field through comparable and translatable evidence. The variability in the populations studied is a barrier to drawing generalizable conclusions about programs, population needs, or intervention opportunities. Importantly, it is also a potential source of ambiguity and confusion between stakeholders in the complex care community, who may have different mental maps regarding what constitutes complexity. However, it may be infeasible to standardize definitions of patients with complex needs because criteria chosen and how they are specified may depend on the setting, data available, and other factors.

Conclusion

Complex care program designers need more information about the comparability of the many distinct ways of identifying complex populations. The aim of this pragmatic review was to develop a detailed characterization of the ways that patients with complex needs are identified and to synthesize these approaches into a framework to better support informed population selection decisions for complex care interventions. Greater specificity in the language used to talk about population selection and more transparency surrounding the details of the population criteria and rationale for their use would be a substantial contribution to the emerging field of complex care and its expanding evidence base.

Footnotes

Acknowledgments

The authors express their gratitude to the following individuals who contributed to the conceptualization and development of this work and shared their insights generously: Melinda Abrams (The Commonwealth Fund); Jose Figueroa (Harvard University); Jessica Hughes (New York Health and Hospitals); Rachel Davis, Karla Silverman, and Stefanie Turner (Center for Health Care Strategies); and Mariana Saenz and Michelle Wong (Kaiser Permanente). We are also grateful to Jacob Quinton for his thoughtful comments on early drafts of this manuscript. Finally, the authors also wish to thank the leaders of the Kaiser Permanente CORAL Initiative and the Kaiser Permanente Care Management Institute's Care for Complex Needs Program.

A preliminary version of this work was presented as a poster at the 2019 AcademyHealth Annual Research Meeting in Washington, DC. A related abstract was selected for presentation at the 2020 AcademyHealth Annual Research Meeting. The abstract will be published as part of meeting proceedings although the meeting has been moved to a virtual format.

Author Disclosure Statement

All authors declare that there are no conflicts of interest.

Funding Information

This project was supported by Kaiser Permanente's Garfield Memorial Fund, under its Complex Care Collaboration: Research, Operations and Leadership (CORAL) portfolio.

Supplementary Material

Supplementary Table S1

References

Long

, Abrams

, Milstein

, et al. Effective care for high-need patients: opportunities for improving outcomes, value, and health. Washington, DC: National Academy of Medicine, 2017.

Blumenthal

, Anderson

, Burke

, Fulmer

, Jha

, Long

. Tailoring complex-care management, coordination, and integration for high-need, high-cost patients: a vital direction for health and health care. NAM Perspective. Discussion Paper, National Academy of Medicine, Washington, DC, 2016. Available at: nam.edu/perspectives/ Accessed September 10, 2020 .

The Better Care Playbook. https://www.bettercareplaybook.org/ Accessed August 8, 2019 .

Engel

GL.

The need for a new medical model: a challenge for biomedicine. Science, 1977; 196:129–136.

National Center for Complex Health and Social Needs. https://www.nationalcomplex.care/ Accessed August 8, 2019 .

Scarpati

Designing Efficient Systematic Reviews Using Economical Allocation, Creation and Synthesis of Medical Evidence. 2014. Available at: rand.org/pubs/rgs_dissertations/RGSD334.html Accessed September 10, 2020 .

Sagliocca

, De Masi

, Ferrigno

, Mele

, Traversa

. A pragmatic strategy for the review of clinical evidence. J Eval Clin Pract, 2013; 19:689–696.

Liberati

, Altman

, Tetzlaff

, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ, 2009; 339:b2700.

Davis

, Shen

, Shah

, et al. Segmentation of high-cost adults in an integrated healthcare system based on empirical clustering of acute and chronic conditions. J Gen Intern Med, 2018; 33:2171–2179.

10.

Anderson

, Balkrishnan

, Camacho

. Risk classification of Medicare HMO enrollee cost levels using a decision-tree approach. Am J Manag Care, 2004; 10(2 Pt 1):89–98.

11.

Ash

, Zhao

, Ellis

, Schlein Kramer

. Finding future high-cost cases: comparing prior cost versus diagnosis-based methods. Health Serv Res, 2001; 36(6 Pt 2):194–206.

12.

Bell

, Krupski

, Joesch

, et al. A randomized controlled trial of intensive care management for disabled Medicaid beneficiaries with high health care costs. Health Serv Res, 2015; 50:663–689.

13.

Bird

, Kurowski

, Dickman

, Kronborg

. Integrated care facilitation for older patients with complex health care needs reduces hospital demand. Aust Health Rev, 2007; 31:451–461; discussion 449–450.

14.

