Background The need for orthopaedic surgery in RA is considered a marker of disease severity¹ & hospital admissions for orthopaedic intervention are frequent events. Length of stay (LOS) is an important driver of costs. Few studies report on clinical factors affecting LOS in RA patients. In the current economic climate of over-pressurised health services, identifying & targeting such factors could result in reduced LOS, improved outcomes & health care utilization planning.

Objectives To investigate rates of orthopaedic surgery in RA, LOS & factors affecting it.

Methods The Early RA Study (ERAS) is an inception cohort study of RA patients from 9 UK rheumatology centres recruited between 1986-1999, followed until 2011 (n=1465). The Early RA Network (ERAN) has a similar design, recruiting from 24 UK centres, 2002-2011 (n=1236). Standard clinical, laboratory & x-ray measures were performed yearly in both cohorts. Source data of all orthopaedic interventions included clinical datasets (patient report & medical records from 1986 & from 2002 onwards) validated with Hospital Episodes Statistics (HES from 1997) & the National Joint Registry (NJR from 2002). Length of follow up was based on the National Death Registry. Univariate & multivariate regression analysis was used to investigate predictors for LOS.

Results 558 (38%) ERAS & 238 (19%) ERAN patients had a total of 1287 & 340 interventions respectively. The commonest interventions in ERAS were hand/forefoot procedures (19%), knee (TKR 17%) & hip replacements (THR 14%). ERAN was similar, except for lower rates for hand/forefoot arthroplasty (6%). The median LOS for major, intermediate & minor procedures were 10, 4 & 2 days respectively (IQRs 7-15, 3-7, 1-4) in ERAS, & in ERAN were 7, 3 & 1 (IQRs 5-8, 1-7, 1-2). The major differences in costs of actual procedures & associated inpatient stay will be displayed graphically. In univariate analysis, major comorbidities increased LOS significantly. In multivariate analyses LOS varied significantly by type of procedure, DAS score, haemoglobin (HB) levels & HAQ scores. In ERAS, having a THR increased LOS by 9.9 days (p=0.001) & a TKR by 5.9 days (p=0.001). Normal baseline HB significantly reduced LOS by 1.8 days (p=0.001), baseline HAQ<1 by 1.8 days (p=0.002), & low 1yr DAS (<3.2) by 1.1 days (p=0.051). Using the same model in ERAN, THR & TKR significantly increased LOS by 6.4 days (p=0.001) & 3.3 days (p=0.033) respectively. Clinical variables did not reach significance in multivariate analysis. LOS in ERAS & ERAN varied significantly by age (p=0.001 & p=0.015) but not by gender or centre.

Conclusions LOS is an important determinant of health care costs. This study found evidence of significant variation in LOS & therefore costs by type of procedure in the care for RA patients. This unique dataset with up to 25 years follow up reflects how orthopaedic intervention has changed over 1986-2011, mainly in less hand/foot surgery & reduced LOS. This could be a reflection of improved medical treatments & overall patient care in the new millennium. Normal baseline HB, low baseline HAQ & low DAS at 1 year were found to significantly influence LOS in ERAS.

1. Anderson RJ. The orthopedic management of rheumatoid arthritis. Arthritis Care Res 1996;9:223–8.

Disclosure of Interest None Declared