Protecting human subjects from harm in medical research: A proposal for improving risk judgments by institutional review boards

摘要

Our concern for the way biomedical research is conducted has been evidenced in codes of ethics and in federal and international guidelines which emphasize the protection of the rights and welfare of research subjects. While there is agreement that research subjects ought to be protected from harm, Institutional Review Boards (IRBs) may be unable to assess meaningfully the risks of harm in relation to the hoped for benefits of research required in regulations and guidelines because they lack the technical and philosophical expertise to do so. In particular, IRBs are unable to make judgments about the risks of harm to subjects. Analysis of fundamental terms such as harm and risk led to the following conclusions. First, there can be several meanings of harm depending on whether a medical, philosophical or jurisprudential theory is adopted. Second, while risk is understood to express the probability and magnitude of a future harm, there are several objective and subjective considerations that affect the identification, estimation and evaluation of risk. Two considerations were found to be especially relevant: the availability of empirical data and the conduct of risk assessment by experts and non-experts. Emphasis should be placed on quantifying the risks of harm where possible and seeking methods for involving IRBs and subjects in risk judgments. Analysis of taxonomies by Jay Katz and Robert Levine showed them to be useful guides for identifying types of harm; however, they offered little insight as to how risk judgments might be improved. Decision analysis was found to be too restrictive for the protocol review environment although its quantitative, explicit and prescriptive features suggest that it is worthy of continued study. A four-cell matrix is developed that accounts for the contributions of IRBs and subjects in the presence and absence of data. Two cases are offered to illustrate the utility of the matrix for improving the risk judgments of IRBs.