Can Targeted Messaging Encourage PCP Contact Before ED Visits?
Published Online: January 21, 2013
Maria C. Raven, MD, MPH, MSc; Scott M. Kotchko, MA; and David A. Gould, PhD
The use of emergency departments (EDs) by consumers who have ready access to a primary care physician (PCP) can increase care fragmentation1 and may also ncrease healthcare expenditures if equally effective care can be provided in an alternative setting.2 Addressing ED utilization can be challenging for capitated plans that must provide high-quality care while accounting for rising healthcare costs.3
Recent studies indicate that Medicaid beneficiaries have rates of ED use disproportionate to other insured populations.4-6 States are currently moving increasing numbers of Medicaid beneficiaries into capitated plans as both a cost control and care coordination effort.7,8 This creates an additional incentive for capitated plans to safely and effectively influence members’ decisions about ED use. The goal of our study was 2-fold. First, we developed messages (Phase I) designed to encourage members of a Medicaid capitated plan who had frequent ED use to contact their PCP practice prior to visiting the ED. Next, we assessed the messages’ impact on patient-reported potential to change care-seeking behaviors.
This study focused on the ED use of enrollees of Health Plus, a notfor-profit health plan based in Brooklyn, New York, that participates in New York’s Medicaid and Child Health Plus programs. Whitman Insight Strategies, a strategic research and communications firm, conducted focus groups, interviews, and surveys. We recruited participants from 2 New York City neighborhoods—Southwest Brooklyn and the South Bronx—because each is home to large numbers of Health Plus (Medicaid capitated plan) members. All Health Plus member participants had an assigned PCP and access to a free 24-hour nurse call line.
The study was reviewed and approved by the institutional review board at the Lutheran Medical Center in Brooklyn.
Phase I: Message Development
To provide a basis for developing messages to help patients seek care in the most appropriate setting, we conducted a qualitative prestudy with Health Plus members identified as having had 5 or more ED visits in the previous 12 months.
Health Plus Member Interviews
We conducted in-depth interviews with participants from Health Plus member families in which multiple family members had visited the ED at least 5 times in the past 12 months. Since we were interested in the use of EDs for non-emergent care, we excluded patients who were subsequently admitted for inpatient care or were pregnant during the 12 months. Contact information was provided by Health Plus and families were contacted at random. Participants received a $25 gift card. Interviews and the subsequent surveys were conducted, using Computer Assisted Telephone Interviewing techniques, by trained interviewers employed by our data collection partner. Interviewers were fluent in English and Spanish, and interviewed the member of the household who identified her- or himself as the family’s medical decision maker. Interviews aimed to: a) gauge participants’ understanding of the role of both the ED and their PCP in their overall care and that of their families; and b) measure awareness of factors relating to ED usage and the decision-making pathway leading to ED visits and PCP visits. After reviewing results from 30 interviews, it was agreed that we had reached a point of theoretical saturation (no new conceptual insights would be generated).9
Phase II: Messages and Patient-Reported Behavioral Change
Phase I findings were used to design and evaluate a series of messages meant to encourage appropriate use of one’s PCP as an ED alternative.
From August 5 to August 15, 2010, messages were evaluated via a quantitative phone survey with 118 Health Plus members. Participants were obtained from a sample of 528 Health Plus member households with frequent ED use (families with 4 or more ED visits in the past 12 months with at least 2 family members making ED visits). The threshold number of ED visits was lowered from 5 per family to 4 per family to ensure a sufficiently quantifiable sample size. Data on non-responders were not collected. Interviews testing messages were conducted with the member of the household who makes medical decisions and were conducted in English or Spanish by trained interviewers. Participants received $25 gift cards.
Interviews were structured in a “prepost” format. First, participants were asked their preferences on visiting their PCP versus the ED (predictor variable). Next, participants were exposed to and asked to evaluate a series of messages, after which their preferences were rechecked to determine whether the messaging had any measurable impact on reported care-seeking preferences related to PCP versus ED use (primary outcome variable). Every participant evaluated and rated every message in a randomized order. By asking an initial “pre-test” question, exposing respondents to numerous individual messages that attempt to influence behavior related to that question, then reasking the same initial question after exposure to messages, we could design a regression model examining the impact of each message on the patient’s stated willingness to contact the PCP for guidance about whether or not to go to the ED.
This methodology allowed the initial PCP versus ED preference to be held constant in order to determine which messages had the strongest impact on persuading respondents to contact their PCP prior to an ED visit.
We employed an ordinary least squares linear regression analysis to determine which message or messages were most highly correlated with an increased likelihood of contacting the PCP instead of the ED. The impact of individual messages with their respective beta coefficients is reported.
Phase I: Message Development
In patient interviews, 3 themes emerged: perceived barriers to PCP use; assumptions about appropriate care setting; and lack of alternatives to the ED.
Theme 1: Perceived Barriers to PCP Use
Overall, wait times and PCP availability were the main barriers patients cited that hindered their acceptance of the PCP as a first choice for medical attention. Many respondents indicated that their PCPs were too far away to visit on short notice, and reported EDs kept them waiting for a shorter period of time than their PCP’s office. Participants indicated a willingness to call their PCPs prior to visiting the ED, and many had tried but were unsuccessful.
Theme 2: Assumptions About Appropriate Care Setting Based on Presenting Problem
Most participants chose to see their PCP for preventive care and health maintenance issues. In general, participants reported specific reasons including prolonged or acute symptoms that, while potentially manageable in an ambulatory setting, led them to choose the ED. Participants reported PCP offices are not equipped to handle certain types of illnesses and injuries, and that only EDs have the ability and equipment necessary for some types of procedures.
Theme 3: Lack of ED Alternatives
Overall, participants reported resistance to using services such as walk-in clinics instead of EDs because many remain unfamiliar with them; however, the concept was appealing to some who use EDs for routine medical care simply because their PCP is not available at the time when care is needed.
Phase II: Messages and Patient-Reported Behavioral Change
Of the 118 Phase I participants, 86% were female. Most (58%) were between the ages of 25 and 44, 93% had at least 1 child in the home, and 69% had less than a college education. Half of the participants self-identified as Hispanic or Latino, 7% were black or African American, and 19% were white.
Most Phase II families (85%) believed their family was in “excellent” or “good” health. The majority indicated they regularly visited their PCP and specialists and were satisfied with the care they received (see Table 1). Consistent with our findings from the Phase I patient interviews, participants stated they go to the ED because of convenience and perceived urgency, and at times they believe their PCP office to be closed. Most (64%) respondents indicated that when a family member needs care the first thing they do is go to the ED or call an ambulance, and 85% of participants indicated satisfaction with their ED visits.
Using Messaging to Change Behavior
A total of 14 messages were developed based on Phase I data. Messages were evaluated based on their ability to influence participants’ report that they would contact their primary care physician instead of the ED.
Messages are listed in Table 2 with their beta (standardized regression) coefficient. The beta coefficients represent the patients’ likelihood of visiting their PCP comparing pre- and post-message exposure. Larger beta coefficients have higher relative importance in a multiple regression model: for example, an item with a beta of 0.3 is twice as powerful as is an item with a beta of 0.15. The model was significant at the 98th percentile.
With a beta coefficient of .29, the message that yielded the greatest likelihood to persuade respondents to contact their PCP instead of the ED was:
PDF is available on the last page.