This study was approved by the Wake Forest Health Sciences Institutional Review Board. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. The Wake Forest Health Sciences Institutional Review Board approved the study with a waiver of written informed consent.

Eligible survivors included those who were at least 21 years of age, diagnosed with nonmetastatic breast cancer, and at least 3 months after potentially curative cancer treatment (ie, surgery, chemotherapy, or radiation), excluding maintenance hormonal therapy. Additional inclusion criteria included no current evidence of disease or a history of cancer recurrence, a working email address, and ability to read medical information in English. Survivors were ineligible for the study if they had visual impairments that prohibited them from viewing material on a tablet device or if they were enrolled in hospice care or had a life expectancy less than 6 months. Survivors were identified through clinic appointment schedules and contacted by a research member prior to their appointment by telephone or immediately before their appointment in the waiting room.

Eligible providers included medical, radiation, gynecologic, and surgical oncologists; nurse practitioners, and physician assistants who provided survivorship care to posttreatment cancer survivors. A list of eligible oncologists, nurse practitioners, and physician assistants was procured from the oncology service line administrators, and providers were emailed an invitation to participate.

All participants provided informed consent prior to participation, and the study was approved by the Wake Forest Health Sciences Institutional Review Board (number 37786). Survivor participants completed a baseline survey, viewed the Vigor-Us tool (Figure 1) with their cardiovascular health information on a tablet computer with the study research coordinator, and then completed a brief postsurvey. The total research visit time was 15 to 20 minutes. Separate from the survivor assessment, provider participants were provided with a prototype of the enhanced AH-HA tool (Figure 1) on a tablet computer, introduced to the manipulation of slider bars and buttons, and asked to use the tool as they might with a cancer survivor. Providers also completed brief assessments before and after viewing the tool. Both survivor and provider participants received a US $10 gift card.

The AH-HA tool (Figure 1) visualizes data regarding the AHA Simple 7 modifiable cardiovascular heath factors—BMI, smoking status, blood pressure, total cholesterol level, hemoglobin A_1c or fasting glucose level, healthy diet, and physical activity [27]—to promote discussions at the point of care between breast cancer survivors and oncology providers. AH-HA was adapted from the SPHERE [30] primary care tool to include information about receipt of potentially cardiotoxic chemotherapies (ie, anthracyclines, antimetabolites, hormone therapy, aromatase inhibitors, monoclonal antibodies, antimicrotubule agents, alkylating agents, and radiation) [32] and was designed to be integrated with electronic health records (EHR) using Fast Healthcare Interoperability Resources [33]. Breast cancer survivors viewed a patient-facing version of the SPHERE tool designed for personal computers and mobile devices, the Vigor-Us app (Figure 1). This app did not contain information about receipt of potentially cardiotoxic chemotherapies because our clinical advisory group felt that this information was best discussed with a medical provider.

Survivor cardiovascular risk information abstracted from the medical record include weight, height, smoking status, blood pressure, total cholesterol level, and hemoglobin A_1c level; self-reported factors included smoking, physical activity, and diet (Table 1). Survivor knowledge of cardiovascular health and perceived importance and appropriateness of heart health discussions during oncology care were evaluated with 6 questions assessing confidence in understanding risk of heart disease, understanding steps needed to improve heart health, perception that cancer (or heart disease) poses a risk to health, and desire to talk to a provider (oncologist or primary care provider) about heart health. Survivors were also asked about the numerical value of each heart health factor (with “I don’t know” as an option) and to rate each health factor as high (poor health), somewhat high (intermediate health), or low-risk (ideal health) according to Table 1. Following their use of the tool, survivors completed the same 6 preassessment questions along with 3 additional questions reflecting acceptability of heart health discussions with oncologists prior to, during, and after treatment completion. Survivor tool acceptability was assessed with 5 questions on a 5-point Likert scale (strongly agree to strongly disagree) regarding liking the tool, helpfulness, ease of understanding, picture/diagram improved understanding, and desire to use this tool with oncologist. Survivors also reported gender, race and ethnicity, years of education, internet and email usage, and health literacy [34].

Provider self-reported demographic and practice data included gender, race/ethnicity, years in practice, and percentage of time spent in patient care. Provider usability was assessed using 6 questions utilized in our previous study of general internal medicine physicians [29] assessing useful information, promotion of effectiveness, ease in accessing needed information, information meets needs, easy to use, and useful format. These questions were rated on a 7-point Likert scale from strongly agree to strongly disagree. Three questions reflecting potential use of tool prior to, during, and after treatment completion were rated on a 4-point Likert scale (never, almost, always, almost always).

