Epidemiology

Multimorbidity is common in stroke and TIA (Tran et al., Reference Tran, Norton, Conrad, Rahimian, Canoy, Nazarzadeh and Rahimi2018; The Academy of Medical Sciences, 2018). Gallacher and colleagues (Reference Gallacher, Batty, Mclean, Mercer, Guthrie, May, Langhorne and Mair2014) collated the evidence for multimorbidity in stroke across 40 conditions in Scotland and found that the overwhelming majority (94.2%) of those with a stroke diagnosis in their GP records (n = 35 690) had one or more existing morbidities excluding stroke, whereas, only 48% of the comparator group (ie, no stroke, n = 1 388 688) had one or more morbidities. The two most frequent physical comorbidities for those with a diagnosis of stroke were hypertension and coronary heart disease, consistent with other studies (Barnett et al., Reference Barnett, Mercer, Norbury, Watt, Wyke and Guthrie2012; Bergman et al., Reference Bergman, Henriksson, Åsberg, Farahmand and Terént2015; O’Donnell et al., Reference O’donnell, Chin, Rangarajan, Xavier, Liu, Zhang, Rao-Melacini, Zhang, Pais and Agapay2016).

A large-scale study of multimorbidity in cardiovascular disease (ie, ischaemic heart disease, and stroke or TIA) across 56 conditions found that of 4.2 million UK adults, 229 205 individuals experienced cardiovascular disease (Tran et al., Reference Tran, Norton, Conrad, Rahimian, Canoy, Nazarzadeh and Rahimi2018). Between 2000 and 2014, rates of multi- and comorbidity increased fourfold (6.3–24.3%) for this population. However, this study did not report the comorbidities associated with stroke/TIA specifically. Also in the UK, Gallacher and colleagues (Reference Gallacher, Mcqueenie, Nicholl, Jani, Lee and Mair2018) found, in a sample of 8 751 stroke/TIA patients, patients with multimorbidity have increased risk of mortality and this risk increases as the number of comorbid conditions increases (regardless of type of condition). Both studies found higher numbers of comorbidities in females, older age groups, and people living in deprived areas (Gallacher et al., Reference Gallacher, Mcqueenie, Nicholl, Jani, Lee and Mair2018; Tran et al., Reference Tran, Norton, Conrad, Rahimian, Canoy, Nazarzadeh and Rahimi2018). Comorbid conditions are also associated with poorer functional outcome post-stroke (Karatepe et al., Reference Karatepe, Gunaydin, Kaya and Turkmen2008). However, pairings or clusters of comorbid conditions – types of co-occurring conditions – in relation to outcomes other than mortality, such as quality of life, were not explored.

The differences between the temporal relationship (ie, pre-stroke, post-stroke, unrelated) and timespan in which comorbid conditions present in relation to stroke/TIA also need further exploration given subsequent implications on treatment, care management, and health service financing and planning (Valderas et al., Reference Valderas, Starfield, Sibbald, Salisbury and Roland2009). One study found an association between stroke and subsequent increased risk of progression to heart disease and diabetes; however, the population was women only and self-report data were used (Xu et al., Reference Xu, Mishra, Dobson and Jones2018). A nuanced understanding of cardiometabolic and non-cardiometabolic comorbidities could help with developing strategies that better take these into account.

Measures of multimorbidity

Measuring multimorbidity is challenging due to a lack of international consensus regarding its conceptualisation, scope, and how conditions are defined. Comparison of multimorbidity prevalence rates across different populations and settings is hindered by availability of data, the variety of measurement tools, and the broad range of conditions included in existing multimorbidity measures (Fortin et al., Reference Fortin, Stewart, Poitras, Almirall and Maddocks2012; Lefèvre et al., Reference Lefèvre, D’ivernois, De Andrade, Crozet, Lombrail and Gagnayre2014).

