PLOS ONE RESEARCH ARTICLE Determinants of the number of days people in the general population spent at home during end-of-life: Results from a population- based cohort analysis Albert Dalmau-Bueno1, Anna Garcı́a-Altés1,2,3, Jordi Amblàs ID4,5*, Joan Carles Contel4,5, Sebastià Santaeugènia ID4,5 1 Catalan Agency for Health Quality and Evaluation (AQuAS), Barcelona, Spain, 2 CIBER de Epidemiologı́a a1111111111 y Salud Pública (CIBERESP), Barcelona, Spain, 3 Institut d’Investigació Biomèdica (IIB Sant Pau), a1111111111 Barcelona, Spain, 4 Central Catalonia Chronicity Research Group (C3RG), Centre for Health and Social Care a1111111111 Research (CESS), University of Vic-Central University of Catalonia (UVIC-UCC), Vic, Spain, 5 Chronic Care a1111111111 Program, Department of Health, Barcelona, Spain a1111111111 * jamblas@chv.cat Abstract OPEN ACCESS Citation: Dalmau-Bueno A, Garcı́a-Altés A, Amblàs J, Contel JC, Santaeugènia S (2021) Determinants Background of the number of days people in the general The number of days spent at home in the last six months of life has been proposed as a population spent at home during end-of-life: Results from a population-based cohort analysis. comprehensive indicator of high-value patient-centered care; however, information regard- PLoS ONE 16(7): e0253483. https://doi.org/ ing the determinants of this outcome is scarce, particularly among the general population. 10.1371/journal.pone.0253483 We investigated the determinants of spending time at home within the six months preceding Editor: Rosa Maria Urbanos Garrido, University death. Complutense of Madrid, SPAIN Received: January 26, 2021 Methods Accepted: June 4, 2021 Population-based, retrospective analysis of administrative databases of the Catalan govern- Published: July 15, 2021 ment. The analysis included adult (�18 years) individuals who died in Catalonia (North-east Spain) in 2017 and met the McNamara criteria for palliative care. The primary outcome was Peer Review History: PLOS recognizes the benefits of transparency in the peer review the number of days spent at home within the last 180 days of life. Other variables included process; therefore, we enable the publication of the cause of death, demographic characteristics, and socioeconomic status, stratified as all of the content of peer review and author very low, low, mid, and high level. responses alongside final, published articles. The editorial history of this article is available here: https://doi.org/10.1371/journal.pone.0253483 Results Copyright: © 2021 Dalmau-Bueno et al. This is an The analysis included 40,137 individuals (19,510 women; 20,627 men), who spent a median open access article distributed under the terms of of 140 days (IQR 16–171) at home within the six months preceding death (women 140 [16– the Creative Commons Attribution License, which 171]; men 150 [100–171]). Female gender was an independent factor of staying fewer days permits unrestricted use, distribution, and at home (OR 0.80 [95% CI 0.77–0.82]; p<0.001). Higher socioeconomic levels were signifi- reproduction in any medium, provided the original author and source are credited. cantly associated with an increasing number of days at home in both genders: among women, ORs of the low, middle, and high levels were 1.09 (0.97–1.22), 1.54 (1.36–1.75), Data Availability Statement: Owing to the local regulatory framework on data protection, raw data and 2.52 (1.69–3.75) (p<0.001), respectively; the corresponding ORs among men were cannot be publicly posted. Data used for this 1.27 (1.12–1.43), 1.56 (1.38–1.77), 2.82 (2.04–3.88) (p<0.001). The presence of dementia PLOS ONE | https://doi.org/10.1371/journal.pone.0253483 July 15, 2021 1 / 13 PLOS ONE Days spent at home during end-of-life trajectory analysis are available from the Government of was a strong predictor of spending less time at home in women (0.41 (0.38–0.43); p<0.001) Catalonia Institutional Data Access (PADRIS and men (0.45 (0.41–0.48); p<0.001). Program) for researchers who meet the criteria for access to confidential data. Inquiries of access to aggregated data should be addressed to AQuAS Conclusions director, Ms. Montserrat Moharra Our results suggest that end-of-life care is associated with gender and socioeconomic (mmoharra@gencat.cat), under the following conditions: http://aquas.gencat.cat/ca/ambits/ inequalities; women and individuals with lower socioeconomic status spend less time at analitica-dades/padris/. home within the last 180 days of life. Funding: The authors received no specific funding for this work. Competing interests: The authors have declared that no competing interests exist. Introduction The increased life expectancy and demographic shift experienced in high-income countries is expected to raise the number of patients in their end-of-life stage [1]. This scenario has encour- aged moving towards a patient-centered care approach in which the alignment between patients’ preferences and healthcare policies gains importance over classical indicators of health―such as life expectancy―for measuring the quality of healthcare delivery [2]. Neverthe- less, evidence suggests that patients’ preferences regarding end-of-life care are not always in line with healthcare policies [3–5]. Besides the upset that these inconsistencies may cause on patients and their relatives, some of