Original Investigation | Substance Use and Addiction

Effect of Exposure to Visual Campaigns and Narrative Vignettes

on Addiction Stigma Among Health Care Professionals
A Randomized Clinical Trial
Alene Kennedy-Hendricks, PhD; Emma E. McGinty, PhD, MS; Amber Summers, PhD; Susan Krenn, BA; Michael I. Fingerhood, MD; Colleen L. Barry, PhD, MPP

Abstract Key Points

Question Are specific communication
IMPORTANCE Stigma toward people with opioid use disorder (OUD) is pervasive in clinical settings,
strategies effective in reducing stigma
impeding delivery of high-quality care. To date, no study has evaluated the effect of different
toward people with opioid use disorder
stigma-reduction messages or messengers among health care professionals.
(OUD) among health care professionals?

OBJECTIVE To evaluate the effect of OUD-related messages delivered by different messengers on Findings In this randomized clinical trial
stigma and attitudes toward people with OUD among health care professionals. involving a national sample of 1842
health care professionals, exposure to
DESIGN, SETTING, AND PARTICIPANTS This randomized clinical trial examined the effects of visual campaigns combined with short
OUD-related messages delivered by a visual campaign alone or in combination with a written narrative vignettes told from the
narrative vignette from the perspective of 1 of 3 messengers. Health care professionals in the US perspective of a patient with OUD that
were recruited from 2 national online survey panels (Ipsos KnowledgePanel and emphasized the harm of stigmatizing
SurveyHealthcareGlobus). A total of 1842 participants completed a web-based survey measuring language or the effectiveness of
stigma toward people with OUD from November 13 to 30, 2020. medications for treating OUD was
associated with reduced levels
INTERVENTIONS Eight groups were exposed to 1 of 2 message frames. One frame (Words Matter) of stigma.
emphasized the harm of stigmatizing language, and the other (Medication Treatment Works) focused
Meaning The findings of this
on the effectiveness of medications approved by the US Food and Drug Administration for the
randomized clinical trial suggest that
treatment of OUD. Message frames were communicated through either a visual campaign alone or a
carefully designed communication
visual campaign in combination with a written narrative vignette from the perspective of a simulated
campaigns may reduce OUD-related
patient with OUD, a clinician, or a health care system administrator.
stigma among health care professionals.

MAIN OUTCOMES AND MEASURES Dimensions of stigma toward people with OUD were
measured on 5-point Likert scales that included items about desire for social distance from people + Invited Commentary
with OUD, perception of individual blame for OUD, perspective of OUD as a medical condition, and
support for increased governmental spending on OUD treatment. The level of warmth felt toward
+ Supplemental content
Author affiliations and article information are
people with OUD was measured by a feeling thermometer (range, 0-100 points).
listed at the end of this article.

RESULTS Among 1842 participants, the mean (SD) age was 47 (13) years; 1324 participants (71.9%)
were female, 145 (7.9%) were Hispanic, 140 (7.6%) were non-Hispanic Black, 1344 (73.0%) were
non-Hispanic White, and 213 (11.6%) were of other non-Hispanic race (ie, individuals who did not self-
report race as Black or White and did not self-report ethnicity as Hispanic). Compared with
nonexposure, exposure to the combination of visual campaign and narrative vignette
communicating the importance of nonstigmatizing language from the perspective of a patient with
OUD was associated with a lower probability (difference, −16.8 percentage points, 95% CI, −26.1 to
−7.4; P < .001) of unwillingness to have a person with OUD marry into the family (a measure of social
distance preference) and a 7.2-point (95% CI, 3.2-11.1; P < .001) higher warmth rating. Participants
exposed to the combined visual campaign and patient vignette about the value of medication

(continued)

Open Access. This is an open access article distributed under the terms of the CC-BY License.

JAMA Network Open. 2022;5(2):e2146971. doi:10.1001/jamanetworkopen.2021.46971 (Reprinted) February 4, 2022 1/14

Downloaded From: https://jamanetwork.com/ by Nguyen Tuan on 05/02/2022

 JAMA Network Open | Substance Use and Addiction Effect of Visual and Narrative Campaigns on Addiction Stigma in Health Care Professionals

Abstract (continued)

treatment for OUD also had significantly lower levels of stigma compared with those in the
nonexposed control group (eg, unwillingness to have a person with OUD as a neighbor: difference,
−15.3 percentage points; 95% CI, −24.6 to −6.0; P = .001).

CONCLUSIONS AND RELEVANCE In this study, messages about nonstigmatizing language and
effective medication for OUD reduced stigma among health care professionals. Stigma-reduction
efforts targeting health care professionals may improve health care system capacity to serve people
with OUD.

TRIAL REGISTRATION ClinicalTrials.gov Identifier: NCT05127707

JAMA Network Open. 2022;5(2):e2146971. doi:10.1001/jamanetworkopen.2021.46971

Introduction
Overdose mortality rates have worsened during the SARS-CoV-2 pandemic.1,2 Stigma toward people
with opioid use disorder (OUD) has been an intractable challenge to ameliorating this problem.3-5
Defined through a sociological perspective, stigma occurs when the processes of labeling,
stereotyping, status loss, and discrimination are enforced through power differentials.6 Stigma
toward people with OUD and other substance use disorders is rooted in racism, classism, and other
systems of oppression.7,8 In the health care system, stigma manifests at multiple levels.9 Expectation
of stigma may reduce treatment seeking and engagement with care.10,11 Stigma enacted by clinicians
may result in worse quality of care and exacerbate patient mistrust.4,5,12,13 On a structural level,
stigma contributes to discriminatory policies and reduced investment in systems that support people
with OUD.5,14 Although interest in stigma-related issues has grown,3 the evidence base for stigma-
reduction strategies remains limited.4,15
Previous studies have reported a connection between language and stigma. Terms such as
addict and substance abuser have been associated with greater stigma relative to person-centered
language, such as person with a substance use disorder.16-18 Clinician use of stigmatizing language
may translate to worse quality of care.13 However, stigmatizing language remains common in clinical
and public discourse.19,20
In addition to discouraging use of stigmatizing language,21,22 communicating about solutions to
OUD and overdose, including effective treatment, may mitigate stigma.4,23-25 Most health care
professionals receive minimal training in addiction and its treatment.26 Health care professionals may
be more aware that they are interacting with patients with OUD when these individuals display OUD
symptoms and may be unaware that they are interacting with patients with OUD when these
individuals are in recovery,12 which could distort the practitioner’s sense of the potential for recovery.
Despite the robust evidence base in favor of medication treatment, stigma has impeded broad
availability of highly effective medications approved by the US Food and Drug Administration to treat
OUD.27 Stigma toward the opioid agonists methadone and buprenorphine is related to the
misperception that these medications replace 1 drug for another.28,29 Among clinicians, higher levels
of stigma have been associated with less interest in prescribing medication for the treatment of
OUD.26,30,31 Sympathetic narratives that highlight solutions such as effective treatment can help
audiences empathize with a person with OUD.4,15,23,25,32 Communication strategies focusing on the
effectiveness of medication in facilitating recovery may be particularly persuasive to an audience of
health care professionals.
Health care professionals are frequently exposed to messages from patients, other clinicians,
and health care system administrators. A health care professional may better relate to
communication from a fellow health care professional but be more emotionally engaged when
messages are delivered by patients. To our knowledge, no study has evaluated the effectiveness of

