EBP, IS, AND QI

Available online at www.sciencedirect.com

Nur s Out l oo k 6 8 ( 2 0 2 0 ) 4 3 0 �4 3 9 www.nursingoutlook.org

An integrated model of multimorbidity and

Th occurr NR018

*Co E-m

0029-6 https:

symptom science

Toni Tripp-Reimer, PhD, RN, FAAN, Janet K Williams, PhD, RN, FAAN*,

Sue E. Gardner, PhD, RN, FAAN, Barbara Rakel, PhD, RN, FAAN, Keela Herr, PhD, RN, AGSF, FGSA, FAAN,

Ann Marie McCarthy, PhD, RN, PNP, FNASN, FAAN, Linda Liu Hand, PhD, Stephanie Gilbertson-White, PhD, APRN-BC, Catherine Cherwin, PhD, RN

College of Nursing, The University of Iowa, Iowa City, IA

e authors are grateful to the anonym ed while the authors were supporte 081, the NINR was not involved in the p rresponding author: Janet K. Williams, ail address: [email protected] 554/$ -see front matter � 2020 Elsevier //doi.org/10.1016/j.outlook.2020.03.003

A B S T R A C T

Background: Prevalence and complexity of persons with multiple chronic condi- tions (MCC), also known as multimorbidity, are shifting clinical practice from a single disease focus to one considering MCC and symptoms. Although symp- toms are intricately bound to concepts inherent in MCC science, symptoms are largely ignored in multimorbidity research and literature. Purpose: Introduce an Integrated Model of Multimorbidity and Symptom Science. Methods: Critical integrative review and synthesis process. Findings: The model comprises three primary domains: 1. Contributing/ Risk Fac- tors; 2. Symptom/Disease/Treatment Interactions; and 3. Patient Outcomes. Discussion: The model highlights the multilevel nature of contributing factors and the recursive interactions among multiple etiologies, conditions, symptoms, therapies, and outcomes. Cite this article: Tripp-Reimer, T., Williams, J.K., Gardner, S.E., Rakel, B., Herr, K., McCarthy, A.M., Hand, L.

L., Gilbertson-White, S., & Cherwin, C. (2020, July/August). An integrated model of multimorbidity and

symptom science. Nurs Outlook, 68(4), 430�439. https://doi.org/10.1016/j.outlook.2020.03.003.

A R T I C L E I N F O

Article history: Received 13 December 2019 Received in revised form 13 March 2020 Accepted 21 March 2020 Available online May 29, 2020.

Keywords: Multimorbidity Multiple chronic conditions Symptoms Precision health Nursing Conceptual model

ous reviewers for their comments and suggestions. Although development of this article d by funding from the National Institute of Nursing Research (NINR) under grant P20 reparation of the article. University of Iowa College of Nursing, 50 Newton Road, Iowa City, IA 52242-1121. u (J.K. Williams). Inc. All rights reserved.

Introduction

A common challenge for clinicians and investigators in symptom andmultimorbidity science is the inability to anticipate (a) those who will develop multiple chronic conditions (MCC); (b) those with MCC who will develop associated symptoms and should be targeted for surveillance, preventive interventions, and/or spe- cific treatments; and (c) treatments or adjustments in

treatments required for various combinations of MCCs and symptom profiles. This challenge is due largely to the heterogeneity in disease and onset of symptom(s), progression, and outcomes, even among patients who receive the same diagnoses and treatments. As a result, clinical practice is often based on treatments that are most suited to the ‘‘average” patient but may not be appropriate for a particular person. Further- more, when people have MCC, the problems created by one condition may be aggravated by the symptoms

Nur s Ou t l o o k 6 8 ( 2 0 2 0 ) 4 3 0�4 3 9 431

or treatments of others (Price, Surr, Gough, & Ashley, 2019). MCC present significant health and health care bur-

dens that, over the past decade, have been identified as a national priority. In 2010, the Agency for Health- care Research and Quality (AHRQ) and U.S. Depart- ment of Health and Human Services (USDHHS) developed a Strategic Framework for MCC with the express purpose of catalyzing a paradigm shift for addressing chronic illnesses in the United States: mov- ing from an approach focused separately on individual chronic diseases to one that examines the interplay of coexisting chronic conditions and their treatments. This shift presents significant challenges for clinicians and researchers given the current state of the science. Although clinical practice guidelines (CPGs) have become widely implemented, few guidelines acknowl- edge MCC let alone provide recommendations to aid clinical decision making when managing coexisting, interacting conditions (Whitson et al., 2016; U.S. Department of Health and Human Services, 2010). As exceptions, the American Geriatrics Society, and the American Thoracic Society pioneered principles for incorporating MCCs into practice and for adapting guidelines (AGS, 2012; Fabbri et al., 2012). To date, how- ever, existing guidelines give little attention to the impact of MCCs on patients’ symptom experience. Although symptoms, either single or co-occurring,

have not been addressed in care of patients with MCC, symptoms should be a central concern in MCC science for several reasons. First, symptoms may emerge during the natural history of disease progression, they may be sequelae of therapies, or develop independently from personal or contextual risk factors (Grembowski et al., 2014; Miaskowski et al., 2015). Second, symptoms may alter the disease trajectory, resulting in new (or resur- gent) conditions, symptoms, or adverse events. Third, uncontrolled symptoms can cause considerable suffer- ing and health care costs (Wallace & Salive, 2013). Finally, symptoms are often the primary focus of con- cern for patients (evenmore than the disease conditions) and may be key drivers in patients’ ratings of their health status and use of health care services (Wallace & Salive, 2013; Willadson et al., 2016). Consequently, MCC science is incomplete without due consideration and inclusion of symptoms. The assessment and management of symptoms are

key aspects of nursing care. Symptom science is defined as research that seeks to understand the bio- logic and behavioral mechanisms underlying a single symptom and/or co-occurring symptoms (Cashion & Grady, 2015; NINR, n.d.) as well as the impact of symp- toms on important outcomes such as functioning, quality-of-life, and health care utilization. Notable gains have been made in both the biological and behavioral areas of symptom science. Molecular mechanisms that contribute to variability in the pre- sentation of symptoms such as pain, fatigue, and sleep disturbance have been identified (Besedovsky, Lange, & Haack, 2019; Kim, Barsevick, Fang, & Miaskowski,

