An official website of the United States government
NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
John A. Spertus, MD, MPH, FACC, Adnan Chhatriwalla, MD, Carole Decker, RN, PhD, Philip G. Jones, MS, and Elizabeth Gialde, RN, MSN.
Author Information and AffiliationsPatients undergoing percutaneous coronary intervention (PCI) may receive drug-eluting stents (DES) or bare-metal stents (BMS). DES reduce the risk of a repeat procedure but necessitate prolonged use of dual antiplatelet therapy (DAPT), which can increase the risk of bleeding, delay elective surgeries, and increase medication costs. Because these trade-offs directly impact patients' experience and outcomes, patients should be engaged in the decision-making process. However, less than a third of patients discuss stent choices with their doctors.
To create and implement a shared decision-making (SDM) tool for stent choice that incorporates individualized estimates of each patient's likely reduction in repeat procedures from DES and describes the drawbacks of prolonged DAPT, and to study the impact of the SDM tool on patient education, stent preference, participation in SDM, and, secondarily, physicians' use of DES or BMS.
To create the SDM tool, we interviewed patients and providers to identify the information perceived to be most important in stent selection. We implemented the tool in a nonrandomized, sequential fashion at 2 US hospitals. In the preimplementation phase, we did not use the SDM tool; we interviewed patients to determine their knowledge of stent types, benefits and drawbacks of DES and BMS, and their stent preference. In the 2 sequential postimplementation phases, all patients received the paper SDM tool before their procedures, with or without decision coaching from a trained nurse.
We assessed experiences with stent selection in 336 patients receiving “usual care” without the SDM tool, 113 receiving the SDM tool with coaching and 137 receiving the tool without coaching. In fully adjusted analyses, the SDM tool with coaching group, as compared with usual care, resulted in 1.8 (95% CI, 1.5-2.1) out of 6 more correct responses to questions about BMS and DES. For patients receiving the SDM tool with coaching, the rate ratio (RR) for discussing stent options with nurses was 3.7 (95% CI, 2.8-4.7) and for discussing with doctors was 1.5 (95% CI, 1.26-1.79). Patients receiving the tool with coaching were more likely to state a stent preference (relative risk [RR], 1.9; 95% CI, 1.6-2.3), to report being involved in the decision (RR, 2.87; 95% CI, 1.8-4.5), and to correctly recall the stent that they received (RR, 1.5; 95% CI, 1.1-1.8). No clinically significant differences were observed between use of the SDM tool without coaching and usual care. No differences in physicians' use of DES and BMS were observed after implementation of the SDM tool, and less than 50% of patients who preferred BMS actually received one.
Implementation of an SDM tool for stent selection, accompanied by decision coaching, was associated with improvements in patient education, participation in SDM, and patient satisfaction, but it resulted in no change in the type of stent used. Use of the SDM tool without decision coaching had no impact on patient education, participation in SDM, or stent selection.
The intervention was designed to be sustainable in routine clinical care. However, the lack of hospital nursing staff to perform decision coaching, despite available funding, was the most significant limitation. To achieve the benefits of SDM with decision coaching, novel incentives will need to be created to enable hospitals to invest resources necessary to support decision coaching.
Percutaneous coronary intervention (PCI) for patients with coronary artery disease (CAD) is performed >600 000 times/year in the United States.1 Patients undergoing PCI can be treated with either drug-eluting stents (DES) or bare-metal stents (BMSs). Although DES are associated with a 50% reduction in repeat PCI, by avoiding restenosis of the coronary artery,2,3 the magnitude of benefit depends on a patient's underlying restenosis risk.4,5 For example, as few as 6 high-risk patients need to be treated with DES to avoid 1 repeat PCI procedure, while more than 100 low-risk patients may need to be treated with DES to avoid 1 repeat procedure. Importantly, if patients receive DES, they must take a prolonged course (6-12 months) of dual antiplatelet therapy (DAPT), compared with only 1 month if they choose BMS.6 This is to minimize the risk of stent thrombosis, a potentially catastrophic complication that can result in heart attack or death. DAPT represents an increased medication burden for patients, can be expensive, can cause increased major and nuisance bleeding,7,8 and can complicate future surgeries.9 In fact, a recent study demonstrated that nuisance bleeding occurs in 37.5% of patients over the year after a heart attack and is associated with a significantly lower quality of life, and that the use of DAPT doubles the risk of nuisance bleeding.7 Despite these trade-offs, DES are currently used in roughly 80% of PCI cases, regardless of risk for restenosis.10,11
In 2010 the American Medical Association advised that when more than 1 clinically acceptable treatment is available, patients should be engaged in SDM, as it can improve the value of US health care.12 The National Academy of Medicine likewise includes patient-centered care as 1 of its 6 domains of quality.13 In the setting of PCI, stent selection should be a preference-sensitive decision because of the offsetting benefits of DES and the drawbacks of prolonged DAPT. Research shows that patients want to be involved in their treatment decisions, with a recent study reporting that only 1 in 10 patients felt that the doctor alone should make the decision. Similarly, in another study that asked patients who made the decision about whether BMS or DES was used, >70% felt that the doctor alone made the decision and only 31% reported discussing stent options with their physician.14 These findings were also consistent with a prior study that suggested that only 10% of patients undergoing PCI were presented with other options, only 19% were presented with a treatment's drawbacks, and only 16% were asked about their treatment preferences.15
Clinicians may change their behavior in response to personalized risk estimates and this can lead to improved care and outcomes. In a previous 9-center prospective study,14 we used the Patient Risk Information Systems Manager (PRISM) to create personalized informed consent documents that included patients' risks of periprocedural mortality16 and bleeding17 as well as restenosis requiring a repeat PCI if patients are treated with DES or BMS.18 PRISM is a computer program that executes multivariable risk models with a patient's specific risk factors. Using PRISM, we can take previously validated risk models and the results of comparative effectiveness research and prospectively deploy them in clinical care.
