Developing a model for rehabilitation in the home as hospital substitution for patients requiring reconditioning: a Delphi survey in Australia

There was consensus for members of a rehabilitation team being the ones to initially identify patients potentially suitable for RITH (See T2-S1.1). However, it was also agreed that acute care teams, aided by the use of a formal screening tool, could identify potential patients (T2-S1.2). Table 3 lists supported items for use in screening potential patients by acute care teams.

Following the identification of potentially suitable patients, eligibility for RITH should be determined by the RITH team or a rehabilitation team acting on behalf of the RITH team (T2-S2.1). The RITH team should have the final say on eligibility (T2-S2.3), with sign-off by a rehabilitation physician (or their delegate) (T2-S2.4) to confirm the appropriateness of RITH as hospital substitution. Table 4 lists supported items for determining patient eligibility and includes items relating to patient function, cognition, motivation, available support, and the home environment. Assessment of the safety and suitability of the home was considered essential, with alternatives to a physical home visit (e.g., checklist or virtual tour) being acceptable (T2-S2.5, S2.6). It was deemed important in considering eligibility that only patients whose goals are likely to be achievable within a RITH program should be selected (T4-13). Consensus was not reached on the patient’s degree of frailty (as scored by the Clinical Frailty Scale [23]) being a determinant for eligibility (A1-6).

‘High falls risk’ in the patient as an exclusion criterion for RITH fell just short of consensus (69.3%, A1-7), with comments in DR2 suggesting that high falls risk is common in rehabilitation patients. Consensus on eligibility was subsequently achieved in DR3 for a patient having a ‘manageable’ (DR3-43) falls risk, alongside consideration of ‘patient autonomy’ (DR3-26) around this risk. Comments supported ‘falls risk mitigation strategies’ (DR3-11) such as ‘falls education’ (DR3-51), ‘the provision of equipment and aids…[and] carer supervision during mobility tasks’ (DR3-11); with eligibility decisions informed by patient and family ‘insight’ (DR3-63) and whether they are likely to be ‘compliant’(DR3-59) with falls mitigation strategies.

Patients and their carers should have a clear understanding of the available rehabilitation options (pros and cons of RITH versus inpatient rehabilitation) (T2-S2.7), and jointly established minimum achievable goals must be agreed prior to admission to RITH (T2-S2.8). A written agreement for carers, making explicit the expectations and roles of a carer during RITH, is desirable (T2-S2.9). Carers must agree that the support available through RITH is sufficient for them to support the patient at home (T2-S2.10). Further, paid support services should be available to patients on RITH programs when needed (e.g., personal care, home help, meal services) (T2-S2.11).

It was agreed that the acute team remains responsible for discharge from acute care (T2-S2.13). The RITH team should work with them to facilitate patient discharge (T2-S2.15) and ensure that liaison with carers occurs (T2-S2.14). Acute hospitals should not discharge medically unstable patients to RITH (T2-S2.16), and medically unstable patients should not be accepted onto RITH programs (T2-S2.17).

DR1 sought the best time for the development of the RITH care plan. Comments viewed the care plan as an ‘evolving process’ (DR1-111) with an ‘initial basic plan’ (DR1-88) in place before the patient leaves acute care ‘reflecting the person’s status on discharge’ (DR1-99). A more comprehensive plan would follow the multidisciplinary team (MDT) case conference, being modified when the patient’s ‘goals become clearer once they are at home’ (DR1-52). Clarity and consensus on who can develop the initial care plan was confirmed in DR2 (T2-S3.1). A brief admission to a rehabilitation ward for assessment and RITH care plan development is not needed (T2-S3.2).

A rehabilitation physician or their delegate should assess the patient before the initial care plan is developed (T2-S3.3). The care plan should include an indicative program duration, and number and type of therapy sessions (T2-S3.4, S3.5). There was consensus for using assessment tools that are, or would be, supported by AROC (T2-S3.6).

Outcome data should include functional measures, degree of achievement of negotiated patient goals, patient reported experience measures (PREMs) and patient reported outcome measures (PROMs) (T2-S5.1–S5.3). Comments raised concerns about the suitability of the FIM ‘as [RITH patients] are usually more independent than inpatient[s]’ (DR1-114). Outcome data should be submitted to AROC for benchmarking (T2-S5.4-S5.5), with consensus achieved for a mandatory set of AROC-supported assessment and outcome tools (S5.6). Comments indicated these needed to be ‘meaningful’ (DR3-2) with ‘input from different stakeholders’ (DR3-51) and allow for ‘comparison…between RITH and inpatient rehabilitation’ (DR3-25).

