Abstract
The medical model of history, examination and investigation forms the bedrock of diagnosis and management of all patients. The essence is the recognition of patterns of symptoms and signs. In the modern era there are an increasing number of non-medical resources ranging from web-based information, computer diagnostic aids and non-specialist healthcare professionals to provide a diagnosis and commence management of a wide range of conditions, including knee problems. We analysed the quality and patterns of clinical presentation in order to answer the question how closely clinical symptoms and examination findings correlate to diagnosis based on MRI scan and/or arthroscopic findings. The analysis was a dataset of a consecutive series of patients, aged 18 to 45, with no past history of knee problems or end stage arthritis, presenting to a single specialist triage physiotherapist, working within an integrated knee service, who fully completed a standardised knee assessment proforma of presenting symptoms and signs at a large district general hospital. The study comprises 86 patients and 98 knees. We analysed this data based on diagnostic findings of MRI scan or arthroscopy to provide definitive intra-articular diagnosis. Based on standard textbook descriptions of common presentations, we went on to define the patients' presentation history and examination as typical or atypical, with typical meaning the symptoms and signs correlated with the diagnosis. The null hypothesis is that patients have a high chance of typical presentations for common knee conditions. In the 75% of patients with a significant intra-articular pathology we found the majority had chondral rather than meniscal tears 1.7 to 1. Forty four percent of patients had atypical symptoms and 71% had atypical clinical signs, 30% and only 26% of the cohort had both typical symptoms and signs together, reflecting a surprisingly low positive predictive probability of symptoms and signs in this group of patients, particularly those with chondral lesions which was 44%. In this cohort, 57% of the cohort has 3 or more multiple diagnoses. In the diagnostically normal group, 43% had symptoms and signs typical for a meniscal tear. We conclude that clinical symptoms and signs surprisingly inaccurate in guiding intra-articular pathology within the knee, even in a sub-set considered the easy and accurate to assess. The number of multiple diagnoses and the incidence of false positive results also means that simplistic interpretations of non-definitive diagnoses and linear causation of pain pathways should be treated cautiously.
