Brain Drain

Data, information, and the patient-physician encounter are at the heart of medicine and health. William Osler (the father of modern medicine) recognized early the role of patient-focused data-driven decision making when he stated that, “the practice of medicine is an art, based on science.” Translated into modern medicine, the physician gathers information from “indirect” sources such as the electronic health record and clinical guidelines and also from “direct” sources – namely the patient – in the form of a history and physical exam. “Direct” patient care is the foundation of the “art of medicine” and it is during this time spent with the patient, that the physician develops the trust that facilitates shared decision making.

However, as physician visits shorten in duration, and concomitantly, the disease burden of an aging population increase, physicians spend the majority of their clinical time in accessing, retrieving, and processing indirect information. During a busy patient day, physicians need to find a balance between managing a large number of patients and thoroughly reviewing a patient’s medical history. Workflow mapping studies have shown that physicians spend only 15-17% of their time on “direct” patient care, whereas, almost 65-67% of their time is spent on “indirect” patient care. The vast majority of time is spent reviewing a patient’s electronic health record and examining varied sources of information ranging from documents and medical images to charts and laboratory values. This imbalance not only takes a toll on the doctor-patient relationship but also represents a lost opportunity to have the patient actively engaged in their care. For example, it has been demonstrated that shorter doctor visits result in an increased likelihood for a prescription for medicine rather than a discussion of behavioral change.

Despite spending the majority of time in “indirect” care, physicians are unable to find the information they need at the point of care. According to studies, 50% of the information needs that clinician’s raise in the context of patient care are unmet. Physicians have approximately eight unanswered questions for every ten ambulatory care visits. They have unmet information on the order of two questions for every three patients and pursue answers for only 30% of those questions. Additionally, the constant barrage of interruptions typical of a clinical day leads to an even more burdensome cognitive load. For example, ED physicians are interrupted on average once every ten minutes with more than 10% of tasks being interrupted.  If you add the inefficiency of information retrieval with the burden of interruptions, it is no surprise to find physician decision making impacted by this increased cognitive load. As we head towards Stage 3 of meaningful use, there will be an increased push towards clinical decision support (CDS). CDS is traditionally only thought of as alerts but a key aspect of CDS will be to organize and present disparate data into logical, intuitive schemas at the point of care. Context-specific dashboards that integrate relevant information across all available sources and differentially display information pertinent to the patient’s presentation will not only help practitioners reduce the time spent searching for relevant data but also assist them in processing this data. Ultimately CDS tools, will not only enhance the decision-making but will also enable physicians to spend more time engaging in the “art of medicine.”

As EHR’s become the primary repository for all generated clinical data, the amount of data stored will increase exponentially. For example, a large integrated health system such as Kaiser Permanente manages up to 44 petabytes of data through its EHR. As we start to incorporate ‘omics and sensor data, the amount of information for a physician to process will continue to increase exponentially. This presents a huge challenge as the quantity and complexity of data will overwhelm human cognitive capacity and render clinical judgment ineffective. However, this also presents a big opportunity to develop enhanced information retrieval and information processing tools within the EHR. These tools will not only improve clinical decision-making but also increase the time for the provider to spend in “direct” patient care, and thereby, improve the patient-physician encounter.