This randomized clinical trial compares rates of adherence to recommended breast cancer screening when using an opt-out automatic referral strategy compared with an opt-in automated telephone message strategy among veterans.
Key Points
Question
Does an opt-out automatic referral strategy improve adherence to breast cancer screening compared with an opt-in automated telephone message strategy among veterans?
Findings
In this randomized clinical trial of 883 veterans, there was no difference in rates of completed mammography between opt-out and opt-in strategies.
Meaning
In this trial, an opt-out compared with an opt-in breast cancer screening strategy did not result in a significant difference in mammography completion.
Abstract
Importance
Optimal strategies for population-based outreach for breast cancer screening remain unknown.
Objective
To evaluate the effect on breast cancer screening of an opt-out automatic mammography referral strategy compared with an opt-in automated telephone message strategy.
Design, Setting, and Participants
This pragmatic randomized clinical trial was conducted from April 2022 to January 2023 at a single Veterans Affairs (VA) medical center. Participants were female veterans aged 45 to 75 years who were eligible for breast cancer screening and enrolled in VA primary care.
Intervention
Veterans were randomized 1:1 to receive either an automatic mammography referral (opt-out arm) or an automated telephone call with an option for mammography referral (opt-in arm).
Main Outcomes and Measures
The primary outcome was completed mammography 100 days after outreach. Secondary outcomes were scheduled or completed mammography by 100 days after outreach and referrals canceled if mammography was not scheduled within 90 days. Both intention-to-treat analyses and a restricted analysis were conducted. The restricted analysis excluded veterans who were unable to be reached by telephone (eg, a nonworking number) or who were found to be ineligible after randomization (eg, medical record documentation of recent mammography).
Results
Of 883 veterans due for mammography (mean [SD] age, 59.13 [8.24] years; 656 [74.3%] had received prior mammography), 442 were randomized to the opt-in group and 441 to the opt-out group. In the intention-to-treat analysis, there was no significant difference in the primary outcome of completed mammography at 100 days between the opt-out and opt-in groups (67 [15.2%] vs 66 [14.9%]; P = .90) or the secondary outcome of completed or scheduled mammography (84 [19%] vs 106 [24.0%]; P = .07). A higher number of referrals were canceled in the opt-out group compared with the opt-in group (104 [23.6%] vs 24 [5.4%]; P < .001). The restricted analysis demonstrated similar results except more veterans completed or scheduled mammography within 100 days in the opt-out group compared with the opt-in group (102 of 388 [26.3%] vs 80 of 415 [19.3%]; P = .02).
Conclusions and Relevance
In this randomized clinical trial, an opt-out population-based breast cancer screening outreach approach compared with an opt-in approach did not result in a significant difference in mammography completion but did lead to substantially more canceled mammography referrals, increasing staff burden.
Trial Registration
ClinicalTrials.gov Identifier: NCT05313737
Introduction
Over 280 000 individuals who receive care from the Veterans Health Administration (VHA) are eligible for breast cancer screening (BCS), and only 66% are up to date on screening.1 Since the COVID-19 pandemic, BCS rates have been persistently lower within the VHA compared with before the pandemic.2 Barriers to BCS among veterans include scheduling, long wait times, and lack of on-site mammography, with the need to coordinate care with outside facilities.3,4,5 Population-based outreach for preventive care (in which patients due for screening are notified outside a primary care visit) may help to address barriers to BCS. However, these strategies can be time intensive and have yielded mixed results both in and outside the VHA.6,7,8,9,10 Considering the null or modest results from these studies, optimal strategies for population-based outreach for BCS remain unknown.
