INTRODUCTION

The population of women ≥ 75 years is rising rapidly and the incidence of breast cancer, the most common life-threatening cancer in women, increases with age.1, 2 While mammography screening is associated with a 19% relative risk reduction in breast cancer mortality in women < 75 years,3, 4 its effectiveness in women ≥ 75 years is uncertain since none of the screening randomized controlled trials (RCTs) included these women.5 Meanwhile, the harms of screening include pain, anxiety, false positives, and overdiagnosis (diagnosis of tumors that otherwise would not have caused problems during a woman’s lifetime) resulting in overtreatment.6 Therefore, guidelines recommend that older women be informed of the benefits and risks of mammography before screening.7,8,9,10

To help inform women ≥ 75 years of the benefits and risks of mammography, we previously developed a pamphlet mammography screening decision aid (DA) for women ≥ 75 years.11 In a pilot study, receipt of the DA was associated with women ≥ 75 years being more knowledgeable about mammography and with fewer women intending to be screened.12 The American Cancer Society (ACS) and experts recommend use of decision aids to help older women make more informed, preference-sensitive, screening decisions.7, 13 However, implementing DAs in practice, particularly in primary care, is challenging.14,15,16 Relying on primary care physicians, nurse practitioners, or physician assistants (herein referred to as PCPs) to give patients DAs results in inconsistent delivery and may not be feasible or sustainable.15, 16 Instead, engaging other members of the healthcare team to provide DAs may increase implementation.17, 18 Little is known about non-PCP healthcare team members’ thoughts on implementing DAs. Therefore, we aimed to use qualitative methods to learn from primary care administrators, nurses, and staff their thoughts on how best to implement our mammography DA.

METHODS

Design and Setting

We conducted a qualitative study using in-person interviews with primary care–based administrators, nurses, medical assistants (MAs), and practice assistants (e.g., front desk staff). We approached potential participants from 11 primary care practices that are participating in a large randomized controlled trial testing the efficacy of a mammography screening DA for women ≥ 75 years.19 The 11 practices include a large Boston-based academic internal medicine practice and geriatrics practice, 7 Boston-area community-based practices, and an academic internal medicine practice and family medicine practice in Chapel Hill, North Carolina (NC). None of the non-PCP staff approached for this study were involved in the trial since research assistants (RAs) identified eligible patients and gave them the DA during the trial. During the timing of this qualitative study, patients participating in the RCT were being followed via chart abstraction to see whether or not they underwent screening mammography.

For this study, we sent non-PCP staff recommended by physicians at each practice a personalized email and/or letter (depending on the practice) informing them of the study. The letter explained that the study involved meeting with an RA for approximately 30 minutes to provide feedback on and to discuss barriers and facilitators to implementing a mammography screening DA for women ≥ 75 years. Participants were offered a $50 incentive and were informed that their medical director approved of the study, that their comments would be confidential, and that participating would not affect their employment.

Data Collection

Interviews were audio-recorded and conducted between July 2016 and January 2017 in private office space by one of three RAs (2 in Boston, 1 in NC) trained in qualitative methods. Initially, the RA asked participants to read the DA and to complete a structured questionnaire on the DA’s length, balance, clarity, and to report their demographics (sex, race/ethnicity, years in practice, age, and professional role). Then, using a semi-structured questionnaire, the RA asked open-ended questions about what they liked about the DA, suggestions for improvement, and their thoughts on how to make the DA available to older women (see online appendix for questionnaire). The RA probed participants to learn their thoughts on how use of the DA would affect their workload and about any training that they would need. The interview guide was modified as interviews were conducted to explore emerging themes.

