This study investigates gender fairness in personalized pain care recommendations using machine learning algorithms. Leveraging a contextual bandits framework, personalized recommendations are formulated and evaluated using LinUCB algorithm on a dataset comprising interactions with $164$ patients across $10$ sessions each. Results indicate that while adjustments to algorithm parameters influence the quality of pain care recommendations, this impact remains consistent across genders. However, when certain patient information, such as self-reported pain measurements, is absent, the quality of pain care recommendations for women is notably inferior to that for men.