Descriptive statistics of the study participants appear in Table . Participants’ had an average of 12.3 years of medical work experience and 6.8 years of EMR experience. As expected, subject matter experts had the longest clinical and EMR experience, while residents had the shortest. Participants were not asked directly for their ages. However, a previous study of a similar population36
has shown that age and time since being graduated from medical school are highly correlated.
Descriptive Statistics of the Study Participants
EMRs and Cognitive Load
Study participants reported that clinical tasks such as diagnosing, reasoning, and treating severe or multiple medical conditions imposed the highest cognitive loads. Overall, study participants felt the EMR system reduced their cognitive loads. They were satisfied with the EMR system, especially with its data-related comprehensiveness, organization, and readability. They also appreciated that it made reviewing patients’ medical histories and test results easier (Table ). These aspects minimized the need to recall information from memory and eliminated the difficulty of reading handwriting. To some extent the EMR system also provided clinical decision aids. Physicians considered the system simple to use, and even those who had been using it for only 3–5 months reported completing many actions automatically (Table ). Researchers’ observations confirmed this view, with study participants quickly and nearly automatically performing system-related actions such as opening and closing charts, navigating between fields, and selecting items from lists.
Main Findings and Supportive Sample Quotations
EMRs and Perceptions of Patient Safety
EMR use exerted both positive and negative impacts on patient safety. On the positive side, physicians reported a reduction of their cognitive loads. They also said that the system’s decision-making aids improved the quality of patient care. These as well as other features of the EMR system—such as alerts of potential adverse drug interactions—were perceived to enhance patient safety.
On the negative side, EMR use provoked new types of medical errors. Typical errors reported by most study participants (>60%) were typos, adding information to the wrong patient’s chart, and unintentionally selecting an erroneous item (diagnosis or medication) from a scroll-down list located above or below the desired item (Table ). Study participants described two common scenarios for adding to the wrong chart. The first arose when they opened a chart by typing a patient’s name instead of using his or her unique magnetic card provided by the HMO. Typing in a patient’s name opened a list of patients, and physicians sometimes accidentally selected the incorrect individual. Some of the more experienced physicians reported knowing most of their patients by name. To save time, they reported, they would open such a person’s chart as soon as he or she entered the office and glance at the record before beginning the clinical interview. The second reported scenario arose when, in response to an interruption (e.g., a nurse asking about another patient or taking a telephone call from a patient), a physician would open another patient’s chart, forget to close it, and then type into it information about the visiting patient.
The study found that pharmacists often were the last safeguards against EMR-related errors. A number of study participants reported discovering they had written in the wrong patient’s chart only after a pharmacist alerted them that the name on a prescription differed from the presenting patient’s name (in Israel, paper prescriptions are mandatory even when generated by an EMR system). Similarly, study participants sometimes realized they had prescribed the wrong medication when concerned pharmacists asked whether they really had intended to prescribe a particular drug.
The findings indicate that study participants were aware of these potential errors. In all the cases observed, study participants reviewed printed prescriptions before signing and handing them to their patients. In several instances, study participants opened second charts as a result of interruptions; however, they always closed those charts before returning to their visiting patients. The only actual error that had been observed was a typo: a study participant had intended to write a letter confirming that a patient could exercise in a gym, but he instead wrote “can NOT exercise.”
EMRs and Patient–Doctor Communication
Of the study participants, 92% felt EMR use disturbed communication with their patients. Two physicians argued that “multitasking is not a problem” and that “those who say the computer interferes with communication simply have a communication problem.” Observational data indicate that physicians’ average screen gaze lasted from 25% to 55% of the visit time.
The physicians who participated in the study were able partially to overcome the negative impact of EMRs on communication by using various strategies and enabling factors. The main strategy entailed separating EMR use from time spent communicating with patients (Fig. ). Seventeen study participants reported the same encounter stages, although the sequence of events sometimes varied among physicians. During our observations we did not detect well-defined patient-visit stages. However, a clear separation between time spent inputting data into EMRs and time spent consulting with patients was noted. With one exception, physicians maintained eye contact with their patients and turned away from the computers. Most also did not touch their keyboards during conversations with their patients.
A typical sequence of a patient visit as described by study participants. Patient-centered stages are separated from EMR-centered stages (black). *Optional stage.
Physicians reported several communication-enabling factors in EMR settings. Computer skills, especially blind typing and the use of keyboard shortcuts and templates, reduced the burden of typing and, therefore, allowed more time for communication (Table ). One of the observed pediatricians used a predefined template for all physical examinations. Prior to an examination he would type a keyboard shortcut to insert the template and then, while the patient settled down after the examination, he would alter the data based on his findings. One study participant commented that a concern for patient safety led her to prefer typing over using templates; in her view, “Pressing Enter, Enter, Enter is a prescription for errors.”
Spatial organization of physicians’ offices also supported patient–doctor communication. Two typical models were observed or reported (Fig. ). In neither organization did the monitor interfere with eye contact. Three study physicians who employed the patient-centered model reported that this spatial organization often facilitated comments from patients, corrections, and greater information exchange.
Typical spatial arrangements of physicians’ offices.
Finally, study participants’ communication skills were another enabling factor that reduced the negative impact of EMR use on communication. The most influential skills observed were reading aloud while typing, maintaining eye contact, using body language to show attention and empathy, using humor to reduce tension, and leaving the computer completely and turning to the patient when conveying important information or discussing sensitive issues. However, in the 69 observed encounters only once did a study participant use the computer for patient education, i.e., to calculate a Framingham score using online software and to explain how a healthy lifestyle could reduce the risk of coronary heart disease.
Elements of Expertise
In the previous sections, a number of observed and reported elements of expertise were identified (Table ). Ensuring documentation comprehensiveness was another element of expertise the majority of study participants noted. They usually achieved this element by typing data into their EMR systems; however, two study participants reported preparing comprehensive templates for common problems or examinations and using them as checklists to ensure anamnesis completeness. Elements of expertise appeared to be individual rather than related to physician accreditation, formal educational stage, or years of experience.