Behavioral economics principles can be used in population-based outreach to help design interventions to align with how people typically make decisions.11,12,13,14 One principle, status quo bias, states that people are more likely to maintain a default (ie, opt-out) option than to actively choose an alternative (ie, opt-in).15 Opt-out programs—those that deliver services to all eligible participants, with the option to decline—have been shown to be effective in increasing influenza vaccination and colon and cervical cancer screening rates compared with opt-in approaches—those that offer services to all eligible participants but deliver them only to those who actively accept.12,13,14,16 Opt-out strategies for BCS outreach may help to decrease access barriers for patients and improve reach by making screening the default option. However, to our knowledge, opt-out vs opt-in BCS outreach approaches have not been evaluated in the VHA.
The objective of this quality improvement analysis was to compare the effectiveness of opt-out vs opt-in population-based outreach strategies for a breast cancer program at a single Veterans Affairs (VA) medical center among veterans who were due for BCS. Our goal was to evaluate strategies to improve BCS rates among veterans.
Methods
Study Design
We conducted a randomized clinical trial from April 2022 to January 2023 to compare the effectiveness of an opt-out vs opt-in BCS outreach strategy (NCT05313737). The trial protocol and statistical analysis plan are in Supplement 1. The VHA Office of Primary Care is committed to the development and principles of the learning health systems model, embedding research teams within operations to improve care delivery for veterans.17 Combining the rigor of clinical trial methods with health systems operations in a way that is feasible and pragmatic can allow for rapid knowledge generation and dissemination.17,18,19 Using this model, we evaluated 2 operational processes for BCS. This trial was completed as nonresearch quality improvement for evaluation of primary care operations under the designation of the VHA Office of Primary Care and was therefore not subject to institutional review board review or exemption. There were no active recruitment or informed consent procedures given the minimal risks and purpose of quality improvement. This study followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline.20
Study participants included veterans who were assigned to a primary care practitioner (PCP) and were due for BCS in the VA Puget Sound Health Care System (VAPS), which like many VA medical centers, does not provide mammography on site but refers patients to breast imaging services in the community and pays for this care.3 All eligible veterans were included.
Setting and Participants
We identified female veterans aged 45 to 75 years who were due for BCS using the VA Breast Care Registry, which provides data for longitudinal tracking of breast care and is sponsored by the VA Office of Women’s Health (Figure).21 We referred to the American Cancer Society (ACS) guidelines, which are used by the VA to determine BCS eligibility.22 Veterans were included if they had no documented screening in the past 1 year (for veterans aged 45-54 years) or 2 years (for veterans aged ≥55 years). The ACS recommends stopping screening when life expectancy is 10 years or less. We opted to use 75 years as the upper age limit for inclusion because it is the age at which VA guidelines recommend health care practitioners discuss the benefits of continued screening with patients. Additional eligibility criteria included at least 1 year of prior data (ie, evidence of at least 1 outpatient visit) available as of April 2022 and no evidence of future mammography scheduled within 12 weeks of outreach. Those with a history of bilateral mastectomy, who were enrolled in hospice, or who had a record of recent death were excluded. Trial eligibility was determined on April 13, 2022. The baseline BCS rate for eligible veterans at VAPS was 67% based on the VA External Peer Review Program BCS quality measure as of October 2021.23
Figure. CONSORT Diagram.
aHung up, line busy, no answer, no way to leave a message, or nonworking number.
bDescribed in the Randomization and Intervention subsection of the Methods section.
Randomization and Intervention
Eligible veterans were randomized in a 1:1 allocation using permuted block randomization with random block sizes of 2 and 4 to either (1) an opt-out automatic referral or (2) an opt-in automated telephone call (Figure). Randomization was stratified within arms by prior screening status (ie, history vs no history of BCS).
In the opt-out group, primary care nurses led by a women’s health nurse manager (C.F.) reviewed records for all patients and placed a referral if the patient was eligible for BCS. Veterans were ineligible if medical record review revealed that the veteran (1) was not due for screening, (2) recently declined screening per documented conversation with a PCP, (3) had a contraindication to screening (ie, undergoing breast cancer treatment or needed diagnostic breast imaging), or (4) was no longer a patient at VAPS. Veterans could request to speak with their doctor or care team about the recommended screening during the scheduling call.
