“I would not go to him”: Focus groups exploring community responses to a public health campaign aimed at reducing unnecessary diagnostic imaging of low back pain

Abstract Background Community awareness of the harms of overdiagnosis remains low. Objective To evaluate community responses to a public health campaign designed for health service waiting rooms that focuses on the harms of unnecessary diagnostic imaging for low back pain. Methods We conducted two focus groups of 19 community members with or without low back pain in Sydney, Australia. This study formed the fourth and final stage of the development process of a public health campaign: (a) initial design, (b) expert review and revision, (c) online experiment and (d) community views & revision. We evaluated reactions to components of the campaign that included digital posters and an information leaflet using strong imagery and messaging about the risk of overdiagnosis. We conducted a qualitative thematic analysis to identify main themes. Results Community members reacted with surprise, initial mistrust, and occasionally anger towards imagery and messaging that suggested diagnostic imaging tests could be unnecessary and harmful. With further reflection and discussion, and after reading longer format information about overdiagnosis, the participants found some of the messages informative and useful. Participants appeared to gain a better understanding of the concept of overdiagnosis and the importance of not rushing to imaging. Conclusions Public health campaigns including posters and leaflets displayed in waiting rooms could raise awareness about overuse of diagnostic imaging and the harms of overdiagnosis more broadly. However, negative reactions are possible and must be managed carefully. Patient or Public Contribution We involved a community participation manager who provided advice on the focus group discussion guide, participant recruitment and manuscript presentation.


| INTRODUC TI ON
Many patients with low back pain receive diagnostic imaging despite guidelines recommending against it. One systematic review of over 19 million low back pain consultations found that one in four patients in primary care and one in three patients in emergency departments received imaging. 1 This is problematic because for the majority (90%-95%) of patients without serious spinal pathology, lumbar imaging will not improve outcomes and can cause unnecessary harms including overdiagnosis and overtreatment. 2 Trials have found that lumbar imaging findings can increase worry 3 and more than double the likelihood of having surgery. 4 There are several reasons why unnecessary diagnostic imaging might happen. These include a clinician's desire to maintain the therapeutic relationship, address a patient's fear or anxiety, and manage their limited consultation time. 5 Beliefs about imaging appear to be particularly important. Our systematic review of 69 qualitative studies found that clinicians and patients believed imaging could locate the source of pain, legitimize the pain experience and reduce risk of litigation. 6 Although clinicians were aware of several harms of unnecessary imaging including increased anxiety and overtreatment, patients were not. Interestingly, no qualitative study had explored the beliefs of general community members. While a survey study found community members had misconceptions about imaging for low back pain-more than half expected to have an imaging test during their consultation-7 such survey studies do not provide indepth understanding of people's beliefs and perspectives. A better understanding of why community members might view imaging as necessary, and how they would react to messaging that discourages the test, could help inform population-wide approaches to reduce overuse.
One simple, scalable and under-investigated approach to reducing medical overuse is a public health campaign displayed in health service waiting rooms. A systematic review 8 of observational and intervention studies found that interventions using waiting room screens (eg TV, tablet, computer) can enhance knowledge about cancer screening, influenza vaccination and contraception in women.
However, effects on health behaviour were unclear. Some studies using waiting room screens found increase in healthy behaviours such as uptake of tetanus booster vaccination and polio vaccination, but they used surrogate endpoints. 9,10 But we still do not know if such interventions reduce medical overuse. To our knowledge, these approaches have not been evaluated in the context of medical overuse.
For the community to engage with public health campaigns, the content should be acceptable and understandable. Nearly 60% of Australian adults have low health literacy, meaning they may have difficulty with understanding and applying the information required for healthcare decision-making. 11, 12 A Cochrane review suggested that interventions with consumer input in the development phase had better uptake and engagement than those without. 13 The understandability of public health campaigns could also influence their effectiveness. Consumers can find health promotion messages difficult to understand. 14 In one study, two-thirds of patients with cancer had difficulty understanding the information provided to them, leading many of them to seek out alternate information sources that they found more accessible. 15 We aimed to evaluate community responses to a public health campaign designed for health service waiting rooms. The campaign combines strong imagery on large LCD screens in the waiting area, with messages such as 'back scans can't heal, they can harm', and written information about -harms of diagnostic imaging. The campaign materials were developed by a creative innovation agency and refined by researchers in the Wiser Healthcare Research Collaboration, 16 and aimed to change behaviour by raising awareness of the harms of imaging. This focus group study is part of a larger body of work that has included pilot testing, qualitative studies with patients with low back pain, 17 Emergency Department doctors, 18 and GPs, 17 and a randomized trial in community members. 19 Our goal was to develop a public health campaign that was acceptable and understandable to community members and that had potential to reduce overuse of diagnostic imaging for low back pain.

