Appropriateness of magnetic resonance imaging requested by primary care physicians for patients with knee pain

Appropriateness of magnetic resonance imaging requested by primary care physicians for patients... Abstract Objective To evaluate the appropriateness of magnetic resonance imaging (MRI) of the knee requested by primary care physicians. Design Retrospective observational study. Setting Six primary care centres in the Elche Department of Health of the Valencian Community, Spain. Participants Three hundred patients with knee pain who were prescribed MRI. Main Outcome Measures Data were collected from the electronic clinical history, which allowed us to assess the appropriateness and inappropriateness of the MRI requests for the knee based on the American College of Radiology (ACR) criteria. A multivariate logistic regression model was used to identify factors associated with an inappropriate request. Results About 45% (41–49%) of knee MRI prescriptions were assessed as inappropriate. The frequency was higher in female patients (odds ratio, OR = 1.96; P = 0.03). A history of knee trauma and urgent use of MRI were associated with a lower frequency of inappropriate requests (OR = 0.14, P < 0.001 and OR = 0.32, P = 0.03, respectively). In 82% of cases, the request for MRI was deemed inappropriate because it was used as the initial imaging test. The availability of a previous radiograph of the knee significantly reduced the rate of inappropriate requests (OR = 0.05, P < 0.001); only 47% of the patients had a previous radiograph. Conclusions The percentage of inappropriate knee MRI prescriptions is high. Protocols should be put in place to improve the appropriateness of MRI requests by promoting understanding of the appropriate use of MRI among primary care physicians. appropriateness, knee jointmagnetic resonance imaging, primary healthcare, Spain Introduction Knee pain represents a frequent reason for consultation in primary care because of the large number of conditions (acute, chronic, traumatic and non-traumatic) that can affect this articulation, one of the largest, most complex and overloaded joints in the body. The prevalence of knee pain has been reported as 15–33% in various populations [1–6]. Radiography remains the first-choice imaging test for most patients with knee pain, although one study [7] has noted a progressive increase in the number of requests for magnetic resonance imaging (MRI) of the knee and a parallel decrease in the use of radiography. Given the importance of ligamentous and cartilaginous components in the structure and function of the knee, it is not surprising that MRI is a frequently ordered imaging technique. Knee MRI is a sensitive and non-invasive imaging test that provides an easy and accurate diagnosis of different anatomical abnormalities. However, the findings may cause confusion when they do not correlate with radiography, results of exploration or the patient’s clinical history [8]. MRI sometimes identifies incidental pathology unrelated to the symptoms, which can lead to unnecessary referral and treatment [9]. A recent review of the available evidence attributes little or no benefit to the use of routine MRI in the diagnosis of knee pathology [10]. Knee MRI is one of the most frequently performed tests in our health department and is exceeded in frequency only by MRI of the lumbar spine. Given the diversity of clinical practice, the progressive increase in the use of MRI in the last decade [11–14] and the low rate of appropriate requests for MRI observed for other musculoskeletal conditions [15], it is possible that this test is becoming a standard before a rigorous evaluation of its use has been published [16]. Furthermore, inappropriate requests for MRI represent a high cost for the health system and can lead to unnecessary interventions [17], which are even more costly and risky for patients. All Spaniards have access to the country’s universal healthcare system called the Spanish National Health System. It covers most healthcares free of charge. Unlike in other health systems, Spanish doctors are not financially incentivised to prescribe expensive prescriptions and treatments, so they are accustomed to recommending the cheapest course of action. However, family doctors can freely prescribe MRI in our region. We found no studies that have investigated the appropriateness of requests for MRI in the Spanish population and we chose to investigate this issue. The objectives of this study were to quantify the percentage of appropriate requests for knee MRI in primary care according to the American College of Radiology (ACR) criteria [18, 19] and to identify the variables associated with inappropriate requests. Methods During 2015, the Elche Health Department (Alicante, Spain) provided coverage to 163 583 people. Any physician from its six primary care centres can order knee MRI, which are performed on a Philips Intera 1.5T device (Koninklijke Philips NV, Amsterdam, the Netherlands) at the Department’s University General Hospital. Given the availability of an electronic medical record (EMR) for every MRI performed, we decided to conduct a retrospective cross-sectional study. Initially, the data for all MRI of the knee performed during 2015 were collected. After discarding those requested by specialists, we found 598 cases of knee MRI that had been requested by primary care physicians. Based on this population size and assuming a rate of 40% inappropriate requests based on a recent study conducted in a similar setting [20], a confidence interval (CI) of ±4% and a confidence level of 95%, a sample size of 300 patients was calculated as necessary. All patients were over 17 years old and there were no pregnant women in the sample. Simple random sampling was performed among all registered MRIs to select the sample. The EMR for each patient was reviewed, and a team member (Medical Doctor, Master in Public Health, General Practitioner for 7 years) collected the independent variables that were considered necessary to assess the appropriateness of MRI according to the ACR criteria: age; duration of knee pain (chronic if ≥3 months, acute if <3 months and indeterminate if unclear in the EMR); existence of patellofemoral symptoms; antecedent of a previous knee injury in the form of a fall, twist injury or traffic accident with suspected dislocation of the knee; ability to walk and support the body weight and knee radiography performed in the 3 months before the MRI request. Other details recorded included whether the patient had sustained a fracture or exhibited joint effusion, degenerative signs, crystals, avascular necrosis, inflammatory signs or any evidence of internal disarray of the joint. In addition, other independent variables not needed to evaluate the MRI request according to the ACR criteria were recorded to check whether they were related to inappropriate request for knee MRI, such as sex, MRI priority (normal or urgent), primary care centre originating the request, number of referrals for every patient and surgical indications. With the exception of age, all variables were categorical. Cases with missing values were discarded. To evaluate the appropriateness of each MRI request, the same team member who collected the data compared the values for the variables for every patient against the ACR criteria for both traumatic [18] and non-traumatic knee [19] injury. The quantitative variable age is expressed as median and interquartile range (IQR, in P25–P75 format). The Shapiro–Wilk test showed that age and sex were non-normally distributed, and these variables were analysed using the Mann–Witney U test. Bivariate analysis was performed, followed by multivariate analysis using backward stepwise regression. Significance was accepted at P < 0.05. The results are expressed as the odds ratio (OR) with respect to a reference group. To evaluate the adjustment of the model developed, Pearson’s chi-square test and the Hosmer–Lemeshow test were used. Collinearity was assessed using the variance inflation factor. Statistical analysis was performed in STATA (Version 14, StataCorp, College Station, TX, USA). Results We included 300 patients: 144 (48%) women and 156 (52%) men. Their median age was 53 years (IQR = 39, 66) for women and 46 years (IQR = 31, 57) for men. The difference in age between men and women was significant (P = 0.002). Table 1 shows the frequencies and percentages of each ACR variant in the sample, grouped according to the appropriateness and inappropriateness of the request for MRI and according to the type of injury—traumatic and non-traumatic. Overall, the MRI prescription was assessed as inappropriate in 135 of cases (45%) and as appropriate in 165 cases (55%). Table 1 Knee MRI appropriateness and inappropriateness causes in our samplea Description ACR variant N (%) Appropriateness causes  Non-traumatic knee pain Adult: non-trauma and non-localised pain. Initial knee radiographs are negative or demonstrate a joint effusion 6 52 (31.5) Child or adult: patellofemoral (anterior) symptoms. Initial knee radiographs are negative or demonstrate a joint effusion 5 40 (24.2) Adult: Initial knee radiographs demonstrate inflammatory, crystalline or degenerative joint disease (uni- to tri-compartmental sclerosis, hypertrophic spurs, joint space narrowing and/or subchondral cysts) 9b 19 (11.5)  Acute trauma to the knee Adult or child >1-year-old. Fall or twisting injury with either no fracture or a Segond fracture seen on a radiograph, suspect internal derangement. Next study 3 32 (19.4) Adult or child >1-year-old. Injury to knee, mechanism unknown. Focal patellar tenderness, effusion, able to walk 5 14 (8.5) Adult or child >1-year-old. Fall or twisting injury, with one or more of the following: focal tenderness, effusion, inability to bear weight. First study 2 4 (2.4) Adult or child >1 year old. Significant