File
Chan M1, Hassan I2,3, Pan B2,3, Hui C4, Defreitas T5, Otto D4, Whittaker J6
1University of Alberta, Glen Sather Sports Medicine Clinic, Edmonton, Canada, 2University of Alberta, EPICORE Centre, Edmonton, Canada, 3University of Alberta, Alberta SPOR Support Unit, Consultation & Research Services, Edmonton, Canada, 4University of Alberta, Department of Surgery, Edmonton, Canada, 5University of Alberta, Family Medicine, Edmonton, Canada, 6University of Alberta, Department of Physical Therapy, Edmonton, Canada
Background: Anterior Cruciate Ligament (ACL) tears are common and are associated with reduced function, physical inactivity, and increased risk of future osteoarthritis. An early accurate diagnosis is vital for initiating timely and appropriate treatment. Only 15% of ACL tears are diagnosed at the initial medical appointment, with many patients waiting months for a correct diagnosis. Misdiagnoses contribute to delayed or misdirected rehabilitation, physician, and specialist visits, surgical and diagnostic-imaging waiting lists, and poor patient outcomes. It is essential that primary care practitioners (physiotherapists, family medicine physicians,) accurately identify ACL tears early after injury to facilitate timely and appropriate treatment. To date, there is no consensus about which combination of patient-reported and/or clinical examination variables are the most valuable for diagnosing an ACL tear.
Purpose: To identify which combination of patient-reported and/or clinical examination variables are the most valuable for diagnosing an ACL tear to inform the development of a future clinical decision tool for use in primary care settings.
Methods: The electronic records of patients aged 15-45 years, with ICD-9 codes corresponding to intra-articular knee injuries and a confirmed (ACL+) or denied (ACL-) full-thickness ACL tear through one-of-three criterion standards (orthopaedic surgeon assessment, MRI and/or surgery) seen at a University-based clinic between January 2014 and July 2016 were identified. ACL status (ACL+ or ACL-) and demographic, patient-reported and clinical examination variables were manually extracted. Demographic and potential diagnostic variables were compared between study groups with univariate analyses (chi-squared and t-test, α=0.001). Univariate results and clinician surveys (n=17) were used to select items for inclusion into multivariable logistic regression models that assessed the odds (OR: 95%CI) of an ACL tear based on patient-reported variables alone, or patient-reported and clinician-generated variables. Models were estimated on a training set (random 70% of data) and evaluated with a test set (30% of data). Model performance measures included accuracy rate, AUC, sensitivity and specificity.
Results: Of 1,512 potentially relevant EMRs 725 were included. Participant median age was 26 years (range 15-45), 48% were female and 60% had an ACL tear. Based on univariable comparisons no single diagnostic criteria for ACL tear emerged. The prioritized items for inclusion into the regression models were age, sport-related injury, immediate swelling, family history of ACL tear and Lachman test result. A combination of patient-reported (age, sport-related injury, immediate swelling, family history of ACL tear) and clinician-generated variables (Lachman test result) was superior in identifying an ACL tear (accuracy rate; 95% (95%CI 90,98), AUC; 0.97, sensitivity; 97%, and specificity; 95%) compared to a combination of patient-reported (age, sport-related injury, immediate swelling, family history of ACL tear) variables alone (accuracy rate; 84% (95%CI 77,89), AUC; 0.86, sensitivity; 60%, and specificity; 95%).
Conclusion(s): A combination of patient-reported and clinician-generated variables best support the differential diagnosis of ACL tears. Future research should be aimed at developing a clinical support tool for ACL tear diagnosis for use in primary care settings.
Implications: Clinical support tools that ensure early and accurate diagnosis are essential for mitigating the long-term consequences of ACL tears.
Keywords: Diagnosis, Primary Care, Knee
Funding acknowledgements: Department of Surgery and Division of Orthopaedic Surgery and the Glen Sather Sports Medicine Clinic, University of Alberta.
Purpose: To identify which combination of patient-reported and/or clinical examination variables are the most valuable for diagnosing an ACL tear to inform the development of a future clinical decision tool for use in primary care settings.
Methods: The electronic records of patients aged 15-45 years, with ICD-9 codes corresponding to intra-articular knee injuries and a confirmed (ACL+) or denied (ACL-) full-thickness ACL tear through one-of-three criterion standards (orthopaedic surgeon assessment, MRI and/or surgery) seen at a University-based clinic between January 2014 and July 2016 were identified. ACL status (ACL+ or ACL-) and demographic, patient-reported and clinical examination variables were manually extracted. Demographic and potential diagnostic variables were compared between study groups with univariate analyses (chi-squared and t-test, α=0.001). Univariate results and clinician surveys (n=17) were used to select items for inclusion into multivariable logistic regression models that assessed the odds (OR: 95%CI) of an ACL tear based on patient-reported variables alone, or patient-reported and clinician-generated variables. Models were estimated on a training set (random 70% of data) and evaluated with a test set (30% of data). Model performance measures included accuracy rate, AUC, sensitivity and specificity.
Results: Of 1,512 potentially relevant EMRs 725 were included. Participant median age was 26 years (range 15-45), 48% were female and 60% had an ACL tear. Based on univariable comparisons no single diagnostic criteria for ACL tear emerged. The prioritized items for inclusion into the regression models were age, sport-related injury, immediate swelling, family history of ACL tear and Lachman test result. A combination of patient-reported (age, sport-related injury, immediate swelling, family history of ACL tear) and clinician-generated variables (Lachman test result) was superior in identifying an ACL tear (accuracy rate; 95% (95%CI 90,98), AUC; 0.97, sensitivity; 97%, and specificity; 95%) compared to a combination of patient-reported (age, sport-related injury, immediate swelling, family history of ACL tear) variables alone (accuracy rate; 84% (95%CI 77,89), AUC; 0.86, sensitivity; 60%, and specificity; 95%).
Conclusion(s): A combination of patient-reported and clinician-generated variables best support the differential diagnosis of ACL tears. Future research should be aimed at developing a clinical support tool for ACL tear diagnosis for use in primary care settings.
Implications: Clinical support tools that ensure early and accurate diagnosis are essential for mitigating the long-term consequences of ACL tears.
Keywords: Diagnosis, Primary Care, Knee
Funding acknowledgements: Department of Surgery and Division of Orthopaedic Surgery and the Glen Sather Sports Medicine Clinic, University of Alberta.
Topic: Musculoskeletal: lower limb; Orthopaedics
Ethics approval required: Yes
Institution: University of Alberta
Ethics committee: Health Research Ethics Board Health Panel
Ethics number: Pro00065514
All authors, affiliations and abstracts have been published as submitted.
