An official website of the United States government
NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
Jennifer Blumenthal-Barby, PhD, Kristin Kostick, PhD, Courtenay Bruce, MA, JD, Robert Volk, MD, Brian Bruckner, MD, PhD, Matthias Loebe, MD, PhD, Jerry D. Estep, MD, Sheryl Ann McCurdy, PhD, and Charles Minard, PhD.
Author Information and AffiliationsStudies indicate that decision-making among patients considering treatment for advanced heart failure (HF) could be improved. Decision aids (DAs) increase knowledge, decrease decisional conflict, increase accurate risk perceptions, increase match between values and choice, and decrease the use of some surgical intervention and screening tests. Left ventricular assist device (LVAD) therapy is the fastest growing treatment for advanced HF. Recognizing this, the American Heart Association recently emphasized the “crucial” need for improved informed consent processes for shared decision-making about LVAD.
Our goal was to develop and test a patient-centered DA for decision-making about receiving an LVAD for advanced HF treatment, using the International Patient Decision Aids Standards (IPDAS) Collaboration guidelines, input from patient stakeholders, and advice from clinical experts.
We developed the DA in line with the Ottawa Decision Support Framework and IPDAS for ensuring quality, patient-centered content. We conducted semistructured open-ended interviews with patient candidates for LVAD treatment (n = 15), patients currently with LVADs (n = 15), caregivers of patients with LVADs (n = 15), LVAD decliners (n = 15), and expert clinicians (n = 11) to generate content based on patient values and decisional needs, and providers' perspectives on knowledge needs for informed consent. Content also included short- and long-term risks and benefits based on the most recent literature. We also received ongoing feedback from patient partners (n = 3) throughout development and testing of the DA. We alpha-tested the aid through cognitive interviews (n = 5) and tested acceptability with LVAD patients (n = 10), candidates (n = 10), and clinicians (n = 13) (eg, cardiologists, cardiothoracic surgeons, social workers, LVAD coordinators, and bioethicists) from 5 LVAD clinics in the United States.
We conducted a multisite randomized trial of the DA above standard education alone among inpatients considering LVAD treatment for advanced HF (1:1 allocation ratio). We tested superiority of our DA using an independent, 2-sample t test of mean differences in our main outcome of LVAD knowledge using a validated scale. Participants included inpatient LVAD candidates at 5 LVAD clinics. We collected data at 1-day, 1-week, and 1-month follow-up.
Of 105 eligible patients, 98 agreed to participate (including both bridge to transplant and destination therapy candidates) and were randomly assigned to the DA or standard education arms. Of the 52 in the DA study arm and the 53 in the control arm, the rate of follow-up contact was 58% and 64% at 1 week, respectively, and 52% and 58% at 1 month, respectively. Patients in the DA arm exhibited significantly greater LVAD knowledge than the standard education group at 1-week follow-up (68% vs 59%; P = .01), but not at 1-month follow-up (64% vs 61%; P = .47). We found no differences between DA and standard education patients in rates of acceptance vs decline of LVAD treatment (85% vs 78%; P = .74). Secondary outcomes show that recipients of the DA reported greater satisfaction with life postimplant than nonrecipients (28 vs 23 out of 30; P = .008), although both arms reported high satisfaction. Patients rated the DA high in acceptability and usability, with most patients reporting that the DA helped them understand more about risks and benefits of LVAD treatment (94%), the DA helped them understand their options for dealing with HF (85%), they learned something new (83%), and they believe the DA helps someone to make an informed decision about LVAD treatment (94%).
Our findings indicate that our DA enhances LVAD knowledge during a crucial decision-making period, and that the aid is neutral (ie, does not encourage direction of decision) and reflects patient, caregiver, and physician preferences for content and format.
Results may differ in clinics that lack the experience and infrastructure of our partnering study clinics. Low follow-up (59% across arms) is an additional limitation.
Left ventricular assist device (LVAD) therapy is the fastest growing treatment for advanced heart failure (HF), with the number of LVAD implants increasing 5-fold between 2009 and 2016, and up to 2500 patients implanted per year.1 This trend is expected to continue2,3 particularly as heart transplantation, the optimal life-prolonging therapy for end-stage HF, is often not an option due to high demand for organs in relation to supply and because of restrictions on recipient age.1,4 The LVAD constitutes an effective, evidence-based treatment used not only as a bridge to transplant (BTT) but also increasingly as destination therapy (DT). For patients who do not use LVAD either to keep them alive until they can get a transplant (BTT) or to make a decision (bridge to decision [BTD]) about whether they want to undergo a transplant or keep the LVAD as a permanent solution to HF (DT), other options, primarily palliative and supportive care, are available. While more HF patients are receiving LVADs (with the number of implants in the United States rising from 2317 implants in 2008-2012 to 5072 implants in 2013-2016),1 many who consider LVAD therapy believe they do not have a choice about whether to accept LVAD treatment and view it as their only option for extending life in the absence of heart transplant availability.5,6 Further, recent studies also indicate poor patient understanding about the capabilities, lifestyle implications, and risks of an LVAD (especially long-term and psychosocial risks).7-9 In one study, 80% of patients “could not recall a discussion of perioperative risk or potential long-term complications.”1,7-9 Taken together, these findings suggest that current approaches to informed decision-making about LVAD treatment need improvement.10
An inadequate decision-making process for LVAD implantation can engender misalignments in expectations and goals among HF teams and patients and their surrogates, creating several ethical challenges. For instance, there is evidence that due to inadequate informed consent processes, some patients do not grasp the nature and likelihood of possible trajectories and outcomes, sometimes leading to requests for device deactivation (and resulting complications that include blood backflow, pooling, regurgitation, thrombosis, and reduced heart contractility) that create moral distress for patients and their care teams.11-13 Additionally, there is evidence that, for patients who do not do as well as expected, surrogates may resist discussions about forgoing life-sustaining therapies, citing preintervention discussions focused on transplantation and the term bridge to transplant.11-13 The exponential increase in LVAD implantation further exacerbates the magnitude of suboptimal informed consent issues and renders the need for decisional support for LVAD treatment more urgent than ever.1
Recognizing these issues, the American Heart Association (AHA) recently published a Scientific Statement about decision-making related to advanced HF, emphasizing the “crucial” need for improved informed consent processes, increased shared decision-making, and development of decision aids not tied to any particular device company.14 This statement was endorsed by the Heart Failure Society of America, American Association of Heart Failure Nurses, and Society for Medical Decision-Making, underscoring the importance and timeliness of developing decision support resources for LVAD candidates.
