Defending Choice in Medicine
What are side effects
All orthodox drugs produce side effects, though their severity can vary enormously from person to person. The side effects occur because a drug does not act only on the particular symptoms for which the person is being treated, but also on other aspects of the body’s function. The existence and magnitude of side effects are a consequence of the individuality of people’s symptoms and responses, and the fact that this individuality is not taken into account in trials. In these trials
Disease definitions are generalised, not individual;
Definitions of effectiveness are generalised, not individual;
Selection criteria for test subjects are not only generalised but have to compromise between homogeneity and generalisability.
The side effects themselves may be caused either by the toxicity of the drug overwhelming cellular reactivity or by the homeostatic reaction of the body to the drug.
The difference between ‘beneficial’ effects and side effects is entirely arbitrary, and depends on which effects are considered desirable by researchers or practitioners. Thus,
Viagra, one of the most successful drug discoveries in recent years, was originally developed to treat angina, but a pilot study showed that it did little to alleviate this condition. However, when researchers decided to stop the trial early and recall any unused pills, they were perplexed by the reluctance of the trial volunteers to return them. Subsequent interviews revealed that Viagra had an unexpected and desirable side-effect. Further trials and safety tests have resulted in Viagra’s current widespread availability for the treatment of impotence. 
It should be noted that homeopathic remedies do not produce side effects, because:
The remedies are selected for their precise correspondence with the symptoms;
The remedies generally do not act chemically, and can only act in those areas already sensitised by the nature of the illness.
Simon Singh and Edzard Ernst, Trick or Treatment? Alternative medicine on trial (London: Bantam Press, 2008), p. 225.
Why it works