Service Delivery Models to Maximize Quality of Life for Older People at the End of Life: A Rapid Review

Design

We conducted a rapid scoping review to systematically search, select, and synthesize knowledge regarding our aims to map key concepts, types of evidence, and gaps in research.26 We planned the review in accordance with the Centre for Reviews and Dissemination27 guidance and reported it in accordance with the Preferred Reporting Items for Systematic Reviews and Meta‐Analysis (PRISMA) statement.28

Eligibility Criteria

We included systematic reviews published between January 2000 and October 2017 that examined service delivery models aimed at maximizing older people's quality of life at the end of life. Our primary intention was to assess the evidence for current models.

Operationally, we defined reviews of older people as those where 50% or more of the included population were 60 or more years old, and were at the end of life; that is, those people described as being in the last one to two years of life, using a service typically accessed during an advanced stage of disease (eg, specialist palliative care, nursing home) or having advanced disease. We defined a service delivery model as “an overarching design for health care service provision with multiple components and interacting elements.”29 Delivery models focusing on postdeath intervention were outside the scope of the review, and we excluded reviews of a single component intervention, for example, the provision of an assistive device. We did include reviews containing data on outcomes relating to quality of life, function, or dignified end‐of‐life care. Eligible reviews had to draw on more than one data source30 in order to identify studies using experimental designs. Narrative reviews, or those describing case studies or series or descriptive studies only, were not eligible.

Search Strategy

We devised an electronic search strategy with an information specialist using a combination of MeSH and full‐text search terms developed for MEDLINE and adapted for other databases as necessary. The MeSH terms included “Terminally ill” or “Palliative care” for the population, “Hospice and palliative care nursing” or “Hospice care” for the intervention, and “Quality of life,” “Pain management,” or “Activities of daily living” for outcomes of interest. The Boolean operators OR and AND were used for MeSH terms within and across the population, interventions, and outcomes of interest. Key search terms were used as free text and also with the use of the truncation symbol to retrieve variations in the terminology (details in Online Appendix 1). We conducted electronic database searches in MEDLINE, CINAHL, EMBASE, and the Cochrane Database of Systematic Reviews. Our searches were restricted to human subjects and to systematic reviews using a filter developed by Lunny and colleagues.31 There was no restriction on the language of the publication. We searched the gray literature using hand searching; scanning reference lists, textbooks, and policy documents; and contacting experts in the field to seek potentially relevant research material, including ongoing and unpublished research.

Study Retrieval

Our search retrieved 2,238 articles. After removing duplicates and screening titles and abstracts, we retrieved 165 full‐text articles for further appraisal, of which 72 articles were eligible to be retained (Figure 1). Table 1 is an overview of these 72 reviews. An increasing number of reviews were published over time (2000‐2009, n = 17; 2010‐onward, n = 56) and the majority (49 of 72, 68%) limited their search strategy to articles published in 2000 or onward. All WHO regions were represented in the reviews, although the reviews predominantly included studies from the Americas (n = 52 of 72), European (n = 46 of 72), and the Western Pacific (n = 28 of 72) regions. Reviews including studies from other WHO regions and LMICs were less common; examples are Africa (Uganda,39 Kenya,40 Zambia, and South Africa39, 41), Southeast Asia (India, Nepal, and Pakistan41, 42), and the East Mediterranean (with studies from Israel43, 44, 45, 46, 47) (details in Online Appendix 3). Most of the reviews (71 of 72, 99%) contained effectiveness data, of which 20 of 72 (28%) reported meta‐analysis. Data from more than 784,983 individuals were available, and when stated, samples sizes ranged from 87 to 254,717 (see Table 1). The included reviews encompassed a breadth of service delivery models for older people nearing the end of life that intended to improve quality of life (see Online Appendix 2 detailing the aim for each included review).