Patient decision aids (DAs) are tools developed to aid patients in making health care decisions in instances when an optimal course is uncertain and patients' individual preferences are central to decision-making.15,16 DAs are now usually web based, or also come as leaflets or booklets, computer programs, or DVDs,17 and may be accessed repeatedly (important for patients who are overwhelmed or have limited recall). DAs aim to present outcomes, risks and benefits, and uncertainties in a clear, comprehensive, scientifically valid, and unbiased manner to help people make value-based health decisions.17 Most also include patient stories or quotes and values clarification exercises, which help patients identify their core values and apply them to the decision at hand. A recent Cochrane Systematic Review concluded that DAs increase knowledge, decrease decisional conflict, increase accurate risk and benefit perceptions, increase match between values and choice, and decrease the use of some surgical intervention and screening tests.15,16 The International Patient Decision Aids Standards (IPDAS) Collaboration has developed a protocol and set of criteria for development, quality, and effectiveness of DAs through evidence reviews (last updated 2012).18-20
Patient DAs help address unwanted variation in preference-sensitive care, defined as treatments for conditions for which legitimate treatment options exist that involve tradeoffs among different possible treatment outcomes (eg, some patients will prefer to accept a small risk of death to improve function, others will not).21 Such variation is especially prevalent in the area of surgery. For example, dramatic geographic variation has been found in the cases of lumpectomy vs mastectomy for early stage breast cancer, coronary bypass vs angioplasty for chronic angina, disc surgery vs spinal fusion for low back pain, joint replacement for arthritis of the knee or hip, and prostatectomy vs active surveillance for early stage prostate cancer.21 This variation occurs because surgeons may recommend surgery on the basis of subjective opinion, different levels of expertise, availability of or access to options, personal experience, or anecdotal evidence, and because patients often defer decision-making to physicians.21 DAs integrate, and present in an understandable way, the most recent data from rigorous scientific studies, ensuring that all patients facing the treatment decision are making a decision based on the same (accurate) information. They also replace deferred decision-making with shared decision-making and informed consent with informed patient choice. This shift to shared decision-making rather than deferred is important, as studies show that patients often choose very differently from their physicians when they are fully informed of their options.21 Other studies provide evidence highlighting the value of DAs and their ability to eliminate wasted time and resources.15,16
LVAD placement is a value-laden, preference-sensitive treatment choice, and among the most expensive interventions in medicine, at $500 000 to $1.4 million per quality-adjusted life-year.22,23 Aside from cost, additional decision-making considerations include length and quality of life (QOL), convenience, preservation of bodily integrity, body image, limitations and changes in activities of daily living and functional or exercise capacity, risk of death or other adverse events, and impact on familial and other relationships.14,24-29 Adverse events and complications from LVADs typically include ischemic or hemorrhagic neurological complications; refractory LVAD-related infections; sepsis; and hepatic, renal, or respiratory failure.30-32 Adequate decisional support could help patients and caregivers become more aware of the potential positive and negative impacts on lifestyle and longevity. Further, a DA that presents comprehensive information about LVAD should contextualize LVAD as one choice among others, including palliative and supportive care options.
Three noteworthy systematic reviews (by the Quality Enhancement Research Initiative; an Evidence-based Synthesis Program by the Department of Veterans Affairs, the Interagency Registry for Mechanical Assisted Circulatory Support (INTERMACS), and the AHA) provide supporting evidence that DAs for LVADs would improve patient experiences, eliminate costs associated with unwanted treatment, and enhance informed consent processes.14,33,34 However, no DA has been trial tested and is ready for dissemination, although one DA for LVAD is currently being tested.35 Standard education typically consists of viewing educational pamphlets created by device manufacturers, and speaking with an LVAD provider, who describes the device and answers questions. The traditional informed consent process may also include viewing and manipulating the actual device and meeting a patient with a device already implanted.36 However, these practices are institution specific and unstandardized across clinics, lacking in quality assurance in decision support.37
We have responded to the gap in decision support for patients considering treatments for advanced HF by developing a DA to enhance informed patient choice. While 1 other LVAD DA exists,36 ours is the only DA for LVAD that has been efficacy tested and is ready for dissemination and implementation into clinical practice. Through our formative/qualitative research we aimed to collect patient, caregiver, and physician preferences regarding format and content of the DA as well as expert clinician perspectives on content criteria for ensuring informed patient choice (results published elsewhere).5,39,40 We then tested the DA in a randomized controlled trial aimed to assess the impact of our DA on informed patient choice. We hypothesized that our DA would enhance patients' knowledge about LVAD and alternative options, thereby enhancing a major component of quality decision-making. Results of this trial are presented below.
We involved patient and caregiver stakeholders (LVAD candidates, patients, caregivers, and LVAD decliners) both as research partners and as study participants at all stages of our research. In addition to the 40 patients and caregivers who offered their perspectives during in-depth interviews in aim 1, we invited patient stakeholders to offer their perspectives throughout the entire study. These stakeholders formally included 3 existing LVAD patients and informally included 1 patient's caregiver, and all who attended the meetings. Two of these stakeholders had experienced significant adverse events with their LVAD, including 1 patient who had initially declined LVAD and was able to explain why. The inclusion of these patients as stakeholders offered a nuanced perspective of patient decisional needs and an understanding of diversity in LVAD experiences. In the formative stages (aim 1), these stakeholders provided feedback about their knowledge and decisional needs related to LVAD therapy, and helped us inductively identify key areas of knowledge and value-based information to include in our primary outcome measure (LVAD knowledge) and in the overall DA. Face-to-face meetings with this group of key stakeholders on a bimonthly basis helped us better understand patients' perspectives and experiences with the decision-making process, including the logistics of achieving LVAD candidacy, engaging in shared decision-making, and arriving at a treatment decision. They also helped identify preferences (eg, timing, context) for receiving the DA in the context of clinical flow.
Combining patients' and caregivers' input with that of expert clinicians, we were able to create patient-centered study procedures, protocols, and outcome measures with optimal construct validity and with content, comprehensibility, and usability features tailored to our study population and balanced with clinical expertise.