A recent study concerning the key factors for consideration when measuring multimorbidity concluded that there is ‘no single “best” measure of multimorbidity (p. 6)’ (Griffith et al., Reference Griffith, Gruneir, Fisher, Nicholson, Panjwani, Patterson, Markle-Reid, Ploeg, Bierman and Hogan2018). Broadly, multimorbidity measures are characterised as non-weighted (ie, frequency counts of co-occurring diseases) or weighted indices (ie, accounting for condition severity, healthcare utilisation) (Huntley et al., Reference Huntley, Johnson, Purdy, Valderas and Salisbury2012). The most commonly used multimorbidity measures are disease counts, weighted measures particularly the Charlson Comorbidity Index, the Cumulative Illness Rating Scale, the Index of Coexistent Disease, and the Adjusted Clinical Groups System (Johnston et al., Reference Johnston, Crilly, Black, Prescott and Mercer2019). However, in stroke epidemiology research, frequency counts are most commonly used (Gallacher et al., Reference Gallacher, Batty, Mclean, Mercer, Guthrie, May, Langhorne and Mair2014; Lefèvre et al., Reference Lefèvre, D’ivernois, De Andrade, Crozet, Lombrail and Gagnayre2014; Bergman et al., Reference Bergman, Henriksson, Åsberg, Farahmand and Terént2015; Gallacher et al., Reference Gallacher, Mcqueenie, Nicholl, Jani, Lee and Mair2018; Tran et al., Reference Tran, Norton, Conrad, Rahimian, Canoy, Nazarzadeh and Rahimi2018; Xu et al., Reference Xu, Mishra, Dobson and Jones2018).

Whilst these indices are useful for a range of purposes (eg, clinical, research, health services planning), some are decades old (Linn et al., Reference Linn, Linn and Gurel1968; Charlson et al., Reference Charlson, Pompei, Ales and Mackenzie1987) and conditions included in these measurement tools might not reflect the contemporary conditions that are known to increase the risk of mortality. Furthermore, some measurement tools have been developed using a limited population (Charlson et al., Reference Charlson, Pompei, Ales and Mackenzie1987). A recent expert panel study has identified that current multimorbidity measures do not account for social determinants of health and mental health, and episodic conditions (ie, recurring conditions or previous diagnoses that have already been treated). Such factors are important to consider when developing models of care and health policy (Griffith et al., Reference Griffith, Gruneir, Fisher, Nicholson, Panjwani, Patterson, Markle-Reid, Ploeg, Bierman and Hogan2018). Therefore, a challenge for multimorbidity research is capturing the complex nature of multimorbidity – going beyond physical diagnoses – and ensuring that the appropriate measurement tools are a good fit with the study purpose and/or stakeholder objectives.

Managing multimorbidity in relation to stroke/TIA

The evidence for increasing multimorbidity in stroke/TIA has implications for treatment burden, including but not limited to polypharmacy (Ostwald et al., Reference Ostwald, Wasserman and Davis2006; Gallacher et al., Reference Gallacher, Batty, Mclean, Mercer, Guthrie, May, Langhorne and Mair2014), healthcare utilisation (Cassell et al., Reference Cassell, Edwards, Harshfield, Rhodes, Brimicombe, Payne and Griffin2018), and treatment adherence (Mair and May, Reference Mair and May2014; Tran et al., Reference Tran, Norton, Conrad, Rahimian, Canoy, Nazarzadeh and Rahimi2018). Importantly, there is evidence demonstrating that variations in multimorbidity patterns across different age groups (Tran et al., Reference Tran, Norton, Conrad, Rahimian, Canoy, Nazarzadeh and Rahimi2018) and socioeconomic backgrounds (Bray et al., Reference Bray, Paley, Hoffman, James, Gompertz, Wolfe, Hemingway, Rudd and Collaboration2018; Gallacher et al., Reference Gallacher, Mcqueenie, Nicholl, Jani, Lee and Mair2018) have profound and wide-ranging impacts on people’s functioning, well-being, and quality of life (Fortin et al., Reference Fortin, Lapointe, Hudon, Vanasse, Ntetu and Maltais2004; Navickas et al., Reference Navickas, Petric, Feigl and Seychell2016). For example, people living in areas of deprivation tend to have a first stroke earlier (Bray et al., Reference Bray, Paley, Hoffman, James, Gompertz, Wolfe, Hemingway, Rudd and Collaboration2018) and are at greater risk of experiencing comorbidities (Barnett et al., Reference Barnett, Mercer, Norbury, Watt, Wyke and Guthrie2012), when compared to their more affluent peers.

