Acceptability of a Patient Decision Aid for Women Aged 70 and Older with Stage I, Estrogen Receptor-Positive, HER2-negative Breast Cancer

Abstract

OBJECTIVES

A comprehensive decision aid (DA) for women ≥70 years with Stage I ER+/HER2-negative breast cancer was developed to support locoregional and systemic treatment decision-making. We aimed to test the acceptability of this novel DA in women newly-diagnosed with breast cancer.

MATERIALS AND METHODS

Women ≥70 diagnosed with Stage I, ER+/HER2- breast cancer were recruited from 3 Boston-area hospitals. They underwent baseline interviews after initial surgical consultation, reviewed the DA, and were surveyed <2 weeks later to determine DA acceptability (e.g., was it helpful?), changes in decisional conflict, stage of decision-making, and knowledge. Participants could optionally complete a three-month follow-up. Paired t-tests and McNemar’s tests were used for statistical comparisons, and thematic analyses were conducted to identify themes in participants’ open-ended comments.

RESULTS

Thirty-three of 56 eligible patients approached completed the baseline and acceptability surveys, and 25 completed the three-month follow-up. Participants mean age was 74.7 years (+/−3.8). Nearly all participants (n=31, 94%) strongly agreed that the DA was helpful, and felt the DA prepared them for treatment decision-making, with a mean decision preparation score of 4.1 (out of 5); 6% (n=2) found it very anxiety provoking. Knowledge improved with a mean of 9.0 out of 14 questions correct at baseline to 10.6 correct on the acceptability survey (p<0.0001).

CONCLUSIONS

A DA tailored to women ≥70 with Stage I, ER+, HER2- breast cancer increased knowledge and was perceived to be helpful by older women. A randomized controlled trial is needed to evaluate its efficacy.

INTRODUCTION

Over 83,000 new breast cancers are diagnosed annually in the U.S. in women age 70 years or older,¹ and as the population ages, the incidence in this age group is expected to increase significantly.² Approximately half of these women are diagnosed with Stage I, estrogen receptor-positive (ER+), human epidermal growth factor receptor two negative (HER2-) disease.³ Randomized trials in this population have demonstrated that omission of certain locoregional therapy options, such as radiation treatment (RT) after breast conserving surgery (BCS)^4–6 and axillary evaluation,^4,7,8 do not decrease overall survival in women who take endocrine therapy (ET). In addition, axillary staging often does not change adjuvant treatment strategies, and the 2016 Society of Surgical Oncology’s Choosing Wisely initiative recommended against routine axillary surgery in women ≥70 years with hormone receptor-positive (HR+) breast cancer.⁹ In women with a life expectancy of <5 years, ET alone may also be a safe option.^10–13 However, sentinel lymph node biopsy (SLNB) rates in older women with low-risk disease remain over 80%.^14,15 Similarly, RT is still given to 60–80% of such older women after BCS in the U.S.^16,17 These women also need to decide whether or not to take ET, which requires balancing the benefits and harms. These decisions are often further complicated by the presence of comorbidities and limited life expectancy.

A 2016 systematic review of available DAs on breast cancer treatment identified 23 instruments,¹⁸ with an additional web-based DA on locoregional breast cancer treatment developed in the interim. However, only two DA were developed specifically for older women; one was developed 25 years ago, and neither are currently available nor do they help women consider their overall health in their treatment decisions.¹⁹

