Introduction

Middle-aged and elderly patients are often referred to an orthopaedic surgeon (OS) for knee pain. The appropriateness of knee imaging prior to referral is not well documented. Our hypothesis is that knee imaging prior to referral is often inappropriate or insufficient and varies by the referral source.

Methods

Six months of data was collected on patients over the age of 44 with knee pain presenting to the senior author’s clinic. Pre-referral imaging studies were determined to be appropriate, inappropriate or insufficient by the senior author. Imaging performed in the clinic were classified using the Kellgren-Lawrence system. Univariant and multivariant analysis was performed to assess differences in imaging study appropriateness among providers.

Results

500 patients were enrolled in this study with a mean age of 62.7 ± 10.7 [45-99]. There were 292 females and 208 males. Knee pain was left in 223 cases, right in 213 and bilateral in 64. 219 patients were referred to the OS by their primary care provider (PCP), 198 were self-referred and 83 were referred by some other source. For the patients referred by PCP (n=219), 100 patients (46.7%) had X-rays and only 14 (14%) were weight-bearing. Among patients referred by PCP with MRIs (n=49), 91.7% were inappropriate and 22.4% affected initial treatment. For the self-referred patients (n=198), 27 patients (13.6%) had X-rays and 7 (25.9%) were weight-bearing. Among self-referred patients with MRIs (n=4), 75% of MRIs were inappropriate and 25% of MRIs affected initial treatment. For the patients referred by others (n=83), 67 patients (80.7%) had X-rays and 44 (65.7%) were weight-bearing. Among patients with MRIs referred by others (n=31), 49.5% were inappropriate and 61.3% affected initial treatment. Compared to self-referred patients, patients referred by PCP were 49% less likely to have a weight-bearing X-ray while patients referred by others were 6.48 times more likely. Compared to self-referred patients, patients referred by PCP and others were 15.5 and 25.19 times more likely to have an MRI, respectively. Odds that PCP referred patients’ MRIs were inappropriate were 46% (p-value = 0.786) greater than self-referred, while patients referred by others were 88% (p-value = 0.189) less likely to have had an inappropriate MRI. The proportion of inappropriate MRIs for patients referred by others was significantly less (49.5%; p = 0.032) than PCP (91.7%; p-value = 0.813) or self-referred patients (88%; control). Patients referred by PCP had a statistically significant difference in mean KL score (2.4) compared to patients that were self-referred (2.1), P-value = 0.015). Patients referred by PCP had a statistically greater proportion of patients with KL score =3, 55.1% (p-value = 0.001), compared to self-referral (40.5%).

Discussion

These authors find a majority of the knee imaging performed on patients over 44 years old prior to referral is either inappropriate or insufficient and minimally impacts initial treatment. Patients referred by PCPs are more likely to have severe OA and inappropriate imaging. We encourage PCPs to order weight-bearing radiographs in this patient population and limit MRIs for patients with moderate to severe OA.