Stroke is a leading cause of disability and many patients have complex needs, including physical, emotional, social, communication, and cognition needs (Stroke Association, 2018). Post-stroke care often requires complex management plans to address patients’ rehabilitation, social care, and stroke prevention needs. Beyond the array of impacts strokes/TIAs can have on individuals, their experiences of primary healthcare present further challenges to service delivery. A qualitative systematic review of patients’ experiences of stroke management describes post-stroke ‘treatment burden’ which is intensified by poor communication and fragmented healthcare (Gallacher et al., Reference Gallacher, Morrison, Jani, Macdonald, May, Montori, Erwin, Batty, Eton and Langhorne2013). Additionally, a systematic review with meta-ethnography by Pindus and colleagues (Reference Pindus, Mullis, Lim, Wellwood, Rundell, Aziz and Mant2018) showed that often stroke survivors and their carers felt marginalised and abandoned due to the passivity of services. This passivity was characterised by constraints to access to care, limited continuity of care, poor communication between healthcare professionals as well as between providers and stroke survivors and their carers, and receiving varied information about stroke (Murray et al., Reference Murray, Ashworth, Forster and Young2003). These are interacting factors that make caring for patients with multimorbidity a challenge to healthcare providers and systems (Mair and May, Reference Mair and May2014).

Multimorbidity is likely to exacerbate this post-stroke treatment burden through multiple uncoordinated appointments, polypharmacy, and lack of continuity of care (Noël et al., Reference Noël, Chris Frueh, Larme and Pugh2005). Multimorbidity can also impact on patients’ rehabilitation and recovery; for example, comorbid knee arthritis impairs patients’ ability to fully engage in stroke rehabilitation which can manifest in frustration and required additional coping strategies (Wood et al., Reference Wood, Connelly and Maly2009). Similarly, anxiety and depression have been found to slow recovery from stroke (West et al., Reference West, Hill, Hewison, Knapp and House2010). Stroke patients with comorbidities are often excluded from rehabilitation clinical trials (Nelson et al., Reference Nelson, Mckellar, Yi, Kelloway, Munce, Cott, Hall, Fortin, Teasell and Lyons2017); therefore, evidence-based interventions and recommendations might not be appropriate for the large proportion of stroke patients with comorbidities.