During the course of our working relationship with key stakeholders, we attended patient appreciation events hosted by our main partnering institution (Houston Methodist Hospital), where we spoke with patients and caregivers and gained further insights into the clinical and social contexts in which our DA was operating. Feedback from patients and caregivers in these environments helped us formulate a bigger picture of the study context and begin considering important features of our target population for dissemination and implementation activities. For example, during our aim 1 research activities, patients and caregivers repeatedly mentioned the importance of patient-to-patient (peer-to-peer) interaction, particularly between LVAD candidates and LVAD patients with first-hand experience with the device. Witnessing these interactions in vivo helped us appreciate their significance and impacts. These meetings with the community also gave us an opportunity to debrief and describe to patients and caregivers how their input and feedback were being actively incorporated into our study materials and DA development.
During the efficacy trial, we met regularly (every few months) with our patient and stakeholder partners. This panel comprised 4 key clinical experts, including 2 cardiothoracic surgeons, 1 HF cardiologist, and 1 ethicist specializing in transplant ethics. Other clinicians, including surgeons, cardiologists, LVAD coordinators, nurses, social workers, and ethicists at our partnering hospitals were also sometimes present. The purpose of these meetings was to discuss ongoing study activities and to elicit feedback about the trial, to address any concerns from those administering the study materials to patients (LVAD coordinators and research assistants), and to monitor overall progress.
Toward the end of the trial, we conducted end-of-study interviews with patients (n = 9) and clinicians (n = 11) to gather further feedback about their experiences using the DA in practice. Their insights helped us understand how the DA fits into clinical flow (we found that it enhances flow rather than slows it down) and which clinic personnel are best suited and preferred by patients to administer the DA in practice (patients preferred LVAD coordinators, with systematic support from clinicians).
We feel that our close engagement with patients, caregivers, and clinicians improved our understanding of factors affecting receptivity and buy-in to implementing our DA in practice. The credibility and validity of our findings rest in our solid engagement with patients, caregivers, and clinician stakeholders at every stage of our research. The strong foundation of our study findings in patients' subjective needs, experiences, and perspectives contributes to the enduring usefulness of our DA for other patients, caregivers, and clinicians involved in decision-making about LVAD treatment.
We conducted semistructured, in-depth qualitative interviews with patients (including DT, BTT, and BTD patients), caregivers, candidates, and decliners of LVAD treatment, to assess users' decisional needs from their own perspectives. These questionnaires are available in the supplementary materials from our other publications presenting our qualitative research.5,41 To generate the question items for these interview protocols, we met with our stakeholder partners, including 3 LVAD patients and 1 caregiver, to generate key questions to elicit decisional needs and knowledge preferences. Research staff used outpatient phone numbers on a patient roster to call (using an IRB-approved phone script) and schedule formative qualitative interviews with eligible patient and candidate subjects for either before or after their clinical appointment with their LVAD doctors. We also asked patients whether their caregiver would be interested in providing a separate interview. We did not treat an arrangement with a patient, candidate, or caregiver over the phone as consent to be interviewed.
We also conducted in-depth interviews with clinicians, including cardiologists, cardiothoracic surgeons, LVAD clinic team coordinators, and palliative and supportive care specialists involved in shared decision-making, to better understand patient informational and decisional needs from a clinical perspective. We used separate interview protocols for each interviewee type. Interview domains across types included perceptions of options, outcomes, and probabilities; values in decision-making; degree of decision-making difficulty and factors contributing to difficulty; usual and preferred decision-making roles; and decisional barriers and facilitators. We used a technique called progressive focusing whereby we modified interview questions iteratively throughout the process of data collection, so that question items with diminishing informational returns were gradually replaced by questions eliciting new information from patient narratives.42
The finalized, paper-based DA presents information about viable alternatives (with corresponding risks and benefits presented side by side), including palliative and supportive care, to clarify that patients have agency in decision-making. The aid also clarifies requirements for receiving a transplant, addressing potential misconceptions or unrealistic expectations among patients who might consider LVAD only as a BTT. Contents are conveyed at an eighth-grade reading level. Other sections of the DA include information on HF in general and LVAD specifically; information on how to decide in line with one's own values; details about presurgery and postsurgery, including what to expect about living with an LVAD and potential adverse effects; narratives from other patients and caregivers living with LVAD; potential questions to ask doctors and other LVAD patients; and resources for patients and caregivers.
The final version of the DA can be found at www.lvaddecisionaid.com.
Incorporating information from all phases of formative research, we developed a first draft of the DA, following IPDAS quality indicators.19 To test users' perceptions and understanding of the content, relevance, and readability of the DA, we first conducted cognitive interviews or “think aloud” exercises with patients (n = 5) and acceptability tested it with LVAD patients, who were asked to give qualitative feedback on each section. We made changes iteratively based on consensus in feedback. Next, we administered validated quantitative questionnaires to patients (n = 10), candidates (n = 10), and clinicians (n = 13) (eg, cardiologists, cardiothoracic surgeons, social workers,52,53 LVAD coordinators, and bioethicists) from 5 LVAD inpatient/outpatient clinics in the United States. We used these questionnaires to evaluate perceptions of the DA's acceptability,16 covering aspects of each section's usability, likability, informative effectiveness, audience appropriateness, format, timing, range of content, and any open-ended suggestions for improvement from participants.
This phase also included the development of an LVAD Knowledge Scale for inclusion in the DA as a tool for patients and caregivers to assess their knowledge about LVAD therapy. Development and validation of this scale is described in our publication in The Journal of Heart and Lung Transplantion.40
We developed question items for the scale based on decisional needs identified by participants and clinician recommendations about essential components of informed decision-making for LVAD therapy.
Using the finalized version of the DA and the validated LVAD Knowledge Scale, we conducted a randomized controlled trial with a parallel design and an allocation ratio of 1:1 (50% intervention group; 50% control group). We made no changes to our methodology after trial commencement. However, we did involve additional sites beyond those stipulated in the original study protocol to increase our sample size, with an original projected sample size of 144. We included a total of 5 sites (2 LVAD clinics in Houston, TX, and 1 each in Cleveland, OH; Oklahoma City, OK; and New Orleans, LA). These sites serve both inpatients and outpatients, and were chosen based on the presence of an established LVAD clinic with high patient volume and their informal associations with physicians at our primary partnering hospital (Houston Methodist Hospital). Patient populations across sites did not differ significantly on either demographic or clinical characteristics, and our specific sample across sites included inpatient DT, BTT, and BTD candidates (although these designations are rarely made by clinicians at the decision-making stage). Figure 1 presents the overall patient flow, along with data collection time points at 1 day (immediately after patient reviews DA), 1 week, and 1 month.
LVAD DA Trial Participant Flow Diagram.
