The Organization of Home Palliative Cancer Care by Primary Health Care: A Systematic Review

Highlights

Public health relevance—How does this work relate to a public health issue?

• Addresses the growing global demand for palliative cancer care amid aging populations and rising chronic disease burden.
• Examines how Primary Health Care can reorganize palliative cancer care to reduce hospital dependence and improve equitable access.

Public health significance—Why is this work of significance to public health?

• Identifies organizational components associated with the delivery of home-based palliative care and selected health-system outcomes.
• Highlights gaps in current care models and the need for system-level integration between generalist and specialized palliative care.

Public health implications—What are the key implications or messages for practitioners, policy makers and/or researchers in public health?

• Reinforces the role of Primary Health Care as the coordinating hub for palliative cancer care, guiding future policy and service planning.
• Calls for investment in training, integrated pathways, and sustainable community-based services, especially in low- and middle-income countries.

Abstract

Background: Home-based palliative cancer care coordinated by primary care has been proposed as a strategy to improve access, continuity, and alignment with patient preferences, but its organizational characteristics and effects remain insufficiently synthesized. Aim: The aim was to characterize the organizational components of home-based palliative care models for patients with advanced cancer coordinated by primary care and to evaluate the effects of these models on patient-related and health system outcomes. Methods: Following a Cochrane review protocol, primary studies were searched in MEDLINE, Scopus, LILACS, Web of Science, the Cochrane Library, and CINAHL to identify randomized controlled trials and analytical observational studies published between January 2018 and May 2026. Results: The narrative synthesis included seven analytical observational studies conducted in high-income countries, most classified as 4b—settings with advanced and integrated palliative care. Across these studies, recurrent organizational components included case management, structured care pathways prioritizing home-based care, and mechanisms linking primary care with specialized palliative care teams. Reported outcomes were associated with potential system-level benefits, such as fewer emergency department visits and higher proportions of home death, but these findings are subject to residual confounding, heterogeneity, and indirectness, which limit confidence in the evidence. Conclusion: The available evidence, which was restricted to seven observational studies conducted in countries with highly developed palliative care systems, suggests that certain organizational components may support home-based palliative cancer care coordinated by primary care. However, substantial methodological limitations, the absence of patient-centred outcomes, and the lack of studies from low- and middle-income countries indicate that further research is needed before drawing practice or policy recommendations.

1. Introduction

People with advanced cancer, defined as an incurable, progressive malignant disease no longer responsive to curative-intent treatment [1], may be subjected to a high burden of treatments that increase suffering and compromise the quality of life. Consequently, there is a need for complex care, which intensifies in frequency throughout the course of the disease. These needs are met mostly in the hospital, with uncertain benefits, and disconnected from primary health care (PHC), especially due to deficiencies in palliative care (PC) in both contexts [2,3,4].

The concentration of care at the tertiary level compromises access to health services in resource-constrained settings, especially in low- and middle-income countries (LMICs), which face high prevalence of cancer and other chronic diseases, alongside accelerated population aging [5]. In addition to inequitable access, the quality and safety of care may be compromised when end-of-life (EOL) care is delivered in hospitals without specialized PC teams or structured dehospitalization strategies [4]. To ensure conceptual clarity, this review draws on published models of home-based PC to define it as PC delivered in the patient’s residence by health professionals, supported by structured mechanisms for symptom management, care planning, and coordination across the health system [6,7].

In most high-income countries, PHC is mainly responsible for managing chronic conditions and frailty in older adults, with PC increasingly recommended based on needs rather than prognosis. However, despite global advocacy for community-based PC as a strategy to expand access, PHC has not been able to act in an integrated and collaborative way to respond to the needs of people with advanced cancer and other life-threatening conditions, particularly in LMICs [5,8]. In this review, “coordinated by PHC” refers to organizational models in which PHC teams hold a defined coordinating role, such as case management, shared care pathways, or structured communication, linking home-based care with specialized PC services and other sectors of the health system [4,7,9].

Even in some high-income countries, PC remains associated primarily with EOL care, and end-of-life care is still predominantly hospital-centred; however, longer duration of PC before death is associated with higher quality of life and lower hospital costs [4,9,10]. To increase the duration of PC, integration across the healthcare network is essential to ensure respect for patient autonomy and equitable access according to patient and family preferences [11]. Many patients prefer to remain at home [12]; to achieve this, PHC professionals are increasingly expected to incorporate palliative approaches into practice and be integrated with specialized PC services [6,13].

Earlier evidence, including a 2015 systematic review highlighting persistent delays in initiating PC, indicated that the time from PC initiation to death remained far below recommended durations, and more recent studies continue to show the same pattern [14]. Current evidence suggests that at least three to four months of PC are needed to achieve meaningful clinical benefits; however, in practice, this duration is consistently shorter. A systematic review with meta-analysis found that PC duration was 15 days for oncological conditions and six days for non-oncological conditions, with more favorable durations in community and home care settings compared with general hospitals (20 days versus six days, respectively) [10].

Recent evidence syntheses have examined PC in LMICs, the roles of PHC professionals in PC delivery, and early intersectoral integration between oncology and general practice. These include a 2024 systematic review of PC in LMICs [5], a 2022 scoping review of PC within the primary care setting in Australia [15], two integrative reviews published in 2022 [16,17], a 2023 scoping review of early integrated PC between oncology and primary care [18], and a 2023 scoping review on EOL care education for general practitioners [19]. However, none of these reviews examined structured home-based organizational models coordinated by PHC for patients with advanced cancer.

Although the literature includes several studies describing components of home-based PC or reporting outcomes associated with such services, these studies are highly heterogeneous. This heterogeneity highlights the need for an integrated understanding of how organizational models within PHC are structured and how their specific organizational features relate to outcomes at both the patient and system levels. Therefore, this systematic review aimed to characterize the organizational components of home-based PC models for patients with advanced cancer coordinated by PHC and evaluate the effects of these models on patient-related and health system outcomes.

2. Methods

This systematic review was structured to address two complementary research questions (RQ), both related to the same phenomenon: (RQ1) Which organizational components characterize home-based PC models for patients with advanced cancer coordinated by PHC? (RQ2) What are the effects of these models on patient-related and health system outcomes?