Capp

, Rosenthal

, Desai

, et al. Characteristics of Medicaid enrollees with frequent ED use. Am J Emerg Med, 2013; 31:1333–1337.

15.

Chakravarty

, Cantor

. Informing the design and evaluation of superuser care management initiatives: accounting for regression-to-the-mean. Med Care, 2016; 54:860–867.

16.

Chechulin

, Nazerian

, Rais

, Malikov

. Predicting patients with high risk of becoming high-cost healthcare users in Ontario (Canada). Healthc Policy, 2014; 9:68–79.

17.

Colla

, Lewis

, Kao

, O'Malley

, Chang

, Fisher

. Association between medicare accountable care organization implementation and spending among clinically vulnerable beneficiaries. JAMA Intern Med, 2016; 176:1167–1175.

18.

Crane

, Collins

, Hall

, Rochester

, Patch

. Reducing utilization by uninsured frequent users of the emergency department: combining case management and drop-in group medical appointments. J Am Board Fam Med, 2012; 25:184–191.

19.

Cunningham

PJ.

Predicting high-cost privately insured patients based on self-reported health and utilization data. Am J Manag Care, 2017; 23:e215–e222.

20.

de Stampa

, Vedel

, Buyck

, et al. Impact on hospital admissions of an integrated primary care model for very frail elderly patients. Arch Gerontol Geriatr, 2014; 58:350–355.

21.

DuBard

, Jackson

. Active redesign of a Medicaid care management strategy for greater return on investment: predicting impactability. Popul Health Manag, 2018; 21:102–109.

22.

Figueroa

, Frakt

, Lyon

, Zhou

, Jha

. Characteristics and spending patterns of high cost, non-elderly adults in Massachusetts. Healthc (Amst), 2017; 5:165–170.

23.

Figueroa

, Joynt Maddox

, Beaulieu

, Wild

, Jha

. Concentration of potentially preventable spending among high-cost medicare subpopulations: an observational study. Ann Intern Med, 2017; 167:706–713.

24.

Freund

, Kunz

, Ose

, Szecsenyi

, Peters-Klimm

. Patterns of multimorbidity in primary care patients at high risk of future hospitalization. Popul Health Manag, 2012; 15:119–124.

25.

Goldhar

, Daub

, Dhalla

, Ellison

, Purbhoo

, Sinha

. Integrated client care for frail older adults in the community: preliminary report on a system-wide approach. Healthc Q, 2014; 17:61–69.

26.

Guilcher

, Bronskill

, Guan

, Wodchis

. Who are the high-cost users? A method for person-centred attribution of health care spending. PLoS One, 2016; 11:e0149179.

27.

Hegney

, Buikstra

, Chamberlain

, et al. Nurse discharge planning in the emergency department: a Toowoomba, Australia, study. J Clin Nurs, 2006; 15:1033–1044.

28.

Hensel

, Taylor

, Fung

, de Oliveira

, Vigod

. Unique characteristics of high-cost users of medical care with comorbid mental illness or addiction in a population-based cohort. Psychosomatics, 2018; 59:135–143.

29.

Heslop

, Athan

, Gardner

, Diers

, Poh

. An analysis of high-cost users at an Australian public health service organization. Health Serv Manage Res, 2005; 18:232–243.

30.

Horn

, Crandall

, Moffett

, et al. The economic impact of intensive care management for high-cost medically complex patients: an evaluation of New Mexico's care one program. Popul Health Manag, 2016; 19:398–404.

31.

Horn

, Crandall

, Binder

, Sklar

. What happens to high-cost patients? An analysis of the trajectories of billed charges over time. Popul Health Manag, 2017; 20:362–367.

32.

Hudon

, Chouinard

, Dubois

, et al. Case management in primary care for frequent users of health care services: a mixed methods study. Ann Fam Med, 2018; 16:232–239.

33.

Iglesias

, Baggio

, Moschetti

, et al. Using case management in a universal health coverage system to improve quality of life of frequent Emergency Department users: a randomized controlled trial. Qual Life Res, 2018; 27:503–513.

34.

Izad Shenas

, Raahemi

, Hossein Tekieh

, Kuziemsky

. Identifying high-cost patients using data mining techniques and a small set of non-trivial attributes. Comput Biol Med, 2014; 53:9–18.

35.

Joynt

, Gawande

, Orav

, Jha

. Contribution of preventable acute care spending to total spending for high-cost Medicare patients. JAMA, 2013; 309:2572–2578.

36.

Kanzaria

, Niedzwiecki

, Montoy

, Raven

, Hsia

. Persistent frequent emergency department use: core group exhibits extreme levels of use for more than a decade. Health Aff, 2017; 36:1720–1728.