For this pilot study, the sample size for the oncology provider survey (n=20) is driven primarily by feasibility concerns. For the survivor survey, we estimated power to test the hypothesis that responses to each Likert scale question are generally positive, which we defined as testing the alternative hypothesis that the mean score for each question is greater than 3.5 (where a score of 3 denotes a neutral response to the question). Assuming a sample of 50 survivors and a standard deviation of 1.0, we will have >80% power provided the true mean score for a particular question is 3.9 or greater (roughly corresponds to an average response of agree).

We conducted descriptive analyses and summarized oncology provider and survivor demographics and survey responses with counts and percentages. Providers’ responses were assessed on a 7-point Likert scale from strongly disagree to strongly agree; we categorized responses of 5-7 as agreeing. Survivors’ responses were assessed on a 5-point Likert scale from strongly agree to strongly disagree; responses of agree or strongly agree were categorized as agreeing. Wilcoxon signed rank tests were used to compare breast cancer survivors’ knowledge regarding their cardiovascular risk factors and perceived importance of cancer and heart disease before and after viewing the tool. Comparisons were made individually for each of the 6 questions included on the questionnaire about cardiovascular risk factor knowledge and perceived importance and appropriateness of heart health discussions during oncology care. Sidak correction for multiple testing were utilized due to the 6 questions; P values <.0085 were considered significant for these outcomes. We calculated the percent of survivors who reported “I don’t know” for each cardiovascular risk factor. Among those who did respond with a value for their risk factor, we calculated percent agreement between categorization of objective EHR data and the survivor’s subjective assessment. Finally, we present survivor and provider data on usability of the tools. Specifically, we calculated the percent of survivors and providers who agreed or strongly agreed with the usability questions, and we presented data on the preferred timing of the intervention according to survivors and providers.

We enrolled 49 breast cancer survivors (Table 2). An additional 13 survivors were screened and not enrolled (4 did not have an email address, 6 were not interested, and 3 could not stay after an appointment). The majority of enrolled survivors (92%) had an early-stage cancer and were on 11 years postdiagnosis (mean 10.8 years, SD 6); all received surgical treatment, 55% (27/49) received chemotherapy, and 69% (34/49) received radiation. With regards to receipt of potentially cardiotoxic cancer treatments, one-third (17/49, 35%) had received treatment with an anthracycline; almost half received hormone therapy (24/49, 49%); 45% (22/49) received aromatase inhibitors, 6% (3/49) received monoclonal antibodies, 29% (14/49) received antimicrotubule agents, 43% (21/49) received alkylating agents, and 8% (4/49) received antimetabolites. Almost half of survivors (23/49, 47%) reported graduating from college, and 96% (47/49) reported adequate health literacy. Most had a cell phone (47/49, 96%), used the internet (43/49, 88%), and used email almost every day (34/49, 69%). Almost all, survivors (47/49, 96%) completed the postvisit assessment and provided cardiovascular health tool usability data.

Prior to viewing the tool, 90% of survivors (44/49) agreed that cancer posed a risk to their health, and 84% (41/49) agreed that cardiovascular disease posed a risk to their health. On average, only 45% (range 0%-86%) of survivors’ known cardiovascular health factors were reported to be at an ideal level. More than 50% of survivors reported smoking status (45/49, 92%) and blood pressure (26/49, 53%) in the ideal category; less than one-third reported BMI, diet, and physical activity in the ideal range (Figure 2).

Prior to viewing the tool, 24% of survivors (12/49) expressed strong agreement that they understood their risk of cardiovascular disease; 58% (28/49) agreed. Yet 65% (32/49) reported not knowing the level for one or more cardiovascular health factors (range 0-4). Cardiovascular risk factors most likely to be self-reported as “not known” (Figure 2) included hemoglobin A_1c (44/49, 90%), blood glucose level (32/49, 65%), cholesterol level (21/49, 43%), blood pressure (7/49, 14%), and BMI (1/49, 2%). When comparing concordance between the EHR and self-report for categorization of cardiovascular health factors as ideal vs nonideal among survivors who knew the categorization of their factor, 90% of survivors (44/49) were concordant for BMI, 47% (23/49) were concordant for blood pressure, 28% (14/49) were concordant for blood glucose level, and 34% (17/49) were concordant for cholesterol level (Figure 3).

Usability ratings of the tool by breast cancer survivors are shown in Figure 4. The majority of breast cancer survivors thought the tool was easy to understand (48/49, 98%), improved their understanding (45/49, 92%), and was helpful (45/49, 92%); 94% (46/49) liked the tool and agreed oncologists should discuss heart health during survivorship care. A majority (34/49, 69%) would like to use the tool with their oncologist at a future appointment. There were no differences in usability statistics by those 65 years and older versus those younger than 65 years.