Quality Appraisal

The median methodological score was 6 (range 0‐11). Most reviews were categorized as moderate (42 of 72, 58%), or low (20 of 72, 28%), with only a few considered as high quality (10 of 72, 14%). The methods used to combine studies were appropriate in most of the reviews (63 of 72, 88%); 53 of 72 (74%) reviews used an a priori design; and 51 of 72 (71%) used a comprehensive literature search strategy. However, only 22 of 72 (31%) reviews detailed the status of publication in the inclusion criteria (eg, gray literature), and 21 of 72 (29%) included a conflict of interest and assessed the likelihood of publication bias (details in Online Appendix 4). Owing to the heterogeneity of studies, we made a post hoc decision not to use quality criteria in any sensitivity analysis.

Target Outcomes

In total, 117 separate outcomes were measured across the reviews (details in Online Appendix 6). Only 25 of 117 (21%) were used in a meta‐analysis. Outcomes were grouped into five main themes: quality of life (including symptoms and well‐being), functional outcomes, dignified EoLC, health service use and costs, and survival (Figure 2). The most commonly targeted outcomes were quality of life (43 of 72, 60%), satisfaction with care (37 of 72, 51%), survival (28 of 72, 39%), physical function (25 of 72, 35%), health care costs (25 of 72, 35%), hospital length of stay (23 of 72, 32%), and psychiatric symptoms (23 of 72, 32%). Both Integrated Geriatric Care and Integrated Palliative Care placed equal emphasis on measures for quality of life (17 of 25, 68%, versus 19 of 30, 63%) and health service use (17 of 25, 68%, versus 19 of 30, 63%). The overarching models showed variation in outcome measures relating to symptoms that were more common in Integrated Palliative Care (21 of 30, 70%, versus 11 of 27, 44%) and to physical function and survival, which were more common in Integrated Geriatric Care (15 of 25, 60%, versus 8 of 30, 27%, and 14 of 25, 56%, versus 12 of 30, 40%, respectively). Reviews across the classifications considered survival in order to assess both increasing longevity and as a proxy for the potential harm of shortening life.

Impact of Service Delivery Models

Table 3 details the results from the meta‐analyses and narrative syntheses for the outcomes and impacts reported. We present these data in relation to the proposed overarching classifications, key processes, and components. Meta‐analyses were more frequently conducted in recent studies (2000‐2010, n = 10; 2011‐onward, n = 18). There was no discernible trend in reported effectiveness according to the date of publication.

Forty‐seven reviews analyzed the impact on quality‐of‐life outcomes, nine of which reported meta‐analyses. Pooled analyses were reported for health‐related quality of life (physical and psychological) (n = 4/9) and individual and/or multiple symptoms (n = 5/9). All pooled estimates of effect demonstrated effectiveness on quality‐of‐life outcomes. Effectiveness on health‐related quality of life was evident for the overarching classifications of Integrated Geriatric Care 40, 53, 57, 59, 94 and Integrated Palliative Care,40, 59, 62 the respective key processes of comprehensive assessment, case management and collaborative working, and key components (see Table 3).

Twenty‐eight reviews reported a narrative synthesis focusing on quality of life (excluding symptoms): 13 reported the models were effective; 11 found inconsistent evidence; and 4 found no effect. Symptoms were reported in narrative form in 37 of the reviews, with 19 finding the model(s) to be effective, 16 finding inconsistent results, and 2 finding no effect.

Impact on function was reported in 25 reviews, with 6 providing a meta‐analysis. Effectiveness on function for the classification of Integrated Geriatric Care was evident in three reviews,52, 53, 57 but two also found inconsistency55 and no effect,52 dependent on the population, the model's characteristics, and the outcome measured (see Table 3). Nineteen reviews provided consistent narrative findings related to function (see Table 3). The impact on dignified end‐of‐life care was reported by 39 reviews. One pooled quantitative satisfaction ratings and found an effect for caregiver but not patient satisfaction.62 Of the remaining 38 reviews that narratively synthesized the impact on dignified EoLC, 24 found an effect; 10 were inconsistent; and 4 reported no effect.

Mortality was assessed in 14 reviews, 5 of which performed meta‐analysis, and all but one found no effect while one found reduced mortality in a subgroup of “younger” older people. Nine other reviews reported on survival narratively. Most reported no effect or inconsistent findings, with only one reporting that the model of care reduced mortality. In no cases were the models found to be associated with increased mortality.

