UBC Faculty Research and Publications

Health care use and costs at the end of life: a comparison of elderly Australian decedents with and without… Reeve, Rebecca; Srasuebkul, Preeyaporn; Langton, Julia M; Haas, Marion; Viney, Rosalie; Pearson, Sallie-Anne Jun 21, 2017

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-12904_2017_Article_213.pdf [ 1.11MB ]
JSON: 52383-1.0348406.json
JSON-LD: 52383-1.0348406-ld.json
RDF/XML (Pretty): 52383-1.0348406-rdf.xml
RDF/JSON: 52383-1.0348406-rdf.json
Turtle: 52383-1.0348406-turtle.txt
N-Triples: 52383-1.0348406-rdf-ntriples.txt
Original Record: 52383-1.0348406-source.json
Full Text

Full Text

RESEARCH ARTICLE Open AccessHealth care use and costs at the end of life:a comparison of elderly Australiandecedents with and without a cancerhistoryRebecca Reeve1,4, Preeyaporn Srasuebkul2,6, Julia M. Langton2,5, Marion Haas1* , Rosalie Viney1,Sallie-Anne Pearson2,3 and On behalf of the EOL-CC study authorsAbstractBackground: There is limited population-level research on end-of-life care in Australia that considers health careuse and costs across hospital and community sectors. The aim of this study was to quantify health care use andcosts in the last 6 months of life in a cohort of elderly Australian decedents and to examine the factors associatedwith end-of-life resource use and costs.Methods: A retrospective cohort study using routinely collected health data from Australian Government Departmentof Veterans’ Affairs clients. The study included two cohorts of elderly Australians who died between 2005 and 2009;one cohort with a recorded cancer diagnosis and a comparison cohort with no evidence of a cancer history. Weexamined hospitalisations, emergency department (ED) visits, prescription drugs, clinician visits, pathology, andprocedures and associated costs in the last 6 months of life. We used negative binominal regression to explore factorsassociated with health service use and costs.Results: The cancer cohort had significantly higher rates of health service use and 27% higher total health care coststhan the comparison cohort; in both cohorts, costs were driven primarily by hospitalisations. Older age was associatedwith lower costs and those who died in residential aged care incurred half the costs of those who died in hospital.Conclusions: The results suggest differences in end-of-life care pathways dependent on patient factors, with younger,community-dwelling patients and those with a history of cancer incurring significantly greater costs. There is a need toexamine whether the investment in end-of-life care meets patient and societal needs.Keywords: End-of-life care, Terminal care, Neoplasm, Veterans health, Health care utilisation, Health care costsBackgroundElderly populations continue to grow, with estimatesthat in Organisation for Economic Co-operation and De-velopment (OECD) countries one in five people will be65 years and older by 2030 [1] and, in developed coun-tries, at least half of the population over 65 have morethan one chronic condition [2–5]. As populations be-come sicker, they have increasing health care needs andit is not surprising that per capita health care costsincrease with age [4]. However, the health care costs as-sociated with ageing are low compared with costs in-curred in the 6–12 months prior to death; someestimates show that per capita health care costs are upto four times higher in those at the end of life comparedwith age-matched persons who are not at life’s end. Forexample, Neuman et al. [4] found that in the USA, aver-age Medicare per capita spending for beneficiaries age96 ($US16,145) was more than double that for benefi-ciaries age 70 ($US7,566) [4, 6].There is a growing literature dedicated to understand-ing patterns of health care use and costs at the end oflife, much of which has been conducted using routinely* Correspondence: marion.haas@chere.uts.edu.au1Centre for Health Economics Research & Evaluation, University ofTechnology Sydney, PO Box 123 Broadway, Sydney, NSW 2007, AustraliaFull list of author information is available at the end of the article© Commonwealth of Australia 2017 Open Access This article is distributed under the terms of the Creative Commons 4.0Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made.Reeve et al. BMC Palliative Care  (2018) 17:1 DOI 10.1186/s12904-017-0213-0collected health care data [6–9]. The research highlightsthe extensive array of health care services delivered atthe end of life and the high costs associated with care.However, significant variability is reported in end-of-lifehealth care and costs which is dependent on a range ofpatient and health system factors; one of the most con-sistently important factors is age at death [8]. Severalstudies show that the chance of receiving intensive orlife-sustaining treatments such as chemotherapy, emer-gency department (ED) visits, and