We also assessed survivors’ perception of cardiovascular risk before and after viewing the tool (Figure 5). For all variables, survivors reported that they were in stronger agreement with the statements after viewing the tool (Figure 5). Significant changes were observed for understanding of cardiovascular risk (S=–65, P=.009), understanding steps to improve cardiovascular health (S=–70.5, P<.001, perception of health risk from cardiovascular disease (S=–45, P=.007), and desire to discuss cardiovascular risk with a primary care provider (S=–121, P<.001). There was no significant change in perception of health risk from cancer or desire to discuss cardiovascular risk with an oncologist.

We enrolled 14 physicians, 2 physician assistants, and 4 nurse practitioners; 60% (12/20) were female, 70% (14/20) were White, 10% (2/20) were Black, and 10% (2/20) identified as Hispanic or Latino (Table 2). Hematology oncology was defined as the practice specialty for 65% of providers (13/20), 60% (12/20) had been practicing as an attending for more than 5 years, and 80% (16/20) spent more than 50% of their time in direct patient care. Only 50% (10/20) reported usually or always talking to their posttreatment patients about cardiovascular health, and 35% (7/20) usually or always initiated discussion about cardiovascular health with posttreatment patients. However, 95% (19/20) reported it was somewhat or very important to discuss cardiovascular health with posttreatment patients. About half of providers (9/20, 45%) reported a high level of comfort with cardiovascular health discussions.

Usability ratings of the tool by providers are shown in Figure 4. Most providers agreed the tool provided useful information (18/20, 90%), would help their effectiveness (18/20, 90%), was easy to use (20/20, 100%), and presented information in a useful format (19/20, 95%); and 85% of providers (17/20) reported they would use the tool most or all of the time when providing survivorship care, with 50% (10/20) reporting the same for initial treatment planning and 45% (9/20) during active treatment.

Overall, our results suggest both the need for and suitability of a tailored cardiovascular health assessment tool to heighten awareness of cardiovascular health among oncology providers and breast cancer survivors. We present the first usability data from breast cancer survivors and oncology providers on the usability of EHR-integrated cardiovascular health assessment tools. On average, only 45% of breast cancer survivors’ known cardiovascular health factors were at an ideal level, most survivors did not know the value or categorization of at least one of their cardiovascular health factors, and 94% of survivors (46/49) thought oncologists should discuss heart health during survivorship care. Nearly all providers indicated that it was either somewhat important or very important to discuss cardiovascular health with posttreatment patients. However, less than half of providers reported a high level of comfort with cardiovascular health discussions, and only half reported usually or always talking to their posttreatment patients about cardiovascular health. Usability data from providers and survivors demonstrate positive perceptions of the cardiovascular health apps; 85% of providers (17/20) reported they would use the tool most or all of the time when providing survivorship care. Thus, we conclude that clinical decision support tools such as AH-HA have potential to provide relevant data to providers at the point of care to initiate discussions and prompt appropriate referrals to primary care and cardiology—settings in which cardiovascular health can be managed effectively.

The use of the AH-HA and Vigor-Us tools are one strategy for improving risk assessment and personalized cardiovascular disease prevention in cancer survivorship programs, a research priority identified by the AHA [2]. A majority of breast cancer survivors did not know one or more of their cardiovascular health risk factors, despite a majority expressing agreement before viewing the tool that they understood their risk of heart disease. In particular, knowledge gaps exist among survivors with respect to their hemoglobin A_1c and cholesterol values, which are strong independent predictors of cardiovascular disease [35]. Self-reported understanding of cardiovascular risk increased among survivors with use of the tool, and survivors increased their interest in discussing their heart health with primary care providers following the use of the tool. This increased awareness and interest may facilitate linking survivors back into primary care so these risks can be addressed.

Our results are consistent with our previous evaluation of general cardiovascular health clinical decision support in the primary care setting. In our previous study [29], providers indicated that the content and the accuracy of the tool met their needs always or most of the time. Primary care providers felt the tool was clear and presented data in a useful format, was easy to use and user-friendly, and provided up-to-date information in a timely manner [29].

Limitations of this study include the smaller sample size, nonrandomized usability assessment, single-institution setting, and the absence of data regarding the impact of the tool on cardiovascular health and health care utilization. Although we focused on breast cancer survivors in this usability study, the tool may also be appropriate for other survivor populations who have significant competing risk from cardiovascular disease. Future testing of this tool should take place in more diverse multi-institutional settings.

The AH-HA point-of-care EHR-based visualization tool brings together personalized cardiovascular health and contextual cancer treatment data to address potential gaps in breast cancer survivorship care. Our previous SPHERE study [29] suggested that cardiovascular health clinical decision support tools are well-received in the primary care setting. Findings from the current study suggest that oncology providers and breast cancer survivors would benefit from and value the integration of cardiovascular health clinical decision support apps in survivorship care. A newly initiated study will test the effectiveness and implementation of the AH-HA app in a clinic-randomized trial in community oncology practices.