Background The acute general surgical unit (AGSU) model has become a standard of efficient acute surgical care. Whilst the AGSU has been compared to the traditional surgical model, there is a lack of research auditing referrals and admissions. This study evaluates abdominal pain referrals to AGSU and the necessity of admission. Methods A retrospective cohort study of adult abdominal pain admissions was conducted over a two-year period at a single centre in metropolitan Victoria, Australia. The data were extracted from electronic medical records and key endpoints of data included the diagnosis, length of stay, investigations and subjective pain outcomes. Results A total of 1587 patients met the study criteria of which 1116 (70.3%) had a non-surgical diagnosis with the majority having non-specific abdominal pain. The non-surgical patients had a lower median length of stay (25.3 h) compared to surgical patients (44.2 h, p < 0.001). They were less likely to have an abnormal haemoglobin (p = 0.004), elevated white cell count (p = 0.02) or elevated C-reactive protein > 50 mg/L (p < 0.001). On multivariable analysis, surgical patients had higher odds of having a CRP > 50 mg/L (p = 0.024) and a positive imaging result (p < 0.001). The patient’s pain control also correlated with length of stay. Conclusion A large population of patients with non-specific abdominal pain are admitted to AGSU. These patients do not require surgery and have a short length of stay. Incorporating a negative CRP result and negative imaging result may be utilised in conjunction with optimised analgesia to help avoid these unnecessary admissions, thereby improving AGSU efficiency and workload.