admission to intensivecare units and hospitals decreases with age [10–16]. Assuch, it is not surprising that corresponding health carecosts decrease steadily with age, particularly in thoseaged 85 years and older [4, 14, 15, 17, 18]. To date, themajority of research on end-of-life care has been under-taken in Europe or North America. Given the complex-ities and jurisdiction-specific features of end-of-life carefor elderly patients, there is a need to conduct studiesthat account for local health system characteristics suchas the relationship between primary and hospital-basedcare, level of subsidisation of services and the availabilityof hospice and palliative care services.The aim of this study is to quantify health care useand costs in the last 6 months of life in the Australiansetting and to examine the factors associated with end-of-life care in an elderly decedent population. We com-pare total health care use and costs in two elderly dece-dent cohorts; a cohort with a history of cancer and acohort without a cancer history. We focused on compar-ing cohorts with and without a previous cancer diagnosisas cancer is the leading cause of death in Australia andother similar countries and cancer patients represent asignificant proportion of patients receiving end-of-lifecare [19, 20]. Additionally, patterns of end-of-life cancercare are more extensively studied and understood thanother terminal diseases which provides an importantcontext within which to interpret our findings [21].MethodsSettingThe Australian Government Department of Veterans’Affairs (DVA) funds the health care of eligible veterans,war widows, war widowers and their dependents. DVAclients have access to the universal health care arrange-ments provided to Australian permanent residents andcitizens plus additional DVA-approved services andpharmaceutical items not subsidised for the generalpopulation; the types of services subsidised depends onthe level of entitlement.Study design and populationThis study used data from the cancer and comparisoncohorts developed for the End of Life in Cancer Care(EOL-CC) study, the details of which have beenpublished elsewhere [22]. Briefly, this is a retrospectivestudy of two decedent cohorts of Australian DVA clientswith full health care entitlements during the last12 months of life to ensure near-complete capture ofhealth-related resource use and costs. Clients were eli-gible for the study if they: resided in New South Wales(Australia’s most populous state) in the 18 months priorto death; were aged 65 years or older at death; died be-tween June 30 2005 and December 31 2009; and re-ceived at least one health service in the last 12 monthsof life. The cancer cohort had a notifiable cancer diagno-sis recorded in the NSW Central Cancer Registry be-tween 1994 and 2009. The comparison cohortcomprised the remaining eligible clients with no evi-dence of a cancer notification or any cancer relatedhealth service use or cancer medicine; for more informa-tion, see our study protocol [22].Data sources and linkageThe data infrastructure comprises DVA data holdingslinked with NSW Ministry of Health data collections.Data were linked by a third party under the custodian-ship of the DVA, the NSW Ministry of Health and theCancer Institute NSW, using best practice, probabilistic,privacy preserving protocols [23]. The linked datasetscapture information on DVA client characteristics, resi-dence in an aged care facility, cancer notification historyand cause of death. However, we did not present infor-mation on cause of death for the two cohorts becauseinformation was only available to 2007. For those whodied in the period 2005–2007, the most common causesof death in the cancer cohort were cancer (59%) and dis-eases of the circulatory system (21%). For the compari-son cohort, the most common causes of death werediseases of the circulatory system (48%), lung disease(7%) and dementia (9%); for more information see ourstudy protocol [23].The linked datasets also include information about allsubsidised health services including hospital admissions,emergency department (ED) presentations and dis-pensed prescription medicines. Clinical consultations,pathology services and procedures; procedures includediagnostic services (e.g., ultrasound, CT, MRI), thera-peutic services (e.g., radiotherapy) and surgery are alsocaptured; unit costs were available for all these services,while costs for the hospital admissions and ED presenta-tions were calculated per admission or presentationusing the methodology described in the NSW Costs ofCare Standards Report 2009/10. Full details of the cost-ing methods are described elsewhere [22].Statistical analysesOutcomes were health service use and associated costsin the last 6 months of life based on six constructedReeve et al. BMC Palliative Care  (2018) 17:1 Page 2 of 10‘months’ consisting of 30 days