JAMA Network Open. 2022;5(2):e2146971. doi:10.1001/jamanetworkopen.2021.46971 (Reprinted) February 4, 2022 2/14

Downloaded From: https://jamanetwork.com/ by Nguyen Tuan on 05/02/2022

 JAMA Network Open | Substance Use and Addiction Effect of Visual and Narrative Campaigns on Addiction Stigma in Health Care Professionals

different stigma-reduction messages or messengers among health care professionals. Understanding

the ways in which different messages and messengers affect stigma and related attitudes among
health care professionals is important to informing the design of effective stigma-reduction
communication campaigns in health care settings. In this study, we enrolled a large national sample
of health care professionals to assess the impact of 2 message framing strategies communicated
through a visual campaign with and without an accompanying written narrative vignette from the
perspective of a patient with OUD, a clinician, or a health care system administrator.

Methods
From November 13 to 30, 2020, we conducted a parallel-group randomized clinical trial involving
1842 health care professionals recruited from a blended sample of 2 online survey panels (Ipsos
KnowledgePanel and SurveyHealthcareGlobus). The study was approved by the institutional review
board of Johns Hopkins Bloomberg School of Public Health. All participants were members of 1 of
the 2 online panels and provided consent to participate in surveys at the time of panel enrollment
(with incentives for participation provided by the panel administrators). Therefore, this study was
deemed exempt from the need for additional informed consent. The trial protocol was registered
retrospectively because the original conception was a survey experiment (Supplement 1). This study
followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline for
randomized clinical trials.

Data
The study sample included 1242 Ipsos KnowledgePanel members who self-reported their current
employment in a health care profession. Ipsos uses probability- and address-based sampling to
create an online panel comprising 60 000 members representative of the US population.33 Among
2703 KnowledgePanel members invited to participate in the study, 1484 completed the online
survey, and 1242 qualified for study inclusion based on current employment in a health care
profession (Figure).
To ensure greater representation of physicians and nurses, we supplemented the
KnowledgePanel sample with 600 members of SurveyHealthcareGlobus, a health care market
research firm with an opt-in online panel of physicians and nurses.34 Among the 111 105
SurveyHealthcareGlobus panelists invited to participate, 10 265 completed the online survey, and
600 qualified for study inclusion. Our sampling approach was informed by an effort to enroll a large
diverse group of health care professionals from across the US. We did not focus on clinicians
dedicated to treating OUD because stigma affects quality of care for patients with OUD across all
health care settings.

Study Design
Ipsos used a probability proportional to size sampling approach to select random samples
corresponding to 8 exposure groups and 1 nonexposure control group. Each of the 9 groups included
approximately 200 participants, providing adequate statistical power to detect medium-sized
differences between groups. The 8 groups were exposed to 2 message frames, Words Matter and
Medication Treatment Works, which were communicated through a visual campaign alone or
together with a written narrative vignette from the perspective of 1 of 3 messengers: a simulated
patient with OUD, a clinician, or a health care system administrator (eTable 1 in Supplement 2). Thus,
for each message frame, exposed participants were randomized to 1 of 4 groups: visual campaign
only, visual campaign plus patient vignette, visual campaign plus clinician vignette, or visual
campaign plus administrator vignette.
The Words Matter message frame discouraged use of stigmatizing language related to
substance use (including OUD) and encouraged use of nonstigmatizing language. The Medication
Treatment Works message frame highlighted the value of US Food and Drug Administration–

JAMA Network Open. 2022;5(2):e2146971. doi:10.1001/jamanetworkopen.2021.46971 (Reprinted) February 4, 2022 3/14

Downloaded From: https://jamanetwork.com/ by Nguyen Tuan on 05/02/2022

 JAMA Network Open | Substance Use and Addiction Effect of Visual and Narrative Campaigns on Addiction Stigma in Health Care Professionals

approved medications in facilitating recovery. These message frames were developed during a 2-day
communication design workshop held in September 2019 involving 25 invited participants, including
individuals receiving treatment for OUD, peer recovery coaches, nurses, physicians, administrators,
communication experts, and researchers.
The visual campaigns for the Words Matter and Medication Treatment Works message frames
both displayed headline text stating, “What we say and do matters for patients with substance use
disorder,” with images of clinicians interacting with patients (eFigure in Supplement 2). The Words
Matter visual campaign also displayed stigmatizing terms to avoid along with preferred alternative
nonstigmatizing terms and invited participants to take a pledge committing to the use of
nonstigmatizing language. The Medication Treatment Works visual campaign included text that
stated, “Medications help patients recover and live full lives. Don’t let misperceptions get in the way.
Learn more about methadone, buprenorphine (also called Suboxone or Subutex), and injectable
extended-release naltrexone (also called Vivitrol).”
Participants randomized to receive exposure to the visual campaign in combination with a
vignette viewed the visual campaign first, followed by a screen displaying the written narrative
vignette. The vignettes were approximately the same length (165 words) and communicated the
message frame from the perspectives of the 3 different messengers (patient with OUD, clinician, or
health care system administrator) (eTable 2 in Supplement 2). For instance, in the patient vignette for

Figure. CONSORT Flow Diagram

113 808 Individuals assessed for eligibility

111 105 From SurveyHealthcareGlobus
2703 From Ipsos KnowledgePanel

111 966 Excluded

110 505 From SurveyHealthcareGlobusa
100 840 Did not complete screening questions or survey
9665 Did not meet health care profession inclusion
criteria
1461 From Ipsos KnowledgePanelb
1219 Did not complete screening questions or survey
242 Did not meet health care profession inclusion
criteria

1842 Randomizedc

201 Randomized to receive no message 819 Randomized to receive words matter message 822 Randomized to receive medication treatment
201 Received no message as randomized 819 Received words matter message as works message
randomized 822 Received medication treatment works
message as randomized

206 Received 613 Received visual plus 205 Received 617 Received visual plus
visual format written narrative format visual format written narrative format
only 205 Received clinician only 208 Received system
vignette leader vignette
205 Received system 205 Received patient
leader vignette vignette
203 Received patient 204 Received clinician
vignette vignette

1842 Included in the final analysis

a
Excluded if not working in health care profession at time of study. Nurses, physicians categorized as general practitioners or family doctors, and physicians categorized as specialists
were included to achieve 600 participants, with approximately one-third represented from each category.
b
Excluded if not working in health care profession at time of study or if working in a nonqualifying health care profession, such as dentist, dental hygienist, dental assistant,
optometrist, veterinarian, veterinary assistant, or massage therapist.
c
Ipsos KnowledgePanel used a probability proportional to size sampling approach to select 9 random samples.