trauma to the knee from motor vehicle accident, suspect knee dislocation 6 4 (2.4) Inappropriateness causes  Non-traumatic knee pain Adult: non-trauma and non-localised pain. Initial examination 3 52 (38.5) Child or adult: patellofemoral (anterior) symptoms. Initial examination 2 43 (31.9) Adult: Initial knee radiographs demonstrate inflammatory, crystalline or degenerative joint disease (uni- to tri-compartmental sclerosis, hypertrophic spurs, joint space narrowing and/or subchondral cysts) 9b 18 (13.3) Adult: patellofemoral (anterior) symptoms. Initial knee radiographs demonstrate degenerative joint disease and/or chondrocalcinosis 8 4 (3.0) Othersc 3 (2.2)  Acute trauma to the knee Adult or child >1-year-old. Fall or twisting injury, no focal tenderness, no effusion; able to walk. First study 1 15 (11.1) Description ACR variant N (%) Appropriateness causes  Non-traumatic knee pain Adult: non-trauma and non-localised pain. Initial knee radiographs are negative or demonstrate a joint effusion 6 52 (31.5) Child or adult: patellofemoral (anterior) symptoms. Initial knee radiographs are negative or demonstrate a joint effusion 5 40 (24.2) Adult: Initial knee radiographs demonstrate inflammatory, crystalline or degenerative joint disease (uni- to tri-compartmental sclerosis, hypertrophic spurs, joint space narrowing and/or subchondral cysts) 9b 19 (11.5)  Acute trauma to the knee Adult or child >1-year-old. Fall or twisting injury with either no fracture or a Segond fracture seen on a radiograph, suspect internal derangement. Next study 3 32 (19.4) Adult or child >1-year-old. Injury to knee, mechanism unknown. Focal patellar tenderness, effusion, able to walk 5 14 (8.5) Adult or child >1-year-old. Fall or twisting injury, with one or more of the following: focal tenderness, effusion, inability to bear weight. First study 2 4 (2.4) Adult or child >1 year old. Significant trauma to the knee from motor vehicle accident, suspect knee dislocation 6 4 (2.4) Inappropriateness causes  Non-traumatic knee pain Adult: non-trauma and non-localised pain. Initial examination 3 52 (38.5) Child or adult: patellofemoral (anterior) symptoms. Initial examination 2 43 (31.9) Adult: Initial knee radiographs demonstrate inflammatory, crystalline or degenerative joint disease (uni- to tri-compartmental sclerosis, hypertrophic spurs, joint space narrowing and/or subchondral cysts) 9b 18 (13.3) Adult: patellofemoral (anterior) symptoms. Initial knee radiographs demonstrate degenerative joint disease and/or chondrocalcinosis 8 4 (3.0) Othersc 3 (2.2)  Acute trauma to the knee Adult or child >1-year-old. Fall or twisting injury, no focal tenderness, no effusion; able to walk. First study 1 15 (11.1) aAmerican College of Radiology. ACR appropriateness criteria. bACR variant 9 considers MRI usually not appropriate, except when preoperative assessment is involved. cMRI prescriptions not matching any ACR criteria. Table 1 Knee MRI appropriateness and inappropriateness causes in our samplea Description ACR variant N (%) Appropriateness causes  Non-traumatic knee pain Adult: non-trauma and non-localised pain. Initial knee radiographs are negative or demonstrate a joint effusion 6 52 (31.5) Child or adult: patellofemoral (anterior) symptoms. Initial knee radiographs are negative or demonstrate a joint effusion 5 40 (24.2) Adult: Initial knee radiographs demonstrate inflammatory, crystalline or degenerative joint disease (uni- to tri-compartmental sclerosis, hypertrophic spurs, joint space narrowing and/or subchondral cysts) 9b 19 (11.5)  Acute trauma to the knee Adult or child >1-year-old. Fall or twisting injury with either no fracture or a Segond fracture seen on a radiograph, suspect internal derangement. Next study 3 32 (19.4) Adult or child >1-year-old. Injury to knee, mechanism unknown. Focal patellar tenderness, effusion, able to walk 5 14 (8.5) Adult or child >1-year-old. Fall or twisting injury, with one or more of the following: focal tenderness, effusion, inability to bear weight. First study 2 4 (2.4) Adult or child >1 year old. Significant trauma to the knee from motor vehicle accident, suspect knee dislocation 6 4 (2.4) Inappropriateness causes  Non-traumatic knee pain Adult: non-trauma and non-localised pain. Initial examination 3 52 (38.5) Child or adult: patellofemoral (anterior) symptoms. Initial examination 2 43 (31.9) Adult: Initial knee radiographs demonstrate inflammatory, crystalline or degenerative joint disease (uni- to tri-compartmental sclerosis, hypertrophic spurs, joint space narrowing and/or subchondral cysts) 9b 18 (13.3) Adult: patellofemoral (anterior) symptoms. Initial knee radiographs demonstrate degenerative joint disease and/or chondrocalcinosis 8 4 (3.0) Othersc 3 (2.2)  Acute trauma to the knee Adult or child >1-year-old. Fall or twisting injury, no focal tenderness, no effusion; able to walk. First study 1 15 (11.1) Description ACR variant N (%) Appropriateness causes  Non-traumatic knee pain Adult: non-trauma and non-localised pain. Initial knee radiographs are negative or demonstrate a joint effusion 6 52 (31.5) Child or adult: patellofemoral (anterior) symptoms. Initial knee radiographs are negative or demonstrate a joint effusion 5 40 (24.2) Adult: Initial knee radiographs demonstrate inflammatory, crystalline or degenerative joint disease (uni- to tri-compartmental sclerosis, hypertrophic spurs, joint space narrowing and/or subchondral cysts) 9b 19 (11.5)  Acute trauma to the knee Adult or child >1-year-old. Fall or twisting injury with either no fracture or a Segond fracture seen on a radiograph, suspect internal derangement. Next study 3 32 (19.4) Adult or child >1-year-old. Injury to knee, mechanism unknown. Focal patellar tenderness, effusion, able to walk 5 14 (8.5) Adult or child >1-year-old. Fall or twisting injury, with one or more of the following: focal tenderness, effusion, inability to bear weight. First study 2 4 (2.4) Adult or child >1 year old. Significant trauma to the knee from motor vehicle accident, suspect knee dislocation 6 4 (2.4) Inappropriateness causes  Non-traumatic knee pain Adult: non-trauma and non-localised pain. Initial examination 3 52 (38.5) Child or adult: patellofemoral (anterior) symptoms. Initial examination 2 43 (31.9) Adult: Initial knee radiographs demonstrate inflammatory, crystalline or degenerative joint disease (uni- to tri-compartmental sclerosis, hypertrophic spurs, joint space narrowing and/or subchondral cysts) 9b 18 (13.3) Adult: patellofemoral (anterior) symptoms. Initial knee radiographs demonstrate degenerative joint disease and/or chondrocalcinosis 8 4 (3.0) Othersc 3 (2.2)  Acute trauma to the knee Adult or child >1-year-old. Fall or twisting injury, no focal tenderness, no effusion; able to walk. First study 1 15 (11.1) aAmerican College of Radiology. ACR appropriateness criteria. bACR variant 9 considers MRI usually not appropriate, except when preoperative assessment is involved. cMRI prescriptions not matching any ACR criteria. The main causes of inappropriateness were the use of MRI as the initial imaging technique when it was not indicated. For example, in the initial evaluation of non-traumatic and non-localised knee pain (inappropriate use of MRI for variant 3 of ACR for non-traumatic knee injury, comprising 39% of inappropriate requests), initial study of patellofemoral symptoms (inappropriate use of MRI for variant 2 of non-traumatic ACR, 32% of inappropriate requests) and pain after a fall or twist without spill or focalised pain (use MRI for ACR variant 1 for traumatic knee, 11% of inappropriate requests). Another important reason was the inappropriate use of MRI for ACR variant 9 for non-traumatic knee injury when not considering surgery (13%). As shown in Table 2, the bivariate analysis of the association between inappropriate requests and independent variables found no significant differences in sex, age, clinical signs (meniscal, arthrosis, blockage or spills) or priority of the MRI request. By contrast, there were significant differences in cases for which a knee radiograph was available before the MRI request (OR = 0.08; P < 0.001), when there was evidence of previous trauma (OR = 0.26, P < 0.001), when the duration of symptoms was not accurately determined (OR = 2.27; P = 0.03) and when the MRI request came from a specific primary care centre (OR = 0.28, P = 0.005). Table 2 Knee MRI Inappropriateness: bivariate analysis Inappropriateness Appropriate MRI (N) Inappropriate MRI (N) Odds ratio CI 95% P-value Patient sex  Men 90 66 1 (ref)  Women 75 69 1.25 0.80–1.98 0.33 Patient age (years) <24 16 15 1 (ref)  25–44 47 40 0.91 0.40–2.06 0.82  45–64 69 54 0.83 0.38–1.84 0.65  65–74 25 16 0.68 0.27–1.75 0.43 >75 8 10 1.33 0.42–4.28 0.63 Patellofemoral symptoms  No 122 110 1 (ref)  Yes 43 25 0.64 0.37–1.12 0.12 Degenerative joint disease  No 150 118 1 (ref)  Yes 15 17 1.44 0.69–3.00 0.33 Joint blockages  No 156 128 1 (ref)  Yes 9 7 0.95 0.34–2.62 0.92 Joint efussion  No 147 125 1 (ref)  Yes 18 10 0.65 0.29–1.47 0.3 Duration of symptoms  Acute 99 80 1 (ref)  Chronic 54 35 0.75 0.45–1.28 0.3  Undetermined 12 20 2.27 1.06–4.86 0.035 MRI priority  Normal 151 124 1 (ref)  Urgent 14 11 0.96 0.42–2.18 0.92 Previous radiograph  No 32 102 1 (ref)  Yes 133 33 0.08 0.04–0.13 <0.001 Previous trauma  No 111 120 1 (ref)  Yes 54 15 0.26 0.14–0.48 <0.001 Primary care centre  1 35 41 1 (ref)  2 48 43 0.76 0.42–1.41 0.39  3 17 13 0.65 0.28–1.53 0.33  4 18 16 0.76 0.34–1.71 0.51  5 20 13 0.55 0.24–1.27 0.17  6 27 9 0.28 0.12–0.69 0.005 GP sex  Men 80 58 1 (ref)  Women 87 75 1.19 0.75–1.88 0.46 GP age (years)a <45 48 39 1 (ref)  45–54 28 15 1.52 0.71–3.23 0.28  55–64 76 54 0.89 0.53–1.52 0.68 Inappropriateness Appropriate MRI (N) Inappropriate MRI (N) Odds ratio CI 95% P-value Patient sex  Men 90 66 1 (ref)  Women 75 69 1.25 0.80–1.98 0.33 Patient age (years) <24 16 15 1 (ref)  25–44 