2012; Lee, Dziadkowiec, & Meek, 2014). With regard to symptom management, effective interventions that can be delivered in person or via technology have been developed. However, one persistent challenge faced by clinicians and researchers is knowing which patients will develop symptoms, when, and who will respond positively to which interventions (Miaskowski, Dodd, West, et al., 2007; Sikorski, 2009; Lei, 2012). Although the symptom experience is well characterized for some specific conditions such as various types of can- cer, the symptom experience of individuals living with MCC has largely been ignored. Furthermore, over the past 15 years, the two fields of

multimorbidity science and symptom science have evolved as parallel, but largely separate entities. In 2017, the National Institute of Nursing Research issued fund- ing calls (RFA-NR-17-002 https://grants.nih.gov/grants/ guide/rfa-files/RFA-NR-17-002.html and RFA-NR-17-004 https://grants.nih.gov/grants/guide/rfa-files/rfa-nr-17- 004.html) for centers in biobehavioral research focused on complex MCC in human adult populations. In the process of developing the proposal for an exploratory center, it was clear that we needed to determine how symptom science relates to multimorbidity science. Consequently, we initiated a synthesis of literatures of multimorbidity and symptom sciences, which estab- lished the lack of a theoretical framework that brings these areas of science together. We then sought to iden- tify a theoretical framework that integrates these two scientific fields.

Background

As a result of scientific advancements, (i.e., antibiotics, public sanitation, and shifting lifestyles, more seden- tary work, abundant/processed foods), the 20th cen- tury saw simultaneous increases in longevity and rates of chronic illnesses. In response, health care shifted emphasis from acute and infectious disease to chronic illnesses. Rates of chronic disease have contin- ued to escalate rapidly: in 2000, 125 million Americans had a chronic condition and by 2009, that number rose to 145 million; between 2000 and 2030, the number of Americans with chronic conditions will have grown by 37%, an increase of 46 million people (Anderson, 2010). Many people with chronic disease have more than one chronic condition. The prevalence of MCC increased from 9.6% in 2000 to 17.1% in 2010. Although the over- all prevalence of MCC is high (affecting one fourth of all Americans), that rate increases dramatically with advancing age: more than 2/3 of persons over 65 have two or more conditions, and one third has four or more conditions (Whitson et al., 2016). The financial burden of MCC is extreme: medical

costs for persons with chronic conditions account for 75% of U.S. health care spending, and more than 90% of Medicare spending on older adults is targeted to per- sons with MCC (CMS Centers for Medicare and

432 Nur s Out l o o k 6 8 ( 2 0 2 0 ) 4 3 0�4 3 9

Medicaid Services, 2012; Parekh & Myers, 2014). Yet, this heavy expenditure has not yielded the desired improvements in quality of life for those affected. Patients with MCC tend to have greater health care costs due to more clinic visits, more and longer hospi- tal stays, greater adverse drug events, and unneces- sary lab tests/hospitalizations (Anderson, 2010; U.S. Department of Health and Human Services, 2010; Par- ekh & Myers, 2014). Correspondingly, patients with MCC have worse health outcomes than those with a single chronic condition, including decreased func- tioning and quality of life as well as greater depression, psychological distress, and mortality (Grembowski et al., 2014). The terms MCC, multimorbidity, and co- morbidity are all used within an overall context of the presence of more than one clinical condition. In this paper we adopt the terms as they were used by

the authors cited. The term ‘‘co-morbidity” was the first term used in the literature to refer tomore than one con- dition occurring concurrently. In 1970, Feinstein (Fein- stein, 1970) proposed restricting use of this term to a primary focus on a single “index disease” for which there are secondary, coexisting conditions; since then, co-morbidity is increasingly defined as the “existence or occurrence of any distinct additional entity during the clinical course of a patient who has the index disease under study” (Radner, Yoshida, Smolen, & Solomon, 2014; Van der Akker, 1996), although this distinction is somewhat less commonly made in the U.S. than in Europe, Canada, and Australia (Bayliss, Edwards, Steiner, & Main, 2008). With co-morbidity, the one index disease is regarded as the most important phenomena, and the second condition is salient only in relation to the first; that relationship may be classified as causal, complica- tion or coincidental (Radner et al., 2014). The terms ‘‘multiple chronic conditions” and ‘‘multi-

morbidity” are widely used interchangeably. Although multimorbidity is themost recent term to evolve, having only been recognized as a MeSH term since 2015 (Tug- well & Knottnerus, 2019), it is becoming the term used most frequently when no index disease is the focus. In contrast to co-morbidity, multimorbidity is regarded as a more patient-centered (than disease centered) construct and is more holistic and complex (Bayliss et al., 2014). Although there is no consensus definition for MCC or multimorbidity (Lef�evre et al., 2014; Willadson et al.,

Table 1 – Defining Features That Have Been Used to Cha

Defining Feature Examples

Number of conditions present in addi- tion to the index condition

1, 2, or 3

Antecedent risk factors present Obesity, hypertension Outcomes affected Function, inability to p

of daily living

Acute conditions present Infections Qualifying characteristics Cannot currently be cu

controlled through m other treatments

Presence of symptoms Pain, fatigue

2016), these terms tend to share the core feature of being characterized by two or more long-term diseases lasting at least 1 year in one individual (WHO, 2016; Willadson et al., 2016; Xu, Mishra, & Jones, 2017; Johnston et al., 2019). Additional defining features of multimorbidity that have been applied less consistently (see Table 1) include a cutoff for the number of other conditions pres- ent (one, two, or three), presence of antecedent risk fac- tors (e.g., obesity, hypertension), types of outcomes affected (e.g., function- inability to perform activities of daily living), presence of acute conditions (e.g., infec- tions), satisfying qualifying characteristics (e.g., cannot currently be cured but can be controlled through medi- cations or other treatments), and co-occurrence with symptoms (e.g., pain, fatigue). In attempting to definemultimorbidity, various ante-

