Study on the related factors of post‐herpetic neuralgia in hospitalized patients with herpes zoster in Sichuan Hospital of Traditional Chinese Medicine based on big data analysis

Although various factors were reported to be related to post-herpetic neuralgia (PHN), studies based on adequate and comprehensive data were absent. Data was extracted from cases of hospitalized patients with herpes zoster in dermatology department, Sichuan hospital of traditional Chinese medicine range from December, 2011 to February, 2018, and then cleaned to build prediction model with TREENET algorithms. Following evaluated the prediction model by ROC and confusion matrix, variables importance ranking and variables dependency analysis were performed, resulting in the importance ranking of factors for PHN and the dependency between factors and PHN. Based on strict inclusion and exclusion criteria, 1303 (571 PHN and 732 normal controls) cases and 2958 indicators were selected. Model evaluation showed high ROC value (training sample = 0.985, test samples = 0.752) and high accuracy value (70.27%), which indicated that the model was predictive. After variables importance ranking and variables dependency analysis, 62 variables in the model were associated with the occurrence of PHN. Our study identified 62 variables related to PHN and revealed that various variables were the important risk factors for PHN, including age, MCHC, sodium and UA.