Client characteristics and outcomes of the Australian short-term restorative care programme: a cohort study

Abstract

Background

Restorative care is a goal-oriented, time-limited, multidisciplinary approach to address functional decline in older adults. Within Australia, one form of restorative care available to community-dwelling older adults is the Short-Term Restorative Care (STRC) Programme. Australian government expenditure on such services is high, yet research on programme outcomes is scarce. The primary aim of this study was to provide a descriptive analysis of STRC participant and outcome data from clients’ first STRC episode with a large, aged care provider in New South Wales (NSW), Australia. Secondary aims were: (i) examine any associations between client demographic and outcome measures routinely collected, and (ii) consider the suitability for broader sector use of the clinical and outcome measures routinely collected by the provider.

Methods

A retrospective, cohort study of STRC client data routinely collected between 1 March 2021 and 28 February 2023 from a large, not-for-profit aged care provider in NSW, Australia. Client demographic data and routinely used clinical and outcome measures (modified Barthel Index, Mini Nutritional Assessment short form, Clinical Frailty Scale, Patient-Reported Outcomes Measurement Information System 10-Question short form - global health, and Goal Attainment Scale light) were collected. Descriptive and inferential analyses were completed to explore available data.

Results

Four hundred and eighty-four STRC clients were included in this study. The mean age of clients was 81.5 (7.6) and majority were female (56%), born in Australia (49%), had a normal nutritional status on entry (54%), and classified as either pre-frail or mildly frail on entry (66%). The modified Barthel Index, Patient-Reported Outcomes Measurement Information System 10-Question short form - global health, and Goal Attainment Scale light all showed statistically significant improvements. Almost all clients remaining in their own home after exit of the STRC Programme (99%).

Conclusions

This study looked at detailed data from the STRC Programme in Australia with a large sample over multiple years from a single aged care provider, and the findings suggest it is an effective program for older adults. The data provide insights into different clinical and outcome measures used, identify those that work well with this population, and those that need further consideration.

Background

Population ageing has led to higher numbers of older adults living with multimorbidity and frailty, resulting in increased requirements for community aged care services [1]. However, aged care services are often reactive, episodic and fragmented, and may not address the underlying source of the functional decline for the services being requested [2, 3]. Examples include traditional home care services that lack a focus on optimising independence, and acute care models that contribute to loss of functional ability through prolonged bed rest [4]. To address these issues, along with the preference of many older adults to stay at home for longer, and the high costs associated with hospitalisation and residential care, government policies in many countries (e.g. Australia, New Zealand, England, Sweden, Norway, Netherlands) have begun focussing on person-centred, early intervention and preventative services [5-7].

Restorative care (a term commonly used in Australia, New Zealand, and the United States of America), often used interchangeably with reablement (a term commonly used in the United Kingdom, Ireland and Denmark), is a goal-oriented, time-limited, multidisciplinary approach to address functional decline in older adults [3, 8]. The intention is to enhance performance of Activities of Daily Living (ADLs), independence, and social participation [8]. Most restorative care programs take place within the home setting and tend to be of longer duration and lower intensity than health-based rehabilitation programs [3]. Supplementary program aims include reducing the need for ongoing community aged care services and premature residential aged care admission, and addressing broader social and psychological needs [3]. In contrast, rehabilitation programs often occur after an acute hospital admission/significant health event within a specialised facility and at a higher intensity [3].

Within Australia, community-dwelling older adults have access to a number of program types that could be deemed ‘restorative care’. Eligibility for these different programs is dependent on a person’s circumstances, such as existing service arrangements and whether or not their functional decline has resulted in a hospitalisation. Australian restorative care programs for older adults include: the Short-Term Restorative Care (STRC) Programme; the Transition Care Programme (TCP); ‘allied health and therapy’ provided under the Commonwealth Home Support Programme (CHSP); and the provision of allied health support within a Home Care Package (HCP) [3, 9]. Australian government expenditure on restorative care and reablement services for older adults is estimated to exceed $640 million annually: expenditure in FY 2021-22 on STRC specifically was over $74.6 million; on TCP over $282.3 million; and ‘allied health and therapy’ accounted for over $283.6 million (9.3%) of CHSP expenditure [10]. These amounts are expected to continue to rise as the population ages and the Australian Government directs a larger proportion of aged care funding towards home-based services [11].

