RESULTS: The repeatability of the Chinese version of HARQ was validated at a 2-week interval with the intraclass correlation coefficients of 0.926. The patients with chronic cough had higher HARQ score than healthy volunteers (20.7±10.4 vs 4.6±2.1, t=3.904, P=0.000). Among all the patients with variable causes of chronic cough, the patients with gastroesophageal reflux-induced cough presented with the highest HARQ score. When used for the prediction of cough due to gastroesophageal reflux, the optimal cut-off point of HARQ was defined at 23.5 according to the highest Youden index of 0.577, with a sensitivity of 85.7%, specificity of 72.0%, positive predictive value of 58.3% and negative predictive value of 92.6%. A sub-analysis only for cough dut to acid reflux showed the predictive accuracy of HARQ was further improved, corresponding to sensitivity of 87.5%, specificity of 83.9%, positive predictive value of 57.1% and negative predictive value of 96.9%. A meaningful cut-off point of HARQ could not be determined for the prediction of cough due to non-acid reflux.