To Err is Human: 15 Years Later

6 min
November 12, 2014

To mark the anniversary of the Institute of Medicine’s watershed report To Err Is Human: Building a Safer Health System, West Health is running a series of interviews with IOM committee members who helped produce the report, as well as other national health experts to examine what progress has been made in reducing medical errors in the US.

The first Q&A in this eight-part series is with one of the report’s co-authors, Molly Joel Coye, MD, MPH, Chief Innovation Officer of UCLA Health at the University of California, Los Angeles.

Tell us what you think in the comments, or send us your stories about medical errors and interoperability at

Joe Smith: The stated goal of the IOM report To Err is Human: Building a Safer Health System was to break the cycle of inaction surrounding medical errors. The report estimated the number of deaths in hospitals due to preventable errors to be 98,000. A more recent report in the Journal of Patient Safety suggests that number may be between 210,000 and 440,000. Do we actually understand the size and scope of the problem? Are we making progress fast enough, and if not, what more should be done?
Molly Coye: It may be daunting to find that the task of improving quality and safety is so much greater than our initial estimates. But the members of the IOM Committee on Quality of Healthcare in America knew the limitations of our sources, and most importantly, we knew that better data would reveal not only underestimates in the rates we reported for inpatient errors, but other types of medical errors not yet quantified. We could not give probable rates for errors in ambulatory settings, or for skilled nursing facilities, or for diagnostic errors, in addition to treatment errors. Today all of these are measured, and a whole field has emerged to design and test interventions.

The result is not yet good enough. Though many organizations are working toward a culture of safety, and have built quality and safety systems, we are still far short of six sigma care. Information systems and electronic medical records were created to document care, but are only beginning to easily produce the reports needed to track and improve care. Driving better performance will require rapid data feedback loops, far more predictive modeling and clinical decision support tools, direct participation by patients in their care plans and health records, and IT ecosystems that test new apps and other tools, integrate them into EHRs and deploy them rapidly across organizations. When clinicians and patients have the right data and support tools at hand, their own intrinsic motivation is a powerful force.

JS: The report discussed the opportunity for technology and automation to prevent errors, but also spoke to the complexity that occurs when operators are asked to manage a variety of opaque and siloed technological elements, and/or do not have the right information at the right time. In the 15 years since the report, where have we seen the greatest progress with respect to the use and integration of technology to reduce errors? Where do we still have the greatest opportunity?

MC: The HiTech Act and Meaningful Use have built important early capabilities for data exchange in primary care practices, yet the vast majority of Americans still receive their care from multiple fragmented sources. Few emergency rooms, for example, routinely receive information about previous care provided elsewhere for new patients. The greatest progress has been made within integrated delivery systems that maintain a single electronic health record (EHR), or in clinically integrated networks that work over time to interface all the disparate flows of data from independent physician practices, home care agencies, networked hospitals, imaging centers and free-standing surgical centers and urgent care centers. In some cases this is supported by health information exchange (HIE) vendors, or health plans that have acquired vendors. Because almost all institutional providers are locked into enterprise solutions, however, it will be a long and painful process to achieve clinically meaningful integration. We are still very far from the vision of a national information highway – even within a city or a region.

JS: Fifteen years ago, the report pointed out that healthcare services is a complex and technological industry prone to accidents, and that some systems are more prone to accidents because of the way the components do or don’t link together. Recently, there has been a great deal of discussion about the lack of interoperability in EHRs, and yet much of the burden of managing/interpreting/reprogramming bedside technology is related to an absence of medical device interoperability, which has gotten relatively little attention. What is the biggest challenge to ensuring that the varied medical devices/technologies engaged in patient care are seamlessly integrated, communicating and coordinated?

MC: What an irony – we rely upon IT-enabled devices to produce data to improve care, and at the same time recognize new errors due to failures in device interoperability and larger issues of siloed data sources. Device manufacturers themselves have recognized the problem, and the industry initiative for interoperability, Continua, has led efforts for common interface design in medical devices. Other industry leaders provide integration hubs and software for multiple independent devices, such as Qualcomm for mobile devices. Remote monitoring for patients in the home and community are increasingly supported by device-agnostic platforms. In the in-patient setting, sophisticated tele-ICU and other data interpretation systems detect early deterioration in patient status and reduce complications and shorten hospital and skilled nursing facility stays. As providers aggregate, their growing market power, and the shifting of financial incentives to reward them for positive outcomes, suggests that they will increasingly reward device manufacturers who build interoperable solutions.

JS: We believe in the potential for an automated, connected and coordinated system (or systems of systems) to help manage the complexity of healthcare, reduce medical errors and save lives and money. As someone who has been a part of the development and adoption of many new medical innovations and technologies, how do you see such an ecosystem evolving? Will we continue to innovate and deploy isolated point-solutions, each individually safe and effective, but each adding to the overall complexity of the enterprise? Will we put additional requirements on such ‘solutions’ – i.e., that they must smoothly integrate and interoperate with our existing systems?

MC: At UCLA Health, we’ve been tracking the evolution of new technologies and services for healthcare closely. Innovation is paying off – the number of new products and services entering the market each year with a high potential to improve quality and safety is rising steadily, and investment dollars are flowing into this sector. The consolidation of provider systems has meant that more delivery systems can afford larger and more sophisticated quality and safety programs, capable of integrating predictive modeling and near-real-time systems for the detection of patient deterioration, and of deploying remote monitoring for ambulatory patients at risk. Many of the innovations reduce the likelihood that patients will need to visit emergency rooms, be admitted or readmitted to hospitals, and in other ways be exposed to the potential for errors and quality gaps in institutional care. Some of them support more effective interventions in the course of chronic disease, from secondary prevention to intensive home-based coordination of multiple chronic diseases or advanced care planning services. Relatively simple solutions that focus on medication adherence, physiological monitoring and behavioral health monitoring and support are directly addressing the silos and gaps that have challenged population health.

JS: A fundamental principle described in the report was a need to respect human limits in process design. The report also called for technology to be recognized as a ‘member’ of the team. What do you see as the next big opportunity to use emerging technologies to help overcome human limitations in our delivery of safe, high-quality healthcare?

MC: In the original IOM committee, we studied airline systems to understand how system design and tools that combine information, communication and device technologies could solve problems inherent in human performance. In the 15 years since our reports, the identification of opportunities has exploded – but we have failed to take advantage of the potential. The President’s Council of Advisors on Science and Technology issued a report earlier this year, Better Health Care and Lower Costs: Accelerating Improvement through Systems Engineering, that gives inspiring examples of this approach, and describes what would be needed to encourage the development of systems engineering approaches more broadly throughout healthcare. One of the elements they emphasized was beginning with patient-centered design – they observed that involving patients in both the definitions of the goals and problems, and the solutions, will be essential to future progress.

Molly Joel Coye, MD, MPH, Chief Innovation Officer of UCLA Health at the University of California, Los Angeles is an internationally recognized leader in healthcare delivery policy and an expert in the use of information and clinical technology to advance the health of communities. As Chief Innovation Officer, Dr. Coye oversees the UCLA Innovates HealthCare Initiative, and is responsible for developing programs and strategies that promote and nurture innovation across the UCLA Health System.

Dr. Coye was elected to the National Academy of Sciences’ Institute of Medicine (IOM) in 1994 and co-authored two landmark reports on healthcare quality, To Err Is Human: Building a Safer Health System and Crossing the Quality Chasm. She also chaired the IOM’s Committee on Access to Insurance for Children, and co-chaired the Committee on Patient Safety Data Standards.