JAMA Network Open. 2022;5(2):e2146971. doi:10.1001/jamanetworkopen.2021.46971 (Reprinted) February 4, 2022 4/14

Downloaded From: https://jamanetwork.com/ by Nguyen Tuan on 05/02/2022

 JAMA Network Open | Substance Use and Addiction Effect of Visual and Narrative Campaigns on Addiction Stigma in Health Care Professionals

the Words Matter message frame, the patient described interacting with clinicians using language
such as addict and feeling as though the clinicians “didn’t see me as a person” and “all they could see
was my addiction.” In the clinician vignette, the clinician described using terms such as addict and
learning that this language made patients feel as though “I didn’t see them as people” and “all I could
see was their addiction.” In each narrative vignette, the patient, clinician, and health care system
administrator evolved in their understanding of the harm of stigmatizing language, and the vignettes
ended by noting that “health care professionals can be role models.” The Medication Treatment Works
vignettes included similar content and language across the 3 different messenger narratives and
shared themes with the Words Matter message frames, including an emphasis on health care
professionals as role models. The Medication Treatment Works patient vignette featured a person
recovering from OUD with the help of medication; the clinician and administrator vignettes described
positive experiences with providing medication for the treatment of patients with OUD.

Measures
After exposure to the visual campaign with or without an accompanying narrative vignette, all
participants answered questions designed to measure stigma toward people with OUD (domain A),
the primary set of outcomes. Participants in the control group answered the questions with no
preceding exposure to either message frame. After completing questions for domain A, participants
randomized to 1 of the 4 Words Matter groups answered questions about the appropriateness of
various terms (domain B). Participants randomized to 1 of the 4 Medication Treatment Works groups
answered questions about OUD medication treatment (domain C). Participants randomized to the
control group answered questions in all 3 domains. For all participants, domain A questions appeared
first; for the control group, the order of domains B and C was randomized. The order of questions
within each domain was randomized for all groups to reduce risk of bias due to priming.
Domain A included items adapted from the General Social Survey stigma module35,36 and from
published scales measuring stigma.14,30 These items assessed preferences for social distance (such
as unwillingness to have a person with OUD marry into their family or be a neighbor),14,30,36,37
perspectives on the causes of OUD,14,30,36 support for increased governmental spending on OUD
programs, and level of warmth felt toward people with OUD.38 Dimensions of stigma were measured
using 5-point Likert scales, and levels of warmth were measured by a feeling thermometer (range,
0-100 points).
Domain B questions asked participants to rate on a 5-point Likert scale the extent to which they
agreed or disagreed that language was appropriate to use in clinical settings. The language included
5 stigmatizing terms (addict, substance abuse, dirty and clean [in reference to drug test results], and
addicted baby) and 5 alternative nonstigmatizing terms (person with substance use disorder,
substance use, negative and positive [in reference to drug test results], and baby born with neonatal
opioid withdrawal syndrome). We selected terms based on previous work suggesting that these
terms elicited different levels of stigma.16-18 We created a stigmatizing terms scale (Cronbach α = .74)
by summing the Likert scale responses to the stigmatizing terms and dividing by the total number of
items. We used the same procedure for the alternative terms to create a nonstigmatizing terms scale
(Cronbach α = .72). Participants in the Words Matter and control groups reported whether they
would sign a Words Matter pledge committing to the use of nonstigmatizing language.
Domain C began by asking participants whether a treatment for OUD existed that was effective
over a long period.39 This question was followed by a short explanation of the Food and Drug
Administration–approved medications for the treatment of OUD. Participants then responded to
adapted versions of several questions from domain A (substituting “a person taking medication to
treat OUD” for “a person with OUD”) and questions about perceived effectiveness of OUD
medication treatment26,30,31 (all survey questions are available in eTable 3 in Supplement 2).

JAMA Network Open. 2022;5(2):e2146971. doi:10.1001/jamanetworkopen.2021.46971 (Reprinted) February 4, 2022 5/14

Downloaded From: https://jamanetwork.com/ by Nguyen Tuan on 05/02/2022

 JAMA Network Open | Substance Use and Addiction Effect of Visual and Narrative Campaigns on Addiction Stigma in Health Care Professionals

Statistical Analysis
We used unpaired 2-sided t tests and χ2 tests to assess differences in observed characteristics across
study groups. We dichotomized Likert scale responses (eg, agree and strongly agree were coded as
1, and other responses were coded as 0) and estimated logistic and linear regression models (for the
feeling thermometer item and scales) to assess differences in responses between the exposure
groups and the control group. In the models, we controlled for participant age, sex, prescriber status
(physician, nurse practitioner, or physician assistant), and race and ethnicity (non-Hispanic White or
non-Hispanic other race). We calculated predicted margins and marginal effects to display results for
binary outcomes as predicted probabilities and changes in predicted probabilities. To correct for
multiple hypothesis testing, we generated sharpened false discovery rate q values and identified
estimates with q values less than .05 as statistically significant (with 2-sided P values reported to aid
interpretation of significance).40 To estimate the marginal effect of the narrative vignettes on
outcomes, we also assessed differences between groups exposed to both the visual campaign and a
narrative vignette vs those exposed to the visual campaign only. In a sensitivity analysis, we
estimated ordered logistic regression models using the nondichotomized Likert scale responses.
Data were analyzed using Stata software, version 14 (StataCorp LLC).

Results
Participants
Among 1842 participants, 1324 individuals (71.9%) were female and 518 (28.1%) were male. With
regard to race and ethnicity, 145 participants (7.9%) were Hispanic, 140 (7.6%) were non-Hispanic
Black, 1344 (73.0%) were non-Hispanic White, and 213 (11.6%) did not self-report race as Black or
White and did not self-report ethnicity as Hispanic. Most participants were nurses (505 individuals
[27.4%]) or physicians (467 individuals [25.4%]) (Table 1).
A total of 201 participants were randomized to the nonexposed control group, 206 to the Words
Matter visual campaign–only group, 203 to the Words Matter visual campaign plus patient vignette
group, 205 to the Words Matter visual campaign plus clinician vignette group, and 205 to the Words
Matter visual campaign plus administrator vignette group. A total of 205 participants were
randomized to the Medication Treatment Works visual campaign–only group, 205 to the Medication
Treatment Works visual campaign plus patient vignette group, 204 to the Medication Treatment
Works visual campaign plus clinician vignette group, and 208 to the Medication Treatment Works
visual campaign plus administrator vignette group (Figure). We observed no significant differences
in participant characteristics across study groups (eg, 133 participants [66.2%] in the control group,
585 participants [71.4%] in the Words Matter group, and 604 participants [73.5%] in the Medication
Treatment Works group were female; 64 participants [31.8%] in the control group, 270 participants
[33.0%] in the Words Matter group, and 249 participants [30.3%] in the Medication Treatment Works
group were prescribing clinicians) (eTable 4 in Supplement 2).