each; all individuals areobserved for exactly 180 days to the date of death.We calculated person-level mean (with 95% confi-dence interval (CIs)) and median (with inter quartilerange (IQR)) service use and costs for the entire six-month period, and by month until death.We allocated unit costs to each item of resource use tocalculate mean (95% CI) and median (IQR) costs in 2009/10 Australian dollars. Total costs were calculated as thesum of costs of health services (excluding any services re-ceived during a hospitalisation which are captured in hos-pital costs), pharmacy, ED and hospital costs (excludingthe pharmacy component of hospital costs for private hos-pital patients which is captured in the prescribing data-base); more details are reported elsewhere [22, 24].Factors associated with resource use and costsWe used negative binomial regression to determine fac-tors associated with health service use and costs. Thecost data can be thought of as counts (of dollars andcents). Cost data are non-negative and skewed so countdata models can be applied to these data, and in ourcase this makes sense because it is consistent with themethod used for utilization [25]. Factors included: age,sex, comorbidity burden, location of residence (remote-ness, areas of socioeconomic disadvantage), year ofdeath, residence in an aged care facility and place ofdeath. These variables were selected based on commonfactors found to predict patterns of health service use asdeath approaches [22]. Comorbidity burden was esti-mated in the periods prior to the last 6 months of lifeusing hospitalisation codes (Charlson index [26]) andprescription dispensing history (RxRisk [27]); while theCharlson is likely to under ascertain morbidity burden itis more likely to capture people with severe morbidity,e.g. those needing admission to hospital for the condi-tion. Two indices were used to capture a more completemorbidity history [22, 28].Separate models were estimated for utilisation and costsof each type of health service (i.e., clinician visits, path-ology and procedures; prescription medicines dispensed;ED visits and hospitalisations) and total health care costs.The strength of associations was represented by adjustedincident rate ratios with 95% CIs, and two-tailed p < 0.05were used as a criterion for statistical significance.We used SAS version 9.3 (SAS Institute) for data ma-nipulation and performed statistical analyses usingSTATA version 12 (StataCorp).ResultsCohort characteristics [Table 1]A total of 9862 decedents met the eligibility criteria forthe cancer cohort and 15,483 for the comparison cohort.The cohorts were similar in relation to age, socioeconomicstatus and geographical location of residence. Comorbidityburden (based on hospitalisations) was higher in the can-cer cohort; 17% of cancer decedents versus 10% of thecomparison cohort had a comorbidity score of three ormore. However, comorbidity burden was similar for bothcohorts when calculated based on prescription medicinesdispensed [Table 1].Resource use and costs in the last 6 months of lifeOn average, decedents in the cancer cohort were dis-pensed 41 medicines (vs. 38 medicines in the compari-son cohort), received 90 clinician visits/procedures (vs.66 in the comparison cohort) and had approximatelythree hospital admissions (vs. two in the comparison co-hort). The mean number of ED visits was similar in bothcohorts at about one visit/decedent in the last 6 monthsof life [Additional file 1: Table S1].The mean total cost associated with health care in thelast 6 months of life was higher in the cancer cohort($28,091 per decedent) than the comparison cohort($19,696 per decedent). Costs were driven primarily byhospitalisations, accounting for about 80% of total costsin both cohorts [Fig. 1].Prescription medicinesThe mean cost of medicines per person in the cancercohort was $1840 compared with $1234 for the com-parison cohort. [Additional file 1: Table S1] The differ-ence was driven by the costs of antineoplastic andimmunomodulation agents (mean of $469 per person inthe cancer cohort vs. $5 in the comparison cohort).Clinician visits and proceduresDifferences in utilisation can be attributed mainly tohigher numbers of medical specialist visits and pathologyservices in the cancer cohort. The greatest difference inhealth service costs related to diagnostic procedures at amean cost of $1041 per person in the cancer cohortcompared to $568 for the comparison cohort. The costof therapeutic procedures also differed substantially be-tween cohorts, with mean costs of $877 and $464 perperson in the cancer and comparison cohorts respect-ively [Additional file 1: Table S1].HospitalisationsThe difference in hospitalisation costs was driven by thehigher number of admissions in the cancer cohort (meanof three episodes per person costing $22,852 in total)than the comparison cohort (two episodes