47 40 0.91 0.40–2.06 0.82  45–64 69 54 0.83 0.38–1.84 0.65  65–74 25 16 0.68 0.27–1.75 0.43 >75 8 10 1.33 0.42–4.28 0.63 Patellofemoral symptoms  No 122 110 1 (ref)  Yes 43 25 0.64 0.37–1.12 0.12 Degenerative joint disease  No 150 118 1 (ref)  Yes 15 17 1.44 0.69–3.00 0.33 Joint blockages  No 156 128 1 (ref)  Yes 9 7 0.95 0.34–2.62 0.92 Joint efussion  No 147 125 1 (ref)  Yes 18 10 0.65 0.29–1.47 0.3 Duration of symptoms  Acute 99 80 1 (ref)  Chronic 54 35 0.75 0.45–1.28 0.3  Undetermined 12 20 2.27 1.06–4.86 0.035 MRI priority  Normal 151 124 1 (ref)  Urgent 14 11 0.96 0.42–2.18 0.92 Previous radiograph  No 32 102 1 (ref)  Yes 133 33 0.08 0.04–0.13 <0.001 Previous trauma  No 111 120 1 (ref)  Yes 54 15 0.26 0.14–0.48 <0.001 Primary care centre  1 35 41 1 (ref)  2 48 43 0.76 0.42–1.41 0.39  3 17 13 0.65 0.28–1.53 0.33  4 18 16 0.76 0.34–1.71 0.51  5 20 13 0.55 0.24–1.27 0.17  6 27 9 0.28 0.12–0.69 0.005 GP sex  Men 80 58 1 (ref)  Women 87 75 1.19 0.75–1.88 0.46 GP age (years)a <45 48 39 1 (ref)  45–54 28 15 1.52 0.71–3.23 0.28  55–64 76 54 0.89 0.53–1.52 0.68 aGP age was unknown in 23 cases. Table 2 Knee MRI Inappropriateness: bivariate analysis Inappropriateness Appropriate MRI (N) Inappropriate MRI (N) Odds ratio CI 95% P-value Patient sex  Men 90 66 1 (ref)  Women 75 69 1.25 0.80–1.98 0.33 Patient age (years) <24 16 15 1 (ref)  25–44 47 40 0.91 0.40–2.06 0.82  45–64 69 54 0.83 0.38–1.84 0.65  65–74 25 16 0.68 0.27–1.75 0.43 >75 8 10 1.33 0.42–4.28 0.63 Patellofemoral symptoms  No 122 110 1 (ref)  Yes 43 25 0.64 0.37–1.12 0.12 Degenerative joint disease  No 150 118 1 (ref)  Yes 15 17 1.44 0.69–3.00 0.33 Joint blockages  No 156 128 1 (ref)  Yes 9 7 0.95 0.34–2.62 0.92 Joint efussion  No 147 125 1 (ref)  Yes 18 10 0.65 0.29–1.47 0.3 Duration of symptoms  Acute 99 80 1 (ref)  Chronic 54 35 0.75 0.45–1.28 0.3  Undetermined 12 20 2.27 1.06–4.86 0.035 MRI priority  Normal 151 124 1 (ref)  Urgent 14 11 0.96 0.42–2.18 0.92 Previous radiograph  No 32 102 1 (ref)  Yes 133 33 0.08 0.04–0.13 <0.001 Previous trauma  No 111 120 1 (ref)  Yes 54 15 0.26 0.14–0.48 <0.001 Primary care centre  1 35 41 1 (ref)  2 48 43 0.76 0.42–1.41 0.39  3 17 13 0.65 0.28–1.53 0.33  4 18 16 0.76 0.34–1.71 0.51  5 20 13 0.55 0.24–1.27 0.17  6 27 9 0.28 0.12–0.69 0.005 GP sex  Men 80 58 1 (ref)  Women 87 75 1.19 0.75–1.88 0.46 GP age (years)a <45 48 39 1 (ref)  45–54 28 15 1.52 0.71–3.23 0.28  55–64 76 54 0.89 0.53–1.52 0.68 Inappropriateness Appropriate MRI (N) Inappropriate MRI (N) Odds ratio CI 95% P-value Patient sex  Men 90 66 1 (ref)  Women 75 69 1.25 0.80–1.98 0.33 Patient age (years) <24 16 15 1 (ref)  25–44 47 40 0.91 0.40–2.06 0.82  45–64 69 54 0.83 0.38–1.84 0.65  65–74 25 16 0.68 0.27–1.75 0.43 >75 8 10 1.33 0.42–4.28 0.63 Patellofemoral symptoms  No 122 110 1 (ref)  Yes 43 25 0.64 0.37–1.12 0.12 Degenerative joint disease  No 150 118 1 (ref)  Yes 15 17 1.44 0.69–3.00 0.33 Joint blockages  No 156 128 1 (ref)  Yes 9 7 0.95 0.34–2.62 0.92 Joint efussion  No 147 125 1 (ref)  Yes 18 10 0.65 0.29–1.47 0.3 Duration of symptoms  Acute 99 80 1 (ref)  Chronic 54 35 0.75 0.45–1.28 0.3  Undetermined 12 20 2.27 1.06–4.86 0.035 MRI priority  Normal 151 124 1 (ref)  Urgent 14 11 0.96 0.42–2.18 0.92 Previous radiograph  No 32 102 1 (ref)  Yes 133 33 0.08 0.04–0.13 <0.001 Previous trauma  No 111 120 1 (ref)  Yes 54 15 0.26 0.14–0.48 <0.001 Primary care centre  1 35 41 1 (ref)  2 48 43 0.76 0.42–1.41 0.39  3 17 13 0.65 0.28–1.53 0.33  4 18 16 0.76 0.34–1.71 0.51  5 20 13 0.55 0.24–1.27 0.17  6 27 9 0.28 0.12–0.69 0.005 GP sex  Men 80 58 1 (ref)  Women 87 75 1.19 0.75–1.88 0.46 GP age (years)a <45 48 39 1 (ref)  45–54 28 15 1.52 0.71–3.23 0.28  55–64 76 54 0.89 0.53–1.52 0.68 aGP age was unknown in 23 cases. Logistic regression analysis was performed on all collected variables to identify those that were independently associated with the inappropriate request for MRI after adjusting for possible confounding factors. The results are shown in Table 3. One variable, female sex, was associated with a higher percentage of inappropriate requests (OR = 1.96; P = 0.03). Three variables were associated with a lower percentage of inappropriate requests: previous knee radiograph (OR = 0.05; P < 0.001), knee trauma (OR = 0.14; P < 0.001) and urgent MRI request (OR = 0.32; P = 0.03). Table 3 Factors associated to knee MRI inappropriateness N Odds ratio CI 95% P-value Previous radiograph  No 134 1 (ref)  Yes 166 0.05 0.026–0.097 <0.001 Previous trauma  No 231 1 (ref)  Yes 69 0.14 0.06–0.31 <0.001 Sex  Men 156 1 (ref)  Women 144 1.96 1.07–3.57 0.028 MRI priority  Normal 275 1 (ref)  Urgent 25 0.32 0.11–0.91 0.033 N Odds ratio CI 95% P-value Previous radiograph  No 134 1 (ref)  Yes 166 0.05 0.026–0.097 <0.001 Previous trauma  No 231 1 (ref)  Yes 69 0.14 0.06–0.31 <0.001 Sex  Men 156 1 (ref)  Women 144 1.96 1.07–3.57 0.028 MRI priority  Normal 275 1 (ref)  Urgent 25 0.32 0.11–0.91 0.033 Table 3 Factors associated to knee MRI inappropriateness N Odds ratio CI 95% P-value Previous radiograph  No 134 1 (ref)  Yes 166 0.05 0.026–0.097 <0.001 Previous trauma  No 231 1 (ref)  Yes 69 0.14 0.06–0.31 <0.001 Sex  Men 156 1 (ref)  Women 144 1.96 1.07–3.57 0.028 MRI priority  Normal 275 1 (ref)  Urgent 25 0.32 0.11–0.91 0.033 N Odds ratio CI 95% P-value Previous radiograph  No 134 1 (ref)  Yes 166 0.05 0.026–0.097 <0.001 Previous trauma  No 231 1 (ref)  Yes 69 0.14 0.06–0.31 <0.001 Sex  Men 156 1 (ref)  Women 144 1.96 1.07–3.57 0.028 MRI priority  Normal 275 1 (ref)  Urgent 25 0.32 0.11–0.91 0.033 The model based on these four variables could explain 32.3% of the variance (calculated using pseudo R2) and was found to provide a satisfactory fit with both Pearson’s chi-square test (P = 0.63) and the Hosmer–Lemeshow test (P = 0.50 with six quantiles). To evaluate collinearity, the variance inflation factor showed a mean value of 1.34. The 300 patients with knee pain included in the sample were associated with 232 referrals to other services. The most frequent were directed to the Orthopaedic Surgery and Traumatology Service (n = 202, 87%) and Rehabilitation Services (n = 28, 12%). Of all referrals, the MRI request was considered appropriate for 147 (i.e. 0.89 per each inappropriate MRI) and 85 referrals were considered inappropriate. In 72 (36%) of the referrals to the Orthopaedic Surgery and Traumatology Service, a surgical solution to the knee pain was proposed: the MRI request was assessed as appropriate for 51 (71%) and as inappropriate for the other 21 (29%). Finally, we analysed the relationship between the MRI findings and the appropriateness of the request. There was some pathological finding in 102 (76%) of the cases for which MRI was considered inappropriate and in 144 (87%) of those for which MRI was considered appropriate. No pathological findings were observed in 15 (11%) of the group for which MRI was considered appropriate and in 17 (10%) of the group for which MRI was considered inappropriate. Discussion About 45% (41–49%) of the knee MRI requested in this primary care setting did not seem justified from the viewpoint of their fulfilling the ACR criteria. Comparison of our results with those reported by others is difficult because there are few studies of the appropriateness of knee MRI and no reports on this in the Spanish population. A recent study by Solivetti et al. [20] from Italy reported a rate of 40% inappropriate requests for MRI, which is similar to that found in our study. In that study, most of the requests for knee MRI came from orthopaedic surgeons (44.3%) or other specialists (19.5%), and only 36.3% of the knee MRI requests came from primary care physicians. The study by Solivetti et al. is consistent with other studies [21, 22] in attributing the higher percentage of appropriate MRI requests to those requested by orthopaedic surgeons, which may explain why their rates are slightly lower than that found in our study. Vedjani et al. [23] estimated a rate of 46.7% for the inappropriate use of knee MRI in an Iranian hospital, which is also similar to the rate found in our study, even though their criteria were different, their sample was smaller and MRI was requested by a specialist in 98.7% of cases. Oikarinen et al. [24] reported a 10% rate of inappropriate MRI requests in one Finnish hospital, but their sample included only 30 knee MRI cases. Petron et al. [25] found that only 12 of 100 MRIs ordered by primary care providers in Utah (USA) would have been ordered by an orthopaedist given the documented data. In our study, the main reason for considering the MRI request inappropriate was its use as the initial imaging test, which accounted for 82% of the requests considered inappropriate. This result suggests that using MRI as the initial imaging mode increases the chance of the request being inappropriate according to the ACR criteria, which recommend radiography as the most appropriate initial imaging mode in patients with knee pain without a history of trauma. If the radiograph demonstrates deposits of crystals or an inflammatory or degenerative pathology of the joint, no further imaging tests are indicated. MRI is indicated only when a possible surgical intervention is being considered or when the findings cannot explain the patient’s symptoms. For traumatic knee pathology, the ACR also recommends radiology as the initial mode, with the sole exception of significant trauma—such as that received in motor vehicle accidents—for which MRI is considered appropriate as the initial mode for assessing the extent of damage to ligaments and other support structures. Despite this recommendation, 47% of our patients did not have any previous radiograph. This percentage is double that observed in another study [26], and this indicates a clear opportunity for improvement in our environment. Roberts et al. [22] already pointed out the concern among the orthopaedic surgical community about an