Extending VBHC to population-based healthcare systems: defining a segmentation model, outcome goals and steering mechanisms

doi:10.21203/rs.3.rs-1640436/v1

Download PDF

Research Article

Extending VBHC to population-based healthcare systems: defining a segmentation model, outcome goals and steering mechanisms

https://doi.org/10.21203/rs.3.rs-1640436/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background: Value-based healthcare (VBHC) is considered the most promising guiding principle for a new generation of health service production. Many countries have attempted to apply VBHC to managerial and clinical decision-making. However, implementation remains in its infancy and varies between countries. The objective of the study is to help health systems in implementing a value-based approach by building a value-based population segmentation model for Health Authorities (HA).

Materials and Methods: First, we define the principles according to which segmentation models in healthcare could be developed. Second, we merge the theoretical characteristics of VBHC with population segmentation dimensions identified in previous literature and design a flow model that establishes population segments from these combinations. We then estimate the size of the segments based on national register data. Finally, we describe how the segmentation model we developed can be used by an HA to steer healthcare service production.

Results: The population can be divided into ten different segments based on relevant outcomes, goals, and the outcome measurement logic. These segments consist of healthy, help, increased risk, mild curable without risk, mild curable with risk, severe curable without risk, severe curable with risk, single chronic, multimorbid, and terminal. The representatives of Finnish HAs found the segments meaningful for evaluating and managing the healthcare system towards improved population health. The HA can use the outcome information for the segments to steer the service providers towards the outcome goals.

Conclusions: A value-based segmentation model for the entire population is needed if a HA wants to steer the healthcare system employing the principles of VBHC. Segmentation should be based on the outcome measurement logic and outcome measurements relevant to each segment and the number of segments has to be limited to get a comprehensive understanding of the whole population.

value-based health care

outcomes-based health care

segmentation

population health

health authority

steering

Value-based healthcare (VBHC) is considered the most promising guiding principle for a new generation of health services. Value in healthcare is defined as patient-relevant outcomes per patient-level costs. Thus, a VBHC-based system would shift focus from outputs to outcomes [1,2]. There are many attempts worldwide to apply VBHC to managerial and clinical decision-making, resource allocation, reimbursement schemes, and performance measurement. However, implementation remains in its infancy, varying between countries [3].

VBHC originally focused on health service production rather than the health of a population or the performance of a healthcare system. According to Mjåset et al. [3], the VBHC framework functions better in some healthcare systems than others.

VBHC best applies to medical conditions for which a clear diagnosis and a corresponding curative process with a definitive endpoint and outcomes exist [2]. The question remains how VBHC can be employed for long-term, chronic, and multimorbid cases - the primary concern of public health. On average 33% of the population in different countries have multiple chronic conditions [4].

We use the term health authority (HA) to refer to an organisation responsible for making health services available to a predefined population. HAs have an interest not only in the costs and outcomes for specific diseases, but in the current and future health of populations. The objectives of HAs parallel those of VBHC: to maximise health and minimise health disparities given specific budget allocations. In Beveridge systems, an HA represents a national or a regional entity. In insurance-based systems and health management organisations, HAs correspond to insurers and the public agencies that regulate them. Common for all HAs is that they have policy objectives and limited resources, and, therefore, value is relevant to them all.

Applying VBHC at the level of a HA requires understanding how outcomes should be defined, and measured, and how outcomes should be used to develop and manage health services. Given that a HA covers an entire population, managing the health of a population requires dividing it into meaningful segments.

Traditionally, segmentation in healthcare has relied on urgency and severity, clinical categories based on organ systems and diseases or diagnostic groups, principal methods, specialisation level, and demographic categories. [5,6,7] Segmentation models can be developed bottom-up by clustering empirical data [8,9]. Several proprietary segmentation models exist, such as Johns Hopkins Adjusted Clinical Groups (ACG) system, 3M’s Clinical Risk Groups (CRG), and Kaiser Permanente’s Senior Segmentation Algorithm (SSA) [10]. Most of the segmentation models (e.g., CRG and ACG) are designed primarily for cost prediction or cost accounting. This means that they have not been developed to support segmentation from the outcomes measurement perspective, and it is unlikely that the segments that behave homogenously with respect to healthcare resource use do so in health outcomes. Also, due to having a service production focus, these segmentations don’t necessarily include the whole population. For instance, CRG focuses on people with chronic conditions and SSA on older adults. Some of these classification systems have limited possibilities to be utilized on the population level simply because they divide the population into hundreds or even thousands of segments. It would be impossible for the HA to draw conclusions on the health status of the population if there are over a thousand segments to consider.