per personcosting $15,893 in total). [Additional file 1: Table S1]Only 15% of the cancer cohort and 3% of the comparisoncohort received a palliative service while in hospital; andReeve et al. BMC Palliative Care  (2018) 17:1 Page 3 of 104% of the cancer cohort and <1% of the comparison cohortwere admitted to a hospice [Additional file 1: Table S2].Resource use and costs by month, during the last 6 monthsof life [Figs. 2 and 3]Rates of health service use and associated costsincreased over the last 6 months of life and peaked inthe last month of life. Total costs for both cohorts andin each month to death are driven by the cost of hospi-talisations [Fig. 1].Factors associated with resource use and costsMultivariable analyses demonstrate that cancer decedentshad significantly higher rates of prescription medicines(adjusted IRR: 1.09; 95% CI: 1.08–1.11, p < 0.001); clin-ician visits and procedures (adjusted IRR: 1.23; 95% CI:1.20–1.25, p < 0.001); hospitalisations (adjusted IRR: 1.26;95% CI: 1.23–1.30, p < 0.001) and ED visits (adjusted IRR:1.05; 95% CI: 1.02–1.08, p < 0.001) than the comparisoncohort. Decedents aged over 90 at death had lower ratesof prescription medicines (adjusted IRR: 0.92; 95% CI:0.90–0.93, p < 0.001); clinician visits and procedures (ad-justed IRR: 0.81; 95% CI: 0.79–0.84, p < 0.001); hospitalisa-tions (adjusted IRR: 0.76; 95% CI: 0.74–0.79, p < 0.001)and ED visits (adjusted IRR: 0.91; 95% CI: 0.88–0.95,p < 0.001) than younger decedents [Fig. 4].Overall, cancer decedents incurred 27% higher healthcare costs than non-cancer decedents, (adjusted IRR:1.27; 95% CI: 1.24–1.30, p < 0.001) [Fig. 5]. The cancercohort had 42% higher costs for prescribed medicines(adjusted IRR: 1.42; 95% CI: 1.39–1.46, p < 0.001), 32%higher costs for clinician visits and procedures (adjustedIRR: 1.32; 95% CI: 1.29–1.35, p < 0.001) and 28% highercosts for hospitalisations (adjusted IRR: 1.28; 95% CI:Table 1 Cohort characteristicsCancer cohort(N = 9862)Comparison cohort(N = 15,483)n (%) n (%)SexFemale 3116 (31.6) 7521 (48.9)Male 6746 (68.4) 7962 (51.4)Age in years: median (IQR) 86 (83–89) 87 (84–90)Age at death65–74 294 (3.0) 254 (1.7)75–84 4075 (41.3) 5028 (32.5)85–94 5215 (52.9) 9232 (59.6)95–104 277 (2.8) 958 (6.2)≥ 105 1 (0.0) 11 (0.1)Year of death2005 1204 (12.2) 1772 (11.4)2006 2236 (22.7) 3199 (20.7)2007 2351 (23.8) 3473 (22.4)2008 2133 (21.6) 3619 (23.4)2009 1938 (19.7) 3420 (22.1)Age at cancer diagnosis:Median (IQR)83 (78–86) Not applicableYears from diagnosis to deathyears: Median (IQR)1.6 (0.2–5.6) Not applicableLocation of residence (remoteness area)Major cities 6147 (62.3) 9530 (61.6)Inner Regional 2777 (28.2) 4400 (28.4)Outer Regional 872 (8.8) 1410 (9.1)Remote 39 (0.4) 81 (0.5)Very Remote 5 (0.1) 2 (0.0)Missing 22 (0.2) 60 (0.4)Socioeconomic disadvantage index(most disadvantaged) 1–2 1160 (11.8) 1862 (12.0)3–4 2831 (28.7) 4470 (28.9)5–6 2032 (20.6) 3085 (19.9)7–8 1418 (14.4) 2248 (14.5)(least disadvantaged) 9–10 2019 (20.5) 3183 (20.6)Missing 402 (4.1) 635 (4.1)Comorbidity burdena (based on hospitalisations)0 3105 (31.5) 4451 (28.8)1–2 1500 (15.2) 2068 (13.4)≥ 3 1713 (17.4) 1578 (10.2)Unable to calculate, nohospitalisations3544 (35.9) 7386 (47.7)Table 1 Cohort characteristics (Continued)Comorbidity burdenb (based on prescriptions)0 461 (4.7) 846 (5.5)1–2 1298 (13.2) 2130 (13.8)3–5 3865 (39.2) 6104 (39.4)≥ 6 4238 (43.0) 6403 (41.4)Living in residential aged care atany time during the last 6 monthsof life3659 (37.1) 8940 (57.7)Place of deathHospital 5740 (58.2) 6861 (44.3)Residential aged care 2507 (25.4) 6435 (41.6)Other 1615 (16.4) 2187 (14.1)© Commonwealth of Australia 2017aCharlson comorbidity index calculated using hospitalisations between 18 and7 months before deathbRx-Risk comorbidity index calculated using dispensing history in the 6 monthperiod before the last 6 months of life (between month 12 and 7before death)Reeve et al. BMC Palliative Care  (2018) 17:1 Page 4 of 101.21–1.35, p < 0.001) than the comparison cohort. Therewas no significant difference in costs associated with EDvisits (adjusted IRR: 1.05; 95% CI: 0.98–1.12, p = 0.166)[Fig. 5]. A number of factors were associated with lowerhealth care costs: decedents aged over 90 at death had20% lower health care costs than decedents aged 80–84(adjusted IRR: 0.80; 95% CI: 0.77–0.82); decedents resid-ing in residential aged facilities incurred 9% lower totalcosts (adjusted IRR: 0.91; 95% CI: 0.88–0.95); and thosewho died in residential aged care incurred less than halfthe costs of those who died in hospital (adjusted IRR:0.47; 95% CI: 0.45–0.49) [Fig. 5].Fig. 1 Total health care costs in the last six months of life, by month and by health service type for the (a) cancer cohort and (b) comparison cohortFig. 