Attitudes Toward People With Opioid Use Disorder

In the control group, 142 participants (70.6%) reported unwillingness to have a person with OUD
marry into their family, and 90 participants (44.8%) reported unwillingness to have a person with
OUD as a neighbor (Table 2). Most participants in the control group disagreed that people with OUD
had only themselves to blame for the problem (118 individuals [58.7%]) and favored increased
spending on OUD treatment (131 individuals [65.2%]). The mean (SD) level of warmth toward people
with OUD was 50.7 (19.1) points on the feeling thermometer scale.
Exposure to the Words Matter and Medication Treatment Works visual campaigns alone was not
associated with statistically significant differences from the control group (eg, unwilling to have a
person with OUD marry into the family: difference, 0.3 percentage points [95% CI, −8.5 to 9.1;
P = .94] in the Words Matter group and −5.3 percentage points [95% CI, −14.3 to 3.7; P = .25] in the
Medication Treatment Works group). However, exposure to the combined Words Matter visual

JAMA Network Open. 2022;5(2):e2146971. doi:10.1001/jamanetworkopen.2021.46971 (Reprinted) February 4, 2022 6/14

Downloaded From: https://jamanetwork.com/ by Nguyen Tuan on 05/02/2022

 JAMA Network Open | Substance Use and Addiction Effect of Visual and Narrative Campaigns on Addiction Stigma in Health Care Professionals

campaign and patient vignette was associated with a lower probability of unwillingness to have a
person with OUD marry into the family (difference, −16.8 percentage points; 95% CI, −26.1 to −7.4;
P < .001) and a 7.2-point (95% CI, 3.2-11.1; P < .001) higher warmth rating than nonexposure.
Exposure to the combined Medication Treatment Works visual campaign and patient vignette was
associated with lower probabilities of unwillingness to have a person with OUD marry into the family
(difference, −14.5 percentage points; 95% CI, −23.8 to −5.3; P = .002) and to have a person with OUD
as a neighbor (difference, −15.3 percentage points; 95% CI, −24.6 to −6.0; P = .001) as well as a
6.4-point (95% CI, 2.3-10.5; P = .002) higher warmth rating than nonexposure.

Table 1. Participant Characteristics

Characteristic Participants, No. (%)

Total participants, No. 1842
Age, mean (SD), y 47 (13)
Educational level
High school or less 87 (4.7)
Some college 431 (23.4)
Bachelor's degree or higher 1324 (71.9)
Sex
Female 1324 (71.9)
Male 518 (28.1)
Race and ethnicity
Hispanic 145 (7.9)
Non-Hispanic Black 140 (7.6)
Non-Hispanic White 1344 (73.0)
Non-Hispanic other racea 213 (11.6)
Annual household income, $
<50 000 261 (14.2)
50 000-99 999 560 (30.4)
100 000-149 999 405 (22.0)
150 000-199 999 215 (11.7)
>200 000 401 (21.8)
Married or living with partner 1285 (69.8)
Household size
1 349 (18.9)
2 621 (33.7)
3 323 (17.5)
≥4 549 (29.8)
Owns home 1505 (81.7)
Lives in metropolitan area 1622 (88.1)
Region of US
South 639 (34.7)
Midwest 512 (27.8)
West 346 (18.8)
Northeast 345 (18.7)
Health care profession
Registered nurse or licensed practical nurse 505 (27.4)
Physician 467 (25.4)
Health aide or assistant 238 (12.9)
Health technician or technologist 219 (11.9)
Nurse practitioner or physician assistant 116 (6.3)
Therapist 106 (5.8)
Other practitioner 191 (10.4)
a
Includes individuals who did not self-report race as Black or White and did not
self-report ethnicity as Hispanic.

JAMA Network Open. 2022;5(2):e2146971. doi:10.1001/jamanetworkopen.2021.46971 (Reprinted) February 4, 2022 7/14

Downloaded From: https://jamanetwork.com/ by Nguyen Tuan on 05/02/2022

 Table 2. Effect of Exposure vs Nonexposure to Words Matter and Medication Treatment Works Message Frames on Stigma Toward People With Opioid Use Disordera

Percentage point differences in attitudes or differences in warmth between exposure group and control group (95% CI)
Words Matter message frame Medication Treatment Works message frame
Visual Visual Visual Visual Visual Visual
Visual campaign + campaign + campaign + Visual campaign + campaign + campaign +
campaign P patient P clinician P administrator P campaign P patient P clinician P administrator P
Survey item Control group only value vignette value vignette value vignette value only value vignette value vignette value vignette value
Attitudes toward
people with OUD
(percentage point
difference)
Unwilling to 142 (70.6) 0.3 (–8.5 .94 –16.8 (–26.1 <.001 –10.2 (–19.4 .03 –6.6 (–15.7 .15 –5.3 (–14.3 .25 –14.5 (–23.8 .002 –7.0 (–16.1 .13 –8.8 (–17.9 .06
have person to 9.1) to –7.4)b to –1.0) to 2.5) to 3.7) to –5.3)b to 2.1) to 0.30)
with OUD
marry into
family
Unwilling to 90 (44.8) –0.6 (–10.2 .91 –12.2 (–21.7 .01 –6.2 (–15.8 .20 –8.0 (–17.6 .10 –5.3 (–14.9 .28 –15.3 (–24.6 .001 –8.2 (–17.7 .09 –9.5 (–18.9 .047
have person to 9.1) to –2.8) to 3.4) to 1.5) to 4.3) to –6.0)b to 1.3) to –0.1)
with OUD as
neighbor
Agree OUD is a 108 (53.7) 6.2 (–3.2 .20 6.2 (–3.1 .19 5.6 (–3.9 .25 9.3 (0 .049 3.1 (–6.3 .52 5.9 (–3.5 .22 4.3 (–5.1 .37 7.0 (–2.4 .15
JAMA Network Open | Substance Use and Addiction

medical to 15.6) to 15.6) to 15.2) to 18.7) to 12.4) to 15.3) to 13.7) to 16.3)

condition
Disagree people 118 (58.7) –7.7 (17.3 .11 –1.7 (–11.2 .74 –5.7 (–15.4 .24 1.9 (–7.6 .70 –6.5 (–15.9 .18 –1.5 (–11.0 .76 –1.0 (–10.4 .85 –4.8 (–14.2 .33
with OUD are to to 1.8) to 7.9) to 3.9) to 11.1) to 3.0) to 8.1) to 8.6) to 4.7)

Downloaded From: https://jamanetwork.com/ by Nguyen Tuan on 05/02/2022

blame
Favor increased 131 (65.2) –2.9 (–12.2 .54 6.1 (–2.9 .18 –1.5 (–10.8 .75 2.0 (–7.1 .67 –2.7 (–11.9 .57 1.3 (–7.9 .79 3.7 (–5.3 .42 1.9 (–7.3 .68
spending on to 6.4) to 15.1) to 7.8) to 11.1) to 6.6) to 10.4) to 12.8) to 11.1)
OUD treatment
Warmth toward 50.7 (19.1) –0.3 (–4.4 .89 7.2 (3.2 <.001 1.2 (–3.2 .72 5.6 (1.6 .006 3.0 (–1.1 .16 6.4 (2.3 .002 1.8 (–2.1 .37 5.9 (2.0 .003
people with OUD, to 3.8) to 11.1)b to 4.7) to 9.6)b to 7.1) to 10.5)b to 5.7) to 9.7)b
mean (SD) (level
of difference)c

