# Feasibility study of the Cambridge knee injury tool (CamKIT): 18-month analysis of acute knee injuries at an urgent treatment centre

## TL;DR

This study evaluated the Cambridge Knee Injury Tool (CamKIT) for assessing acute knee injuries and found it effective in identifying injured patients, though with room for improvement.

## Contribution

The study prospectively validated CamKIT's performance in an urgent treatment center setting for acute knee injuries.

## Key findings

- CamKIT had 100% sensitivity and 100% negative predictive value for identifying knee injuries.
- Scores were significantly higher in injured and surgically treated patients compared to non-injured and conservatively treated ones.
- The tool shows potential for early triage but requires further refinement and multicenter validation.

## Abstract

Soft tissue knee injuries are associated with many short- and long-term consequences. Currently, obstacles in the diagnostic pathway lead to inefficiencies in the management process, resulting in suboptimal patient outcomes. Early stratification of knee injury severity is essential for guiding timely investigations and management decisions. This study aimed to prospectively evaluate how the Cambridge Knee Injury Tool (CamKIT) performs in stratifying knee injuries at initial presentation.

This prospective observational study involved eighty-five participants presenting with acute knee injuries in a Major Regional Urgent Treatment Centre (UTC) over 18 months. Patients recorded information regarding patient factors, external factors, injury mechanisms and signs and symptoms at index presentation. CamKIT scores were then calculated. Patients then continued the management stream determined by the treating healthcare team. Management decisions were not influenced by the results of the CamKIT score. Patient outcomes were analysed by extracting time-to-event outcomes, injury classification, and management information. Statistical analysis involved sensitivity and specificity calculations, descriptive statistics, and Mann-Whitney Test calculations between injured and non-injured cohorts, and surgical and conservative cohorts.

The tool achieved a sensitivity of 100 % and specificity of 31.7 %, with a negative predictive value (NPV) of 100 % and a positive predictive value (PPV) of 34.9 %. CamKIT scores were higher in patients with injuries (median 8; IQR: 7–9) than in non-injured patients (median 6; IQR: 4–7) (p < 0.0001). Similarly, patients who proceeded to surgical intervention (n = 9) had higher scores (median 8; IQR: 7.5–10) compared to those managed conservatively (median 6; IQR: 4.25–7) (p = 0.0001).

The CamKIT continues to show promise in guiding early triage and management of acute knee injuries. Future research involves a multicenter study as well as the integration of machine learning to develop a more robust prediction tool and create a risk-stratified management protocol.

Image 1