The review protocol followed the Cochrane Handbook for Systematic Reviews of Interventions [20], was registered in the PROSPERO database (CRD42023392997), and was published prior to data extraction. The protocol presented the research question in a single, broad formulation, encompassing both organizational components and outcomes [21]. Although expressed as one question, the protocol already established a dual analytical focus by aiming to examine both the organizational aspects of home-based palliative cancer care and the effects of these models on patient- and system-level outcomes.

During the development of the full review, this original scope was retained, but the review was reported as two explicit and complementary questions. This adjustment was made purely as a post hoc clarification to ensure greater conceptual clarity and alignment between the protocol’s intended focus and the structure of the available evidence, rather than a structural modification of the protocol itself. The resulting report was guided by the Preferred Reporting Items for Systematic Reviews (PRISMA) [22].

All complete search strategies, including exact search strings, platforms, limits, and numbers of records retrieved, are fully reported in the published protocol [21] and available in Supplementary Materials. In this review, the search strategy was replicated exactly as published, and the only methodological update was the extension of the search period to include studies published from 2023 onward. This update did not modify the search structure; the original search strings were rerun identically, with the temporal filter serving solely to capture newly published studies.

2.1. Information Sources and Search Strategy

The academic databases used to retrieve the studies were MEDLINE^® (through PubMed^®), Scopus (through the Elsevier interface), LILACS, Web of Science, the Cochrane Library, and CINAHL (through EBSCOhost). Studies published from January 2018 to May 2026 were included using the definition of palliative care of the International Association for Hospice and Palliative Care in 2018 [23]. Because this review focuses specifically on organizational models of home-based palliative cancer care coordinated by PHC, restricting the search to the post-2018 period ensured conceptual consistency with contemporary understandings of PC delivery, integration, and coordination. Bibliographies of primary studies were searched manually to identify further eligible studies.

2.2. Eligibility Criteria

Studies were included based on participants, interventions, context, outcomes, and type of study. The eligibility criteria were structured to operationalize the two research questions: RQ1 through the intervention and context criteria, which required PHC-coordinated home-based PC models for patients with advanced cancer; and RQ2 through the outcome criteria, which captured both patient-related and health-system effects of these models.

2.2.1. Participants

The studies included adult participants, not limited by gender, ethnicity, or type of cancer. Studies could include additional participants (e.g., caregivers, volunteers, healthcare professionals), provided that adult patients with cancer receiving home-based PC were included. Pediatric studies and those that in any way did not address oncology were excluded. Studies that combined oncological and non-oncological diseases were included.

2.2.2. Interventions

Studies examining organizational models of home-based PC for adults with advanced cancer coordinated by PHC were included. These models could involve any combination of organizational components, such as care coordination mechanisms, case management, multidisciplinary teams, integration between generalist and specialized PC, structured care pathways, shared records, or tools to support home-based care, if they involved PHC in a defined role.

2.2.3. Context

Studies that had the patients’ home as the care environment and monitored by healthcare teams due to advanced cancer in PC, not receiving curative-intent treatments, were included. Countries were included regardless of the level of economic development, from countries with no known PC activity to those with advanced and integrated services, according to the classification of the Global Atlas of Palliative Care [24].

2.2.4. Outcomes

Primary outcomes included patient-related and health system outcomes associated with home-based PC models, such as symptom control, quality of life, relief of suffering, time spent at home, use of emergency or hospital services, and place of death.

Secondary outcomes included measures related to the organization and functioning of the care model, such as the level of integration between PHC and specialized PC, coordination processes, community support structures, and economic or policy considerations.

2.2.5. Types of Studies

Experimental studies (e.g., randomized controlled trials—RCTs) and analytical observational studies with a control group (e.g., cohort and case–control studies) were eligible for inclusion. Studies of any design were excluded if they were considered to be at critical risk of bias in any domain of the risk-of-bias tool applicable to their design. This criterion was defined a priori and applied consistently to all eligible studies, including RCTs.

2.3. Data Extraction (Selection and Coding)

EndNote^® reference manager was used; subsequently, data were recorded in Microsoft Office EXCEL^® 2019 spreadsheets. First, all the titles and abstracts of the identified articles were analyzed concerning the established inclusion criteria. An abstract-level eligibility review was conducted by three team members (M.M.S., T.A.B.S.T., and M.C.A.C.); each abstract was independently screened by two reviewers (M.M.S. and T.A.B.S.T.). The same procedure was implemented at the full-text level. Data were extracted independently by two team members (M.M.S. and T.A.B.S.T.), and the extracted data were verified by a third reviewer (C.L.B.). Disagreements at all stages were resolved by consensus.

The following variables were extracted: study characteristics (duration, location and environment, type of study, population, and sample), organizational components of home-based PC models, the role of PHC within these models, and patient-related and health system outcomes associated with the models. Reported bias or limitations were also extracted. To assess the risk of bias, the Cochrane risk-of-bias tool for randomized trials (RoB 2, 2019) and the Cochrane risk-of-bias tool for non-randomized studies of interventions (ROBINS-I V2, 2024) were applied. In accordance with the prespecified eligibility criteria, following the Cochrane guidance, studies judged to have a critical risk of bias in any domain of these tools were excluded from the synthesis.

2.4. Strategy for Data Synthesis

The following data were synthesized: organizational components of the home-based PC models, mechanisms of coordination and integration with PHC, time spent at home, use of emergency services, death at home, number of home consultations by the professional category, applied resources (material and human), means of communication and integration of the service with the health network, and cost-related outcomes. Patient-centred outcomes such as quality of life, relief of suffering, and symptom control were pre-specified in the protocol but were rarely or not reported in the included studies; this absence was incorporated into the synthesis as an indicator of gaps in the literature.

The results were synthesized narratively according to the following characteristics:

• Organizational components of the model, such as care coordination mechanisms, case management, multidisciplinary teams, shared records, and structured pathways.
• Characteristics relevant to interpreting service organization, such as country income classification (low-/middle-/high-income) and level of PC development [24].
• Patient-related and health system outcomes, such as emergency department (ED) visits, hospitalizations, time spent at home, place of death, and costs.
• Model content, including consultations, health professionals involved, communication strategies, and duration and intensity of case management in home care.