JAMA Network Open. 2022;5(2):e2146971. doi:10.1001/jamanetworkopen.2021.46971 (Reprinted)

b
Abbreviation: OUD, opioid use disorder. Estimate reached threshold of statistical significance as defined by a sharpened false discovery rate q value of
a <.05. Sharpened q values rather than P values were used to correct for multiple hypothesis testing.
Analysis included all 1842 participants. Logit regression models were used to estimate differences between
c
exposure groups and nonexposure control group. Percentage point differences and mean differences in warmth Score range, 0-100 points.
level were calculated using postestimation marginal effects. Model estimates were adjusted for age
(continuous), female sex, prescriber status (physician, nurse practitioner, or physician assistant), and race and
ethnicity (non-Hispanic White or non-Hispanic other race).

February 4, 2022
Effect of Visual and Narrative Campaigns on Addiction Stigma in Health Care Professionals

8/14
 JAMA Network Open | Substance Use and Addiction Effect of Visual and Narrative Campaigns on Addiction Stigma in Health Care Professionals

Exposure to the combined Words Matter visual campaign and administrator vignette was
associated with a 5.6-point (95% CI, 1.6-9.6; P = .006) higher warmth rating than nonexposure.
Exposure to the combined Medication Treatment Works visual campaign and administrator vignette
was associated with a 5.9-point (95% CI, 2.0-9.7; P = .003) higher warmth rating than nonexposure.
For both message frames, exposure to the combined visual campaign and clinician vignette was not
associated with significantly different levels of stigma compared with nonexposure.
Relative to exposure to the visual campaign alone, exposure to the combined visual campaign
and patient vignette was associated with lower levels of stigma in both the Words Matter and
Medication Treatment Works groups (eg, unwilling to have a person with OUD as a neighbor:
difference, –12.1 percentage points [95% CI, –21.5 to –2.8; P = .01] in the Words Matter group and –9.9
percentage points [95% CI, –19.0 to –0.7; P = .04] in the Medication Treatment Works group)
(eTable 5 in Supplement 2).

Stigmatizing Language and Alternative Terms

In the control group, 93 participants (46.3%) agreed that addict was an appropriate term to use in a
clinical setting, and 163 participants (81.1%) agreed that substance abuse was an appropriate term
(eTable 6 in Supplement 2). A total of 148 participants in the control group (73.6%) were willing to
sign a pledge to use nonstigmatizing language.
Compared with nonexposure, exposure to the Words Matter visual campaign combined with
the patient vignette was associated with lower probabilities of endorsing addict (difference, −23.1
percentage points; 95% CI, −32.1 to −14.2; P < .001), substance abuse (difference, −23.3 percentage
points; 95% CI, −31.9 to −15.0; P < .001), dirty (difference, −9.0 percentage points; 95% CI, −15.7 to
−2.3; P = .009), and clean (difference, −22.9 percentage points; 95% CI, −33.2 to −13.5; P < .001) as
acceptable terms as well as a 0.5-point lower rating (95% CI, −0.7 to −0.4; P < .001) on the 5-point
scale measuring perceived acceptability of stigmatizing terms. We observed similar reductions
among those exposed to the visual campaign combined with the clinician vignette (eg, endorsement
of addict as an appropriate term: difference, −16.4 percentage points; 95% CI, −25.6 to −7.1; P = .001)
or the administrator vignette (eg, endorsement of addict as an appropriate term: difference, −20.4
percentage points; 95% CI, −29.5 to −11.3; P < .001) compared with the control group. None of the
exposures was associated with a greater willingness to sign a pledge to use nonstigmatizing language
(visual campaign only: difference, −3.0 percentage points [95% CI, −15.4 to 9.4; P = .63]; visual
campaign plus patient vignette: difference, 0.4 percentage points [95% CI, −7.9 to 8.8; P = .92];
visual campaign plus clinician vignette: difference, −1.6 percentage points [95% CI, −23.0 to 2.0;
P = .10]; visual campaign plus administrator vignette: difference, 3.4 percentage points [95% CI, −4.7
to 11.6; P = .41]) compared with the control group. The Words Matter visual campaign and vignette
combinations were associated with greater recognition of stigmatizing language than the visual
campaign alone (eg, endorsement of addict as an appropriate term: difference, −18.5 percentage
points [95% CI, −27.3 to −9.7; P < .001) for visual campaign plus patient vignette, −11.8 percentage
points [95% CI, −20.9 to −2.6; P = .01] for visual campaign plus clinician vignette, and −15.7
percentage points [95% CI, −24.7 to – 6.8; P < .001] for visual campaign plus administrator vignette)
(eTable 7 in Supplement 2).

Attitudes About Medications for Opioid Use Disorder

The Medication Treatment Works visual campaign combined with the patient vignette was associated
with a 4.6-point (95% CI, 0.3-9.0; P = .04) higher warmth rating toward people receiving medication
to treat OUD, with a similarly higher warmth rating (difference, 4.4 points; 95% CI, 0.2-8.6; P = .04)
among those exposed to the combined visual campaign and administrator vignette (eTable 8 in
Supplement 2). However, exposure to the Medication Treatment Works message frames was not
associated with any other significant differences in attitudes related to medication treatment
compared with nonexposure (eg, unwillingness to have a person receiving OUD medication
treatment as a neighbor: difference, −5.3 percentage points; 95% CI, −12.8 to 2.3; P = .17). Sensitivity

JAMA Network Open. 2022;5(2):e2146971. doi:10.1001/jamanetworkopen.2021.46971 (Reprinted) February 4, 2022 9/14

Downloaded From: https://jamanetwork.com/ by Nguyen Tuan on 05/02/2022

 JAMA Network Open | Substance Use and Addiction Effect of Visual and Narrative Campaigns on Addiction Stigma in Health Care Professionals

analyses estimating ordered logistic regression models with nondichotomized Likert scale response
options produced qualitatively similar results (eTable 9 to eTable 11 in Supplement 2).

