Financial analysis of revision hip surgery at a tertiary referral centre as classified using the British Hip Society Revision Hip Complexity Classification

Introduction

Total hip arthroplasty (THA) continues to be one of the most common procedures performed in the NHS, with 109,624 procedures recorded in the National Joint Registry (NJR) in 2019 prior to the ongoing COVID-19 pandemic.¹ It is recognized to be one of the most clinically effective and cost-effective interventions performed based on the criteria set by the National Institute for Health and Care Excellence (NICE).²

As a result, revision total hip arthroplasty (rTHA) surgery continues to be a sizeable burden, consistently accounting for 8% of all hip arthroplasty procedures over the last few years.¹ In the USA, the burden of revision hip surgery is expected to double between 2007 and 2026.³ It is a more complex, technically demanding, and time-consuming intervention,⁴ with outcomes and survivorship well below that of the primary intervention.^1,5 Failure of the primary THA may be due to a wide variety of reasons, with the most common being aseptic loosening, infection, periprosthetic fracture, and dislocation. The cost of rTHA surgery is much greater than that of primary surgery owing to the increased theatre time, hospital length of stay (LOS), and implant costs. Additionally, the indication for rTHA also has a direct effect on cost, with infected revision being significantly more expensive than aseptic cases.^6,7 Remuneration in England is currently based on Healthcare Resource Groups (HRG) data, whereby the type of work, as well as the patient’s comorbid state, is grossly grouped and assigned a value. From 2019 to 2020, the average remuneration for an elective, “very complex hip procedure for non-trauma” was £13,184 to £23,571, dependent upon comorbid state.⁸ But patients requiring rTHA surgery are a very heterogeneous group, and thus categorization in this way is unlikely to truly reflect the cost of providing such specialized services.

In the UK, there is a drive to create regional network systems. Such networks are already well established in the UK in cancer and trauma care, demonstrating improved patient outcomes.^9-11 A three-tier system for revision arthroplasty has been proposed, with the Major Revision Centre (MRC) managing the multidisciplinary team (MDT) meeting, supported by Revision Units (RU) and Primary Arthroplasty Units (PAU). The aim is to ensure that the more complex or low-volume cases are performed by fewer specialist units in order to achieve excellent outcomes for patients, with low complication rates. The potential risk of such a system is that the burden of the more complex, high-risk, and consequently high-cost, work will be borne by the MRC without adequate remuneration. This has been reported previously for revision hip and knee surgery under previous iterations of NHS payment schemes.^6,12

To help guide triage of patient care to the appropriate unit within the revision network, the British Hip Society (BHS) has proposed a classification system, the Revision Hip Complexity Classification (RHCC) (Table I).¹³ This was developed using a modified Delphi method of consensus opinion among experts. The aim is to reflect complexity based upon mechanism of failure, bone loss on both the femoral and acetabular side of the joint (Paprosky Classification),^14,15 and patient factors such as BMI and American Society of Anesthesiologists (ASA) grade. The RHCC classifies patients into three groups of increasing complexity from the relatively simple aseptic loosening (H1), through to the highly complex (H3) situations of the multiply revised patient for infection, dislocation, or complex bone loss. Within each category there is the provision to denote high-risk patients on the basis of BMI ≥ 40 kg/m² or ASA grade ≥ III with an asterisk appended (e.g. H3*). The validation work on the RHCC has been completed, demonstrating that the RHCC is reliable in predicting complexity of hip revision.¹³

Methods

A retrospective analysis was performed of all revision hip arthroplasty cases carried out between 5 April 2018 and 5 April 2020, constituting two consecutive financial years. All episodes were identified from prospectively collected data in the unit’s comprehensive arthroplasty database. This is a tertiary referral centre for the region and therefore sees a high volume of patients, representing the full spectrum of revision cases from the least to the most complex. As such, the unit also reflects any likely future proposal for a MRC.

Each episode was individually reviewed by the lead authors (DIH, MJP) to ensure that they fit the inclusion criteria of revision hip procedure performed in the reference timeframe, either debridement, antibiotics, and implant retention (DAIR), single-, first-, or second-stage. Exclusions included fixation only of periprosthetic fractures and staged procedures for septic arthritis of native joints. During the review process, the clinical notes were obtained from the electronic patient record (EPR) (Lorenzo EPR Solutions; DXC Technology, USA) to identify basic patient demographics, BMI, comorbidities, ASA grading, and the mode of failure (infected or non-infected). The radiological imaging from the hospital picture archiving and communication system (PACS) was obtained in order to determine the Paprosky Classification of acetabular and femoral bone loss.^14,15 Theatre use data were collected from the Operating Room Management Information System (ORMIS) (DXC Technology).

The complexity of each case was classified using the BHS RHCC (Table I). The RHCC classification takes into account the indication for revision (i.e. aseptic loosening, trauma, infection), as well as the previous history (re-revision), and the complexity of the reconstruction (Paprosky Classification), categorizing the case into one of three groups (H1 to H3) based upon increasing complexity. Further denotation can be applied within each group with an asterisk suffix to reflect the risk profile of the patient themselves based upon BMI ≥ 40 kg/m² and ASA grade ≥ III.