cedents or consequences of MCC have been consid- ered, including severity (Willadsen et al., 2016), episodes of illness (Griffith et al., 2018), access or unco- ordinated care (Yarnell et al., 2017), polypharmacy (Yarnall et al., 2017), social determinants of health (Northwood, Ploeg, Markle-Reid, & Sherifalli, 2018), frailty (Yarnall, et al., 2017), presence of mental disor- ders overlapping with physical disorders (Yarnall et al., 2017), and how patients prioritize outcomes (Griffith et al., 2018). One result of having this hetero- geneity in antecedents and consequences of MCC is that adapting CPGs to treat MCC is challenging because sparse data exist to provide direction. Most interven- tion studies, particularly randomized clinical trials, limit the sample population variability, and specifi- cally exclude MCC and potentially “confounding” interacting treatments. Furthermore, when subjects with MCC are included in intervention studies, such as pragmatic clinical trials, the additional conditions are rarely identified nor systematically considered during analysis. Thus, the body of knowledge on MCC includes a wide scope of contributing factors all influ- encing an equally wide range of patient outcomes. In addition, there has been a lack of consistent ter-

minology used to describe the clinical practice of tai- loring treatment based on specific aspects of an individual patient. Several relatively new terms are used, often interchangeably: Personalized, Precision, Stratified, and P4 (predictive, preventive, personalized, and participatory) medicine. Precision science

racterize Multimorbidity

Citation

Wallace and Salive (2013)

Willadson et al. (2016) erform activities NINR (2017; RFA-NR-17-002 https://

grants.nih.gov/grants/guide/rfa- files/RFA-NR-17-002.html)

Le Reste et al. (2013) red but can be edications or

Yarnall et al. (2017)

Willadson et al. (2016)

Nur s Ou t l o o k 6 8 ( 2 0 2 0 ) 4 3 0�4 3 9 433

encourages one to identify subgroups of these risk/ responder profiles and symptom expression so that prevention and treatment interventions can be cus- tomized for patients within each subgroup. The terms do not imply or portray the difficulty in treating patients with MCC and they often are not used consis- tently, either within each term or across terms. Histor- ically, the term Precision Medicine referred to the tailoring of medical treatment to the individual char- acteristics of each patient, originally based on a patient’s genetic make-up or other molecular or cellu- lar analysis. According to the National Library of Medicine (2017),

Precision Medicine has been re-envisioned more broadly to includemore individual and contextual factors; to tar- get prevention as well as treatment, and to determine and target subgroups of similar patients (Whitson et al., 2016). The individualization or personalization comes as clinicians incorporate patients’ values, priorities, prefer- ences, lifestyles, and resources into clinical decision- making. More recently, DeVon, Rice, Pickler, Krause-Par- ello, and Richmond (2015) and Grady (2017) use the term Precision Health to apply the concept of precision medi- cine to health caremore broadly. In 2018, The University of Iowawas funded to establish

an exploratory center (P20 NR018081) to integrate symp- tom science with the study of MCC and focus on predic- tion of both risk and personalized treatment response. In the center, we approach the study of MCC through the lens of precision health and system science (e.g., an approach to understanding the complex, interconnected set of factors that result in identifiable patterns over time). Because adults who present with the same set of MCC may have different symptom expression and risk/ responder profiles, precision health science encourages us to identify subgroups of these risk/responder profiles and symptom expression, so that prevention and treat- ment interventions can be customized for patients within each subgroup. System science guides one to think about how a current clinical problem fits in the context of multimorbidity—and helping prioritize which problem to address first—being mindful of interactions among all diseases, symptoms, and treatments. To our knowledge, this is the first model that links together the concepts of MCC, precision health, and symptom sci- ence. The purpose of this paper is to describe the Inte- grated Model of Multimorbidity and Symptom Science, developed for the center, that can guide research, prac- tice, and policy.

Critical Literature Interpretive Synthesis

To create the Integrated Model of Multimorbidity and Symptom Science, we used the Critical Interpretive Syn- thesis (CIS) approach of reviewing the literature that was developed by Dixon-Woods and colleagues in 2006, and methodologically elaborated by Wilson et al. (2014) and McDougall (2015). The CIS approach has the core objective

of developing a theoretical framework based on insights and interpretation drawn from a broad range of relevant sources. In a CIS literature review, the ‘‘compass ques- tion” drives the literature review and evolves over time as the literature becomes more specified and themes emerge. The initial question is characterized as more like a compass than an anchor (Dixon-Woods et.al., 2006) in that it provides direction but does not limit searches. Over time, the compass question is refined. For our CIS review, the original compass question was:What is the fit between MCC and symptom science? The refined ques- tion that emerged was: How are symptom science and multimorbidity characterized and how are they related? A CIS review employs purposive literature sampling

and an inductive analytic approach. Unlike typical sys- tematic reviews, in a CIS review there are no rigid, lim- iting criteria for literature selection, for instance, focusing exclusively on studies employing randomized clinical trials (Dixon-Woods et al., 2006; McDougall, 2015; Wilson et al., 2014). Literature is critically ana- lyzed as a whole, rather than through a critique of individual papers. The process includes summarizing key substantive points with the goal of producing a theory or conceptual model. In developing the Integrated Model of Multimorbidity

and Symptom Science, purposive sampling was used initially to select papers that were clearly concerned with MCC and symptom science. As the review process proceeded, theoretical sampling was used to add, test, and elaborate on the emerging analysis and the devel- oping model. Because we were interested in similari- ties and differences in multimorbidity and symptoms science, our initial selection criteria were minimal. We did not exclude any papers on the basis of research methodology but did exclude papers if they were spe- cific to only one disease (Dixon-Woods, et al., 2006). Search strategies included searching electronic data- bases, searching websites, reference chaining, and contacting experts. We also used expertise within the team to identify relevant literature from adjacent fields not immediately or obviously relevant to the purpose of the review. In consultation with a library scientist, PubMed, Goo-

gle Scholar, and PsycINFO were searched, beginning with four key search terms for morbidity and symptom science: MCC, multimorbidity, comorbidities, and symptoms (Figure 1A). The purpose of the literature search was to identify

potentially relevant papers to provide a sampling frame, which eventually yielded approximately 1,200 records. As the search progressed, articles were grouped into six topical domains: Epidemiology, Meth- ods, Conceptual/Theoretical Models, Clinical Interac- tions, Interventions, and Outcomes (Figure 1B). As we became more familiar with the multimorbid-

ity literature, we extended our search to additional terms that resulted in an additional 1,658 abstracts and articles. Articles within the six domains were then clustered into smaller topical categories such as “risk condition, progression, treatment response,”

Figure 1 –Stepwise creation of a literature map organizing multimorbidity concepts. Note: Panel A: Four key search terms. Panel B: Emergence of six topical domains. Panel C: Development of final literature map reflect- ing the flow and clustering of content.