In the aforementioned 9-center prospective study, we demonstrated significant improvements in the informed consent process with PRISM, including more often reading the consent form and better knowledge transfer, suggesting a better patient experience with care.14 We also demonstrated that the inclusion of risk estimates for target vessel revascularization (ie, the risk of needing a repeat procedure to treat the same artery) supported greater patient participation in SDM with respect to stent selection, which increased from 31% to 58% of patients discussing stent types with their doctors. Despite this improvement, 42% of patients, even after signing consent forms explaining their target-vessel revascularization (TVR) risk with BMS and DES, did not discuss stent choices with their physicians. To better engage patients undergoing PCI in SDM, the current study sought to create a personalized SDM tool through collaboration with multiple stakeholders, including patients, nurses, and interventional cardiologists. We also trained the medical and nursing staff to serve as decision coaches to support and educate patients. The successful implementation of this technology and documented improvements in knowledge transfer could underscore the feasibility and potential of providing a novel infrastructure and process of care to engage patients undergoing PCI in SDM.
One of the goals of health care is to improve quality while limiting costs. While reducing costs was not a primary goal of this intervention, aligning treatment with patient preference may, indirectly, lower the costs of PCI by using less DES, which are more expensive than BMS. In a prior analysis,10 we calculated the cost savings and increased rate of TVR if the rate of DES use was decreased in patients at low risk for TVR. We found that if the rate of DES use was cut in half (from 74% to 37%) in patients at the lowest risk for restenosis, an estimated $205 000 000/year could be saved in the United States (adding together the cost savings for stents and DAPT and subtracting the costs of repeat procedures). Interestingly, even if DES were entirely withheld from low-risk patients, the absolute increase in TVR would be <1%, with a savings of >$400 000 000/year. Accordingly, we postulated that a higher-quality SDM process might not only improve patient participation in SDM, but that, if patients with less benefit from DES preferred BMS, savings would occur from respecting patient preferences and not from rationing care.
Communicating these concepts to patients, however, can be difficult. Therefore, the goal of this study was to create a SDM tool to enable patients to understand their personal benefits from DES in preventing restenosis, as compared with BMS, and to inform them of the need for prolonged DAPT if DES were used. This would enable patients to choose the stent type that is most aligned with their personal goals and values.
Given that the decision regarding a patient's optimal stent is likely to involve many considerations, we created our SDM tool based on the critical issues relevant to patients and doctors, as defined by focus groups, in weighing the benefits and risks of DES. It is noteworthy that the investigators had no preferences for which type of stents patients chose, although we were very concerned that, in current practice, the vast majority of patients do not discuss stent options with their physicians. Given the long-term implications of this choice on the need to adhere to DAPT, we believe that patients should be informed and empowered to participate in communicating their personal goals and values. Thus, this novel SDM tool was designed to achieve this goal without directing treatment to one stent or another.
We recruited patients from 4 distinct practices to participate in our focus groups (n = 35). First, we recruited patients from Truman Medical Center (TMC), Kansas City's Safety Net Hospital and provider for a primarily African American and low-income population. Second, we recruited patients from Saint Luke's Mid America Heart Institute (MAHI), which has a more affluent patient population that also includes many rural farmers. We also recruited patients from the Women's Heart Center at MAHI to ensure that women were well-represented. Finally, we recruited patients from the Cabot Clinic, which serves a primarily Hispanic patient population from which we could recruit an additional segment of patients for focus groups. We used information obtained from these focus groups to design, revise, and finalize the SDM tool for stent selection.
The Society for Participatory Medicine (SPM) (http://participatorymedicine.org/) is a 501(c)3 public charity devoted to promoting the concept of participatory medicine among patients, caregivers, and their medical teams. SPM advances the ideal of empowered and engaged patients making informed decisions about their care and treatment, and it fosters communication among medical, communication, patient advocacy, and public health subspecialties to exchange a wide range of ideas related to participatory medicine and to support information and research programs in participatory medicine. SPM provided guidance to the research team via Dave deBronkart (e-patient Dave), a nationally recognized patient advocate. He was on the advisory committee that assisted in identifying appropriate comparators and outcomes, helped to coordinate feedback on the SDM tool from additional patients and patient advocates, and assisted in analysis and publication of preimplementation study findings, for which he was credited with authorship.
HHN represents a coalition of community leaders within the Kansas City African American community, including pastors and their churches, community-based organizations, and safety net health care facilities. HHN recently received substantial resources from an infrastructure grant in community-based participatory research (RC4 MD005738). We used HHN to engage the New Bethel Church, an African American Pentecostal Church in Kansas City, Kansas, to recruit patients for the focus groups that addressed the SDM tool for stent selection. A member of New Bethel Church also served on our steering committee.
We recruited three patients and patient advocates as members of our steering committee. The committee's launch, wrap-up, and quarterly meetings informed decision-making throughout the study and monitored study conduct and progress. Committee members provided feedback on study design and methodology, creation and implementation of the SDM tool, and analysis and dissemination of study results.
Following the final focus group, we posted the tool to the e-Patients.net website (website no longer active) for 10 days for review and comments before sharing a final version with the steering committee.
We recruited interventional and noninterventional cardiologists from MAHI and TMC (n = 8) to participate in interviews that led to the creation of the SDM tool for stent selection. These cardiovascular practices have extensive experience in the management of CAD/PCI patients among a diverse population of patients. We recruited members from the American College of Cardiology (n = 2), American Heart Association (n = 1), and Society for Cardiovascular Angiography and Interventions (n = 1) to serve on our advisory committee and contribute to the study design and methodology. Further, the steering committee also included 2 interventional cardiologists and 1 noninterventional cardiologist. We are continuing to work with these same organizations to disseminate the study findings.
The chief medical officer for United Healthcare in Kansas City served on the steering committee, helped design this proposal, and contributed to study design and methodology. He is advising us on strategies for disseminating our tool and how it might be appealing to payers.