3. Results

3.1. Study Selection

The search yielded a total of 17,181 articles. Duplicates were removed first using EndNote and then through manual verification, leaving 16,302 records for title and abstract screening. At this stage, most records were excluded because they did not address the research question or were unrelated to PC or home-based services. Of the 86 reports sought for retrieval, seven were initially inaccessible through institutional resources. The corresponding authors were contacted by e-mail; one provided the full text, whereas the remaining six did not respond and were classified as reports not retrieved. A total of 80 full-text reports were assessed for eligibility, of which seven met all inclusion criteria and were included in the final analysis. The flow of study selection is presented in Figure 1 and in Supplementary Materials Table S1.

3.2. Study Characteristics

Regarding contextual characteristics, one study was conducted in a country classified as level 3b (Finland), and the remaining six studies (86%) were conducted in countries classified as level 4b, with advanced and integrated PC [24]. All countries had a very high Human Development Index (HDI). We examined all included studies for potential overlap in populations or datasets and found no evidence of duplicates. Although two studies were conducted in Canada, they originated from different provinces and did not share participants, recruitment periods, or datasets (

Table 1. General characteristics of the seven selected studies.

Title, Year	Country, HDI and Level of PC	Objective	Type of Study and Risk of Bias	Sample Size/Participants	Purpose of the Intervention/Comparator	PHC Role	Main Outcomes/Effects Estimate (95% CI)
Community Palliative Care Initiatives to Reduce End-of-Life Hospital Utilization and In-Hospital Deaths: A Population-Based Observational Study Evaluating Two Home Care Interventions, 2019 [25]	Ontario, Canada, HDI very high Level 4b	To evaluate the impact of interventions aimed at planning for a home death (Yellow Folder) and managing symptoms in the home (Symptom Response Kit) on place of death and hospital utilization among palliative home care patients	Ecological and retrospective cohort study ROBINS-I V2-Moderate	Final cohort 5900 patients; 4538 decedents	Yellow Folder to facilitate EOL planning at home and Symptom Response Kit; the comparator was the absence of both interventions	Program-led with PHC integration. Home PC organized by care coordinators; later supported by a Supportive Care Team with palliative training	Death in the community: - Yellow Folder vs. neither intervention—RR 1.66; CI not reported. - Symptom response vs. neither intervention—RR1.79 (1.55 to 2.05) Both vs. neither intervention—RR 2.20 (2.05 to 2.36)
Palliative home care and emergency department visits in the last 30 and 90 days of life: a retrospective cohort study of patients with cancer, 2021 [26]	Calgary, Canada, HDI very high Level 4b	To evaluate the association of specialist palliative home care on emergency department visits in the 30 and 90 days prior to death	Retrospective cohort study ROBINS-I V2-Moderate	6976 adults who died of cancer; 51% were men, 49% were women	Home PC vs. generalist home care vs. no home care. ED use measured in the last 30 and 90 days of life; prior ED use assessed over 90–365 days	Home care (HC) delivery included palliative HC and generalist HC; timing of first palliative contact captured PC access	ED visit in last 30 days: - Generalist HC vs. Palliative HC—OR 1.19 (1.06 to 1.34) - No HC vs. Palliative HC—OR 1.54 (1.31 to 1.82) ED visit in last 90 days: - Generalist HC vs. Palliative HC—OR 1.48 (1.32 to 1.67) - No HC vs. Palliative HC—OR 1.66 (1.39 to 1.99)
Impact of the regional palliative care pathway on emergency room visits and hospitalizations, 2021 [27]	Finland, HDI very high Level 3b	To explore the effect of palliative care pathway including home care and end-of-life care beds in the community hospitals supported by a palliative care unit in the secondary hospital on emergency room visits and hospitalizations	Retrospective cohort study ROBINS-I V2-Moderate	112 patients with advanced cancer and a palliative treatment goal visited the ED in both 2009 and 2015; median age 69 years, 46% women	Regional PC pathway including home PC teams, EOL beds in community hospitals, a general practitioner/nurse coordinator in each community, and a PC outpatient unit. Comparator: before vs. after pathway implementation in 2009 and 2015	Integrated Tiered Network; pathway explicitly linked primary and secondary care; communities provided PHC via general practitioner-led community hospitals and home care units	Revisit ED rate 59% to 38% (p < 0.01); secondary hospital admissions 56% to 45% (p = 0.016); community hospital admissions 12% to 28% (p < 0.001). No numerical effect estimate or CI was reported; the study states differences in revisits were insignificant and hospitalizations increased in one setting, but without extractable estimates.
The impact of patient identification on an integrated program of palliative care in the Basque Country, 2019 [28]	Basque Country, Spain, HDI very high Level 4b (Spain)	To evaluate the process and the economic impact of an integrated palliative care program	Comparative cross-sectional study ROBINS-I V2-Moderate	Population of 160,000; deceased patients in 2012 (1023) and 2015 (1142)	To ensure that patients with PC needs received appropriate support in the EOL. Comparator: pre-intervention (2012) vs. post-intervention (2015)	Specialist-led with PHC coordination; integrated model using primary care centers, general practitioner/nurse coordination, and home care contacts	2015 vs. 2012: higher likelihood of V66.7 identification—OR 1.14 (1.13 to 1.16¨); increased opioid prescribing—OR 1.48 (1.45 to 1.50); slight increases in hospital death—OR 1.03 (1.01 to 1.05) and hospital admissions OR 1.03 (1.01 to 1.05), and higher total costs—OR 1.04 (1.00 to 1.09)
Impact of palliative home care support on the quality and costs of care at the end of life: a population-level matched cohort study, 2019 [29]	Belgium, HDI very high Level 4b	To evaluate the impact of palliative home care support on the care quality and costs in the last 14 days of life	Retrospective cohort study ROBINS-I V2-Moderate	A total of 8837 individuals who received home PC during the final 720 to 15 days of life were matched, via propensity score, to 8837 individuals who received standard care	To advise family doctors, other healthcare professionals, informal carers, and volunteers involved in providing home PC. Exposure: palliative home care support vs. none; sensitivity analyses separated home PC, multidisciplinary team visit, and palliative nursing/physiotherapy at home	Generalist-led with Specialist Support; some quality indicators counted family physician or other primary care professional contacts in the last 14 days	The home-support group had more contacts with family physicians: mean 3.1 vs. 0.8 in the control group, with SD 6.5 vs. 1.2. Home death—OR 4.08 (3.86 to 4.31); ED use—OR 0.54 (0.51 to 0.57); Hospital admission—OR 0.45 (0.43 to 0.46); Diagnostic testing—OR 0.43 (0.41 to 0.45); Blood transfusion—OR 0.47 (0.40 to 0.54); Surgery—OR 0.19 (0.14 to 0.26)
Accessibility of home palliative care and place of death of cancer patients: data analysis of the Emilia-Romagna Regio, 2018 [30]	Bologna, Italy, HDI very high Level 4b	To evaluate the possible impact of the ANT Foundation program on health data relating to the location where cancer patients die within the Local Health Care system area (AUSL) of Bologna, compared to what happens in the other AUSLs of the Emilia-Romagna Region (ERR) where it is not present	Comparative cross-sectional study ROBINS-I V2-Moderate	1023 decedents in 2012 (control) and 1142 in 2015 (intervention).	The model was tailored to each patient’s needs, primarily for those with a short-term prognosis, with the main objective of promoting home death	Strong primary care involvement plus hospital/home-hospitalization pathways and shared electronic records.	Among patients followed by ANT, home death reached 64.6% vs. 14.6% among those not followed; hospital death was 18.7% vs. 42.3%, and hospice death 16.6% vs. 40.7%. No numerical effect estimate or CI was reported
Effects of a Population Health Community-Based Palliative Care Program on Cost and Utilization, 2019 [31]	United States of America, HDI very high Level 4b	To evaluate utilization and cost outcomes of a standardized, population health community-based palliative care program provided by nurses and social workers	Retrospective propensity-adjusted study ROBINS-I V2-Moderate	804 high-risk Medicare Advantage members identified; 176 in the PC group and 570 in standard-care controls for evaluation	Nurse and social worker community-based PC vs. standard telephonic health-plan case management.	Specialist-led with PHC Coordination. Community-based program delivered by nurses and social workers; patients were selected by a proprietary predictive model	Total medical cost, pharmacy cost, hospital admissions, hospital days, ED visits. 20% reduction in total medical costs ($619 per enrolled member per month), 33% reduction in hospital admissions, 38% reduction in ICU admissions, and 12% reduction in hospital days. No numerical effect estimate or CI was reported