Scalability and Sustainability

Forty‐four reviews (44 of 72, 61%) provided information to assess the potential scalability and sustainability of the service delivery models. However, only four reviews included studies from LMICs incorporating Zambia,41 South Africa,39, 41 Uganda,39 Kenya,40 India,41, 42 Pakistan,42 and Nepal.42 All reported Integrated Palliative Care, with only two classified also as Integrated Geriatric Care.40, 42 This limits conclusions about the provision of geriatric service delivery models and how they could be provided. Scalability and sustainability require an understanding of the population, services, and resources available to align models of care with these contextual details.45 Important contextual factors for scalability and sustainability were the social and cultural characteristics of the populations studied, the health structural systems, and the definitions of palliative care.14, 36, 65, 70, 77, 80, 84, 90, 95

The reviews offered important insights into key components to support the scalability and sustainability of service delivery models. A public health approach in India became a promising model for increasing access to Integrated Palliative Care through formal health systems, family and community support, and public engagement.42 Three reviews focused on sustainability in rural areas and cited the pivotal role of community services to develop palliative care.41, 46, 81 A close working relationship between palliative care and other primary care providers was associated with improved quality of life and more dignified EoLC.41, 46 The reviews identified approaches to widening the provision of palliative and geriatric care, such as education and training for primary care clinicians and other generalists in the community,41, 46 and the role of family physicians to develop and implement strategic health programs for aging communities.81

Centrality of Person‐Centered Care

Care that is driven by the person's priorities and goals requires impeccable assessment to encompass the breadth of needs, an understanding of the goals of “what matters to you” (the person), and involving the person in decision making about care and treatment.112 Such care enables a pursuit of quality of life by attuning care provision and its goals to the person's own needs and priorities. The use of person‐centered outcome measures in routine care is advocated to assess systematically individual needs, priorities, and goals; to direct the provision of care; and to measure the outcomes of care from the perspectives of the patients and their families, thereby promoting autonomous decision making, quality, and equity.114 The likely impact on clinical care, however, requires the use of measures developed and validated for the respective population and setting and the staff's engagement.115, 116 The Integrated Palliative care Outcome Scale for Dementia (IPOS‐Dem) is an example of a person‐centered measure for palliative geriatric care with clinical validation for use in care homes for people with dementia and multimorbidities.117, 118 The measure is an outgrowth of the established and validated Palliative care Outcome Scale (POS) family of measures,119, 120 widely used internationally, including for LMICs in the African and Thai POS, for example.121, 122

To live life as well as possible in old age and enable healthy dying require the integration and coordination of care between services across the care continuum.123 A model of comprehensive palliative care integrated with other specialties like geriatric care is advocated for meeting the multidimensional needs of individuals (and their families) living with multiple long‐term conditions.112 With advancing age, a person's needs often increase in both prevalence and severity associated with living with several morbidities and frailty. Consequently, the attainment of these goals often requires more resources and greater effort, by both the patient and his or her caregivers. In turn, this necessitates balancing the management of symptoms and concerns and preserving autonomy, especially by maintaining the person's physical function.112

Our analysis conceptualizes Integrated Geriatric Care as person‐centered care with an emphasis on strengthening and maintaining the person's intrinsic capacity and functional ability, and/or reversing the causes of acute decline.123 Attention to both maintaining function and managing acute decline is important. Acute decline is common when living with frailty, which is associated with diminishing physiological reserves and increasing risk to a seemingly minor health event like an infection leading to poor outcomes, such as end of life.3 However, emphasis on an integrated approach must encompass management of symptoms and concerns and not neglect care of the dying as part of the continuum from healthy aging to healthy dying. The dying phase often is the time of greatest need for the person and his or her family.112 Integrated Palliative Care encompasses both specialist and generalist palliative care. Specialist palliative care includes hospice care (with inpatient hospice, day hospice, hospice at home) as well as a range of other specialist advice, support, and care for patients with complex problems.124 Generalist palliative care is provided by professionals with basic training in palliative care supporting patients with life‐threatening conditions.15