Segmentation models can also be developed top-down by theoretically postulating segmentation variables for specific purposes. Models for management purposes have been presented based on demand–supply combinations [11], governance [12], and predictability [13,14]. In addition, intuitively meaningful categories have been presented in the Bridges to Health model [15].

The segmentation models previously suggested are not directly applicable to Has, which typically rely on public health objectives, nor do they consider VBHC. Thus, our objectives are to develop a value-based population segmentation model for HAs, demonstrate the volume of each segment using register data, and to discuss how a HA can use the model to steer service production.

Our methodology follows the principles of design science [16], which focuses on creating methods and tools to solve practical problems. We combine a top-down, purpose-driven approach with verification through a bottom-up, empirical validation.

In the selection of segmentation dimensions and the development of the segmentation model, we use a methodology similar to Lillrank et al. (2010). First, the goal of segmentation is defined; second, the segmenting variables are selected and defined; third, the variables are cross-checked for combinations and overlap; and, finally, the key questions by which segments are identified are expressed as a flow model.

Once the segments have been defined, the volume estimates for each segment are calculated. As a case example, we calculated the volumes (proportions) of the segments in the population of Finland. In the last step, we describe the practical implications of the segmentation model by combining the objectives from different segments with governance mechanisms. We use the common classification of governance mechanisms [17,18], and describe how a HA can use each mechanism to steer healthcare providers towards the objectives of each population segment.

During the study, both segments and their outcomes as well as the implementation of outcome goals into governance mechanisms were validated in three expert workshops. The first workshop was held in June 2019, with invitations sent to the management and development personnel from regional HAs in Finland. In total, 50 participants were divided into six groups in which participants commented on and developed the segmentation model and outcome goals for each segment. Based on the results from that workshop, we refined the segmentation model. Two additional workshops were held in October 2019 with the management personnel of the Keski-Suomi and Pirkanmaa regions. Groups consisted of 10 to 15 individuals, including chief medical officers, directors of nursing, director of health services, and administrative staff members.

Availability of data and materials

The data generated and analyzed during this study is available from the corresponding author on request. The data of national registers related to the calculation of proportions of the segments is in individual level and requires research permit to be assessed.

Dimensions of Segmentation in Healthcare

Segmentation as a systematic methodology emerged within marketing during the post-WWII expansion of markets and consumer choices. [19] In earlier studies [6,7,15,20], the following patient-related segmentation dimensions have been used: health status, health-related risk, capabilities, multimorbidity, severity, and urgency. These dimensions and their relevance must be considered from the VBHC point of view. Supply-based dimensions, such as clinical methods, are irrelevant in population-level segmentation.

Health status

Traditionally, the World Health Organisation (WHO) has defined health as a ‘State of complete physical, mental, and social well-being, and not merely the absence of disease or infirmity’. Huber and colleagues [21] proposed a new definition for health as the ability to adapt and self-manage in the face of social, physical, and emotional challenges.

Health status can range from healthy to a terminal condition, with intermediate categories being curable and chronic conditions [6].

Health status can be viewed as the basis of value-based segmentation since the primary focus of VBHC is the outcome, typically measured as a change in one’s health status [1]. In addition, the number of simultaneous conditions (i.e., multimorbidity) is essential to evaluate. Multimorbidity is the presence of several diseases or conditions, often with a cut-off of two or more [22]. Multimorbidity is relevant to discern from single chronic conditions because VBHC can be applied to single conditions, but the value in multimorbid conditions requires a different measurement [2].

