Electronic Medical Records (EMR) can potentially improve quality and safety of ambulatory care. However, little research systematically documents the effect of EMRs on patient-centered care. Studies of the EMR's effect on patient-provider communication have been observational and had small sample sizes. Overall, these studies reported varied success regarding providers integrating the EMR into office visits, and suggest that further research is needed to evaluate the effectiveness of training providers in patient-centered communication and EMR use.
The PACE aims were to study how EMR use affects patient-provider communication behaviors, and patient-centered care and related health outcomes; to develop a unique provider training program tailored to patient-centered EMR use; and to evaluate the effect of the training intervention on patient-provider communication, patient-centered care, and provider EMR use.
The study used a quasi-experimental (pre-post intervention design) carried out in three phases:
(1) Pre-intervention: Twenty-three primary care providers and 126 patients were enrolled. A pre-intervention patient-provider visit was conducted for each patient-provider pair. Visits were video recorded and reviewed for verbal and nonverbal patient-provider communication. MORAE software was used to record provider-EMR interaction data, including page views, navigation, and mouse clicks. Data were collected for related outcomes (patient and provider satisfaction).
(2) Training: Findings from pre-intervention data guided development of a multifaceted provider training intervention promoting patient-centered EMR appropriation. The training intervention was delivered via a full day training workshop and individual feedback sessions.
(3) Post-intervention: A second round of visits (n=77) was conducted with the same patient-provider pairs. Data collection (described in the pre-intervention phase) was performed per protocol. Within group analyses (pre-post) were used to test whether the training intervention resulted in significant improvements in (a) patient-centered EMR use and (b) related outcomes (patient and provider satisfaction).
Of the 23 primary care providers enrolled in the pre-phase, 20 completed the post-phase (3 left the VA). The study enrolled 126 patients in the pre-phase, and 77 completed the post-phase. Reasons for patient drop out were as follows: 9 not interested; 3 passed away; 14 provider dropped out; 9 cancelled appointment/no show; 3 relocated; 3 changed PCP; and 8 we were unable to contact.
EMR Usage vs. Patient Engagement: The average time each provider spent during the visit performing EMR activity was 38.6% pre-intervention (sd=17.5%, median=33.5%, range=(0, 81.2%), as compared to 38.5% (sd=16.4%, median=39.2%, range=(8.8%, 69.3%) in post-intervention. The average time spent on patient engagement was 36% pre-intervention (sd=16%, median=35%, range=(5%, 76%), n=125), and improved to 38.9% (sd=17%, median=35.5%, range=(6%, 81%) post-intervention. Essentially, providers spent as much time engaged with the EMR as they did with patients. In interviews, providers expressed great concern regarding the need for multitasking between the EMR and patient. Providers recognized that existing EMRs are inefficient, which takes time away from patient-provider communication. Patients reported similar concerns but felt that technology use was inevitable and a good thing.
EMR Usage - Total Mouse Clicks: The average total mouse clicks per visit was 192 (sd=151, median=156, range=(0, 685), n=119) pre-intervention compared to 189 (sd=136, median=158, range=(31, 645), n=63) post-intervention. Providers shared the EMR screen with patients in 24.8% of pre-intervention visits. This improved to 28.9% in post-intervention visits. Mouse clicks are a quantifiable measure of EMR usage, and in both pre and post visits, the high volume of mouse clicks are evidence of poor usability and high task burden imposed by EMRs.
Patient-Centered Communication: Despite the heavy EMR burden on providers, 86.5% (pre) and 89% (post) of patients reported high satisfaction with provider's patient-centered communication. Pre-visit average score is 4.71 (sd=0.56, median=5, range=(1.67, 5); post-visit average score is 4.75 (sd=0.54, median=5, range=(2.17, 5). Similarly, 97.6% (pre) and 96% patients (post) were satisfied with provider's interpersonal skills. Pre-visit average score is 4.86 (sd=0.32, median=5, range=(3, 5), n=126); post-visit average score is 4.87 (sd=0.36, median=5, range=(3, 5).
Quality of Patient-provider Relationship: Providers, as a group, reported less satisfaction with visits than did patients. Only 70.6% (pre) and 74.3% (post) reported satisfaction with the quality of patient-provider relationships. Pre-intervention average score is 4.17 (sd=0.70, median=4.25, range=(1.75, 5); post-intervention average score is 4.23 (sd=0.64, median=4.25, range=(2, 5). Similarly, only 58.3% (pre) and 61.8% (post) of providers were satisfied with their ability to collect required data during visit. Pre-intervention average score is 3.95 (sd=0.86, median=4, range=(1.67, 5); post-intervention average score is 4.08 (sd=0.84, median=4, range=(1.33, 5).
In this study, a provider-coaching intervention was unable to significantly improve provider ability to incorporate the EMR in ways that facilitated patient-centered care. Contributing factors include EMR usability issues and inefficient workflows that shift the provider's focus from patient to computer. Pressed for time, providers multitask between patient interview and complex and inefficient EMR tasks. To populate notes, clinicians used various laborious mechanism, including direct typing, copy-paste from other parts of the EMR, and formatting the note. Providers use menu-driven interfaces for order entry of medications, lab tests, consults, and imaging. Such interfaces have multiple nested menus, long lists, and numerous data fields that require input. If EMRs are to support patient-centered care, they need to be redesigned to ease providers' EMR workload burden, thus increasing time spent in face-to-face communication with patients.
PACE findings emphasize the need to address EMR usability in next generation of EMR technologies being developed by the VHA hi2 (Health Informatics Initiative) and iEHR team. Drs. Agha and Calvitti have served as subject matter experts for hi2 HMP (Health Management Platform) User Centered Design team since 2011. A concrete impact from PACE is the hi2's $1 million commitment to fund usability research to support HMP and to develop software requirements for a usability analytics platform. PACE findings are also driving EHR redesign within the VA. Based on PACE data, Dr. Agha is leading a team of engineers, cognitive scientist, and clinicians at San Diego to prototype novel EHR user interfaces for hi2 HMP.
None at this time.
Technology Development and Assessment, Treatment - Comparative Effectiveness, Research Infrastructure
Behavior (provider), Informatics, Practice patterns