Effectiveness of Models of Care

Improvement in older people's quality of life is apparent from receipt of both Integrated Geriatric Care40, 57 and Integrated Palliative Care.40 Effectiveness is also evident in the two approaches’ main focus on outcomes to improve function52, 53, 57 and to reduce the severity of symptoms like pain.59, 62 One systematic review of the Acute Care for Elders model showed effectiveness for both outcomes of function and symptom severity, as well as expenditure data on health service use.53 However, all the meta‐analysis data were derived from high‐income countries with no representation of LMICs. Data for LMICs were limited to narrative data from two systematic reviews supporting the effectiveness of Integrated Geriatric Care42 and Integrated Palliative Care41, 42 on quality of life. The effectiveness and sustainability of a service delivery model are critically influenced by how they are implemented in each context, yet these areas are often poorly considered in research evaluating models of care.125

The findings reported in the systematic reviews rarely considered the relationship between the outcome of quality of life and how this may be impeded by contextual factors such as poverty. Resource use was rarely collected or reported beyond individuals and health care systems, which are often financed by taxes, social insurance, and private insurance. Informal caregiver costs and the opportunity costs of lost work and lost productivity were not reported. Care and treatment may be initiated by decision making in the clinician‐patient relationship, but its delivery is influenced by the resources available in a society, as well as the configuration and capacity of the delivery system and finance mechanisms,126 for example, the availability of national health insurance programs to provide affordable health care.20 Personal resources for older people in LMICs are often limited, with the proportion living in poverty above the national average in, for example, 11 out of 15 African countries.127 Wider policies and publicly funded safety nets are vital to enable individuals to pursue personal goals and to prevent families from sacrificing basic needs to care for loved ones at the end of life.113 Key to maintaining quality of life in old age is the extent to which older people's lives continue to contribute to their families, communities, and environment, however small.1 This requires action across sectors to enable individuals to adjust to changes in social positioning, for example, to connect to social networks, maintain function to preserve autonomy, and prevent families from plunging into bankruptcy as they care for their aging relatives.

Model Delivery

Our findings showed that palliative and geriatric care were delivered through two main overarching processes: collaborative work among settings, services, and disciplines; and case management to coordinate with and tailor care provision to individual needs and priorities. Multidisciplinary teams were common, with three or more disciplines. Multidisciplinary work is vital to maximize function and optimally manage symptoms and concerns. This requires integration across the continuum of EoLC for older people to maintain quality of life and achieve a dignified death. However, the service delivery models minimized consideration of continuity of care across settings such as between secondary and primary care, integration between specialists in palliative and geriatric care, or between specialists and generalists in palliative care.

Transitions between care settings are common for older people nearing the end of life, particularly moving from home to hospital in high‐income countries.128, 129 The degree of integration among settings, specialties, and services is a key system feature that influences both the quality of care delivered and the outcomes. Poor integration leads to fragmentation of services. In turn, this can impede the coordination of services and late delivery of services, like palliative care, as well as the duplication of services, such as diagnostic tests.126 Countries, health care systems, and institutions, therefore, must define “integration” as aligned with the local population's needs and characteristics and primary resources.15, 130 This definition needs to be guided by a common understanding of what is implied by integration for person‐centered and demand‐led health care.111, 131 Arriving at a common understanding can contribute to a sharper focus and thus direct attention to the most critical investments and policy reforms.