We developed a paper-based SDM tool with input from the study team, steering committee, 4 patient focus groups, provider interviews, and patients. Three phases of development ensued. First, the study team and steering committee determined the selection strategy for focus groups, drafted a preliminary SDM tool, and developed the interview guide. The interview guide (Appendix A) articulated preliminary considerations in stent selection and consisted of questions regarding patients' considerations in stent selection and the framing of risks and outcomes. Multiple presentation formats for the communication of risk were developed and assessed in the patient and physician focus groups as part of phase 2. Examples of each presentation type were used along with appropriate explanatory text. We held a second phase of focus groups to review a revised SDM tool incorporating initial patient input from the first phase. We shared insights from these focus groups with the steering committee before we prepared a final version of the tool to present to a subset of participants in the last, confirmatory phase of tool development.
To elicit the most important aspects of the treatment decisions to include in our SDM tool, we first assembled focus groups with key stakeholders. We conducted focus groups with diverse patients and providers until saturation was attained at MAHI, the Women's Heart Center at MAHI, TMC, the Cabot Clinic, and New Bethel Church. Participants were compensated for their time. Physician and catheterization (cath) lab staff were interviewed and provided feedback on the tool development.
We conducted focus groups on site, depending on the group.
We developed interview guides based on an ancestry approach, where potential topics to be explored are tracked from one study to another until redundancy occurs (see Appendix A). We tailored each guide slightly for each cohort, to optimize relevancy for the diverse groups and to carry forward learnings from prior focus groups.
None.
Factors of importance in stent selection identified by stakeholders included the risk of repeat procedures, need for additional medication, medication costs, side effects of medication, and the need to delay future procedures or surgeries due to DAPT-related bleeding concerns. For each outcome, the optimal mode of risk presentation was discussed (eg, absolute vs relative risk reduction, positive vs negative framing, expression of risks as frequencies vs percentages). Focus groups collated the most common questions patients have about the choice of stents, to create an option grid to be used in both the SDM tool and the decision coach training. We also elicited opinions regarding the optimal timing and setting for patients to use the SDM tool.
Following IRB approval, we conducted purposive patient sampling among patients who recently underwent PCI. We developed an interview guide (see Appendix A) to foster discussions on framing and presentation of benefits and risks of DES; discussing the drawbacks to DAPT (ie, medication taking, delaying future surgery, bleeding, bruising); and finally having the participants rank the importance of the 4 main drawbacks to DAPT. Experienced researchers (CD, BG) facilitated the focus groups, with a Spanish facilitator for the Spanish-only group at Cabot Clinic. Each focus group comprised 6 to 7 subjects/meeting, lasted 40 to 55 minutes, and was digitally recorded. Similarly, provider interviews were assembled from purposive sampling of the >50 cardiologists, >30 cath lab staff, and >25 nurses working in the cath lab.
Using qualitative research methods, team members reviewed the transcribed interviews independently and identified patterns and themes.
These data were manually coded according to passages that exemplified key concepts or ideas from each transcript. This iterative 2-phase process captured the meaning behind the transcribed text from the interviews, with the goal of creating an increasingly sophisticated and rich description of participants' perspectives regarding important outcomes.19-21
After the preliminary codes were identified, team members created a visual display of key nodes and attributes and transformed it into the taxonomy of themes used to assign final coding of the raw narrative data.22 They reviewed conclusions drawn from the qualitative data in the context of the entire data set, with the goal of finding discrepant information and modifying the themes when discrepancies were found. To establish the construct validity of the coding, coders from diverse disciplines were used (ie, nurses, a physician, and an experienced research associate). After separately summarizing and interpreting the findings from patients and clinicians, the research team discussed similarities and differences to reach consensus.
We held focus groups over the course of 2 months, allowing for information gained at 1 group to refine the materials before meeting with the next group. The 4 patient focus groups included 7 patients from MAHI, 7 from TMC, 17 from the New Bethel Church, and 6 from Cabot Clinic. Table 1 shows participants' demographic characteristics. After each focus group was completed, the transcribed audio recording was reviewed and coded. The tool was then modified and edited to incorporate feedback before the next focus group's meeting. The study timeline is included as Appendix B.
Demographic Characteristics of Patients Participating in Focus Groups.
Each time a “new version” was shared, the original tool was also provided so that member checking (ie, accurate integration of focus group participant feedback into the evolving tool) was ongoing. Following the final focus group, we posted the tool to the e-Patients.net website (website no longer active) for review before sharing a final version with the steering committee.
In the postimplementation phases, patients presenting for PCI or coronary angiography before PCI received the SDM tool and decision coaching while in the preparatory area for their procedure (first interventional phase of the study), or the SDM tool alone, without decision coaching (second interventional phase of the study). The patient then underwent the procedure, and following it, once sedation had worn off, the patient was approached to complete a postprocedure interview to assess knowledge transfer, patient assessment of the quality of SDM, and his or her preferences for stent type.
We developed a template of the final SDM tool into which the PRISM system could populate the personalized restenosis risk estimates; those estimates could be delivered on paper to the patients at the time that informed consent for PCI was obtained (see Appendix C). This SDM template included text that describes, in lay terms, the nature of CAD, stents, TVR and DAPT obligations, costs, and risks. All such components were generated from the information obtained from the focus groups.
We analyzed the impact of the SDM tool on patient outcomes in a pre-post fashion, by conducting an interview before hospital discharge that assessed patients' knowledge and experience. In addition, we trained nurses to provide decision coaching to patients receiving the SDM tool. We included in the analytic data sets only patients who underwent PCI and compared outcomes for patients who did and did not receive the SDM tool and for patients who did or did not receive decision coaching.