Note: PHC involvement was classified based on the reported role of PHC professionals in the organizational model. “Generalist-led” indicates models primarily delivered by PHC professionals with generalist palliative care training. “Specialist-led” indicates models primarily delivered by specialized PC teams. “With PHC coordination” or “with specialist support” indicates integration between PHC and specialized PC services.

).

When broken down across studies, depending on their objectives, the inclusion criteria were patients (oncological or non-oncological) who died in a certain region and in a certain time frame [28], or correlated to the place of death, whether in the hospital, hospice, long-stay institution or home [30]. Other studies defined the sample based on records of use of at least one type of PC service at home [25] between the last 720 and 15 days of life [29], patients who visited ED [27], patients with incurable cancer for a period greater than or equal to 180 days after diagnosis [26], and who could benefit from the PC [31].

Among the studies that included participants with other chronic conditions, cancer accounted for more than half of the sample in both the intervention and control groups [25,28,29]; only one study reported a smaller proportion of cancer patients (23.8%) compared with other chronic diseases [31]. None of the studies examined the family or community profile.

The timeframe for data collection across the seven studies spanned from 2008 to 2016, with data obtained from varied sources, including electronic medical records, administrative databases, and health plan claims data.

3.3. Types of Study Interventions

A 2019 Canadian study [25] implemented two organizational components—the ‘yellow folder’ and the ‘symptom response kit’—designed for patients with a Palliative Performance Scale score of 50% or less. The first component, introduced in April 2010, consisted of a folder containing materials intended to support patients and families. The second component, introduced in February 2012, provided medications and supplies tailored to the patient’s needs, enabling nurses to manage predictable disease-related emergencies during home visits.

Another Canadian study conducted in 2021 [26] evaluated two organizational models of home-based PC (specialized and generalist PC) delivered in the homes of cancer patients. In this service model, specialized and generalist services operated separately and could not be accessed simultaneously. Both offered personal support for daily care (such as dressing and bathing) as well as professional assistance (including home nursing visits for symptom management). These services were available 24/7 and aimed to promote quality of life, control symptoms, address emotional and psychological needs, and prevent trips to the ED.

Within the specialized PC model, a case manager—typically a nurse—provided oversight, supported by advanced expertise and a smaller caseload to ensure quality of care. For patients enrolled in either model, the primary physician was the oncologist or the family doctor, who could consult a separate PC team of specialist physicians and nurses whenever needed. Assignment to either the generalist or specialized team was based on the complexity of each case [26].

In the study carried out in Finland [27], the tiered care model integrated hospital-based and community-based services. At the PHC level, the structure included a community hospital with wards run by general practitioners and dedicated home care units. Regarding the organizational model, in 2011 a PC pathway was established in the hospital district, incorporating home PC teams and EOL beds in community hospitals. Each community appointed a general practitioner and a nurse responsible for coordinating and arranging PC at home.

Patients with incurable cancer had access to a team providing PC at home, if desired, and a guaranteed place in the community hospital ward when EOL care was indicated. They could be admitted either through the ED or directly from home, according to patient preference and at any time. In addition, an outpatient PC unit was established in the secondary hospital, staffed by a part-time physician with specialized expertise in PC and a nurse. A standardized document was developed to guide referrals along the PC pathway and support the work of the consultative PC unit [27].

In the Basque Country, the initiative involved implementing an integrated PC model based on PHC physicians identifying patients with International Classification of Diseases—ICD-9 codes indicating suitability for PC. The organizational model also included joint care pathways with standardized criteria across primary and tertiary care, incorporating home hospitalization and clinic-based services, and training for PHC professionals [28].

