Who says what to whom through what channel? Formative communication research on antibiotic resistance messaging for urgent care patients

Rachel B. Wade; Monique M. Turner; Rana F. Hamdy; Youjin Jang; Ruth J. Heo; Cindy M. Liu

doi:10.1017/ash.2024.429

Who says what to whom through what channel? Formative communication research on antibiotic resistance messaging for urgent care patients

Published online by Cambridge University Press: 15 October 2024

Ruth J. Heo and

Rachel B. Wade: Affiliation:
Department of Communication, School of Arts and Sciences, The Ohio State University, Columbus, OH, USA
Monique M. Turner: Affiliation:
Department of Communication, College of Communication Arts and Sciences, Michigan State University, East Lansing, MI, USA
Rana F. Hamdy: Affiliation:
Children’s National Medical Center, Washington, DC, USA
Youjin Jang: Affiliation:
Communicating in Health Impact (CHI) Lab, University of North Carolina, Chapel Hill, NC, USA
Ruth J. Heo: Affiliation:
Department of Communication, College of Communication Arts and Sciences, Michigan State University, East Lansing, MI, USA
Cindy M. Liu*: Affiliation:
Antibiotic Resistance Action Center, Department of Environmental and Occupational Health, Milken Institute School of Public Health, George Washington University, Washington, DC, USA
*: Corresponding author: Cindy M. Liu; Email: cindyliu@email.gwu.edu

Article contents

Abstract
Objective:
Methods:
Participants:
Results:
Conclusions:
Introduction
Methods
Results
Discussion
Conclusion
Supplementary material
Author Contributions
Financial support
Competing interests
Research transparency and reproducibility
Footnotes
References

Abstract

Objective:

To explore the source, message, channel, and receiver effects on patient concern for antibiotic resistance, willingness to reduce antibiotic use, and expectations for an antibiotic prescription in a prepandemic sample.

Methods:

We used data reported from a national cross-sectional survey of adults who had visited an urgent care center within the last year. Data were collected from April 4 to April 9, 2017. The survey included an embedded experimental design to test changing effects before versus after message exposure.

Participants:

A national sample of adult participants (n = 610) who had used urgent care at least once in the past year were recruited through GfK’s KnowledgePanelTM. KnowledgePanel survey response rates are typically about 65%. Respondents ranged in age from 18 to 85 and were more likely to be female (377/610; 62%), White (408/610; 67%), and covered by private insurance (414/610; 68%).

Results:

Outcome variables were measured on 4-point scales 1–4 scale, and t-tests were conducted for measures that were collected pre and postmessaging. The majority of participants trusted their doctor and desired them as the source for information regarding antibiotic resistance, followed by field experts (eg, CDC). Direct messaging (eg, email) and targeted advertisements were least preferred.

Conclusions:

This study provides foundational data on patient communication preferences in terms of source, message content, and channel when receiving information on antibiotics and antibiotic resistance, as well as how these factors affect patient concern, willingness, and expectations. Follow-up work is needed to replicate these findings in a postpandemic sample.

Type: Original Article
Information: Antimicrobial Stewardship & Healthcare Epidemiology , Volume 4 , Issue 1 , 2024 , e177

DOI: https://doi.org/10.1017/ash.2024.429 [Opens in a new window]
Creative Commons: This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright: © The Author(s), 2024. Published by Cambridge University Press on behalf of The Society for Healthcare Epidemiology of America

Introduction

Antibiotic resistance is a major global public health crisis. Annually, 2.8 million people in the U.S. are infected with an antibiotic-resistant strain, and approximately 35,000 people die due to resistance.¹ The increase in antibiotic resistance and creation of antibiotic-resistant bacteria result from the overuse of antibiotics, inappropriate prescribing, extensive agricultural use, and a limited number of new antibiotics and alternatives. This leads to antibiotic-resistant bacteria that are resistant to even last-resort antibiotics, resulting in infections that become extremely difficult to treat and run the risk of spreading. Unfortunately, 1 in every 3 antibiotic prescriptions in outpatient settings were inappropriate for the presenting condition.^{Reference Ventola2}