increasing number of knee MRIs being ordered in the primary care setting without prior radiographs. Although requesting a knee MRI before performing other tests and examinations may be an inappropriate request, the primary care physician may be tempted to do so at an early stage in the evolution of the knee pain to provide the specialist doctor the results at the patient’s first visit, usually several weeks after the referral is made. MRI requests are frequently made by specialists in the private healthcare sector, so the patient may pressure the primary care physician to receive this service at no cost through the public health system. Our observations agree with those of some studies [20, 27, 28], but further studies are needed to assess the role of these and other factors in the prescribing physician’s decision to request an MRI for patients with knee pain. We consider that the high percentage of inappropriate requests for MRI may be a starting point for assessing the costs associated with inappropriate prescription and considering the number of referrals. Other data obtained from the EMR were also associated with a higher percentage of inappropriate requests for knee MRI. However, some of these are not modifiable, such as female sex and the existence of previous trauma. Regarding the female sex, in our sample there were no pregnant women, so we considered that the decision to prescribe MRI was not influenced by the fear of radiation. A lower percentage of inappropriate MRI requests was observed when the request was urgent, but it seems likely that such a priority would be reserved for the most serious cases. Other significant and non-modifiable factors associated with higher percentages of inappropriate MRI requests were knee pain whose duration could not be established from the EMR or history that did not generate any referral and that did not eventually lead to a surgical indication. A limitation of this study is the choice of the ACR criteria for assessing the appropriateness of knee MRI requests. The US health system encourages the use of diagnostic procedures, which may introduce some laxity in other countries in terms of their criteria for determining appropriateness [29] and could increase the inappropriate use of MRI, as observed in this study. Measures to reduce the over-prescription of MRI for knee injury include reassessment of the knee MRI prescription procedure, development of clear guidelines and protocols, improving the training of primary care physicians to encourage their adherence to these guidelines, meetings for case discussion and providing individualised feedback to professionals. Conclusions Our study shows, for the first time in Spain that a high percentage of knee MRI requests by primary care physicians are inappropriate. These findings suggest that having a radiograph of the knee available before the MRI request may reduce the number of inappropriate requests for MRI because the risk is greatly reduced when a previous knee radiograph is available; this happened in only 47% of the sample reported here. However, these results should be interpreted within the context of the specific circumstances of our health system: freedom of MRI prescription, without incentives for prescribers and at no cost for patients. References 1 Noormohammadpour P , Mansournia MA , Koohpayehzadeh J et al. . Prevalence of chronic neck pain, low back pain and knee pain and their related factors in community-dwelling adults in Iran: a population-based national study . Clin J Pain 2017 ; 33 : 181 – 7 . Google Scholar CrossRef Search ADS PubMed 2 Turkiewicz A , Gerhardsson de Verdier M , Engström G et al. . Prevalence of knee pain and knee OA in southern Sweden and the proportion that seeks medical care . Rheumatology (Oxford) 2015 ; 54 : 827 – 35 . Google Scholar CrossRef Search ADS PubMed 3 United States Bone and Joint Initiative: The Burden of Musculoskeletal Diseases in the United States (BMUS). [Internet]. 2014 . Third edition: http://www.boneandjointburden.org/2014-report/ (03 October 2018, date last accessed). 4 Yoshimura N , Akune T , Fujiwara S et al. . Prevalence of knee pain, lumbar pain and its coexistence in Japanese men and women: the longitudinal cohorts of motor system organ (LOCOMO) study . J Bone Miner Metab 2014 ; 32 : 524 – 32 . Google Scholar CrossRef Search ADS PubMed 5 Jhun HJ , Sung NJ , Kim SY . Knee pain and its severity in elderly Koreans: prevalence, risk factors and impact on quality of life . J Korean Med Sci 2013 ; 28 : 1807 – 13 . Google Scholar CrossRef Search ADS PubMed 6 Thiem U , Lamsfuß R , Günther S et al. . Prevalence of self-reported pain, joint complaints and knee or hip complaints in adults aged ≥ 40 years: a cross-sectional survey in Herne, Germany . PLoS One 2013 ; 8 : e60753 . Google Scholar CrossRef Search ADS PubMed 7 Britt H , Miller G , Valenti L et al. . Evaluation of imaging ordering by general practitioners in Australia, 2002–03 to 2011–12 . General Practice Series no.35. Sydney : Sydney University Press , 2014 . 8 ACR–SPR–SSR Practice parameter for the performance and interpretation of magnetic resonance imaging (MRI) of the knee [Internet]. 2015 . Available from: https://www.acr.org/-/media/ACR/Files/Practice-Parameters/mr-knee.pdf (03 October 2018, date last accessed). 9 Pompan DC . Appropriate use of MRI for evaluating common musculoskeletal conditions . Am Fam Physician 2011 ; 83 : 883 – 4 . Google Scholar PubMed 10 Karel YHJM , Verkerk K , Endenburg S et al. . Effect of routine diagnostic imaging for patients with musculoskeletal disorders: A meta-analysis . Eur J Intern Med 2015 ; 26 : 585 – 95 . Google Scholar CrossRef Search ADS PubMed 11 Solomon DH , Katz JN , Carrino JA et al. . Trends in knee magnetic resonance imaging . Med Care 2003 ; 41 : 687 – 92 . Google Scholar PubMed 12 Espeland A , Natvig NL , Løge I et al. . Magnetic resonance imaging of the knee in Norway 2002–2004 (national survey): rapid increase, older patients, large geographic differences . BMC Health Serv Res 2007 ; 7 : 115 . Google Scholar CrossRef Search ADS PubMed 13 Baker LC , Atlas SW , Afendulis CC . Expanded use of imaging technology and the challenge of measuring value . Health Aff (Millwood) 2008 ; 27 : 1467 – 78 . Google Scholar CrossRef Search ADS PubMed 14 Smith-Bindman R , Miglioretti DL , Larson EB . Rising use of diagnostic medical imaging in a large integrated health system . Health Aff (Millwood) 2008 ; 27 : 1491 – 502 . Google Scholar CrossRef Search ADS PubMed 15 Emery DJ , Shojania KG , Forster AJ et al. . Overuse of magnetic resonance imaging . JAMA Intern Med 2013 ; 173 : 823 . Google Scholar CrossRef Search ADS PubMed 16 Brealey S , Russell I , Gilbert F et al. . Value of knee imaging by GPs requires rigorous assessment . BMJ 2002 ; 325 : 1242 . Google Scholar CrossRef Search ADS PubMed 17 Ben-Galim P , Steinberg EL , Amir H et al. . Accuracy of magnetic resonance imaging of the knee and unjustified surgery . Clin Orthop Relat Res 2006 ; 447 : 100 – 4 . Google Scholar CrossRef Search ADS PubMed 18 ACR Appropriateness Criteria®. Acute Trauma to the Knee. [Internet]. American College of Radiology. 2014 . Available from: https://acsearch.acr.org/docs/69419/Narrative/ (03 October 2018, date last accessed). 19 ACR Appropriateness Criteria®. Nontraumatic Knee Pain. [Internet]. American College of Radiology. 2012. Available from: https://acsearch.acr.org/docs/69432/Narrative/ (03 October 2018, date last accessed). 20 Solivetti FM , Guerrisi A , Salducca N et al. . Appropriateness of knee MRI prescriptions: clinical, economic and technical issues . Radiol Med 2016 ; 121 : 315 – 22 . Google Scholar CrossRef Search ADS PubMed 21 Wylie JD , Crim JR , Working ZM et al. . Physician provider type influences utilization and diagnostic utility of magnetic resonance imaging of the knee . J Bone Joint Surg Am 2015 ; 97 : 56 – 62 . Google Scholar CrossRef Search ADS PubMed 22 Roberts TT , Singer N , Hushmendy S et al. . MRI for the evaluation of knee pain: comparison of ordering practices of primary care physicians and orthopaedic surgeons . J Bone Joint Surg Am 2015 ; 97 : 709 – 14 . Google Scholar CrossRef Search ADS PubMed 23 Vejdani M , Molavi Taleghani Y , Kachoei AR et al. . Appropriateness of knee MRI prescriptions in a hospital of Birjand, 2014 . Int J Sci Rep 2015 ; 1 : 299 . Google Scholar CrossRef Search ADS 24 Oikarinen H , Karttunen A , Pääkkö E et al. . Survey of inappropriate use of magnetic resonance imaging . Insights Imaging 2013 ; 4 : 729 – 33 . Google Scholar CrossRef Search ADS PubMed 25 Petron DJ , Greis PE , Aoki SK et al. . Use of knee magnetic resonance imaging by primary care physicians in patients aged 40 years and older . Sports Health 2010 ; 2 : 385 – 90 . Google Scholar CrossRef Search ADS PubMed 26 George E , Tsipas S , Wozniak G et al. . MRI of the knee and shoulder performed before radiography . J Am College Radiol 2014 ; 11 : 1053 – 8 . Google Scholar CrossRef Search ADS 27 Robling M , Kinnersley P , Houston H et al. . An exploration of GPs’ use of MRI: a critical incident study . Fam Pract 1998 ; 15 : 236 – 43 . Google Scholar CrossRef Search ADS PubMed 28 Cohen O , Kahan E , Zalewski S et al. . Medical investigations requested by patients: how do primary care physicians react? Fam Med 1999 ; 31 : 426 – 31 . Google Scholar PubMed 29 Rodríguez Recio FJ , Sanz JC , Peiró S et al. . Utilización inapropiada de la resonancia magnética lumbar en un área de salud . Radiología 1999 ; 41 : 553 – 6 . © The Author(s) 2018. Published by Oxford University Press in association with the International Society for Quality in Health Care. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png International Journal for Quality in Health Care Oxford University Press

Appropriateness of magnetic resonance imaging requested by primary care physicians for patients with knee pain

Loading next page...
