Prolonged Wait Time Prior to Entry to Home Care Packages Increases the Risk of Mortality and Transition to Permanent Residential Aged Care Services: Findings from the Registry of Older South Australians (ROSA)

The term home care package (HCP) is used in this paper to describe the community-based services funded by the Commonwealth Government. The packages before August 2013 consisted of increasing levels of care from community aged care packages (CACP) through to extended aged care at home (EACH) and extended aged care at home dementia (EACH-D) packages. After August 2013, these packages were replaced by a 4 level system of packages, where Levels 1 and 2 in this paper is equated to the historical CACP and Levels 3 and 4 are equated to the EACH/EACH-D (9).

A retrospective cohort study was conducted using data from the National Historical Cohort of the Registry of the Older South Australians (ROSA), which comprised linked data obtained from the Australian Institute of Health and Welfare (AIHW) National Aged Care Data Clearinghouse. Specifically, this study used datasets from the Australian Commonwealth Aged Care Assessment Program (ACAP), HCPs details, residential aged care services details, and National Death Index (NDI) (5, 10, 11). The ACAP de-identified dataset includes every assessment performed by an Aged Care Assessment Team (ACAT) when determining eligibility for aged care services in Australia. These ACAT assessments were undertaken for people seeking permanent or respite residential aged care, HCPs, short-term restorative care or transition care support services. The ACAP dataset contains information on the assessor, service approvals, and aged care seeker. The HCP and residential aged care datasets provide details on the dates (i.e. entry and exits) and levels of services people received. The NDI provide the mortality status for this national cohort.

Our cohort includes people aged 65 years old and older or 50 years old and over for Aboriginal and Torres Strait Islander peoples that had approval for and entered a HCP between July 1st, 2003 and June 30th, 2013 (N=178,924) as the first service received. Individuals who had a wait time of more than 5 years for a HCP and those who may have died whilst waiting or never took up a HCP despite approval were not included.

Wait time from the ACAT assessment approval to the beginning of a HCP was the main exposure of interest. For people who had several ACAT assessments, we considered the assessment closest to the date of beginning of HCP as the one for this study. Wait time was calculated as the difference in the time between the date of approval and the date of first access to a HCP. The study cohort was divided into four groups (quartiles) according to their wait time and this equated to 0–30 days, 31–59 days, 2–6 months days and >6 months.

Mortality after beginning to receive a HCP was the primary outcome of interest. Other outcomes of interest were entry into permanent residential aged care after beginning a HCP and the risk of mortality after transition to permanent residential aged care. Mortality was assessed for the cohort between July 1st, 2003 and June 30th, 2015, yielding a minimum 2 year follow up for the entire cohort, entry into permanent residential aged care was assessed between July 1st, 2003 and June 30th, 2014, yielding a 1 year minimum follow up.

Individuals’ sociodemographic characteristics, living situations, activity limitations, health conditions and geriatric syndrome conditions were evaluated. Specific variables evaluated were: age, sex, English proficiency index/migrants level of English proficiency derived from country of birth (0=Australian born, 1= Countries rating ≥98.5% on the English Proficiency index, 2= Countries rating ≥84.5%, 3= Countries rating 57.5% to less than 84.5%, 4= Countries rating less than 57.5%)(12), country of birth (Australia or born overseas), indigenous status (Aboriginal/Torres Strait Islander or neither), Department of Veterans’ Affairs Card Status (no card, gold card, other), remoteness location (major city vs. other), state, living arrangements (lives alone or with someone), usual accommodation (private owned, retirement village, hospital, short term temporary referring to individuals sheltered in shortterm crisis or emergency or transitional or public places), activity limitations (communication, domestic assistance, health care tasks, home maintenance, meals, movement activities, self-care, social and community participation, transport, moving around places at or away from home), priority category (within 48 hours, between 3 and 14 days, more than 14 days, no priority), other approval (permanent, respite or transition care), HCP entry care level (CACP, EACH or EACH-D), financial year, recommended government assistance services (community aged care, home and community care, veterans’ community aged care service), carer availability, carer co-residency status, carer gender, geriatric syndromes (falls, fractures, delirium, dementia, depression), other health conditions (hypertension, arthritis, diseases of the skin and subcutaneous tissue, cancer, diabetes, diseases of the eye, incontinence, malnutrition, deafness/hearing loss, osteoporosis, chronic lower respiratory diseases, kidney & urinary system disorders, chronic lower respiratory diseases), eligibility assessors’ professional training (medical vs. nursing vs. health vs. social welfare), and current assistance, source of current assistance, and recommended assistance for limitations.