Based on the need, we developed a novel and comprehensive workbook DA on breast cancer treatment for women 70 years and older with primary Stage I ER+/HER2- breast cancer. Design of the DA was informed by the Ottawa Decision Support Framework. The Framework involves identifying patients’ decisional needs, providing tailored decision support, and evaluating their decisional outcomes.^20–22 DAs based on this framework have been shown to increase patient knowledge, reduce decisional conflict, and increase patient engagement. Based on international standards for DAs, we included information on breast cancer, treatment options, and outcomes relevant to health status.^23,24 In addition, the DA was written using low literacy principles, considers the decision making processes of older adults, and included a validated 10-item mortality index to group women by their health status.²⁵ The included mortality index is recommended by both the National Comprehensive Cancer Network (NCCN) and the American Society of Clinical Oncology (ASCO) to estimate patient life expectancy in cancer treatment decisions.^26,27 Women who score <10 on the index are categorized as being in good health (associated with a >10-year life expectancy on the index). Women who score >14 points are categorized as being in poor health, since their scores on the index are associated with <five-year life expectancy. Women who score 10–14 points are categorized as being in fair health (estimated life expectancy of five to ten years). Although the DA informs users of their health group, it does report estimated life expectancy, since some older women find this information objectionable.²⁸ Throughout the DA, users are informed how likely they are to experience benefit or harm from a treatment based on their health status. The DA also assesses women’s treatment preferences and offers sample questions to facilitate doctor-patient communication.

The DA was iteratively revised based on feedback from 4 different Boston-area multidisciplinary breast cancer groups including surgical, radiation, and medical oncologists, nurses, social workers, patient advocates, as well as experts in medical decision making, health literacy, and geriatric oncology (see Supplementary Material for a copy). It was further shaped with feedback on its content and design from 35 women with a history of breast cancer in the past 6 to 24 months.²⁹ Based on stakeholder feedback we divided the DA into two parts. The first part includes information on surgical treatment options and RT, and the second part includes information on ET. The aim of the current study was to learn whether the DA was acceptable to women ≥70 years newly diagnosed with breast cancer (i.e., would ≥80% of users find it helpful?) and to obtain preliminary information on the DA’s effects to inform design of a future clinical trial to test the DA’s efficacy.

METHODS

Study Population and Recruitment

Women 70 years and older presenting for a treatment consultation for newly-diagnosed clinical anatomic Stage I invasive breast cancer (≤2 cm with clinically uninvolved nodes) with ER+ and HER2- receptor phenotype were recruited between August 2018 and February 2020 from the Dana-Farber Cancer Institute (DFCI), Beth Israel Deaconess Medical Center (BIDMC) (both in Boston, MA), and Mount Auburn Hospital (MAH) (Cambridge, MA) at the time of initial consultation with a surgical oncologist. Research staff (MK, ML, GA) at DFCI and BIDMC, and a surgeon (SP) at MAH screened patients’ medical records prior to the visit via a HIPAA waiver. We aimed to recruit 30 participants because we wanted to achieve thematic saturation on participants’ comments on DA improvement; qualitative studies have generally found that thematic saturation is achieved with approximately 25 participants.²⁰ Once an eligible patient was identified, the treating surgeon was contacted to approve approach of the patient at the time of their consultation. Research staff approached patients either before or after the surgical encounter, based on each clinic’s preferences. Written informed consent was obtained from all participants. This study was approved by the institutional review boards of the Dana-Farber/Harvard Cancer Center and of Mount Auburn Hospital.

Women were excluded if they had documented dementia, scored >9 on the Orientation-Memory-Concentration Test (indicative of dementia) during the recruitment process,³⁰ were non-English speaking, had a known pathogenic genetic mutation predisposing to breast cancer development, or who underwent pre-operative systemic therapy. Participants were permitted to remain in the study even if later radiologic or pathologic evaluation found the patient to have larger tumors.