The modified Barthel Index (mBI) [12] is a mandated outcome measure completed with every client when they enter and exit some government subsidised restorative care programs, such as the STRC Programme [13]. In the case of STRC specifically, the mBI is the only mandated outcome measure. Aged care providers may (but are not required to) collect additional internal clinical assessment and outcome data to help monitor the effectiveness of the restorative care programs they deliver. If aged care providers do routinely collect other outcome measures (in additional to the mBI) then these data are not available to the aged care sector as they are not reported. In addition, there is likely to be a wide variation in any tools (in terms of type and suitability) used by different aged care providers, which would make benchmarking difficult in any case. Data on the effectiveness of restorative care programs in Australia are therefore limited [14], despite the significant government funding underpinning the STRC Programme. Globally, research on the effectiveness of restorative care is also scarce and results are conflicting [15]. Given high government expenditure on these community-based care services and the preference of many older adults to remain in their own homes, comprehensive analysis of the outcomes achieved by restorative care programs is warranted [6, 7].

The primary aim of this study was to provide a descriptive analysis of STRC participant and outcome data from clients’ first STRC episode with a large, aged care provider in New South Wales (NSW), Australia. Secondary aims were: (i) examine any associations between client demographics and the available outcome measures routinely collected by the provider as part of the STRC Programme, and (ii) consider the suitability for broader sector use of the clinical and outcome measures routinely collected by the provider. It is hypothesised that clients who completed the STRC Programme will have demonstrated a functional improvement, based on positive changes in the outcome measures routinely collected.

Methods

Design and setting

A retrospective, cohort study of STRC client data routinely collected between 1 March 2021 and 28 February 2023 from a large, not-for-profit aged care provider in NSW, Australia. The regions of NSW covered by the aged care provider’s STRC team were Northern Sydney, South Eastern Sydney, South Western Sydney, and Inner West Sydney. This study was approved by the St Vincent’s Hospital Human Research Ethics Committee (reference 2023/ETH02083).

Participants

Older adults (aged 55 years and older) who completed their first STRC episode with the aged care provider between 1 March 2021 and 28 February 2023. Clients were excluded if they left the program early (e.g. due to hospitalisation, passed away) as they would not have completed the STRC outcome measures on discharge, limiting data available for analysis. In addition, clients that had previously completed a STRC episode with the aged care provider prior to the study period were excluded to maintain data consistency by focusing the descriptive analysis on the first STRC episode only.

Intervention

The STRC Programme is a 56-day community-based program available to older adults with potentially reversible functional decline [3, 9]. Access to the STRC Programme requires assessment and pre-approval from an Aged Care Assessment Team (ACAT) [9]. Health professionals and services that can be accessed as part of the STRC Programme include Occupational Therapists (OTs), physiotherapists, nursing support, Exercise Physiologists (EPs), dieticians, personal carers, provision of equipment and in-home technology to help with daily activities, and minor home modifications [13]. There was consistency in the ‘dollar value’ of the eight -week STRC Programme per client, as per STRC Programme guidelines. The mix of services delivered per client was individualised based on the client’s assessed need. Table 1 provides a detailed description of the interventions provided by the aged care provider as part of the STRC Programme, the approximate duration of each intervention, and mean count of sessions per AHP per STRC episode. The aim of STRC is to prevent hospitalisation and delay the need for more intensive aged care supports such as a HCP or residential care [13]. Older adults are ineligible if they have been discharged from hospital in the previous three months for a condition related to their STRC needs, they are receiving or had an episode of TCP within the last six months, are currently receiving a higher level HCP (levels 3 or 4), receiving end-of-life care, or had previously received two STRC episodes in a 12-month period [13].

Table 1 Aged care provider’s services (including approximate duration)

During the study period the aged care provider’s STRC Programme was attended by a multidisciplinary team which consisted of physiotherapists, OTs, EPs, dieticians, and Allied Health Assistants (AHAs). If other disciplines (e.g. nurse, psychologist, podiatrist) were required for an individual’s STRC Programme, this was sourced via other teams within the aged care provider or from outside agencies. The input of any other additional disciplines was still managed by the STRC coordinator, who was also an Allied Health Professional (AHP). All interventions were provided face-to-face within the client’s home. During the study period the aged care provider had capacity for 62 approved STRC places at any one time, with an average occupancy rate of 70% across both years. Both study years were impacted to a variable degree by the Covid-19 pandemic, which resulted in lower than usual occupancy.

