Abstract This study measures the differences in access to healthcare for female patients in France in three medical specialties (dentistry, gynecology and psychiatry) according to two criteria: the African ethnicity of the patient and the benefit of having means‐tested health insurance coverage. To this purpose, we conducted a nationally representative field experiment on more than 1500 physicians. We do not find substantial discrimination against the patient of African origin. However, the results indicate that patients with means‐tested health insurance coverage are less likely to get an appointment. Differentiating between two types of coverage, we show that the lesser‐known coverage (ACS) is more penalized than the other (CMU‐C) as poor knowledge of the program increases the physician's expectation of additional administrative tasks and is an important element to explain cream‐skimming. We also find that, for physicians who are free to set their fees, the opportunity cost of accepting a means‐tested patient increases the penalty. Finally, the results suggest that enrollment in OPTAM, the controlled pricing practice option that incentivizes physicians to accept means‐tested patients, reduces cream‐skimming.