Discussion
This randomized clinical trial assessed the effects of 2 message frames delivered through visual
campaigns and written narrative vignettes from the perspectives of a patient with OUD, a clinician,
or a health care system administrator among a large national sample of health care professionals. Our
findings suggested that visual campaigns alone may have limited impact on stigma toward people
with OUD. A potentially more effective communication strategy may combine these visual
campaigns with narrative vignettes featuring the voices of people with OUD.
Findings were consistent with other research suggesting that messages emphasizing
nonstigmatizing language16-18 and highlighting effective treatment23,25 through sympathetic
stories23,25,32 can reduce stigma among the public and that narrative-based messaging can be an
effective means of communicating important information to clinicians.41 Both the Words Matter and
Medication Treatment Works message frames were associated with lower levels of stigma when they
incorporated vignettes, particularly vignettes featuring narratives from the perspective of a patient
with OUD. Exposure to narrative vignettes from all messengers (patients, clinicians, and
administrators) was associated with greater recognition of stigmatizing language. Health care
systems seeking to reduce stigma may consider highlighting these voices in their communication
efforts. Hiring and integrating people with lived experience into the health care professional
workforce is a contact-based approach that may also be an effective stigma-reduction strategy.
Several high-profile communication campaigns have used images and stories of real individuals
with OUD who received treatment with medication.42,43 Notably, our Medication Treatment Works
vignettes were not associated with lower levels of medication-related stigma. Other strategies are
likely needed to reduce stigma toward these highly effective medications, including structural and
policy reforms to improve access.44 Several health care systems have implemented campaigns to
discourage use of stigmatizing language and encouraged staff to sign language pledges.21,22 Although
the results of the present study suggest that these pledges have the potential to improve recognition
of stigmatizing terms that are inappropriate for use in clinical settings, we did not find that the Words
Matter message frame was associated with greater willingness to sign a language pledge.
Communication campaigns likely need to be implemented together with other interventions to
change behavior rather than knowledge and attitudes only.
Stigma toward people with OUD and other substance use disorders has been constraining our
ability to resolve the persistent drug overdose problem.5 Clinical settings are meant to be welcoming
and safe environments in which people with substance use disorders can seek care and connect to
services. Given that stigma among health care professionals can have serious consequences for
patients with OUD, stigma-reduction communication campaigns are needed to target this
population. Encouraging changes in language and improving understanding of the effectiveness of
medication treatment may reduce stigma toward people with OUD.

Limitations
This study has several limitations. First, although samples were recruited from 2 national online
survey panels, those panels are not probability-based samples of health care professionals, which
may affect generalizability. Sample characteristics indicate geographically diverse representation
from a variety of health care professions. Although 71.9% of participants were female, this
percentage is consistent with the composition of the broader health care professional workforce.45
Second, we did not use vignettes from real individuals, although message design was informed by
input from health care professionals and people with lived experience. We constructed the
exposures to minimize confounding and enhance our ability to isolate the impact of the most
important elements of the messages and messengers. When designing campaigns for health care

JAMA Network Open. 2022;5(2):e2146971. doi:10.1001/jamanetworkopen.2021.46971 (Reprinted) February 4, 2022 10/14

Downloaded From: https://jamanetwork.com/ by Nguyen Tuan on 05/02/2022

 JAMA Network Open | Substance Use and Addiction Effect of Visual and Narrative Campaigns on Addiction Stigma in Health Care Professionals

systems, the inclusion of real people in communication materials could increase the authenticity of
the campaign; whether this inclusion changes the effects of these messages is unknown. Third,
participants answered questions directly after viewing the visual campaign and/or vignettes a single
time. The durability of effects and the impact of repeat exposure (during an extended campaign) are
uncertain. Fourth, we conducted the study during the SARS-CoV-2 pandemic, which strained the
health care workforce, potentially affecting responses. Fifth, we could not translate estimated
changes in attitudes into measures of clinical impact; future work is warranted to explore these
connections.

Conclusions
In this randomized clinical trial, messages about nonstigmatizing language and effective medication
for the treatment of OUD were associated with reductions in stigma toward people with OUD among
health care professionals. Stigma-reduction efforts targeting health care professionals may improve
health care system capacity to serve people with OUD.

ARTICLE INFORMATION
Accepted for Publication: November 25, 2021.
Published: February 4, 2022. doi:10.1001/jamanetworkopen.2021.46971
Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2022
Kennedy-Hendricks A et al. JAMA Network Open.
Corresponding Author: Alene Kennedy-Hendricks, PhD, Department of Health Policy and Management,
Johns Hopkins Bloomberg School of Public Health, Hampton House 408, 624 N Broadway, Baltimore, MD 21205
(alene@jhu.edu).
Author Affiliations: Department of Health Policy and Management, Johns Hopkins Bloomberg School of Public
Health, Baltimore, Maryland (Kennedy-Hendricks, McGinty, Barry); Johns Hopkins Center for Mental Health and
Addiction Policy, Baltimore, Maryland (Kennedy-Hendricks, McGinty, Barry); Department of Mental Health, Johns
Hopkins Bloomberg School of Public Health, Baltimore, Maryland (Kennedy-Hendricks, McGinty, Fingerhood,
Barry); Johns Hopkins Center for Communication Programs, Baltimore, Maryland (Summers, Krenn); Division of
Addiction Medicine, Johns Hopkins School of Medicine, Baltimore, Maryland (Fingerhood).
Author Contributions: Dr Kennedy-Hendricks had full access to all of the data in the study and takes responsibility
for the integrity of the data and the accuracy of the data analysis.
Concept and design: All authors.
Acquisition, analysis, or interpretation of data: McGinty, Barry.
Drafting of the manuscript: Kennedy-Hendricks, McGinty, Fingerhood.
Critical revision of the manuscript for important intellectual content: Kennedy-Hendricks, McGinty, Summers,
Krenn, Barry.
Statistical analysis: Kennedy-Hendricks.
Obtained funding: McGinty, Barry.
Administrative, technical, or material support: McGinty, Summers, Krenn, Fingerhood, Barry.
Supervision: McGinty, Barry.
Conflict of Interest Disclosures: Dr Kennedy-Hendricks reported receiving grants from Johns Hopkins Bloomberg
School of Public Health during the conduct of the study and grants from Arnold Ventures, the National Institute
on Alcohol Abuse and Alcoholism, and the National Institute on Drug Abuse outside the submitted work. Dr
McGinty reported receiving grants from Arnold Ventures, the Centers for Disease Control and Prevention, the
National Institute of Mental Health, and the National Institute on Drug Abuse outside the submitted work. No
other disclosures were reported.
Funding/Support: This study was supported by funding from the the Bloomberg American Health Initiative at the
Johns Hopkins Bloomberg School of Public Health.

JAMA Network Open. 2022;5(2):e2146971. doi:10.1001/jamanetworkopen.2021.46971 (Reprinted) February 4, 2022 11/14

Downloaded From: https://jamanetwork.com/ by Nguyen Tuan on 05/02/2022

 JAMA Network Open | Substance Use and Addiction Effect of Visual and Narrative Campaigns on Addiction Stigma in Health Care Professionals

Role of the Funder/Sponsor: The funding institution had no impact on the design and conduct of the study;
collection, management, analysis, and interpretation of the data; preparation, review, or approval of the
manuscript; or decision to submit the manuscript for publication.
Data Sharing Statement: See Supplement 3.