Procedure

Retrospective data (demographic and outcome variables) were extracted from the provider’s electronic client record system using a study-specific reporting service developed by the aged care provider’s information technology team. The start date of the study period was predetermined based on the date from when the included clinical and outcome measures were implemented into routine practice as part of the aged care provider’s STRC Programme. The end date was set to ensure all relevant data had been inputted into client records prior to data extraction to reduce the amount of missing data. Extracted data were checked against the selection criteria, coded, and identifiable markers removed. Prior to analysis, study data were cross-checked manually by two members of the research team and two members of the STRC team. Clinical and outcome measures with high levels of missing data (20% or more) were further analysed by review of clients’ initial ACAT assessments, and feedback collected from STRC clinical care managers to identify any contributing factors towards the missing data.

Clinical and outcome measures

Routinely used STRC clinical measures (assessed on entry to the STRC Programme only) collected for this study were the: Mini Nutritional Assessment Short Form (MNA-SF); and Clinical Frailty Scale (CFS). Routinely used STRC outcome measures (assessed on entry and exit of the STRC Programme) collected for this study were the: mBI; Patient-Reported Outcomes Measurement Information System 10-Question Short Form (PROMIS10 Global Health); and Goal Attainment Scale (GAS) Light.

The MNA-SF is a screening tool used to identify older adults who are malnourished or at risk of malnourishment, supporting early nutritional intervention when needed [16]. Scores range from zero to 14, with a lower score indicating malnourishment or risk of malnourishment [17]. The MNA-SF categories based on the score are: Malnourished (0–7), at risk of malnutrition (8–11), and normal nutritional status (12–14) [17]. It is a validated screening tool for use with older adults in a variety of settings [16, 17]. The MNA-SF is used by the aged care provider’s STRC teams when a client enters the STRC Programme to prompt appropriate and timely referrals to the dietician, if the client agrees. It is important to note that a low MNA-SF is not the sole requirement for the aged care provider’s STRC team to refer to the dietician; they may also refer to the dietician based on other relevant information such as their clinical judgement and client input.

The CFS is a frailty screening tool used to quantify a person’s overall health status [18]. Scores range from one to nine, with a higher score indicating higher levels of frailty [19]. Each point represents a different category: Very fit (1), well (2), managing well (3), pre-frail (originally referred to as vulnerable to fragility) (4), mildly frail (5), moderately frail (6), severely frail (7), very severely frail (8), and terminally ill (9) [19]. It is a validated tool for use with older adults and is strongly correlated to the Frailty Index [18]. The CFS is completed by the aged care provider’s STRC team when a client enters the STRC Programme to obtain an overall measure of frailty, with this information used in the formulation of the STRC care plan. The CFS was selected by the STRC team as they considered it a suitable clinical assessment tool to identify frailty and thus help inform the content of the STRC Programme. It was also selected as it provides a simple classification, allowing greater insights into the nature of the clients referred to STRC, and because the tool is easy to administer, without any additional burden on clients.

The mBI assesses a person’s ability and level of independence when completing 10 activities of daily living [20]. Scores range from zero to 100, with a higher score representing greater levels of independence [12]. Categories to support score interpretation are: Total dependency (0–20), severe dependency (21–60), moderate dependency (61–90), slight dependency (91–99), and independent (100) [12]. The mBI is recommended for use with people with stroke, Parkinson’s disease, brain injury, cancer, people admitted to intensive care units, and older adults [21]. Aged care providers are required to complete the mBI with all clients at entry to and exit from the STRC Programme [9].

The PROMIS10 Global Health is an outcome measure that assesses self-reported (or proxy when needed) global function and well-being in the physical, mental, and social domains of health, producing two summary scores (range 0–20) and t-scores (mean of 50 and standard deviation of 10): global physical health and global mental health [22, 23]. A higher score indicates greater self-perceived global physical/mental health [24]. The t-scores for each subdomain can be interpreted based on the following categories, though they each have different cutoff scores: Poor (global physical health < 35, global mental health < 29); fair (global physical health 35–41, global mental health 29–39); good (global physical health 42–49, global mental health 40–47); very good (global physical health 50–57, global mental health 48–55); and excellent (global physical health > 57, global mental health > 55) [22]. Studies have shown it to be a valid tool among the general population as well as those with chronic conditions [22]. The PROMIS10 Global Health measure is completed by the aged care provider’s STRC clients on entry and exit of the STRC Programme. The PROMIS10 Global Health measure was selected by the STRC team as it they wanted an overall measure of perceived client global physical health as well as their perceived mental health and provide insights into their psychosocial needs.