Analysis

Interview audio recordings were transcribed verbatim by a professional transcriptionist and analyzed using NVivo 11 (QSR International) qualitative software. We conducted a thematic analysis to identify themes in our data.20, 21 All investigators read the first 5 interviews and coded phrases, sentences, or longer segments of text.20, 22 Codes were generated inductively (i.e., they emerged from the text). After the open coding process, our team met to organize codes into larger categories to reflect major themes (axial coding). Disagreement about the meaning of themes or codes was discussed by the research team and resolved by consensus. Once a codebook was established, subsequent interviews were read by at least 4 investigators and were coded in detail by at least 2 investigators. Differences in coding were reconciled by consensus until 100% agreement was reached. As new themes emerged, we developed new codes and recoded previously coded interviews. Ultimately, six iterations of the codebook were used. We stopped interviewing participants when no new themes emerged from interviews and thematic saturation was reached.20 During interpretive analysis, we re-grouped transcripts by participants’ professional role and re-read them to see if themes varied by professional role.23 We present direct quotes from the data to illustrate themes that emerged. The study was IRB approved at the Beth Israel Deaconess and University of North Carolina Medical Centers.

RESULTS

At least one staff member and/or nurse participated from 8 of 11 practices participating in the randomized controlled trial of the DA. Of 46 non-PCP staff approached, 69.6% (n = 32) agreed to participate; 8 never responded and 6 refused to participate. Of the 32 participants, 68% (n = 21) worked in the Boston area, 34% (n = 11) worked at community practices, and 53% (n = 18) had been in their position for ≥ 10 years. Participants’ professional roles varied: 22% (n = 7) were administrative supervisors, 31% (n = 10) were nurses, 16% (n = 5) were medical assistants, and 31% (n = 10) were practice assistants. Of the 32 participants, 88% (n = 28) were female and 59% (n = 19) were non-Hispanic White (see Table 1 for our sample’s characteristics).

Table 1 Sample Characteristics (n = 32)
Full size table

DA Feedback

Overall, 68% (n = 21) of participants thought the DA’s length was just right while 32% (n = 10) thought it was too long (1 did not respond). Approximately half (53%, n = 17) thought the DA was balanced, while 25% (n = 8) thought it was slanted towards not getting a mammogram and 22% (n = 7) thought it was slanted towards getting a mammogram. In open-ended comments, participants felt that the DA would help older women with screening decisions, that it was informative, and that it was easy to understand. “It is written very well, it is very respectful, and it’s not talking over anybody’s head.” They also liked that the health score made it “specific to the person.” However, some felt that the pictographs in the DA would be confusing to women with low health literacy.

Facilitators to DA Use

Several themes emerged regarding facilitators to using the DA which we grouped into three categories: motivators for use, ease of use, and factors that would facilitate its reach to more patients (see Table 2). Motivators included that participants felt it was important for older women to make a personal choice about mammography screening. “I would suggest that it becomes available to all women in that age range to help people make the decision.” Participants also felt that a medical director’s approval and a champion within the practice would serve as motivators. In addition, they felt it would be helpful if the DA was recommended by a trusted physician, if guidelines supported its use, and if it was considered standard of care. “[It would help] if your doctor said this is what we are now doing, I want you to read it carefully.”

Table 2 Facilitators and Barriers to DA use
Full size table

In terms of ease of use, most felt that existing systems could be used to implement the DA and felt that for MAs, delivering the DA was analogous to tasks they were already doing. In terms of reach, participants noted that older patients are seen more frequently and that Medicare Annual Wellness Visits (AWVs) would be an opportune time for the DA to be delivered. They also felt it would be helpful if the DA were available in a “variety of ways” (e.g., on the web, the electronic medical record, a patient portal, and on paper).

Barriers to DA Use

Several themes also emerged regarding challenges to using the DA and factors that would lower motivation and limit its reach (see Table 2). MAs and practice assistants thought that delivering the DA may increase their workload slightly. “It might take a few more minutes because they might have a few more questions.” Administrators were also concerned about overburdening staff, especially MAs. Participants felt that staff would need some training, albeit minimal, to deliver the DA. Specifically, staff would need to be informed of the DA’s purpose and given a brief script to introduce the DA and to respond to patient questions. Participants also felt that it may be challenging to identify appropriate patients to receive the DA, to keep the DA stocked, and to track its use.

While participants commented that Medicare AWVs would trigger the DA’s use, participants also noted that patients already received a lot of paper work during these visits. Others also questioned whether patients would actually read the DA. “They may just take it, throw it in their bag and never look at it again.” A couple of participants noted that the DA was only in English.