these practices have high healthcare costs; hence, despite affecting a low proportion of the population, they might compromise the entire system in the advent of the so-called geriatric tsunami [6, 7]. The place of care in the event of a terminal illness is, along with the therapeutic approach (e.g., life-extending vs. palliative), a cornerstone of patients’ preferences and healthcare policy regarding end-of-life care. Surveys consistently indicate that most patients with advanced and progressive illnesses prefer to die and spend their last days at home [3, 4]. Although some patients may change these preferences at some point during their end-of-life trajectory, pro- spective studies suggest that care preferences are rather stable as the illness progresses [8, 9]. The determinants of the place of care during the end-of-life process have been extensively investigated and seem to result from a complex interaction between illness-, individual-, and environmental-related factors [10, 11]. However, most of these studies have been focused on either the place of death or a traditional concept of end-of-life care, typically confined within the last few days of life [12]. Alternatively, it has been recently proposed that the number of days spent at home in the last six months of life might be a more comprehensive indicator of high-value patient-centered care [13]. Since this concept was proposed, few data have become available regarding the determinants of spending more days at home within the last six months [14–17]. The results of these studies―most of them focused on cancer patients―showed het- erogeneous results, which are not always aligned with those reported in studies investigating factors influencing end-of-life. In this analysis, we have investigated the factors that influence staying at home within the six months preceding death in the general population. Materials and methods Study design This was a population-based, retrospective analysis of end-of-life trajectories of all adult (i.e., 18 years or older) individuals who died in Catalonia (North-east Spain) in 2017 due to advanced and progressive illness. Individuals were selected from the mortality registry of the Catalonian Ministry of Health; the analysis included all individuals of the registry whose cause PLOS ONE | https://doi.org/10.1371/journal.pone.0253483 July 15, 2021 2 / 13 PLOS ONE Days spent at home during end-of-life trajectory of death met the diagnostic criteria proposed by McNamara et al. to define palliative care pop- ulations [18]. The Catalan Health Service provides universal and free healthcare (including primary, hos- pital, and intermediate care) to the entire population. The sole exception is drug prescription, which is based on a co-payment system calculated according to the individual’s income [19]. Long-term care in our area is provided by public and private (nearly half) facilities, including nursing homes and mental health facilities for people with severe mental disorders. Each citi- zen is assigned a unique personal healthcare identification code, which was used to gather information regarding the socioeconomic and clinical characteristics of these individuals, as well as the use of social and healthcare resources during the 180 days preceding death. This was a retrospective analysis of data anonymously collected in the setting of routine management by the national healthcare provider; no data from hospital electronic health rec- ords were included. Since the analysis did not involve any intervention, the Agency for Health Quality and Assessment of the Catalan Minstry of Health waived ethical approval for analysis conduct. Analyzed data were recorded in various central databases of the Catalan Ministry of Health between January and December 2017, and extracted on March 2020. All data were han- dled according to the General Data Protection Regulation 2016/679 on data protection and privacy for all individuals within the European Union and the local regulatory framework regarding data protection. Data sources Clinical and healthcare services use information was retrieved from the healthcare registry of the Catalan Ministry of Health (CMBD, Conjunt Mínim Bàsic de Dades). The Catalan Health Service provides universal healthcare to the entire population (i.e., 7.5 million people). The CMBD systematically collects information regarding resource utilization, including acute care hospitals, intermediate and long-term care, emergency services, primary care, and mental health centers. The CMBD has an automated data validation system that checks data consis- tency and identifies potential errors. Furthermore, as this information is used for provider pay- ment purposes, external audits are regularly conducted to ensure the quality and reliability of the data [20, 21]. All datasets are linked through a unique insurance number. Further details on the datasets retrieved from the CMBD are listed in the S1 File. Data regarding pharmaceutical prescription were retrieved from the pharmacy activity reg- istry, which also identifies individuals institutionalized in a health and social care facilities for intermediate and long-term care (irrespective of whether they are private or public), as well as non-institutionalized individuals. Demographic and socioeconomic information were retrieved from the