Health-related risks

Among others, a sedentary lifestyle, obesity, and tobacco use strongly associate with a variety of long-term adverse health outcomes [23,24]. When making resource allocation decisions, healthcare delivery organisations must assess modifiable health risks [25]. For a HA, understanding the population segment with increased health risks is relevant from the viewpoint of prevention and, therefore, should be included in the adapted VBHC model.

Capabilities

The capabilities of the population are related to health-seeking behaviour and how individuals engage in the co-creation of health. Capabilities represent combinations of knowledge (health literacy) and motivation (willingness to co-create) [26].

The capabilities of the individual must be considered in service production since they reflect the patient’s ability for self-care and willingness to engage in the co-creation of services. However, they do not affect the outcome goals or measurement logic and, therefore, we do not include them in the VBHC segmentation model.

Severity and urgency

Severity and urgency are key elements in many segmentation models [11,20]. Severity refers to the extent of damage or harm resulting from an incident or state if nothing is done. Urgency refers to the severity of an outcome as a function of waiting time.

Urgency is relevant to planning health services. However, to evaluate outcomes, no clear distinction is made between urgent and nonurgent conditions. Many pathways may begin with an urgent need, but then evolve similarly to nonurgent pathways. As an example, a curable trauma is initially urgent, but the outcome measurement logic is similar to conditions cured through elective surgery. Therefore, we do not include urgency in the segmentation model. Severity is essential to consider, however, since the expected outcomes and measurement logic differ significantly between minor and major cases.

Designing the Segmentation Model

Segmentation is completed for a defined purpose and specific goals. For a HA, the goals consist of promoting the health of a population and minimising health disparities [27-29]. In VBHC, the goal is to maximise value for patients, whereby value is defined as patient-relevant health outcomes achieved over a full cycle of care per dollar spent [1]. Combining these goals with the general segmentation principles, we define the primary criteria of value-based segmentation as follows:

Segmentation must be relevant in terms of managing health outcomes.
The individuals within each segment must have sufficiently similar expected outcomes measurable in a similar way.
Each segment must be sufficiently different in terms of outcome goals or measurement logic to justify separate consideration.
The identified population segments should include everyone; every person at every point in their life should fit into one of these categories.

The goals of VBHC and HAs align in maximising the health of a population while minimising inequalities within a given budget. Thus, segmentation must begin with the health goals of the population, and the health status of a population is the most relevant segmentation dimensions.

The health statuses are described in terms of the number of health conditions, the severity of the condition, and health behaviour–associated risks. These dimensions affect the goals of the patient groups and are thus important for a HA to consider. These combinations are presented in Table 1.

Table 1

Dimensions of the segmentation model and their relevance in terms of outcome measurement

Health status	Severity	Single / multimorbid	Health behaviour / risks
Healthy	Not relevant	Not relevant	Relevant: A healthy person may have behavioural issues and an elevated risk
Curable condition	Relevant: For severe: disease-specific frameworks For mild: generic assessment of outcomes	Not relevant: Measurement logic can follow disease-specific frameworks	Relevant: A person may have an elevated risk for a chronic condition
Chronic condition	Not relevant: Severity is an outcome measure	Relevant: For a single, chronic condition: disease-specific frameworks For multimorbid conditions: general health status	Relevant: A person may have elevated risk for another chronic condition
Terminal condition	Not relevant: Always severe	Not relevant	Not relevant

The health statuses are first categorised by severity. Healthy people cannot have a severe condition, while terminal cases are always severe. Curable conditions can be separated into mild and severe cases. Care processes for severe, curable conditions follow a disease-specific logic, and outcomes can be measured with disease-specific metrics, such as those provided in the International Consortium for Health Outcomes Measurement (ICHOM) standard sets. For mild, curable conditions (e.g., upper respiratory infections), it may not be necessary to have disease-specific measurement sets, since such an approach might be too burdensome. For chronic patients, severity can indicate the progression of a chronic condition. In this case, the severity of a disease can also be used as an outcome measure of how well the disease is managed. It is possible, however, that a chronic condition is severe from the outset, which must be considered when defining the outcome goals.