| ME THODS
Development of the public health campaign has been a multi-stage process. Campaign messages were determined by back pain researchers (AT, CM, IH) who pitched the idea of increasing public awareness about harms of unnecessary imaging for low back pain to a creative innovation agency. The agency then produced draft versions of the materials-5 posters and a health promotion leaflet. The overarching model of behaviour change was based on behavioural economics, that is, considering the effects of cognitive biases on decision-making processes. The behavioural cues used in the campaign materials included: framing, loss aversion, anchoring, chunking, status quo bias and suggested alternatives (Box 1).
Draft versions of the campaign materials were revised based on feedback from experts in back pain care including clinicians and researchers. We tested the revised version of the leaflet in an online randomized controlled trial that found the leaflet changed community members intention to request an imaging test. 19 The current study was a final step in the larger project, to ensure that these materials are acceptable to members of the community. More details on the steps in the larger campaign project are provided in Table 1 and Figure 1.

| Design
We conducted a qualitative study using two focus groups of community members. We chose the focus group format to allow opportunities for community members to ask questions and clarify sources of confusion. We wanted to provide the opportunity for participants to hear each other's views and exchange ideas. Another advantage of focus groups is that they shift control of conversation topics from researchers to participants and allow participants to express views, in their own words, on topics not previously anticipated by the researcher.

| Ethics
All study procedures were approved by the South Western Sydney

| Participant selection and recruitment
We recruited a sample of community members who were over the age of 18 to participate. Community members with or without low back pain were eligible to participate. Those who were not fluent in English were excluded. We aimed for a minimum 2, 8-person focus groups, based on research suggesting that the majority (two-thirds) of themes in focus groups studies are often generated from the first group, and more than 80% of the themes can be generated after 2-3 focus groups. 21 We conducted the focus groups in a socioeconomically and cul- The population selected via the Network is representative of the intended audience of the campaign, that is, community members with or without low back pain, from a diverse area of Sydney.

| Focus groups and data collection
We conducted focus groups on June 24 (n = 9) and June 28 (n = 10) 2019 at Liverpool Hospital. The focus groups were audio recorded and a researcher (SS) and a research assistant took field notes. Each 90-minute focus group session was divided into three parts. Part 1 explored participants' understanding of diagnostic imaging for low back pain. Parts 2 and 3 evaluated reactions to components of the public health campaign (5 digital posters; then one leaflet in hard copy and displayed as a Powerpoint slideshow). Participants were shown the posters for discussion (part 2) before reading the health promotion leaflet (part 3). The groups were facilitated by one of the authors (AT, a researcher with physiotherapy qualifications). Basic demographic information was collected with participants with their consent.
We assessed participants' health literacy using a single item lit- This literacy screener has been found to perform moderately well in identifying adults with limited reading ability (sensitivity = 54% (95% CI = 47% to 61%) and specificity = 83% (95% CI = 81% to 86)). 23,24 Box 1 Behavioural cues included in a health promotion leaflet to reduce unnecessary diagnostic imaging of low back pain