Fall-back admission to inpatient rehabilitation should be available where patients fail to progress on RITH (S5.7), with a key performance indicator of ≤ 10% for subsequent admission to inpatient rehabilitation considered acceptable (T2-S5.8). Also, in a well-functioning RITH program, acute hospital readmission rates should be no higher than occurs following inpatient rehabilitation (T2-S5.9). Discharge from RITH should occur as soon as a patient achieves their goals (T2-S5.10).

The patient’s clinical care needs were seen as more important than the available budget for their RITH program (T2-BF.1), but participants also agreed that RITH costs should be no more than comparable inpatient episodes (T2-BF.2) and that RITH programs/models need an endpoint for service delivery (T2-BF.3). Budget flexibility between individual patient programs was supported if the overall RITH service was able to work to its budget (T2-BF.4). Appropriate processes for approval of additional expenditure and RITH program extension are needed (T2-BF.5-BF.6).

Budget parameters are necessary for funders (T2-BF.7). Any RITH service (publicly or privately funded) must work to a budget (either an individual budget per patient or an overall budget for the RITH service) (T2-BF.8). Predictive models used to develop a costing model for RITH can be helpful at an overall program level but should not be strictly applied at the individual patient level (T2-BF.9). Key decision makers should receive formal education on setting, monitoring and managing individual program budgets, and on using a schedule of costs or pricing tool to aid in the development and costing of patient RITH care plans (T2-BF.10-BF.11). RITH could be an attractive option in both the public and private systems (T2-BF.12-BF.13).

There was strong support for RITH for reconditioning being widely available for appropriate patients (T2-PU.1). Participants did not believe that the design of a model of care for RITH would be too complex (T2-PU.2). Anticipated utilisation of RITH was tested in DR3 by asking participants to estimate its likely utilisation for each of the six reconditioning case-mix classes used in Australia (AN-SNAP classes 4AR1 through 4AR6 [1]). These classes are explained with Table 5, along with participant responses. Twenty-one participants (of the 78 who participated in DR3) who felt they were sufficiently experienced with AN-SNAP to respond to the questions estimated that around two thirds of patients in the first two classes (4AR1 and 4AR2, which account for over one half of all inpatient reconditioning episodes in Australia [24]) would likely be suitable for RITH. They also estimated that around three-quarters of these suitable patients might want to take part in RITH if it were available (Table 5).

Considerable variation in views amongst inpatient staff about patient suitability for rehabilitation [31], and for early supported discharge programs [32] has been reported in the literature, which impacts equity of access for patients and makes the adoption of agreed service models problematic. Our findings support a two-step process (i.e., screening, followed by a more detailed eligibility assessment), with agreement being reached on relevant items for both (Tables 3 and 4). This should assist in addressing referral inconsistency and in improving the efficiency of patient referrals to RITH.

Eligibility items include a range of factors which are generally not considered in determining eligibility for inpatient rehabilitation. For example, whether the home environment is safe for both patient (while undergoing their RITH program) and staff (no safety risk posed); and whether the patient will be able to manage the necessities of living (such as meal preparation, doing laundry), or if not, that these tasks can be provided for under the RITH program (e.g., using external brokerage if required). The willing assistance of the patient’s carer (where there is one) was also seen as key to eligibility, and participants supported written documentation for carers to consider and agree to, which makes explicit the expectations and roles of a carer when a patient enters RITH. This would address what Dow and McDonald [33] identified, that carers are usually consigned a marginal status within home-based rehabilitation programs and feel bound by an “invisible contract” to carry out substantial care-work that would otherwise have been undertaken by hospital staff. Finally, while high falls risk is generally not an exclusion criterion for inpatient rehabilitation [34], our findings suggest that when balancing falls risk with program benefit in the RITH setting, the situation may be more nuanced. Survey participants recognised that patients in the reconditioning impairment category are likely to have a degree of falls risk and the Delphi process settled on that risk being a ‘manageable’ one, rather than there being a specific level of falls risk that would exclude someone from a program, with patient autonomy, insight, compliance, and mitigation strategies to be considered. Further, as frailty and falls risk are strongly related [35, 36] it is consistent that survey participants did not also agree on a level of frailty as an inclusion or exclusion criterion for acceptance onto RITH.