central population registry [22]. Finally, deaths―and the main cause―and availability of advanced directives (living wills) were identified from the mortality registry of Catalonia and the advanced directives registry of the Catalan Ministry of Health. Variables and outcomes The primary study outcome was the percentage of days spent at home within the last 180 days of life. Other environments included nursing homes, acute care hospitals, emergency depart- ments, emergency medical services, out-of-hours urgent primary care services, and intermedi- ate care facilities, either for rehabilitation or palliative care purposes. The stay of the individual in all settings except home was directly retrieved from the various CMBDs or the pharmacy registry level; time spent at home was defined as the complementary outcome of not being in any other environment. PLOS ONE | https://doi.org/10.1371/journal.pone.0253483 July 15, 2021 3 / 13 PLOS ONE Days spent at home during end-of-life trajectory Characteristics of study participants included gender, age at death, the cause of death, citi- zenship, socioeconomic status (very low, low, mild, and high, based on annual income catego- ries defined for pharmaceutical copayment), living alone, having completed their advanced directives, and being a recipient of a specific local healthcare program: domiciliary care, pallia- tive care, complex chronic patient (PCC, Pacient Crònic Complex), and advanced chronic dis- ease (MACA, Malaltia Crònica Avançada) (S1 File provides further details on the variables collected). Analyses The characteristics of the study population and causes of death were described according to gender and reported using absolute and relative frequencies. The number of days spent in each place was first described using measures of central tendency (mean and median) and the cor- responding measures of dispersion (standard deviation [SD] and interquartile range [IQR], defined by the 25th and 75th percentiles). To determine the factors associated with spending more days at home during the last 180 days, we carried out a robust generalized multivariate linear model with a logit link and the binomial family, with the percentage of days spent at home as a dependent variable. The analy- sis of the determinants of spending time at home environment was performed for sociodemo- graphic, clinical variables, and causes of mortality separately (unadjusted), and by adjusting for all these factors. Results were presented as the estimated percentage of days spent at home and the odds ratio (OR) associated with each factor, along with the corresponding 95% confi- dence interval (95% CI) and its corresponding p-value. All analyses were performed for each gender separately. Complementarily, the relevance of gender in spending more days at home during the 180 days was explored by including this variable in a fully robust adjusted general- ized model. The significance threshold was set at a two-sided alpha value of 0.05 for all analy- ses, which were performed with the Stata Program, version 14.2 (UCLA: Statistical Consulting Group). Results Characteristics of study patients The analysis included 40,137 individuals (19,510 women and 20,627 men) aged 18 years or more who died in Catalonia in 2017 and met the McNamara criteria for palliative care (S1 Fig in S1 File, S1 File). Table 1 summarizes the demographic, socioeconomic, and clinical charac- teristics of women and men included in the analysis. Most individuals (94% of women and 95% of men) had an annual income below 18,000€ or between 18,000€ and 100,000€. Cancer (35%) and dementia (27%) were the leading causes of death among women, whereas men more frequently died from cancer (51%), followed by and COPD (15%). Living alone was more frequent among women (12%) than men (5%). Days spent at home Overall, the study population spent a median of 140.0 days at home (IQR 16.0–171.0) (mean 106.8; SD 71.5) (S2 Fig in S1 File). Fig 1 summarizes the proportion of individuals of the study population staying at home and the investigated facilities along the last 180 days of life. The number of individuals staying at home progressively decreased throughout the entire period and dropped rapidly within the last 60 days. Correspondingly, the number of individuals stay- ing in the hospital increased progressively as approaching the end of life. Fig 2 shows the tran- sitions in the place of stay that occurred in the last ten days of life (i.e., day 171 of the PLOS ONE | https://doi.org/10.1371/journal.pone.0253483 July 15, 2021 4 / 13 PLOS ONE Days spent at home during end-of-life trajectory Table 1. Demographic, socioeconomic, and clinical characteristics of the study population. Women Men (n = 19,510) (n = 20,627) No. (%) No. (%) Age (years) <55 843 (4.32) 1173 (5.69) 55–64 1179 (6.04) 2269 (11) 65–69 865 (4.43) 1781 (8.63) 70–74 1103 (5.65) 2182 (10.58) 75–79 1445 (7.41) 2508 (12.16) 80–84 3054 (15.65) 3836 (18.6) 85–89 4508 (23.11) 3788 (18.36) 90–95 4301 (22.05) 2306 (11.18) �95 2212 (11.34) 784 (3.8) Citizenship Spanish 19033 (97.56) 20032 (97.12) Foreign 477 (2.44) 595 (2.88) Socioeconomic level Recipients of social services1 873 (4.47) 708 (3.43) Annual income <18,000€ 15448 (79.18) 14352 (69.58) Annual income 18,000€ -100,000€ 2960 (15.17) 5285 (25.62) Annual income >100,000€ 89 (0.46) 123 (0.6) Not available 140 (0.72) 159 (0.77) Cause of death Tumors 6833 (35.02) 10417 (50.5) Heart failure 3588 (18.39) 2589 (12.55) Kidney disease 675 (3.46) 542 (2.63) Liver failure 222 (1.14) 492 (2.39) COPD 2123 (10.88) 3179 (15.41) Nervous system 737 (3.78) 846 (4.1) Dementia 5316 (27.25) 2518 (12.21) AIDS 16 (0.08) 44 (0.21) Advanced directives available 77 (0.39) 107 (0.52) Living alone 2324 (11.91) 1088 (5.27) Recipient of specific care programs2 Complex chronic patient 4492 (23.02) 4007 (19.43) Advanced chronic disease 5147 (26.38) 4674 (22.66) Domiciliary care 6716 (34.42) 4868 (23.6) Palliative care 2196 (11.26) 2145 (10.4) 1 Includes individuals perceiving a minimum integration income, unemployment allowance, unemployment benefit or not qualifying for either of the previous. 