434 Nur s Out l o o k 6 8 ( 2 0 2 0 ) 4 3 0�4 3 9

“symptom science,” “function,” and “patient-centered practice guidelines.” Each category was reviewed for substantive content, and then integrated with other topics. The flow and clustering of content identified is reflected in the Literature Map (Figure 1C) that outlines and summarizes this process. Finally, the overall body of literature was synthesized within each category and inspected for relationships across categories. These were sorted into descriptive and critical analytic state- ments and the theoretical model emerged.

Description of the Model

The Integrated Model of Multimorbidity and Symptom Science highlights the multilevel nature of contribut- ing risk factors for multimorbidity as well as the com- plexity of the iterative interactions among multiple conditions, symptoms, and treatments (Figure 2). Our model adds an emphasis on articulating and examin- ing outcomes. It is well accepted that symptoms and chronic diseases arise from the combination of several risk factors ranging from genetic characteristics to environmental exposures (Arden et al., 2015; Grady, 2017; Papathomas, Molito, Richardson, Riboli, & Vineis, 2011; Radner et al., 2014; Xu et al., 2017). These broader individual and contextual factors have appeared in prior symptom models in nursing (Armstrong, 2003; Brant, Beck, & Miaskowski, 2010; Brant, Dudley, Beck, & Miaskowski, 2016; Dodd et al., 2001; Henley, Kallas, Klatt, & Swenson, 2003; Lenz, Pugh, Milligan, Gift, & Suppe, 1997), although without extensive elaboration. These broader factors are absent from the increasingly used NIH Symptom Science Model (Cashion & Grady,

2015), which emphasizes the contributions of genetic/ omic factors that are crucial, but alone are not suffi- cient to account for the complex interactions among conditions, symptoms and treatments. A recent paper on precision health by Hickey et al. (2019) exemplifies a Precision Health model that accounts for these types of complex interactions by including components of measurement, phenotype, genotype and other bio- markers, and intervention target discovery. The Inte- grated Model of Multimorbidity and Symptom Science is an even broader risk conceptual model that approaches risk more comprehensively. This broader approach reflects the transition from early formula- tions of precision medicine and the emphasis on genetics and biomarkers to precision health, that also encompasses broader contextual features. The model distinguishes itself through the following features: (a) Inclusion of Symptoms, (b) Specification of model domains that account for the complex interplay between risk factors, disease interactions, and patient outcomes, and (c) Risk/responder profiles.

Inclusion of Symptoms

Symptoms/co-occurring symptoms and multimorbidity have important similarities. Neither multimorbidity nor symptoms/co-occurring symptoms are unitary con- structs: they exhibit great heterogeneity in number and composition of conditions/symptoms clustered, the severity of illness, and types of functional limitations (Miaskowski et al., 2015; U.S. Department of Health and Human Services, 2010). Among the four most frequently used lists of multimorbidity (including the Medicare Chronic Condition Data Warehouse and the Charlson comorbidity index), the number of conditions in each

Figure 2 – Integrated model of multimorbidity and symptom science

Nur s Ou t l o o k 6 8 ( 2 0 2 0 ) 4 3 0�4 3 9 435

list ranges from 15 to 119 (Dattalo et al., 2017). The same variability is true for characterizing symptoms/ co- occurring symptoms, with little consistency across stud- ies in the number or composition of symptoms placed within a cluster (Miaskowski et al., 2017). Ultimately, the identification of which conditions or symptoms should be incorporated into multimorbidity science is contin- gent upon the research question being addressed and the purpose of the study (Wallace & Salive, 2013). While many single chronic diseases and some symptoms have been well characterized epidemiologically, this is not true for either combinations of MCC, or symptoms. In the literature, there is no consensus regarding (a) their overall incidence or prevalence; (b) the relative fre- quency of symptoms or co-occurring symptoms; (c) their major causal or contributing risk factors; or (d) their associated adverse outcomes. Although a conceptual model to guide research on

MCC and symptom science was not found in the litera- ture, teams of scholars in the United Kingdom, Canada, Denmark and have begun to link these concepts (Griffith, et al., 2018; Willadsen et al., 2016; Yarnall et al., 2017). While symptoms/co-occurring symptoms are intricately bound to the concepts inherent in MCC science through risk factors, chronic disease entities, treatment sequelae, self-management, and patient centered outcomes, symptoms are largely ignored in the context of MCC research in the United States. Exceptions to this observa- tion primarily involve reports examining the correlation of symptom counts with adverse outcomes. This neglect

may result from several factors: (a) symptoms are some- times considered ‘‘soft” or not rigorous data because they are primarily known through patient self-reports; (b) they tend to be difficult to access through electronic health records; unlike diseases and treatments with unique codes (e.g., ICD-10), symptoms are generally available only through difficult to retrieve free-text as ‘‘nursing notes”; and (c) they are often seen only as sec- ondary to a disease process and not a primary focus of interest (Wallace and Salive, 2013).

Model Domains

The Integrated Model of Multimorbidity and Symptom Science comprises three primary domains: Contribut- ing/Risk Factors; Symptom/Disease/Treatment Interac- tions; and Patient Outcomes. The complex interactions within and between these domains are a critical aspect of themodel. It is well accepted that symptoms and con- ditions arise from the combination of several risk factors ranging from genetic characteristics to environmental exposures (Arden et al., 2015; Grady, 2017; Papathomas et al., 2011; Radner et al., 2014; Xu et al., 2017).