We recruited patients for this study at 2 PCI centers: MAHI and TMC. Cardiologists at MAHI and TMC are employed by their respective health systems, with separate administrative organizations that are not affiliated. MAHI has achieved nursing magnet status, but TMC has not. We identified patients based on review of the procedure schedule, and patients were approached in the procedural preparatory area before coronary angiography to obtain informed consent for participation in the study at TMC. MAHI's IRB considered the study to be quality Improvement and granted a waiver of written informed consent. There were 1282 patients who underwent PCI during the time period of the study who were not included, mostly due to the lack of a decision coach during the phase in which decision coaching was intended. When a coach was not available, the patient did not receive the SDM tool or coaching and was not enrolled in the study. We do not believe that this affected our study findings, because the availability of a decision coach was a random variable based on the allocation of work days and not associated with patient characteristics.
The SDM tool, along with the PRISM-generated informed consent, was provided to patients in the catheterization laboratory preparatory area. Post-procedure interviews were conducted in the patient recovery unit.
In the postimplementation phase, patients received the paper-based SDM tool (Appendix C) with or without decision coaching. While MAHI already used the PRISM tool for obtaining informed consent in routine clinical care, TMC did not. Therefore, the PRISM platform was implemented at TMC before patients were recruited for the study. Drs Kathleen Goggin and Delwyn Catley, experts in motivational interviewing, designed a training curriculum for decision coaches to impartially solicit patients' preferences and empower patients to express them to their physicians. In the postimplementation phases, decision coaches initiated the SDM process by sharing the SDM tool with patients. Coaches also used frequently asked questions (FAQs)/option grids, thus blending supportive training for the coaches with personalized estimates for the patients. Finally, 2 interventional cardiologists (1 at MAHI and 1 at TMC) were trained in motivational interviewing techniques to serve as physician champions and to engage their colleagues at their institutions.
Postprocedure interviews were conducted in the patient recovery unit.
Because the goal of this study was to develop a tool to increase SDM for stent selection in PCI, the primary outcome in this trial was whether patients discussed stent choices with their physicians. We assessed other measures of SDM, including patient preference for a stent type and discussion quality score based on the Health Care Climate Questionnaire,23 as secondary end points. Secondary goals of the study were to improve patient education and knowledge retention and to assess the impact of the SDM tool on actual stent selection at time of procedure. Therefore, secondary outcomes included patients' recall of individual aspects of the stent discussion, stent knowledge score (based on number of correct answers to 6 questions assessing knowledge transfer), concordance of patient stent preference and type of stent received, and alignment of patient values with stent type received.
A research associate conducted post-PCI patient interviews to collect data. Data collected included answers to questions regarding knowledge transfer surrounding stents and stent types, patients' value of factors that may play a role in stent selection, preferences for patients' level of engagement in SDM, and patients' self-report of their level of engagement in the stent selection decision. Patient demographic, clinical, and procedural characteristics were obtained from the study data collection form (see Appendix D). The data were entered into our electronic data collection system, REDCap, and regularly audited for completeness.
To assess implementation of the SDM tool, we surveyed the sites to understand the barriers and benefits of introducing an SDM tool into clinical practice. To evaluate the effect of the SDM intervention, we compared the study's outcomes (described above) among patients receiving personalized consents for their PCI procedure only (with no SDM tool), patients receiving the SDM tool along with decision coaching, and patients receiving the SDM tool without decision coaching. We made unadjusted comparisons of continuous variables using 1-way analysis of variance, and comparisons of categorical variables using a chi-square or Fisher exact test. We made adjusted comparisons between groups (usual care, SDM tool with coaching, SDM tool without coaching) using estimates from generalized linear models (normal distribution/identity link for mean differences, binomial or modified Poisson distribution with log link for rate ratios), and adjusting for site and patient characteristics between the 3 phases of the study when the standardized differences of the patients between the 3 groups were >10 (race, diabetes, peripheral arterial disease, prior CABG, prior PCI, admission status, and enrolling site).
Comparisons of patients' correct recollection of the stent type received and of concordance with their stent preference we further adjusted for stent type received. We obtained the data for this cross-sectional study from chart abstraction and patient interviews, without follow-up assessments, resulting in minimal missing data (≤5% for all outcomes). Statistical significance was denoted by 2-sided P values < .05. We performed analyses using SAS version 9.4 (SAS Institute) and R version 3.3.1 (R Core Team).
After IRB approval, we began the patient survey at both centers during the preintervention phase. Study coordinators received training on survey administration and clinical data collection. The survey ran concurrent with the focus groups developing the SDM tool. Following completion of the focus groups, the SDM tool was programmed into PRISM. Study coordinators at both sites then recruited decision coaches for training. At TMC, bedside nurses from the catheterization lab served as decision coaches; at MAHI, a similar plan to use bedside nurses was pursued, but personnel limitations meant that clinical research nurses served as decision coaches. The decision coach curriculum included a didactic portion, with a proposed script highlighting the Ottawa Decision Support framework that addresses decisional needs, decision quality, and decision support. The curriculum also included the use of motivational interviewing techniques and provided opportunities for role playing (see Appendix E). Training was conducted in a 2-hour classroom setting with video examples (https://www.youtube.com/watch?v=8oELU5vGNxw), then modeled by Beth Gialde RN, MSN, in a patient setting. Gialde then supervised initial coaching sessions to ensure consistency. Non-coach research nurses conducted postimplementation surveys. Decision coaching was conducted at MAHI from March 12, 2015, until February 10, 2016, and then attempted by bedside nurses from June 27, 2016, until September 5, 2016. Patient coaching was conducted at TMC from December 8, 2015, until December 16, 2016. The survey portion of the study concluded at both sites on December 16, 2016.
Findings from the patient focus groups included specific feedback on the visual representation and ranking of the 4 main DAPT drawbacks identified from patient and clinician feedback. We organized all coding according to the 3 main interview guide constructs: framing of the risks and benefits, discussion of the 4 DAPT drawbacks, and ranking of the value of each drawback as to the importance or impact on the individual's life (Table 2). Additionally, we held a clinician meeting, during which participants viewed the first and final versions of the SDM tool.