The health statuses can further be categorised as single versus multimorbid conditions. For multimorbid patients, it may be impossible to discern the effects of different diseases vis-à-vis functionality and self-reported health status. Thus, general quality-of-life measurements may be more relevant than disease-specific measurement frameworks. Even though multiple chronic conditions are common in the population, especially older people [30], outcome measurement studies focusing on multimorbid conditions remain limited [31].

A similar problem also exists, if a patient simultaneously has two or more curable conditions. For example, a person can have severe osteoarthritis, while simultaneously suffering from a cataract. When following the disease-specific measurement logic of VBHC, the other condition affects the measurement of the individual’s health status and functionality, and should somehow be taken into consideration.

The last relevant segmentation dimension is health behaviour, which refers to the risk faced by a healthy person of developing a chronic condition or by a chronically ill person of developing complications from a disease or another chronic condition. Preventing diseases is an important goal for a HA. In order to achieve this goal, it is important to follow the development of health risks in a population, many of which relate to the behaviour of individuals. Thus, classifying all groups (excluding terminal) with respect to health risk is relevant for segmentation.

The categories defined in Table 1 are arranged as a flowchart algorithm that produces mutually exclusive segments, as illustrated in Figure 1. The flowchart starts with the most severe condition and moves towards mild conditions and healthy individuals. Patients with severe conditions may also experience mild conditions, such as a multimorbid patient with a curable condition and health risks, although the goals for an HA are determined by the most severe condition. The assessments of health status is limited by what data can be collected from the population based on service use, or through screening, and surveys. This inevitable uncertainty is expressed with the phrase ‘not that I know of’. Thus, the result of segmentation is never 100% accurate—there are always individuals with diseases that are undiagnosed as well as individuals with elevated risks of which they themselves or their healthcare provides remain unaware.

The first question in the algorithm is: Does the person have a terminal condition? If so, palliative care is in order. Patient-relevant goals are associated with the quality of death, such as its timeliness and peacefulness [32]. If the answer to the first question is, ‘not that I know of’, the algorithm advances to the next question.

If there is no indication of multiple chronic conditions [33], the algorithm moves to examining the existence of a single chronic condition. If no chronic condition has been identified for the individual, the remaining options consist of a curable condition, an elevated risk, or being healthy.

If no disease has been identified, the individual can have an elevated health risk or can be healthy. The question ‘Does the patient have a health-related risk?’ can be asked at any point, since an elevated risk can co-exist with any medical condition (excluding a terminal condition). Therefore, the ‘increased risk’ segment can be divided further: 1) people who have no diagnosis, whereby risk is their only known medical issue and 2) people who have a curable or chronic medical condition, but are also at an increased risk for some other chronic condition.

The last question is: ‘Does the person use services without falling into any of the previous segments?’ People using services without a discernible reason are assumed to be seeking help for a perceived problem which has no externally observable manifestation or the need is such that it cannot be fulfilled through the means available within the healthcare system.

The exact outcome goals need to be defined in each organisation in order to align with the organisation’s strategy and goals. We have defined example goals to illustrate the differences between segments (see Table 2). Either the outcome goals or the outcome measurement logic for separate segments differ from each other (see Table 2), indicating that they are meaningful segments for an HA and fulfil the criteria for segmentation. For instance, the measurement logic for severe and mild curable conditions diverge: for severe, curable conditions specific measurement sets can be employed, but for mild conditions a simpler approach is justified. The goals for chronic conditions are the same regardless of the number of chronic conditions (single chronic versus multimorbid). However, the measurement logic is different for these groups.

Table 2

Goals and outcome measurement logic for each segment

Segment	Outcome goal	Outcome measurement logic
Healthy	Keep healthy	Routine surveys concerning health behaviour and health status
Help	Help to find valuable services	Routine surveys considering health behaviour and health status
Increased risk	Mitigate the risk	Risk-specific measures
Mild curable without risk	Solve the health problem	Light generic assessment if problem is solved (Patient Reported Outcome Measure, PROM)
Mild curable with risk	Solve the health problem Mitigate the risk	Light generic assessment if problem is solved (PROM) Risk-specific measures
Severe curable without risk	Recover from episodes of illness or injury	Health condition–specific measurement sets
Severe curable with risk	Recover from episodes of illness or injury Mitigate the risk	Health condition–specific measurement sets Risk-specific measures
Single chronic	Maintain (or improve) health status and functioning	Health condition–specific measurement sets
Multimorbid	Maintain (or improve) health status and functioning	Assessment of health status and functioning and possibly health condition–specific sets
Terminal	Quality of death	Further research is required