 
/lp/ou_press/appropriateness-of-magnetic-resonance-imaging-requested-by-primary-lurfsjA0o8
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press in association with the International Society for Quality in Health Care. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
1353-4505
eISSN
1464-3677
D.O.I.
10.1093/intqhc/mzy067
Publisher site
See Article on Publisher Site

Abstract

Abstract Objective To evaluate the appropriateness of magnetic resonance imaging (MRI) of the knee requested by primary care physicians. Design Retrospective observational study. Setting Six primary care centres in the Elche Department of Health of the Valencian Community, Spain. Participants Three hundred patients with knee pain who were prescribed MRI. Main Outcome Measures Data were collected from the electronic clinical history, which allowed us to assess the appropriateness and inappropriateness of the MRI requests for the knee based on the American College of Radiology (ACR) criteria. A multivariate logistic regression model was used to identify factors associated with an inappropriate request. Results About 45% (41–49%) of knee MRI prescriptions were assessed as inappropriate. The frequency was higher in female patients (odds ratio, OR = 1.96; P = 0.03). A history of knee trauma and urgent use of MRI were associated with a lower frequency of inappropriate requests (OR = 0.14, P < 0.001 and OR = 0.32, P = 0.03, respectively). In 82% of cases, the request for MRI was deemed inappropriate because it was used as the initial imaging test. The availability of a previous radiograph of the knee significantly reduced the rate of inappropriate requests (OR = 0.05, P < 0.001); only 47% of the patients had a previous radiograph. Conclusions The percentage of inappropriate knee MRI prescriptions is high. Protocols should be put in place to improve the appropriateness of MRI requests by promoting understanding of the appropriate use of MRI among primary care physicians. appropriateness, knee jointmagnetic resonance imaging, primary healthcare, Spain Introduction Knee pain represents a frequent reason for consultation in primary care because of the large number of conditions (acute, chronic, traumatic and non-traumatic) that can affect this articulation, one of the largest, most complex and overloaded joints in the body. The prevalence of knee pain has been reported as 15–33% in various populations [1–6]. Radiography remains the first-choice imaging test for most patients with knee pain, although one study [7] has noted a progressive increase in the number of requests for magnetic resonance imaging (MRI) of the knee and a parallel decrease in the use of radiography. Given the importance of ligamentous and cartilaginous components in the structure and function of the knee, it is not surprising that MRI is a frequently ordered imaging technique. Knee MRI is a sensitive and non-invasive imaging test that provides an easy and accurate diagnosis of different anatomical abnormalities. However, the findings may cause confusion when they do not correlate with radiography, results of exploration or the patient’s clinical history [8]. MRI sometimes identifies incidental pathology unrelated to the symptoms, which can lead to unnecessary referral and treatment [9]. A recent review of the available evidence attributes little or no benefit to the use of routine MRI in the diagnosis of knee pathology [10]. Knee MRI is one of the most frequently performed tests in our health department and is exceeded in frequency only by MRI of the lumbar spine. Given the diversity of clinical practice, the progressive increase in the use of MRI in the last decade [11–14] and the low rate of appropriate requests for MRI observed for other musculoskeletal conditions [15], it is possible that this test is becoming a standard before a rigorous evaluation of its use has been published [16]. Furthermore, inappropriate requests for MRI represent a high cost for the health system and can lead to unnecessary interventions [17], which are even more costly and risky for patients. All Spaniards have access to the country’s universal healthcare system called the Spanish National Health System. It covers most healthcares free of charge. Unlike in other health systems, Spanish doctors are not financially incentivised to prescribe expensive prescriptions and treatments, so they are accustomed to recommending the cheapest course of action. However, family doctors can freely prescribe MRI in our region. We found no studies that have investigated the appropriateness of requests for MRI in the Spanish population and we chose to investigate this issue. The objectives of this study were to quantify the percentage of appropriate requests for knee MRI in primary care according to the American College of Radiology (ACR) criteria [18, 19] and to identify the variables associated with inappropriate requests. Methods During 2015, the Elche Health Department (Alicante, Spain) provided coverage to 163 583 people. Any physician from its six primary care centres can order knee MRI, which are performed on a Philips Intera 1.5T device (Koninklijke Philips NV, Amsterdam, the Netherlands) at the Department’s University General Hospital. Given the availability of an electronic medical record (EMR) for every MRI performed, we decided to conduct a retrospective cross-sectional study. Initially, the data for all MRI of the knee performed during 2015 were collected. After discarding those requested by specialists, we found 598 cases of knee MRI that had been requested by primary care physicians. Based on this population size and assuming a rate of 40% inappropriate requests based on a recent study conducted in a similar setting [20], a confidence interval (CI) of ±4% and a confidence level of 95%, a sample size of 300 patients was calculated as necessary. All patients were over 17 years old and there were no pregnant women in the sample. Simple random sampling was performed among all registered MRIs to select the sample. The EMR for each patient was reviewed, and a team member (Medical Doctor, Master in Public Health, General Practitioner for 7 years) collected the independent variables that were considered necessary to assess the appropriateness of MRI according to the ACR criteria: age; duration of knee pain (chronic if ≥3 months, acute if <3 months and indeterminate if unclear in the EMR); existence of patellofemoral symptoms; antecedent of a previous knee injury in the form of a fall, twist injury or traffic accident with suspected dislocation of the knee; ability to walk and support the body weight and knee radiography performed in the 3 months before the MRI request. Other details recorded included whether the patient had sustained a fracture or exhibited joint effusion, degenerative signs, crystals, avascular necrosis, inflammatory signs or any evidence of internal disarray of the joint. In addition, other independent variables not needed to evaluate the MRI request according to the ACR criteria were recorded to check whether they were related to inappropriate request for knee MRI, such as sex, MRI priority (normal or urgent), primary care centre originating the request, number of referrals for every patient and surgical indications. With the exception of age, all variables were categorical. Cases with missing values were discarded. To evaluate the appropriateness of each MRI request, the same team member who collected the data compared the values for the variables for every patient against the ACR criteria for both traumatic [18] and non-traumatic knee [19] injury. The quantitative variable age is expressed as median and interquartile range (IQR, in P25–P75 format). The Shapiro–Wilk test showed that age and sex were non-normally distributed, and these variables were analysed using the Mann–Witney U test. Bivariate analysis was performed, followed by multivariate analysis using backward stepwise regression. Significance was accepted at P < 0.05. The results are expressed as the odds ratio (OR) with respect to a reference group. To evaluate the adjustment of the model developed, Pearson’s chi-square test and the Hosmer–Lemeshow test were used. Collinearity was assessed using the variance inflation factor. Statistical analysis was performed in STATA (Version 14, StataCorp, College Station, TX, USA). Results We included 300 patients: 144 (48%) women and 156 (52%) men. Their median age was 53 years (IQR = 39, 66) for women and 46 years (IQR = 31, 57) for men. The difference in age between men and women was significant (P = 0.002). Table 1 shows the frequencies and percentages of each ACR variant in the sample, grouped according to the appropriateness and inappropriateness of the request for MRI and according to the type of injury—traumatic and non-traumatic. Overall, the MRI prescription was assessed as inappropriate in 135 of cases (45%) and as appropriate in 165 cases (55%). Table 1 Knee MRI appropriateness and inappropriateness causes in our samplea Description ACR variant N (%) Appropriateness causes  Non-traumatic knee pain Adult: non-trauma and non-localised pain. Initial knee radiographs are negative or demonstrate a joint effusion 6 52 (31.5) Child or adult: patellofemoral (anterior) symptoms. Initial knee radiographs are negative or demonstrate a joint effusion 5 40 (24.2) Adult: Initial knee radiographs demonstrate inflammatory, crystalline or degenerative joint disease (uni- to tri-compartmental sclerosis, hypertrophic spurs, joint space narrowing and/or subchondral cysts) 9b 19 (11.5)  Acute trauma to the knee Adult or child >1-year-old. Fall or twisting injury with either no fracture or a Segond fracture seen on a radiograph, suspect internal derangement. Next study 3 32 (19.4) Adult or child >1-year-old. Injury to knee, mechanism unknown. Focal patellar tenderness, effusion, able to walk 5 14 (8.5) Adult or child >1-year-old. Fall or twisting