2 Categories are not mutually exclusive. https://doi.org/10.1371/journal.pone.0253483.t001 investigated period). At that time point, most individuals were at home (n = 15,441; 38.47%), followed by acute care hospital (n = 7,760; 19.63%), and nursing homes (n = 6,384; 15.91%). From that time point on, most transitions occurred from home to acute care hospital, except the last day before death, when nearly half of the patients who left home moved to emergency services/department. Conversely, individuals institutionalized in nursing homes showed a PLOS ONE | https://doi.org/10.1371/journal.pone.0253483 July 15, 2021 5 / 13 PLOS ONE Days spent at home during end-of-life trajectory Fig 1. Percentage of patients staying at home and at each facility within the last six months of life. S1 Fig in S1 File shows the distribution of cases across the primary variable (i.e., days spent at home within the last 180 days of life). https://doi.org/10.1371/journal.pone.0253483.g001 more heterogeneous transfer destination, including hospital, comprehensive community palli- ative care services, and home, with no dominance of a particular destination. Although most individuals admitted to acute care hospitals remained in this place until death, we observed some transitions to home within the few last days of life. Overall, women and men spent a mean (SD) of 106.8 (71.5) and 125.1 (60.0) days at home in the last 180 days of their life, respectively (Table 2). The number of days spent at home pro- gressively declined with age and raised with the socioeconomic status of both women and men. The overall trend of men spending more time than women at home became evident across all sociodemographic characteristics. For instance, women aged <55 years spent more time at home than those aged 90–95 years (mean of 133.2 and 97.4 days, respectively); the Fig 2. Transition of place of stay during the last 10 days before death. Vertical boxes show the place of stay at the given day, counting from 180 days before death. https://doi.org/10.1371/journal.pone.0253483.g002 PLOS ONE | https://doi.org/10.1371/journal.pone.0253483 July 15, 2021 6 / 13 PLOS ONE Days spent at home during end-of-life trajectory Table 2. Descriptive analysis of the number of days spent at home within the last 180 days of life among women and men, according to clinical and socioeconomic characteristics. Women Men Mean (SD) Median (IQR) Mean (SD) Median (IQR) Total 106.77 (71.52) 140.00 (16.00–171.00) 125.10 (60.00) 150.00 (100.00–171.00) Age (years) <55 133.22 (49.90) 150 (113.00–170) 143.12 (44.06) 157.00 (131.00–176.00) 55–64 135.75 (47.05) 152.00 (118.00–169.00) 135.86 (47.94) 151.00 (123.00–170) 65–69 126.13 (55.72) 147.00 (106.00–168.00) 130.70 (51.73) 149.00 (115.00–168.00) 70–74 121.20 (59.98) 146.00 (91.00–168.00) 130.62 (53.19) 150 (113.00–169.00) 75–79 115.82 (63.45) 141.00 (74.00–168.00) 125.39 (56.89) 147.00 (103.00–168.00) 80–84 107.84 (69.56) 138.00 (30–168.00) 122.78 (60.75) 149.00 (92.00–170) 85–89 100.87 (73.98) 133.00 (1.00–170) 118.83 (66.24) 150 (72.00–172.00) 90–95 97.37 (77.08) 132.00 (0–173.00) 114.38 (70.27) 149.50 (37.00–174.00) �95 89.40 (80.06) 105.50 (0–174.00) 111.24 (73.34) 150 (15.50–176.00) Citizenship Spanish 106.20 (71.79) 139.00 (12.00–170) 124.75 (60.27) 150 (99.00–171.00) Foreign 124.29 (56.33) 145.00 (102.00–168.00) 132.59 (50.52) 148.00 (118.00–169.00) Not available 145.17 (48.11) 167.50 (126.00–179.00) 150.03 (40.42) 164.00 (139.50–179.00) Socioeconomic status Recipients of social services� 106.98 (68.19) 135.00 (31.00–167.00) 121.39 (61.53) 148.00 (93.00–170) <18.000€ 103.13 (72.57) 135.00 (5.00–170) 122.42 (61.29) 149.00 (93.00–170) 18.000€ -100.000€ 123.10 (64.93) 153.00 (88.00–175.00) 131.50 (56.22) 154.00 (114.00–173.00) >100.000€ 142.49 (54.64) 168.00 (132.00–179.00) 152.09 (40.25) 171.00 (142.00–179.00) Not available 139.29 (52.91) 164.00 (115.50–179.00) 149.70 (39.90) 163.00 (139.00–179.00) Cause of death Tumors 124.43 (56.30) 146.00 (101.00–166.00) 130.32 (51.19) 149.00 (114.00–167.00) Heart failure 118.89 (69.82) 154.50 (54.00–177.00) 140.48 (56.49) 167.00 (131.00–179.00) Kidney disease 115.88 (70.50) 154.00 (33.00–174.00) 128.90 (59.16) 156.00 (106.00–174.00) Liver failure 121.38 (57.09) 141.00 (97.00–166.00) 139.23 (45.79) 155.00 (124.00–171.00) COPD 104.92 (74.31) 141.00 (2.00–173.00) 122.21 (63.59) 150 (90–173.00) Nervous system 98.84 (77.89) 135.00 (0–175.00) 116.70 (70.96) 156.00 (40–176.00) Dementia 75.91 (78.14) 32.00 (0–168.00) 90.63 (75.79) 109.50 (0–168.00) AIDS 124.75 (59.03) 140 (99.00–176.50) 124.02 (53.95) 138.00 (105.50–164.50) Living wills available Yes 128.34 (54.43) 149.00 (103.00–170) 121.75 (57.18) 145.00 (93.00–165.00) No 106.68 (71.57) 140 (15.00–171.00) 125.12 (60.01) 150 (100–171.00) Living alone Yes 99.66 (74.26) 131.00 (0–169.00) 115.84 (65.97) 146.00 (64.00–170) No 108.90 (70.54) 141.00 (28.00–171.00) 127.34 (58.25) 151.00 (106.00–171.00) � Includes individuals perceiving a minimum integration income, unemployment allowance, unemployment benefit or not qualifying for either of the previous nor other categories. The number of days spent in other resources is summarized in S1 Table in S1 File (women) and S2 Table in S1 File (men) (S1 File) https://doi.org/10.1371/journal.pone.0253483.t002 corresponding times for men at the same age groups were 143.1 and 114.4, respectively. Simi- larly, women in the lowest socioeconomic status spent fewer days at home than those with higher annual income than 100,000 € (mean 107.0 vs. 142.5), while the corresponding values for men were 121.4 and 152.1, respectively. Finally, Spanish women tended to stay less time at PLOS ONE | https://doi.org/10.1371/journal.pone.0253483 July 15, 2021 7 / 13 PLOS ONE Days spent at home during end-of-life trajectory home than foreign women (mean 106.2 vs. 124.3 days); men showed a similar trend, although with systematically higher time than women (132.6 vs. 124.8). Regarding the cause of death, individuals who died from dementia stood fewer days at home in both genders. The gender imbalance was also pervasive in this factor, with women and men who died from dementia spending a mean of 75.9 and 90.6 days at home, respec- tively, and those who died from cancer 124.4 and 130.3. Factors influencing time spent at home The analysis of the percentage of days spent at home for the entire population showed that women had lower odds to stay at home (adjusted OR 0.80 [95% CI 0.77–0.82]; p<0.001) (S3 Table in S1 File). Fig 3 summarizes the results of the adjusted model for men and women. In both genders, higher age and lower socioeconomic status were associated with spending less time at home. Compared with women younger than 55 years, those in the age group of 90–95 spent 36% less time at home (OR 0.65; 95% CI 0.55–0.72); the corresponding decrease among men was 44% (OR 0.56; 0.50–0.63). Regarding the socioeconomic status, women perceiving more than 100,000 € a year had more than twice higher odds to spend time at home compared with those in the lower stratum of socioeconomic status (OR 2.52 [95% CI 1.69–3.75]); the cor- responding OR for men was 2.82 (2.04–3.88). The unadjusted analysis showed the same trends as the adjusted analysis for age, gender, and socioeconomic status (S4 Table in S1 File). In both genders, compared with tumors, dying from dementia was a strong predictor of spending less time at home (OR 0.41 [95% CI 0.38–0.43] in women and 0.45 [0.41–0.48] in men). COPD and nervous system diseases were also associated with less time at home, albeit Fig 3. Adjusted model of percentage of days spent at home within the last 180 days of life, according to gender. Squared dots and lines show the odds ratio and 95% confidence interval of spending a higher percentage of days at home compared with the reference category. The model has been adjusted for all investigated variables. (a) Includes individuals perceiving a minimum integration income, unemployment allowance, unemployment benefit or not qualifying for either of the previous nor other categories. https://doi.org/10.1371/journal.pone.0253483.g003 PLOS ONE | https://doi.org/10.1371/journal.pone.0253483 July 15, 2021 8 / 13 PLOS ONE Days spent at home during end-of-life trajectory to a lesser extent. In men, dying from heart and liver failure were significantly associated with spending more time at home. In women, none of these diagnoses significantly contributed to the time spent at home in the adjusted analysis (although unadjusted analysis showed a signifi- cant influence). Living alone was associated with spending less time at home in both men and women. Discussion In this retrospective analysis of a population-based cohort, we found that older people who died from an advanced and progressive illness spent nearly a quarter of the last six months of life in care facilities. The frequency of transition from home to these facilities became particu- larly frequent within the 60 days preceding death. Male gender and high socioeconomic status were significantly associated with spending more time at home. The cause of death also influ- enced the place of stay, with dementia―more prevalent among women―being a strong predic- tor of spending less time at home. Individuals who lived alone were less likely to spend time at home during the last six months of life. While a body of evidence has been published regarding the determinants of the place of death, to our knowledge, only four studies have specifically investigated the factors that influ- ence time spent at home within the last six months of life: three in population-based cohorts of cancer patients and one in a community-based cohort of 754 patients [14–17]. The population in our analysis was selected using the McNamara criteria for defining palliative care popula- tions in the research setting [18]. Of note, the number of patients eligible for palliative care is typically higher than