All participants were recruited through LVAD coordinators at each participating hospital, who contacted them via telephone using existing patient rosters at each clinic. Eligibility criteria mirrored those widespread for LVAD implantation:
There were no additional exclusion criteria. We collected data from participants within the LVAD clinics at which they were currently being treated, including 1 primary LVAD clinic and 4 partnering LVAD clinics across the United States. LVAD coordinators and participating clinics obtained all clinical characteristics, including baseline and outcomes data, via secure email.
Study procedures were approved by the IRBs at all participating research sites. Senior LVAD cardiologists and cardiosurgeons at each site were contacted for collaboration through the senior cardiologist at our primary participating site (Houston Methodist), and not only gave their consent to host our study but also actively helped facilitate coordinator involvement and administration of our study materials and DA. We consulted all clinicians for their approval and feedback on the DA before the study began. All recruiters at our primary site were credentialed by the department of Cardiovascular Sciences at Houston Methodist Hospital, involving the completion of Human Biomedical Research and a Conflict of Interest course. Recruiters at the other 4 sites consisted of already-credentialed LVAD coordinators. To identify potential participants, our research staff, including coordinators at all 5 sites, referred to an appointment schedule and LVAD Patient Census that was distributed to the transplant/LVAD team via email 3 times a week. Each email detailed the LVAD inpatients at each clinic and their room number. Research staff and coordinators recruited these inpatients and their caregivers by approaching them directly on site.
Once identified, capacitated inpatient LVAD candidates completed standard institutional LVAD education. They then met with the LVAD coordinator or other research staff, who obtained their consent to participate. At this time, patients were randomized (before baseline) and administered Baseline Study Packet 1, which includes a survey of all baseline measures and a patient demographics questionnaire. Patients randomized to the intervention arm then received Baseline Study Packet 2, which included the LVAD DA and the Ottawa Acceptability Scale. Envelopes containing packets 1 and 2 (control vs intervention) were identical. The LVAD candidate was asked to read the DA and then complete the Acceptability Scale. After completion, the study staff or coordinator collected the study packets for data entry (at our primary site) or the site's LVAD coordinator faxed the completed surveys to the secure fax line in the Center for Ethics at BCM.
To monitor progress, our research staff at BCM contacted the LVAD coordinator educator and/or the remote research assistant at our partnering hospitals 2 to 3 times a week to see if any eligible patients were ready to be educated that day or later in the week. Before patients underwent education procedures, the research assistant obtained their informed consent and administered project materials.
All participants were administered a 1-week follow-up and 1-month follow-up postrandomization. Study staff completed the 1-week follow-up either in person or over the phone 7 days after baseline contact. Candidates scheduled to be implanted before the 1-week follow-up had their 1-week follow-up the day before LVAD implant. We chose the 1-week follow-up because this is typically the time frame within which patients make their decisions and receive their chosen treatment. Study staff completed the 1-month follow-up over the phone, 30 days after baseline contact. Patients who were still in the hospital had the 1-month follow-up on-site. We chose the 1-month time frame because of informal indications from clinicians that patients usually form strong opinions about their treatment decisions after 1 month. This time frame also conformed to the timeline and budget constraints of our project. No follow-ups occurred after the study completion date. All participants of aim 1 and aim 2 received a $25 Kroger gift card by mail.
Patients in the intervention group received a paper copy (with content written at an eighth-grade level) of the DA from their LVAD coordinator, who informed them (and their caregivers, if present) of the purpose of the aid. Patients and caregivers were encouraged to take notes and to bring any questions to their appointments with their clinical teams. Patients and caregivers were also directed to the website lvaddecisionaid.com, where they could view an online version of the aid, as well as watch supplementary videos created by our research team in collaboration with an experienced videographer. All participants in this group also received standard education as well as the informed consent process provided by their LVAD clinic (see below).
Patients in the nonintervention group received the standard education process, which is institution-specific and unstandardized across our participating LVAD clinics. Standard education typically consists of viewing educational pamphlets created by device manufacturers, and speaking with an LVAD coordinator, who describes the device and answers any questions LVAD candidates have. The traditional informed consent process may also include viewing and manipulating the actual device and meeting a patient with a device already implanted.
Table 1 presents each of our study measures for aim 2 and our basis for choosing them. Our primary outcome measure was patients' LVAD knowledge at 1 week and 1 month, using a scale developed and validated by our research team.40 While we do not impose a cutoff to determine what constitutes “enough” knowledge, we approach the results with the perspective that more knowledge on the continuum between 0 and 100 is better. We selected our secondary measures in accordance with IPDAS recommendations, which call for validated measures of key aspects of quality decision-making, including the following:
Primary and Secondary Outcome Measures.
We also included as a secondary measure whether participants intended to choose vs actually chose LVAD as their preferred treatment, as well as the match between intentions and ultimate choices (preferred treatment). Table 2 presents time points for administration of all measures during the trial. All measures were administered by LVAD coordinators at various time points throughout the course of patients' review and decision about LVAD treatment.
Time Points for Administration of Study Measures.
For aim 1 (qualitative analysis), we determined sample size according to thematic saturation, defined as diminishing returns in novel information with each subsequent interview. We reached saturation after interviewing 15 participants in each subgroup (patients, caregivers, candidates, decliners). Interviews were audio-recorded, transcribed, and analyzed collaboratively using the well-established Grounded Theory22 method of analysis in ATLAS.ti, a computer-assisted qualitative data analysis software.
For aim 2 (quantitative trial analysis), we determined sample size by conducting power analyses for various sample sizes per study group (see Figure 2 for results). We calculated that a number of ~29 per group (58 total) would be sufficient to detect a mean difference of 15 points on a 100-point scale with 80% power. We selected the 15-point difference threshold because, on a scale with 20 question items, a 15-point difference would indicate a difference of 3 correct answers—that is, 15 correct answers in 1 study arm vs 12 correct answers in the other arm. We further selected the 15-point difference on the basis of our power analysis showing that detecting anything smaller than a 15-point difference would require a sample size that was beyond the means of our study resources. It should also be noted that our study was originally powered for a sample of 144, but a study with 105 participants has the same ability to detect a 15-point difference as 1 with 144 participants. We did not conduct any interim analyses with smaller subgroups.
Results of Power Analysis.