The GAS Light is a mathematical technique used to identify and quantify the achievement of client-centred goals [25]. The GAS Light uses a 5-point scale from − 2 to + 2, and an aggregated t-score (mean of 50 and standard deviation of 10) allows for the assimilation of several goals into one overall score and comparison between individuals [25]. The three categories used to interpret the GAS Light t-score are: all goals not achieved to expected level (< 50); all goals achieved to the expected level (50); and all goals achieved to a better than expected level (> 50) [25]. The GAS Light has shown to be a valid and reliable tool when used with older adults, including those with chronic conditions, in primary care, residential care, and inpatient settings [26-28]. Each AHP involved in the care of the STRC client completes at least one GAS Light goal with the client when they enter the STRC Programme, which is reassessed on exit of the STRC Programme. This outcome measure was implemented by the aged care provider to support clients to set appropriate goals with the support of the STRC team. It was also felt that the measurement of goal attainment was more client-specific than global outcome tools, and therefore a useful additional tool.

Data analysis

Descriptive analyses were used to detail the client characteristics and STRC outcomes of study participants, as well as any contributing factors to high levels of missing clinical and outcome measure data. Data were not normally distributed, therefore non-parametric tests were used for further analysis. The Wilcoxon signed-rank test was used to compare median change in the mBI, PROMIS10 - physical health, and PROMIS10 - mental health from entry to exit of the STRC Programme. Median changes in the outcome measures were also compared between groups based on client demographic variables using the Mann-Whitney U test (gender, English as first language, having a carer) and Kruskal-Wallis test (age category, marital status, region of birth, living arrangements, discharge destination). Spearman’s correlation was used to check for associations between all of the clinical and outcome measures (MNA-SF and CFS scores at entry, and change in scores from entry to exit for the mBI, PROMIS10 – physical health, PROMIS10 – mental health, and GAS-light t score). The study data did not meet all requirements to run multiple linear regression. To allow for comparison to the available national data, the mean score for the mBI has also been reported. Data analysis was conducted using SPSS Version 27 (IBM Corp. in Armonk, NY). The significance level of all tests was set at p-value < 0.05.

Results

Client demographics

Four hundred and eighty-four STRC clients were included in this study (Fig. 1), their demographics are detailed in Table 2. The mean age of clients was 81.5 (SD 7.6) and the majority were female (56%), born in Australia (49%), married (55%), and lived with their partner (44%). After completion of the STRC Programme 97% of clients remained living in their own homes. OTs and physiotherapists were the most common AHPs who provided services as part of the STRC Programme during the study period.

Fig. 1

Participant flow chart. *Reasons clients requested to finish early: no reason provided (n = 2),travel plans (n = 1), anxiety issues (n = 2),concerned about COVID-19 (n = 3), no longer interested in STRC (n = 1), carer responsibilities (n = 1)

Table 2 STRC client demographics for study population (n = 484)

MNA

Clients were most commonly (54%) categorised as having a normal nutritional status on entry (Table 3). When MNA data were considered alongside AHP involvement: 79% (n = 37) of clients identified as being ‘malnourished’; 66% (n = 100) of clients identified as ‘at risk of malnourishment’; and 39% (n = 91) of clients identified as having a ‘normal nutritional status’, received dietician input during the STRC Programme.

Table 3 STRC clinical and outcome measures for study population (n = 484)

CFS

Majority (60%) of the clients were considered frail (mild, moderate, and severe categories combined), and approximately one third (33%) were considered pre-frail (Table 3). No clients were classified as very fit (lowest score), very severely frail or terminally ill (the two highest scores).

mBI

The mBI demonstrated a median improvement in total score of 2.5 (range − 28 to 46) by the end of the STRC Programme, which was statistically significant (p < 0.001) (Table 3). For the majority (65%) of clients this statistically significant improvement did not result in a change of category. Clients who had a carer demonstrated a significantly greater change in median mBI than those who did not (3 [-28 to 46] and 2 [-14 to 32] respectively, p < 0.001). Supplementary Table 1 provides additional descriptive data based on carer status. No other between-group differences (based on gender, English as first language, age category, marital status, region of birth, living arrangements, discharge destination) in median mBI were significant. The mean mBI score on entry was 87 (14), which increased to 90 (13), and had a mean change of 1 (18) (mBI data was not normally distributed, mean only reported to allow for comparison to national STRC data).