REFERENCES
1. Friedman J, Beletsky L, Schriger DL. Overdose-related cardiac arrests observed by emergency medical services
during the US COVID-19 epidemic. JAMA Psychiatry. 2021;78(5):562-564. doi:10.1001/jamapsychiatry.2020.4218
2. Ahmad FB, Rossen LM, Sutton P. Provisional drug overdose death counts. National Center for Health Statistics;
2021. Accessed February 7, 2021. https://www.cdc.gov/nchs/nvss/vsrr/drug-overdose-data.htm
3. Volkow ND. Stigma and the toll of addiction. N Engl J Med. 2020;382(14):1289-1290. doi:10.1056/
NEJMp1917360
4. McGinty EE, Barry CL. Stigma reduction to combat the addiction crisis—developing an evidence base. N Engl J
Med. 2020;382(14):1291-1292. doi:10.1056/NEJMp2000227
5. Tsai AC, Kiang MV, Barnett ML, et al. Stigma as a fundamental hindrance to the United States opioid overdose
crisis response. PLoS Med. 2019;16(11):e1002969. doi:10.1371/journal.pmed.1002969
6. Link BG, Phelan JC. Conceptualizing stigma. Annu Rev Sociol. 2001;27:363-385. doi:10.1146/annurev.
soc.27.1.363
7. Morone JA. Enemies of the people: the moral dimension to public health. J Health Polit Policy Law. 1997;22(4):
993-1020. doi:10.1215/03616878-22-4-993
8. Kim JW, Morgan E, Nyhan B. Treatment versus punishment: understanding racial inequalities in drug policy.
J Health Polit Policy Law. 2020;45(2):177-209. doi:10.1215/03616878-8004850
9. van Boekel LC, Brouwers EPM, van Weeghel J, Garretsen HFL. Stigma among health professionals towards
patients with substance use disorders and its consequences for healthcare delivery: systematic review. Drug
Alcohol Depend. 2013;131(1-2):23-35. doi:10.1016/j.drugalcdep.2013.02.018
10. Kulesza M, Larimer ME, Rao D. Substance use related stigma: what we know and the way forward. J Addict
Behav Ther Rehabil. 2013;2(2):782. doi:10.4172/2324-9005.1000106
11. Simon R, Snow R, Wakeman S. Understanding why patients with substance use disorders leave the hospital
against medical advice: a qualitative study. Subst Abus. 2020;41(4):519-525. doi:10.1080/08897077.2019.
1671942
12. Mendiola CK, Galetto G, Fingerhood M. An exploration of emergency physicians’ attitudes toward patients
with substance use disorder. J Addict Med. 2018;12(2):132-135. doi:10.1097/ADM.0000000000000377
13. Goddu AP, O’Conor KJ, Lanzkron S, et al. Do words matter? stigmatizing language and the transmission of bias
in the medical record. J Gen Intern Med. 2018;33(5):685-691. doi:10.1007/s11606-017-4289-2
14. Kennedy-Hendricks A, Barry CL, Gollust SE, Ensminger ME, Chisolm MS, McGinty EE. Social stigma toward
persons with prescription opioid use disorder: associations with public support for punitive and public health–
oriented policies. Psychiatr Serv. 2017;68(5):462-469. doi:10.1176/appi.ps.201600056
15. McGinty E, Pescosolido B, Kennedy-Hendricks A, Barry CL. Communication strategies to counter stigma and
improve mental illness and substance use disorder policy. Psychiatr Serv. 2018;69(2):136-146. doi:10.1176/appi.ps.
201700076
16. Kelly JF, Westerhoff CM. Does it matter how we refer to individuals with substance-related conditions? a
randomized study of two commonly used terms. Int J Drug Policy. 2010;21(3):202-207. doi:10.1016/j.drugpo.
2009.10.010
17. Ashford RD, Brown AM, Curtis B. Substance use, recovery, and linguistics: the impact of word choice on explicit
and implicit bias. Drug Alcohol Depend. 2018;189:131-138. doi:10.1016/j.drugalcdep.2018.05.005
18. Ashford RD, Brown AM, Curtis B. The language of substance use and recovery: novel use of the go/no-go
association task to measure implicit bias. Health Commun. 2019;34(11):1296-1302. doi:10.1080/10410236.2018.
1481709
19. McGinty EE, Stone EM, Kennedy-Hendricks A, Barry CL. Stigmatizing language in news media coverage of the
opioid epidemic: implications for public health. Prev Med. 2019;124:110-114. doi:10.1016/j.ypmed.2019.03.018
20. Kelly JF, Wakeman SE, Saitz R. Stop talking ‘dirty’: clinicians, language, and quality of care for the leading cause
of preventable death in the United States. Am J Med. 2015;128(1):8-9. doi:10.1016/j.amjmed.2014.07.043
21. Dell Medical School. Reducing stigma education tools (ReSET). Dell Medical School, The University of Texas at
Austin. July 21, 2020. Accessed March 25, 2021. https://vbhc.dellmed.utexas.edu/courses/course-v1:ut+cn01+2020-
21/about

JAMA Network Open. 2022;5(2):e2146971. doi:10.1001/jamanetworkopen.2021.46971 (Reprinted) February 4, 2022 12/14

Downloaded From: https://jamanetwork.com/ by Nguyen Tuan on 05/02/2022

 JAMA Network Open | Substance Use and Addiction Effect of Visual and Narrative Campaigns on Addiction Stigma in Health Care Professionals