In line with recommendations from the National Center for Biotechnology Information,45 we used an online statistical computing program (available at www.graphpad.com/quickcalcs) to generate our randomization schedule using a “block” randomization design, in which the blocks were balanced with predetermined group assignments (with a nonpermuted block size of 4). Our statistician generated the allocation sequence and assigned participant numbers to 1 of the 2 study arms before receiving baseline study materials. We then passed these allocations on to the LVAD coordinators to implement the allocation sequence by enrolling participants using the predetermined vector of participant numbers, which in turn determined which set of study materials the patient would receive.
While we made every attempt to ensure a 1:1 allocation ratio, differing rates of attrition resulted in slightly unequal numbers across study arms.
Using allocation concealment, we ensured that the research staff and LVAD coordinators enrolling participants did not know in advance which study arm the next participants would be assigned to. We did this by ensuring that both study arm materials were provided in identical, sealed envelopes that would be opened only by study participants. The rationale behind our allocation concealment method was that those who handed out the study materials might consciously or unconsciously transmit bias or interact differently with study participants in ways that our research group could not predict or control. Allocation was concealed at baseline but not at 1 week or 1 month out of logistical necessity, given that coordinators were often the ones administering both the survey materials and the DA intervention. After baseline measures were administered, coordinators had to know who received the DA to effectively facilitate questions and discussion between patients and clinical staff.
We conducted a trial to test the superiority of receiving the DA over standard LVAD education related to their effects on LVAD knowledge and other aspects of decision-making. We summarized subject characteristics and baseline information by means with standard deviations or frequencies with percentages, as appropriate. We stratified summary statistics by treatment group. We used an independent, 2-sample t test to compare knowledge test scores postintervention between the intervention group and control group. We also used a general linear model to test for significant differences between treatment groups while adjusting for baseline test scores and other clinically important variables. We assessed statistical significance at the .05 level. We summarized cumulative scores for acceptability by means with standard deviations. All hypotheses tests were 2 sided. The statistical approach was the same for both primary and secondary outcomes.
We did not link patient data from other registries, data networks, or other outside sources. We analyzed all data as intention to treat (ITT) based on the initial treatment assignment and not on the treatment eventually received. ITT analysis is intended to avoid various misleading artifacts that can arise in intervention research, such as nonrandom attrition of participants from the study or crossover. We assumed missing data to be missing at random, including for patients who could not be contacted for a follow-up survey. We scored surveys with missing responses in accordance with the scoring instruction for each survey.
For repeated measures, we used mixed effects models that allowed us to use all available data. We ran the analyses with and without the participants that had missing data to check for consistency in responses and effects. We did not use any imputation methods (ie, replacing missing data with substituted values) for the primary analysis. The approach was the same for both primary and secondary outcomes.
We created a detailed enrollment and patient data tracking spreadsheet that indicated whether patients unenrolled, as well specific reasons for unenrollment, who decided that the participant would unenroll, and whether patients who unenrolled had completed any or none of the measures. We accounted for all participants who were enrolled, regardless of unenrollment, in our records, whether or not we included them in the analysis. Our efforts to maximize enrollment and follow-up entailed communicating regularly with LVAD coordinators to track patient visits and to ensure up-to-date contact information.
There were no changes in the study protocol over the course of the study.
We interviewed a total of 45 participants: 15 patients, 15 candidates, and 15 family caregivers. Four pairs of patient-caregiver interviews exist in our data set. Our sample approximated larger demographic trends among LVAD patients and candidates, including distribution subgroups by gender (80% male, 20% female); age (mean, 58 years); hospital status (53% inpatient, 47% outpatient); length of time with LVAD (range, 16-1894 days); monthly household income (range, $423-$10 833); and ethnic subgroups, including 31% African American, 20% Hispanic/Latino, and 49% White. See Table 3 for details of sample characteristics.
Aim 1 Sample Characteristics: LVAD Patients, Candidates, and Caregivers.
We identified 7 major themes across participants. The first themes relate the perceived nature of a patient's choice and decision-making about treatment, the most abundant of which was that most participants did not feel that they or their loved one really had a “decision” to make (patients n = 11, candidates n = 10, caregivers n = 7). They felt as if they did not have any viable options since the alternative was likely death. Participants reported that their decision-making process was mostly reflexive, devoid of heavy deliberation and discussion (patients n = 13, candidates n = 10, caregivers n = 7). In rare cases in which patients heard about the LVAD option earlier and medical management was still an option, they preferred to postpone the decision until they absolutely had to. Participants also expressed a high level of trust in and deferral to their physicians and did not do independent research outside of what their physician told them (patients n = 9, candidates n = 10, caregivers n = 3).
Another theme relates to patients' awareness of their values and expectations about outcomes. Findings demonstrated patient optimism in their expectations for transplant. Of our 15 patient interviews, 9 patients indicated that they believed they were on the transplant trajectory (BTT), but 7 of these patients were on a trajectory of DT. We found a high level of values clarity in our participants. Most patients easily identified what was important to them (patients n = 12, candidates n = 13). They wanted to live longer—especially for family members, including spouses, children, and grandchildren (patients n = 10, candidates n = 14)—and to regain mobility and engagement in the world (patients n = 7, candidates n = 13), which for some meant going to the grocery store and to others meant travel.
Another theme concerns social support needs for decision-making. Participants (patients n = 12, candidates n = 14, caregivers n = 5) reported that they needed an LVAD community of patients and caregivers that they can see and speak with—to “hear others' stories” during the decision-making process and after the LVAD is implanted if they choose in favor of it. Finally, participants reported the need for a supportive, informed, and prepared caregiver who can help receive, remember, and translate information (patients n = 10, candidates n = 11, caregivers n = 7), and a readily available clinical team that they can call at any time for even minor questions (patients n = 6, candidates n = 4, caregivers n = 6). These findings are elaborated in our publication in The Journal of Heart and Lung Transplantation.5
We conducted cognitive interviews with 3 patients and 2 candidates to assess content clarity and readability of the entire DA. Cognitive interviewing involves think aloud exercises in which the patient is asked to review the storyboards and describe in his or her own words the meaning of the information. During this process, changes in wording, informational content, and format were suggested to ensure the messages in the aid are communicated clearly.