37.

Kobb

, Hoffman

, Lodge

, Kline

. Enhancing elder chronic care through technology and care coordination: report from a pilot. Telemed J E Health, 2003; 9:189–195.

38.

Kodner

DL.

Managing high-risk patients: the Mass General care management programme. Int J Integr Care, 2015; 15:e017.

39.

Lee

, Whitman

, Vakharia

, Taksler

, Rothberg

. High-cost patients: hot-spotters don't explain the half of it. J Gen Intern Med, 2017; 32:28–34.

40.

Luptak

, Dailey

, Juretic

, et al. The Care Coordination Home Telehealth (CCHT) rural demonstration project: a symptom-based approach for serving older veterans in remote geographical settings. Rural Remote Health, 2010; 10:1375.

41.

Malseptic

, Melby

, Connolly

. Complex care models to achieve accountable care readiness: Lessons from two community hospitals. Healthc (Amst), 2018; 6:74–78.

42.

Mattke

, Han

, Wilks

, Sloss

. Medicare home visit program associated with fewer hospital and nursing home admissions, increased office visits. Health Aff, 2015; 34:2138–2146.

43.

Matzer

, Wisiak

, Graninger

, et al. Biopsychosocial health care needs at the emergency room: challenge of complexity. PLoS One, 2012; 7:e41775.

44.

Meenan

, Goodman

, Fishman

, Hornbrook

, O'Keeffe-Rosetti

, Bachman

. Using risk-adjustment models to identify high-cost risks. Med Care, 2003; 41:1301–1312.

45.

Mion

, Palmer

, Anetzberger

, Meldon

. Establishing a case-finding and referral system for at-risk older individuals in the emergency department setting: the SIGNET model. J Am Geriatr Soc, 2001; 49:1379–1386.

46.

Moore

, Gerdtz

, Manias

, Hepworth

, Dent

. Socio-demographic and clinical characteristics of re-presentation to an Australian inner-city emergency department: implications for service delivery. BMC Public Health, 2007; 7:320.

47.

Muratov

, Lee

, Holbrook

, et al. Regional variation in healthcare spending and mortality among senior high-cost healthcare users in Ontario, Canada: a retrospective matched cohort study. BMC Geriatr, 2018; 18:262.

48.

Murphy

, Castro

, Sylvia

. Predictive modeling in practice: improving the participant identification process for care management programs using condition-specific cut points. Popul Health Manag, 2011; 14:205–210.

49.

Newman

, Bikowski

, Nakayama

, Cunningham

, Acker

, Bradshaw

. Outcomes of embedded care management in a family medicine residency patient-centered medical home. Fam Med, 2017; 49:46–51.

50.

Noel

, Parchman

, Williams

Jr , et al. The challenges of multimorbidity from the patient perspective. J Gen Intern Med, 2007; 22 Suppl 3:419–424.

51.

Orueta

, Garcia-Alvarez

, Aurrekoetxea

, Garcia-Goni

. FINGER (Forming and Identifying New Groups of Expected Risks): developing and validating a new predictive model to identify patients with high healthcare cost and at risk of admission. BMJ Open, 2018; 8:e019830.

52.

Porath

, Irony

, Borobick

, et al. Maccabi proactive Telecare Center for chronic conditions—the care of frail elderly patients. Isr J Health Policy Res, 2017; 6:68.

53.

Price-Haywood

, Petersen

, Burton

, et al. Outpatient complex case management: health system-tailored risk stratification taxonomy to identify high-cost, high-need patients. J Gen Intern Med, 2018; 33:1921–1927.

54.

Radcliff

, Cote

, Duncan

. The identification of high-cost patients. Hosp Top, 2005; 83:17–24.

55.

Rais

, Nazerian

, Ardal

, Chechulin

, Bains

, Malikov

. High-cost users of Ontario's healthcare services. Healthc Policy, 2013; 9:44–51.

56.

Raven

, Doran

, Kostrowski

, Gillespie

, Elbel

. An intervention to improve care and reduce costs for high-risk patients with frequent hospital admissions: a pilot study. BMC Health Serv Res, 2011; 11:270.

57.

Rea

, Kenealy

, Horwood

, et al. Integrated systems to improve care for very high intensity users of hospital emergency department and for long-term conditions in the community. N Z Med J, 2010; 123:76–85.

58.

Reid

, Evans

, Barer

, et al. Conspicuous consumption: characterizing high users of physician services in one Canadian province. J Health Serv Res Policy, 2003; 8:215–224.

59.