The hospital finance department provided the costs based on the Patient Level Information and Costing System (PLICS) and remuneration figures for each of the inpatient episodes based on the HRG codes. These were then analyzed to compare the relative costs between the RHCC groups, infected versus non-infected revisions, and low- versus high-risk patient groups (BMI ≥ 40 kg/m² and ASA grade ≥ III). The primary outcome measure was the difference between the cost of provision of the service and the remuneration to the hospital (profit/loss) per patient per inpatient episode.

Results

Complete demographic, clinical, and financial data were identified for 199 rTHA episodes in 168 patients, during the two consecutive financial years from 2018 to 2020, that met our defined inclusion criteria. There were 95 females (56.5%). The mean age at the time of surgery was 69.7 years (28 to 98). Of the 199 rTHA episodes, 47% had previously undergone primary or revision surgery within this unit. A heat map which depicts our current referral network is shown in Figure 1.

Fig. 1

Heat map depicting density of referrals to our unit.

There were a total of 25 episodes (13%) in the H1 group, 110 (55%) in the H2 group, and 64 (32%) in the most complex H3 group. High-risk patients accounted for 78 episodes (39%). Infection was the primary diagnosis in 76 episodes (38%), of which 27 (36%) were second-stage procedures. For the infection group, 42 (55%) were H2 and 34 (45%) were H3 complexity groups. The remaining 123 episodes (61.8%) were non-infected due to all other causes including wear and aseptic loosening (n = 62, 50%), dislocation (n = 37, 30.1%), implant failure (n = 10, 8.1%), trauma (n = 9, 7.3%), and pain (n = 5, 4.1%). The indications for revision in each group are presented in Figure 2.

Fig. 2

Primary indication for revision by complexity group.

The median LOS demonstrated a significant increase between complexity groups: H1 (4 days, IQR 3 to 9); H2 (6 days, IQR 4 to 11); and H3 (8 days, IQR 4 to 17) (p = 0.006). The LOS also significantly increased in the infected group compared with the non-infected group (9 (IQR 6 to 15) vs 5 days (IQR 3 to 11), respectively, p < 0.001) and low- versus high-risk patient groups (6 (IQR 3 to 11) vs 8 (IQR 4 to 15), respectively, p = 0.044) (Figure 3).

Fig. 3

Median length of stay (days) by Revision Hip Complexity Classification (RHCC), mode of failure, and patient risk category. LOS, length of stay.

The median total cost per episode of all rTHAs over the two financial years was £13,222 (IQR £9,676 to £19,203). Moreover, for the least complex H1 group the median cost was £10,057 (IQR £7,312 to £13,558), H2 £12,756 (IQR £9,188 to £17,854), and H3 £16,105 (IQR £11,982 to £25,704), demonstrating a significant increase in cost to the unit as case complexity rises according to the RHCC (p < 0.001). Costs were significantly increased between the non-infected group (£11,594, IQR £8,926 to £16,366) and infected group (£16,234, IQR £12,620 to £21,591), (p < 0.001), which is demonstrated in Figure 4. Remuneration did increase significantly (p = 0.019) with the RHCC complexity from the H1 group (£9,868, IQR £7,560 to £12,297) to the H2 group (£11,912, IQR £8,852 to £16,219) and the H3 group (£12,557, IQR £9,450 to £22,100), as it did between non-infected (£11,135, IQR £8,609 to £16,322) and infected groups (£14,070, IQR £9,767 to £17,360) (p = 0.012). A median deficit was demonstrated in all groups but rose significantly with complexity from £97 deficit (IQR £1,224 deficit to £1,308 profit) for H1, £1,050 deficit (IQR £3,189 deficit to £1,382 profit) for H2, and £2,887 deficit (IQR £5,947 deficit to £676 profit) for H3 (p = 0.006). The deficit for non-infected cases was £635 (IQR £2,749 deficit to £1,021 profit), rising to £2,629 (IQR £5,837 deficit to £1,438 profit) for infected cases (p = 0.032) (Figure 5). There was no statistically significant difference in the cost, remuneration, or final profit/deficit between the high- and low-risk groups.

Fig. 4

Median cost based on the Revision Hip Complexity Classification (RHCC), mode of failure, and patient risk category.

Fig. 5

Median deficit based on the Revision Hip Complexity Classification (RHCC), mode of failure, and patient risk category.

Over the course of the two financial years, the rTHA service generated a deficit to the unit of £512,202.