In response to participant comments about how practice has had to adapt to the COVID-19 pandemic, we tested and received confirmation of the idea that a “virtual” tour of the patient’s home to assess the home environment is feasible. Interestingly, a scoping review of the literature published prior to the pandemic identified that the use of technology within occupational therapy home assessments had been generally underutilised [37], but that its use appeared to be feasible and acceptable, with benefits in terms of rapid visual access of a property, enhanced resource utilisation (e.g. elimination of travel time) and potential cost-effectiveness [37, 38].

Participants supported three RITH service delivery models—intensive and short term (mirroring closely an inpatient episode, which for example averages 12.1 days and 15.2 days for AN-SNAP classes 4AR1 and 4AR2, respectively [24]); variable and longer term, generally to a maximum of ten weeks duration; and intensive single discipline where the patient cannot be discharged from hospital without such a program. These models suggest the possibility of better tailoring programs to client needs. For example, Model 2 may allow some patients (e.g., with frailty or comorbid cognitive impairment or dementia) to engage in a program of therapy to achieve their goals at a slower pace and with less fatigue [39, 40] than may be experienced with Model 1. Consensus was for all three models to be delivered by a RITH service provider because of the expertise within the RITH team to monitor progress and take corrective action when needed. Indeed, co-ordinated MDT input was found to provide better outcomes for stroke patients in a Cochrane Review of early discharge with rehabilitation at home [13].

While not supporting any specific discipline as being the ideal RITH case manager, participants were of the view that a suitably skilled nurse should be part of the RITH MDT. Six roles for the nurse within the multi-professional rehabilitation team have been described for the inpatient setting, as follows: ‘assessment, co-ordination and communication, technical and physical care, therapy integration and therapy carry-on, emotional support, and involving the family’ [41]. While similar rehabilitation nurse roles will likely exist in RITH, these roles will probably be shared across RITH team members where an interdisciplinary approach to care is adopted.

There was strong support by participants for a mandatory agreed set of assessment and outcomes tools to enable benchmarking of RITH services. Around 98% of all inpatient rehabilitation episodes in Australia are reported to AROC [42]. This enables a robust analysis of clinical information for the purposes of reporting and improving outcomes, with AROC publishing comparative data biannually for inpatients [42]. However, the collection of ambulatory data by AROC remains in its infancy in Australia, and is hindered by the diverse range of service models in operation and the lack of agreed outcome measures [14]. Further, the opportunity to compare outcomes from inpatient and ambulatory settings may be lost where tools appropriate for both settings cannot be identified. The impact of heterogeneity of measurement tools on comparative research is illustrated by a systematic review of inpatient versus home-based rehabilitation for older adults with musculoskeletal disorders, which found 21 different assessment tools across 12 studies, with the most frequent tools only reported in three studies [43].

The identification of tools for RITH is challenging, particularly so for the reconditioning impairment category which covers a diverse clinical group. More patient-centric outcome tools may be needed, such as PREMs and PROMs and goal attainment scales. These were all supported by survey participants. Goal attainment is likely to be especially important because RITH patients may experience ceiling effects on standard functional measures (e.g., FIM), and reaching goal attainment was an agreed point of readiness for discharge from RITH.

A clear and widely understood system of clinical governance that ensures the continuing delivery of safe and high-quality care as the patient transitions from the hospital environment to the community was strongly supported. This included consensus on specific roles for the acute care team, the rehabilitation physician, the RITH team, and the case manager. Specific standards for RITH to be developed by the Australasian Faculty of Rehabilitation Medicine were supported. These could be similar to those that already exist for inpatient rehabilitation [44]. A clear medical escalation plan is also needed. Actions to integrate primary care (such that the patient’s GP can provide ongoing general medical care to the patient at home) into the RITH as hospital substitution program were supported but challenged by existing funding models. While there has been little research in Australia on vertical integration (e.g., between state-funded public hospitals and largely federally-funded practice in the community), significant funding model barriers to more integrated care models have been noted [45].