We then tested the DA for acceptability among 10 patients, 10 candidates, and 13 clinicians. Of these, 13 participants from aim 1 were revisited for follow-up. Findings from this phase of testing indicated an overwhelmingly positive response to the DA, with 100% of patients/candidates saying they would recommend the aid to other people considering treatment options for HF, and 92% of clinicians saying they would recommend the aid to patients making the decision. Likewise, 100% of clinicians and patients/candidates agreed that the DA would help patients understand more about the risks and benefits of LVAD treatment, and to think about what aspects of HF treatment matter most to patients. Only 5% of patients/candidates felt that they had unanswered questions after reading the DA, and 70% of patients/candidates reported learning something new from the aid. Acceptability testing confirmed that the DA was appropriately targeted to its intended audience and reading level (eighth-grade level). A total of 95% of patients/candidates said they could relate to the people portraits in the patients' stories and photographs, and 100% said they could easily understand the information they were presented with.
In terms of format, participants shared their preferences for print with supplementary online delivery over other formats. Regarding distribution timing, most (55%) candidates preferred to receive the DA when visiting a local heart doctor before referral to an LVAD program, and only a quarter (25%) preferred to receive it after being evaluated and offered LVAD therapy.
This phase of testing highlighted a need to further clarify and develop certain sections (eg, elaborate on palliative and supportive care) for a more informed and balanced presentation of alternative treatment options. While patients/candidates (100%) and clinicians (92%) agreed that the DA covered both positive and negative aspects of LVAD treatment, most patients felt the DA “clearly” (50%) or “slightly” (20%) favored LVAD treatment.
The DA took an average of 59 minutes to fully review, with 95% of participants voluntarily reading the aid from cover to cover. A total of 75% of patients and candidates thought the length was “about right.” A detailed elaboration of our alpha testing can be found in our publication in The VAD Journal.36
We assessed a total of 115 individuals for eligibility for participation in the trial. Of these, 105 were eligible and subsequently randomized (10 were excluded for not meeting inclusion criteria or declining to participate). During the process of randomization, we excluded 7 more participants due to ineligibility or declining to participate, resulting in a total of 98 participants who were analyzed for the primary outcome (LVAD knowledge). Of these, 52 patients were randomly allocated to the intervention group (to receive the DA) and 53 patients were randomly allocated to the control group (no DA). In the DA group, 47 (90%) completed baseline, 37 (71%) completed 1-day follow-up (to test for acceptability and usability), 29 (56%) completed 1-week follow-up, and 27 (52%) completed 1-month follow-up. In the no DA control group, 51 (96%) completed baseline, 0 completed 1-day follow-up (no study materials administered to this group at 1 day), 34 (64%) completed 1-week follow-up, and 31 (58%) completed 1-month follow-up (see Figure 1).
After randomization to the DA group, 5 participants did not complete baseline measures for reasons of feeling overwhelmed (n = 2), being implanted before they could complete the survey (n = 1), or being unreachable. Of the 47 remaining, only 37 completed 1-day follow-up, after a total of 10 unenrolled due to asking not to be contacted further (n = 2), enrolling in another study (n = 1), feeling overwhelmed (n = 3) or being unreachable (n = 4). For the control group (no DA), 2 patients unenrolled before completing the baseline questionnaire because they became ineligible (n = 1) or enrolled in another study that disallowed participation in multiple trials (n = 1). Of the 51 remaining participants, 17 unenrolled before completing the 1-week survey due to enrollment in another study (n = 1), becoming ineligible for LVAD (n = 1), being excluded for age (n = 1), being noncompliant with terms of participation (ie, they did not agree to be contacted for follow-up; n = 1), being implanted before the survey could be administered (n = 2), withdrawing participation (n = 2), being lost to follow-up (n = 8), or death (n = 1). Of the 31 remaining in the control group, 3 more participants became unreachable for completion of the 1-month follow-up survey. In the 2 study arms, a total of 98 baseline surveys, 37 one-day surveys, 63 one-week surveys, and 58 one-month surveys were completed.
Formative qualitative research and alpha testing occurred from July 31, 2014, to July 31, 2015. The subsequent trial period ran from July 31, 2015, to January 31, 2017. The trial ended as scheduled and was not stopped for any unintentional or unexpected reasons.
Tables 4 and 5 present baseline demographic and clinical characteristics for each group. A total of 98 participants were included in the analysis, including 47 (48%) participants in the DA (intervention) group and 51 (52%) participants in the no DA (control) group. We conducted the analysis with all participants in their original assigned groups. Intervention and control patients did not differ significantly on either baseline or clinical characteristics. We found no significant demographic differences for any of our secondary, primary, or ancillary outcomes.
Aim 2 Trial Sample Patient Characteristics.
Clinical Characteristics of Trial Participants.
Table 6 presents each outcome by study arm group. We found that patients in the DA arm exhibited significantly greater LVAD knowledge than the control group at 1-week follow-up (68% vs 59%; P = .02). There was no significant difference between treatment arms at 1-month follow-up (64% vs 61%; P = .29). Results were similar when we compared LVAD knowledge between treatment arms at 1 week (P = .01) and 1 month (P = .47) using a general linear mixed model adjusting for baseline LVAD knowledge. The observed difference in LVAD knowledge did not change the direction of decision-making, with no differences between DA and standard education participants in their rates of acceptance vs decline of LVAD treatment (85% vs 78%; P = .74).
Results for Primary and Secondary Outcomes: Means (SD).
Participants who viewed the DA reported significantly greater satisfaction with life at 1 month (28 vs 23 out of 30; P = .006), based on a linear mixed model that includes fixed effects for treatment arm, time, and the arm × time interaction term. After adjusting for baseline satisfaction, the result was still significant (P = .014). The DA group reported better QOL over the control group at both baseline (56 vs 43 out of 100) and at 1 month (81 vs 69), with significant differences at both time points (P = .02; .03), adjusting for baseline scores. We observed an overall increase in QOL from baseline through 1 month for both standard education (P = .03) and the DA group (P = .003). While this difference is significant, the change from baseline to 1 month is not significantly different between arms. On average, QOL scores increased by 27 points for the non-DA group compared with 25 points for the DA group. Regarding decisional conflict, we observed the only significant difference within the Informed subscale, with the DA group exhibiting greater “informedness” at baseline than the control group (27 vs 35 out of 100, with lower scores indicating greater informedness), and no other difference observed at either follow-up.
While groups did not differ significantly for any other secondary outcomes, 75% of individuals in the intervention group vs only 43% in the control group reported that their LVAD outcomes were “very close to what I expected,” a difference that approached significance at P = .08 (DA: 75% [95% CI, 51-91]; control: 43% [95% CI, 22-66]). Regarding whether participants intended to choose vs actually chose LVAD as their preferred treatment, we found no difference between groups in their baseline or <1-week stated intentions to choose LVAD (1 week: 61% vs 62% for DA vs controls; 1 month: 75% vs 79%), and no difference between ultimate choices for implantation (85% vs 78%).