First, a HA needs to determine the size of the population for each segment described in Table 2. The size of the segments represents the first means of analysing the health of the population. At a system level, the objective is to prevent people from having multimorbid or chronic conditions and to keep people in the healthy segment as long as possible. Thus, changes in the proportions of segments can be utilised to assess the performance of the healthcare system and to manage the system.

As a case example, we calculated the volumes (proportions) of the segments in the adult (18+ year-old) population of Finland (4.4 million people) in 2018. Finland has unique Social Security ID as well as national registries for healthcare services, which enables reliable population level analysis. The estimates were based on individual level data from the national health care registers and cause of death register (The Finnish Institute of Health and Welfare, Statistics Finland).

First, the volume of terminal conditions was estimated through cause of death statistics from Statistics Finland. We excluded acute and trauma-based causes of death as they often do not imply a terminal condition as defined in this model, and considered only classes of malignant tumours, neurological causes other than strokes, and cardiac conditions other than ischemic heart disease.

Second, the number of chronic patients was defined as the number of patients with an ICD-10 code or ICPC-2 code in primary care implying a chronic condition as defined in Calderón-Larrañaga et al [34]. Multimorbid conditions were defined as patients having two or more of these chronic conditions and single chronic as having only one.

The number of severe cure conditions was defined as patients without chronic conditions, who had an inpatient episode in specialist care during one year. Mild cure and prevention were defined as the patients who had a contact with healthcare services during 2018 but did not belong in the above-mentioned segments. Prevention could not be distinguished from mild cure based on the information available in our database.

The group of people with an increased risk was difficult to estimate. Based on studies by the Finnish Institute for Health and Welfare [35], the prevalence of risk factors in the Finnish population ranges from 10% (regular smoking) to 40% (an elevated total cholesterol exceeding 6.5 mmol/l). Such studies do not consider whether individuals with an elevated risk for a specific disease already have another chronic condition or sought health services for curable conditions. Therefore, it is impossible to say how individuals with an elevated risk are divided among the ‘healthy’, ‘curable’ or ‘chronic’ categories.

The largest segments were multimorbid with 1.19 million patients (27 % of the population) and the single chronics with 1.59 million patients (36 % of the population) (see Figure 2).

The next step involves measuring the outcomes for each segment based on the goals and logic summarised in Table 2. Although the ultimate goal for the HA is to reach the health goals for the population segments, it can only try to obtain them through the actions of healthcare service providers, who have direct access to the population. What a HA can do is to utilise the health status and outcome information through integrating it in the governance mechanisms for each provider. This includes normative guidance, information sharing, and resource allocation principles as well as reimbursement schemes for providers. The challenge here is that one service provider may have several segments as customers and one segment may need multiple service providers.

Resource steering serves to allocate money and/or resources. In Table 3, we show how the outcome goals for different segments could be tied to providers’ incentives. E.g., in severe curable conditions, bundled payments [3,36] may support a value-based approach better than fee-for-service since the episode may consist of several service events (i.e., contacts, visits, and inpatient episodes).

A HA can set norms for service providers such as requirements to follow care guidelines, measure outcomes, or even set minimum requirements for outcomes depending on the segment. Setting norms for outcomes may raise concerns that providers will select patients with better expected outcomes, which needs to be handled using case-mix adjustments or other norms that limit possibilities for selection. Furthermore, information steering may be directed at both service providers and the population. Simply collecting and distributing outcomes information to providers and the general public may affect the behaviour of the service providers or the population.