2 Health service use and associated costs in the last 6 months of life, month. a Prescription medicines dispensed; and (b) Clinician visits,pathology and proceduresReeve et al. BMC Palliative Care  (2018) 17:1 Page 5 of 10Fig. 3 Health service use and associated costs in the last 6 months of life, by month. a Hospital admissions and (b) Emergency Department visitsFig. 4 Multivariable analysis examining the associations between cohort characteristics and costs in the last 6 months of life, by health servicetype and total cost1Reeve et al. BMC Palliative Care  (2018) 17:1 Page 6 of 10DiscussionThis study used routine health care data collections toquantify and characterise factors associated with healthcare use and costs at the end of life in elderly decedents.Our results suggest wide variation in end-of-life care,particularly when comparing decedents with and with-out a cancer history. Decedents with a cancer historyhad higher rates health services use and higher associ-ated costs than non-cancer decedents. Age at death wasalso a determinant of end-of-life care, with older dece-dents in both cohorts using fewer health care servicesthan younger decedents. This is the first study of its kindin Australia, yet the results align with studies conductedinternationally in terms of the impact of age on healthservices use and distribution of health care costs, withhospitals being the main driver of end-of-life costs [8].Our results suggest systematic differences in the carereceived by elderly populations, particularly those aged85 years and over. The patterns of care received by thisgroup may suggest a different attitude towards active orintensive treatment for this group compared with theiryounger counterparts. Our finding that older decedentswere less likely to use hospital and ED services are simi-lar to another study examining the entire NSW popula-tion which reported that decedents aged 90 and olderwere 60% less likely to have more than three hospital ep-isodes and 85% less likely to spend time in the ICU thanthose aged 60–79 years at death [29]. Our results showthat not only were elderly decedents (both cancer andnon-cancer) less likely to be admitted to hospital, butalso received fewer clinician visits and prescription med-icines at the end of life. The reduced services in elderlydecedents may be compounded by the fact that thisgroup are more likely to live in residential aged care set-tings; those who died in residential aged care incurredhalf the health care costs of those who died in hospital.While some literature suggests that lower rates of hos-pital and ED use at the end of life may represent qualitycare, [30] there is a need to further investigate whetherthe reductions in end-of-life health service use in elderlyFig. 5 Multivariable analysis examining the associations between cohort characteristics and health service use in the last 6 months of lifeReeve et al. BMC Palliative Care  (2018) 17:1 Page 7 of 10decedents translate into appropriate treatment that isconsistent with patient preferences [31, 32]. We suggestthat the differences in care observed in this and otherstudies does not necessarily represent inequities in treat-ment. Rather, the very elderly may have their end-of-lifeneeds met with other sources of care [33, 34].Services such as palliative hospital admissions and useof hospice care are regarded as important for deliveringquality end-of-life care [7]. The proportion of people re-ceiving palliative services in hospital in the last 6 monthsof life was low (14.7% in the cancer cohort and 3.1% forthe comparison cohort); hospice admissions were alsolow (4.1% and 0.5% respectively, for the same cohorts).-While our findings are consistent with other studiesreporting the challenges of elderly patients accessing pal-liative services in acute care settings, [35] a limitation ofour study is that rates of palliative care services deliveredmay be underestimated due to coding practices inAustralia not being tied to hospital payment [36]. Never-theless, the low rates of palliative services in this studyare important from a health care planning perspectiveand warrant further examination. National data demon-strate that less than half of all patients who die in hos-pital receive any palliative care service [36].Our findings have important policy implications withinthe broader social and economic context of end-of-lifecare. For instance, international studies and surveys of theAustralian population indicate that most people with aterminal condition would prefer to die at home and prefera symptom management approach rather than intensivetreatment or heroic life-saving measures [37–39]. How-ever, we found that hospital was the most common placeof death and also placed the greatest burden on the healthsystem in terms of costs. This indicates the need for morewidespread discussion between patients, providers, care-givers, and health system managers about end-of-life plan-ning (e.g., measures such as advanced care directives).Indeed, patients’ awareness that their condition is terminaland end-of-life discussions with clinicians have beendemonstrated to result in the delivery of care in-line withpatient preferences [31, 40]. There is also evidence thatsuggests there is merit in increasing resources forcommunity-based palliative care services [41, 42]. Whileour study is comprehensive in that it captures nearly allhealth services from a health payer perspective, recentstudies considering costs more broadly found thatbetween one-fifth and one-third of overall end-of-lifecare costs can be attributed to informal care givers[43]. Costs of informal caregiving and patient prefer-ences are not captured in routinely collected dataholdings but this information is needed to fullyunderstand how to maximise the value of end-of-lifecare services at a population-level in a health careenvironment with limited resources.There are a number of strengths of our study, includingour comprehensive patient-level analysis using multiplelinked routine data collections. Our findings relating toDVA clients may not be generalisable to the elderlyAustralian cancer population. However, there is evidencethat DVA clients have similar rates of health service usewhen compared with Australians of a similar age [44, 45]and comparison with recent studies on end-of-life hospitalcare in the general population suggest similar rates ofhospital admissions and ED visits [29, 46].A study limitation is that we do not provide informa-tion on patterns of end-of-life care in people under65 years of age, a group that represents about 20% of alldeaths in Australia [47] and up to one third of cancerdeaths [20, 48]. Also there may have been changes inpractice in Australia since our study period ended in2009; however, the data reflect the most recent availableat the time this research was undertaken. Anotherpotential limitation of our study is the retrospectivestudy design that does not necessarily reflect the chal-lenges associated with predicting survival time in realworld clinical practice [49]. However, some researchcomparing prospective and retrospective designs foundsimilar patterns of end-of-life health care [13, 50].ConclusionDecedents with a history of cancer used more servicesand incurred higher costs in the last 6 months of life thandecedents without a cancer history. We also found evi-dence of a shift in treatment patterns by age with dece-dents aged 65–84 years receiving a greater number andrange of services (particularly hospital based services) thanthose aged85 years and older. Future research is requiredto understand which patterns of care are associated withimproved patient-reported care and to understand whichpatterns of care represent the best value in terms ofappropriateness, quality and patient preferences.Additional fileAdditional file 1: Table S1. Mean and median health service use andcosts in the last 6 months of life, by health service type, and cohort.Table S2. Mean and median health service use and costs in the last 6months of life, by health service type, and cohort. (DOCX 24 kb)AbbreviationsCT: Computed tomography; DVA: Department of Veterans Affairs;ED: Emergency Department; EOL-CC: End-of-life cancer care study;ICU: Intensive care unit; MRI: Magnetic resonance imaging; NSW: New SouthWalesAcknowledgementsThe authors would like to thank the Centre for Health Record Linkage forundertaking the data linkage process.Reeve et al. BMC Palliative Care  (2018) 17:1 Page 8 of 10FundingThis research is supported, in part, by a National Health and Medical ResearchCouncil Health (NHMRC) Services Research Capacity Building Grant (ID: 571,926),NHMRC Centre of Research Excellence in Medicines and Ageing (ID: 1,060,407),Cancer Australia Grant (ID: 568,773) and a University of Sydney, Faculty ofPharmacy Small Project Grant. SP is supported by a Cancer Institute New SouthWales Career Development Fellowship (ID: 12/CDF/2–25).Availability of data and materialsOur service agreement with the Australian Department of Veterans’ Affairsdoes not permit the authors to share the data with other parties.Authors’ contributionsRV, MH and SP designed the research and supervised the analysis. RR and PSperformed the statistical and econometric analyses. All authors reviewed theresults and helped to draft the manuscript. All authors read and approvedthe final manuscript.Competing interestsThe authors report no actual, potential, or perceived competing interestswith regard to the submission of this manuscript.Consent for publicationNot applicable.Ethics approval and consent to participateThe NSW Population and Health Services Ethics Committee (approvalnumber 2013/11/494) and the Department of Veterans’ Affairs HumanResearch Ethics Committee (E013/015) approved this project.DisclaimerThe Australian Government Department of Veterans’ Affairs is the copyrightowner of the data presented in this manuscript including all tables, figuresand supplementary content. The authors have obtained permission from theAustralian Government Department of Veterans’ Affairs to publish thismanuscript. The views expressed in this version of the work do notnecessarily represent the views of the Minister for Veterans’ Affairs or theDepartment of Veterans’ Affairs. The Commonwealth of Australia does notgive any warranty nor accept any liability in relation to the contents of thiswork.Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.Author details1Centre for Health Economics Research & Evaluation, University ofTechnology Sydney, PO Box 123 Broadway, Sydney, NSW 2007, Australia.2Faculty of Pharmacy, The University of Sydney, Sydney, NSW 2006, Australia.3Centre for Big Data Research in Health, Faculty of Medicine, UNSW Australia,Sydney, Australia. 4Centre for Social Impact, UNSW Australia, Sydney, NSW2000, Australia. 5Centre for Health Services and Policy Research, TheUniversity of British Columbia, Vancouver, BC V6T 1Z3, Canada. 6Departmentof Developmental Disability Neuropsychiatry, Faculty of Medicine, UNSWAustralia, Sydney, NSW 2052, Australia.Received: 1 May 2016 Accepted: 8 June 2017References1. OECD. Economic, environmental and social statistics. Paris: OECD; 2009.2. Barnett K, Mercer SW, Norbury M. Epidemiology of multimorbidity andimplications for health care, research, and medical education: a cross-sectional study. Lancet. 2012;380:37–43.3. Fortin M, Stewart M, Poitras M. A systematic review of prevalence studies onmultimorbidity: toward a more uniform methodology. Ann Fam Med. 2012;10:142–51.4. Neuman P, Cubanski J, Damico A. Medicare per capita spending by age andservice: new data highlights oldest beneficiaries. Health Aff. 2014;34:335–9.5. Tinetti ME, Fried TR, Boyd CM. Designing health care for the most commonchronic condition-multimorbidity. JAMA. 2012;307:2493–4.6. Yabroff KR, Warren JL, Brown ML. Costs of cancer care in the USA: adescriptive review. Nat Clin Pract Oncol. 2007;4:643–56.7. Barbera L, Seow H, Sutradhar R, Chu A, Burge F, Fassbender K, et al. Qualityindicators of end-of-life care in patients with cancer: What rate is right?J Oncol Pract. 2015;11(3):e279–87.8. Langton JM, Blanch B, Drew AK. Retrospective studies of end-of-liferesource utilization and costs in cancer care using health administrativedata: a systematic review. Palliat Med. 2014;28:1167–96.9. Luta X, Maessen M, Egger M. A., S. Measuring intensity of end of life care:review. PLoS One. 2015;10:e0123764.10. Burge FI, Lawson BJ, Johnston GM. A population-based study of ageinequalities in access to palliative care among cancer patients. Med Care.2008;46(12):1203–11.11. Gagnon B, Mayo NE, Hanley J. Pattern of care at the end of life: does agemake a difference in what happens to women with breast cancer? J ClinOncol. 2004;22(17):3458–65.12. Ho TH, Barbera L, Saskin R. Trends in the aggressiveness of end-of-life cancercare in the universal health care system of Ontario, Canada Oncol. J Clin Oncol2011, 29:1587-1591.13. Setoguchi S, Earle CC, Glynn R. Comparison of prospective and retrospectiveindicators of the quality of end-of-life cancer care. J Clin Oncol. 2008;26:5671–8.14. Shugarman LR, Bird CE, Schuster CR. Age and gender differences in Medicareexpenditures at the end of life for colorectal cancer decedents. J Women’sHealth. 2007;16:214–27.15. Shugarman LR, Bird CE, Schuster CR. Age and gender differences inmedicare expenditures and service utilization at the end of life for lungcancer decedents. Womens Heal Issues. 2008;18:199–209.16. Virnig BA, McBean A, Kind S. Marshall hospice use before death: variabilityacross cancer diagnoses. Med Care. 2002;40:73–8.17. Kaul P, McAlister FA, Ezekowitz JA. Resource use in the last 6 months of lifeamong patients with heart failure in Canada. Arch Intern Med. 2011;171:211–7.18. Unroe KT, Greiner MA, Hernandez AF. Resource use in the last 6 months oflife among medicare beneficiaries with heart failure, 2006-2007. Arch InternMed. 2011;171:196–203.19. Australian Institute of Health and Welfare (AIHW). Cancer survival andprevalence in Australia: period estimates from 1982 to 2010. Canberra:AIHW; 2013.20. Australian Institute of Health and Welfare (AIHW). Cancer in Australia: anoverview. Canberra: AIHW; 2014.21. Lunney JR, Lynn J, Hogan C. Profiles of older medicare decedents. J AmGeriatr Soc. 2002;50:1108–12.22. Langton JM, Srasuebkul P, Reeve R, Parkinson B, Gu Y, Buckley NA, et al.Resource use, costs and quality of end of life care: observations in a cohortof elderly Australian cancer decedents. Implement Sci. 2015;10:25.23. Centre for Health Record Linkage (CHeReL). http://www.cherel.org.au/.24. Reeve R, Haas M. Estimating the cost of emergency department presentationsin NSW. CHERE working paper 2014/01. In. Sydney: CHERE; 2014.25. Jones A. Models for health care. HEDG working paper 10/01. York: Universityof York; 2010.26. Sundararajan V, Henderson T, Perry C. New ICD 10 version of the Charlsoncomorbidity index predicted in-hospital mortality. J Clin Epidemiol. 