Representative Statements from Focus Group Participants Grouped by Topic.
Following completion of the focus groups and clinician meeting, having met saturation to the point that no new themes were identified during analysis of the final focus group transcript (see the “Analytical and Statistical Approaches” section), we posted the tool to the e-Patients.net website (website no longer active) for 10 days of review and comment before sharing a final version with the steering committee. We shared this with the steering committee and patient advisors for their input, with only minor edits being made—except in the case of 1 cardiologist on the steering committee (not an interventional cardiologist) who spent approximately 15 minutes reading and reviewing the tool and who subsequently encouraged additional text at the top of the page to stress that “there is no difference in survival or future heart attacks between the 2 types of stents.” This statement was added in boldface type. Once the tool was finalized (Figure 1), we sent the template to Health Outcomes Sciences for programming into PRISM. It was successfully tested for accuracy, appropriate visual display, and printing capabilities.
Draft of the SDM Tool.
Revised SDM Tool.
Final SDM Tool.
Following implementation of the tool and coaching at MAHI, we conducted clinician surveys (n = 30) at MAHI, but not at TMC because the only interventional cardiologist practicing there at the time was a co-investigator. Interventional cardiologists stated they used visual materials in 100% of their discussions with patients, although only 76% included a discussion of the stent type and elicited a patient preference. Most physicians (97%) responded that nearly three-quarters of the patients had no stent preference.
With the assistance of Health Outcomes Sciences, the entity that supports PRISM, we successfully programmed PRISM by adding the SDM tool template to the existing PRISM-generated, personalized informed consent form and populating the SDM tool with personalized TVR risk estimates with BMS and DES. In the postimplementation phase, at both sites, patients undergoing PCI received a paper-based SDM tool with their personalized consent form.
Patients in our focus groups reached general consensus that they would like information about their procedure as early as possible within the process of care. Therefore, we proposed to cardiology office administrators that the nursing staff should use the SDM tool in the outpatient setting at the time that coronary angiography is ordered.
Alternatively, having identified a gap in knowledge, we postulated that patients could be scheduled for another appointment before their procedure, to have an opportunity to speak to a clinician about any questions or concerns. After many discussions regarding possible implementation strategies, it was noted that the most pressing concern was the lack of consistency in the staff who communicated with patients about the need for angiography (often after the results of stress testing were available) and that the pool of nurses was too large to train and deliver the intervention consistently and with adequate fidelity. It was also not possible to add a visit before the procedure, which was estimated to add upward of 1000 visits per year, based mostly on the lack of providers and space.
The original intent of the study was to use bedside nurses as decision coaches to help patients better understand their preferences in medical care. After being unsuccessful with implementation in the cardiology office, we pursued training bedside nurses in the cath lab prep/recovery unit. After approval by nursing administration, we trained and certified 5 bedside nurses in decision coaching. These nurses attempted to coach for a period of 8 weeks; however, the unit was experiencing higher-than-usual staff turnover and was reliant on temporary nurse support, which decreased the time available for decision coaches to engage patients in SDM. Lack of support from interventional cardiologists was also reported as a barrier. Ultimately, we deemed these barriers too difficult to overcome, and we elected to have research staff conduct the coaching, with the intention that, if this proved successful, we could reengage hospital administration and nursing staff to implement the tool in routine care.
After concluding that bedside nurses would not be able to provide decision coaching, the study team trained designated research nurses as decision coaches, and 4 nurses became certified. This approach proved to be the most successful; however, it also had limitations and barriers. Because these nurses were not directly caring for the patients, they had to insert themselves into the process of care and find time with each patient to review the SDM tool and provide decision coaching. This was often difficult because patient care takes priority, and there is often little or no time for additional education/coaching. For example, some patients could not complete decision coaching due to other clinicians needing to speak with them, nursing care being carried out, and patients undergoing other testing (eg, echocardiograms).
At TMC, catheterization laboratory nurses served as decision coaches for the duration of the study.
The consensus of the study team, steering committee, and advisory committee was that a pre-post study design was most appropriate, largely because of concerns that if a randomized study were performed, there would be a “bleeding over” effect of the additional coaching and discussions to nonrandomized patients. Due in part to the barriers outlined above regarding the unavailability of nursing staff to serve as decision coaches, we conducted the study in 3 distinct phases: (1) preimplementation, in which patients undergoing coronary angiography and possible PCI received a PRISM-generated, personalized consent form with embedded personalized risk estimates for mortality, bleeding, and TVR with BMS and DES, but not the SDM tool; (2) postimplementation with decision coaching, in which patients received the PRISM consent form, the SDM tool for stent selection, and decision coaching from a nurse trained in motivational interviewing; and (3) postimplementation without decision coaching, in which patients received the PRISM-generated consent form and the SDM tool as an additional page of that consent form, with no decision coaching. This provided a unique opportunity to assess whether the coaching was important or if merely presenting a personalized SDM tool with the consent would suffice. We assessed experiences with stent selection in 336 patients not receiving the SDM tool, 113 receiving the SDM tool with decision coaching, and 137 receiving the SDM tool without coaching (Figure 2). Characteristics of patients enrolled in these 3 study phases were generally similar, although we observed statistically significant differences in site of enrollment, race, history of peripheral arterial disease, and admission status (Table 4). The standardized differences denote the difference in means (or proportion) between the 2 groups being compared, divided by their pooled SDs (and multiplied by 100 to represent percent of the standard deviation), which rendered a unitless measure of similarity that can be compared across variables with different scales.
Patient Enrollment.
Patient and Procedural Characteristics of Patients in the Preimplementation and Postimplementation Phases,.
