A comparison of a referral triage tool for knee osteoarthritis with surgeon’s decision making for total knee arthroplasty

Abstract

s / Osteoarthritis and Cartilage 22 (2014) S57–S489 S385 683 A COMPARISON OF A REFERRAL TRIAGE TOOL FOR KNEE OSTEOARTHRITIS WITH SURGEON’S DECISION MAKING FOR TOTAL KNEE ARTHROPLASTY M. Harrison y, D.V. Cooke z, W. Hopman y, M. Brean x, J. Hope y. yQueens Univ., Kingston, ON, Canada; zOaisys Inc., Kingston, ON, Canada; xOaisys Inc, Kingston, ON, Canada Objective: To assess the ability of a triage tool, based on patient selfreport disability assessment and standardized radiographic score to determine appropriateness for TKA surgery for patients with knee osteoarthritis (OA). Design: A prospective study was performed to assess a triage tool to clinical decision making by the orthopedic surgeon. Participants: Patients with knee pain seen during their initial referral to one of four orthopedic surgeons over a 1-year period. Methods: The study was approved by the institutional research ethics board. Patients were enrolled into the study by an advanced practice physiotherapist (APP). Informed consent was obtained and subjects completed a standardized battery of selfreport disability questionnaires, namely; SF 12, WOMAC, Tegner & Lysholm questionnaires, Functional Comorbidity Index and an Inflammatory Disorder Questionnaire. The APP also collected demographic data and assessed the patient based on a standardized osteoarthritis referral questionnaire WCWL-PRS. Standardized knee radiographs were scored based on joint space width (0-3), femoral osteophytes (0-3), tibial erosion (0-4) and subluxation (0-3) with a total score from zero to 13. 5 Radiographic assessment was performed by trained readers who were blinded to the clinical data. The diagnosis and decision for surgery was performed by the orthopedic surgeon, independently from the data collection of the clinical disability questionnaires. Information from the orthopedic consultation letter to the referring physician was abstracted by a single trained observer. The diagnosis and treatment recommendations and decision for surgery, as well as any reasonswhy the subject did notwant surgery were recorded. Statistical analysis was performed to assess correlations between the diagnosis, appropriateness for surgery and whether the subjects had surgery with the various disability questionnaires and radiographic assessment. Results: Of the 173 patients enrolled, 46 were assessed by the orthopedic surgeon as being not to be appropriate for surgery (Group 1) and 127 were deemed as being appropriate candidates for surgery (TKA) (Group 2). Statistical analysis showed that most disability questionnaires were positively correlated with subjects being appropriate for surgery (Table 1). A step-wise logistic regression analysis was performed to determine what variables would best distinguish between groups. The independent variables in the final model were, age, WOMAC total score and total radiographic score. The odds ratio for being appropriate for surgery increased by 1.90 for every one point increase in the total radiographic score, by 2.03 for every ten year increase in age and by 1,03 for every 10 point increase in WOMAC score (Table 2). There were 48 subjects who were felt to be appropriate for TKA , but who refused surgery for a variety of reasons. One subjects might have multiple reasons why they declined surgery. The common reasons to decline surgery were; minimal symptoms (37), brace treatment (6), obesity (6), severe medical condition(6), age (6) and other MSK condition (1). We subdivided the subjects who were deemed appropriate for surgery into two groups , those who declined surgery and those who underwent surgery. We then repeated the statistical analysis with three groups (Group 1 not appropriate, Group 2 appropriate but declined and Group3 had surgery)(Table 3). As with the previous statistical analysis the disability questionnaires were correlated with the three groups. Post-Hoc analysis was performed using the Tukey test. This identified that Group 1 was statistically different from Group 3 for patient age. Group 3 was statistically different from Groups 1 & 2 for the physical component of the SF-12, all components of the WOMAC and the total WCWL score. The radiographic scores for joint space narrowing and total radiographic score were the only variables that stratified the subjects into three separate groups, with group 1 different from group 2 which was also different from group 3. Conclusion: These results suggest that an evidence based triage tool based on the combination of patient age, total WOMAC score and total radiographic score may be useful to determine the need for TKA surgery, since it correlates highly which surgeon’s decision for surgery. Further optimization of the triage tool may include simple questions to address the common reasons why patients declined surgery (significant medical condition, obesity and patient desire to avoid surgery). 684 SUBJECTS WITH SELF-REPORTED BASELINE INJURY HAVE AN INCREASED RISK OF SYMPTOMATIC RADIOGRAPHIC KNEE OSTEOARTHRITIS OVER TWENTY-YEARS K.M. Leyland y, M.T. Sanchez-Santos y, S. Kluzek y, D. Prieto-Alhambra y, T.D. Spector z, D.J. Hart z, J. Newton y, N.K. Arden yx. yUniv. of Oxford, Oxford, United Kingdom; z St. Thomas’ Hosp., London, United Kingdom; xUniv. of Southampton, Southampton, United Kingdom Purpose: Knee injury has been previously identified as a strong predictor of both radiographic and symptomatic knee-osteoarthritis (ROA), primarily in cross-sectional, short-term longitudinal studies (<10 years) or in young adult cohorts. This research investigated the association between self-reported knee injuries in middle-aged women and the risk of developing clinically relevant symptomatic radiographic knee osteoarthritis over the subsequent nineteen years. Methods: The Chingford study, a UK population based prospective cohort of women recruited in 1989 was used for the analysis. All subjects without evidence of pain or radiographic osteoarthritis (ROA) in either knee and at least one follow-up visit were included. Subjects with any history of rheumatoid arthritis were excluded. The injury question at baseline asked “have you ever injured your knees, enough to rest them for a week?” Knee ROA was defined as Kellgren and Lawrence grade 2 or above (including knee replacements), and pain was assessed as 15 or more days in the previous month. A subject was positive for symptomatic radiographic osteoarthritis (SROA) if they had both pain and ROA in the same knee on the same visit. A proportional odds model within a discrete-time survival analysis was used to examine the risk of SROA using 5-year intervals over twenty-years between subjects with and without reported previous injury at baseline. The model was adjusted for age, BMI and weekly walking distance. Results: 64.3% (n1⁄4645) out of the original 1003 subjects were knee pain and ROA free at baseline, had at least one follow-up point and did not have rheumatoid arthritis. Mean follow-up for subjects with injury (n1⁄4191) was 14.9 years (SD 4.9), and 15.2 years (SD 5.2) for subjects without previous self-reported injury (n1⁄4454). The risk of developing SROA over any five-year period was significantly increased for women who reported previous injury at baseline compared to injury-free women (OR 2.22 [95% CI 1.39, 3.54]). The time-varying hazard rate was significantly different between groups at year 10 and year 15 (p<0.05) (see figure). Conclusions: Self-report of even mild knee injury is significantly associated with an increased risk of developing clinically relevant symptomatic knee osteoarthritis up to twenty years after the original injury. Efforts to identify and prevent knee injuries of all levels of severity at a community level may help reduce the long-term risk of incident knee osteoarthritis.