The emerging field of biomedical microelectromechanical systems (bioMEMS) has the potential to revolutionize how drugs are discovered, how diseases are diagnosed, and how treatments are administered. Nevertheless, that potential is unlikely to translate into commercial success unless bioMEMS designers can prove that such novel functionality can be delivered each and every time without failure. The life-and-death ramifications of a system failure make absolute reliability a moral and regulatory requirement for many medical applications. This paper reveals the reliability hurdles that must be overcome when systems are used for pharmaceutical research, clinical diagnostics, or implantation into humans. Special attention is given to challenges that are likely to be exacerbated for miniaturized devices. Techniques useful for characterizing, minimizing, and monitoring for failures are also described. A strategy for addressing reliability problems is also presented. By refining bioMEMS for the research laboratory first and then applying the experience gained from that environment to clinical diagnostics applications, bioMEMS does not need to confront head-on the regulatory hurdles and established technologies that could otherwise impede its entry. Similarly, development of more robust, miniaturized clinical diagnostics systems can simplify the entry of implanted bioMEMS.