those receiving it―as was the case of our study sample, with only 10% of individuals receiving palliative care [23, 24]. Hence, rather than palliative care recipients, the population included in our analysis corresponded to individuals who died from advanced and progressive illness eligible for planned end-of-life care. Overall, the median time spent at home in our study population (140 days) was on the lower side of the range found in previous studies (156 to 164 days, both extremes corresponding to cancer populations) [14–16]. Patients who died from cancer in our cohort tended to spend more days at home, though the average time remained below that of previous studies on cancer patients. Gender came up as an important determinant of staying at home during the last six months, with men showing persistently higher percentages of days at home across all variables analyzed. This trend, consistent with previous reports of population-based cohorts of cancer patients [15, 16], illustrates how gender inequalities in healthcare provision observed elsewhere [25] persist during the end-of-life trajectory. In line with the sustainable development goals of the WHO [26], the Catalan government urged to explicitly incorporate a gender perspective in all phases of health planning [27]. Hence, we investigated the factors influencing the amount of time spent at home in women and men separately. The mechanisms underlying gender inequalities are complex and may overlap with other factors such as longevity (women in our population tended to be older), socioeconomic status (women tended to receive lower annual incomes), and social environment (living alone was twice more frequent among women than men), among others. Still, our findings reinforce the need for incorporating a gender perspec- tive in the design of health policies regarding end-of-life trajectory and care programs for patients with advanced diseases. Another important contributor to spending more time at home within the last six months of life was the socioeconomic status, with higher income associated with more days at home in both genders. A similar trend was reported by Anderson et al., although they found a more modest contribution of this factor [14]. The type of population (i.e., cancer patients vs. com- munity-based population) may have contributed to the different weight of this factor. PLOS ONE | https://doi.org/10.1371/journal.pone.0253483 July 15, 2021 9 / 13 PLOS ONE Days spent at home during end-of-life trajectory However, it is worth mentioning that Anderson et al. stratified individuals by income quin- tiles, whereas we used the strata established for drug reimbursement purposes, which result in more extreme differences between income categories and may affect the sensitivity of the anal- ysis. Regardless of the weight of the influence of the socioeconomic status on time spent at home, our analysis depicts a clear gradient across socioeconomic categories. Like in the case of gender, this finding indicates that the socioeconomic inequities found in other healthcare indi- cators (e.g., emergency care visits and hospitalization rate) [28] among older persons also per- sist in end-of-life pathways. The impact of socioeconomic status on multiple aspects of care reinforces the need for designing integrated health and social care strategies. In addition to the patient- or disease-related characteristics, the likelihood of staying at home might be reasonably influenced by context-specific factors such as the organization of the healthcare system. In our environment, patients with chronic and advanced diseases may be eligible for specific care programs aimed at providing multidisciplinary, patient-centered care. However, interactions between these programs (i.e., one patient may be a recipient of two or more care programs simultaneously) challenge the analysis and interpretation of the specific contribution of each of them to the percentage of days spent at home. Future studies designed ad hoc to investigate the influence of each of these programs and their combination on this outcome are warranted. Our results are strengthened by the population-based approach, which allowed us to gather basic socioeconomic and clinical information (both, primary care and specialized) from all individuals covered by the national system of health―virtually the entire population, consider- ing the universal coverage of the healthcare system in our country. On the other hand, retro- spective studies based on administrative databases are constrained by the number and type of variables recorded. In this regard, we could not adjust for disease severity (i.e., the CMBD rec- ords the diagnostic code only), and cultural factors (e.g., religious, relationship patterns, etc.) with known influence on patients’ preferences and behaviors regarding end-of-life trajectories [29]. Likewise, the availability of advanced directives could be identified in less than 1% of our sample, and family preferences regarding end-of-life care, which strongly influences the place of stay [11], could not be assessed. Another constrain regarding the data sources was the prede- fined cutoffs of socioeconomic status, which are tailored to classy the population into levels of pharmaceutical co-payment. Nevertheless, this approach has allowed identifying socioeco- nomic