Table 3

How to build governance mechanisms for service providers in different segments

Segment	Resource / incentives	Norms	Information
Healthy	Incentives for the promotion of a healthy lifestyle	Not relevant	Health literacy–related information
Help	Global budget, incentives for the promotion of a healthy lifestyle	Not relevant	Health literacy–related information, information on other services
Increased risk	Outcome-based, incentives for the promotion of a healthy lifestyle	Systematic evaluation of health-related risks of a population	Health literacy–related information
Mild curable (with or without risk)	Fee-for-service, out-of pocket with an outcome bonus	Access to care, requirements for outcome measurements	Outcome comparisons between providers for patients and professionals
Severe curable (with or without risk)	Bundled payment with an outcome bonus	Minimum requirements for outcome measurements	Outcome comparisons between providers for patients and professionals
Chronic	Regular fee with an outcome bonus	Clinical care guidelines, requirements for outcome measurements	Health literacy–related information, information supporting self-care, outcomes comparisons
Multimorbid	Regular fee with an outcome bonus	Clinical care guidelines, requirements for outcome measurements	Health literacy–related information, information supporting self-care, outcome comparisons
Terminal	Global budget	Quality of resources in end-of-life care	Quality of death information

The decision-makers of HAs as well as representatives of ministries found the model useful for steering the regional health services. There is a growing need of

The concept of VBHC and its application on the healthcare system level have not been widely studied. Mjåset et al (2020) found government involvement to be a key factor for VBHC implementation [3]. The HA needs a systematic segmentation and measurement framework to assess the value of the healthcare system and to implement a value-based steering model. Many countries have been building outcomes measurement based disease-specific quality registers, but the problem is that there may exist hundreds or thousands different patient groups which makes it impossible for HA to get a comprehensive picture of outcomes in the whole population.

The VBHC approach best suits single-disease health problems and treatment episodes. Yet, at the system level, the unit of analysis is the individual rather than an episode. Managing a healthcare system requires a top-down approach to segmentation to include everyone; every person at every point in their life should fit into one segment. The VBHC approach has its purpose: measuring outcomes for individual diagnoses or treatments enables comparisons across providers and treatments to support meaningful and actionable choices for patients [37]. Building the measurement logic bottom-up, disease-by-disease helps service providers improve care outcomes for specific patients, but it may not be the best solution for the system-level measurement of outcomes: either you end up with thousands of segments or you run the risk of omitting some individuals. Furthermore, all multimorbid cases fall into several segments. How outcome measurements should be organised in such cases has not been captured in disease-specific measurement systems.

The commonly used population-level segmentation model by Lynn et al. [15] focuses on the descriptive health service needs of a population and is not directly suitable for outcome measurement. The segmentation model of Vuik et al. [8] divides the population based on observed service use. This offers insights into the different needs of segments (e.g., emergency care, home care, and specialised care versus primary care), but as such is not applicable to defining relevant outcomes for each segment. For example, patients with chronic conditions fall into several segments, and, within all segments, there are individuals with no chronic conditions. The earlier segmentation models presented [8,15] are somewhat similar to ours, but they are data driven classifications and therefore they do not necessarily divide the population in mutually exclusive groups. Also, neither of their segmentation logics are based on the outcomes measurement logic.

More detailed grouping or segmentation models (SSA, CRG, ACG, etc) can be used on the service provider level. Based on the validation of our model in the expert groups consisting of management of regional they have too many groups or do not cover the whole population and therefore cannot be used on the population level analysis. Also, since they are based on homogenous resource use patterns, they do not follow the logic of homogenous outcome goals. CRGs or ACGs can be used for subgroups within the main segments of the model for more in-depth understanding of patient groups on the service provider level.

Additional studies are required to deliver more accurate estimations of the size of the different segments, particularly the ‘increased risk’ segment. In this study, only aggregate data could be used for the rough estimation of segment sizes. In the future, the sizes, outcomes, and costs of different segments should be examined using patient/citizen level data to test the validity and benefits of the segmentation model.

A value-based segmentation model for the entire population is needed if a health authority wants to steer the healthcare system employing the principles of value-based care. Segmentation should be based on the outcome measurement logic and outcome measurements relevant to each segment and the number of segments has to be limited to get comprehensive understanding of the whole population. We proposed a segmentation model which divides the entire population into mutually exclusive segments that differ from each other in terms of outcome measurements and outcome goals. We concretise the model by quantifying the segments and suggesting ways to incorporate the segment-specific outcome goals into traditional steering mechanisms used by HAs. This extension of the VBHC theory supports the implementation of VBHC principles at the healthcare system level. The implementation of VBHC has been limited so far [3].

