The reform of the Italian insolvency law in 2005 introduced the troubled debt restructuring (TDR) procedure as a means to restore companies that are in financial distress and avoid potential liquidation. The success of this procedure depends strictly on the timeliness of intervention. Therefore, the availability of a prediction tool appears to be crucial. This paper focuses on the ability of accounting ratios to predict the financial distress status of a firm as defined by the law. Based on a linear discriminant analysis, we formulate the probability of a firm filing for TDR in one year, as well as other quantitative techniques that are intended to monitor financial health. Specifically, we begin with a test of Altman's Z, Z' and Z'' scores for bankruptcy on the listed Italian companies that filed for TDR in the period 2005-12. The test results do not completely satisfy the TDR prediction. We then introduce a new score formula based on seven accounting ratios and specific coefficients. Several confirmative analyses are also conducted to validate the predictive accuracy and the generalization power of our score formula.