Evaluation

Immediately after or within three days of their first surgical consultation, a research assistant (MK, GA) administered a baseline questionnaire, either in person or by telephone, to determine participant demographic information, life expectancy (using the Schonberg index, scores ≥10 suggest a <10-year life expectancy),³¹ decisional conflict (using the four-item SURE scale),³² stage of treatment decision-making (using a one-item index that assesses readiness to engage in treatment decision- making),³³ decision self-efficacy (using an eleven-item validated index),³⁴ and treatment knowledge (using a fourteen-item test [thirteen true/false and one multiple choice] used successfully in our prior study).^29,35–44 After completing the baseline questionnaire, participants were instructed to read the DA, and the research assistant (RA) scheduled a time to complete the DA post-visit “acceptability” questionnaire via telephone in the next one-to-two weeks. This questionnaire was developed based on the Ottawa Decision Support Framework⁴⁵ and assessed whether women found the DA helpful, clear, understandable, anxiety-provoking, whether they would recommend it or needed help from others to use it.⁴⁶ It also assessed participants’ treatment preferences, preferred role in treatment decision-making,⁴⁷ perceived involvement in treatment decisions with their surgeon (using the four-item Perceived Involvement in Care Index),⁴⁸ their perceptions on how well the DA prepared them for treatment decisions (assessed using a ten-item validated index)⁴⁹ and re-assessed participant decisional conflict,³² stage of decision-making,³³ decision self-efficacy,³⁴ and treatment knowledge. Participants were also asked to report what they liked about the DA, how they would improve it, to describe any information that was missing or unnecessary, or for any other comments; all open-ended comments were transcribed by RAs. The option to participate in a follow-up interview three months later was also offered. The three-month follow-up questionnaire re-assessed participant perceptions of the DA, including Part 2 of the DA that focuses on ET, and asked participants their actual role in treatment decision-making and their satisfaction with treatment decision making (four items on a five-point Likert scale).⁵⁰ All study questionnaires are available in the Supplementary Material.

Statistical Analysis

Paired t-tests and McNemar’s tests were used to compare participants’ responses to the knowledge test, the decisional conflict SURE test, and their self-efficacy to engage in treatment decisions by baseline patient characteristics. We also conducted thematic analyses to identify themes in participants’ open-ended comments on the DA.⁵¹ Three investigators (MAS, CM, MK) reviewed all participants’ open-ended comments and the text from visit notes to identify themes which were sorted using Microsoft Word and Excel. Discrepancies in themes identified by investigators were resolved by consensus. Statistical analyses were completed using SAS statistical software, version 9.4. Direct quotes from the data are included in the Results to illustrate emergent themes.

RESULTS

Supplemental Figure 1 demonstrates participant recruitment flow. Of 300 patient records reviewed, 241 were ineligible, the majority due to ineligible tumor characteristics or prior history of breast cancer. Of 56 eligible patients approached, 38 consented to participate but four could not be reached by phone after consenting, and 1 patient withdrew. Of the eighteen patients who refused to participate, seven refused because they were too overwhelmed with their diagnosis, six were not interested, two were too busy, one self-reported feeling too sick, one felt that she did not need help with treatment decisions, and one gave no reason. The 33 participants who completed the baseline also completed the acceptability survey, and all agreed to be contacted for the three-month follow-up survey. However, seven patients could not be reached at three months, and one other declined to participate, leaving 25 patients who completed three-month follow-up. In total, the recruitment rate was 69% and the retention rate was 80%. Mean age of the 33 participants was 74.7 years (+/− 3.8 years, range 70–84 years); 31 (94%) were non-Hispanic white; 24 (72%) were college graduates; and 4 (15%) had <10-year life expectancy. (Table 1). The median time between surgical appointment to completion of the baseline survey was 5 days (interquartile range: 5 days).

Table 1:

Sample Characteristics (n=33)^a

Characteristics
Mean age, years (SD) (age range 70–84)	74.7 (+/−3.8)
Median time from surgical consultation to baseline survey completion (days) (interquartile range)	5 (+/−5)
Race: Non-Hispanic white, n (%)	31 (94%)
American Indian or Alaskan Native, n (%)	1 (3%)
Asian, n (%)	1 (3%)
Education: Completed high school, n (%)	2 (6%)
Some college, n (%)	7 (21%)
Completed college, n (%)	11 (33%)
Beyond college, n (%)	13 (39%)
Currently married, n (%)	19 (58%)
Lives alone, n (%)	10 (30%)
<10 year life expectancy, n (%)b	5 (15%)
Recruited fromc: BIDMC, n (%)	3 (9%)
DFCI, n (%)	20 (61%)
Mount Auburn Hospital, n (%)	10 (30%)
Mean subjective numeracy (1–7)c	4.3 (1.2)
MacArthur socioeconomic scale ladder (1 [worst off] to 10 [best off])d	3.1 (1.7)
Confidence completing medical forms by yourself:
Extremely confident, n (%)	25 (76%)
Quite a bit confident, n (%)	5 (15%)
Somewhat confident, n (%)	1 (3%)
A little confident, n (%)	1 (3%)
Missing, n (%)	1 (3%)
Comfort using the internet:
Use the internet very comfortably, n (%)	16 (48%)
Use the internet somewhat comfortably, n (%)	11 (33%)
Use the internet but not comfortably, n (%)	3 (9%)
Do not use the internet, n (%)	3 (9%)
Preference for format of health educational materials:
On paper, n (%)	26 (79%)
On computer/internet, n (%)	3 (9%)
No preference, n (%)	4 (12%)
Treatment e
Mastectomy+ET+RT	1 (3%)
BCS+ET+chemotherapy	1 (3%)
BCS+ET+RT	12 (36%)
BCS+ET	14 (42%)
BCS+RT (lumpectomy with radiation therapy)	3 (9%)
BCS (Lumpectomy only)	2 (6%)
Sentinel lymph node biopsy	14 (42%)
Number of lymph nodes positive	1 (3%)
Tumor Characteristics
Clinical tumor stage (enrollment):
T1a (<0.5 cm)	6 (18%)
T1b (≥0.5 to <1 cm)	15 (45%)
T1c (≥1 cm to <2 cm)	12 (36%)
Grade (enrollment):
1	12 (36%)
2	13 (39%)
3	1 (3%)
Not available	7 (21%)
Pathologic tumor stage (post-surgery)f:
T1a (<0.5 cm)	5 (15%)
T1b (≥0.5 to <1 cm)	14 (42%)
T1c (≥1 to <2 cm)	11 (33%)
T2 ≥2 cm	2 (6%)
Missing	1 (3%)
Grade (post-surgery)f:
1	10 (30%)
2	18 (55%)
3	4 (12%)
Not available	1 (3%)

^a.Not all proportions add to 100% due to rounding.

^b.Calculated using the Schonberg mortality index. Scores of 10 or more suggest >50% mortality in <10 years and thus are estimated to have <10-year life expectancy.

^c.Abbreviations: BIDMC: Beth Israel Deaconess Medical Center, Boston, Massachusetts, DFCI: Dana-Farber Cancer Institute, Boston, Massachusetts

^d.Subjective Numeracy Scale scores range from 1 (not at all good with numbers) to 7 (extremely good with numbers⁴⁶

^e.On the Macarthur Scale of Subjective Social Status, individuals are asked to place an X on 1 of 10 rungs of a ladder (1 is the least well-off, and 10 is the most well-off) to represent how well-off (based on money/educational level/job) they feel compared with others in the United States⁴⁷

^f.Abbreviations: ET=Endocrine Therapy; BCS=Breast Conserving Surgery; RT=Radiotherapy.

^g.Largest tumor or highest grade if multiple tumors were found

Thirty-one (94%) participants found the DA helpful, 2 (97%) would recommend it; 32 (97%) felt its length was just right, and 27 (82%) felt the amount of information was just right (18% [n=6] felt more information was needed), Table 2. Twenty-one participants (64%) felt the DA was balanced, with ten (30%) feeling as if it was slanted towards RT. Participants reported understanding all (82%, n=27) or most (18%, n=16) of the DA’s content. Twelve women (36%) reported slight anxiety from reading the DA; only two (6%) reported greater anxiety. None needed help from someone else to read the DA; but, ten (30%) shared the DA with a family member. Overall, participants felt the DA prepared them for treatment decision making quite a bit, with a mean decision preparation score of 4.1 (out of 5).