PROMIS10 – physical health and mental health subscales t-scores

Both the PROMIS10 – physical health and mental health subscales demonstrated significant improvements (p < 0.001 for both) in t-score by the end of the STRC Programme; median improvements of 3.2 (range − 13.4 to 21.7), and 2.5 (range − 12.6 to 33.9) respectively (Table 3). This statistically significant improvement in score was associated with the following changes in category: for physical health, 48% improved in category, 41% experienced no change, and 10% deteriorated; for mental health, 37% improved, 48% experienced no change and 15% deteriorated. A statistically significant greater change in the PROMIS10 (mental subscale) was observed in clients who did not have a carer compared to clients who did have a carer (2.5 [-12.6-17.4] and 0 [-9.7-33.9] respectively, p 0.035), and in clients whose first language was not English compared to clients whose first language was English (7.3 [-9.7 to 19.9] and 2.5 [-12.6 to 33.9] respectively, p 0.002). Supplementary Tables 1 and 2 provide additional descriptive data based on carer status and English as first language respectively. There were no other significant between-group differences (based on gender, age category, marital status, region of birth, living arrangements, discharge destination) observed.

Missing data were notably higher for both subscales of the PROMIS10 - Global Health measure (21% on entry and 41% on exit for both subscales) than the other clinical and outcome measures (ranged from 0.6 to 11%). This was investigated to identify potential contributing factors. The presence of a cognitive impairment (41%) or language barrier (3%) was identified in the analysis of initial ACAT assessments of clients with missing PROMIS10 Global Health measure data (Table 4). This was further supported by feedback collected from 14 STRC clinical care coordinators.

Table 4 PROMIS10 global health measure missing data review

GAS-light t-score

The median GAS-light t-score on exit was 50 (range 30-74.8) (Table 3). On average each client, in partnership with their treating STRC AHPs, set 2.5 (SD 1) goals, examples are shown in Fig. 2. Goals were achieved (either at the expected, or better than expected level) by 68% (n = 318) of clients (Fig. 3). No significant differences between groups (based on gender, English as first language, having a carer, age category, marital status, region of birth, living arrangements, discharge destination) were observed.

Fig. 2

Examples of GAS-light goals set with STRC clients during study period

Fig. 3

Proportion of clients per GAS-light category on exit of STRC Programme (n = 470)

Correlation between clinical and outcome measures

Spearman’s correlation demonstrated positive associations between the change in scores of all outcome measures that were assessed at entry to and exit from the STRC Programme (mBI, PROMIS10 – physical health and mental health subscales, and GAS-light t score). All were significant, strong correlations except between the mBI and PROMIS10 – mental subscale (Table 5). Significant, strong correlations were also observed between higher CFS (more frail) on entry and change in mBI (r_s[452] = 0.294, p < 0.001) and between lower CFS (less frail) on entry and higher MNA (better nutritional status) on entry (r_s[421] = − 0.251, p < 0.001).

Table 5 Correlation matrix of STRC clinical and outcome measures

Table 6 outlines the agreement on categorisation (improvement and no improvement) between GAS-light and other outcome measures completed on entry to and exit from the STRC Programme (mBI, PROMIS10 – physical subscale, PROMIS10 – mental subscale). For the GAS-light, the categories all goals achieved to expected level and all goals achieved to higher than expected level were classed as an improvement; the category all goals not achieved to expected level were classed as no improvement. For the other three outcome measures, a change to a higher category was classed as an improvement; no change in categorisation or change to a lower category was classed as no improvement. The greatest lack of agreement was when clients were classified as showing an improvement on the GAS-light but not on the mBI.

Table 6 Agreement between GAS-light and other outcome measures on exit of the STRC programme

Discussion

This study provides a descriptive analysis of characteristics and outcomes of a cohort of older adults in Australia who had undergone a government funded, community-based, restorative care programme (the STRC Programme), delivered by a multidisciplinary allied health team. The study presents in-depth, real world data from a large, Australian aged care provider’s STRC Programme. Published STRC data from the Australian Government Department of Health and Aged Care’s Aged Care Data Snapshot − 2022, provides only limited information on STRC: basic national and state-based data (gender, age, Aboriginal and Torres Strait Islander status, cultural and linguistic diversity, remoteness area, and mBI on entry and exit of the STRC Programme) [10]. Where comparative data are available, the findings in this study are broadly consistent with those in the national data: females predominated (56% in this this study compared to 62% in the national data); English was participants’ first/native language (70% and 77% respectively); and the majority of participants had a mean mBI score that categorised them as moderately dependent on both entry into and exit from the STRC Programme [10]. However, the STRC population in this study were marginally older (greater proportion of clients aged 85 and older compared to greater proportions aged 84 and younger in the national data); had slightly higher functioning according to the mBI (mean mBI score of 87 on entry and 90 on exit compared to 80 and 84 respectively in the national data); and 100% of participants in the present study lived in a major city (by virtue of the geographical spread of the aged care provider) compared with 72% in the national data [10].

