Developing and testing a brief clinic‐based lung cancer screening decision aid for primary care settings

Abstract Background Cancer screening‐related decisions require patients to evaluate complex medical information in short time frames, often with primary care providers (PCPs) they do not know. PCPs play an essential role in facilitating comprehensive shared decision making (SDM). Objective To develop and test a decision aid (DA) and SDM strategy for PCPs and high‐risk patients. Design The DA was tested with 20 dyads. Each dyad consisted of one PCP and one patient eligible for screening. A prospective, one‐group, mixed‐method study design measured fidelity, patient values, screening intention, acceptability and satisfaction. Results Four PCPs and 20 patients were recruited from an urban academic medical centre. Most patients were female (n = 14, 70%), most had completed high school (n = 15, 75%), and their average age was 65 years old. Half were African American. Patients and PCPs rated the DA as helpful, easy to read and use and acceptable in terms of time frame (observed t = 11.6 minutes, SD 2.7). Most patients (n = 16, 80%) indicated their intent to be screened. PCPs recommended screening for most patients (n = 17, 85%). Conclusions Evidence supports the value of lung cancer screening with LDCT for select high‐risk patients. Guidelines endorse engaging patients and their PCPs in SDM discussions. Our findings suggest that using a brief, interactive, plain‐language, culturally sensitive, theory‐based DA and SDM strategy is feasible, acceptable and may be essential to effectively translate and sustain the adoption of LDCT screening recommendations into the clinic setting.


| INTRODUCTION
Cancer screening-related decisions can require patients to evaluate complex medical information in short time frames, often with primary care providers (PCPs) they do not know. Intense emotions about the possibility of a cancer diagnosis may create anxiety and affect decision making. In the United States (U.S.), the Centers for Medicare and Medicaid Services (CMS) decision to include low-dose computed tomography (LDCT) for lung cancer screening as a reimbursable service with the requirement that a "counselling and shared decision making" (SDM) visit with a PCP precede the screening is evidence of the anticipated complexity of making this decision and the importance of weighing its risks and benefits. 1,2 PCPs play an essential role in facilitating comprehensive SDM. 3,4 SDM was endorsed by the U.S. Preventive Services Task Force in its Screening for Lung Cancer Recommendation Statement and is viewed as a way to minimize concerns related to implementation variances. 5 SDM is a collaborative communication strategy that allows patients and their PCPs to make health-care decisions together, taking into account the best clinical evidence available as well as the patient's values and preferences. SDM incorporates the patient's voice in health-care decisions and is described as the pinnacle of patientcentred care. [6][7][8] The principles of self-determination (the freedom of patients to make their own choices) and relational autonomy (an understanding that decisions are made in the context of interpersonal relationships and mutual dependencies) are important precursors to the integration of SDM into clinical practice. Even though the importance of these principles has been well documented, there is a lack of practical guidance about how to implement SDM between PCPs and patients in routine clinical settings where time is limited and there are often competing priorities. 4,[9][10][11] Elwyn and colleagues 4 describe a three-step SDM model that is practical, is easy to remember and can act as a guide to skill development: (i) introducing choice; (ii) describing options, often by integrating the use of decision support tools or aids; and (iii) helping patients explore preferences and make decisions. This model involves deliberation, with the understanding that decisions should be influenced by respecting "what matters most" to patients as individuals and that the decision should also rely on patients developing informed preferences. 4 SDM emphasizes building a good relationship in the clinical encounter so that information is shared and patients are encouraged and supported to express their preferences during the decision-making process. 4,12,13 Evidence strongly indicates that decision support tools or aids lead to improved knowledge, reduced decisional conflict and decisions that are compatible with the patient's value system. [14][15][16][17][18][19]  cancer; however, they are not all interactive, written in plain language, sensitive to diverse cultures or designed to be used in brief clinical encounters (see Table 1). [16][17][18][19][20][21][22][23] The purpose of this pilot study was to test the feasibility and acceptability of implementing a brief, clinic-based DA written with lower readability and enhanced cultural sensitivity and a SDM strategy developed for PCPs to use in clinical settings as they discuss the pros and cons of lung cancer screening with their high-risk patients.

| DA development
A DA-based on Janis and Mann's conflict theory of decision making, 24 18,28 In 2003, an international, collaborative group was established for the sole purpose of developing standards for patient DAs. This guideline, the IPDASi (the fourth version of which was released in 2014), includes 44 standards and 3 broad categories of criteria: (i) qualifying criteria, (ii) certification criteria and (iii) quality criteria. 26 The qualifying criteria category is considered definitional, and the criteria are essential for designation as a DA. The certification category includes criteria that enhance avoidance of risk of harmful bias. Lastly, the quality criteria are designated non-essential but are known to enhance a DA.
Results from a statewide survey of PCPs about knowledge, attitudes and use of lung cancer screening also informed DA development. 27 Among 101 physicians surveyed, knowledge gaps existed about approved guidelines and reimbursement. Major physician concerns included unnecessary procedures, radiation exposure and patient anxiety. 15,27 These three physician concerns, among others, were addressed in the DA.
Using a model development process, components of the decisionbalance portion of the DA were developed by the primary author, and consensus was reached with the help of an interdisciplinary research team (composed of the coauthors), three nationally known content experts in decision making and lung cancer screening, plus three laypersons. 29 The DA incorporated IPDASi standards, including all 7 of 7 defining criteria, 8 of 9 criteria to reduce bias, 2 of 4 criteria related to screening and 10 of 13 quality criteria (see Table 1 for comparisons).
The DA developed for this study is an eight-page, 5.5 inch-by-8  Table 2). 31,32 The final page of the DA asked patients to rate, on a 1-item scale of A prospective, one-group, mixed-method design was used to evaluate the DA implementation and a SDM strategy. The setting was an urban academic family health centre in the south-eastern United States serving a large proportion of low-income individuals.

