Multilevel Factors Impacting Mammography Screening Decisions in the Elderly: A Clinician and Patient Pilot

For there is evidence to recommend as the benets may not outweigh harms. The objective of this study was to identify variables that clinicians consider inuential when making screening mammography recommendations for these women and to assess the acceptability and feasibility of a patient print intervention designed to support patient decision-making and patient-clinician communication about stopping mammography. guidelines, Cancer Society (ACS)

3) The implications are that this study would be feasible to reproduce in other clinic or medical setting where women of this target population can be screened.

Background
More than a quarter of new US breast cancer cases diagnosed each year occur among women ages 75 years or older. 1 The increasing life expectancy of women and attendant rise in breast cancer cases in older women will likely lead to an increasing absolute number of mammograms performed in this group of patients. 2 However, the US Preventive Services Task Force (USPSTF) currently does not recommend screening mammography among women in this group due to insu cient evidence. 3 Given that survival bene t from screening is not typically observed among women with life expectancy of < 10 years, many older women with co-existing chronic illnesses may not live long enough to bene t from the procedure.
Current guidelines emphasize the need to tailor screening mammography to life expectancy and screening preferences.
This need for tailoring poses challenges for effective communication among primary care clinicians and older women. However, there is a paucity of tools to facilitate effective communication and decisionmaking in the rapidly growing older population. [4][5][6] Factors that may hinder effective communication and decisions include provider factors such as uncertainty regarding the effectiveness of preventive healthcare in the elderly [7][8][9] and discomfort in discussing life expectancy 10,11 and how it impacts care decisions [12][13][14] , as well as patient factors such as habits regarding screening mammography usage and related cancer worry [15][16][17] . Identifying variables most in uential to patient-provider communication in this setting is key to determining how to optimize patient outcomes.
In this study, we assessed factors that clinicians perceive as most in uential to their recommendation to stop screening mammography. We then applied these data and existing literature 4,5 to develop a tailored print patient activation intervention to support patients' understanding of their breast cancer risk, life expectancy, and account for potential bene ts and harms of mammography. We also included questions for patients to consider asking their primary care clinician in order to facilitate shared decision-making about continuing or ceasing mammography. We then piloted this intervention to assess feasibility, acceptability, and initial e cacy.

Participants
Clinician and patient recruitment and retention are displayed in Fig. 1.

Clinician survey
Clinicians were recruited from a primary care research network in the Mid-Atlantic US from February-May 2019. Of the eligible clinicians contacted (N = 131), 22 completed the survey via email (17% response rate) and received a $50 gift card upon completion.

Patient pilot
Female patients aged 75-84 years were recruited from three primary care practices in Maryland and Washington, DC between June 2019-November 2019. Eligible patients (N = 20) had received a mammogram within the past two years and not been diagnosed with invasive breast cancer, DCIS, LCIS, or had received prior chest radiation therapy. Patients were screened for eligibility and 14 completed a baseline phone interview (70% recruitment). Baseline data were used to create tailored decision-making aid booklets that were then mailed to patients. 11 patients completed post-intervention interviews (79% retention) two weeks later to provide feedback and received a $20 gift card for each survey.

Measures
Clinician survey Clinician variables Clinician demographics (age, gender, race/ethnicity) were obtained via self-report, as were the percentage of their patients in their current practice who were low income, on public insurance, age ≥ 75, and among those 75 or older, the percentage who received mammograms (< 25%, 25-50%, > 50%). Clinicians also reported the age at which they recommend stopping screening mammography. Research network registry data were used to describe practice setting and clinician specialty.
Variables that in uence recommendation to stop screening By adapting methods from prior work [18][19][20][21] utilizing the literature on overscreening in the elderly, [22][23][24][25][26] clinicians were asked to indicate the extent to which patient variables affect their recommendation to stop screening mammography at age 75 (1 = Never in uences to 5 = Always in uences). Scores of 4 or 5 were considered in uential. Factors included functional status, overall comorbidity, family history of cancer, pain from the procedure, and a recent death due to cancer in the patient's family. Other factors included communication (with colleagues, the practice's policies/norms, the patient's family and friends' opinions about mammography screening, media representations of screening mammography, time in the consultation, a language barrier between the clinician and patient, clinicians' di culty explaining potential screening mammography bene ts and harms, the patient's di culty understanding screening mammography bene ts and harms, and the nature of screening results) and clinical evidence (new research evidence concerning screening mammography in the elderly, USPSTF guidelines, American College of Radiology (ACR) guidelines, and American Cancer Society (ACS) guidelines) Baseline We collected patients' age, race/ethnicity, marital status, education, income, and insurance status, as well as variables needed to create intervention materials (below). Patients also indicated their 5-year and lifetime perceived breast cancer risk using Personal Absolute Percentage Risk items (0-100 scale) 22 and their cancer worry oriented to the decision to continue or discontinue mammography. 27 Interventional patient materials Patients received a tailored infographic booklet using baseline data input into NCI's Breast Cancer Risk Assessment Tool (BCRAT), [28][29][30][31] to show current and future breast cancer risk. The Lee Schonberg Index within the ePrognosis breast cancer screening module [18][19][20][21] was used to calculate the potential screening bene ts and harms. Infographics were used to convey numerical information about potential risks and harms while question prompts were utilized to help patients consider their preference and respective bene ts of their choice to continue or discontinue mammography. To convey bene ts, patients were provided a ve-item questionnaire and asked to rate the importance of each item (0 = Not at all important to 10 = Especially important). Potential harms from mammograms in the rst year, deaths from breast cancer avoided after 10 years, and deaths from other causes after 10 years, were all conveyed as percentages in the infographics.