Table 7 presents results of efficacy-related acceptability and usability measures by study arm group. Acceptability testing (prespecified) demonstrated that most patients felt that the DA helped them understand more about risks and benefits of LVAD treatment (94%), felt it helped them understand their options for dealing with HF (85%), felt they learned something new (83%), and believed the DA helps someone make an informed decision about LVAD treatment (94%).
Results of Acceptability and Usability Testing Among Participants in DA Arm at 1 Day (n = 37).
Other ancillary outcomes include results from our end-of-study interviews with patients and clinicians. End-of-study interviews with trial participants provided further insights into experiences using our DA in the context of individuals' clinical and treatment experiences. We found that patients do not have a strong preference for who delivers the DA (only that the person is “knowledgeable about LVADs”), and that the preferred timing for receiving the DA is “as early as possible” after LVAD candidacy. Patients in the intervention group did not report discerning a difference between standard and enhanced education (receiving our DA), as the administration of our DA was successfully integrated into clinical flow without any observable complications. Patients also shared their appreciation of the patient stories section of the DA, as well as of the informational insert tailored for caregivers. Nearly all patients expressed the significance of meeting someone with an LVAD before making their decision. Many patients also said that they did not bring their DA to their appointments.
We have no harms or unintended effects to report across any stages of our research (Table 8). We inquired about harms only when patients refused to participate or declined participating in follow-up, at which point they were asked to share their rationale. While some patients reported feeling “overwhelmed” as a rationale for not participating in follow-up, they explained that this feeling was due to their overall context of undergoing major treatment, not due to the features of our study. Some patients described feeling unable to handle “one more thing” in addition to undergoing decision-making and LVAD treatment.
Adverse Events (n = 0).
Previous studies demonstrate a need for significant improvements in quality decision-making among patients considering LVAD treatment.7-10 The exponential increase in LVAD implantation in recent years adds further urgency to test strategies and tools designed to enhance informed patient choice.1-3 Responding to this need, we created a DA evidenced to increase patient knowledge about key decisional needs among LVAD patients and their caregivers. Our findings demonstrate that patients who received our DA had significantly higher LVAD knowledge after 1-week follow-up than patients who did not receive the aid, and were also better able than individuals in the control group to accurately anticipate their outcome experiences post-implantation. These results provide supporting evidence that our DA enhances preparedness and accurate forecasting among patients.
IPDAS18-20 as well as a recent Cochrane Systematic Review46 call for DAs that effectively convey information in unbiased ways. We found that individuals in both study arms did not differ in their actual treatment decisions. We thus conclude that the DA is a neutral support tool that does not bias individuals toward acceptance or rejection of LVAD treatment. That many patients decided in favor of pursuing LVAD treatment after reviewing the DA may indicate that patients perceive the evidence-based material contained in the DA to support a (potentially preexisting) decision to be implanted.
The AHA14 and others46 have drawn attention to the existing lack of an LVAD DA and called for guidelines to ensure that informed consent practices achieve greater uniformity and consistency across clinics providing LVAD therapy. As we described in an earlier publication,47 practices for patient-to-patient encounters are variable, inconsistent, lacking in ethical safeguards, and subject to mismatches between patient values and choice. As the AHA concluded, “the increasing prevalence, high symptom burden, and possible disease-exchanging therapies (ie, transplantation and mechanical circulatory support) for patients living with advanced heart failure mandate a systematic and thoughtful approach to decision-making.”14 Our results showed significant, positive effects on patients' LVAD knowledge, with no significant differences across sites in the effect size of the DA intervention, suggesting that our DA constitutes a positive step toward reducing variability in decision support for those considering LVAD therapy.
Only about half of the DAs included in the Cochrane Review46 of treatment and screening patient DAs reported involving patients in the development process. The authors indicated a need for eliciting user feedback at various stages of development, including alpha testing for usability and acceptability using cognitive interviewing and mixed methods. In line with these recommendations, our LVAD DA is based on multiple, iterative stages of development, including formative interviews with end-users (LVAD candidates, patients, and caregivers) who offered rich qualitative insights into their preferences for informational content and format, as well as interviews with physicians who provided clinical expertise about content criteria for informed patient treatment choice. The aid underwent alpha testing, including cognitive interviews and quantitative assessments of usability and acceptability, with user feedback iteratively applied toward the final product. Results from our trial show that patients who received the DA rated it exceptionally high in acceptability and usability, with an overwhelming majority of patients reporting that the DA taught them something they did not know before, helped them understand more about risks and benefits of LVAD treatment, and helped them understand their options for dealing with HF. Recipients of the DA almost unanimously believed the DA can help someone make an informed decision about LVAD treatment. Further details of the iterative nature of our DA development and patient-centered outcome measures are published in 2 recent methodological papers.39,40
We recognize that DAs are not administered in a vacuum but are received in a context complicated by patient- and clinic-level variables. In an effort to better disseminate understanding of the complex social and psychological aspects of decision-making about LVAD therapy, we published a number of papers exploring contextual factors affecting patient and caregiver decision-making. In addition to our publication presenting main decisional needs among patients and caregivers,5 we published a systematic review of psychosocial risk factors and their impact on LVAD outcomes,50 a commentary considering benefits and challenges of early introduction of LVAD placement from a patient-centered perspective,51 a qualitative analysis of reasons why eligible candidates decline LVAD placement,41 a mixed methods analysis of the relationship between caregiver characteristics and LVAD outcomes,52 a qualitative analysis of patients' perspectives on why LVAD therapy may be preferable to cardiac transplants,53 and a case study highlighting ethical considerations related to use of life-prolonging technology among patients at high risk for postimplant complications.54 Finally, we published a content analysis of patients' use of and preferences for social media related to LVAD.55 This publication offers an understanding of how patients and caregivers receive information about LVADs and provides a platform for strategizing future dissemination of our DA via user-preferred social media.