| 'Scan your options' posters
Five different digital posters about overdiagnosis and potential harms of unnecessary imaging (Appendix S2) were shown to participants.
Each poster incorporated the slogan 'Scan your options, not your back'.
The posters were designed to be displayed on a 55-inch digital screen in the waiting room of a hospital emergency department. The prominent messages and images on the posters were as follows: • Poster 1: 'Back scans can lead to dangerous and unnecessary treatments'. Image: Surgical scissors.
• Poster 2: 'Over 2/3 of people who have a back scan will get a false alarm'. Image: X-ray film of a spine.
• Poster 3: 'Back scans can't heal-they can harm'. Image: X-ray film of a spine.

| Health promotion leaflet
The 2-page health promotion leaflet had 6 panels of information (Appendix S2). Each page contained cues to emphasize the potential harms of unnecessary imaging for low back pain. The leaflet had a reading Grade Level of 7, that is, fairly easy to read; suitable for ages 11-13 years and above. We pilot tested the leaflet with eight consumers and clinicians before recruiting participants. Stage 1 This stage involved the design of a draft health promotion strategy in collaboration with a creative innovation agency. Researchers pitched the idea of raising awareness of the harms of unnecessary imaging for low back pain. A key goal of the strategy was to encourage community members to ask questions of their doctor and engage in the decision-making process. The agency produced draft versions of 5 digital posters and one patient leaflet.

Stage 2
This stage involved revision of the health promotion strategy materials based on expert feedback. The experts included an orthopaedic surgeon, emergency physicians, physiotherapists and back pain researchers. During Stage 2 experts decided that the agency's initial preference to focus on the harms of radiation to deter patients from imaging could not be supported well enough by data (Appendix S1). Following this feedback, the agency shifted focus to the harms of overdiagnosis, including incidental findings that cause worry and increase the risk of unnecessary surgery. Revised versions of the 5 digital posters and patient leaflet were provided to researchers for further testing.

Stage 3
It involved conducting an online randomized trial of the patient leaflet with a sample of 418 community members in Australia. 19 Stage 3 testing revealed that the leaflet with messages about overdiagnosis could reduce intention to request imaging for low back pain compared with a control leaflet with neutral information about imaging. The patient leaflet also reduced the belief that everyone with low back pain should have imaging.

Stage 4
This stage is the focus of this study and was a qualitative study with 19 community members which aimed to gather their views of Stage 3 versions of the intervention materials which included five posters and a leaflet (Appendix S2).

Stage 5
This stage will be rigorous evaluation in large-scale randomized trial in a real-life clinical practice setting.

TA B L E 1 Overview of the public health campaign development and evaluation
Consumers and clinicians gave feedback on overall aspects of the leaflet such as the utility, readability and content. We made changes to the leaflet based on their feedback. We tested this revised version of the leaflet in a randomized controlled trial in 418 members of the public and made further changes to wording to improve understandability. 18

| Coding and analysis of qualitative data
We conducted thematic analysis of the transcripts to identify main themes. Four members (SS, AT, KP, JKH) of the research team independently documented salient observations from the transcripts and field notes. The initial impressions of the data, combined with the discussion guide (Appendix S3), formed the basis of the coding framework. The authors (SS and AT) met to discuss the findings and arrive at a final set of themes, which were then reviewed by all authors.

| RE SULTS
Of the 19 participants, 12 were female, 14 were born in Australia, 4 had university education, 15 had experienced low back pain in the past 12 months, 11 had undergone an imaging test for back pain in the past ( to be more open to the campaign messaging and spent more time discussing the potential of the campaign to help start conversations between patients and healthcare professionals.
Below we present our key findings with supporting quotes.