Post-intervention Survey
In addition to perceived cancer risk and cancer worry, patients provided their intentions to continue to get mammograms (1 = Not at all likely to 4 = Very likely) They also indicated intervention acceptability using seven items 'Yes/No' items 32 and their overall satisfaction with the materials (1-4 scale).

Data Analysis
Clinician and patient responses were summarized using frequencies and descriptives. Clinician characteristics were compared using prevalence odds ratios. Paired t-tests were used to compare patient cancer risk and cancer worry across time points.

Clinician survey
Primary care clinician characteristics are presented in Table 1 . Approximately half of clinicians indicated that the opinion of a patient's friends and family [10/22, 45.5%] in uenced their recommendation. Notably, clinicians reported being minimally in uenced by both di culty in their explaining potential screening mammography bene ts and harms to their patients [3/22, 13.6%] and, conversely, by their patient's di culty understanding this information [5/22, 22.7%] ( Table 2). Clinicians reported that, on average, they recommend stopping screening mammography when a patient is 77 years old (range = 70-85). Clinician association with gender, race, and patient population socioeconomic status are detailed in Table 2.   Patient characteristics are presented in Table 1. Patients (N = 14) were all female, predominantly White (N = 8, 57.8%), and were on average 79 years old. Patient's perceived lifetime risk of breast cancer, meaning the chance from 0-100 a patient believes they will develop cancer in their lifetime, decreased from pre-to post-intervention (M = 22.91 vs. 4.18, t = 2.79, p = 0.02), with a score more aligned with actual breast cancer risk for these patients (M = 2.69, SD = 0.79, N = 13). Patients were satis ed with the intervention materials (M = 3.3/4) and endorsed the overall length (100%), that the materials were easy to read and understand (64.3%), that the booklet would help them make decisions (57.1%), and that they learned new information from the booklet (57.1%). Few indicated that it was upsetting to use the booklet (7.1%), which is in line with previous surveys in which older women would avoid being reminded of or discussing mammography 33 . Patients indicated that they were somewhat to very likely to talk to their doctor about whether they would continue screening mammography (M = 3.2/4, 4 = very likely). Those who wanted to continue indicated concerns about their speci c breast cancer risk factors.

Discussion
Among women aged 75 and older, patient and clinician perspectives both in uence the decision to stop screening mammography. 34 Our results support that a clinician's recommendation to stop screening in older patients is in uenced by a combination of the patient's health, cancer risk factors and clinical guidelines, speci cally those issued by the USPSTF.
We observed that patients are open to using this information to make decisions about stopping screening, with a signi cant reduction in perceived breast cancer risk that suggests processing of the information about cancer risk and weighed bene ts and harms from mammography. Women often overestimate their breast cancer risk 15,16,35−37 and this led to anxiety and disease-speci c worry. 15 In response, family and friends can push women to continue screening past the point where patient would have stopped. 33 Further, women in our study held fairly strong intentions to talk to their primary clinician about stopping mammography. This suggests that shared decision-making encounters that factor in the opinions of their patient's friends and family could be fruitful in reducing overall screening rates in this population. 38 This study was limited by the small sample size and recruitment from one primary care research network.
The clinician population of this study was weighted towards family physicians and lacks input from physicians in other specialties who might be in a position to have this conversation in the primary care setting. We relied on self-report of the perceived in uence on communication and did not address potential language barriers between the clinician and patient, given that our inclusion criteria required both groups to be uent in English.

Conclusions
Our results suggest that clinicians and patients are prepared take part in shared decision-making discussions about discontinuing mammography screening for women over 75. Replication of this study in a larger population has the potential to assist primary care clinicians in incorporating conversations around screening and recommendations in routine care. Primary care clinicians report barriers such as uncertainty in predicting prognosis, di culty in discussing prognosis, and concern about patient reactions. Patients stand to gain a more accurate understanding of their breast cancer risk, experience reduced cancer related distress, and utilize information from the intervention in shared-decision making.
For these reasons, research should be conducted to promote communication between physician and patients around the of screening mammography in elder patients.

Declarations
Ethics approval and consent to participate The study protocol was reviewed and approved by the Institutional Review Board of Georgetown University.

Consent for publication
Not applicable Availability of data and materials The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.