Low value care and inpatient rehabilitation after total knee replacement

A benchmark for the referral rate to inpatient rehabilitation could reduce the evidence–practice gap

There are several reasons why inpatient rehabilitation after total knee replacement (TKR) should be reserved for those most in need. Randomised clinical trials1,2 and observational studies3,4 indicate that inpatient rehabilitation for patients after an uncomplicated TKR is not usually necessary. It is, however, considerably more expensive than the alternatives,1,3,4 whether the cost is borne by the patient, their insurer, or the government. These observations together suggest that the inpatient pathway is often low or no value health care. Further, there is low-level evidence from the United States that rates of re-admission and adverse events, after adjusting for relevant factors, are higher among patients discharged to inpatient rehabilitation or skilled nursing facilities.5,6

Despite this evidence, the article by Schilling and colleagues7 in this issue of the MJA shows not only that inpatient rehabilitation in Australian private hospitals is not reserved for the few, but that rates have increased markedly in recent years, from 31% in 2009 to 45% in 2016. Equally perplexing, rates varied between hospitals from 0 to 100%, with provider characteristics explaining three times as much variation as patient characteristics such as age, comorbid conditions, complications, and residential status. The authors concluded that the degree of variation unexplained by patient-related factors potentially signals low value health care.

Annual volumes of TKR surgery in Australia have increased since 2003.8 In 2016, 52 126 primary procedures were undertaken, 70% of them in the private sector. With increasing volumes and growing out-of-pocket expenses for private health care patients,9 it is time to curtail costs and to deliver care that is aligned with the evidence.

In the US, bundled care payments for TKR that cover the first 90 days after surgery have driven significant reductions in inpatient rehabilitation rates.6,10 This approach cannot be discussed here in detail; it suffices to note that Australia may not be ready or suitable for such an approach. An alternative solution in our health care context may be to establish an upper limit for referral rates that hospitals (or surgeons) should endeavour to keep.

Rates of inpatient rehabilitation after TKR are not high in health care systems without sizeable private sectors, as in the United Kingdom and Canada. Public hospital rates are a logical starting point for defining a benchmark in Australia, as referral in the public system is dictated by need because of the limited number of rehabilitation beds. Need in this context is defined by lack of social support or poor progress because of multimorbidity or a significant hospital-acquired complication. The Arthroplasty Clinical Outcomes Registry National — the only source of comparative data of this kind — reported an overall rate of 14% for patients with osteoarthritis who had undergone primary, unilateral TKR in 2016 (and 15% for hip replacement)11 (Box). It is unlikely that this rate is inappropriately low, as rates of re-admission caused by difficulties in coping with activities of daily living — a proxy measure of inappropriate or premature discharge home — are very low, and because the rate is similar to those reported overseas.6

Schilling and colleagues suggest that a private hospital referral rate of 31% (the 2009 level) would achieve substantial cost savings. Although considerably higher than the public hospital rate, it would accommodate the many rural residents who travel to metropolitan private hospitals for surgery who may not have access to local community-based rehabilitation facilities. Even after adjusting for these patients, it is likely the benchmark could reasonably be set much closer to the public sector rate. However, a conservative approach to reducing potential overuse, may, at least initially, gain more traction in a sector that is heavily invested in the inpatient rehabilitation model.

Realisation of a system-wide benchmark requires a collaborative, multipronged approach that brings all stakeholders to the table. Formulation of criteria for referrals to inpatient rehabilitation, similar to those in the public sector, would be integral to promoting transparency and fairness. Incorporating a measure of function that quantifies poor progress or testifies to a level of impairment that could be noticeably reduced by inpatient treatment is key. Other supportive strategies include developing and promoting rehabilitation in the home1^,12 (akin to community packages available to eligible public patients), insurer commitment to removing out-of-pocket expenses for community-based rehabilitation so that clinic-based options are more attractive, and informing patients prior to surgery about the efficacy of the rehabilitation options available.12 Importantly, these arguments and approaches apply equally to rehabilitation after hip replacement.1

Box – Referral to inpatient rehabilitation of publicly insured patients with osteoarthritis after primary unilateral total knee replacement in public hospitals, 2016*

---

* 149 of 1066 patients. Based on Arthroplasty Clinical Outcomes Registry National data,11 with permission.