Table 2:

Acceptability of the Decision Aid (n=33)

Measure of Acceptability	n (%)
Length – Just right, n (%)	32 (97%)
Too long, n (%)	0
Too short, n (%)	1 (3%)
Information – Just right, n (%)	27 (82%)
More than needed, n (%)	0
Less than needed, n (%)	6 (18%)
All information in DA clear, n (%)	22 (67%)
Most information in DA clear, n (%)	10 (30%)
Some information in DA clear, n (%)	1 (3%)
Balanced, n (%)	21 (64%)
Slanted towards radiation, n (%)	10 (30%)
Slanted against radiation, n (%)	1 (3%)
Missing, n (%)	1 (3%)
Understood all of the information, n (%)	27 (82%)
Understood most of the information, n (%)	6 (18%)
Helpful
Strongly agree/agree, n (%)	31 (94%)
Neutral, n (%)	2 (6%)
Would recommend it
Strongly agree/agree, n (%)	32 (97%)
Neutral, n (%)	1 (3%)
Reading DA made me: Very anxious, n(%)	2 (6%)
A little anxious, n (%)	12 (36%)
Not at all anxious, n (%)	19 (58%)
Prepared me for decision making (mean, +/− SD)	4.1 (0.9)
Would need help from others to use the DA:
Strongly disagree/disagree, n (%)	33 (100%)
When would you like to receive DA:
Mailed/emailed ahead of surgical visit, n (%)	10 (30%)
Doctor give during the visit, n (%)	11 (33%)
Staff give after a visit, n (%)	6 (18%)
At time of diagnosis (from PCP), n (%)	5 (15%)
At time of diagnosis (from radiologist), n (%)	1 (3%)
I asked my surgeon to explain breast cancer treatments and/or procedures to me in greater detail:
Strongly agree, n (%)	13 (39%)
Agree, n (%)	7 (21%)
Neutral, n (%)	0 (0%)
Disagree, n (%)	4 (12%)
Strongly disagree, n (%)	7 (21%)
Missing, n (%)	2 (6%)
I asked my surgeon a lot of questions about my breast cancer treatment options:
Strongly agree, n (%)	12 (36%)
Agree, n (%)	8 (24%)
Neutral, n (%)	4 (12%)
Disagree, n (%)	1 (3%)
Strongly disagree, n (%)	8 (24%)
My surgeon asked me about my worries about breast cancer
Strongly agree, n (%)	13 (39%)
Agree, n (%)	2 (6%)
Neutral, n (%)	0 (0%)
Disagree, n (%)	3 (9%)
Strongly disagree, n (%)	11 (33%)
Missing, n (%)	4 (12%)
My surgeon encouraged me to give my opinions about my treatments/options
Strongly agree, n (%)	17 (52%)
Agree, n (%)	7 (21%)
Neutral, n (%)	0 (0%)
Disagree, n (%)	1 (3%)
Strongly disagree, n (%)	8 (24%)
How do you like to make decisions about your medical treatment?
The doctor decides what should be done and does it, n (%)	1 (3%)
The doctor presents his or her recommendations to me to accept or reject, n (%)	10 (30%)
The doctor discusses alternatives with me and we decide together how to proceed, n (%)	17 (52%)
The doctor presents all available options and allows me to decide, n (%)	5 (15%)
Has one of your family members read the workbook?
Yes, n (%)	10 (30%)
No, n (%)	23 (70%)
Do you anticipate that a family member or friend will help you with decisions about treatment?
Yes, n (%)	22 (67%)
No, n (%)	11 (33%)

Preferences regarding when women should receive the DA varied; eleven (33%) participants would prefer to receive it from the doctor during their visit, ten (30%) for it to be mailed ahead of the visit, six (18%) for the staff to provide it after the visit, and five (15%) wanted to receive it from their primary care physician at diagnosis. Most (n=26, 79%) prefer to receive educational materials on paper rather than on a computer, even though sixteen (48%) reported being very comfortable using the internet.

Regarding their involvement in treatment decision-making with their surgeon, twenty (62%) reported asking their surgeon to explain their breast cancer treatment options in greater detail, and twenty (62%) reported asking their surgeon a lot of questions about treatment. Twenty-five (73%) reported that their surgeon encouraged them to give their own opinions. In terms of preferred decision making-role, seventeen (52%) preferred that their doctor discuss their treatment options and that they decide together on treatment, while ten (30%) preferred that their surgeon make a recommendation which they could accept or reject. Five women (15%) wanted their doctor to present all the available options and let the patient decide. Only one woman (3%) wanted the doctor to decide on treatment without her input. Nine women (27%) also wanted family involved in treatment decision-making.