In this model, the shared electronic medical record and electronic prescription system served as key communication facilitators. Educational activities included general workshops on PC, symptom control, the use of opioids and subcutaneous access, palliative sedation, and EOL communication for physicians and nurses. A PC Committee was established for coordinating, evaluating, and overseeing the program [28].

The study conducted in Belgium evaluated the impact of earlier initiation of home-based PC as an organizational model. In this retrospective cohort, patients who had used any of three types of PC services in 2012 were included in the intervention group: a multidisciplinary home PC team comprising at least one general practitioner, two nurses, and an administrative assistant; palliative nursing or physical therapy at home; and a fixed subsidy of €647.16 for palliative home patients (in 2012). The main aim of the multidisciplinary teams was to advise family doctors, other healthcare professionals, informal carers, and volunteers involved in providing home PC [29].

The study conducted in Italy [30] evaluated the impact of a comprehensive home-based care model for cancer patients, providing 24/7 care delivered by physicians, nurses, and psychologists specialized in PC. The organizational model operated within a private, non-profit initiative recognized and integrated into the public health system. Activation was initiated at the request of the patient and/or family, with approval from the family doctor.

The study intervention in the United States [31] consisted of a nursing and social work organizational model of community PC to identify patients at risk of an overly medicalized death, as defined by information in the medical record at the time of death: chemotherapy within 28 days before death, unplanned hospitalization within 60 days before death, more than one ED visit within 30 days before death, intensive care unit admission within 30 days before death, or use of life support treatment within 30 days before death.

Nurses and social workers were responsible for home consultations and telephone calls. A practical nurse and a physician participated in weekly case-management meetings and visited patients and families with more complex care needs. The team received training, including training in coordinating care with the case manager and PHC professionals. Patients classified as low intensity (symptom controlled, treatment adherence, low caregiver stress, high social support, and no ED visits in the last 30 days) received safety assessments, medication education, disease management, and goal discussions through a monthly call or visit. Mid-level patients were contacted and visited twice a month to assess the need for additional community support services, manage symptoms, or assist with advance care planning [31].

PHC involvement was greatest in the care of patients requiring high-intensity intervention. Because these patients had more complex needs, they were advised to contact their family doctor or other professionals to address urgent issues after program hours and on weekends. Within the organizational model, high-intensity patients received weekly or more frequent calls or visits for crisis intervention, medication management, illness self-management education, caregiver and social support services, or evaluation for hospice care. To facilitate care integration and case management, the PC program team directly contacted the family doctor to communicate any concerns about the case. Nurses and social workers were responsible for securing social support resources, such as transportation, adult day care, food assistance, and other support groups [31].

3.4. Main Results of the Primary Outcomes

The primary outcomes were related to patient-related and health-system outcomes of the organizational models, including the number of ED visits and changes in service utilization, including the use of specialized home-based PC [26,27,28,29]. Another outcome concerned the place of death, along with time spent at home before death [25,30]. Other studies reported cost-related outcomes [29,31].

In more detail, the primary outcome of one of the Canadian studies [25] was the place of death. After adjusting for confounding factors, the results suggested a higher probability of dying at home among those who received the ‘yellow folder’ (relative risk—RR 1.66), the ‘symptom response kit’ (RR 1.79), or both organizational components (RR 2.20), compared with those who received no component.

In the second Canadian study [26], the primary outcome examined the association between the provision of home-based PC (specialized PC, generalist PC, or no PC) and ED visits in the 30 and 90 days before death. Both early referral to PC services (more than 90 days before death) and specialized home-based PC were independently associated with lower ED visit rates. Patients who received home care from a generalist PC team 90 days before death, or who did not receive home-based PC, had a higher RR of ED visits, with RR values of 1.48 and 1.66, respectively, compared with those who received specialized PC at home.

The cohort study conducted in Finland [27] suggested a positive impact of home-based PC model, even with modest investment. The total number of ED visits was higher in 2009 (control group, n = 277) than in 2015 (intervention group, n = 187). Sixty-six patients (59%) in 2009 and 42 (38%) in 2015 revisited the ED. These findings supported the hypothesis that patients receiving home-based PC were less likely to visit ED or be admitted to the hospital than those receiving standard care.

In the Basque Country, the integrated PC model focused on patient identification was associated with an increase in patient identification with the ICD for PC in 2015 (intervention group); more patients were identified with the ICD for PC. Identification increased from 7% in 2012 (control group) to 16% in 2015 among non-oncology patients, and from 19% to 33% among cancer patients in the last three months of life. This earlier identification was associated with greater use of community-based PC services. The number of contacts with nurses and family doctors (in person, at home, or by telephone) increased by 35% (from 4.56 to 6.14 contacts) and 24% (from 6.71 to 8.30 contacts), respectively. The largest increase (68%) was observed in the use of home hospitalization, which rose from 0.19 to 0.32 episodes per patient. In contrast, hospital admissions increased by only 13% (from 0.54 to 0.61 per patient). Overall, the increased use of home hospitalization was the main factor contributing to a 26% rise in per-patient costs [28].

Regarding costs, the study conducted in the United States [31] found that patients enrolled in the organizational model had, on average, monthly per-patient costs that were US$619 (20%) lower than those of controls, driven mainly by reductions in physician-related costs. Pharmacy costs increased slightly by US$79 per patient per month. The community-based PC model, staffed by nurses and social workers, was associated with lower overall costs, a 38% reduction in intensive care admissions, a 33% reduction in hospitalizations, and a 12% reduction in hospital days.

In Belgium [29], 56% of individuals who received home-based PC died at home, compared with 13.8% of those who did not (RR of 4.08). On average, patients in the home-based PC cohort had nine contacts with primary caregiver and three contacts with family doctors in the last two weeks of life, compared with two primary caregiver contacts and fewer than one family doctor contact in the unexposed cohort. Total hospitalization costs were lower among those who received home PC, while outpatient costs were higher. However, overall costs remained lower in the exposed cohort, with an average reduction of €1617 per person. More than four times as many patients receiving home PC died at home compared with those without home PC. Additionally, fewer individuals in the exposed cohort were admitted to hospitals, the ED, or an intensive care unit, and fewer underwent diagnostic tests, blood transfusions, or surgery in the last two weeks of life.