The opt-in group received an automated telephone call with the options to press 1 if they would like a referral for mammography, 2 if they would like to discuss BCS with their PCP or had completed mammography within the past 12 months, or 3 if they declined mammography (Box). If they opted in for BCS, a nurse reviewed the patient’s medical record to confirm screening eligibility and placed a mammography referral. Nurses who ordered mammography referrals were blinded to the trial arm except if an individual requested a telephone call to discuss screening or to document prior screening.
Box. Automated Telephone Call Script (Opt-In Arm).
Greeting
This is the VA Puget Sound Health Care System calling about your cancer screening.
Message Content
According to our records, you are due for breast cancer screening with a mammogram. Screening can now detect breast cancer earlier when it is easier to treat. Your primary care doctor recommends that you complete a mammogram.
-
Please select 1 of the following options:
Please press 1 if you would like us to order your mammogram at this time. Someone from our breast care team will contact you within 2 weeks to schedule your mammogram.
Please press 2 if you would like to discuss cancer screening with your doctor or if you completed a mammogram within the last 12 months outside of the VA.
Please press 3 if you do not want us to order your mammogram at this time.
Closing
If you have questions or think this message may be an error, please call your primary care team at [primary care team contact telephone number]. Thank you for your service and have a great day.
For both groups, once a nurse placed a mammography referral in the electronic health record, the usual operational procedures followed. A VA department tasked with scheduling and coordinating care received at outside facilities coordinated mammography scheduling. The outside facility scheduled screening directly with the veteran. The VA confirmed scheduled mammography with the outside facility and reminded veterans to schedule if needed. Multiple outreach attempts by the VA team were spaced over several weeks. The referral remained active for 90 days, after which no additional reminder calls were made and the referral was canceled. The VA requested and received mammography results from the outside facility and uploaded results into the VA record; PCPs relayed results to veterans. The referral was deemed complete once results were received and documented in the VA record.
Outcomes
Our primary outcome of interest was mammography completion at 100 days after outreach. Our secondary outcomes of interest were (1) mammography completed or scheduled at 100 days after outreach and (2) canceled referrals (those that had been placed but not scheduled within 90 days). We retrieved primary and secondary outcomes from electronic health record data. We chose 100 days to accommodate the challenges and delays in coordinating care with outside facilities.3 We evaluated mammography scheduling as a proximal or earlier outcome on the pathway to completion. Canceled referrals were a measure of excess administrative burden.
Covariates
In addition to outcome variables and prior screening status used in stratified randomization, covariates in the analysis included age, race and ethnicity, neighborhood disadvantage, veteran rurality, and medical comorbidities. Race and ethnicity were defined using a validated algorithm developed in the VA and were included in the analysis given evidence of lower BCS rates and higher breast cancer mortality in marginalized racial and ethnic populations.24,25 Race and ethnicity groups included American Indian or Alaska Native; Asian, Pacific Islander, or Native Hawaiian; Hispanic; non-Hispanic Black; non-Hispanic White; and multiple or other race and ethnicity (further delineation was not possible using the available data). Neighborhood disadvantage, associated with several poor health outcomes, was measured using US census data and deciles of neighborhood socioeconomic status.26 Veteran rurality was defined using a classification of rural, highly rural, or insular island based on Rural-Urban Commuting Area Codes and was included given the potential for transportation barriers.27 Medical comorbidities were measured using the Gagne score.28 All covariate data except for neighborhood disadvantage came from the national VA Corporate Data Warehouse, a repository of VA clinical and administrative data.
Statistical Analysis
We estimated that we would need to include 871 individuals to detect an effect size of 0.2 with 80% power and a 5% 2-sided significance level while also allowing for 10% attrition. We decided on a small effect size (Cohen h = 0.2) based on prior research suggesting modest effects of BCS outreach strategies. We examined descriptive characteristics of the sample by study assignment using the Pearson χ2 test for dichotomous variables and the Student t test for continuous variables.
