Objective To evaluate joint counts in patients with oligoarticular psoriatic arthritis (PsA) for their correlation with treatment decisions and to determine the proportion of patients in whom active disease was missed using reduced joint counts. Methods The international Group for Research and Assessment of Psoriasis and Psoriatic Arthritis Composite Exercise cohort was used for this study. Oligoarthritis was defined as <5 tender and/or swollen joints. At baseline, a tender joint count (TJC) using 68 joints and a swollen joint count (SJC) using 66 joints (66/68-joint counts) were assessed. Reduced joint counts designed for use in rheumatoid arthritis (RA), including 28- and 44-joint counts, were analyzed. In addition, the following proposed joint counts for PsA were tested: the PsA-44 (which includes elbows, wrists, metacarpophalangeal joints, finger proximal interphalangeal [PIP] joints, distal interphalangeal joints, knees, and metatarsophalangeal joints) and the PsA-56 (which includes the same joints as those assessed for the PsA-44 plus ankles and toe PIP joints). Receiver operating characteristic (ROC) curve analysis was used to assess whether joint counts predict treatment changes. The proportion of patients in whom active disease was missed using reduced joint counts designed for RA was also assessed. Results Among 503 patients recruited to the study, 266 (53%) had oligoarthritis. ROC curve analysis revealed that no TJC or active joint count (AJC), even a 66/68-joint count, predicted treatment change (for the TJC, area under the curve [AUC] 0.57, P = 0.125; for the AJC, AUC 0.56, P = 0.159). Use of the SJC in 66 joints did predict treatment change (AUC 0.62, P = 0.006), as did the SJC using the PsA-44 and the PsA-56 (P < 0.03). Neither of the reduced joint counts designed for RA predicted treatment change. A 28-joint count designed for RA missed 21% of patients with tender joints (n = 29) and 27% of patients with swollen joints (n = 23). The PsA-44 and PsA-56 joint counts missed tender joints in 25 patients and 13 patients, respectively, and missed swollen joints in 11 patients and 2 patients, respectively. Conclusion Patients with oligoarticular PsA cannot be accurately assessed for active disease using reduced joint counts designed for RA. Full 66/68-joint counts should be performed to assess patients with PsA.