Is Google good enough for medicine?

Furthermore, available data are clearly limited to Google users, and hence a non-representative sampling bias may have resulted from various factors such as accessibility, socioeconomic status, age, disability and preferred search engine. [...]it is important to bear in mind other confounding factors that may influence the 'seasonality' of such search volumes, such as the effects of political reform on health and disease, health campaigns, and the influence from pop culture and media. [...]it is not surprising that 'medically trained' users more frequently establish the correct diagnosis using Google compared with their 'less informed' counterparts (50-60% compared with 22%).