differences in other healthcare outcomes, with consistent results across studies [28, 30, 31]. Finally, studies investigating healthcare outcomes in end-of-life trajectory are challenged by the boundaries of the population to be investigated in this setting [18, 23, 24]. We consid- ered the McNamara criteria, based on the ICD-10 causes of death, the most suitable system for selecting patients with advanced and progressive illnesses, who could be, therefore, candidates to planned end-of-life care. Conclusions Our results show how despite some efforts recently implemented to move towards patient- centered care, individuals with chronic, advanced, and progressive illnesses spend a remark- able amount of time in health and social care facilities within the last six months of life. More worrisome, the significant influence of gender and socioeconomic status on this outcome sug- gests that gender and social inequalities often observed in healthcare may also affect the end- of-life trajectory. These characteristics (and with living alone) should drive prioritization of end-of-life care policies. Taken together, these findings encourage including the outcome "days spent at home during the last six months of life" in healthcare programs as a quality indi- cator of patient-centered care. The inclusion of this outcome may contribute to designing PLOS ONE | https://doi.org/10.1371/journal.pone.0253483 July 15, 2021 10 / 13 PLOS ONE Days spent at home during end-of-life trajectory integrated health and social care pathways flexible enough to provide adequate care during the end-of-life stage. This approach might also offer a more comprehensive perspective of care in this stage of life that goes beyond efficiency indicators such as cost, hospitalization rates, or admissions to the emergency department. Future works to investigate how these factors behave in each cause of death are warranted. Supporting information S1 File. (PDF) Acknowledgments The authors would like to thank Cristina Colls, Raul González, and Ingrid Bullich for their helpful insights on the study conception. Gerard Carot-Sans provided medical writing assis- tance during the manuscript preparation. Author Contributions Conceptualization: Albert Dalmau-Bueno, Anna Garcı́a-Altés, Jordi Amblàs, Joan Carles Contel, Sebastià Santaeugènia. Formal analysis: Albert Dalmau-Bueno. Investigation: Anna Garcı́a-Altés, Jordi Amblàs, Joan Carles Contel, Sebastià Santaeugènia. Methodology: Albert Dalmau-Bueno. Supervision: Anna Garcı́a-Altés. Writing – original draft: Albert Dalmau-Bueno, Anna Garcı́a-Altés, Jordi Amblàs, Joan Carles Contel, Sebastià Santaeugènia. Writing – review & editing: Albert Dalmau-Bueno, Anna Garcı́a-Altés, Jordi Amblàs, Joan Carles Contel, Sebastià Santaeugènia. References 1. Sherman FT. The good news: It’s our 60th birthday—The bad news: A giant, geriatric tsunami! Geriat- rics. 2006. pp. 10–11. 2. World Health Organization (WHO). Palliative Care for Older People: better practices. 2004 [cited 10 Jul 2020]. Available: https://www.euro.who.int/__data/assets/pdf_file/0009/98235/E82933.pdf 3. Barnato AE, Herndon MB, Anthony DL, Gallagher PM, Skinner JS, Bynum JPW, et al. Are regional vari- ations in end-of-life care intensity explained by patient preferences?: A study of the US medicare popu- lation. Med Care. 2007; 45: 386–393. https://doi.org/10.1097/01.mlr.0000255248.79308.41 PMID: 17446824 4. Higginson IJ, Sen-Gupta GJA. Place of care in advanced cancer: A qualitative systematic literature review of patient preferences. J Palliat Med. 2000; 3: 287–300. https://doi.org/10.1089/jpm.2000.3.287 PMID: 15859670 5. Wright AA, Keating NL, Ayanian JZ, Chrischilles EA, Kahn KL, Ritchie CS, et al. Family perspectives on aggressive cancer care near the end of life. JAMA—J Am Med Assoc. 2016; 315: 284–292. https://doi. org/10.1001/jama.2015.18604 PMID: 26784776 6. Penders YWH, de Vleminck A. Serious Illness and Out-of-Pocket Spending. Textbook of Palliative Care. Springer International Publishing; 2018. https://doi.org/10.1007/978-3-319-31738-0_104–1 7. Aldridge MD, Kelley AS. The myth regarding the high cost of end-of-life care. American Journal of Public Health. American Public Health Association Inc.; 2015. pp. 2411–2415. https://doi.org/10.2105/AJPH. 2015.302889 PMID: 26469646 PLOS ONE | https://doi.org/10.1371/journal.pone.0253483 July 15, 2021 11 / 13 PLOS ONE Days spent at home during end-of-life trajectory 8. Auriemma CL, Nguyen CA, Bronheim R, Kent S, Nadiger S, Pardo D, et al. Stability of end-of-life prefer- ences: A systematic review of the evidence. JAMA Intern Med. 2014; 174: 1085–1092. https://doi.org/ 10.1001/jamainternmed.2014.1183 PMID: 24861560 9. Gomes B, Calanzani N, Gysels M, Hall S, Higginson IJ. Heterogeneity and changes in preferences for dying at home: A systematic review. BMC Palliat Care. 2013; 12. https://doi.org/10.1186/1472-684X- 12-7 PMID: 23414145 10. Gomes B, Higginson IJ. Factors influencing death at home in terminally ill patients with cancer: System- atic review. British Medical Journal. BMJ Publishing Group; 2006. pp. 515–518. https://doi.org/10.1136/ bmj.38740.614954.55 PMID: 16467346 11. Costa V, Earle CC, Esplen MJ, Fowler R, Goldman R, Grossman D, et al. The determinants of home and nursing home death: A systematic review and meta-analysis. BMC Palliat Care. 2016; 15: 8. https://doi.org/10.1186/s12904-016-0077-8 PMID: 26791258 12. Lunney JR, Lynn J, Foley DJ, Lipson S, Guralnik JM. Patterns of functional decline at the end of life. Jama. 2003; 289: 2387–2392. https://doi.org/10.1001/jama.289.18.2387 PMID: 12746362 13. Groff AC, Colla CH, Lee TH. Days spent at home—A patient-centered goal and outcome. N Engl J Med. 2016; 375: 1610–1612. https://doi.org/10.1056/NEJMp1607206 PMID: 27783911 14. Andersen SK, Croxford R, Earle CC, Singh S, Cheung MC. Days at Home in the Last 6 Months of Life: A Patient-Determined Quality Indicator for Cancer Care. J Oncol Pract. 