Table 3 presents how outcome measures changed from the baseline (completed within three days of surgical encounter) to acceptability questionnaire (completed a median of thirteen days [interquartile range: 12] after baseline). Participant knowledge increased, with women scoring a mean of 9.0 (out of fourteen) questions correctly at baseline and 10.6 correctly at the time of acceptability (p<0.0001). Supplementary Table 1 shows the change in women’s response for each question. For seven of the fourteen questions, >80% of women answered the question correctly at baseline leaving little room for improvement. Participant knowledge significantly improved for understanding that RT and axillary exploration were unlikely to improve older women’s survival and that axillary exploration is unlikely to change their treatment choices. Few women (<20%) had answered questions related to these issues correctly at baseline. Women reported high self-efficacy for treatment decision-making both at baseline and after reading the DA (mean score on eleven-item index, 95.7 versus 96.8, p=0.23). Similarly, women felt sure in their treatment decision-making at baseline and after reading the DA (mean score, 3.4 versus. 3.6, p=0.44). Supplementary Table 2 shows women’s responses to each question on the decisional conflict SURE test. After reading the DA, women tended to be more likely to report that they knew the benefits and harms of their treatment options. Only one patient underwent mastectomy. Thirty-two patients underwent BCS (97%) and fifteen (47%) of those also received RT. Fourteen patients (45%) underwent SLNB (1–4 lymph nodes were sampled), of which one had a single positive node; none needed an axillary lymph node dissection. On final pathology, 32 patients (97%) had tumors ≤2 cm, and 28 (85%) had grade one or two tumors. Twenty-eight patients (85%) received ET.

Table 3:

Outcome Changes from Baseline to Patient Acceptability and Patient Acceptability to Three Month Follow-Up Questionnaires

Outcome	Baseline (n=33)	Patient Acceptability (n=33)	p-value*
Knowledge Score, mean (SD)	9.0 (1.5)	10.6 (1.6)	<0.0001
Decision Conflict Score, mean (SD)	3.4 (0.8)	3.6 (0.8)	0.44
Decision Self-Efficacy, mean (SD)	95.7 (7.0)	96.8 (6.2)	0.23
Preparation for Decision Making, mean (SD)	-	4.1 (0.9)	-
Stage of treatment decision making, n (%)			0.53
Haven’t begun to think/currently considering treatment options	10 (30)	8 (24)
Close to or already made a decision	23 (70)	25 (76)

^*p-values for paired t-test or McNemar’s test where appropriate

The 3-month follow-up survey was completed by 25 women a median of 105 days after the acceptability questionnaire (interquartile range: 15 days). Women reported no significant changes in their perceptions of the DA’s acceptability; 24 women (96%) still found the DA helpful (Supplementary Table 3) and felt the DA helped prepare them for decision-making quite a bit (mean 4.0 [+/− 0.9] out of 5). Sixteen (64%) women reported no anxiety and nine (36%) reported only a little anxiety using the DA at three months. Fourteen women (56%) reported using the DA a few times or more during follow-up, nine (36%) read it at least once more, and two (8%) did not read it again. Participants reported feeling very satisfied with their treatment decisions (mean score 18.6 [+/− 2.0] on a twenty-point scale).

In open-ended comments, participants reported that the DA was “very easy to understand, digest, and not too complicated.” (Table 4). Participants also reported that the DA increased their awareness of the complexity of their treatment choices and that their overall health should affect their treatment decisions. The DA helped them feel supported in their treatment decisions and they felt that it was also a useful resource for family. Some felt it would help women to receive the DA at the time of their diagnosis from a primary care provider, while others felt it would be best to receive from the surgeon’s office: “Best to come from physician treating cancer but also someone could be more comfortable with a PCP.” Women had suggestions for improving the DA, including providing more information on the decision between mastectomy AND lumpectomy alone OR lumpectomy and RT. They also wanted more details on the benefits and harms of axillary evaluation and on the role of ET on memory.