MAs and practice assistants tended to be more confused about the DA’s purpose then administrators or nurses. “I am assuming the point of the pamphlet is to get women to have a mammogram.” Some MAs and practice assistants recognized that the DA diverged from how they were usually trained to think about screening which generally focused on increasing uptake, “it makes me think about all the outreach that we’re doing, getting patients in for mammograms.” They questioned whether patients would perceive the DA as the government trying to ration care. “It raises the suspicion that someone wants to save money.” Along this line, a few participants also questioned why the DA focused on stopping screening rather than on tapering screening.

Ideas for Implementation

Table 3 presents participants’ suggestions for implementing the DA. Since “doctors may all have a different idea of what is the best” and since “you are talking about a reduction in medical care,” most staff felt, regardless of role, that a patient’s PCP would need to approve the DA and support its use with individual patients. “I feel strongly that if the physician agrees with this and s/he wants to educate their patient it’s a great tool.”

Table 3 Ideas for Implementing the DA
Full size table

Nearly all participants felt that older women should read the DA in the waiting or exam room before discussing mammography with their PCPs. Thus, many participants brainstormed ways to have patients come early to visits. However, they acknowledged that this can be challenging and noted that even if a patient does come early the PCP may be ready for her and the patient may not have time to read the DA. They also noted that it can be difficult to identify appropriate patients to receive the DA before a visit. Therefore, participants also suggested that the DA be given to patients at the end of a visit either by the PCP or by staff to be read at home before the next visit. However, in these cases, participants were concerned that women would forget.

Participants tended to present ideas for implementing the DA related to their professional role. For example, practice assistants suggested that women receive the DA while in the waiting room and MAs suggested women receive the DA while being “roomed.” Nurses suggested that the DA be used during group preventive health visits or with a health educator, especially for those with low health literacy. “Even for those with less [education], someone could be guided through it.”

Administrators and nurses tended to review the pros and cons of multiple methods of delivering the DA including population health approaches such as mailing the DA to all women at the time of their 75th birthday, or to all women ≥ 75 years before a Medicare AWV, or that it be sent to patients through a patient portal. Even in these cases participants emphasized the importance of having a PCP approve and support use of the DA, “If you mailed it to them have a cover letter supported by their physician.” Participants also recommended that the DA be made available at mammography suites; however, they realized that mammography generates revenue for radiology which may be a conflict. “You have to make sure it’s not taking away from the company.” Other suggestions included making the DA available for viewing or printing from the web, through the electronic medical record, and/or to make a video. Within medical records, participants noted that existing alerts could be changed to accommodate the new practice and be used to identify patients who should receive the DA and to document its use. “You change the alert after a certain age rather than just eliminating it.”

DISCUSSION

Primary care–based nurses, administrators, MAs, and practice assistants felt that a mammography screening DA for women ≥ 75 years was readable, clear, and would help older women with mammography screening decisions. To facilitate implementation, they felt it would be best if the DA was approved for use by PCPs in their practice and that it be given, typically by an MA, before a visit—especially before a Medicare Annual Wellness Visit. Due to the strength of the patient-PCP relationship, staff also felt it would help if PCPs encouraged its use with individual patients. Staff noted that systems needed to be developed to identify appropriate patients to receive the DA, to keep it stocked, and to document its use. They also felt that staff would need to be informed of the purpose of the DA and given a brief script to introduce it to patients and on how to respond to patients’ questions. While their comments were focused on how to implement a specific DA, their thoughts may be helpful when considering implementing other DAs.

The need to figure out how best to implement DAs in practice is increasingly necessary as national policy is progressively more supportive of shared decision-making. Specifically, the Affordable Care Act includes several provisions on shared decision-making,24 the National Quality Forum is developing quality metrics for use of decision tools,25 and the Centers for Medicare and Medicaid Services have been adding shared decision-making requirements before providing coverage for medical interventions with uncertain benefit to risk ratios (e.g., lung cancer screening26). We and others foresee a time when Medicare would require shared decision-making before provision of mammography screening for women ≥ 75 years since most guidelines already recommend individualized decision-making.27, 28 However, few studies have examined implementation strategies to increase adoption of DAs in primary care29 and we are unaware of studies focused on the perspectives of non-PCP primary care team members’ even though several studies have recommend engaging non-PCP team members in DA implementation.17, 18 Barriers to DA implementation identified in other studies include time constraints, insufficient provider training, lack of applicability, lack of patient understanding, and inadequate clinical information systems16, 30,31,32; facilitators include provider motivation and the positive impact on patient outcomes.16 Participants in our study brought up these barriers and facilitators; however, most felt the DA was clear and applicable to women ≥ 75 years and felt that they would need minimal training to provide the DA. While system-based approaches result in more patients receiving DAs15, 33,34,35 and many of our participants suggested system-based approaches; they also stressed that regardless of who delivered the DA that it would need to be approved and its use supported by a patient’s PCP.