Along with a greater understanding of client characteristics and content of the Australian STRC Programme, the data from this study also provided some insights into the potential range of additional (i.e., beyond the mBI) clinical and outcome measures that may have utility with this population, guiding future research to better understanding the effectiveness of the STRC Programme. The aged care provider’s STRC team complete additional, non-mandatory clinical and outcome measures as part of routine care (detailed in the Methods section above) to provide a better understanding of their clients and of programme efficacy, identifying opportunities for improvement. An example is the MNA-SF that is assessed on entry to the STRC Programme (noting that the measure is a screening tool and reassessment of community-dwelling older adults is advised annually) [17] not only to provide baseline information on clients’ nutritional status, but to also prompt clinicians to discuss with the client the need for a referral to a dietician member of the STRC team. The fact that only 66% of those identified as malnourished received dietician input during the STRC Programme requires further evaluation, in terms of staff training in the use and utility of the tool, the acceptance by clients that dietician referral could be of benefit, and potential issues with the availability of a dietician during the STRC Programme. While data to this effect were not in-scope for this study, it is important going forwards for healthcare teams collecting clinical data routinely to fully understand the utility of the tools used and barriers to fully actioning the data thus derived.

The other clinical measure only assessed upon STRC Programme entry was the CFS. Similar to the MNA-SF, it was used by the aged care provider’s STRC team to guide treatment plans. Specifically, it supported the early identification of older adults who were frail or at risk of frailty; an important consideration in light of the association between frailty and high rates of adverse outcomes and increased healthcare utilization [29, 30]. In this study, the prevalence of frailty (60%) was slightly higher and pre-frailty (33%) similar to that reported in another Australian study that was conducted with older adult who received a HCP (details on participants’ package level were not provided) in rural NSW: based on the Edmonton Frail Scale, 45% of participants were classified as frail (mild, moderate or severe) and 32% pre-frail [31]. In contrast, a third Australian study that used the modified Fried Frailty Phenotype to conduct a secondary analysis of a national cohort of community-dwelling older adults (no details were provided on the use of community aged care services by participants) reported a frailty prevalence of 21% and pre-frail prevalence of 48% [32]. The different in frailty and pre-frail prevalence across all three studies is likely accounted for by differences in the study populations. This study and the one conducted in rural NSW were focused on older adults who received some form of community aged care service, STRC and HCP respectively [31]. The national study, in contrast, included community-dwelling older adults who may or may not have received any community aged care services, and therefore could potentially be more highly functioning. The variability in frailty and pre-frail prevenance across all studies could also be, in part, due to the variation in frailty assessment tools used [30, 33].

While standardisation in the choice of frailty assessment tool would be helpful for benchmarking purposes, it is more important for providers of restorative care programs (such as the STRC Programme) to accurately identify at program commencement those older adults who are showing features of frailty, or who are at risk of frailty. Allied health programs can then be better designed to begin to address the physical sequelae of frailty, especially those resulting from sarcopenia, such as muscle weakness, reduced mobility and falls risk [34]. One concern with the STRC Programme itself is its limited duration, being of only 56 days, which in terms of adequately addressing frailty may be too short [35, 36]. Though one study that examined the effectiveness of the STRC Programme in South Australia did find frailty was reduced on completion, as observed by improved scores in the Vitality Questionnaire [14]. The STRC Programme could support changes in client behaviour through the introduction of sustainable dietary modifications (such as ensuring adequate protein intake) and self-efficacy when it comes to ongoing exercise adoption [35, 37]. Further research into the longer term benefits of episodic restorative care programs (such as the STRC Programme) are required to determine whether client benefits, especially when it comes to frailty, continue to accrue over time.

The mBI is the only mandatory outcome measure to be completed at entry to and exit from the STRC Programme, and it is required in order for the aged care provider to claim the government STRC subsidy [12, 13]. However, a concern related to mBI is the well-known ceiling effect, which limits it’s suitability to measure meaningful clinical change (as opposed to statistically significant change) in more able individuals [38, 39]. Older adults who participate in the STRC Programme could be considered more able as eligibility criteria includes no recent hospital admission related to the STRC need, and they are not so disabled as to be receiving a higher level HCP (levels 3 or 4; noting that a person may have been assessed as being eligible for a Level 3 or 4 HCP, but are on a waiting list for receipt of the higher HCP, thus still being eligible for STRC). This is supported by the findings of this study which showed the majority of the study population did not have a carer (64%) and scored in the second highest category of the mBI (slight dependency) on entry (46%); and 5% (n = 23) were already achieving the highest possible score for the mBI (100 = independent) on entry.