First, the data was cleaned, coded, checked for distributions and then wait time was computed and described by each independent variables. Graphs describe wait time variations by age, gender, financial year and state. To identify factors associated with wait for a HCP, individuals were grouped by quartiles of wait time as 0–30 days vs. 31–59 days vs. 2–6 months vs. > 6 months. Second, a multinomial logistic regression model was fitted using quartiles of wait time as the dependent variable and covariates as a predictor to characterise cohorts according to their wait times. A stepwise variable selection approach was used to determine factors associated with wait time and the best model was determined using the Akaike Information Criterion (AIC). To assess the impact of wait time on mortality, univariate survival analysis was performed followed by a multivariable survival analysis (Cox regression) estimating the risk of mortality in older people who waited for 31–59 days, 2–6 months or > 6 months compared to those who waited 0–30 days. A Cox regression model was also used to examine the effect of wait time on risk of entering residential aged care permanently after a HCP, and evaluate on mortality risk once permanent residential aged care permanently. The findings are described using survival estimates (95% confidence intervals (CI)), Kaplan Meier curve (KM) plots and hazard ratios (HR: 95% CI). Proportional hazard assumptions were evaluated using a scaled Schoenfeld residuals plot and Schoenfeld test, and time-dependent effects were calculated when assumptions were not met (i.e. for the mortality estimates). Cox regression models were adjusted for covariates related to individual’s sociodemographic and personal characteristics known to be associated with wait time (i.e. were known to be associated with wait times). Both adjusted and unadjusted effect estimates (95% CIs) were calculated for estimates of mortality, risk of transition to permanent residential aged care and risk of mortality after entry to permanent residential aged care. Sensitivity analyses were conducted to determine if HCP entry care level (CACP, EACH or EACH-D) was an effect modifier (i.e. interaction) in each of the models. All tests were two-sided and alpha=0.05 was considered statistically significant. The analysis was performed using R programming language version 3.5.1.

We analysed data from 178,924 older Australians who received a HCP between 2003 and 2013. The mean (SD) age of the cohort was 81.6 (7.0) years. The majority were female (65.2%), born in Australia (66.4%), living in major cities (67.3%), and residing in a private or rental accommodation (87.5%), (Table 1 and Supplementary Table 1). 5855 people entered the newer HCP levels of packages, Level 1 and Level 2 and this was equated to a CACP. 32 people received Level 3 and level 4 packages, and this was equated to EACH/EACH-D.

Of the 178,924 older people who received a HCP as their first service during the study period, 33.2% of them received services with 30 days, 41.2% between 30 days to 6 months and 25.6% after 6 months. The overall median wait time was 62 (IQR=21-187) days. Wait time to obtain HCP varied by sex, country of birth, indigenous status, living arrangements, current accommodation, the existence of activity limitations and health conditions, across years and by state (Table 1, Supplementary Table 1, Supplementary Figure 1a-e). Of the 142,755 people who were approved for CACP packages, 137,790 (96.5%) received this package at entry. Among people who were approved for EACH (N=15,712), 4030 (25.6%) received a lower level CACP and of the 7087 approved for EACH-D, 931 (13.1%) received a lower level package initially (Supplementary Figure 1a).