Participants in our study felt that MAs were the best suited members of the healthcare team to deliver DAs. Yet, administrators were concerned about increasing MA workload. With PCP shortages and an aging population, MAs are increasingly being utilized.36 However, MAs generally receive minimal training and have a continuously expanding scope of responsibilities.37 In this context, the benefits of adding new tasks to MAs will need to be weighed against the cost of expanding their duties. Reassuringly, MAs in our study felt it would be feasible for them to implement the DA as long as they had the support of the PCPs they were serving.

Based on our study’s findings, to increase the likelihood of successful implementation of our DA, we plan the following. First, we have made the DA available online for viewing or printing.38 Once we made the DA available online, we realized that it would be better if the DA were available at a website more familiar to clinicians. Therefore, we are now developing a tab within the widely used ePrognosis website for links to DAs like this one that consider patient prognosis. Since participants in our study felt it would be helpful to have a brief explanation of the DA’s contents, its purpose, and an example script to use when providing the DA, we will include this information on ePrognosis. We also recently had this DA professionally translated to Spanish due participants’ concerns that it was only in English. We further recommend that practices print some DAs in color and store them where other health educational materials are kept. We also recommend that healthcare systems planning to implement the DA create a medical record alert to flag patients who may be appropriate to receive the DA and space to document its use. Since participants felt that Medicare AWVs would be the ideal encounter for older women to receive the DA, provision of the DA should be incorporated into processes for implementing these visits.39, 40 It may also be innovative to think of ways to incorporate group visits to cover the health educational component of Medicare AWVs.41, 42 In addition, our DA will need to be endorsed by professional organizations, since staff suggested it would be important for the DA to be considered standard of care. Such endorsement would also likely lead to approval by individual PCPs.

While our DA was written at a 6th grade reading level using low literacy principles,11, 43 many participants felt that it was too long and some felt that the pictographs would be hard for older women with low health literacy to understand. While the pictograph is the graphical format most recommended for patients with low literacy,44, 45 few DAs have been developed specifically to meet the needs of vulnerable populations and it may be that even pictographs are too challenging for older adults with low literacy.46 We are currently developing and testing an even lower literacy version of the DA without pictographs.

Our study has several limitations that must be acknowledged. First, our findings are based on participant self-report rather than experience providing the DA. However, conducting qualitative research to determine how best to implement an intervention is a recommended first step for successful implementation.47 Due to purposeful sampling, participants may be more satisfied in their work and thus likely to be supportive of implementing DAs than non-participants. Also, staff, especially MAs and practice assistants, may not have felt comfortable expressing a lack of motivation to deliver the DA. Our sample size was too small to examine difference in perceptions by participant age, geography, or sex. As a next step, we will be surveying PCPs to learn their thoughts on implementing the DA.

This is one of the first studies to interview non-PCP healthcare team members on how best to implement a DA in primary care. Based on their comments, there is likely no one-size-fits-all approach to implementing this or other DAs. However, most non-PCP staff felt that as long as use of the DA was approved and supported by PCPs that women should receive the DA before a visit from staff (usually medical assistants) so that patients could ask their PCP questions during the visit. Non-PCP healthcare team members also shared many insights to help DA developers think about what they can do to increase DA implementation (e.g., make it available in multiple formats, in multiple languages, and at multiple literacy levels, and to seek endorsement from professional organizations). They also reported that small modifications of existing systems could increase the likelihood of implementation (e.g., medical record alerts to identify appropriate patients and space for documentation). These insights are timely and needed to increase implementation of high-quality DAs.