Concern about discouraging necessary imaging
There was some concern expressed about the potential for the posters to discourage people from necessary imaging. For example, participants worried that patients might decline imaging after an accident or trauma. Some interpreted the messages to mean that imaging is unnecessary for all, rather than the intended message, which was that imaging is unnecessary for those without signs of a serious condition: 'People like me, I would ignore it (the posters

| Main findings
This study identified the potential for public health campaigns to raise awareness about harms of medical overuse and overdiagnosis, but also elicit strong negative reactions from community members. We found that the community members seemed to trust what their family doctor says more than what a poster says. Community members reacted with surprise and initial mistrust, but with further reflection found some of the messages informative and useful

| Strengths and limitations
This research has several important strengths. We conducted focus groups in a socioeconomically and culturally diverse area in Australia. We developed public health campaign materials in stages, incorporating feedback from a multidisciplinary team of experts and consumers. We developed the materials to be suitable for people with an average reading level (grade 7), and we have separately demonstrated their potential to influence behavioural intentions regarding imaging. 19 In terms of limitations, some vocal participants may have in-

| Comparison with previous research
Several previous studies have examined the role of beliefs in the overuse of lumbar imaging. A systematic review of 1747 patients and clinicians found that overuse might happen because many believe imaging helps identify the source of low back pain and rule out serious pathology. 6  There is an ongoing debate about using appeals to emotions, such as fear, to promote public health. In a recent article, authors argued that using fear to improve health behaviour could be harmful and may not even be effective. 26  There are, however, important differences between those campaigns and the current campaign. Rather than using emotion to encourage a 'positive' health behaviour such as safe sex or quitting smoking, the current campaign was discouraging people from asking for, or accepting, a medical test from their doctor. Despite the differences in the campaigns, the effects were similar: campaigns using fear can raise awareness of health issues, including the harms of unnecessary medical care.
These campaigns have also shown that measuring the impact of public health initiatives is not without its challenges. The Australian HIV campaign led to a drastic increase in HIV antibody testing and increased awareness about the spread of the virus. 28 However, there were also unintended impacts. The nature of the HIV campaign messaging-where a TV advertisement shows the 'Grim Reaper' bowling down women and children in a bowling alley as a metaphor for mortality from AIDS-sparked substantial pushback. Some argued that the advertisement sensationalized the illness, provided very little information, scared children, and stigmatized a group who were already facing widespread discrimination. 29 Appeals to fear in public health should consider measuring not just the impact on health but also the potential unintended impacts on public perception and trust.
Our study suggests that a delicate balance in public health, between maintaining community trust and improving health outcomes, could apply to communicating about overdiagnosis. Such communication is challenging because the public can find the concept of overdiagnosis confusing and there is a risk of negatively affecting those who have been diagnosed already. 30  Overuse of imaging is a complex phenomenon driven by clinician and patient beliefs, fear of litigation, and patient pressure. 5,6 This campaign directly targeted only part of the patient-clinician dyad.
However, posters can raise awareness among clinicians indirectly, 8 and the campaign was designed to target clinicians indirectly via discussions with the patient. Future work could explore whether this happens in practice. Behavioural interventions to reduce overuse are likely to need a multi-faceted approach and may need to more closely target clinicians. 36 Nudge-interventions that encourage changes to clinical behaviour without restricting choice, such as default option nudges in the Electronic Medical Record, have potential to reduce overuse. 36 More restrictive policy-based interventions such as removing Medicare funding for ineffective tests and treatments also have potential to reduce overuse. 37

| CON CLUS IONS
This study suggests that public health campaigns using strong messaging and imagery could help raise awareness about the harms of overdiagnosis and overuse of medical tests but could also generate negative reactions among community members. Strong community beliefs in favour of diagnostic imaging, scepticism about overdiagnosis and anger at the concept of reducing testing could all be barriers to an effective campaign to reduce overuse.

| ROLE OF THE FUNDER /S P ON SOR
The funder had no role in the design and conduct of the study; collection, management, analysis and interpretation of the data; preparation, review or approval of the manuscript; and decision to submit the manuscript for publication.

| ADDITIONAL CONTRIBUTIONS
None.

| DATA S HARING S TATEMENT
All data relevant to the study is included in the article or uploaded as supplementary information.