Domain 1: Contributing/Risk factors The Contributing/Risk Domain is based on the World Health Organization Commission on Social Determi- nants of Health conceptual framework (2010) and con- sists of two clusters of factors: individual and contextual. Individual factors encompass (a)

436 Nur s Out l o o k 6 8 ( 2 0 2 0 ) 4 3 0�4 3 9

biophysiological elements (including omics, cellular mechanisms, organ systems (Bayliss, 2014) as well as interacting metabolic, psychoneuroimmunological, and mitochondrial networks (Lehman, David, & Gruber, 2017; Sturmberg, Bennett, Martin, & Picard, 2017) and (b) personal characteristics (demographics, lifestyles, priorities/personal values; Bayliss et al., 2014). Contex- tual factors encompass Environmental, Socio-cultural, and Health System elements (Bayliss et al., 2014; North- wood et al., 2018; WHO 2016).

Domain 2: Condition x Symptom x Treatment Interactions In the Condition x Symptom x Treatment Interaction domain, we depict that conditions, symptoms, and treatments may all act as antecedents to other symp- toms and conditions and outcomes over time. These recursive interactions add further complexity to the study of these phenomena. Our model highlights the multilevel nature of contributing factors and the com- plexity of the recursive interactions among multiple conditions, symptoms, and treatments.

Domain 3: Outcomes The Outcomes domain is primarily patient-centered and includes clinical and patient focused outcomes such as Morbidity, Mortality, Functional Status, Quality of Life, Symptom/Condition or Caregiver Burden, and Self-man- agement Optimization (Boyd et al., 2019; Krumholz et al., 2006; Parekh & Myers, 2014). Although there are other outcomes that are of interest to organizations and sys- tems, our model is intentionally focused on patient cen- tered outcomes. Outcomes are represented as having feedback loop interactions with the other domains. These recursive interactions add complexity to the study of these phenomena. The Outcomes domain links back to impact Contributing/Risk factors, then the Condition x Symptom x Treatment Interaction domain, and ulti- mately the Outcomes themselves. The relationships depicted here are consistent with principles of a dynamic systems model (Lehman et al., 2017; Pincus, Kiefer, & Beyer, 2017; Sturmberg et al., 2017).

Risk/Responder Profiles

A major focus of the Integrated Model of Multimorbid- ity and Symptom Science and of our center is to develop risk profiles for identifying which subpopula- tions of patients are most susceptible to developing conditions, symptoms, and poor outcomes, as well as which subpopulations are most likely to respond to specific treatments. Patient Risk Susceptibility Profiles and Treatment Responder Profiles encase the domains of Contributing/Risk factors and Symptoms x Condi- tions x Treatments in this model, with brackets to indicate the data from which they are constructed. This process brings precision health to an integrated symptom andmultimorbidity science framework. Risk profiles address two primary questions: what fac-

tors contribute to susceptibility and what factors predict

heterogeneous treatment response, including adverse patient outcomes. The first set of risk factors produce Risk Susceptibility Profiles that designate risk factors and mechanisms that contribute to patient susceptibility for (a) development, (b) recurrence, or (c) progression of a condition, symptom, or poor outcomes. The second pro- file type—Treatment Responder Profiles—predict what subgroups of patients will likely receive most benefit and least harm from interventions. The targeting of treat- ments (both pharmacological and nonpharmacological interventions), according to the individual and contextual risk factors shared by subgroups of patients, will provide a first step in aiding clinicians to manage multimorbidity and symptoms (Whitson et al., 2016). These two types of profiles are differentiated because previous work has shown that predictors for susceptibility to developing symptoms/conditions often differ from the predictors of treatment responders (Arden et al., 2015).

Model Summary

Taken as a whole, our emerging model will guide in predicting two sets of patients: those who will develop MCC and/or symptoms/ co-occurring symptoms (and should be targeted for surveillance or preventive inter- ventions) and those patients who will respond to par- ticular therapies. Further research is needed to test how well the model will foster identification of strati- fiers (individual/contextual as well as other conditions, symptoms, treatments, and outcomes) for risk suscep- tibility and treatment responders that can be used to create profiles to classify patients into appropriate subgroups for prevention or treatment interventions.

Managing Complexity

Because of the complexity of combining multimorbidity and symptom science constructs in a model that fosters prediction and analyzes the impact of multiple cluster- ing (both comorbidities and symptoms), advanced ana- lytic techniques are essential. To date, much of the MCC research has used latent class analysis that often has only demonstrated relatively obvious or common-sense results, or relied on linear assumptions, using explor- atorymodels that are insufficient for studyingmultimor- bidity (Sturgiss, Boeckxstaens, & Clark, 2019). It is also important to note that given the multilevels of risk fac- tors and the complexity of the recursive interactions among multiple conditions, symptoms, and therapies, research methods often need to go beyond assumptions of straightforward linear relationships. An analytical approach that does not investigate complex multifactor effects is likely to be ineffective in explaining the onset or sequalae of symptoms and chronic diseases. The investigation of interactions with traditional statistical methods has limitations related to low power andmodel constraints (Papathomas et al., 2011). Consequently, cap- turing the interplay between individual, biological and contextual risk factors requires new computational models and analytic techniques. These methods can

Nur s Ou t l o o k 6 8 ( 2 0 2 0 ) 4 3 0�4 3 9 437

help “move beyond understanding what works on aver- age to understanding what works for whom and in what situation” (Bayliss et al., 2008) which is essential in the examination of symptoms in people withMCC.

Adaptation of Clinical Practice Guidelines

Finally, although professional societies have largely taken the responsibility for clinical practice guideline development, clinicians, and investigators must be sen- sitized to the current deficits in the single disease focused guidelines, as well as the recommendations from AHRQ and professional societies about strategies for developing or adapting guidelines for clinical practice (Fabbri et al., 2012; Goodman et al., 2014; Shekelle,Woolf, Grimshaw, Sch€unemann, & Eccles, 2012; Uhlig et al., 2014; National Guideline Centre, 2016). The adaptation of CPGs for MCC is problematic

because sparse data exist to provide direction. The Inte- grated Model of Multimorbidity and Symptom Science is intended to facilitate investigators’ abilities to capture essential elements to track all multimorbidity and symp- tom patterns among study populations, incorporate individual level data on MCC into large data sets, and analyze the most common complex chronic conditions and symptoms for informative patterns. The prevalence and complexity of managing MCC, cou-

pled with increasing poor health outcomes and costs, is shifting clinical practice from a single disease focus to one considering multiple conditions and symptoms. The problem in considering MCC in customizing therapies requires that our science must transform from a single disease focus to MCC focus by incorporating advanced analytics and symptom science with research designs and computational models. This transformation also requires a retooling of clinical researchers who in general have not been prepared to integrate these new computa- tional models in formulating and answering different types of research questions and will allow us to identify subgroups of MCC combinations that will drive the cus- tomization of therapy.