In an age in which longevity increasingly becomes a norm, health system reforms must focus on accelerating the expansion of health care from primarily the prevention of disease and mortality toward a more comprehensive approach that includes functional ability, quality of life, and a dignified end of life. A diversity of service provision is required to manage effectively the complex and multifaceted needs of older people in order to enable them to live life well and have a healthy death. Figure 3 details the proposed breadth of possible packages of services to deliver the continuum of care across illness trajectories, care settings, and context. These packages attempt to widen access to palliative care for people with chronic progressive conditions across care settings by combining the requirements for UHC to ensure access to affordable palliative care for all, and they realize the World Health Assembly (WHA) Palliative Care Resolution to strengthen palliative care as a comprehensive component of services throughout the life course.11

The commonality of education as a key component of service delivery models, reported in 72% of reviews (52 of 72), provides an opportunity to capitalize on the specialties’ reciprocal learning. Palliative care training and education for all staff are essential to realizing the WHA resolution on the integration of palliative care in all health services.11 The relief of suffering and the pursuit of an optimal quality of life can and should be offered to older people by all providers of care, including specialties like geriatrics and generalist providers, notably in primary and community care as the main providers of care for older people.113 However, most of the evidence underpinning the model in Figure 3 were from high‐income countries with an assumption of available services, for example, primary and community care.

Aging is a relatively new and rapidly increasing epidemiological phenomenon in LMICs, and LMIC health care services are already burdened, for example, with a high prevalence of HIV and the demand for testing and treatment. This places different stressors on health systems and resources compared with those of high‐income countries.1 Family caregivers also contend with many competing demands, such as being a subsistence earner and the main carer of orphaned grandchildren.133, 134, 135, 136 We must consider the applicability, potential effectiveness, and sustainability of system change and service models across low‐, middle‐ and high‐income settings.137 Demographic transitions will change the ways in which we stay healthy, live, and die.138 Significant gains in life expectancy mean that more and more older people will live with multiple progressive conditions,139 and for many, quality of life may start to matter more than quantity of life.1 Accordingly, health systems will be driven to move from a single‐disease framework focusing on cure and prevention toward ensuring equal emphasis on continuity of care across services and specialties and on treatments and care that maintain function, quality of life, and dignity in dying, to enable older people to live meaningful lives despite progressive diseases.111, 139

Realigning Health and Social Care Systems

In the 20th century, health care systems developed to improve maternal and child health.131 When this was achieved, more middle‐aged persons, having survived childhood, then developed noncommunicable diseases (NCDs).140 At this point, health systems altered their focus to tackle early avoidable deaths from NCDs, notably cardiac disease. Many countries, like India, were also newly developing to middle‐income countries. With advances from programs of prevention and cure, most people now die in their late 70s and 80s.138 This led to the next wave of NCDs, neurological disorders and dementias, arriving. Health care systems readjusted yet again, moving away from investment in hospitals in favor of community and primary care–based approaches.141 Such approaches are one of the principal means of ensuring the continuity and coordination of care required for multimorbid progressive conditions and for the optimal use of resources to meet the growing demand.113, 142

The integrated approach we have presented in this article is an integral facet of health care systems that all countries will require in order to increase performance and align health and social care with the needs of their aging populations.1 The linear health care system's developmental approach of attributing change to specific inputs and outputs (eg, immunization programs) may have been effective in the 20th century, but in the 21st century, a systemwide program is required. Because the global health care system faces many challenges, all health systems, whether in low‐, middle‐, or high‐income countries, must plan their health care systems, allocate their resources, develop health in all sector policies, and train their staff in a different way. It is incumbent on 21st‐century health care systems to advance with people‐centered perspectives in mind and as part of their planning strategies.111 The values that underpin people‐centered approaches to health care systems will serve to maximize quality of care. This is applicable for health care systems managing high child‐disease burdens, as well as those in countries with more equal ratios of older people to children,111 a key change as societies experience increasing longevity.138

Our findings indicate crucial components for models in LMICs, notably the development of community services, particularly in rural areas,41, 46 as well as health and education policies to integrate palliative and geriatric care across health and social care services by training primary care providers and broadening public engagement.42, 81 Volunteers play a vital role in community‐oriented programs and public engagement. They act as an indicator of community involvement. Their presence can enable members of the community to engage with palliative and geriatric care services, particularly in communities in which access to services is limited by low levels of health literacy and likely a concurrent lack of understanding of older populations’ needs, especially in late life.112, 143 Nonetheless, only five systematic reviews reported service delivery models that involved volunteers, and only two included LMICs.41, 42