Complete data for patient preferences about factors related to stent choice were available for 311 patients in the preimplementation phase. Information about the final stent placed was missing in 15 (4.7%) patients and could not be captured due to IRB requirements of working with deidentified data. We found that 14.4% of patients considered avoiding a repeat revascularization procedure as the single most important factor when deciding on a choice of stent, while 20.6% considered avoiding one of the DAPT drawbacks as most important (Figure 3). The remaining 65% of patients considered avoiding at least 1 of the drawbacks of DAPT as important as avoiding repeat revascularization, underscoring the importance of SDM. These results were recently published.24 Figure 4 shows the percentage of patients who ranked each consideration as one of their top concerns. Table 5 shows the characteristics of patients based on what they most valued.
Distribution of the Most Important Considerations to Patients in Selecting a Stent.
Distribution of DAPT Drawbacks That Patients Ranked Most Important to Avoid.
Factors Most Important in Stent Selection in Patient Subgroups.
Although statistically statistical differences in patient preferences were observed according to gender, race, and elapsed time since prior PCI, only a minority of patients in all subgroups valued avoidance of repeat revascularization as the single most important factor in stent selection. Among the 60.6% of patients who subsequently underwent PCI, 85% were treated with DES, and no difference in DES use was observed among patients who most valued avoiding repeat revascularization, who most valued avoiding DAPT drawbacks, or who valued both equally (78.7% vs 86.2% vs 85.6%, respectively, P = .56; Figure 5).
Final Type of Stent Placed Stratified by What Patients Valued the Most (Preimplementation).
Based on postprocedure interviews, patients who received the SDM tool and decision coaching were more likely to participate in decision-making regarding stent selection (60% vs 30%, P < .001) and to state that they had a stent-type preference (68.8% vs 35.4%, P < .001), compared with patients in the preimplementation phase (Figures 6 and 7). There was no difference in these parameters between patients in the preimplementation phase and those who received the SDM tool without decision coaching. In fully adjusted analyses (Table 6), patients who received the SDM tool and decision coaching were significantly more likely to participate in decision-making regarding stent selection (odds ratio [OR], 2.87 [1.84, 4.49]; P < .001) and to state a stent preference (OR, 1.91 [1.57, 2.33]; P < .001).
Patients' Participation in Decision-Making Regarding Stent Selection: Answer to the Question, “Who Chose the Type of Stent?”.
Patients' Stated Stent Preference.
Adjusted Analyses of Knowledge Transfer and SDM Items Answered Correctly.
Stent knowledge score (0-6) increased from 2.3 ± 1.4 in the preimplementation phase to 4.3 ± 1.5 in patients who received the SDM tool and decision coaching (Table 3; P < .001). The proportion of patients who answered all 6 questions correctly increased from 1.8% in the preimplementation phase to 24.8% in patients who received the SDM tool with decision coaching (P < .001). No differences were observed between patients in the preimplementation phase and those who received the SDM tool without decision coaching. In fully adjusted analyses, patients who received the SDM tool and decision coaching demonstrated a higher stent knowledge score (mean difference +1.8 [1.5, 2.1]; P < .001) and remained more likely to answer all 6 questions correctly (relative risk, 11.7 [4.9, 27.9]; P < .001). Similarly, patients who received the SDM tool with decision coaching reported higher nurse and physician discussion quality scores than patients who did not receive the SDM tool or those who received the SDM tool without decision coaching (Table 6).
Representative Quotes From Providers Regarding the Benefits and Limitations of the SDM Tool.
Figure 8 depicts patients' stated stent preferences and the concordance between stent preference and stent received. A greater proportion of patients preferred to receive DES as compared with BMS in all 3 phases of the study. DES were used in 86.7% of cases, and Figure 9 shows the concordance between stent preference and stent received stratified by stent preference, which demonstrates that DES use was nearly universal (98%) in patients who preferred DES and very high (85%) in patients who did not have a stent preference. Conversely, patients who stated a preference for BMS received DES 47% of the time. No differences in the concordance of preference and treatment were observed across phases of the study.
Patients' Stent Preference, and Concordance Between Stent Preference and Stent Received.
DES Use Stratified by Patients' Stent Preference.
Following completion of the implementation study (survey, coaching, and tool), we interviewed several interventional and general cardiologists to get their perspective of the implementation and value of the tool to patient-physician discussion. We also interviewed the decision coach. Table 3 presents representative quotes from the clinicians, grouped by topic.
The clinicians interviewed found both value and frustration with the tool and patient engagement. The nurse decision coach found value, though was she was frustrated with the clinical milieu of apathy for active patient engagement. Two clinicians had suggestions for improvement and enhancement of the tool as well as a comment cautioning that the tool itself should not replace discussion; they emphasized that SDM should occur only when the tool is used in conjunction with a discussion between the clinician and the patient.
In the setting of PCI and because of the offsetting benefits of DES and the drawbacks of prolonged DAPT, stent selection should be a preference-sensitive decision. However, DES are currently used in approximately 80% of PCI cases in the United States, regardless of patient risk for restenosis, and evidence suggests that patients are not typically engaged in SDM with respect to stent selection.
While decision aids have been demonstrated to increase patient knowledge and promote shared decision-making in patients with coronary artery disease and other conditions, they are inconsistently used and typically do not incorporate personalized risk information.25 Prior studies have also demonstrated the benefits of decision coaching—both as an alternative to usual care and as a supplement to patient decision aids.26 This study is the first to elicit and describe patient preferences regarding factors surrounding stent choice, and it provides a unique insight into whether the current use of DES is aligned with patient preferences—or whether more effort to support SDM is needed. In the preimplementation phase, we found that the majority of patients (65%) valued both the benefits of DES and the drawbacks of prolonged DAPT equally, and only a minority (14.4%) were most concerned about having to return for repeat revascularization. In fact, a similar proportion of patients (20.6%) were most concerned about DAPT drawbacks. Despite this evidence, we found no association between patients' value of DES benefits and DAPT drawbacks with the actual stent received. Thus, the predominant use of DES in this study suggests that physicians are prioritizing the avoidance of repeat revascularization procedures rather than engaging patients in SDM so that treatment can more closely align with patient preferences. These data underscore the need to more systematically elicit patient preference and engage patients in SDM before stent selection during PCI.