Conclusions

Multimorbidity science is incomplete without inclu- sion and considerations of symptoms. The Inte- grated Model of Multimorbidity and Symptom Science depicts the intricate and complex interac- tions among multiple domains and iterative phases across Contributing & Risk Factors; Condition, Symp- tom, & Treatment Interactions; and Outcomes. This complexity calls for better means to document and track symptoms, inclusion of symptom management in Guidelines for people with MCC, and advanced analytics in order to manage the multilevel variables under consideration.

Acknowledgments

The authors wish to thank Jennifer DeBerg, OT, MLS for providing expertise and assistance throughout the litera- ture search process.

Supplementary materials

Supplementary material associated with this article can be found in the online version at doi:10.1016/j.out look.2020.03.003.

R E F E R E N C E S

American Geriatrics Society (AGS) Expert Panel on the Care of Older Adults with Multimorbidity. (2012). Guid- ing principles for the care of older adults with multi- morbidity: An approach for clinicians. Journal of the American Geriatrics Society, 60(10), E1–E25, doi:10.1111/ j.1532-5415.2012.04188.x.

Anderson, G. (2010). Chronic care: Making the case for ongoing care. Princeton, NJ: Robert Wood Johnson Foundation. Retrieved from: https://www.rwjf.org/content/dam/ farm/reports/reports/2010/rwjf54583.

Arden, N., Richette, P., Cooper, C., Bruy�ere, O., Abadie, E., Branco, J., . . ., Reginster, J. Y. (2015). Can we identify patients with high risk of osteoarthritis progression who will respond to treatment? A focus on biomarkers and frailty. Drugs & Aging, 32(7), 525–535, doi:10.1007/ s40266-015-0276-7.

Armstrong, T. (2003). Symptoms experience: A concept analysis. Oncology Nursing Forum, 30, 601–606, doi:10.1188/03.ONF.601-606.

Bayliss, E. A., Edwards, A. E., Steiner, F. J., & Main, D. S. (2008). Processes of care desired by elderly patients with multimorbidities. Family Practice, 25(4), 287–293.

Bayliss, E. A., Bonds, D. E., Boyd, C. M., Davis, M. M., Finke, B., Fox, M. H., . . ., Stange, K. C. (2014). Under- standing the context of health for persons with multi- ple chronic conditions: Moving fromwhat is to what matters. Annals of Family Medicine, 12(3), 260–269.

Besedovsky, L., Lange, T., & Haack, M. (2019). The sleep- immune crosstalk in health and disease. Physiological Reviews, 99(3), 1325–1380.

Boyd, C., Smith, C. D., Masoudi, F. A., Blaum, C. S., Dodson, J. A., Green, A. R., . . ., Tinetti, M. E. (2019). Deci- sionmaking for older adults withmultiple chronic condi- tions: Executive summary for the american geriatrics society guiding principles on the care of older adults with multimorbidity. Journal of the American Geriatrics Society, 67(4), 665–673, doi:10.1111/jgs.15809 Epub 2019Mar 10.

Brant, J. M., Beck, S., & Miaskowski, C. (2010). Building dynamicmodels and theories to advance the science of symptommanagement research. Journal of Advanced Nursing, 66, 228–240, doi:10.1111/j.1365-2648.2009 05179.x.

Brant, J. M., Dudley, W. N., Beck, S., & Miaskowski, C. (2016). Evolution of the dynamic symptomsmodel. Oncology Nursing Forum, 43(5), 651–654, doi:10.1188/16. ONF.651-654.

Cashion, A. K., & Grady, P. A. (2015). The National Insti- tutes of Health/National Institutes of Nursing Research

438 Nur s Out l o o k 6 8 ( 2 0 2 0 ) 4 3 0�4 3 9

intramural research program and the development of the National Institutes of Health Symptom Science Model. Nursing Outlook, 63(4), 484–487, doi:10.1016/j. outlook.2015.03.001.

CMS Centers for Medicare andMedicaid Services. (2012). Chronic conditions amongmedicare beneficiaries. Chart- book. Baltimore, MD: Centers for Medicare andMedicaid Services. Baltimore, MD: CMS. Center for Strategic Plan- ning http://www.cms.gov/Research-Statistics-Data-and- Systems/Statistics-Trendsand-Reports/Chronic-Condi tions/Downloads/2011Chartbook.pdf.

Dattalo, M., DuGoff, E., Ronk, K., Kennelty, K., Gilmore- Bykovskyi, A., & Kind, A. J. (2017). Apples and oranges: Four definitions of multiple chronic conditions and their relationship to 30-day hospital readmission. Jour- nal of the American Geriatrics Society, 65(4), 712–720, doi:10.1111/jgs.14539 PMCID: PMC5397355.

DeVon, H. A., Rice, M., Pickler, R. H., Krause-Parello, C. A., & Richmond, T. S. (2015). Setting nursing science priori- ties to meet contemporary health care needs. Nursing Outlook, 64(4), 399–401. http://dx.doi.org/10.1016/j.out look.2016.05.007.

Dixon-Woods, M., Cavers, D., Agarwal, S., Annandale, E., Arthur, A., Harvey, J., . . ., Sutton, A. J. (2006). Conduct- ing a critical interpretive synthesis of the literature. BMCMedical Research Methodology, 6, 35–42.

Dodd, M., Janson, S., Facione, N., Faucett, J., Froelicher, ES., Humphreys, J., . . ., Taylor, D. (2001). Advancing the science of symptommanagement. Jour- nal of Advanced Nursing, 33(5), 668–676, doi:10.1046/ j.1365-2648.2001.01697.