2019; 15: e308–e315. https:// doi.org/10.1200/JOP.18.00338 PMID: 30849006 15. Cheung MC, Croxford R, Earle CC, Singh S. Days spent at home in the last 6 months of life: a quality indicator of end of life care in patients with hematologic malignancies. Leuk Lymphoma. 2020; 61: 146– 155. https://doi.org/10.1080/10428194.2019.1654095 PMID: 31437075 16. Gill TM, Gahbauer EA, Leo-Summers L, Murphy TE, Han L. Days Spent at Home in the Last Six Months of Life Among Community-Living Older Persons. Am J Med. 2019; 132: 234–239. https://doi.org/10. 1016/j.amjmed.2018.10.029 PMID: 30447203 17. Bjørnelv GMW, Edwin B, Fretland ÅA, Deb P, Aas E. Till death do us part: the effect of marital status on health care utilization and costs at end-of-life. A register study on all colorectal cancer decedents in Nor- way between 2009 and 2013. BMC Health Serv Res. 2020; 20: 115. https://doi.org/10.1186/s12913- 019-4794-6 PMID: 32054492 18. McNamara B, Rosenwax LK, Holman CDAJ. A Method for Defining and Estimating the Palliative Care Population. J Pain Symptom Manage. 2006; 32: 5–12. https://doi.org/10.1016/j.jpainsymman.2005.12. 018 PMID: 16824980 19. World Health Organization. Regional Office for Europe. Catalonia—General overview. In: Regions for health network [Internet]. [cited 15 Apr 2021]. Available: https://www.euro.who.int/__data/assets/pdf_ file/0004/373387/rhn-catalonia-eng.pdf 20. Catalan Health Service. Minimum basic data set (CMBD). 2017 [cited 9 May 2020]. Available: https:// catsalut.gencat.cat/ca/proveidors-professionals/registres-catalegs/registres/cmbd/index. html#googtrans(ca%7Cen) 21. Farre N, Vela E, Cleries M, Bustins M, Cainzos-Achirica M, Enjuanes C, et al. Medical resource use and expenditure in patients with chronic heart failure: a population-based analysis of 88,195 patients. Eur J Heart Fail. 2016;In Press: 1–9. https://doi.org/10.1002/ejhf.549 PMID: 27108481 22. Catalan Health Service. Central Population Register of CatSalut. 2020 [cited 9 May 2020]. Available: https://catsalut.gencat.cat/ca/proveidors-professionals/registres-catalegs/registres/central-poblacio/ 23. Murtagh FEM, Bausewein C, Verne J, Groeneveld EI, Kaloki YE, Higginson IJ. How many people need palliative care? A study developing and comparing methods for population-based estimates. Palliat Med. 2014; 28: 49–58. https://doi.org/10.1177/0269216313489367 PMID: 23695827 24. Scholten N, Günther AL, Pfaff H, Karbach U. The size of the population potentially in need of palliative care in Germany—an estimation based on death registration data. BMC Palliat Care. 2016; 14: 1–7. https://doi.org/10.1186/s12904-016-0099-2 PMID: 26957121 25. Fernandez E, Schiaffino A, Rajmil L, Badia X, Segura A. Gender inequalities in health and health care services use in Catalonia (Spain). J Epidemiol Community Health. 1999; 53: 218–222. https://doi.org/ 10.1136/jech.53.4.218 PMID: 10396547 26. World Health Organization (WHO). Goal 5: Achieve gender equality and empower all women and girls. In: Sustainable Development Goals [Internet]. Available: https://www.un.org/sustainabledevelopment/ gender-equality/ 27. World Health Organization (WHO). Thirty-year retrospective of Catalan health planning. 2020 [cited 17 Jul 2020]. Available: https://www.euro.who.int/en/health-topics/Health-systems/health-systems- financing/publications/2020/thirty-year-retrospective-of-catalan-health-planning-2020 PLOS ONE | https://doi.org/10.1371/journal.pone.0253483 July 15, 2021 12 / 13 PLOS ONE Days spent at home during end-of-life trajectory 28. Garciá-Altés A, Ruiz-Munõz D, Colls C, Mias M, Martı́n Bassols N. Socioeconomic inequalities in health and the use of healthcare services in Catalonia: Analysis of the individual data of 7.5 million residents. J Epidemiol Community Health. 2018; 72: 871–879. https://doi.org/10.1136/jech-2018-210817 PMID: 30082426 29. Searight HR, Gafford J. Cultural diversity at the end of life: Issues and guidelines for family physicians. Am Fam Physician. 2005; 71: 515–522. PMID: 15712625 30. Cainzos-Achirica M, Capdevila C, Vela E, Cleries M, Bilal U, Garcia-Altes A, et al. Individual income, mortality and healthcare resource use in patients with chronic heart failure living in a universal health- care system: A population-based study in Catalonia, Spain. Int J Cardiol. 2019; 277: 250–257. https:// doi.org/10.1016/j.ijcard.2018.10.099 PMID: 30413306 31. Oliver-Parra A, Dalmau-Bueno A, Ruiz-Muñoz D, Garcı́a-Altés A. Relationship between parents’ mental disorders and socioeconomic status and offspring’s psychopathology: A cross-sectional study. Cerni- glia L, editor. PLoS One. 2020; 15: e0240681. https://doi.org/10.1371/journal.pone.0240681 PMID: 33064781 PLOS ONE | https://doi.org/10.1371/journal.pone.0253483 July 15, 2021 13 / 13