Regarding other implications for dissemination and implementation, our qualitative end-of-study interviews with clinicians and clinic staff who reflected on the trial implementation of our DA in practice indicated that use of our DA not only enhanced patient shared decision-making and knowledge about LVADs but also helped speed clinical flow by standardizing patient education at early stages of a patient's candidacy and by reducing information variability in the education process. Responding to calls56 for the timely translation of evidence-based DAs into clinical practice, we presented at the 2017 meeting of the International Society for Heart and Lung Transplantation a set of actionable items to enhance readiness for dissemination and implementation of our DA into practice. These steps include efforts at the clinical level to (1) elicit patients' reviewing scenario preferences on their own, with their families, and/or with their doctors; (2) engage with and designate administration responsibilities to coordinators and staff, with clear roles for coordinators/staff to review the DA with patients and caregivers to avoid mixed messages and to strategically reinforce information; and (3) optimize patient reception and retention by reviewing in small doses to avoid information overload and during times when patients are most lucid. These recommendations stem directly from our interpretation of aggregate comments from partnering clinicians. These clinicians also pointed out that a primary limitation to implementation and dissemination is physicians' limited time, hence the need for alternate clinic staff to administer the DA.
End-of-study interviews with trial participants revealed key insights for planning dissemination and implementation efforts. Patients did not voice strong preferences for who delivers the DA (only that the person is “knowledgeable about LVADs”), suggesting some flexibility in selecting a point person based on other considerations such as cost and feasibility. Further, patients prefer to receive the DA “as early as possible” after LVAD candidacy, suggesting that patients may be open to receiving information about LVAD even before being diagnosed with HF, with local clinics as a potential venue for dissemination and implementation. Nearly all patients expressed the significance of meeting someone with an LVAD before making their decision, indicating that successful implementation of our DA could be contextualized within a larger education and informed consent process that includes peer-to-peer encounters. Many patients also said that they did not bring their DA to their appointments, indicating a need for implementing clinicians and clinic staff to encourage active use of the tool in shared decision-making encounters.
Our DA is ready for use in practice in its current state, with a modular design that allows for updates to certain sections without having to reprint the entire aid. As part of our larger effort to implement our DA more widely and with fidelity (ie, according to user preferences for timing, administration, encouraging questions, eliciting control preferences, etc), we constructed a fidelity checklist for administrators that we will make available online, in addition to a brief implementation training video with step-by-step guidance on how to implement the DA with fidelity in practice. We have submitted a detailed dissemination and implementation plan currently under review for funding by PCORI. To date, 448 unique users have viewed our online DA.
Our study was not powered to test for differences in the effect of the DA in patient subgroups.
Potential limitations to the current study include possible lack of generalizability to LVAD clinics that lack the same experience and infrastructure as our partnering study clinics. Our main partnering clinic is a pioneer in LVAD implantation and patient care, and our other partnering clinics have substantial expertise in LVAD therapy. Results may differ across clinics with varying resources and experience. However, we would expect to see even greater effect sizes among LVAD clinics without a standard education package already in place, assuming that the many challenges of implementation in these comparatively resource-limited settings can be effectively addressed. Further, we did not document difference or potential contradictions between standard vs educational materials, which may be important to document during the process of implementation across LVAD clinics in the future.
Another potential limitation is that the DA is designed to target candidates considering a HeartMate II device. More advanced devices may have different implications for patient decision-making that are not covered by our DA. However, to enhance future usability of our aid, we purposefully ensured that our descriptions of the device were generic and addressed patient and caregiver issues that will likely apply with future-generation devices, and we designed the layout of our DA to be modular (ie, with removable sections), so that certain portions may be updated without reprinting the entire DA. The DA will be revised on a timeline consistent with advancements in the field, especially in LVAD technology, with significant potential for impacting patient decision-making and outcomes.
An additional limitation is the low rates for 1-week and 1-month follow-up (58% and 64% at 1 week, respectively, and 52% and 58% at 1 month), which reduce our study power as well as our confidence that our follow-up patients were typical of the entire study population.
Given that allocation was concealed at baseline but not at 1 week or 1 month out of logistical necessity (coordinators were often the ones administering both the survey materials and the DA intervention), lack of a double-blind method was an additional limitation, with the potential that coordinators could have unwittingly influenced patient responses by knowing their treatment allocation.
While recipients of the DA did not demonstrate any significant differences from the control group related to other aspects of the decision-making process (eg, decisional regret, decisional conflict, shared decision-making, and other secondary outcomes), we found that individuals who received the DA reported significantly greater satisfaction with life than those in the control group postimplant. Further research is needed to accurately understand this outcome in the context of patients' expectations and experiences postimplant. However, the observed difference offers some promise that informed decision-making (generally, or among LVAD recipients specifically) may have more widespread positive implications for global evaluations of life.
Future research in this area should also focus on institution- and clinic-level factors that may affect implementation success (use of our efficacy-tested DA) to reduce variability in how our DA is used in clinical practice and to optimize uptake by patients, caregivers, and clinicians.
We conclude that our DA was subjectively judged to be useful and objectively observed to increase LVAD knowledge over and above standard LVAD education (which differed across partnering sites) among participants who were not lost to follow-up. The results indicate that the main utility of the DA is to inform patients about what to expect from LVAD therapy vs its alternatives. Use of the DA was associated with higher satisfaction with life and QOL after 1 month (postimplant) controlling for baseline scores. The DA did not reduce decisional conflict or alignment of decision with patient values, nor did it increase preparedness for decision-making or satisfaction with the decision-making process. This is the only available trial-tested decision support tool for LVAD that integrates qualitative insights about patients' experiences and clinical expertise and is designed in line with clinicians' insights into optimizing DA dissemination within existing clinical workflow—a crucial but rarely explored factor in DA development.56
Clinicians providing care for patients with advanced HF will likely benefit from using our validated, patient-centered, and clinically informed aid to support decision-making. Positive impacts of our DA are expected to be substantial among LVAD sites that lack developed resources for education toward informed treatment choice and decision-making. The standard LVAD education offered at our participating sites was already substantial, which may have reduced the effect of our DA. We recommend augmenting standard education using our DA to maximize gains in users' LVAD knowledge. Given the surge in national rates of LVAD implantation, we believe that our DA constitutes an important tool in the decisional challenges faced by patients seeking treatment for advanced HF.
Research reported in this report was [partially] funded through a Patient-Centered Outcomes Research Institute® (PCORI®) Award (#CDR-1306-01769) Further information available at: https://www.pcori.org/research-results/2013/does-decision-aid-help-patients-learn-about-their-treatment-options-advanced
Blumenthal-Barby J, Kostick K, Bruce C, et al. (2019). Does a Decision Aid Help Patients Learn about Their Treatment Options for Advanced Heart Failure? Patient-Centered Outcomes Research Institute (PCORI). https://doi.org/10.25302/4.2019.CDR.130601769
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...