Regarding death at home, the study carried out in Italy showed that, under the home care model, cancer patients followed by the Foundation appeared to be half as likely to die in hospital as patients in other regions, and, conversely, they were twice as likely to die at home [30].

Among the seven included studies, none reported patient-centred outcomes; one study assessed quality of care rather than quality of life [29]. It is important to note, however, that this absence reflects the characteristics of the included evidence rather than the entirety of the available literature. Some randomized trials, such as studies [7,32], which were excluded due to critical risk of bias, have reported patient-centred outcomes, including improvements in physical, psychological, emotional, and social domains, with better pain and gastrointestinal symptom relief and higher satisfaction [7]. Therefore, the lack of patient-centred outcomes in this review should be interpreted as a limitation of the evidence that met the methodological criteria for inclusion, rather than a complete absence of such outcomes in the broader literature.

3.5. Main Results of Secondary Outcomes

In the study conducted in Finland, secondary outcomes included the primary reasons for ED visits, such as pain (20%), decreased general health (19%), and dyspnea (14%) [27]. Another secondary outcome was the proportion of patients requiring admission to a secondary hospital, which decreased from 56% to 45% between the study years, while admission to community hospitals increased from 12% to 28% [27]. In the Basque Country, secondary outcomes examined the association between sociodemographic and clinical variables and the likelihood of identifying patients with ICD codes indicating PC needs [28].

3.6. Organizational Components of Home-Based Palliative Care Models Coordinated by PHC

Across the seven included studies, the organizational components of home-based PC models coordinated by PHC showed substantial convergence.

Table 2. Matrix of Organizational Components Across Included Studies.

Organizational Component	25	26	27	28	29	30	31
Case management	X	✓	X	X	X	X	✓
Multidisciplinary teams	X	X	X	X	✓	X	X
24/7 availability	X	✓	X	X	X	✓	X
Anticipatory care planning	✓	X	✓	X	X	X	✓
Shared records/electronic records	X	X	X	✓	X	✓	X
Anticipatory medication/emergency kits/medication education	✓	X	X	X	X	X	✓
Professional training/capacity building	✓	X	X	✓	X	✓	✓
Structured referral pathways	X	X	✓	✓	X	X	X
Specialist support integrated with PHC	X	✓	✓	✓	✓	✓	X

Note: components were coded as present when explicitly reported or reasonably inferred from the intervention description. Partial compliance was coded as present when the component was implemented in at least one care setting or for at least one subgroup of patients.

summarizes the presence of each component across studies.

The best-documented organizational components of home-based PC models coordinated by PHC were (1) specialist support integrated with PHC, (2) professional training and capacity building for PHC teams, and (3) anticipatory care planning. These components appeared most frequently across the included studies and therefore represent the core organizational features of PHC-coordinated home-based PC for patients with advanced cancer.

3.7. Risk of Bias and Reported Limitations

Among the limitations identified across studies, restricted team availability, including the absence of 24/7 home-based services [27], and limited coverage in urban areas [26] were frequently noted. Other limitations related to the instruments used to identify PC patients, which in some cases relied solely on cause of death rather than clinical criteria [28] or depended on the physician’s judgment for cohort entry [25].

In retrospective studies, limited knowledge about program goals at the time of implementation hindered the interpretation of results [28]. Studies examining place of death also faced constraints due to the absence of information on patient and caregiver preferences at the EOL [25,29,30]. Clinical guidelines followed in each country can influence patient recognition and early referral to PC; for example, in Belgium, physicians can assign ‘home PC status’ only when life expectancy is estimated at three months or less [29]. Conversely, studies restricted to cancer patients highlighted this narrow focus as a limitation [26].

Administrative data sources imposed additional constraints, particularly in cost analyses, since non-reimbursed services were not captured [25,29]. Some datasets lacked key variables, such as race or socioeconomic status, that could affect outcomes [31]. Limited follow-up capacity also affected the accuracy of outcomes such as place of death [31]. Moreover, heterogeneity in home-based PC delivery models reduced the generalizability of findings [25].

Applying ROBINS-I, all seven non-randomized studies aimed to assess attribution to the intervention, but none met all criteria for a low risk-of-bias judgment. Confounding bias was rated as moderate in all studies, followed by concerns related to outcome measurement and participant selection. In contrast, reporting selection bias was consistently low across studies.

Key confounders included distance to the ED, cancer type, symptom burden, caregiver capacity to manage crises, prior family experiences, duration of follow-up with the PC team, home-service coverage, and integration between generalist and specialized PC. Some studies attempted to control for confounding, such as the Canadian study [26], which standardized ED use based on the total number of visits in the preceding 365 days. Another relevant confounder was the limited knowledge about PC among patients, caregivers, and health care professionals, which could delay referral; to mitigate this, the Belgian study [29] considered any home PC used within the last 14 days of life.

4. Discussion

The included evidence consisted of seven analytical observational studies evaluating organizational models of home-based PC coordinated by PHC. These studies consistently reported system-level benefits associated with PHC-coordinated home care, including reduced ED use, fewer hospital admissions, increased time spent at home, and higher proportions of home deaths. Despite heterogeneity in team composition, follow-up intensity, and integration mechanisms, several organizational components emerged repeatedly across models, such as collaboration between PHC and specialist PC.

Although randomized trials exist in the broader literature [7,32,33], some of which report patient-centred outcomes, they did not meet the methodological criteria for inclusion in this review. Their exclusion does not imply that such outcomes are absent from the field; rather, it indicates that these outcomes were not represented in the evidence retained for the present synthesis.

Challenges inherent to PC research, such as attrition, death, and difficulties with blinding, may contribute to risk of bias, but they do not neutralize it. Instead, these challenges should inform the interpretation and certainty of the available evidence, which in this review is based primarily on observational studies.

Regarding the organizational characteristics of the included models, which differed substantially in team composition, follow-up intensity, integration mechanisms and health-system context, we identified several common characteristics, including strategies intended to support care at home and reduce potentially avoidable ED visits, with the implementation of strategies to maintain symptom control and avoid visits to the ED. Relevant strategies include the presence of a case manager, usually a nurse, and the integration of PC into the health care network, with the definition of pathways that, in this case, emphasize the home as a care environment in order to increase time spent at home, although this heterogeneity limits the ability to determine which specific components are most strongly associated with the reported outcomes.