We conducted an interim analysis to assess efficacy and futility of the opt-out strategy after 50% of the estimated sample size was enrolled and reached the primary end point (ie, 100 days after randomization). We calculated a z statistic from the proportional difference between the treatment groups and compared this z statistic to the futility boundary to determine whether it was crossed. We used a 1-sided nonbinding futility boundary of 0.764 (using an O’Brien-Fleming α spending design and Pocock-type β spending function).29,30,31 Interim analysis results did not meet the predefined significance threshold, so the trial was completed as planned.
We conducted an intention-to-treat (ITT) analysis using logistic regression adjusting for prior screening as a stratification variable and age, race and ethnicity, and Gagne score as precision variables.32 We defined the final significance level as 2-sided P < .048, which incorporates the interim analysis.29 We also conducted a restricted analysis that excluded veterans who were unable to be reached by telephone (eg, nonworking telephone number, hang-up) or who were found to be ineligible after randomization based on medical record review. In the restricted analysis, there were more opportunities for exclusion in the opt-out group because a nurse reviewed all medical records compared with the opt-in group in which a nurse reviewed medical records only of those who selected screening. The investigator responsible for statistical analysis (C.W.) was blinded to intervention group while conducting the analysis.
We conducted subgroup analyses to evaluate potential heterogeneity of intervention effects across race and ethnicity, neighborhood disadvantage, and veteran rurality by including interactions with the assigned treatment group and statistically testing the interactions. In the case of a low number of veterans within a category (<10), categories were pooled. We also conducted a subgroup analysis stratifying groups by age (45-54 years and ≥55 years), considering that ACS screening guidelines change starting at age 55 years.22 Participants were excluded from subgroup analyses if they were missing any data. We considered subgroup analyses exploratory and therefore made no adjustment for multiplicity.
Finally, we conducted a sensitivity analysis extending the follow-up period to 180 days. Given delays in scheduling and records retrieval for mammography performed outside the VHA, a cutoff of 100 days may underestimate effects of the intervention.3
All descriptive and inferential analyses were performed using R, version 4.1.1 or subsequent versions (R Project for Statistical Computing). We adjusted P values via the O’Brien-Fleming method to account for the interim analysis; P values for adjusted analyses were also adjusted using the Hochberg method to account for multiple comparisons.29,33 There were no missing data for primary or secondary outcomes; we conducted complete case analysis with respect to any missing covariate data. The upper limit for any missing covariate data was approximately 5%, except for neighborhood socioeconomic status (approxmiately 20%).
Results
Participant Characteristics
We included 883 veterans eligible for BCS (441 in the opt-out arm and 442 in the opt-in arm). The mean (SD) age was 59.13 (8.24) years. Forty-one veterans (4.6%) had missing race and ethnicity data; of the remaining 842, 13 (1.5%) were American Indian or Alaska Native; 60 (7.1%), Asian, Pacific Islander, or Native Hawaiian; 41 (4.9%), Hispanic; 123 (14.6%), Black non-Hispanic; 579 (68.8%), White non-Hispanic; and 26 (3.1%), multiple or other race and ethnicity. Veterans in the opt-out and opt-in groups were similar in all characteristics measured, and 328 veterans in each group (656 [74.3%] overall) had evidence of prior BCS (Table 1).
Table 1. Participant Characteristics.