Discussion

Remuneration of secondary care services in the NHS is based upon the National Tariff with a complex system of HRG codes, which aim to group together similar activities and assign them a value.¹⁶ These HRGs have a basic descriptor such as “Very Major Hip Procedures for Non-Trauma” (HN12) or, “Very Complex Hip or Knee Procedures for Non-Trauma” (HN80), suffixed with a letter (A to F) to further denote complexity based on comorbid status. Additionally, there is then a degree of uplift to account for a prolonged LOS and the market forces factor (MFF), which accounts for the differing costs of services geographically. The National Tariff is adjusted to take into account inflationary costs, but this is done by way of a three-year cycle of cost data returns by hospital trusts on the true costs of treatments which are then averaged out nationally. As such, remuneration payments on the National Tariff are always likely to lag behind. In addition to this, as demonstrated in this paper, rTHA is a very heterogeneous group of procedures in which no two cases are the same and the costs vary wildly. To try and use such a blunt tool as HRGs, applied by non-clinical staff relying on limited information, to generalize a group of procedures as complex as rTHA, is extremely challenging and unlikely to give accurate results.

Our institution is a tertiary referral centre for revision hip and knee arthroplasty. It has a busy local referral pattern as well as a wide-reaching and long-established tertiary referral catchment area. The unit undertook 199 revision procedures in 168 patients over the two financial years of 2018 to 2020. Of these, more than 38% were for infection (DAIR, single-stage, first-, or second-stage), significantly higher than the average of 14.4% reported by the NJR. We have shown that revision procedures for infection carry the greatest cost and deficit to the treating unit and therefore account for much of the financial burden.

Median LOS varied widely, from one to 122 days. There was a statistically significant difference between all complexity groups, infected versus non-infected cohorts, and the high- versus low-risk cohorts. This is likely to account, in part, for some of the higher costs in these groups. However, in the high-risk group (ASA ≥ III and BMI ≥ 40 kg/m²), there was a trend towards higher costs and remuneration as would be expected, without reaching statistical significance (p = 0.129 and p = 0.072, respectively). There was no suggestion of a difference in the profit/deficit (p = 0.572), which would imply that the comorbidity element of the National Tariff and HRG coding system is working well to account for any true increase in costs that these patients represent.

The routine practice within our unit is to limit the length of time patients are treated with parenteral antibiotics. Typically, for the first-stage procedure, the focus is on strict surgical technique of debridement and local delivery of high-dose antibiotics in the form of cement beads, with no further parenteral antibiotics unless there is significant soft-tissue infection. For single-stage procedures, a short two-week course of parenteral antibiotics is preferred.^17,18 For this reason, we would expect our unit to have a shorter average LOS (and therefore lower costs) for infected cases than most other units treating infection. Shortened courses of antibiotics are the subject of the Short or Long Antibiotic Regimes in Orthopaedics (SOLARIO) study,¹⁹ a multicentre randomized controlled trial which is currently ongoing. By using this approach, there is the opportunity to control costs through both antibiotic stewardship and patient LOS.

The RHCC was designed by the British Hip Society to reflect the complexity of the revision surgery being undertaken. It was formulated by performing a Delphi method of consultation with a panel of experts in the field. The aim of the RHCC was to identify the complexity of a case, and triage which revision centre or unit within a network would be most appropriate to undertake a particular revision procedure. We demonstrate that there is a correlation between the RHCC group and the increasing cost of a rTHA episode. However, the range of costs to the trust still vary wildly within each of the three groups. This, in part, will be due to first-stage revision episodes that are more costly due to the antibiotic use and increased LOS in a relatively unwell, potentially septic patient. A second-stage procedure is more predictable in terms of LOS, surgical time, and costs. The RHCC is not designed to differentiate between the two.

There are limitations to this study. It is recognized that the RHCC is a new classification system that is not yet widely adopted. It has, however, been developed using the recognized Delphi method of consensus opinion and has good inter- and intraobserver reliability. It has also been demonstrated in this paper that it correlates with the patient LOS, as well as the cost to the trust via PLICS data and the remuneration via the National Tariff and HRG coding systems. This paper has used PLICS data to evaluate the cost of the admission but, although this is far more accurate than the HRG coding system, it is not possible to capture each individual consumable used on a single patient. Assumptions are made within the system of the average daily cost per patient to encompass these. This paper also does not capture readmissions, or complications such as ongoing wound discharge or return to theatre for minor procedures unless they led to another revision episode. It is therefore likely that the true costs of managing a revision hip service are even greater than we have reported.

This paper has demonstrated that the current system of remuneration for the complex rTHA is inadequate and leads to a significant financial deficit. With the impending development of revision networks for rTHA, there seems to be little incentive for hospitals to volunteer to offer services for the most complex cases, particularly those with infection. There needs to be further development of funding strategies to cope with the heterogeneity of rTHA, particularly related to case complexity, so that centres treating the bulk of the more complex cases are not financially disincentivized. A more appropriate mechanism of payment may be the centralized specialist commissioning system that funds specialized services such as cancer care and thrombectomy services for stroke patients.²⁰

The current NHS reimbursement for revision hip surgery does not adequately reflect the true costs of undertaking this type of work, particularly the more complex revision cases and those associated with infection. This has the greatest impact on the tertiary referral centres that are likely to become the hubs of any future revision networks and is therefore a financial disincentive to the effective creation of networked care.