Search at PubMed for peer-reviewed journal articles containing your search terms. Many abstracts are available.

However, be aware that there have been serious questions about the independence of much pharmaceutical research, and of phsychopharmaceutical research in particular. The story of British psychiatrist David Healy provides a good introduction to the difficulties facing researchers whose allegience is to science rather than to the pharmaceutical industry which funds much of academic research. David Healy accepted a senior position at the Centre for Addiction and Mental Health (CAMH) and the Department of Psychiatry at University of Toronto. Months after his acceptance of the offer, while he was making final preparations to move from England to Canada, the offer was withdrawn, apparently because a lecture he gave was offensive to the pharmaceutical industry.

Just because most of the articles or the most recent articles that you find through a PubMed search claim to support a particular hypothesis does not mean that the few articles which challenge this hypothesis should be dismissed as unreliable. The rigor of the experimental design, the independence of the research from corporate funding, and the integrity of the journal in which the research was published must also be taken into account when trying to understand what the research indicates.

However, PubMed is useful in many situations, especially if you are investigating a specific question, and use the sorts of search terms that people familiar with the area would be likely to choose. For example, if you are concerned about whether zyprexa is associated with conditions which can cause death, you must be careful in choosing your terms. If you choose

olanzapine mortality

as your search terms, you will miss that olanzapine (the generic name for Zyprexa) is associated with leukopenia and agranulocytosis, which can be life-threatening. In order to obtain this information, you will have to use leukopenia and agranulocytosis each as search terms in place of mortality, which is too general.

Frequently, claims are made that newer drugs do not have risks which are associated with older drugs. Sometimes doing a PubMed search can give you a "heads up" that these claims are premature, as reports trickle in of cases in which the new drug has been associated with these adverse effects, despite the earlier claims to the contrary.