Despite multimorbidity being commonplace, clinical guidelines remain, by and large, disease specific and can be potentially harmful to patients and burdensome for healthcare professionals (Boyd and Fortin, Reference Boyd and Fortin2010; Parekh and Barton, Reference Parekh and Barton2010; Guthrie et al., Reference Guthrie, Payne, Alderson, Mcmurdo and Mercer2012). Integrated care approaches for people with multimorbidity can offer opportunities to treat conditions that have common management strategies together (eg, hypertension and coronary heart disease, risk factors for stroke) including mental health comorbidities that could co-occur with physical conditions. However, a recent large-scale cluster-randomised controlled trial in GPs in England and Scotland tested a 3D approach to care (ie, a patient-centred way of managing dimensions of health, depression, and drugs) of patients with multimorbidity found no improvements in health-related quality of life. Nevertheless, the study found that such an approach enhanced patient-centred care, including patient satisfaction (Salisbury et al., Reference Salisbury, Man, Bower, Guthrie, Chaplin, Gaunt, Brookes, Fitzpatrick, Gardner and Hollinghurst2018). To date, the evidence for integrated care in primary care for people with multimorbidity is equivocal (Smith et al., Reference Smith, Soubhi, Fortin, Hudon and O’dowd2012; Salisbury et al., Reference Salisbury, Man, Bower, Guthrie, Chaplin, Gaunt, Brookes, Fitzpatrick, Gardner and Hollinghurst2018) and requires further investigation particularly in relation to stroke/TIA. Future studies testing complex interventions such as that of Salisbury et al.’s (Reference Salisbury, Man, Bower, Guthrie, Chaplin, Gaunt, Brookes, Fitzpatrick, Gardner and Hollinghurst2018) should include process evaluations, where feasible and cost effectiveness analysis if the findings are positive (Ramsey et al., Reference Ramsey, Willke, Briggs, Brown, Buxton, Chawla, Cook, Glick, Liljas and Petitti2005; Moore et al., Reference Moore, Audrey, Barker, Bond, Bonell, Hardeman, Moore, O’cathain, Tinati and Wight2015). Furthermore, an understanding of the temporal relationship between conditions offers the opportunity for preventative interventions, for example, to prevent the cluster of diabetes, heart disease, and stroke (Xu et al., Reference Xu, Mishra, Dobson and Jones2018).

Multimorbidity care practices vary and could be broadly classified into three groups: (i) those that are focussed on a specific disease (ie, ‘index disease’) and include other conditions as comorbid; (ii) those that are focussed on specific combinations of a number of chronic diseases; and (iii) those that are not confined to specific combinations of chronic diseases (Rijken et al., Reference Rijken, Hujala, Van Ginneken, Melchiorre, Groenewegen and Schellevis2018). Moreover, although interprofessional collaboration is a reported key feature of integrated care models (Lalonde et al., Reference Lalonde, Goudreau, Hudon, Lussier, Duhamel, Bélanger, Lévesque and Martin2012; Valentijn et al., Reference Valentijn, Schepman, Opheij and Bruijnzeels2013), a multitude of terms are used interchangeably in integrated care and interprofessional collaboration literature (Atwal and Caldwell, Reference Atwal and Caldwell2002), making our understanding of how care pathways and collaborative care practices are implemented in practice to address multimorbidity in stroke/TIA a challenge. As such, the range of approaches to managing multimorbidity in primary care can also introduce gaps in care and service delivery, with knowledge concerning best practices and health organisation remaining limited (Rijken et al., Reference Rijken, Hujala, Van Ginneken, Melchiorre, Groenewegen and Schellevis2018; The Academy of Medical Sciences, 2018). These require illumination, as identifying markers of successful integrated care and interprofessional collaboration can impact care and professional practice.

Few randomised controlled trials have investigated the effectiveness of interventions to enhance multimorbidity management; nevertheless, the available literature has focussed on interventions geared toward changes to the organisation of care or enhancing interprofessional collaboration (Smith et al., Reference Smith, Soubhi, Fortin, Hudon and O’dowd2012; Salisbury et al., Reference Salisbury, Man, Bower, Guthrie, Chaplin, Gaunt, Brookes, Fitzpatrick, Gardner and Hollinghurst2018). Given that stroke and TIA survivors frequently seek medical support through primary care, it is imperative that current practices, particularly in the context of multimorbidity, are better understood. This is in line with the shift toward integrated care models, which involve multiple healthcare professionals working together to address multimorbidity, such as GPs, district or community nurses and pharmacists (Rijken et al., Reference Rijken, Hujala, Van Ginneken, Melchiorre, Groenewegen and Schellevis2018). A particular challenge is determining which of these integrated care approaches are predominantly applied (including where and how), which are most effective, and which are most acceptable to stroke/TIA patients experiencing multimorbidity.