2004, 57SRC - GoogleScholar:1288–1294.27. Sloan KL, Sales AE, Liu C. Construction and characteristics of the RxRisk-V: a VA-adapted pharmacy-based case-mix instrument. Med Care. 2003;41:761–74.28. Lu CY, Barratt J, Vitry A. Charlson and Rx-risk comorbidity indices werepredictive of mortality in the Australian health care setting. J Clin Epidemiol.2011;64:223–8.29. Goldsbury DE, Connell DL, Girgis A, C. BM. O’ Acute hospital-basedservices used by adults during the last year of life in New South Wales,Australia: a population-based retrospective cohort study. Health ServRes. 2015; 15:537.30. Earle CC, Neville BA, Landrum MB. Evaluating claims-based indicators ofthe intensity of end-of-life cancer care. Int J Qual Health Care. 2005;17:505–9.31. Mack JW, Cronin A, Taback N. End-of-life care discussions among patientswith advanced cancer: a cohort study. Ann Intern Med. 2012;156:204–10.32. You JJ, Dodek P, Lamontagne F. What really matters in end-of-lifediscussions? Perspectives of patients in hospital with serious illness andtheir families. Can Med Assoc J. 2014;186:E679–87.33. Burt J, Plant H, Omar R, Raine R. Equity of use of specialist palliative care by age:cross-sectional study of lung cancer patients. Palliat Med. 2010;24(6):641–50.Reeve et al. BMC Palliative Care  (2018) 17:1 Page 9 of 1034. Burt J, Raine R. The effect of age on referral to and use of specialistpalliative care services in adult cancer patients: a systematic review. AgeAgeing. 2006;35(5):469–76.35. Gardiner C, Cobb M, Gott M, Ingleton C. Barriers to providing palliative carefor older people in acute hospitals. Age Ageing. 2011;40(2):233–8.36. Australian Institute of Health and Welfare (AIHW). Palliative care services inAustralia. Canberra: AIHW; 2014.37. The National Palliative Care Strategy 2010. Supporting Australians to livewell at the end of life. Canbera: Commonwealth of Australia; 2010.38. Gomes B, Calanzani N, Gysels M, C BM. Heterogeneity and changes inpreferences for dying at home: a systematic review. BMC Palliat Care. 2013; 12:7.39. Gomes B, Higginson IJ, Calanzani N. Preferences for place of death if facedwith advanced cancer: a population survey in England, Flanders, Germany,Italy, the Netherlands, Portugal and Spain. Ann Oncol. 2012;23:2006–15.40. Zhang B, Wright AA, Huskamp HA. Health care costs in the last week of lifeassociations with end-of-life conversations. Arch Intern Med. 2009;169:480–8.41. Abernethy AP, Currow DC, Shelby-James T, Rowett D, May F, Samsa GP, et al.Delivery strategies to optimize resource utilization and performance status forpatients with advanced life-limiting illness: results from the “palliative care trial”[ISRCTN 81117481]. J Pain Symptom Manag. 2013;45(3):488–505.42. Almaawiy U, Pond GR, Sussman J. Are family physician visits and continuityof care associated with acute care use at end-of-life? A population-basedcohort study of homecare cancer patients. Palliat Med. 2014;28:176–83.43. McCaffrey N, Cassel JB, Coast J. Bringing the economic cost of informalcaregiving into focus. Palliat Med. 2015;29:866–7.44. Australian Institute of Health and Welfare (AIHW). Health care usage andcosts: A comparison of veterans and war widows and widowers with therest of the community Canberra Australia. Canberra: AIHW; 2002.45. Tooth LR, Hockey R, Treloar S. Does government subsidy for costs ofmedical and pharmaceutical services result in higher service utilization byolder widowed women in Australia? BMC Health Serv Res. 2012;12:179.46. Rosenwax LK, McNamara BA, Murray K. Hospital and emergencydepartment use in the last year of life: a baseline for future modifications toend-of-life care. Med J Aust. 2011;194:570–3.47. Age at death (AIHW). http://www.aihw.gov.au/deaths/age-at-death/.Accessed 19 Mar 2015.48. Australian Cancer Incidence and Mortality (ACIM) books. http://www.aihw.gov.au/acim-books/. Accessed 19 Mar 2015.49. Bach PB, Schrag D, Begg CB. Resurrecting treatment histories of dead patients:a study design that should be laid to rest. JAMA. 2004;292:2765–70.50. Pyenson B, Connor S, Fitch K. Medicare cost in matched hospice andnon-hospice cohorts. J Pain Symptom Manag. 2004;28:200–10.•  We accept pre-submission inquiries •  Our selector tool helps you to find the most relevant journal•  We provide round the clock customer support •  Convenient online submission•  Thorough peer review•  Inclusion in PubMed and all major indexing services •  Maximum visibility for your researchSubmit your manuscript atwww.biomedcentral.com/submitSubmit your next manuscript to BioMed Central and we will help you at every step:Reeve et al. BMC Palliative Care  (2018) 17:1 Page 10 of 10


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items