Use of artificial neural networks to identify the predictive factors of extracorporeal shock wave therapy treating patients with chronic plantar fasciitis.

The purpose of our study is to identify the predictive factors for a minimum clinically successful therapy after extracorporeal shock wave therapy for chronic plantar fasciitis. The demographic and clinical characteristics were evaluated. The artificial neural networks model was used to choose the significant variables and model the effect of achieving the minimum clinically successful therapy at 6-months' follow-up. The multilayer perceptron model was selected. Higher VAS (Visual Analogue Score) when taking first steps in the morning, presence of plantar fascia spur, shorter duration of symptom had statistical significance in increasing the odd. The artificial neural networks model shows that the sensitivity of predictive factors was 84.3%, 87.9% and 61.4% for VAS, spurs and duration of symptom, respectively. The specificity 35.7%, 37.4% and 22.3% for VAS, spurs and duration of symptom, respectively. The positive predictive value was 69%, 72% and 57% for VAS, spurs and duration of symptom, respectively. The negative predictive value was 82%, 84% and 59%, for VAS, spurs and duration of symptom respectively. The area under the curve was 0.738, 0.882 and 0.520 for VAS, spurs and duration of symptom, respectively. The predictive model showed a good fitting of with an overall accuracy of 92.5%. Higher VAS symptomatized by short-duration, severer pain or plantar fascia spur are important prognostic factors for the efficacy of extracorporeal shock wave therapy. The artificial neural networks predictive model is reasonable and accurate model can help the decision-making for the application of extracorporeal shock wave therapy.