injury, with one or more of the following: focal tenderness, effusion, inability to bear weight. First study 2 4 (2.4) Adult or child >1 year old. Significant trauma to the knee from motor vehicle accident, suspect knee dislocation 6 4 (2.4) Inappropriateness causes  Non-traumatic knee pain Adult: non-trauma and non-localised pain. Initial examination 3 52 (38.5) Child or adult: patellofemoral (anterior) symptoms. Initial examination 2 43 (31.9) Adult: Initial knee radiographs demonstrate inflammatory, crystalline or degenerative joint disease (uni- to tri-compartmental sclerosis, hypertrophic spurs, joint space narrowing and/or subchondral cysts) 9b 18 (13.3) Adult: patellofemoral (anterior) symptoms. Initial knee radiographs demonstrate degenerative joint disease and/or chondrocalcinosis 8 4 (3.0) Othersc 3 (2.2)  Acute trauma to the knee Adult or child >1-year-old. Fall or twisting injury, no focal tenderness, no effusion; able to walk. First study 1 15 (11.1) Description ACR variant N (%) Appropriateness causes  Non-traumatic knee pain Adult: non-trauma and non-localised pain. Initial knee radiographs are negative or demonstrate a joint effusion 6 52 (31.5) Child or adult: patellofemoral (anterior) symptoms. Initial knee radiographs are negative or demonstrate a joint effusion 5 40 (24.2) Adult: Initial knee radiographs demonstrate inflammatory, crystalline or degenerative joint disease (uni- to tri-compartmental sclerosis, hypertrophic spurs, joint space narrowing and/or subchondral cysts) 9b 19 (11.5)  Acute trauma to the knee Adult or child >1-year-old. Fall or twisting injury with either no fracture or a Segond fracture seen on a radiograph, suspect internal derangement. Next study 3 32 (19.4) Adult or child >1-year-old. Injury to knee, mechanism unknown. Focal patellar tenderness, effusion, able to walk 5 14 (8.5) Adult or child >1-year-old. Fall or twisting injury, with one or more of the following: focal tenderness, effusion, inability to bear weight. First study 2 4 (2.4) Adult or child >1 year old. Significant trauma to the knee from motor vehicle accident, suspect knee dislocation 6 4 (2.4) Inappropriateness causes  Non-traumatic knee pain Adult: non-trauma and non-localised pain. Initial examination 3 52 (38.5) Child or adult: patellofemoral (anterior) symptoms. Initial examination 2 43 (31.9) Adult: Initial knee radiographs demonstrate inflammatory, crystalline or degenerative joint disease (uni- to tri-compartmental sclerosis, hypertrophic spurs, joint space narrowing and/or subchondral cysts) 9b 18 (13.3) Adult: patellofemoral (anterior) symptoms. Initial knee radiographs demonstrate degenerative joint disease and/or chondrocalcinosis 8 4 (3.0) Othersc 3 (2.2)  Acute trauma to the knee Adult or child >1-year-old. Fall or twisting injury, no focal tenderness, no effusion; able to walk. First study 1 15 (11.1) aAmerican College of Radiology. ACR appropriateness criteria. bACR variant 9 considers MRI usually not appropriate, except when preoperative assessment is involved. cMRI prescriptions not matching any ACR criteria. Table 1 Knee MRI appropriateness and inappropriateness causes in our samplea Description ACR variant N (%) Appropriateness causes  Non-traumatic knee pain Adult: non-trauma and non-localised pain. Initial knee radiographs are negative or demonstrate a joint effusion 6 52 (31.5) Child or adult: patellofemoral (anterior) symptoms. Initial knee radiographs are negative or demonstrate a joint effusion 5 40 (24.2) Adult: Initial knee radiographs demonstrate inflammatory, crystalline or degenerative joint disease (uni- to tri-compartmental sclerosis, hypertrophic spurs, joint space narrowing and/or subchondral cysts) 9b 19 (11.5)  Acute trauma to the knee Adult or child >1-year-old. Fall or twisting injury with either no fracture or a Segond fracture seen on a radiograph, suspect internal derangement. Next study 3 32 (19.4) Adult or child >1-year-old. Injury to knee, mechanism unknown. Focal patellar tenderness, effusion, able to walk 5 14 (8.5) Adult or child >1-year-old. Fall or twisting injury, with one or more of the following: focal tenderness, effusion, inability to bear weight. First study 2 4 (2.4) Adult or child >1 year old. Significant trauma to the knee from motor vehicle accident, suspect knee dislocation 6 4 (2.4) Inappropriateness causes  Non-traumatic knee pain Adult: non-trauma and non-localised pain. Initial examination 3 52 (38.5) Child or adult: patellofemoral (anterior) symptoms. Initial examination 2 43 (31.9) Adult: Initial knee radiographs demonstrate inflammatory, crystalline or degenerative joint disease (uni- to tri-compartmental sclerosis, hypertrophic spurs, joint space narrowing and/or subchondral cysts) 9b 18 (13.3) Adult: patellofemoral (anterior) symptoms. Initial knee radiographs demonstrate degenerative joint disease and/or chondrocalcinosis 8 4 (3.0) Othersc 3 (2.2)  Acute trauma to the knee Adult or child >1-year-old. Fall or twisting injury, no focal tenderness, no effusion; able to walk. First study 1 15 (11.1) Description ACR variant N (%) Appropriateness causes  Non-traumatic knee pain Adult: non-trauma and non-localised pain. Initial knee radiographs are negative or demonstrate a joint effusion 6 52 (31.5) Child or adult: patellofemoral (anterior) symptoms. Initial knee radiographs are negative or demonstrate a joint effusion 5 40 (24.2) Adult: Initial knee radiographs demonstrate inflammatory, crystalline or degenerative joint disease (uni- to tri-compartmental sclerosis, hypertrophic spurs, joint space narrowing and/or subchondral cysts) 9b 19 (11.5)  Acute trauma to the knee Adult or child >1-year-old. Fall or twisting injury with either no fracture or a Segond fracture seen on a radiograph, suspect internal derangement. Next study 3 32 (19.4) Adult or child >1-year-old. Injury to knee, mechanism unknown. Focal patellar tenderness, effusion, able to walk 5 14 (8.5) Adult or child >1-year-old. Fall or twisting injury, with one or more of the following: focal tenderness, effusion, inability to bear weight. First study 2 4 (2.4) Adult or child >1 year old. Significant trauma to the knee from motor vehicle accident, suspect knee dislocation 6 4 (2.4) Inappropriateness causes  Non-traumatic knee pain Adult: non-trauma and non-localised pain. Initial examination 3 52 (38.5) Child or adult: patellofemoral (anterior) symptoms. Initial examination 2 43 (31.9) Adult: Initial knee radiographs demonstrate inflammatory, crystalline or degenerative joint disease (uni- to tri-compartmental sclerosis, hypertrophic spurs, joint space narrowing and/or subchondral cysts) 9b 18 (13.3) Adult: patellofemoral (anterior) symptoms. Initial knee radiographs demonstrate degenerative joint disease and/or chondrocalcinosis 8 4 (3.0) Othersc 3 (2.2)  Acute trauma to the knee Adult or child >1-year-old. Fall or twisting injury, no focal tenderness, no effusion; able to walk. First study 1 15 (11.1) aAmerican College of Radiology. ACR appropriateness criteria. bACR variant 9 considers MRI usually not appropriate, except when preoperative assessment is involved. cMRI prescriptions not matching any ACR criteria. The main causes of inappropriateness were the use of MRI as the initial imaging technique when it was not indicated. For example, in the initial evaluation of non-traumatic and non-localised knee pain (inappropriate use of MRI for variant 3 of ACR for non-traumatic knee injury, comprising 39% of inappropriate requests), initial study of patellofemoral symptoms (inappropriate use of MRI for variant 2 of non-traumatic ACR, 32% of inappropriate requests) and pain after a fall or twist without spill or focalised pain (use MRI for ACR variant 1 for traumatic knee, 11% of inappropriate requests). Another important reason was the inappropriate use of MRI for ACR variant 9 for non-traumatic knee injury when not considering surgery (13%). As shown in Table 2, the bivariate analysis of the association between inappropriate requests and independent variables found no significant differences in sex, age, clinical signs (meniscal, arthrosis, blockage or spills) or priority of the MRI request. By contrast, there were significant differences in cases for which a knee radiograph was available before the MRI request (OR = 0.08; P < 0.001), when there was evidence of previous trauma (OR = 0.26, P < 0.001), when the duration of symptoms was not accurately determined (OR = 2.27; P = 0.03) and when the MRI request came from a specific primary care centre (OR = 0.28, P = 0.005). Table 2 Knee MRI Inappropriateness: bivariate analysis Inappropriateness Appropriate MRI (N) Inappropriate MRI (N) Odds ratio CI 95% P-value Patient sex  Men 90 66 1 (ref)  Women 75 69 1.25 0.80–1.98 0.33 Patient age (years) <24 16 15 1 (ref)  25–44 47 40 0.91 0.40–2.06 0.82  45–64 69 54 0.83 0.38–1.84 0.65  65–74 25 16 0.68 0.27–1.75 0.43 >75 8 10 1.33 0.42–4.28 0.63 Patellofemoral symptoms  No 122 110 1 (ref)  Yes 43 25 0.64 0.37–1.12 0.12 Degenerative joint disease  No 150 118 1 (ref)  Yes 15 17 1.44 0.69–3.00 0.33 Joint blockages  No 156 128 1 (ref)  Yes 9 7 0.95 0.34–2.62 0.92 Joint efussion  No 147 125 1 (ref)  Yes 18 10 0.65 0.29–1.47 0.3 Duration of symptoms  Acute 99 80 1 (ref)  Chronic 54 35 0.75 0.45–1.28 0.3  Undetermined 12 20 2.27 1.06–4.86 0.035 MRI priority  Normal 151 124 1 (ref)  Urgent 14 11 0.96 0.42–2.18 0.92 Previous radiograph  No 32 102 1 (ref)  Yes 133 33 0.08 0.04–0.13 <0.001 Previous trauma  No 111 120 1 (ref)  Yes 54 15 0.26 0.14–0.48 <0.001 Primary care centre  1 35 41 1 (ref)  2 48 43 0.76 0.42–1.41 0.39  3 17 13 0.65 0.28–1.53 0.33  4 18 16 0.76 0.34–1.71 0.51  5 20 13 0.55 0.24–1.27 0.17  6 27 9 0.28 0.12–0.69 0.005 GP sex  Men 80 58 1 (ref)  Women 87 75 1.19 0.75–1.88 0.46 GP age (years)a <45 48 39 1 (ref)  45–54 28 15 1.52 0.71–3.23 0.28  55–64 76 54 0.89 0.53–1.52 0.68 Inappropriateness Appropriate MRI (N) Inappropriate MRI (N) Odds ratio CI 