Nonetheless, and similar to all other outcome measures routinely used by the aged care provider’s STRC team at entry to and exit from the STRC Programme in this study (PROMIS10 global health measure – physical and mental subscales, and GAS-light), the mBI still showed a statistically significant change in score over time, which positively correlated with the PROMIS10 global health measure (physical and mental subscales), and GAS-light. Similarly, an Australian study that conducted an observational study of retrospective STRC Programme data found a statistically significant improvement in the mBI median score from 78.5 to 95 [14]. However, the mBI category did not change for 65% of the STRC clients after they participated in the STRC Programme in this study, suggesting no clinically significant change. The high proportion of clients who scored in the top two categories of the mBI at the start of the STRC Programme potentially contributed to the low sensitivity of the mBI in detecting clinically significant changes (i.e. change in functional category) in this study population. Considering this is the only mandatory outcome measure for the STRC Programme, issues with the sensitivity and ceiling effects of the mBI thus limit its effectiveness as a tool to effectively benchmark the outcomes of the STRC Programme. Greater consideration of what domains of health and functioning are most impacted by the STRC Programme, including the domains most important to clients, are required. In the regard to the latter, this is where Patient-Reported Outcome Measures (PROMs) may have a role.

The PROMIS10 global health measure was one of the non-mandatory PROMs used by the aged care provider’s STRC team. Both the PROMIS10 physical subscale and PROMIS10 mental subscale showed statistically significant changes over time (median change of 3.2 and 2.5 respectively). However, the Minimally Important Change (MIC) is often considered more relevant than the Minimal Clinical Important Difference (MCID) for this measure, as instead of based on statistically detected change, the MIC is a within-client change over time that is considered important to the client [40]. A systematic review of the PROMIS outcome measures reported a MIC for non-surgical clients as two to six points [40]. Whilst consensus from a PROMIS leadership meeting suggested a threshold of three points [41]. There is growing support for the use of PROMs to help determine the effectiveness of interventions in practice and research [40], and based on the suggested MIC thresholds, both the PROMIS10 physical health and mental health subscales showed a MIC in this study [40, 41]. These findings might suggest the PROMIS10 global health measure is effective in monitoring change over time for STRC clients.

The ability to measure MIC is not the only factor under consideration when determining utility of an outcome measure. In this study there were high levels of missing data for the PROMIS10 global health measure (21% on entry and 41% on exit for both subscales) which limits the interpretation of these findings. This observation highlights the opportunity for clinicians to use routinely collected data to not only monitor STRC clients, but also monitor suitability of outcome measures routinely used. As the proportion of missing data for both PROMIS10 global health subscales were greater than the 20% threshold set for this study (as outlined in the Methods section above) potential contributing factors were investigated by the research team. After further examination of client records, 44% clients who did not have a PROMIS10 global health measure score on entry and/or exit had either a cognitive impairment or language barrier (Table 4), which may have impacted their ability to complete this outcome measure. These findings were supported by feedback from the STRC care coordinators who reported they were often unable to complete the PROMIS10 global health measure with clients who had a cognitive impairment or language barrier (Table 4). These findings suggest there are issues with the utility of this outcome measure as part of the STRC Programme as it cannot be used routinely with all clients. An alternative outcome measure not limited by an individual’s cognition or language may need to be considered to ensure the STRC team have an understanding of perceived physical and mental health status for all clients.

The use of PROMs can also support appropriate goal setting, for example the GAS(-light), which is integral to restorative care programs [4]. The GAS-light outcome measure is routinely used by the aged care provider’s STRC team, but unlike the PROMIS10 global health measure does not appear to be limited by cognitive impairment or language barriers as suggested by much greater completion rates (3% missing data for the GAS-light compared to 21–41% for the PROMIS10 global health measure). This is further supported by a previous study which showed the GAS-light can be used to set effective treatment goals with older adults living with dementia who participate in rehabilitation and reablement interventions [42]. The main limitation reported with the use of GAS-light, particularly in clinical practice, is the ability of clinicians to effectively set relevant and achievable goals. This can be overcome with ongoing training and support to help maintain the accuracy of goal setting and scaling [43]. In this study the majority (41%) of the goals were achieved to the expected level (t-score of 50), and there was an even distribution of those that were not achieved and over-achieved (32% and 27% respectively). Based on the GAS practical guidelines, this distribution across the categories suggests the aged care provider’s STRC team and clients were able to set appropriate goals in an unbiased fashion [25]. This demonstrated the effective use of GAS-light in a clinical setting and supports ongoing utility of this outcome measure in the STRC Programme.

