From Pharmacovigilance to Clinical Care Optimization

Abstract

In order to ensure the continued, safe administration of pharmaceuticals, particularly those agents that have been recently introduced into the market, there is a need for improved surveillance after product release. This is particularly so because drugs are used by a variety of patients whose particular characteristics may not have been fully captured in the original market approval studies. Even well-conducted, randomized controlled trials are likely to have excluded a large proportion of individuals because of any number of issues. The digitization of medical care, which yields rich and accessible drug data amenable to analytic techniques, provides an opportunity to capture the required information via observational studies. We propose the development of an open, accessible database containing properly de-identified data, to provide the substrate for the required improvement in pharmacovigilance. A range of stakeholders could use this to identify delayed and low-frequency adverse events. Moreover, its power as a research tool could extend to the detection of complex interactions, potential novel uses, and subtle subpopulation effects. This far-reaching potential is demonstrated by our experience with the open Multi-parameter Intelligent Monitoring in Intensive Care (MIMIC) intensive care unit database. The new database could also inform the development of objective, robust clinical practice guidelines. Careful systematization and deliberate standardization of a fully digitized pharmacovigilance process is likely to save both time and resources for healthcare in general. Limitations of Clinical TrialsIn the quest for the safe and effective use of pharmaceuticals, there is a pressing need for improved surveillance to reduce the risk of preventable morbidity and mortality after products are released. Clearly, such pharmacovigilance represents a form of care optimization as postrelease safety cannot be maximally assured even by excellent premarket studies. This is particularly so because these agents are prescribed to patient populations that are often not completely represented in the original studies required for market approval. Randomized controlled trials (RCTs), the gold standard for defining drug efficacy and safety profiles, are often of inadequate duration, are too narrow in indication and scope, and consist of relatively few test subjects.1 Therefore, RCTs are ill-equipped to fully unravel heterogeneity of treatment

Artificial intelligence, machine learning and health systems

“…These methods have been deployed in the research context in screening and diagnosis and prediction of future events ( Table 1 ). These deployments are in disparate areas, typically in hospital rather than community setting, and in the vast majority of cases based on data from single centers, with implications for reproducibility [ 11 ] and generalizability [ 12 ]. However, the rapid pace of development of machine learning continues both within health care and more broadly across all information processing tasks in society [ 13 ].…”

State of the art review: the data revolution in critical care

“…The essential starting point of the individual patient can be enhanced by the knowledge present in population-level databases, and the resulting information combinations and comparisons used to make informed clinical decisions. In turn, the information accumulated from individuals benefits the healthcare of the entire population.

Figure 4 Clinical care optimization: a Big Data model for efficient targeting of tests and treatments and vigilance for adverse events (figure courtesy of Kai-ou Tang and Edward Moseley, from [21] with permission).

…”

Challenges and Opportunities in Secondary Analyses of Electronic Health Record Data

“…Pharmaceuticals are often prescribed to a large, diverse patient population that may have not been adequately represented in pre-release clinical trials. In fact, RCT cohorts may deliberately be relatively homogeneous in order to capture the intended effect(s) of a medication without "noise" from co-morbidities that could modulate treatment effects [28]. Humphreys and colleagues (2013) reported that in highly-cited clinical trials, 40 % of identified patients with the condition under consideration were not enrolled, mainly due to restrictive eligibility criteria [29].…”

“…Humphreys and colleagues (2013) reported that in highly-cited clinical trials, 40 % of identified patients with the condition under consideration were not enrolled, mainly due to restrictive eligibility criteria [29]. Variation in trial design (comparators, endpoints, duration of follow-up) as well as trial size limit their ability to detect low-frequency or long-term side-effects and adverse events [28]. Post-market surveillance reports are imperfectly collected, are not regularly amalgamated, and may not be publically accessible to support clinical-decision making by physicians or inform decisionmaking by patients.…”