Good communication is essential to good clinical care. As much of the management of health and wellness occurs outside of episodic clinical visits, patients and clinical teams must engage in a continuous conversation to support information sharing, titrate treatments, support patient self-management, and establish rapport. Communication behaviors can occur in face-to-face visit, on telephone calls, using letters, and most recently using asynchronous electronic communication channels (such as secure messaging within a patient portal).
Measurement of the patient-centered aspects of Patient Aligned Care Teams is a high priority for VA. Although communication behaviors are critical to high quality healthcare, they are difficult to measure across the VA. Communication behaviors have been rigorously measured in research studies using direct audio-video tapes of in-person and telehealth communications. These approaches are expensive, requiring infrastructure, transcription, and manual coding. With the advent of Secure Messaging through patient portals, VA has a new opportunity to directly measure communication behaviors as the written communication between patients and clinical teams is included in the VA clinical records.
In this project, we propose to advance knowledge and methods related to communication behaviors measurable through asynchronous secure messages.
Specific Aim 1: Mine communication codes.
Specific Aim 2: Define communication behavior indicators (CBIs) that represent clinically meaningful measures of Secure Message communication patterns between Veterans and Clinical Teams.
Specific Aim 3: Understand experiences of clinical teams with high (and low) rates of CBIs in SMs.
Using a longitudinal cohort design, with a nested case-comparison mixed
methods study, we propose the following methods for each specific aim:
Specific Aim 1: Using a national corpus of secure messages, we will use natural language processing techniques to detect communication behaviors (information seeking, information sharing, socio-emotional exchanges) in patient messages, and messages from primary care doctors, nurses, and other PACT staff.
pecific Aim 2: At the level of the individual PACT team, assess the association of Secure Messaging communication behavior codes with important healthcare processes including:
Aim 2.1 Look for convergent validity of measured communication in secure messages by linking codes mined in Specific Aim 1 with Survey of Health Experiences of Patients (SHEP) measures related to patient perceptions of communication with their clinical team.
Aim 2.2 Patient access and use of healthcare services (both primary care and emergency department/urgent care)
Aim 2.3 Patient adherence to medications (measured by medication possession ratios)
Specific Aim 3: Conduct a study of positive deviance, using a mixed-methods Case-Comparison design, Among teams with high-volume secure messaging use, conduct site visits (N = 5) with PACT teamlets with the highest positive values for Secure Messaging communication and also highest values on SHEP measures and a comparison group of teamlets (N = 5) with similar volume of Secure Messaging, but lower levels of communication behaviors and SHEP scores.
There are no findings yet.
The long-term goal of this research is to develop NLP systems that use machine learning and other strategies to reliably identify communication behaviors in SMs with a precision comparable to that of trained human coders. Using the system, the VA will be able to directly measure what is now un-measurable: communication between Veterans and clinical teams. Using the system, we can create indicators of positive and negative communication behaviors.
None at this time.
TRL - Applied/Translational, Treatment - Observational, Treatment - Implementation
Information Management, Surveillance, Technology Development