| Recruitment
For the purpose of this pilot feasibility study, the recruitment goal was 20 patient-PCP dyads. A list of 485 patients between the ages of 55 and 77 who listed Medicare or Medicaid as their primary health insurance was collected from the health centre's electronic medical record system. Two hundred of these patients were randomly selected to receive a mailed invitation at their home address. Patients were offered the opportunity to opt out of a recruitment telephone call by sending an email or leaving a voicemail message for the study coordinator. Of the patients called by the study coordinator, 57 had a disconnected telephone number. The recruitment plan included calling the remaining 143 patients up to four times in an attempt to establish contact.
The first 20 interested patients who met these criteria were scheduled for an appointment during one of the three pilot testing events held at the health centre.
During the recruitment telephone calls, the study coordinator determined eligibility. Eligibility criteria followed the Medicare guidelines, specifying that the patient be (i) a Medicare beneficiary; (ii) between 55 and 77 years old; (iii) without current symptoms of lung cancer; (iv) interested in learning more about lung cancer screening; (vi) a current smoker, or a former smoker who had quit within the

| Study procedures
A few patients required transportation, the cost of which was included in the grant budget. Transportation was provided by a private local taxi service known to the research team. At study completion, patients received a $25 cash "thank you" gift.
An invitational email recruited four PCPs to participate as part of the planned 20 PCP-patient dyad interactions. It was not the intent to match the patients with their usual PCPs during this feasibility study.
Therefore, some patients engaged in a SDM discussion with a PCP who was not their usual physician.
Approval was obtained from the local university hospital's institutional review board affiliated with its school of medicine, and each patient gave informed consent to participate after having the opportunity to read the informed consent form and ask questions of the study staff before meeting with their assigned PCP. The testing times were scheduled at one location over a 6-week period.  Table 3)

| RESULTS
All 20 patients were current cigarette smokers or former ciga- In response to a scale on the DA, a majority of patients (n = 16, 80%) rated the importance of a decision about lung cancer screening as a 10 of 10 before and after the discussion with the PCP. The others (n = 4, 20%) indicated that they "preferred to make a final decision after discussing lung cancer screening with their usual PCP." After the discussions with patients, the PCPs wanted the majority of patients (n = 17, 85%) to consider screening.
Overall, the DA was rated highly according to all 6 acceptability criteria (see Table 4). Patients felt the DA was easy to read, easy to use and helpful for making a decision and discussing their personal values with a PCP. All of the PCP responses were positive as well. For the most part, the time involved with using the DA (the first DA some had ever used)-measured by self-report and direct observation (see Table 4)-was acceptable. The majority of patient-PCP interactions (n = 14, 70%) took 10 minutes or less. The longest interaction took 17.2 minutes. It was determined that this particular patient required a longer SDM interaction because of poor vision ( Table 5).
All patients agreed to participate in a brief exit interview.

| DISCUSSION
The most interesting finding was that a high percentage of patients  threat of lung cancer. We agree that individuals need to understand the best available medical evidence relevant to a screening decision. 15 Well-designed DAs that enhance health literacy will promote an understanding of complex health information and an interest in preventive behaviours. Understanding the barriers to the use of DAs and a SDM strategy in the clinic setting will help to enhance their integration. We assumed the greatest barrier to implementation would be the time commitment involved. This study (in which a majority of patient-PCP interactions took 10 minutes or less), together with evidence from more than 100 randomized control trials, provides no indication that an additional or unacceptable time commitment is required to engage in SDM in a clinical practice setting. 35 One of the greatest strengths of this study was the sample's socioeconomic characteristics. Patients were diverse in gender, racial background and educational level, and they fit the qualifying characteristics for Medicare reimbursement of lung cancer screening (Medicare billing code G0296). A second strength was that fidelity was carefully monitored, and implementation of the DA was timed by direct, in-person observation.
The findings are limited by the one-group design and sample size.
Testing the DA with the patient's usual PCP would have provided a clearer picture of how the DA would work when integrated into the typical flow of the patient's care. A second limitation was that follow-up was not conducted to determine whether the patients received a lung cancer screening or met with their usual PCP to discuss screening and/or obtain a referral. The decision to forego this follow-up was based on our desire to avoid circumventing the discussion between patients and their PCPs. We encouraged each patient to take the DA to an appointment with his or her PCP to discuss scheduling a lung cancer screening.

| CONCLUSION
Strong evidence exists that patients exposed to DAs feel more knowledgeable, better informed, more certain about their own values and engaged in a more active role in decision making about their health choices. Our findings suggest that using a brief, interactive, theorybased DA written in plain language at the fifth-grade reading level in a clinical setting is acceptable to diverse group of patients and PCPs.
These pilot study results will be used to refine and enhance the use and delivery of the DA in clinical settings with a more varied PCP workforce (nurse practitioners and physician assistants in addition to medical doctors). The results will also help determine a sample size for a full-powered study that further explores the integration of SDM strategies between PCPs and their high-risk patients with diverse health literacy skills and their adherence to their chosen decisions.

| Practice implications
Now that evidence supports the value of lung cancer screening with LDCT and it is a covered service by most private and public insurance plans in the U.S. for selected patients described as high-risk, it is important that implementation processes proceed in a way to maximize benefits, minimize harms and enhance sustainability. Demand for this service will increase as awareness about lung cancer screening increases. Brief clinic-based strategies, with acceptable readability levels, 37 are essential. Effectively translating and sustaining the use of DAs and SDM in clinic settings will require careful attention to implementation approaches. 39