Going forward

Considering our current knowledge of multimorbidity in stroke/TIA and how these are managed, we outline several suggestions for advancing our understanding of these. First, a more detailed understanding of multimorbidity in stroke/TIA is needed. Specifically, pairings and/or clusters of multimorbidity in stroke/TIA need to be identified, and their relationships with and impact on clinical and patient-reported outcomes assessed. Understanding the impact of multimorbidity on stroke patients’ rehabilitation and recovery is necessary to improve health and social care post-stroke. Measures of multimorbidity and their burden are heterogeneous and thus have implications on how associations and outcomes are investigated, which need to be considered in relation to available data (Huntley et al., Reference Huntley, Johnson, Purdy, Valderas and Salisbury2012).

Studies of multimorbidity trends in stroke/TIA need to select measures of multimorbidity based on study purpose and stakeholder groups involved. Continuity of care, treatment burden, and patients’ experience could potentially be improved by efficient exchange of information between interdisciplinary healthcare providers, particularly specialists, GP, and the third sector. However, multiple barriers for optimal communication exist, such as lack of integration of electronic health records (Sadler et al., Reference Sadler, Porat, Marshall, Hoang, Curcin, Wolfe and Mckevitt2017). New models of care are required to address poor communication between healthcare providers to improve transition of care and follow-up for stroke/TIA patients in the context of multimorbidity.

Importantly, multimorbidity trends in stroke/TIA need to be explored in different contexts (eg, urban versus rural communities) in different countries (ie, high-income versus low- and middle-income), as well as different care settings (ie, primary care versus secondary care) to contribute to the ‘global atlas’ of multimorbidity (The Academy of Medical Sciences, 2018). Future research should explore the settings in which multimorbidity in stroke/TIA is managed. This needs to be better understood, especially in low- and middle-income countries, as multimorbid conditions have implications on healthcare resources allocation and patient-centred care (Valderas et al., Reference Valderas, Starfield, Sibbald, Salisbury and Roland2009). Indeed, the vast majority of the multimorbidity literature is derived from high-income countries, despite most of non-communicable diseases occurring in low- and middle-income countries (Xu et al., Reference Xu, Mishra and Jones2017). Specifically, the definitions, components, provider profiles, and quality indicators of integrated care approaches for stroke/TIA need to be better characterised. The evidence base for multimorbidity management and risk factors need to be built alongside epidemiological evidence on multimorbidity in low- and middle-income countries, which is limited (Xu et al., Reference Xu, Mishra and Jones2017). A further step would be to relate these integrated care approaches to health outcomes (eg, quality of life) and assess these approaches’ cost-effectiveness, which will require more prospective, longitudinal research. Extensive research output from high-income countries can inform future research in low- and middle-income countries (Xu et al., Reference Xu, Mishra and Jones2017). For example, the influence of social determinants of health on multimorbidity rates in these contexts should be explored. The burden of both communicable diseases and non-communicable diseases on multimorbidity should be assessed, particularly as communicable diseases remain a challenge in many countries.

Concluding thoughts

The number of stroke and TIA survivors has been increasing, and with this increase comes multiple challenges – including multimorbidity – to the healthcare system especially primary healthcare services. A more considered understanding of the combinations of diseases that co-occur with stroke/TIA in terms of how these vary across different patient subgroups living in different contexts and their associations with clinical and patient-reported outcomes serve the important purpose of: (i) identifying their risk factors for developing further morbidity, and/or experiencing another stroke or TIA and (ii) developing context-appropriate strategies for addressing wide-ranging and long-term needs, such as cognitive impairment and rehabilitation. From this understanding, new models of integrated care practices can be developed, informed by evidence and current effective integrated care practices. The landmark Alma-Ata Declaration’s 40^th anniversary brings to the spotlight the need for high-quality, integrated primary healthcare in the context of an ageing population and the rise of multimorbidity globally. Patient-centred, individualised, and well-coordinated care is required to optimise stroke recovery and reduce treatment burden in relation to patients’ morbidities. There is a tension within longer term stroke care as to the extent to which this is best delivered by generalists and the extent to which it should be delivered by specialists. A more nuanced understanding of the epidemiology of multimorbidity in stroke/TIA will help inform this debate.