Rice

, Barnes

, Rastogi

, Hillstrom

, Steinkeler

. Tackling 30-day, all-cause readmissions with a patient-centered transitional care bundle. Popul Health Manag, 2016; 19:56–62.

60.

Riley

GF.

Long-term trends in the concentration of Medicare spending. Health Aff, 2007; 26:808–816.

61.

Roe

, Normand

, Wren

, Browne

, O'Halloran

. The impact of frailty on healthcare utilisation in Ireland: evidence from the Irish longitudinal study on ageing. BMC Geriatr, 2017; 17:203.

62.

Rose

, Hatzenbuehler

, Gilbert

, Bouchard

, McGill

. A population health approach to clinical social work with complex patients in primary care. Health Soc Work, 2016; 41:93–100.

63.

Ryan

, Abrams

, Doty

, Shah

, Schneider

. How high-need patients experience health care in the United States. Findings from the 2016 commonwealth fund survey of high-need patients. Issue Brief (Commonw Fund), 2016; 43:1–20.

64.

Sabik

, Bazzoli

, Carcaise-Edinboro

, Chandan

, Harpe

. The impact of integrated case management on health services use and spending among nonelderly adult Medicaid enrollees. Med Care, 2016; 54:758–764.

65.

Sarnak

, Ryan

. How High-Need Patients Experiences the Health Care System in Nine Countries. Issue Brief (Commonw Fund), 2016; 1:1–14.

66.

Schoen

, Osborn

, How

, Doty

, Peugh

. In chronic condition: experiences of patients with complex health care needs, in eight countries, 2008. Health Aff, 2009; 28:w1–w16.

67.

Scuffham

, Byrnes

, Pollicino

, Cross

, Goldstein

, Ng

. The impact of population-based disease management services on health care utilisation and costs: results of the CAPICHe trial. J Gen Internal Med, 2019; 34:41–48.

68.

Shippee

, Shippee

, Mobley

, Fernstrom

, Britt

. Effect of a whole-person model of care on patient experience in patients with complex chronic illness in late life. Am J Hosp Palliat Care, 2018; 35:104–109.

69.

Spoorenberg

SLW

, Wynia

, Uittenbroek

, Kremer

HPH

, Reijneveld

. Effects of a population-based, person-centred and integrated care service on health, wellbeing and self-management of community-living older adults: A randomised controlled trial on Embrace. PLoS One, 2018; 13:e0190751.

70.

Stokes

, Kristensen

, Checkland

, Cheraghi-Sohi

, Bower

. Does the impact of case management vary in different subgroups of multimorbidity? Secondary analysis of a quasi-experiment. BMC Health Serv Res, 2017; 17:521.

71.

Vojta

, Vojta

, TenHave

, Amaya

, Lavizzo-Mourey

, Asch

. Risk screening in a Medicare/Medicaid population administrative data versus self report. J Gen Intern Med, 2001; 16:525–530.

72.

Waxmonsky

, Giese

, McGinnis

, et al. Colorado access' enhanced care management for high-cost, high-need Medicaid members: preliminary outcomes and lessons learned. J Ambul Care Manage, 2011; 34:183–191.

73.

Wilber

, Allen

, Shannon

, Alongi

. Partnering managed care and community-based services for frail elders: the care advocate program. J Am Geriatr Soc, 2003; 51:807–812.

74.

Wrathall

, Belnap

. Reducing health care costs through patient targeting: risk adjustment modeling to predict patients remaining high cost. EGEMS (Wash DC), 2017; 5:4.

75.

Xing

, Goehring

, Mancuso

Care coordination program for Washington State Medicaid enrollees reduced inpatient hospital costs. Health Aff 2015;34:653–661.

76.

Yoon

, Chang

, Rubenstein

, et al. Impact of primary care intensive management on high-risk veterans' costs and utilization: a randomized quality improvement trial. Ann Intern Med, 2018; 168:846–854.

77.

Zulman

, Pal Chee

, Ezeji-Okoye

, et al. Effect of an intensive outpatient program to augment primary care for high-need veterans affairs patients: a randomized clinical trial. JAMA Intern Med, 2017; 177:166–175.

78.

Zulman

, Pal Chee

, Wagner

, et al. Multimorbidity and healthcare utilisation among high-cost patients in the US Veterans Affairs Health Care System. BMJ Open, 2015; 5:e007771.

79.

Hatef

, Rouhizadeh

, Tia

, et al. Assessing the availability of data on social and behavioral determinants in structured and unstructured electronic health records: a retrospective analysis of a multilevel health care system. JMIR Med Inform, 2019; 7:e13802.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.16 MB