In addition, given that during the oncological disease patients may require ED visits due to symptom exacerbation, the outcome after assessment and treatment of the acute episode depends on organization of the care network, since adequate territorial management can facilitate the patient’s return home [34]. The presence of a case manager emerges as a promising approach to coordinate this journey on an individual basis and to concentrate care within PHC. Furthermore, the case manager can facilitate the identification of palliative needs, especially when they are still fluctuating, providing timely referrals, including the integration of specialized PC [2,26,31].

The integration between generalist PC, those services provided by PHC professionals, mainly family doctors and nurses, supported by appropriate foundational knowledge and training, and multidisciplinary teams specialized in PC is essential for increasing the time spent at home by people with advanced cancer [4]. It also optimizes the care capacity of specialized teams which may be particularly relevant for complex situations involving difficult-to-control physical symptoms, as well as psycho-emotional, social, and spiritual distress [35]. Nonetheless, because all included studies were observational, these findings should be interpreted as consistent associations rather than evidence of a causal effect. Moreover, the substantial variation in how these organizational arrangements were operationalized across studies reduces comparability and prevents isolating the effect of individual components.

The absence of patient-centred outcomes limits the ability to assess how these organizational arrangements influence subjective experiences of care. This gap is particularly important in PC, where the quality of an organizational model cannot be assessed solely through service utilization, costs, hospital admissions, or place of death. The absence of patient- and family-reported measures limits the assessment of care experiences and other dimensions of quality that cannot be captured by health-system indicators alone. Future studies should therefore combine health-system indicators with validated patient- and caregiver-reported outcomes.

Greater PHC involvement in coordinating care depends on professional education. Lack of knowledge about PC is a frequently reported problem in the literature [35,36], yet only one of the studies in this review included education for PHC professionals [28]. One of the Canadian studies [25] highlighted the need for team training, but this was limited to the intervention team and did not extend to other professionals in primary or tertiary care.

Many organizational models appeared to be pilot programs funded specifically for research purposes, without clear indication of sustainability, an aspect that would require a more active role for PHC. In LMIC contexts, this sustainability is frequently hindered by intermittent funding and an irregular supply of medications, factors that may lead patients to incur substantial out-of-pocket expenditures for essential palliative support [37,38]. It is necessary to advance toward a model in which PHC resources are consistently available to all individuals who need and prefer to remain at home while managing chronic conditions and EOL care [2,35].

The potential for sustaining these models lies in the ability to develop integrated and collaborative care [4]. The results of this review highlighted the potential value of the case manager, the clear definition of team roles, and the establishment of predefined pathways integrating PHC and hospital care whenever necessary, while recognizing that some features, such as case management and structured care pathways, appear consistently across models, whereas others depend heavily on local infrastructure, workforce availability, and funding arrangements. It is also important to ensure the availability of beds in community hospitals, preferably to standardize PC practices at the EOL [27,28].

The integration of specialized PC within PHC appears to be a key factor associated with achieving goals such as avoiding ED visits, increasing time spent at home, and supporting home death [38]. Evidence from the Canadian study conducted in 2021 [26] reinforces this point: patients who received home care from generalist PC teams had higher odds of visiting the ED 30 and 90 days before death compared with those who received home care from specialized PC teams. These results suggest an association between specialized home-based PC and lower ED use, but do not allow inferring causality. However, the variation in how integration was structured across studies further limits the ability to attribute these associations to specific organizational components.

Although less complex patients may be considered appropriate for generalist PC, the high rate of ED visits observed in this study suggests that their clinical needs may have been more complex than initially assumed. It is also possible that the optimal timing for referral to the specialized team was missed, highlighting the potential value of ongoing assessment, particularly in cancer, where rapid progression and clinical instability can be anticipated [26].

Even in models that demonstrated greater integration between teams, such as the participation of family doctors in scheduling patients identified with ICD codes for PC [28], the availability of 24/7 services may be an important contextual factor influencing these outcomes. However, ensuring 24/7 coverage remains a challenge in PHC-based PC, as it requires substantial investment in human resources, even in high-income countries. In this context, a promising resource is the availability of telephone support, which in some models accounts for more than 20% of the actions carried out by family doctors and nurses [34].

The availability of human resources with skills in palliative cancer care, even at a basic level, remains a central challenge in LMICs. In regions such as Sub-Saharan Africa and South Asia, where physician density is often reported to be ≤ 10 per 10,000 population, specialist-centric models become less feasible. Consequently, the pragmatic implementation of generalist PC in these contexts frequently relies on task-shifting, the strategic redistribution of clinical functions to nursing cadres and community volunteers, leveraging their higher availability to extend the scope of care [38,39]. Therefore, we consider that some characteristics of the models, such as 24/7 PHC coverage, although strongly recommended, cannot be generalized across different health systems that lack such workforce infrastructure.

In view of this, it is important to emphasize, based on the results of this review, that increasing the use of community-based services does not necessarily translate into a higher number of home deaths. This outcome depends, among other factors, on the wishes and preferences of patients and their families/caregivers, as well as on bureaucratic and legal arrangements and the guarantees provided in each country for home death planning [40,41,42]. The evidence highlights the importance of creating conditions that may allow patients to spend more time at home [29], while acknowledging the challenges of integrating specialized PC into health systems in LMICs. These findings are consistent with observations from studies examining resource use and costs [28,29,31]. The heterogeneity of organizational arrangements observed across settings underscores the need to distinguish shared structural elements from strategies that are context-specific.

Although some results indicated the maintenance of hospital deaths [28], including deaths occurring in community hospitals [27], these events were not associated with reduced time spent at home. In PC, particularly in oncological diseases, greater time at home should guide the goals of home care programs that rely on stronger PHC participation in home-based strategies [34]. Early referral to specialized PC at home may be relevant when clinically indicated as both generalist and specialized PC delivered at home have been independently associated with lower ED visit rates [26]. Again, these associations should not be interpreted as causal effects, given the observational nature of the evidence.