Fabbri, L. M., Boyd, Boschetto, C., Rabe, P., Buist, K. F., . . ., Yawn, A. S., & ATS/ERS Ad Hoc Committee on Integrat- ing and Coordinating Efforts in COPD Guideline Devel- opment. (2012). How to integrate multiple comorbidities in guideline development: Article 10 in Integrating and coordinating efforts in COPD guideline development. An official ATS/ERS workshop report. Proceedings of the American Thoracic Society, 9(5), 274–281, doi:10.1513/pats.201208-063ST.

Feinstein, A. (1970). The pre-therapeutic classification of comorbidity in chronic disease. Chronic Disease, 23, 456–489.

Goodman, R. A., Boyd, C., Tinetti, M. E., Von Kohorn, I., Parekh, A. K., & McGinnis, J. M. (2014). IOM and DHHS meeting onmaking clinical practice guidelines appro- priate for patients with multiple chronic conditions. Annals of Family Medicine, 12(3), 256–259.

Grady, P. A. (2017). Advancing science, improving lives: NINR’s new strategic plan and future of nursing sci- ence. Journal of Nursing Scholarship, 49(3), 247–248, doi:10.1111/jnu.12286.

Grembowski, D., Schaefer, J., Johnson, K. E., Fischer, H., Moore, S. L., . . ., Tai-Seale, M., & AHRQMCC Research Network. (2014). A conceptual model of the role of complexity in the care of patients with multiple chronic conditions. Medical Care, 52(Suppl 3), S7–S14, doi:10.1097/MLR.0000000000000045.

Griffith, L. E., Gruneir, A., Fisher, K. A., Nicholson, K., Panjwani, D., Patterson, C., . . ., Upshur, R. (2018). Key factors to consider whenmeasuring multimorbidity: Results from an expert panel and online survey. Journal of Comorbidity, 8, 1–9.

Henly, S. J., Kallas, K. D., Klatt, C. M., & Swenson, K. K. (2003). The notion of time in symptom experiences. Nursing Research, 52(6), 410–417.

Hickey, K. T., Bakken, S., Byrne, M. W., Bailey, D. E., Demiris, G., Docherty, S., . . ., Grady, P. A. (2019).

Precision health: Advancing symptom and self-man- agement science. Nursing Outlook, 67(4), 462–475, doi:10.1016/j.outlook.2019.01.003.

Johnston, M. C., Crilly, M., Black, C., Prescott, G. J., & Mercer, S. W. (2019). Defining andMeasuringmultimor- bidity: a systematic review of systematic reviews. Euro- pean Journal of Public Health, 29(1), 182–189.

Kim, H. J., Barsevick, A. M., Fang, C. Y., & Miaskowski, C. (2012). Common biological pathways underlying the psychoneurological symptom cluster in cancer patients. Cancer Nursing, 35(6), E1–E20, doi:10.1097/ NCC.0b013e318233a811.

Krumholz, H. M., Currie, P. M., Riegel, B., Phillips, C. O., Peterson, E. D., Smith, R., . . ., Faxon, D. P. (2006). A taxonomy for disease management. Circulation, 114, 1432–1445.

Lee, K. A., Dziadkowiec, O., & Meek, O. (2014). A systems science approach to fatigue management in research and health care. Nursing Outlook, 62(5), 313–321.

Lef�evre, T., d’Ivernois, J. F., De Andrade, V., Crozet, C., Lombrail, P., & Gagnayre, R. (2014). What do wemean by multimorbidity? An analysis of the literature on multimorbidity measures, associated factors, and impact on health services organization. Revue d’ Epide- miologie et de Sante Publique, 62(5), 305–314.

Lehman, B. J., David, D. M., & Gruber, J. A. (2017). Rethink- ing the biopsychosocial model of health: Understand- ing health as a dynamic system. Social and Personality Psychology Compass, 11, e12328. https://doi.org/10.1111/ spc3.12328.

Lenz, E. R., Pugh, L. C., Milligan, R. A., Gift, A., & Suppe, F. (1997). Themiddle-range theory of unpleasant symp- toms: An update.Advances in Nursing Science, 19(3), 14–27.

Lei, H., Nahum-Shani, I., Lynch, K., Oslin, D., & Murphy, S. A. (2012). A "SMART" design for building individualized treatment sequences. Annual Review of Clinical Psycholology, 8, 21–48.

Le Reste, J. Y., Nabbe, P., Manceau, B., Lygidakis, C., Doerr, C., Lingner, H., . . ., Lietard, C. (2013). The Euro- pean General Practice Research Network presents a comprehensive definition of multimorbidity in family medicine and long-term care, following a systematic review of relevant literature. Journal of the American Medical Directors Association, 14(5), 319–325, doi:10.1016/ j.jamda.2013.01.001.

McDougall, R. (2015). Reviewing literature in bioethics research: Increasing rigour in non�systematic reviews. Bioethics, 29(7), 1–7.

Miaskowski, C., Dodd, M., West, C., Paul, S. M., Schumacher, K., Tripathy, D., et al. (2007). The use of a responder analysis to identify differences in patient outcomes following a self-care intervention to improve cancer pain management. Pain, 129(1-2), 55–63.

Miaskowski, C., Dunn, L., Ritchie, C., Paul, S. M., Cooper, B., Aouizerat, B. E., . . ., Yates, P. (2015). Latent class analysis reveals distinct subgroups of patients based on symptom occurrence and demographic and clinical characteristics. Journal of Pain and Symptom Management, 50(1), 28–37, doi:10.1016/j.jpainsym- man.2014.12.011.

Miaskowski, C., Barsevick, A., Berger, A., Casagrande, R., Grady, P. A., Jacobsen, E., . . ., Xiao, C. (2017). Advancing symptom science through symptom cluster research: Expert panel proceedings and recommendations. JNCI: Journal of the National Cancer Institute, 109(4), djw253. doi:10.1093/jnci/djw253.