| Variable | Veteransa | ||
|---|---|---|---|
| Overall (N = 883) | Opt-in (n = 442) | Opt-out (n = 441) | |
| Age, mean (SD), y | 59.13 (8.24) | 58.85 (8.09) | 59.40 (8.39) |
| Race and ethnicity, No./total No. (%) | |||
| American Indian or Alaska Native | 13/842 (1.5) | 6/422 (1.4) | 7/420 (1.7) |
| Asian, Pacific Islander, or Native Hawaiian | 60/842 (7.1) | 27/422 (6.4) | 33/420 (7.9) |
| Hispanic | 41/842 (4.9) | 22/422 (5.2) | 19/420 (4.5) |
| Non-Hispanic Black | 123/842 (14.6) | 63/422 (14.9) | 60/420 (14.3) |
| Non-Hispanic White | 579/842 (68.8) | 293/422 (69.4) | 286/420 (68.1) |
| Multiple or otherb | 26/842 (3.1) | 11/422 (2.6) | 15/420 (3.6) |
| Missing | 41/883 (4.6) | 20/442 (4.5) | 21/441 (4.8) |
| Marital status | |||
| Married | 326 (36.9) | 162 (36.7) | 164 (37.2) |
| Other | 557 (63.1) | 280 (63.3) | 277 (62.7) |
| Prior breast cancer screening | |||
| No | 227 (25.7) | 114 (25.8) | 113 (25.6) |
| Yes | 656 (74.3) | 328 (74.2) | 328 (74.4) |
| Gagne score, mean (SD) | 0.24 (1.03) | 0.24 (1.12) | 0.24 (0.94) |
| Socioeconomic status index, decile, No./total No. (%) | |||
| 0 | 14/683 (2.0) | 8/338 (2.4) | 6/345 (1.7) |
| 1 | 26/683 (3.8) | 14/338 (4.1) | 12/345 (3.5) |
| 2 | 75/683 (11.0) | 37/338 (10.9) | 38/345 (11.0) |
| 3 | 59/683 (8.6) | 33/338 (9.8) | 26/345 (7.5) |
| 4 | 97/683 (14.2) | 45/338 (13.3) | 52/345 (15.1) |
| 5 | 88/683 (12.9) | 39/338 (11.5) | 49/345 (14.2) |
| 6 | 106/683 (15.5) | 58/338 (17.2) | 48/345 (13.9) |
| 7 | 79/683 (11.6) | 43/338 (12.7) | 36/345 (10.4) |
| 8 | 71/683 (10.4) | 33/338 (9.8) | 38/345 (11.0) |
| 9 | 68/683 (10.0) | 28/338 (8.3) | 40/345 (11.6) |
| Missing | 200/883 (22.7) | 104/442 (23.5) | 96/441 (21.8) |
| Rurality, No./total No. (%) | |||
| Urban | 664/881 (75.4) | 341/442 (77.1) | 323/439 (73.6) |
| Rural | 208/881 (23.6) | 99/442 (22.4) | 109/439 (24.8) |
| Highly rural or insular islands | 9/881 (1.0) | 2/442 (0.5) | 7/439 (1.6) |
| Missing | 2/883 (0.2) | 0 | 2/441 (0.5) |
Data are presented as the number (percentage) of veterans unless otherwise indicated.
Further delineation of other races and ethnicities was not possible using the available data.
BCS Outreach
Fifty-three veterans in the opt-out group (12.0%) and 27 in the opt-in group (6.1%) were deemed ineligible through nurse medical record review and were not included in the restricted analysis. Among those in the opt-in group, 235 (53.3%) responded to the automated telephone call; of those, 126 (53.6%) selected option 1 to receive mammography, 61 (26.0%) selected option 2 to request a telephone call, and 48 (20.4%) selected option 3 to decline BCS (eTable 1 in Supplement 2).
BCS Scheduling and Completion
In the unadjusted ITT analysis, at 100 days after outreach, 67 veterans in the opt-out arm (15.2%) and 66 in the opt-in arm (14.9%) completed mammography (P = .90) (Table 2). When evaluating secondary outcomes, 106 veterans in the opt-out arm (24.0%) and 84 in the opt-in arm (19.0%) either completed or scheduled mammography (P = .07); 104 mammography referrals in the opt-out arm (23.6%) and 24 in the opt-in arm (5.4%) were canceled (P < .001).