Prior research has attempted to understand the patient’s role in the antibiotic prescribing process both for themselves and their children.^{Reference Cockburn and Pit3,Reference Mangione-Smith, McGlynn, Elliott, Krogstad and Brook4} Higher antibiotic expectations among patients have been shown to increase the likelihood of an antibiotic prescription.^{Reference Davey, Pagliari and Hayes5} A patient’s previous experience with antibiotics can also result in self-diagnosis if they present with the same conditions. This results in increased patient expectations which, in turn, leads to patients accessing antibiotics through physician prescriptions, over-the-counter online, other countries, or a previous prescription. This issue is particularly prevalent among urgent care patients, who are much more likely to be inappropriately prescribed antibiotics than patients with similar afflictions seen in retail health clinics.^{Reference Hyun and Zetts6} Though one strategy is to address physician overprescribing practices, another tactic for understanding and mitigating overuse is patient-focused antibiotic stewardship.

Prior research has found patients are open to receiving targeted communication messages regarding antibiotic resistance.^{Reference Van Hecke, Butler, Wang and Tonkin-Crine7} Moreover, providing information regarding antibiotic resistance has been shown to effectively mitigate inappropriate use.^{Reference Thorpe, Sirota, Orbell and Juanchich8} However, much of the work in this space has focused on making message content effective,^{Reference Roope, Tonkin-Crine and Herd9,Reference Sirota and Juanchich10} with less being known about patients’ preferred methods for receiving these communications. To address this, we take a sender-message-channel-receiver (SMCR) approach. Though simplistic, this model first proffered by D.K. Berlo is foundational in communication studies research today.^{Reference Roope, Tonkin-Crine and Herd9}

Communication scholars and practitioners believe it to be axiomatic that before an effective campaign can be launched, it is critical to understand preferences for who (source) says what (message) through what channel (modality) to whom (target audience) to achieve particular effects.^{Reference Roope, Tonkin-Crine and Herd9} Preferences for SMCR vary greatly across audiences and contexts; but campaign strategies cannot be developed in full without understanding who the audience trusts, the arguments they will find convincing, and the channels they prefer. Therefore, the variance of SMCR, and subsequent effects, was the focus of this formative research conducted with a national sample of urgent care patients. Urgent care patients were selected given that inappropriate prescription for respiratory conditions is highest in this patient population.^{Reference Palms, Hicks and Bartoces12} Our objective was to explore each of these components to determine which elements resonate with patients to change their expectations. This study additionally explored the changing levels of antibiotic resistance concern, willingness to reduce use, and expectations for prescription before versus after message exposure.

Methods

Study design

A cross-sectional quantitative survey was conducted using a national sample of participants recruited through GfK’s KnowledgePanel^TM, a probability-based online panel of adults that is recruited using address-based sampling. The survey was designed by the authors and Strategies 360, a Washington D.C.-based research firm contracted to implement the study. As part of their demographic profiling, GfK had collected data about urgent care use. This information was used in targeting potential survey participants, as well as in setting demographic quotas. Live telephone surveys were conducted between April 4 and 9, 2017.

The study received ethical approval from the George Washington University Human Subjects Review Board. Post hoc power analysis using G*Power (two-tail; d = .03; ɑ = 0.05; Power =.95) suggests a sample size of 580 to conduct means difference t-tests on independent groups. Considering Strategies 360’s predetermined margin of error for a survey of 610 is ±4.0% at the 95% confidence interval, the study was sufficiently powered. Though an exact response rate was not provided, the typical response rate for KnowledgePanel^TM is 65%.^{Reference Poushter, Gubbala and Lippert13}

All participants were asked to confirm they had been an urgent care customer within the last year, and about their health insurance and most frequently visited medical services (Table 1). The survey included one embedded, randomized, experimental component (Figure 1). This design allowed us to make some inferences about the persuasive potential of the messages. Half of the participants (Group A) were asked about their expectations regarding antibiotic prescription before viewing persuasive evidentiary messages. The other half (Group B) received these questions after viewing the messages. Moreover, those in Group A were asked about their willingness to reduce antibiotic use for “bacterial infections” and trust in “your doctor” and Group B participants were asked the same questions but about “viral infections” and “an urgent care doctor,” respectively (Table 2). For all other questions, Groups A and B were treated the same. In addition, all participants responded before and after viewing persuasive messages regarding their concerns and willingness to reduce use.

Table 1. Sample demographics

All n values are out of n = 610.

Figure 1. Flow chart of survey order and experimental design for group A and B.

Table 2. Source trust ratings

^a Indicates question that only Group A answered.