National Guideline Centre. (2016).Multimorbidity: Clinical assessment and management. LondonUK: National

Nur s Ou t l o o k 6 8 ( 2 0 2 0 ) 4 3 0�4 3 9 439

Institute for Health and Care Excellence (NICE) Guide- line [NG56]. Retrieved from: https://www.nice.org.uk/ guidance/ng56.

National Library of Medicine (2017). Genetics home refer- ence.Retrieved from: https://ghr.nlm.nih.gov/

National Institute of Nursing Research (NINR; 2019, November 4). NINR symptom science center. National Institutes of Health. Website. https://www.ninr.nih. gov/researchandfunding/dir#symptom-science.

Northwood, M., Ploeg, J., Markle-Reid, M., & Sherifalli, D. (2018). Integrative review of the social determinants of health in older adults with multimorbidity. Journal of Advanced Nursing, 74(1), 45–60, doi:10.1111/jan.13408.

Papathomas, M., Molito, J., Richardson, S., Riboli, E., & Vineis, P. (2011). Examining the joint effect of multiple risk factors using exposure risk profiles: Lung cancer in nonsmokers. Environmental Health Perspectives, 119(1), 84–91, doi:10.1289/ehp.1002118.

Parekh, A. K., & Myers, D. S. (2014). The strategic frame- work on multiple chronic conditions. Medical Care, 52 (Suppl 3), S1–S2, doi:10.1097/MLR.0000000000000094.

Pincus, D., Kiefer, A. W., & Beyer, J. I. (2017). Nonlinear dynamical systems and humanistic psychology. Journal of Humanistic Psychology, 58(3), 1–24, doi:10.1177/ 0022167817741784.

Price, M. L., Surr, C. A., Gough, B., & Ashley, A. (2019). Experiences and support needs of informal caregivers of people with multimorbidity: A scoping literature review. Psychology & Health, 19, 1–34. https://doi.org/ 10.1080/08870446.2019.1626125.

Radner, H., Yoshida, K., Smolen, J. S., & Solomon, D. H. (2014). Multimorbidity and rheumatic conditions— enhancing the concept of comorbidity. Nature Reviews Rheumatology, 10, 252–256, doi:10.1038/ nrrheum.2013.212.

Shekelle, P.,Woolf, S., Grimshaw, J. M., Sch€unemann, H. J., & Eccles, M. P. (2012). Developing clinical practice guide- lines: reviewing, reporting, and publishing guidelines; updating guidelines; and the emerging issues of enhanc- ing guideline implementability and accounting for comorbid conditions in guideline development. Imple- mentation Science, 7, 62 doi.org/10.1186/1748-5908-7-62.

Sikorskii, A., Given, C.W., You, M., Jeon, S., & Given, B. A. (2009). Response analysis formultiple symptoms revealed differences between arms of a symptommanagement trial. Journal of Clinical Epidemiology, 62(7), 716–724.

Sturgiss, E. A., Boeckxstaens, P., & Clark, A. M. (2019). Mul- timorbidity and patient-centred care in the 3D trial. The Lancet, 393, 127.

Sturmberg, J. P., Bennett, J. M., Martin, C. M., & Picard, M. (2017). Multimorbidity as the manifestation of network disturbances. Journal of Evaluation in Clinical Practice, 23 (1), 199–208.

Tugwell, P., & Knottnerus, J. A. (2019). Multimorbidity and comorbidity are now separate MESHheadings. Journal of Clinical Epidemiology, 105, vi–viii, doi:10.1016/j.jcli- nepi.2018.11.019.

Uhlig, K., Leff, B., Kent, D., Dy, S., Brunnhuber, K., Burgers, J., et al. (2014). A framework for crafting clini- cal practice guidelines that are relevant to the care and management of people with multimorbidity. Journal of General Internal Medicine, 29(4), 670–679, doi:10.1007/ s11606-013-2659-y.

U.S. Department of Health and Human Services. (2010). Multiple chronic conditions—A strategic framework: Opti- mum health and quality of life for individuals with multiple chronic conditions. Washington, DC: U.S. Department of Health and Human Services. Retrieved from: https:// www.hhs.gov/sites/default/files/ash/initiatives/mcc/ mcc_framework.pdf.

van den Akker, M. (1996). Comorbidity or multimorbidity: What’s in a name? European Journal of General Practice, 2, 65–70.

Wallace, R. B., & Salive, M. (2013). The dimensions of mul- tiple chronic conditions: Where do we go from here? A commentary on the Special Issue of Preventing Chronic Disease. Preventing Chronic Disease, 10, E59, doi:10.5888/pcd10.130104.

Whitson, H. E., Johnson, K. S., Sloane, R., Cigolle, C., Pieper, C. F., Landerman, L., et al. (2016). Identifying patterns of multimorbidity in older Americans: Appli- cation of latent class analysis. Journal of the American Geriatrics Society, 64(8), 1668–1673, doi:10.1111/ jgs.14201.

Willadsen, T. G., Bebe, A., Koster-Rasmussen, R., Ejg Jarbel, D., Domit Guassora, A., GochWaldorff, F., . . ., de Fine Olivarius, N. (2016). The role of diseases, risk fac- tors and symptoms in the definition of multimorbidity- A systematic review. Scandanavian Journal of Primary Health Care, 34(2), 112–121.

Wilson, M., Ellen, M. E., Lavis, J., Grimshaw, J., Kaelan, A., Moat, A., . . ., Samra, K. (2014). Processes, contexts, and rationale for disinvestment: A protocol for a critical interpretive synthesis. Systematic Reviews, 3, 143. http:// www.systematicreviewsjournal.com/content/3/1/143.

World Health Organization. (2016).Multimorbidity: Techni- cal series on safer primary care Geneva: 2016. Licence: CC BY-NC-SA 3.0 IGO.

Xu, X., Mishra, G. D., & Jones, M. (2017). Evidence onmulti- morbidity from definition to intervention: An overview of systematic reviews. Ageing Research Reviews, 37, 53– 68, doi:10.1016/j.arr.2017.05.003.

Yarnall, A. J., Sayer, A. A., Clegg, A., Rockwood, K., Parker, S., & Hindle, J. V. (2017). New horizons in multimorbidity in older adults. Age and Ageing, 46, 882–888.