Table 2. Unadjusted Primary and Secondary Outcomes at 100 Days After the Intervention.
| Outcome | Veterans, No./total No. (%) | P valuea | |
|---|---|---|---|
| Opt-in (n = 442) | Opt-out (n = 441) | ||
| Intention-to-treat analysis | |||
| Mammography completed | 66/442 (14.9) | 67/441 (15.2) | .90 |
| Mammography completed or scheduled | 84/442 (19.0) | 106/441 (24.0) | .07 |
| Mammography cancelled | 24/442 (5.4) | 104/441 (23.6) | <.001 |
| Restricted analysis | |||
| Mammography completed | 62/415 (14.9) | 64/388 (16.5) | .50 |
| Mammography completed or scheduled | 80/415 (19.3) | 102/388 (26.3) | .02 |
| Mammography cancelled | 23/415 (5.5) | 100/388 (25.8) | <.001 |
Pearson χ2 test.
In the restricted analysis, of the veterans who were eligible after medical record review and had accurate contact information, at 100 days after outreach, 64 of 388 veterans in the opt-out arm (16.5%) and 62 of 415 in the opt-in arm (14.9%) had completed mammography (P = .50) (Table 2). For secondary outcomes, more veterans in the opt-out arm (102 of 388 [26.3%]) either completed or scheduled mammography than in the opt-in arm (80 of 415 [19.3%]) (P = .02).
Adjusting for age, race and ethnicity, medical comorbidities, and prior BCS, we did not find a significant difference in the odds of BCS completion in the opt-out group compared with the opt-in group in either the ITT analysis (odds ratio [OR], 1.01; 95% CI, 0.69-1.49) or the restricted analysis (OR, 1.10; 95% CI, 0.74-1.63) (Table 3). There was also no significant difference in odds of scheduled or completed screening within 100 days in the adjusted ITT analysis (OR 1.34; 95% CI, 0.96-1.87). There remained a significant difference in odds of scheduled or completed screening within 100 days in the adjusted restricted analysis (OR, 1.47; 95% CI, 1.05-2.08). In subgroup analyses, we did not detect significant heterogeneity of treatment effects in mammography completion based on race and ethnicity, neighborhood disadvantage, rurality, or age.
Table 3. Adjusted Primary and Secondary Outcomes at 100 Days After the Intervention.
| Characteristic | Odds ratio (95% CI) | P value | q Valuea |
|---|---|---|---|
| Intention to treat | |||
| Mammography completed | |||
| Opt-in | 1 [Reference] | .90 | .90 |
| Opt-out | 1.01 (0.69-1.49) | ||
| Mammography completed or scheduled | |||
| Opt-in | 1 [Reference] | .08 | .08 |
| Opt-out | 1.34 (0.96-1.87) | ||
| Restricted analysis | |||
| Mammography completed | |||
| Opt-in | 1 [Reference] | .60 | .60 |
| Opt-out | 1.10 (0.74-1.63) | ||
| Mammography completed or scheduled | |||
| Opt-in | 1 [Reference] | .03 | .03 |
| Opt-out | 1.47 (1.05-2.08) | ||
False discovery rate correction for multiple testing.
Sensitivity Analysis
At 180 days after outreach, there were no significant differences between the opt-out and opt-in groups in either primary or secondary outcomes in both ITT and restricted analyses (eTable 2 in Supplement 2). Adjusted analysis findings were similar to the unadjusted findings (eTable 3 in Supplement 2).
Discussion
This pragmatic randomized clinical trial compared the effect of an opt-out approach with that of an opt-in outreach approach on mammography completion for BCS. Overall mammography completion was low; however, rates were similar or modestly lower compared with other screening interventions among patients overdue for screening.12,18,19,34 We did not find differential response rates between the 2 arms in either the ITT analysis or the restricted analysis. Proportions of patients with scheduled or completed mammography (a secondary outcome) at 100 days were larger in the opt-out group than in the opt-in group in the restricted analysis but not the ITT analyses. There were substantially more canceled mammography referrals in the opt-out group than in the opt-in group. Our results were robust to a sensitivity analysis in which we evaluated outcomes 180 days after initial outreach.