The findings for a model of integrated palliative care align with the Lancet Commission's essential package of palliative care to deliver community‐oriented programs for LMICs.113 The package is designed to be sustainable by locating changes to health care systems within the wider policy context to address, for example, poverty and by incorporating the lowest cost, such as for medication, and emphasizing practitioners’ competencies rather than their disciplines.113 Emphasis is placed on building competencies and including the principles of palliative care in other specialized (eg, geriatrics) and generalists’ services, notably family physicians and community health workers, by providing the necessary training and medical supervision to enable the integration of palliative care across services. There is evidence of the effectiveness of this integrated model in, for example, training generalist nurses to deliver palliative care in HIV care and treatment in Kenya.144

It is imperative for LMICs to increase the provision of geriatric care through training of generalists in geriatric care, both in primary care and as part of community‐based services. The reviews that encompassed LMICs focused mainly on the provision of palliative care models of service delivery, with little discussion of the role of geriatric services. Geriatric services are a relatively recent priority in LMICs. For example, Nigeria opened its first dedicated service in 2012, but provision remains sparse to meet the needs of a burgeoning older population of approximately 17 million people.20 To achieve the current and projected aging in LMICs and to meet the UHC145 and the WHA resolution,11 the WHO's public health palliative care strategy (and the putative model of integrated geriatric and palliative care proposed) may need to include the African model of policy and education and the Indian model of community integration. However, this strategy can be achieved only through local research146 that takes account of local systems, cultural preferences, and resources.

Methodological Reflections

Aging, quality of life, and end of life are broad concepts, and we defined and operationalized each in this review. We used the WHO's definitions and international consensus statements to guide our choices and then worked with information specialists to limit the extent to which our choices narrowed our scoping of the evidence. We limited our search strategy from 2000 to assess current service delivery models and the evidence base for them. By drawing on systematic reviews, we determined that the date limitation would not prohibit the inclusion of earlier studies. However, in keeping with our contemporary focus, we found that the majority (49 of 72, 68%) of our included reviews limited their search strategy to 2000 and beyond.

Although our electronic search included publications in all languages, it was limited to three databases that were primarily indexed on English‐language publications. Relying exclusively on database searching is unreliable when trying to identify literature regarding complex interventions,147 so our extended search strategy included scanning reference lists, using personal knowledge, and making external contacts. Specifically, we called on experts and active researchers across Africa and Latin America to share gray literature and to scope local databases on our behalf. We also drew on resources to translate non‐English language reviews as required, for example, Ruiz Ingez and colleagues published in Spanish.93 The relative gap in knowledge reported from LMICs may reflect the perspective taken by our review teams, as well as the limited dissemination of relevant evaluation studies impeded by the resources and expertise required to undertake effectiveness evaluations. Our main findings on the overarching processes to enhance outcomes of care, however, were corroborated by an international review of integrated health services promoting a people‐centered approach.111

Our choice to scope systematic reviews allowed us to incorporate multiple search strategies, including our own search strategies and those from the 72 systematic reviews we included. We are therefore confident that this represents a comprehensive review of the published literature on the effectiveness of service delivery models to maximize the quality of life for older people in their last 1 or 2 years of life. There was potential for primary studies to be included in more than one review, and we did not explicitly check for duplication among the reviews, which may have caused us to overestimate their effectiveness. Drawing on systematic reviews rather than primary studies also limited our ability to describe the granularity of the model components and processes of delivery, for example, referral criteria or triggers, and the relationships between outcomes such as physical and mental suffering. Often the reviews reported on different service delivery models grouped by their setting or overarching aim, thus requiring us to interpret model components from the narrative descriptions in the text. The case examples we have provided do offer more information and help service providers understand how care might be integrated. In addition, we demonstrated the potential effectiveness of our proposed classification on quality of life and the key processes and components. We could not, however, confidently compare model components or directly link different models to the reported outcomes. These are important future areas of research to inform policy and practice and to examine our proposed conceptual model of integrated palliative care and integrated geriatric care.