Managing the Risks and Potential of High-tech Innovations-in-use: Predictive Analytic Modeling with Big Data and a Longitudinal Field Study

Abstract

Healthcare, like many other industry sectors, is increasingly becoming high tech innovation enabled. Success and growth of many firms in high tech industry segments such as medical device, automotive, electronics, telecommunication, and aerospace are dependent on rapid pace of technology innovation. High tech innovation provides functionalities and benefits that are not feasible otherwise. As an illustration, in the healthcare delivery segment, rapid innovation in three dimensional imaging has made the disease detection process much more accurate, fast and consistent as compared to the past. However, notwithstanding the potential benefits of high tech innovations, high tech innovations entail risks of failure while in use in the market. Failures of high tech innovations-in-use can cause severe harm to the users. As an illustration, a recent incident of failure of a cardio-vascular defibrillator caused severe injuries including several fatalities to many patients. Hence, firms and regulators need to manage the downside risks of failures of high tech innovations-in-use in a timely manner. Realizing the potential benefits of high tech innovation in many usage areas depend on how well firms and regulators can manage the potential downside risks of high tech innovations-in-use. Also, realizing the potential benefits of a high tech innovation-in-use depend on how well users can learn to use the high tech innovations. In my dissertation, I investigate how firms can best manage the downside risks of high tech innovations-in-use as well as how users can realize the potential benefits of high tech innovations. The dissertation consists of three inter-related studies. The first two studies are aimed towards managing the downside risks of high tech innovations. The last study looks at a specific high tech innovation in healthcare delivery, namely, surgical robots and detail out a field study to understand the factors that lead to the realization of the benefits of high tech innovations in health care. The first step in managing risks of failures of high tech innovations-in-use is to be able to detect signals of failure from the market. In the first study, we show that it is possible to use user feedback of adverse events related to medical devices to detect signals of device failures originating from either design failures, or supply chain failures or manufacturing process failures. Using text mining and machine learning based predictive analysis methods on a 'big' unstructured data-set of adverse events reported by users of medical devices, we show that firms can detect failures of medical devices with precision and consistency. We also identify that firms exhibit substantial judgment bias in interpreting and reacting to market signals of failures. Either they under-react or they over-react to market signals of failure under certain conditions. We use the theoretical lenses of signal detection and system neglect to setup the study and identify sources of judgment bias. In the second study we extend the first study by identifying product related, firm related and industry related conditions under which firms are more likely to systematically under-react or over-react to market signals of failures of high tech innovations-in-use. An acknowledgement of these sources would help firms and regulators to bring in greater consistency in their detection and decision process. We integrate the theoretical perspectives of signal detection, system neglect and attention based view of firms to propose a framework of judgment bias in the context of detection of failures of high tech innovations-in-use from user reports of adverse events. In the third study, we undertake a field research in a large multi-specialty hospital in the United States to investigate factors that lead to development of technology capability in healthcare delivery in the context of usage of a surgical robot, namely, da Vinci robot. We identify conditions related to surgeon and team learning that lead to improved usage of the robot. More importantly, we show that with surgeon and team learning, technology mediation can help reduce surgical outcome variation in spite of input heterogeneity in the form of surgeons' experience and skill heterogeneity, patient heterogeneity and team heterogeneity.