Freezing of gait (FOG) is a debilitating symptom of Parkinson's disease (PD) that remains difficult to assess. Wearable movement sensors and associated algorithms can be used to quantify FOG in laboratory settings, but the utility of such methods for real world use is unclear. We aimed to determine the suitability of our wearable sensor-based FOG assessment method for real world use by assessing its performance during in-clinic simulated real world activities. Accuracy of the sensor-based method during simulated real-world tasks was calculated using expert rated video as the gold standard. To determine feasibility for unsupervised home use, we also determined correlations between the percent of active time spent freezing (%ATSF) during unsupervised home use and in-clinic activities. Nineteen people with PD and FOG participated in this study. Results from our sensor-based method demonstrated an accuracy above 90% compared to gold-standard expert review during simulated real-world tasks. Additionally, %ATSF from our sensor-based method during unsupervised home use correlated strongly with %ATSF from our sensor-based method during in-clinic simulated real-world activities (ρ = 0.73). Accuracy values and correlation patterns suggest our method may be useful for FOG assessment in the real world.