Table 4.

Themes Related to Decision Aid Acceptability and Thoughts at Three Month Follow-Up

Theme	Sub-theme	Example quotes
Usability	Clear	“It was concise and very clear and just really went over many scenarios and options…it was much better than any other information I received.”
	Well-organized	“It is just very clearly laid out.”
	Varying view on graphs	“I think the diagrams are worth their weight in gold.” “There were too many graphs. Some of the information in the graphs could have been [written] out in short paragraphs.”
	Font	“The print size was wonderful.”
	Suggestions for improving Clarity	“[It says] lumpectomy + radiation is the same as mastectomy, but then [it] goes on to say that not all women need radiation…”
Helpful for decisions	Increased awareness of complex choices	“It made me more aware of the complexity of the choices which I was not aware of”
	Promotes patient involvement	“…this showed that patients are allowed to ask questions. I think it makes the patient feel like two adults are talking.”
	Increased awareness of overall health	“The health questions were interesting and…got you thinking.”
	Felt supported in decision-making	“The workbook really provided the data I needed to feel comfortable about doing something else.”
	Helpful for family	“It was also helpful for my family when I wanted to stop the hormone pills; they referred back to the workbook for statistics. It became a tool for the entire family.”
	Helpful resources to have at home	“Having a resource at home that you can look at in your own time with your own thoughts in a peaceful way is beautiful.”
Implementation	Varied preferences regarding timing of DA receipt	“Since some people go off the wall, I might say get it from [one’s] gynecologist or primary care provider depending on the person’s reaction at the time of diagnosis…” “I think it should be given by the surgeon’s office.”
	Translate	“I suggest it be in Spanish or other languages.”
Varied effect on anxiety	Varied anxiety levels in response	“The workbook made me very anxious because I didn’t like the side effects for the medication that I read about.” “I was an anxious mess and the information relieved that.”
Additional topics/detail to cover	More discussion of mastectomy versus lumpectomy (with or without RT)	“For me, I had to decide whether I was going to [undergo] a mastectomy or lumpectomy, so I don’t think it discussed those procedures enough.” “It would be helpful for the workbook to address life after surgery and questions people might have after the surgery.”
	More information needed on axilla dissection	“One part I had a hard time with was the sentinel lymph nodes…I still didn’t understand after [it] reading the workbook.” “There is no mention at all about dissection of the axillary nodes but it was explained to me as a possibility depending on the sentinel nodes and I assume that would further impact your recovery.”
	More information on the effects of radiation	“I would have liked to see a page on radiation lowering the chance of it coming back in other parts of your body.” “Will the pain from radiation subside?”
	More information on side effects of endocrine therapy	“There were some questions I had about the side effects and how they progress over time: will they build up overtime or go away if I stop taking the pills? It would be nice to have a little bit more information about these side effects even including stuff about memory loss.”
	Psychological issues	“Include information about psychological issues- who to consult/spiritual guidance etc.”
	More detail on effect of health status on treatment outcomes	“The workbook gives you a very handy tool to assess your own health whether it is good or fair or poor and then it gives you charts but it doesn’t break the charts up to good health, fair or poor.”
	Explain hormone receptivity	“It might be helpful to include what ER+/PR+ is because I had looked it up and knew that was my type but I don’t know what that means.”
	Oncotype DX information	“…there is genomic testing but it doesn’t list “if genes have x, y, or z then you might benefit from X.”

DISCUSSION

Grappling with decisions about breast cancer treatment can be difficult for patients and fraught with anxiety. While 94% of women in our study found the DA helpful, women varied with regards to whether anxiety was provoked or allayed by the DA, consistent with the mixed findings of other evaluations of cancer-related DAs.⁵² One of our past studies, which examined the use of a DA focused on the decision to undergo screening mammography among women ≥75, found that even seemingly emotionally-neutral educational materials (e.g. a home safety pamphlet), can engender anxiety in up to 15% of older women.⁵³ During this psychologically challenging time, any materials regarding potential future treatments may induce anxiety, but we have taken this into consideration and the DA emphasizes the good prognosis for this particular patient population. Reassuringly, only two participants (6%) reported that the DA made them very anxious. Without a control group in this study, it is impossible to know whether reading any materials about their health would have made these women anxious; however, future work should continue to address strategies to mitigate anxiety during this process.