^b Indicates question that only Group B answered. All other questions were reported by both Group A and B. Likert scaled 1 “very much” to 4 “not at all.”

Measures

The survey included 69 questions, including nine demographic variables, nine questions about urgent care perceptions, two questions capturing health literacy, and 49 questions about antibiotics.

Source

Credibility

Participants indicated how trustworthy they found message sources to be from 1 “very much” to 4 “not at all.” Examples include doctors, the World Health Organization (WHO), and friends/family (Table 2).

Message

Persuasiveness of evidentiary statements

Nine potentially persuasive arguments were presented halfway through the survey. Messages were intended to be diverse and spanned from biological to economic and social reasons. To gauge the persuasiveness of these evidentiary statements, participants were asked to mark how “convincing a reason” each statement was in helping to reduce antibiotic resistance on a scale from 1 “very convincing” to 4 “not convincing at all” (Table 3).

Table 3. Messages: ratings of how convincing each message was and ranking of benefits to help reduce antibiotic resistance

Likert scaled 1 “very convincing” to 4 “not convincing at all.”

Benefits of antibiotic resistance reduction

Participants were asked, “In your own words, what is the most important reason to help reduce antibiotic resistance?” Responses were recorded verbatim and coded to fit nine a priori thematic codes (Table 4).

Table 4. Preferences for channels and benefits of antibiotic resistance reduction

All n values for the preferred methods of communication are out of n = 610. Participants could choose up to 3 channel options. All the n values for the perceived benefits of antibiotic resistance reduction are out of n = 416 due to 194 participants opting not to give an open-ended response. Participant responses were recorded verbatim and coded into one of nine predetermined thematic codes.

Channel

Method of communication

Participants were asked, “Which 3 of the following ways would you most like to get information about antibiotic resistance?” then shown a list of 13 potential communication channels in a randomized order (Table 4).

Antibiotic resistance attitudes and behaviors

Concern

Concern about antibiotic resistance was measured before and after exposure to evidentiary statements from 1 “Very concerned” to 4 “Not concerned at all” but then reverse-coded such that greater values indicated more concern (Table 5).

Table 5. Evidentiary message effects of antibiotic expectation by symptoms, antibiotic resistance concern, and action step willingness

Group A n = 302. Group B n = 308. Expectation for antibiotics by symptom was measured on a 4-point scale and reverse-coded so higher numbers indicate greater expectation. Concern and willingness were also measured on 4-point Likert scales and reverse-coded so higher values indicate greater concern/willingness.

Expectations

Six distinct symptomsFootnote ¹ (ie, ear pain, a deep cut/bad scrape, sore throat, fever, cough, stomachache) were presented to participants in random order. Participants were asked how likely they would be to expect antibiotics for each symptom from 1 “definitely expect” to 4 “definitely not expect” (reverse-coded; Table 5). Half of the sample reported expectations before exposure to evidentiary statements (Group A) and the other half reported after exposure (Group B).

Willingness

Willingness to help reduce antibiotic resistance was measured before and after exposure to persuasive messaging from 1 “Very willing” to 4 “Not willing at all” (reverse-coded; Table 5). Participants were asked, “To help reduce antibiotic resistance, how willing would you be to use antibiotics less often for [bacterial/viral] infections?” Half of the sample answered the bacterial question (Group A), the other half (Group B) answered the viral question.

Action steps

Participants indicated how likely they would be to take behavioral steps (eg, asking a provider if there are steps to take to feel better without antibiotics) after exposure to evidentiary statements. Responses ranged from 1 “very likely” to 4 “not likely at all.”

Statistical analysis

All analyses were completed with SPSS 29. For measures that were collected pre- and postmessaging, a paired samples t-test was used to test for changes in response to messages. For measures that were collected across Groups A and B, an independent samples t-test was used to test for changes in response to messages. Our predetermined α level was .05.

Results

Sample results

Adults (n = 610) who had used urgent care at least once in the past year completed this survey (Table 1). Most respondents were 18–44 years old (341/610; 56%), and 53% (322/610) were married. Females comprised 62% (377/610) of the sample, 67% (408/610) were White non-Hispanic, and 26% (158/610) were parents to a 0- to 17-year-old child at the time of the study. Almost all were covered by health insurance (583/610; 95%) with most having private insurance (411/610; 67%).