Statistically significant changes in the scores of the outcome measures used in this study did not always reflect a change in category, as was observed for the majority of the clients for the PROMIS10 mental health subscale (48%) and mBI (65%). This may suggest that even though there was a numerical change in the score, it may have not resulted in an observed/perceived change (as defined by the different categories of the outcome measures) that would be considered meaningful to the client or deemed clinically significant. When compared to the other outcome measures, the GAS-light appeared to be the most sensitive to change: It is determined by what the client considers to be meaningfully important to them [25], and identified the highest (68%) proportion of change in category during the STRC Programme (Fig. 3). The sensitivity (to categorical change) hierarchy of outcome measures in this study after the GAS-light was: PROMIS10 physical subscale; PROMIS10 mental subscale; mBI (Table 3), which was the same order as that for agreement with the GAS-light in identifying clients who improved (Table 6).

An Australian study of older adults who participated in a rural ambulatory rehabilitation service compared the responsiveness of the GAS to the Lawton Scale (assesses clients’ needs in completion of instrumental ADLs) [43]. Similar to this study, they found both showed statistically significant changes over time but the GAS(-light) was more responsive for detecting (perceived) clinically meaningful change [43]. Both studies support the use of GAS(-light) in clinical practice, not only to facilitate client-centred goals but also to assess service effectiveness [43]. Unlike the other outcome measures considered in both studies (PROMIS10 Global Health measure, mBI and Lawton Scale) the GAS(-light) does not have a fixed functional focus. The greater sensitivity of the GAS(-light) could be associated with its ability to measure other factors that are potentially more important to the client, or more relevant to restorative care and rehabilitation that are not picked up in standard outcome measures. The focus of restorative care and rehabilitation is partly to enhance performance of ADLs and independence [8]. Therefore clinicians are more likely to select outcome measures that also have this focus. By including additional outcome measures like the GAS(-light), clinicians have the ability to assess clients more holistically and consider the impact of other aspects relevant to restorative care and rehabilitation, such as social participation and quality of life.

Limitations

The main limitation of this study was associated with missing data. Similar to other studies that relied on retrospectively collected, real world data there were varying levels of missing data for each variable. Most instances of missing data in this study were relatively small; the only one likely to have impacted interpretation of the finding was for the PROMIS10 global health, which had 41% missing on exit from the STRC programme. However, this issue was able to highlight potential limiting factors with this outcome measure in clinical practice with community-dwelling older adults who have a cognitive impairment or language barrier (accounting for almost half of the instances where PROMIS10 data were missing). Though it might limit the interpretation of the results from a research perspective, it is beneficial to identify and address such issues to inform clinical practice, including outcome measure selection. The outcome data suite used by this aged care provider did not include a standalone cognitive screening tool, with the team instead relying on the assessment agency’s (ACAT’s) report that resulted in a referral for STRC. However, ACAT data on client cognitive status proved to be variable. In clinical practice a cognitive screening tool, such as the Mini-Cog or Abbreviated Mental Test Score, would be a beneficial addition to the outcome data suite to ensure consistent data are collected on the cognitive status of clients to help inform care management.

Another limitation was the inability to provide analysis of repeated STRC Programme episodes. There were two key reasons why only first STRC episode data were included: The first was data integrity - there were limited data available from previous STRC participation prior to the study period, thereby making comparison with a subsequent STRC episode less meaningful. The second was that the research team felt that a subsequent STRC episode may have been fundamentally different to the first, and potentially more of a continuation of the first (especially if it were shortly after the first episode) rather than a standalone episode. However, exploring the additional gains of a subsequent episode is an important future study, and may provide information as to the optimal length of an STRC episode. Finally, the generalisability of the findings is also limited as data was only collected from one provider in one state of Australia.

Conclusions

This study looked at clinical and outcome data from the STRC Programme in Australia with a large sample over multiple years from a single aged care provider. The findings suggest that STRC is an effective program for older adults in achieving function gain and meeting individualised goals. The data provide some insights into how the different clinical and outcome measures were used, and the utility of those tools in this population, as well as those, such as the only mandated tool, the mBI, that need further consideration. These findings go beyond informing the practice of the involved provider by strengthening the research base on community-based restorative care for older adults experiencing functional decline more generally. A broader suite of appropriate outcome tools, consistently collected by programme providers, will allow greater opportunities for benchmarking outcomes. Consistent national provider benchmarking will lead to opportunities to further inform policy and practice, and importantly, consumers.