Similar to other DA studies, knowledge improved after DA use. Notably, women were better able to understand that RT and axillary exploration are unlikely to improve survival and that axillary exploration is unlikely to affect treatment choices after reading the DA. These are challenging concepts to explain and to understand and <20% of women answered questions related to these concepts correctly before reading the DA. Highlighting older women’s need for information on these issues, women wanted even more information about the benefits and harms of SLNB, which will be added.

Decisional conflict did not improve significantly after use of the DA. To reduce participant burden, we used the 4-item SURE test to measure decisional conflict rather than the standard 16-item scale that includes five subscales. This may have led to a ceiling effect since nearly all women reported feeling SURE in their decision at baseline. In future work we will use the 16-item decisional conflict scale instead. While reduction in decisional conflict has traditionally been an important goal of DAs, there is increasing recognition that decisional conflict may promote appropriate deliberation.⁵⁴ As participants noted, our DA helped them realize the complexity of their breast cancer treatment decisions.

While an online version of this DA would allow it to better individualize outcomes in relation to health status for different treatment options, the women in our current and previous studies preferred a paper-based tool.^29,53 Even in this highly-educated patient sample, only half of these women felt very comfortable using the internet. However, an online tool would likely be helpful in implementing this DA in wider settings. As the women in our study suggested, this DA will also have to be translated to other languages in order to be accessible to a wider demographic of women.

Of note, this DA was provided to the patients after their initial surgical visit, without integration into the visit itself. When asked how they would like to receive the DA, 49% of patients preferred to receive it at some point prior to their surgical visit, and 33% preferred to receive it from their doctor during their visit. Online decision support tools,⁵⁵ that have been designed to be integrated into clinical treatment conversation have been shown to acceptable and helpful in older adults with early-stage breast cancer.⁵⁶ Thus, having the DA either prior to or during the visit could allow physicians to better elicit older women’s preferences regarding treatment and engage them in shared decision making. We plan to further examine the timing of distribution and efficacy of this DA in a randomized controlled trial.

There are several limitations to this study. First, the study population was mostly comprised of non-Hispanic white, college-educated women, treated in one geographic locale in the U.S. Although >90% of patients diagnosed with Stage I HR+ breast cancer after 70 years of age are non-Hispanic white³, the subsequent lack of generalizability to other patient populations, including international populations, must be acknowledged, Although the DA was designed with low-literacy principles, the predominance of college-educated women in our study sample suggests that further testing is needed in populations with limited health literacy to understand if further revision is needed or if use of the DA among women with low health literacy requires support from family, caregivers, and/or health educators. Second, this study was not designed to study the DA’s effect on women’s treatment decisions. However, it is notable that receipt of RT after BCS (46%) and of SLNB (42%) were lower than of national rates suggesting a potential effect of the DA on older women’s treatment decisions. Third, although saturation of theme was reached, there were still only a relatively small number of patients assessed in this analysis. Fourth, as we aim to support decision-making in current practice settings, and as geriatric assessments (GA) are not integrated into the surgical practices in our study, a full GA was not performed at any point. However, patient life expectancy and self-perceived health were assessed which, together, may approximate some components of a GA²⁷ Finally, although there were no reports of patients involving non-family caregivers in treatment decision-making, we did not specifically assess for this.

In summary, this DA designed to assist women ≥70 years with Stage I ER+/HER2- breast cancer with their treatment decisions was shown to be acceptable to this patient population and is a promising decision support tool. Future steps in its development include modifying the DA in response to the comments we received, further investigating the optimal delivery time, and testing its efficacy in a randomized controlled trial with a diverse population.