By developing and implementing a personalized, evidence-based SDM based on stakeholder feedback, we found significant improvements in patient knowledge, the process of SDM through discussions with doctors and nurses, and participation in and perceived quality of SDM when the tool was supplemented with decision coaching. In both unadjusted and adjusted analyses, patients who received the SDM tool and decision coaching were more likely to participate in decision-making regarding stent selection and to state a preference for the stent type they would like to receive, compared with patients who received neither. Of note, no differences in the proportions of patients receiving DES were observed, and concordance between stent preference and stent received remained modest (just over 50%) in patients who preferred to receive BMS.
Among the most important findings from this study is that when the SDM tool was provided without coaching, there were no differences in patients' participation in SDM or their likelihood to state a stent preference. These data clearly suggest that for a high-quality SDM to occur, a personalized, evidence-based SDM tool needs to be supplemented with decisional coaching. Given the benefits, from patients' perspectives, this study provides a justification for investing in the infrastructure of health care delivery to support decision coaching.
To align with the stated goals of the American Medical Association and National Academy of Medicine to honor patients' preferences and for patients to participate in treatment decisions, a transformation in the process of PCI and stent selection is needed. In this study, through collaboration with patients and patient advocates, providers, and professional societies, we developed a shared decision-making tool for stent selection to better educate patients about the benefits and drawbacks of DES and BMS and to better engage patients in selecting treatment. In addition, we trained nursing staff in motivational interviewing techniques so that they could serve as decision coaches and better engage patients in SDM. We found that patients who received the SDM tool and decision coaching were significantly more likely to participate in SDM with respect to stent selection, to voice a preference for which type of stent they would prefer to receive, and to retain information about stents and stent types. Nevertheless, participation in SDM did not significantly impact physicians' use of DES and BMS, and what is particularly striking is that while concordance between stent preference and stent received was impressively high (98%) for patients who preferred to receive DES, it was significantly lower (53%) for patients who preferred to receive BMS. These findings suggest that even when patients have participated in SDM and voiced a preference for treatment, physicians are likely to resort to their own preferences when making final treatment decisions.
Implementation of the SDM tool into the context of clinical care was more difficult than anticipated. The barriers to implementation largely related to limitations in office and hospital personnel, space, and time. Ultimately, the decision to use research nurses as decision coaches at one of the sites allowed us to conduct the study and to analyze the impact of the SDM tool and decision coaching on patient engagement in SDM, knowledge transfer, and stent selection. However, if the results of this study are to be replicated in clinical practice, the same barriers of personnel, space, and time must be addressed. Innovative methods to overcome these barriers could involve telemedicine, videotaped or web-based educational presentations, which could prepare patients for more limited face-to-face discussion of the personalized information provided in the SDM tool, and communication of their preferences. In any scenario, resources to provide nurse and physician training in motivational interviewing should be considered, to educate providers on the benefits of SDM and to train them on methods to best engage patients in SDM.
Despite input from patients, patient advocates, and providers that the SDM tool should be reviewed with patients as far “upstream” of their procedure as possible, to allow patients to review and ponder the relevant information, we were unable to implement the SDM tool into patient care on an outpatient basis, again due to barriers of personnel, space, and time. Innovative methods to overcome these barriers could involve separating the provision of the SDM tool from decision coaching; for example, the SDM tool could be mailed or e-mailed to the patient or viewed online by the patient ahead of the procedure and reviewed on an outpatient basis, while decision coaching and elicitation of patient preferences could be performed on the day of the procedure, as in this study.
Given that we observed no benefits to providing the SDM tool to patients without decision coaching in this study, we feel that if the SDM tool is to be implemented in clinical practice with the goal of engaging patients in SDM, considerable effort may be needed to convince clinic and hospital administrative bodies to provide the resources necessary to provide decision coaching to patients—and to provide this information to patients upstream from their procedure. This could be accomplished by payers providing a payment to offset the resources required to redesign care, but the payer representative on the steering committee indicated that there would be little interest from payers to do so.
The 2 sites of enrollment for this study, Saint Luke's Mid America Heart Institute and Truman Medical Center, serve different patient populations and were strategically selected so as to encourage recruitment of patients with a broad range of demographic and socioeconomic characteristics. The majority of patients enrolled in this study were Caucasian and male, with a high prevalence of traditional cardiovascular risk factors, as expected. However, the percentage of women enrolled in this study is higher than in most studies, and we enrolled a small but significant proportion of African American patients as well. These findings suggest that the strategy to recruit a diverse group of patients was successful.
Accordingly, it was of interest to evaluate the impact of the SDM tool and decision coaching on primary and secondary outcomes in these diverse patient groups (Table 7). There was little observed impact of enrollment site on knowledge transfer and patient participation in SDM, both of which improved substantially with provision of the SDM tool and decision coaching at MAHI and TMC. The proportion of patients reporting that they discussed stent types with nurses and physicians also improved at both sites, although the proportion of patients reporting a discussion with the physician was somewhat higher at TMC than at MAHI. The majority of patients who voiced a stent preference preferred DES at MAHI, while the majority of patients at TMC preferred BMS. At both sites, the proportion of patients voicing a stent preference increased significantly following provision of the SDM tool and decision coaching. DES use was higher at MAHI than at TMC, and concordance with patient preference was similarly high at both sites among those preferring DES, with no significant differences observed after provision of the SDM tool and decision coaching. This suggests that the selection of a diverse population to inform development of the SDM tool resulted in a tool that was similarly efficacious across a diverse population of patients, supporting the generalizability of our approach.
Site-Specific Comparisons.
