Shoulder disorders are common and associated with high societal costs, especially for a small group of patients. Prognostic factors can help identify high-cost patients, which is crucial to optimize early identification and develop tailored interventions. We aimed to identify prognostic factors for high societal costs, to examine whether the prognostic factors were similar for high healthcare costs and high costs of sick leave, and to investigate the model's robustness across 4 diagnostic categories. Using national Danish registers, potential prognostic factors (age, sex, educational level, long-term sick leave, admission, visits to general practitioner and physiotherapist, comorbidity, diabetes, low back pain, and neck pain) were included in a logistic regression model with high societal costs, defined by the top 10th percentile, as the main outcome. The model's prognostic accuracy was assessed using the Nagelkerke R2 and its discriminative ability using area under the receiver operating curve (AUC). Data on 80% of the patients (n = 449,302) were used to develop the model and 20% (n = 112,363) to validate the model. By far the strongest prognostic factor for high societal costs and high costs of sick leave was sick leave at the time of diagnosis (OR: 20.2, 95% CI: 19.5-20.9). Prognostic factors for high healthcare costs were high age, comorbidity, and hospital admission the year before diagnosis. The model was robust across diagnostic categories and sensitivity analyses. In the validation sample, the primary model's discriminative ability was good (AUC = 0.80) and the model explained 28% of the variation in the outcome (Nagelkerke R2).