22. Grayken Center for Addiction. Reducing stigma: why words about addiction matter. Boston Medical Center;
2021. Accessed March 25, 2021. https://www.bmc.org/addiction/reducing-stigma
23. Kennedy-Hendricks A, McGinty EE, Barry CL. Effects of competing narratives on public perceptions of opioid
pain reliever addiction during pregnancy. J Health Polit Policy Law. 2016;41(5):873-916. doi:10.1215/03616878-
3632230
24. Barry CL, McGinty EE, Pescosolido BA, Goldman HH. Stigma, discrimination, treatment effectiveness, and
policy: public views about drug addiction and mental illness. Psychiatr Serv. 2014;65(10):1269-1272. doi:10.1176/
appi.ps.201400140
25. McGinty EE, Goldman HH, Pescosolido B, Barry CL. Portraying mental illness and drug addiction as treatable
health conditions: effects of a randomized experiment on stigma and discrimination. Soc Sci Med. 2015;126:73-85.
doi:10.1016/j.socscimed.2014.12.010
26. McGinty EE, Stone EM, Kennedy-Hendricks A, Bachhuber MA, Barry CL. Medication for opioid use disorder:
a national survey of primary care physicians. Ann Intern Med. 2020;173(2):160-162. doi:10.7326/M19-3975
27. Mancher M, Leshner AI, eds. National Academies of Sciences, Engineering, and Medicine; Health and Medicine
Division; Board on Health Sciences Policy; Committee on Medication-Assisted Treatment for Opioid Use Disorder.
Medications for Opioid Use Disorder Save Lives. National Academies Press; 2019.
28. Olsen Y, Sharfstein JM. Confronting the stigma of opioid use disorder—and its treatment. JAMA. 2014;311(14):
1393-1394. doi:10.1001/jama.2014.2147
29. Wakeman SE, Barnett ML. Primary care and the opioid-overdose crisis—buprenorphine myths and realities.
N Engl J Med. 2018;379(1):1-4. doi:10.1056/NEJMp1802741
30. Stone EM, Kennedy-Hendricks A, Barry CL, Bachhuber MA, McGinty EE. The role of stigma in U.S. primary care
physicians’ treatment of opioid use disorder. Drug Alcohol Depend. 2021;221:108627. doi:10.1016/j.drugalcdep.
2021.108627
31. Kennedy-Hendricks A, Barry CL, Stone E, Bachhuber MA, McGinty EE. Comparing perspectives on medication
treatment for opioid use disorder between national samples of primary care trainee physicians and attending
physicians. Drug Alcohol Depend. 2020;216:108217. doi:10.1016/j.drugalcdep.2020.108217
32. Heley K, Kennedy-Hendricks A, Niederdeppe J, Barry C. Reducing health-related stigma through
communication approaches. Health Commun. 2019:1-12.
33. Ipsos Group. KnowledgePanel: a methodological overview. Ipsos Group. Accessed November 1, 2020. https://
www.ipsos.com/sites/default/files/ipsosknowledgepanelmethodology.pdf
34. SurveyHealthcareGlobus. Panel counts. SurveyHealthcareGlobus; 2020. Accessed November 1, 2020.
https://www.surveyhealthcareglobus.com/our-panel/united-states/
35. Pescosolido BA, Martin JK, Long JS, Medina TR, Phelan JC, Link BG. “A disease like any other”? a decade of
change in public reactions to schizophrenia, depression, and alcohol dependence. Am J Psychiatry. 2010;167(11):
1321-1330. doi:10.1176/appi.ajp.2010.09121743
36. Link BG, Phelan JC, Bresnahan M, Stueve A, Pescosolido BA. Public conceptions of mental illness: labels,
causes, dangerousness, and social distance. Am J Public Health. 1999;89(9):1328-1333. doi:10.2105/AJPH.89.
9.1328
37. Corrigan PW, Green A, Lundin R, Kubiak MA, Penn DL. Familiarity with and social distance from people who
have serious mental illness. Psychiatr Serv. 2001;52(7):953-958. doi:10.1176/appi.ps.52.7.953
38. Nelson SC. Feeling thermometer. In: Lavrakas PJ, ed. Encyclopedia of Survey Research Methods. Vol 1. Sage
Publications; 2008:276.
39. Blendon RJ, Benson JM. The public and the opioid-abuse epidemic. N Engl J Med. 2018;378(5):407-411. doi:
10.1056/NEJMp1714529
40. Anderson ML. Multiple inference and gender differences in the effects of early intervention: a reevaluation of
the Abecedarian, Perry Preschool, and Early Training projects. J Am Stat Assoc. 2008;103(484):1481-1495. doi:10.
1198/016214508000000841
41. Meisel ZF, Metlay JP, Sinnenberg L, et al. A randomized trial testing the effect of narrative vignettes versus
guideline summaries on provider response to a professional organization clinical policy for safe opioid prescribing.
Ann Emerg Med. 2016;68(6):719-728. doi:10.1016/j.annemergmed.2016.03.007
42. Department of Behavioral Health and Intellectual Disability Services. Bupe works. City of Philadelphia; 2021.
Accessed March 25, 2021. https://dbhids.org/addiction-services/bupe-works
43. NYC Health. Living proof campaign. Government of New York City; 2019. Accessed March 25, 2021. https://
www1.nyc.gov/assets/doh/downloads/pdf/basas/opioid-living-proof-campaign.pdf

JAMA Network Open. 2022;5(2):e2146971. doi:10.1001/jamanetworkopen.2021.46971 (Reprinted) February 4, 2022 13/14

Downloaded From: https://jamanetwork.com/ by Nguyen Tuan on 05/02/2022

 JAMA Network Open | Substance Use and Addiction Effect of Visual and Narrative Campaigns on Addiction Stigma in Health Care Professionals

44. Allen B, Nolan ML, Paone D. Underutilization of medications to treat opioid use disorder: what role does
stigma play? Subst Abus. 2019;40(4):459-465. doi:10.1080/08897077.2019.1640833
45. Day JC, Christnacht C. Women hold 76% of all health care jobs, gaining in higher-paying occupations. United
States Census Bureau. August 14, 2019. Accessed December 11, 2021. https://www.census.gov/library/stories/2019/
08/your-health-care-in-womens-hands.html

SUPPLEMENT 1.
Trial Protocol

SUPPLEMENT 2.
eTable 1. Structure of Randomized Clinical Trial
eTable 2. Narrative Vignettes Communicating Words Matter and Medication Treatment Works Message Frames
eTable 3. Questions Included in Each Domain
eTable 4. Comparison of Characteristics of Survey Participants Across Study Arms
eTable 5. Effects of Exposure to the Words Matter and Medication Treatment Works Narrative Vignette and Visual
Campaign Combination on Stigma Toward People With Opioid Use Disorder Relative to the Visual Campaign–
Only Groups
eTable 6. Effects of Exposure to the “Words Matter” Visual Campaign and Vignette Frames on Perceptions of
Clinically Appropriate Language Related to Substance Use Relative to the No-Exposure Control Group
eTable 7. Effects of Exposure to the Words Matter Narrative Vignette and Visual Campaign Combination on
Perceptions of Clinically Appropriate Language Related to Substance Use Relative to the Visual Campaign–Only
Group
eTable 8. Effects of Exposure to the “Medication Treatment Works” Visual Campaign and Vignettes on Stigma
toward People with Opioid Use Disorder Receiving Medication Treatment Relative to the No-Exposure Control
Group
eTable 9. Effects of Exposure to the Words Matter and Medication Treatment Works Message Frames on Stigma
Toward People With Opioid Use Disorder Relative to the Nonexposed Control Group Estimated With Ordinal
Logistic Regression Models
eTable 10. Effects of Exposure to the Words Matter Narrative Vignette and Visual Campaign Combination on
Perceptions of Clinically Appropriate Language Related to Substance Use Relative to the Visual Campaign–Only
Group Estimated With Ordinal Logistic Regression Models
eTable 11. Effects of Exposure to the Medication Treatment Works Visual Campaign and Narrative Vignettes on
Stigma Toward People With Opioid Use Disorder Receiving Medication Treatment Relative to the Nonexposed
Control Group Estimated With Ordinal Logistic Regression Models
eFigure. Words Matter and Medication Treatment Works Visual Campaigns

SUPPLEMENT 3.
Data Sharing Statement

JAMA Network Open. 2022;5(2):e2146971. doi:10.1001/jamanetworkopen.2021.46971 (Reprinted) February 4, 2022 14/14

Downloaded From: https://jamanetwork.com/ by Nguyen Tuan on 05/02/2022