Our findings differ from those of studies showing benefits of opt-out approaches for other types of cancer screening outreach.14,15,16 Comparable screening rates among those receiving the opt-in strategy may in part be due to the design, which required individuals to make an active choice to screen. Prior research demonstrated that active choice or affirming decisions were associated with increased likelihood of follow-through and may be more effective than opt-in strategies without forced choices.35 Another reason for conflicting results may be the complexity of BCS at VAPS without on-site mammography. The complicated ordering and completion process may attenuate the benefit of an opt-out strategy compared with simpler screening processes (eg, self-testing for colorectal cancer screening). Future research should evaluate opt-out approaches in health systems with integrated BCS, including VA sites with on-site mammography.
The administrative burden of the opt-out approach, including medical record review of all veterans prior to outreach to confirm BCS eligibility, at least 1 telephone call to eligible veterans, and an increased number of canceled referrals following outreach, likely outweighs any potential added benefit. Notably, following the study, primary care leadership at VAPS reported that the opt-out strategy was not sustainable with current staffing. The automated telephone outreach received by those in the opt-in group required little staffing or administrative burden; only 126 (53.3%) of the veterans who responded to the telephone call were interested in BCS, received medical record review, and were contacted for scheduling, and an additional 61 (26.0%) requested a telephone call for questions or to report prior screening. We did not find any significant heterogeneity of treatment effect based on race and ethnicity, neighborhood socioeconomic status, or rurality, although further dedicated evaluation is warranted. While our results did not indicate greater effectiveness of an opt-out strategy for BCS compared with an opt-in strategy, the comparable results for a less burdensome outreach approach may be valuable to VA operational leaders interested in population-based BCS outreach.
Limitations
The study findings should be interpreted considering several limitations. We did not have sufficient nursing staff time to complete medical record reviews for veterans who did not select BCS in the opt-in group, reflecting clinical constraints in a pragmatic trial with limited resources. Likely as a result, there were more individuals excluded from the opt-out group than from the opt-in group in the restricted analysis. Considering the stability of findings in restricted and sensitivity analyses, we anticipate that the difference in exclusions across groups is unlikely to appreciably change conclusions. The opt-out group did not receive a message preceding the scheduling call, which could have made the intervention more effective. Advance messaging, known as prenotification, has been shown in other preventive care services to improve engagement.36 Subanalyses of age, race and ethnicity, neighborhood socioeconomic status, and rurality are limited in interpretability as they were both exploratory and underpowered to detect significant heterogeneity.
Conclusions
This randomized clinical trial did not find higher mammography completion for opt-out vs opt-in outreach among veterans receiving primary care at a large VA medical center, and the opt-out approach was significantly more burdensome. The findings suggest that health systems should consider both potential effects and excess administrative burden when deciding between opt-out and opt-in outreach strategies.
Trial Protocol and Statistical Analysis Plan
eTable 1. Opt-in Group Responses to Automated Phone Call Outreach
eTable 2. Sensitivity Analysis and Primary and Secondary Outcomes at 180 Days
eTable 3. Sensitivity Analysis and Adjusted Primary and Secondary Outcomes at 180 Days After Intervention
Data Sharing Statement
Footnotes
VA indicates Veterans Affairs.