Because decision coaching could not be inserted into existing processes of patient care at MAHI (coaching was performed by a research nurse rather than by clinical staff, in contrast to what was done at TMC), generalizing our findings of the benefits of the SDM tool with coaching requires that a sustainable infrastructure for coaching be developed—ideally in an outpatient consultation with a patient's provider before a PCI is scheduled.
However, despite our desire to provide the coaching upstream, doing so just before treatment is a pragmatic time to elicit patient preferences and is a natural part of the conversation, even if not done well in usual practice. The information contained in the SDM tool was current at the time of the study but should be updated as data change. For example, information regarding the cost of medications should be updated periodically as costs change. Similarly, the recommended duration of DAPT following DES and BMS needs to be updated as guidelines change. Furthermore, the US FDA has approved novel stent platforms such as bioabsorbable vascular scaffolds, and the association of these technologies with TVR should be included, as the data emerges. Finally, if newer risk models for restenosis with DES and BMS or bleeding complications with DAPT27 are derived in the future, then the SDM tool can be updated to incorporate the most up-to-date information.
Further subgroup analysis of the benefits of these interventions may provide additional information on the benefits of the SDM tool and decision coaching as a function of age, gender, race, and socioeconomic status. However, such analyses have not yet been performed.
The first limitation of this study is that the SDM tool was provided to patients in the catheterization laboratory prep area instead of beforehand. As outlined above, patients and providers in focus groups noted that it would have been preferable to provide this information further upstream from the time of the procedure. On the one hand, had we been able to provide the information sooner, patients may have had more time to reflect on the information and become even more engaged in SDM. On the other hand, it is possible that if the discussion was held too far in advance of the procedure, patients would forget the information over time and knowledge transfer would have been negatively impacted.
The second limitation of this study is that the resources to provide decision coaching to patients in this study were limited, particularly by the frenetic workload of bedside nurses at MAHI, which made it difficult for them to prioritize their opportunity to support the education of patients for SDM. As a result, patients in the postimplementation phase did not receive decision coaching if a nurse was unavailable at that time. While it is possible that relying on coaching to have occurred may have resulted in a selection bias (physicians and nurses more committed to SDM may have better engaged patients in shared decision-making), the need to supplement usual care with a decision coach at MAHI and the lack of any observed benefits of providing the SDM tool without decision coaching minimizes the possibility of such bias.
A final limitation of this study is that providers may not have been fully engaged in the process of SDM with patients in this study. Physician follow-up was not directly obtained, and as a result, we could not directly evaluate factors associated with stent concordance or discordance, such as the presence or length of discussion regarding stent selection. Similarly, the study did not record the length of decision coaching and the length of time that patients spent reviewing the SDM tool.
While the results of this study appeared largely positive, future research should focus on implementation strategies to overcome the barriers of personnel, location, and time outlined above. Novel implementation strategies might allow for this information to be shared with more patients. Our finding that stent concordance was 98% for patients who preferred to receive DES and only 53% for patients who preferred to receive BMS suggests that physicians are prone to use DES and may be reluctant to change their decisions when patient preferences differ from their own. This finding warrants further investigation to overcome such physician biases and to better deliver treatment that is preferred by patients.
Given the growing demand for patient engagement in medical decision-making, we described, from the patient perspective, the relative importance of different factors related to stent choice. We found that most patients regarded avoiding the drawbacks of DAPT at least as important as avoiding TVR, and only a minority regarded avoiding TVR as the single most important consideration when selecting a stent. The prominent use of DES, which necessitates a longer duration of DAPT, therefore appears to signal that physicians' goals of therapy may not always be the same as patients' and underscores the importance of explicitly eliciting patients' preferences to better tailor stent choices to the goals and values of individual patients.
To improve the process of stent selection, we developed an SDM tool for BMS vs DES through collaboration with patients, patient advocates, providers, and professional societies, with the goal of better informing patients and engaging them in SDM, and we successfully trained nurses in motivational interviewing techniques so that they could serve as decision coaches. We met substantial barriers in the implementation of the SDM tool, mainly with respect to resources available to provide decision coaching to patients.
Nevertheless, we found significant improvements in the processes of—and patients' engagement in—SDM as well as greater knowledge transfer when the SDM tool was provided to patients along with decision coaching. However, no benefits were observed when the SDM tool was provided to patients without decision coaching. No significant differences in the proportion of DES and BMS were observed in the study, and while patients who preferred to receive DES nearly always received DES, patients who preferred BMS often received DES, suggesting that physician-level barriers to SDM may exist. Resources to educate providers on the benefits of SDM and to train providers on methods to best engage patients in SDM, along with investments by hospitals in supporting decision coaching, are likely necessary to further improve the quality of SDM with respect to stent selection at the time of PCI.
Research reported in this report was [partially] funded through a Patient-Centered Outcomes Research Institute® (PCORI®) Award (#CE-1304-6448) Further information available at: https://www.pcori.org/research-results/2013/does-information-benefits-and-risks-help-patients-decide-between-stent-options
Patient Focus Group Discussing Guide (PDF, 153K)
Study Timeline (PDF, 264K)
DECIDE Decision Tool (PDF, 168K)
Patient Survey Case Report Form (PDF, 270K)
Decision Coach Curriculum (PDF, 518K)
Spertus JA, Chhatriwalla A, Decker C, Jones PG, Gialde E. 2019. Does Information on Benefits and Risks Help Patients Decide Between Stent Options? Patient-Centered Outcomes Research Institute (PCORI). https://doi.org/10.25302/5.2019.CE.13046448
The [views, statements, opinions] presented in this report are solely the responsibility of the author(s) and do not necessarily represent the views of the Patient-Centered Outcomes Research Institute® (PCORI®), its Board of Governors or Methodology Committee.
This book is distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License which permits noncommercial use and distribution provided the original author(s) and source are credited. (See https://creativecommons.org/licenses/by-nc-nd/4.0/
Your browsing activity is empty.
Activity recording is turned off.
See more...