95% P-value Patient sex  Men 90 66 1 (ref)  Women 75 69 1.25 0.80–1.98 0.33 Patient age (years) <24 16 15 1 (ref)  25–44 47 40 0.91 0.40–2.06 0.82  45–64 69 54 0.83 0.38–1.84 0.65  65–74 25 16 0.68 0.27–1.75 0.43 >75 8 10 1.33 0.42–4.28 0.63 Patellofemoral symptoms  No 122 110 1 (ref)  Yes 43 25 0.64 0.37–1.12 0.12 Degenerative joint disease  No 150 118 1 (ref)  Yes 15 17 1.44 0.69–3.00 0.33 Joint blockages  No 156 128 1 (ref)  Yes 9 7 0.95 0.34–2.62 0.92 Joint efussion  No 147 125 1 (ref)  Yes 18 10 0.65 0.29–1.47 0.3 Duration of symptoms  Acute 99 80 1 (ref)  Chronic 54 35 0.75 0.45–1.28 0.3  Undetermined 12 20 2.27 1.06–4.86 0.035 MRI priority  Normal 151 124 1 (ref)  Urgent 14 11 0.96 0.42–2.18 0.92 Previous radiograph  No 32 102 1 (ref)  Yes 133 33 0.08 0.04–0.13 <0.001 Previous trauma  No 111 120 1 (ref)  Yes 54 15 0.26 0.14–0.48 <0.001 Primary care centre  1 35 41 1 (ref)  2 48 43 0.76 0.42–1.41 0.39  3 17 13 0.65 0.28–1.53 0.33  4 18 16 0.76 0.34–1.71 0.51  5 20 13 0.55 0.24–1.27 0.17  6 27 9 0.28 0.12–0.69 0.005 GP sex  Men 80 58 1 (ref)  Women 87 75 1.19 0.75–1.88 0.46 GP age (years)a <45 48 39 1 (ref)  45–54 28 15 1.52 0.71–3.23 0.28  55–64 76 54 0.89 0.53–1.52 0.68 aGP age was unknown in 23 cases. Table 2 Knee MRI Inappropriateness: bivariate analysis Inappropriateness Appropriate MRI (N) Inappropriate MRI (N) Odds ratio CI 95% P-value Patient sex  Men 90 66 1 (ref)  Women 75 69 1.25 0.80–1.98 0.33 Patient age (years) <24 16 15 1 (ref)  25–44 47 40 0.91 0.40–2.06 0.82  45–64 69 54 0.83 0.38–1.84 0.65  65–74 25 16 0.68 0.27–1.75 0.43 >75 8 10 1.33 0.42–4.28 0.63 Patellofemoral symptoms  No 122 110 1 (ref)  Yes 43 25 0.64 0.37–1.12 0.12 Degenerative joint disease  No 150 118 1 (ref)  Yes 15 17 1.44 0.69–3.00 0.33 Joint blockages  No 156 128 1 (ref)  Yes 9 7 0.95 0.34–2.62 0.92 Joint efussion  No 147 125 1 (ref)  Yes 18 10 0.65 0.29–1.47 0.3 Duration of symptoms  Acute 99 80 1 (ref)  Chronic 54 35 0.75 0.45–1.28 0.3  Undetermined 12 20 2.27 1.06–4.86 0.035 MRI priority  Normal 151 124 1 (ref)  Urgent 14 11 0.96 0.42–2.18 0.92 Previous radiograph  No 32 102 1 (ref)  Yes 133 33 0.08 0.04–0.13 <0.001 Previous trauma  No 111 120 1 (ref)  Yes 54 15 0.26 0.14–0.48 <0.001 Primary care centre  1 35 41 1 (ref)  2 48 43 0.76 0.42–1.41 0.39  3 17 13 0.65 0.28–1.53 0.33  4 18 16 0.76 0.34–1.71 0.51  5 20 13 0.55 0.24–1.27 0.17  6 27 9 0.28 0.12–0.69 0.005 GP sex  Men 80 58 1 (ref)  Women 87 75 1.19 0.75–1.88 0.46 GP age (years)a <45 48 39 1 (ref)  45–54 28 15 1.52 0.71–3.23 0.28  55–64 76 54 0.89 0.53–1.52 0.68 Inappropriateness Appropriate MRI (N) Inappropriate MRI (N) Odds ratio CI 95% P-value Patient sex  Men 90 66 1 (ref)  Women 75 69 1.25 0.80–1.98 0.33 Patient age (years) <24 16 15 1 (ref)  25–44 47 40 0.91 0.40–2.06 0.82  45–64 69 54 0.83 0.38–1.84 0.65  65–74 25 16 0.68 0.27–1.75 0.43 >75 8 10 1.33 0.42–4.28 0.63 Patellofemoral symptoms  No 122 110 1 (ref)  Yes 43 25 0.64 0.37–1.12 0.12 Degenerative joint disease  No 150 118 1 (ref)  Yes 15 17 1.44 0.69–3.00 0.33 Joint blockages  No 156 128 1 (ref)  Yes 9 7 0.95 0.34–2.62 0.92 Joint efussion  No 147 125 1 (ref)  Yes 18 10 0.65 0.29–1.47 0.3 Duration of symptoms  Acute 99 80 1 (ref)  Chronic 54 35 0.75 0.45–1.28 0.3  Undetermined 12 20 2.27 1.06–4.86 0.035 MRI priority  Normal 151 124 1 (ref)  Urgent 14 11 0.96 0.42–2.18 0.92 Previous radiograph  No 32 102 1 (ref)  Yes 133 33 0.08 0.04–0.13 <0.001 Previous trauma  No 111 120 1 (ref)  Yes 54 15 0.26 0.14–0.48 <0.001 Primary care centre  1 35 41 1 (ref)  2 48 43 0.76 0.42–1.41 0.39  3 17 13 0.65 0.28–1.53 0.33  4 18 16 0.76 0.34–1.71 0.51  5 20 13 0.55 0.24–1.27 0.17  6 27 9 0.28 0.12–0.69 0.005 GP sex  Men 80 58 1 (ref)  Women 87 75 1.19 0.75–1.88 0.46 GP age (years)a <45 48 39 1 (ref)  45–54 28 15 1.52 0.71–3.23 0.28  55–64 76 54 0.89 0.53–1.52 0.68 aGP age was unknown in 23 cases. Logistic regression analysis was performed on all collected variables to identify those that were independently associated with the inappropriate request for MRI after adjusting for possible confounding factors. The results are shown in Table 3. One variable, female sex, was associated with a higher percentage of inappropriate requests (OR = 1.96; P = 0.03). Three variables were associated with a lower percentage of inappropriate requests: previous knee radiograph (OR = 0.05; P < 0.001), knee trauma (OR = 0.14; P < 0.001) and urgent MRI request (OR = 0.32; P = 0.03). Table 3 Factors associated to knee MRI inappropriateness N Odds ratio CI 95% P-value Previous radiograph  No 134 1 (ref)  Yes 166 0.05 0.026–0.097 <0.001 Previous trauma  No 231 1 (ref)  Yes 69 0.14 0.06–0.31 <0.001 Sex  Men 156 1 (ref)  Women 144 1.96 1.07–3.57 0.028 MRI priority  Normal 275 1 (ref)  Urgent 25 0.32 0.11–0.91 0.033 N Odds ratio CI 95% P-value Previous radiograph  No 134 1 (ref)  Yes 166 0.05 0.026–0.097 <0.001 Previous trauma  No 231 1 (ref)  Yes 69 0.14 0.06–0.31 <0.001 Sex  Men 156 1 (ref)  Women 144 1.96 1.07–3.57 0.028 MRI priority  Normal 275 1 (ref)  Urgent 25 0.32 0.11–0.91 0.033 Table 3 Factors associated to knee MRI inappropriateness N Odds ratio CI 95% P-value Previous radiograph  No 134 1 (ref)  Yes 166 0.05 0.026–0.097 <0.001 Previous trauma  No 231 1 (ref)  Yes 69 0.14 0.06–0.31 <0.001 Sex  Men 156 1 (ref)  Women 144 1.96 1.07–3.57 0.028 MRI priority  Normal 275 1 (ref)  Urgent 25 0.32 0.11–0.91 0.033 N Odds ratio CI 95% P-value Previous radiograph  No 134 1 (ref)  Yes 166 0.05 0.026–0.097 <0.001 Previous trauma  No 231 1 (ref)  Yes 69 0.14 0.06–0.31 <0.001 Sex  Men 156 1 (ref)  Women 144 1.96 1.07–3.57 0.028 MRI priority  Normal 275 1 (ref)  Urgent 25 0.32 0.11–0.91 0.033 The model based on these four variables could explain 32.3% of the variance (calculated using pseudo R2) and was found to provide a satisfactory fit with both Pearson’s chi-square test (P = 0.63) and the Hosmer–Lemeshow test (P = 0.50 with six quantiles). To evaluate collinearity, the variance inflation factor showed a mean value of 1.34. The 300 patients with knee pain included in the sample were associated with 232 referrals to other services. The most frequent were directed to the Orthopaedic Surgery and Traumatology Service (n = 202, 87%) and Rehabilitation Services (n = 28, 12%). Of all referrals, the MRI request was considered appropriate for 147 (i.e. 0.89 per each inappropriate MRI) and 85 referrals were considered inappropriate. In 72 (36%) of the referrals to the Orthopaedic Surgery and Traumatology Service, a surgical solution to the knee pain was proposed: the MRI request was assessed as appropriate for 51 (71%) and as inappropriate for the other 21 (29%). Finally, we analysed the relationship between the MRI findings and the appropriateness of the request. There was some pathological finding in 102 (76%) of the cases for which MRI was considered inappropriate and in 144 (87%) of those for which MRI was considered appropriate. No pathological findings were observed in 15 (11%) of the group for which MRI was considered appropriate and in 17 (10%) of the group for which MRI was considered inappropriate. Discussion About 45% (41–49%) of the knee MRI requested in this primary care setting did not seem justified from the viewpoint of their fulfilling the ACR criteria. Comparison of our results with those reported by others is difficult because there are few studies of the appropriateness of knee MRI and no reports on this in the Spanish population. A recent study by Solivetti et al. [20] from Italy reported a rate of 40% inappropriate requests for MRI, which is similar to that found in our study. In that study, most of the requests for knee MRI came from orthopaedic surgeons (44.3%) or other specialists (19.5%), and only 36.3% of the knee MRI requests came from primary care physicians. The study by Solivetti et al. is consistent with other studies [21, 22] in attributing the higher percentage of appropriate MRI requests to those requested by orthopaedic surgeons, which may explain why their rates are slightly lower than that found in our study. Vedjani et al. [23] estimated a rate of 46.7% for the inappropriate use of knee MRI in an Iranian hospital, which is also similar to the rate found in our study, even though their criteria were different, their sample was smaller and MRI was requested by a specialist in 98.7% of cases. Oikarinen et al. [24] reported a 10% rate of inappropriate MRI requests in one Finnish hospital, but their sample included only 30 knee MRI cases. Petron et al. [25] found that only 12 of 100 MRIs ordered by primary care providers in Utah (USA) would have been ordered by an orthopaedist given the documented data. In our study, the main reason for considering the MRI request inappropriate was its use as the initial imaging test, which accounted for 82% of the requests considered inappropriate. This result suggests that using MRI as the initial imaging mode increases the chance of the request being inappropriate according to the ACR criteria, which recommend radiography as the most appropriate initial imaging mode in patients with knee pain without a history of trauma. If the radiograph demonstrates deposits of crystals or an inflammatory or degenerative pathology of the joint, no further imaging tests are indicated. MRI is indicated only when a possible surgical intervention is being considered or when the findings cannot explain the patient’s symptoms. For traumatic knee pathology, the ACR also recommends radiology as the initial mode, with the sole exception of significant trauma—such as that received in motor vehicle accidents—for which MRI is considered appropriate as the initial mode for assessing the extent of damage to ligaments and other support structures. Despite this recommendation, 47% of our patients did not have any previous radiograph. This percentage is double that observed in another study [26], and this indicates a clear opportunity for improvement in our environment. Roberts et