OBJECTIVE The Instability Severity Index (ISI) Score was developed to preoperatively assess the risk of recurrent shoulder instability after an arthroscopic Bankart repair. This study aims to validate the use of ISI Score for predicting the risk of recurrence after an arthroscopic Bankart repair in a heterogeneous population and proposes an appropriate cut-off point for treating patients with an arthroscopic Bankart repair or otherwise. METHODS This study analysed 99 shoulders after a traumatic dislocation that underwent arthroscopic Bankart repair with at least 3 years follow-up. Patients were divided into subcategories based on their respective ISI Score. Recurrence includes either a postoperative dislocation or perceived instability. RESULTS The overall recurrence rate was found to be 26.3%. A significant correlation was identified between ISI Score and the recurrence rate (odds ratio [OR]: 1.545, 95% confidence interval [CI]: 1.231-1.939, p < 0.001). Furthermore, ISI Score 4-6 (OR: 4.498, 95% CI: 1.866-10.842, p < 0.001) and ISI Score > 6 (OR: 7.076, 95% CI: 2.393-20.924, p < 0.001) both had a significantly higher risk of recurrence compared to ISI Score 0-3. In ISI Score subcategories 0-3, 4-6 and >6, the recurrence rate was, respectively, 15.4%, 40.7% and 71.4%. CONCLUSIONS ISI Score has predictive value in determining the recurrence risk of shoulder instability following an arthroscopic Bankart repair in a heterogeneous population. Based on the findings of this study, we recommend using arthroscopic Bankart repair in patients with ISI Score 0-3. Clinical and shared decision-making are essential in the group with ISI Score 4-6, since the recurrence rate is significantly higher than in patients with ISI Score 0-3. Arthroscopic Bankart repair is not suitable for patients with ISI Score > 6. METHODS Level III.