Godes provides a taxonomy of computer-related injuries in medical settings and suggests liability principles for each category.

Directly caused computer injuries occur when, for example, a program controlling X-ray administration has a bug and dispenses too much radiation. In this case, Godes suggests that manufacturers and distributors be held strictly liable, but that providers (e.g., hospitals) and practitioners (e.g., physicians) not be held liable at all.

Indirectly caused computer injuries occur when, for example, a computer-controlled cardiac monitor malfunctions and a practitioner relies on its incorrect readings. In this case, the author suggests strict liability for manufacturers and distributors and professional negligence (malpractice) liability for providers and practitioners.

Negligence injuries can occur if the provider is “negligent in the selection, installation or maintenance of the program” or “the physician misreads the information, misapplies the program or fails to apply it when it should have been applied.” In each of these cases, Godes suggests that the provider and practitioner be held liable under the professional negligence principle.

Finally, no-fault injuries occur when “the computer functions exactly as expected, and the hospital and physicians act in a reasonable fashion,” so that “the injury cannot reasonably be prevented.” An example of this is an automated diagnostic test’s false negative (a treadmill test is Godes’s example) caused by the imperfection of the test, not by either a defect in the product or negligence in its use. In this case, Godes suggests that no liability should result.

The author justifies his suggestions by considering the aims of both health care provision and tort law and successfully shows that his proposal more nearly meets those aims than does existing law. The paper also extensively discusses whether computer-controlled equipment in a hospital is sold to a patient as a product, in which case the strict liability standard would normally apply, or is merely incidental to the practitioner’s services, in which case the professional negligence standard would normally govern. This interesting paper is well-documented and convincing.