The costs of home care were lower than those of hospital-based care, provided that home hospitalization was excluded, since this service is coordinated by the hospital [28]. According to available cost data, particularly the study from the Basque Country [28], the increased use of PHC resources resulting from the identification of more patients with palliative needs could shift demand toward hospital-based resources if community care is not available. This is especially relevant in cancer, where disease progression is predictable and resource use tends to increase progressively [41].

This inference is supported by the results of the study conducted in Belgium [29], which reported a reduction in medical costs in the exposed cohort and, above all, a higher proportion of appropriate and fewer inappropriate care practices at the EOL among those who received PC at home. Notably, total medical care costs in the last two weeks of life were lower in this group [29].

However, although the results of the study carried out in the United States also pointed to cost savings, attention must be paid to the planned exposure time and cohort follow-up period. Short follow-up durations can compromise the results if they are insufficient to monitor patients in the exposed cohort or to allow individuals initially in the control group to meet eligibility criteria over time, particularly in cancer, where disease progression is expected, and clinical deterioration may occur rapidly [31].

Advance care planning was a central component of several models and was intended to support time spent at home and, potentially, home death. The results highlight the importance of anticipating future situations and communicating effectively [27,42,43], underscoring the role of specialized PC teams at home, which, depending on availability, can also act as consultants for PHC teams providing generalist palliative care [44]. The absence of, or delays in, advance care planning can have markedly negative repercussions on time spent at home, place of death, and, consequently, on service use and costs. Such planning fosters the development of trusting relationships [33,43].

It is well established that, given the manifestations and repercussions of serious illness, especially at the EOL, psychological suffering is closely associated with the exacerbation of physical symptoms [44]. This interplay can influence patients’ and families’ preferences regarding home versus hospital care, depending on the level of support available to manage insecurities, anxieties, and fears related to what dying at home may entail.

5. Limits and Strengths

The limitations of this review include the unclear definition of usual care in unexposed cohorts, the lack of description regarding the effective involvement of PHC professionals in some PC program interventions, and whether specialized PC teams were considered part of PHC resources. Most studies were retrospective, which aligns with the challenges inherent to conducting experimental or analytical observational studies in palliative cancer care, particularly when outcomes involve death and EOL follow-up. Another limitation was the absence of patient-centered outcomes in the included studies. This reflects a limitation of the primary evidence that met the methodological criteria for inclusion, rather than a design flaw of this review.

Because all included studies were observational and conducted within specific health-system contexts, the generalizability of the proposed nine-component framework should be interpreted with caution. Observational designs allow identification of recurring organizational features, but these may reflect context-dependent arrangements shaped by local resources, policies, and service integration, which can vary across health systems.

Although the included studies provide valuable information on health system outcomes such as ED use, hospitalizations, costs, time spent at home, and place of death, they almost entirely omit patient-centred outcomes, including quality of life, symptom intensity, and suffering. Given the centrality of these outcomes in PC, this omission represents one of the most important limitations of the current health-system-level evidence on home-based PC coordinated by PHC.

Another important limitation is the restriction to patients with advanced cancer, considering that much of palliative epidemiology concerns non-oncological conditions and that palliative needs and symptom burden are comparable between patients with and without cancer [4]. The focus on advanced cancer also reinforces the tendency to allocate specialized PC based on prognosis rather than on expressed or perceived need.

Despite the movement to avoid narrative syntheses when meta-analysis is not required, we maintained this method because the review did not synthesize quantitative data; instead, the analysis of intervention effects relied on qualitative evidence related to the organization of PC within PHC. Thus, narrative synthesis was justified by the aim of coherently integrating different strategies toward a single objective, based on heterogeneous data, namely, highlighting the role of PHC in responding to an emerging demand. The presentation of different types of interventions has intersectoral implications, spanning educational, political, and social domains. However, because these interventions do not allow for pooled effect estimates and given the risk of bias, we did not assess the certainty of evidence.

The sample included only high-income countries with well-developed PC systems, underscoring the need for studies in countries with different characteristics. The main strength of this review lies in the identification of key elements of care models for delivering PC within PHC, offering insights that can guide future decisions aimed at strengthening this type of response across health systems.

Although the protocol defined the search period as 2018–2022 [21], the final review extended the search to May 2026 to incorporate more recent evidence. This update represents a post hoc clarification of the search window rather than a structural modification of the protocol; however, no additional studies met the eligibility criteria, particularly due to high risk-of-bias assessments, and therefore the expanded search window did not alter the final set of included studies or the conclusions of this review. We acknowledge that organizational models reported before 2018 may still be relevant and that this temporal restriction carries a risk of missing earlier evidence. Nonetheless, the decision was retained to preserve coherence with the current global definition of palliative care and to ensure that included studies reflect contemporary organizational frameworks.

6. Conclusions

This review suggests that organizing PC through PHC may be a promising path for LMICs, given the limited resource capacity in tertiary care, particularly for hospital admissions. Primary care professionals, including family physicians and nurses, could play a more central role in providing generalist PC and coordinating resource use according to their knowledge of the most appropriate care pathway for each patient. This includes referring patients to specialized PC whenever necessary, based on pre-established criteria aligned with the resources available in each country. It is a collective and integrated process that involves educating both health professionals and the population; for this reason, PC must be placed on the political agenda.

Specialized PC can be coordinated by primary or tertiary care, or both, and delivered in community or hospital settings. Regardless of the model, consultation by specialized professionals can help protect scarce specialized PC resources and reinforce generalist PC within PHC.

The risk-of-bias assessment showed that the selected studies present judgments that substantially reduce confidence in the results. Accordingly, the findings of this review should be interpreted as consistent associations rather than evidence of causal effects. Therefore, we recognize the need for more studies in all countries, particularly in LMICs, to better understand how these settings can respond to growing palliative care needs and realize the potential of palliative cancer care in PHC, as well as the management of non-oncological diseases in this setting.

In addition, the absence of patient-centred outcomes in the included studies—such as quality of life, relief of suffering, symptom control scores, and the grieving process—reflects a limitation of the current primary literature and limits the ability to fully assess the impact of home-based PC models coordinated by PHC. We recommend prospective and comparative studies using standardized patient- and caregiver-centred outcomes, along with clearer descriptions of organizational interventions, to generate more robust and transferable evidence and guide the development of models capable of meeting the growing demand for PC across diverse health systems.