Identification of patients with suboptimal results after hip arthroplasty: development of a preliminary prediction algorithm

BACKGROUND: The ability to predict preoperatively the identity of patients undergoing hip arthroplasty at risk of suboptimal outcomes could help implement interventions targeted at improving surgical results. The objective was to develop a preliminary prediction algorithm (PA) allowing the identification of patients at risk of unsatisfactory outcomes one to two years following hip arthroplasty. METHODS: Retrospective data on a cohort of 265 patients having undergone primary unilateral hip replacement (188 total arthroplasties and 77 resurfacing arthroplasties) from 2004 to 2010 were collected from our arthroplasty database. Hip pain and function, as measured by the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) were collected, as well as self-reported hip joint perception after surgery. Demographic and clinical variables recorded at the time of the surgery were considered as potential predictors. Patients were considered as having a suboptimal surgical outcome if they were in the worst quartile of the postoperative total WOMAC score and perceived their operated hip as artificial with minimal or major limitations. The PA was developed using recursive partitioning. RESULTS: Mean postoperative surgical follow-up was 446 ± 171 days. Forty patients (15.1 %) had a postoperative total WOMAC score in the worst quartile (>11.5/100) and perceived their joint as artificial with minimal or major restrictions. A PA consisting of the following variables achieved the most acceptable level of prediction: gender, age at the time of surgery, body mass index (BMI), and three items of the preoperative WOMAC (degree of pain with walking on a flat surface and during the night as well as degree of difficulty with putting socks or stockings). The rule had a sensitivity of 75.0 % (95 % CI: 59.8-85.8), a specificity of 77.8 % (95 % CI: 71.9–82.7), a positive predictive value of 37.5 % (95 % CI: 27.7–48.5), a negative predictive value of 94.6 % (95 % CI: 90.3–97.0) and positive and negative likelihood ratios of 3.38 (95 % CI: 2.49–4.57) and 0.34 (95 % CI: 0.19–0.55) respectively. CONCLUSIONS: The preliminary PA shows promising results at identifying patients at risk of significant functional limitations, increased pain and inadequate joint perception after hip arthroplasty. Clinical use should not be implemented before additional validation and refining. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (doi:10.1186/s12891-015-0720-1) contains supplementary material, which is available to authorized users.