al. [22] already pointed out the concern among the orthopaedic surgical community about an increasing number of knee MRIs being ordered in the primary care setting without prior radiographs. Although requesting a knee MRI before performing other tests and examinations may be an inappropriate request, the primary care physician may be tempted to do so at an early stage in the evolution of the knee pain to provide the specialist doctor the results at the patient’s first visit, usually several weeks after the referral is made. MRI requests are frequently made by specialists in the private healthcare sector, so the patient may pressure the primary care physician to receive this service at no cost through the public health system. Our observations agree with those of some studies [20, 27, 28], but further studies are needed to assess the role of these and other factors in the prescribing physician’s decision to request an MRI for patients with knee pain. We consider that the high percentage of inappropriate requests for MRI may be a starting point for assessing the costs associated with inappropriate prescription and considering the number of referrals. Other data obtained from the EMR were also associated with a higher percentage of inappropriate requests for knee MRI. However, some of these are not modifiable, such as female sex and the existence of previous trauma. Regarding the female sex, in our sample there were no pregnant women, so we considered that the decision to prescribe MRI was not influenced by the fear of radiation. A lower percentage of inappropriate MRI requests was observed when the request was urgent, but it seems likely that such a priority would be reserved for the most serious cases. Other significant and non-modifiable factors associated with higher percentages of inappropriate MRI requests were knee pain whose duration could not be established from the EMR or history that did not generate any referral and that did not eventually lead to a surgical indication. A limitation of this study is the choice of the ACR criteria for assessing the appropriateness of knee MRI requests. The US health system encourages the use of diagnostic procedures, which may introduce some laxity in other countries in terms of their criteria for determining appropriateness [29] and could increase the inappropriate use of MRI, as observed in this study. Measures to reduce the over-prescription of MRI for knee injury include reassessment of the knee MRI prescription procedure, development of clear guidelines and protocols, improving the training of primary care physicians to encourage their adherence to these guidelines, meetings for case discussion and providing individualised feedback to professionals. Conclusions Our study shows, for the first time in Spain that a high percentage of knee MRI requests by primary care physicians are inappropriate. These findings suggest that having a radiograph of the knee available before the MRI request may reduce the number of inappropriate requests for MRI because the risk is greatly reduced when a previous knee radiograph is available; this happened in only 47% of the sample reported here. However, these results should be interpreted within the context of the specific circumstances of our health system: freedom of MRI prescription, without incentives for prescribers and at no cost for patients. References 1 Noormohammadpour P , Mansournia MA , Koohpayehzadeh J et al. . Prevalence of chronic neck pain, low back pain and knee pain and their related factors in community-dwelling adults in Iran: a population-based national study . Clin J Pain 2017 ; 33 : 181 – 7 . Google Scholar CrossRef Search ADS PubMed 2 Turkiewicz A , Gerhardsson de Verdier M , Engström G et al. . Prevalence of knee pain and knee OA in southern Sweden and the proportion that seeks medical care . Rheumatology (Oxford) 2015 ; 54 : 827 – 35 . Google Scholar CrossRef Search ADS PubMed 3 United States Bone and Joint Initiative: The Burden of Musculoskeletal Diseases in the United States (BMUS). [Internet]. 2014 . Third edition: http://www.boneandjointburden.org/2014-report/ (03 October 2018, date last accessed). 4 Yoshimura N , Akune T , Fujiwara S et al. . Prevalence of knee pain, lumbar pain and its coexistence in Japanese men and women: the longitudinal cohorts of motor system organ (LOCOMO) study . J Bone Miner Metab 2014 ; 32 : 524 – 32 . Google Scholar CrossRef Search ADS PubMed 5 Jhun HJ , Sung NJ , Kim SY . Knee pain and its severity in elderly Koreans: prevalence, risk factors and impact on quality of life . J Korean Med Sci 2013 ; 28 : 1807 – 13 . Google Scholar CrossRef Search ADS PubMed 6 Thiem U , Lamsfuß R , Günther S et al. . Prevalence of self-reported pain, joint complaints and knee or hip complaints in adults aged ≥ 40 years: a cross-sectional survey in Herne, Germany . PLoS One 2013 ; 8 : e60753 . Google Scholar CrossRef Search ADS PubMed 7 Britt H , Miller G , Valenti L et al. . Evaluation of imaging ordering by general practitioners in Australia, 2002–03 to 2011–12 . General Practice Series no.35. Sydney : Sydney University Press , 2014 . 8 ACR–SPR–SSR Practice parameter for the performance and interpretation of magnetic resonance imaging (MRI) of the knee [Internet]. 2015 . Available from: https://www.acr.org/-/media/ACR/Files/Practice-Parameters/mr-knee.pdf (03 October 2018, date last accessed). 9 Pompan DC . Appropriate use of MRI for evaluating common musculoskeletal conditions . Am Fam Physician 2011 ; 83 : 883 – 4 . Google Scholar PubMed 10 Karel YHJM , Verkerk K , Endenburg S et al. . Effect of routine diagnostic imaging for patients with musculoskeletal disorders: A meta-analysis . Eur J Intern Med 2015 ; 26 : 585 – 95 . Google Scholar CrossRef Search ADS PubMed 11 Solomon DH , Katz JN , Carrino JA et al. . Trends in knee magnetic resonance imaging . Med Care 2003 ; 41 : 687 – 92 . Google Scholar PubMed 12 Espeland A , Natvig NL , Løge I et al. . Magnetic resonance imaging of the knee in Norway 2002–2004 (national survey): rapid increase, older patients, large geographic differences . BMC Health Serv Res 2007 ; 7 : 115 . Google Scholar CrossRef Search ADS PubMed 13 Baker LC , Atlas SW , Afendulis CC . Expanded use of imaging technology and the challenge of measuring value . Health Aff (Millwood) 2008 ; 27 : 1467 – 78 . Google Scholar CrossRef Search ADS PubMed 14 Smith-Bindman R , Miglioretti DL , Larson EB . Rising use of diagnostic medical imaging in a large integrated health system . Health Aff (Millwood) 2008 ; 27 : 1491 – 502 . Google Scholar CrossRef Search ADS PubMed 15 Emery DJ , Shojania KG , Forster AJ et al. . Overuse of magnetic resonance imaging . JAMA Intern Med 2013 ; 173 : 823 . Google Scholar CrossRef Search ADS PubMed 16 Brealey S , Russell I , Gilbert F et al. . Value of knee imaging by GPs requires rigorous assessment . BMJ 2002 ; 325 : 1242 . Google Scholar CrossRef Search ADS PubMed 17 Ben-Galim P , Steinberg EL , Amir H et al. . Accuracy of magnetic resonance imaging of the knee and unjustified surgery . Clin Orthop Relat Res 2006 ; 447 : 100 – 4 . Google Scholar CrossRef Search ADS PubMed 18 ACR Appropriateness Criteria®. Acute Trauma to the Knee. [Internet]. American College of Radiology. 2014 . Available from: https://acsearch.acr.org/docs/69419/Narrative/ (03 October 2018, date last accessed). 19 ACR Appropriateness Criteria®. Nontraumatic Knee Pain. [Internet]. American College of Radiology. 2012. Available from: https://acsearch.acr.org/docs/69432/Narrative/ (03 October 2018, date last accessed). 20 Solivetti FM , Guerrisi A , Salducca N et al. . Appropriateness of knee MRI prescriptions: clinical, economic and technical issues . Radiol Med 2016 ; 121 : 315 – 22 . Google Scholar CrossRef Search ADS PubMed 21 Wylie JD , Crim JR , Working ZM et al. . Physician provider type influences utilization and diagnostic utility of magnetic resonance imaging of the knee . J Bone Joint Surg Am 2015 ; 97 : 56 – 62 . Google Scholar CrossRef Search ADS PubMed 22 Roberts TT , Singer N , Hushmendy S et al. . MRI for the evaluation of knee pain: comparison of ordering practices of primary care physicians and orthopaedic surgeons . J Bone Joint Surg Am 2015 ; 97 : 709 – 14 . Google Scholar CrossRef Search ADS PubMed 23 Vejdani M , Molavi Taleghani Y , Kachoei AR et al. . Appropriateness of knee MRI prescriptions in a hospital of Birjand, 2014 . Int J Sci Rep 2015 ; 1 : 299 . Google Scholar CrossRef Search ADS 24 Oikarinen H , Karttunen A , Pääkkö E et al. . Survey of inappropriate use of magnetic resonance imaging . Insights Imaging 2013 ; 4 : 729 – 33 . Google Scholar CrossRef Search ADS PubMed 25 Petron DJ , Greis PE , Aoki SK et al. . Use of knee magnetic resonance imaging by primary care physicians in patients aged 40 years and older . Sports Health 2010 ; 2 : 385 – 90 . Google Scholar CrossRef Search ADS PubMed 26 George E , Tsipas S , Wozniak G et al. . MRI of the knee and shoulder performed before radiography . J Am College Radiol 2014 ; 11 : 1053 – 8 . Google Scholar CrossRef Search ADS 27 Robling M , Kinnersley P , Houston H et al. . An exploration of GPs’ use of MRI: a critical incident study . Fam Pract 1998 ; 15 : 236 – 43 . Google Scholar CrossRef Search ADS PubMed 28 Cohen O , Kahan E , Zalewski S et al. . Medical investigations requested by patients: how do primary care physicians react? Fam Med 1999 ; 31 : 426 – 31 . Google Scholar PubMed 29 Rodríguez Recio FJ , Sanz JC , Peiró S et al. . Utilización inapropiada de la resonancia magnética lumbar en un área de salud . Radiología 1999 ; 41 : 553 – 6 . © The Author(s) 2018. Published by Oxford University Press in association with the International Society for Quality in Health Care. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

International Journal for Quality in Health CareOxford University Press

Published: Aug 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off