ACG: Adjusted Clinical Groups

CRG: Clinical Risk Groups

HA: Health authority

ICHOM: International Consortium for Health Outcomes Measurement

VBHC: Value-Based Health Care

WHO: World Health Organization

SSA: Senior Segmentation Algorithm

Ethics approval and consent to participate:

The study used only aggregate data from national registers, thus no ethics approval was required

Consent for publication:

Not applicable. The manuscript does not include data from individual persons

Availability of data and materials:

Data is available through a formal data request from the Finnish Institute for Health and Welfare

Competing interests:

The authors declare that they have no competing interests

Funding:

The study was partly funded by Prime Minister’s Office of Finland. The funder had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.

Authors' contributions:

PT: Conseptualization, Methodology, Analysis, Writing – original draft; RL: Conseptualization, Methodology, Analysis, Writing – original draft; PM: Conseptualization, Writing – review and editing, ML: Data curation, Analysis; PL: Conseptualization, Writing – review and editing, Supervision. All authors read and approved of the final manuscript.

Acknowledgements:

Laura Pitkänen, Helena Tolkki and Mikko Valtakari

Porter ME. What is value in health care? N Engl J Med 2010;363(26):2477–81.
Porter ME, Larsson S, Lee TH. Standardizing Patient Outcomes Measurement. N Engl J Med 2016;374(6):504–6.
Mjåset C, Ikram U, Nagra NS, Feeley TW. Value-Based Health Care in Four Different Health Care Systems. NEJM Catalyst November 10, 2020.
Nguyen H, Manolova G, Daskalopoulou C, Vitoratou S, Prince M, Prina AM. Prevalence of multimorbidity in community settings: A systematic review and meta-analysis of observational studies. Journal of Comorbidity, 2019;9.
Vuik SI, Mayer E, Darzi A. Patient Segmentation Analysis Offers Significant Benefits For Integrated Care And Support. Health Aff. 2016a; 5:769–75
Lillrank P. The Logics of Healthcare. CRC Press/Taylor and Francis, Boca Raton, Fl. 2018.
Peltokorpi A, Linna M, Malmström T, Torkki P, Lillrank P. Five Focus Strategies to Organize Health Care Delivery, Int J Health Care Qual Assur 2016;29:2.
Vuik SI, Mayer E, Darzi A. A quantitative evidence base for population health: applying utilization based cluster analysis to segment a patient population. Popul Health Metr 2016b;14:44.
Shi Yan, Yu Heng Kwan, Chuen Seng Tan, Julian Thumboo and Lian Leng Low. A systematic review of the clinical application of data-driven population segmentation analysis. BMC Med Res Methodol 2018;18:121
Chong JL, Lim KK, Matchar DB. Population segmentation based on healthcare needs: a systematic review. Syst Rev. 2019;8:202
Lillrank P, Groop J, Malmström T. Demand and Supply–based operating modes—A Framework for Analyzing Health Care Service Production. Milbank Q, 2010;88(4):595–615.
Glouberman S, Mintzberg H. Managing the care of health and the cure of disease, Part I: Differentiation. Health Care Management Rev. 2001;26(1):56–92.
Christensen CM, Grossman JH, Hwang J. The Innovator's Prescription: A Disruptive Solution for Health Care. McGraw-Hill; 2009.
Hopp, WJ, Lovejoy WS. Hospital operations: Principles of high efficiency healthcare. Upper Saddle River, NJ: FT Press; 2012.
Lynn J, Straube BM, Bell KM, Jencks SF, Kambic RT. Using population segmentation to provide better health care for all: The “Bridges to Health” model. Milbank Q. 2007;85(2):185-208.
Dresch A, Lacerda DP, Antunes JAV. Design Science Research. In: Design Science Research. Springer, Cham; 2015.
Lundquist L. Förvaltningen i det politiska systemet. Studentlitteratur, Malmö. 2.uppl. 1977.