Wait time was associated with age, sex, country of birth, whether an individual lived with someone or alone, owned their accommodation, had an activity limitation (communication, domestic assistance, health care tasks, meals, self-care, social and community participation, transportation, and social activity), and health conditions (Table 2). Compared with people who waited 0–30 days (OR, 95%CI), individuals who waited >6 months were more likely to be born overseas (1.07, 1.04-1.10); live with family (1.65, 1.61-1.70) or others (1.36, 1.24-1.49) compared to living alone; from ACT (1.31, 1.17-1.47), NSW (1.52, 1.45-1.6), TAS (1.46, 1.33-1.61),VIC (1.09, 1.04-1.15) compared to being in the state of SA; and have approvals also for residential aged care (1.11, 1.08-1.15), respite care (1.38, 1.33-1.43), transition care program (4.34, 3.94-4.78). Older people were alternatively less likely to wait longer than 6 months if they were male (0.90, 0.87-0.92); lived in retirement villages (0.75, 0.72-0.79) or short-term/ temporary accommodation (0.75, 0.64-0.88) compared to owning a private accommodation; had activity limitations for communication (0.91, 0.88-0.94), domestic assistance (0.85, 0.79-0.92), health care tasks (0.87, 0.85-0.90), meals (0.77, 0.74-0.80), self-care (0.67, 0.65-0.69), social and community participation (0.88, 0.84-0.91), or transport (0.84, 0.8-0.87); had medical health conditions such as cancer (0.85, 0.82-0.88); and geriatric syndromes such as falls (0.89, 0.86-0.93) or delirium (0.65, 0.54-0.78). Similar trends of associations were observed between the individual’s characteristics and whether they waited 31–59 days or 60 days-6 months (Table 2).

The cohort was followed for 4.0 years (interquartile range (IQR:1.8-7.2)) and 38% were alive at the end of the study period (Supplementary Table 2). Individuals waiting >6 months had an 18% higher risk of death after 2 years of starting their HCP, when contrasted to individuals receiving a HCP within 30 days (adjusted hazard ratio (aHR, 95%CI=1.18, 1.16-1.21)). Those who waited 31–59 days had a 6% higher mortality risk (1.06, 1.03-1.09) after 2.5 years of entry and the risk of mortality for those who waited 2–6 months was a 4% higher (1.04, 1.00-1.08) when compared to those who waited < 30 days after five years (Table 3a, Figure 1a). The risk of mortality following transition into permanent residential aged care for people who waited > than 6 months was 8% higher (1.08, 1.05-1.10) compared to people receiving HCP within 30 days. Similar associations were noted for the groups who waited 2–6 months (1.03, 1.01-1.05) and 31–59 days (1.03, 1.01-1.05) (Table 3b, Figure 1b).

Of the 178,924 older people who had a HCP, 92,987 (52.0%) transitioned to a permanent residential aged care service. The median time (IQR) of transition was 2.8 (1.0-8.0) years. The cumulative incidence of transition, % (95%CI), from a HCP to permanent residential aged care service at 30 days, 3 months, 6 months, 1 year, 2 years and 5 years was 1.0 (1.0-1.0), 6.0 (6.0-6.0), 14.0 (13.0-14.0), 26.0 (26.0-26.0), 42.0 (41.0-42.0) and 65.0 (64.0-65.0), respectively (Supplementary Table 3, Figure 2). After 2 years of stay in a HCP, people who waited > 6 months had 10% (aHR, 95%CI = 1.10, 1.06 -1.13) higher risk of transition to permanent residential aged care. The risk for those who waited for 2–6 months and 31–59 days was higher by 7% after 2 years (1.07, 1.04-1.11) and 4% after one year (1.04, 1.02-1.07) compared to individuals receiving the HCP within a month (Table 3c).

HCP entry care level was not an effect modifier for risk of mortality (in both models) and transition into permanent care (data not shown), except in the case of people who were approved for EACHD and waited for >6 months (N=2679) when compared to <=30 days (N=2756). Within this group the risk of mortality after entry into HCP was higher within 2 years (aHR=1.13, 95%CI= 1.03-1.24), which was not observed in the other groups where the risk was only observed 2 years after entry into HCP.