Bemelmans-Videc ML, Rist RC, Vedung E. (eds). Carrots, sticks & sermons: policy instruments & their evaluation. Routledge 2003.
Wedel M, Kamakura WA. Market segmentation: conceptual and methodological foundations. Springer Science+Business Media LLC; New York: 2000. 382p.
Garfield SR. The delivery of medical care. Perm J. 2006;2(10):46–55.
Huber M, Knottnerus JA, Green L, Horst H, Jadad AR, Kromhout D et al. How should we define health? BMJ 2011;343:d4163
Johnston MC, Crilly M, Black C, Prescott GJ, Mercer SW. Defining and measuring multimorbidity: a systematic review of systematic reviews, Eur J Public Health. 2019;29(1):182–9
Keeler E, Manning W, Newhouse J, Sloss E, Wasserman J. The external costs of a sedentary life-style. Am J Public Health. 1989;79:975–81.
Manning W, Keeler E, Newhouse J, Sloss E, Wasserman J. The Costs of Poor Health Habits. Cambridge, Mass: Harvard University Press. 1991:223.
Pronk NP, Goodman MJ, O'Connor PJ, Martinson BC. Relationship Between Modifiable Health Risks and Short-term Health Care Charges. JAMA 1999;282(23):2235–9
Ruger P. Health Capability: Conceptualization and Operationalization. Am J Public Health 2010;100:41–9.
Verweij, M, Dawson A. The Meaning of ‘Public’ in Public Health. In Dawson, A. and Verweij, M. (eds) Ethics, Prevention, and Public Health. Oxford: Oxford University Press. 2007.
Brülde B.. Dawson A. Health, Disease and the Goal of Public Health, Public Health Ethics, Cambridge University Press; 2008.
Münthe C. The Goals of Public Health: An Integrated, Multidimensional Model. Public Health Ethics, 2008;1(1):39–52.
Tinetti ME, Fried TR, Boyd CM. Designing Health Care for the Most Common Chronic Condition—Multimorbidity. JAMA 2012;307(23):2493–4.
Smith SM, Wallace E, O'Dowd T, Fortin M. Interventions for improving outcomes in patients with multimorbidity in primary care and community settings. Cochrane Database of Systematic Reviews 2016:3.
Olthuis G, Dekkers W. Medical education, palliative care and moral attitude: some objectives and future perspectives. Med Educ 2003;37:928–33.
NICE. Multimorbidity: clinical assessment and management NICE guideline Published: 21 September 2016 www.nice.org.uk/guidance/ng56
Calderón-Larrañaga A, Vetrano DL, Onder G, et al. Assessing and Measuring Chronic Multimorbidity in the Older Population: A Proposal for Its Operationalization. J Gerontol A Biol Sci Med Sci. 2017;72(10):1417-23.
Koponen P, Borodulin K, Lundqvist A, Sääksjärvi K and Koskinen S. (eds.) Terveys, toimintakyky ja hyvinvointi Suomessa. Finterveys 2017 -tutkimus. Raportti 4/2018. Helsinki: Finnish Institute for Health and Welfare. https://www.julkari.fi/handle/10024/136223
Dummit LA, Kahvecioglu D, Marrufo G, et al. Association between hospital participation in a Medicare bundled payment initiative and payments and quality outcomes for lower extremity joint replacement episodes. JAMA. 2016;316(12):1267–78
Pross C, Geissler A, Busse R. Measuring, Reporting, and Rewarding Quality of Care in 5 Nations: 5 Policy Levers to Enhance Hospital Quality Accountability. The Milbank Q 2017;95(1): 136–83.
Sakib MN, Shooshtari S, St John P, Menec V. The prevalence of multimorbidity and associations with lifestyle factors among middle-aged Canadians: an analysis of Canadian Longitudinal Study on Aging data. BMC Public Health 2019;19(1):243.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Extending VBHC to population-based healthcare systems: defining a segmentation model, outcome goals and steering mechanisms

Status:

Version 1

Abstract

Figures

Background

Methods

Results

Discussion

Conclusion

Abbreviations

Declarations

References

Additional Declarations

Status:

Version 1