References
- 1.US Department of Veterans Affairs Office of Analytics and Performance Integration. Quality and patient safety. Accessed June 24, 2023. https://www.va.gov/QUALITYANDPATIENTSAFETY/api/index.asp
- 2.Meyer AND, Singh H, Zimolzak AJ, et al. Cancer evaluations during the COVID-19 pandemic: an observational study using national Veterans Affairs data. Am J Prev Med. 2022;63(6):1026-1030. doi: 10.1016/j.amepre.2022.07.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Monty GR, Benson SK, Deeds SA, et al. “We are working harder, not smarter”: a qualitative inquiry into care coordination for Department of Veterans Affairs mammograms referred to the community. Womens Health Issues. 2023;33(4):414-421. doi: 10.1016/j.whi.2022.11.004 [DOI] [PubMed] [Google Scholar]
- 4.Kvasnovsky CL, Kesmodel SB, Gragasin JL, et al. Expansion of screening mammography in the Veterans Health Administration: implications for breast cancer treatment. JAMA Surg. 2013;148(11):999-1004. doi: 10.1001/jamasurg.2013.3738 [DOI] [PubMed] [Google Scholar]
- 5.Ferras M, Dye J, Ayala GX, Schmied E. An examination of factors that influence receipt of reproductive health screenings among female veterans. Mil Med. 2023;188(1-2):42-48. doi: 10.1093/milmed/usac036 [DOI] [PubMed] [Google Scholar]
- 6.Lu M, Moritz S, Lorenzetti D, Sykes L, Straus S, Quan H. A systematic review of interventions to increase breast and cervical cancer screening uptake among Asian women. BMC Public Health. 2012;12(1):413. doi: 10.1186/1471-2458-12-413 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Copeland VC, Kim YJ, Eack SM. Effectiveness of interventions for breast cancer screening in African American women: a meta-analysis. Health Serv Res. 2018;53(Suppl)(suppl 1):3170-3188. doi: 10.1111/1475-6773.12806 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Lairson DR, Chan W, Chang YC, del Junco DJ, Vernon SW. Cost-effectiveness of targeted versus tailored interventions to promote mammography screening among women military veterans in the United States. Eval Program Plann. 2011;34(2):97-104. doi: 10.1016/j.evalprogplan.2010.07.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Vernon SW, del Junco DJ, Tiro JA, et al. Promoting regular mammography screening II: results from a randomized controlled trial in US women veterans. J Natl Cancer Inst. 2008;100(5):347-358. doi: 10.1093/jnci/djn026 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Dalessandri KM, Cooper M, Rucker T. Effect of mammography outreach in women veterans. West J Med. 1998;169(3):150-152. [PMC free article] [PubMed] [Google Scholar]
- 11.Emanuel EJ, Ubel PA, Kessler JB, et al. Using behavioral economics to design physician incentives that deliver high-value care. Ann Intern Med. 2016;164(2):114-119. doi: 10.7326/M15-1330 [DOI] [PubMed] [Google Scholar]
- 12.Winer RL, Lin J, Tiro JA, et al. Effect of mailed human papillomavirus test kits vs usual care reminders on cervical cancer screening uptake, precancer detection, and treatment: a randomized clinical trial. JAMA Netw Open. 2019;2(11):e1914729. doi: 10.1001/jamanetworkopen.2019.14729 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Mehta SJ, Khan T, Guerra C, et al. A randomized controlled trial of opt-in versus opt-out colorectal cancer screening outreach. Am J Gastroenterol. 2018;113(12):1848-1854. doi: 10.1038/s41395-018-0151-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Huf SW, Asch DA, Volpp KG, Reitz C, Mehta SJ. Text messaging and opt-out mailed outreach in colorectal cancer screening: a randomized clinical trial. J Gen Intern Med. 2021;36(7):1958-1964. doi: 10.1007/s11606-020-06415-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Samuelson W, Zeckhauser R. Status quo bias in decision making. J Risk Uncertain. 1988;1(1):7-59. doi: 10.1007/BF00055564 [DOI] [Google Scholar]
- 16.Chapman GB, Li M, Colby H, Yoon H. Opting in vs opting out of influenza vaccination. JAMA. 2010;304(1):43-44. doi: 10.1001/jama.2010.892 [DOI] [PubMed] [Google Scholar]
- 17.Nelson K, Reddy A, Stockdale SE, et al. The primary care analytics team: integrating research and clinical care within the Veterans Health Administration Office of Primary Care. Healthc (Amst). 2021;8(suppl 1):100491. doi: 10.1016/j.hjdsi.2020.100491 [DOI] [PubMed] [Google Scholar]
- 18.Deeds S, Liu T, Schuttner L, et al. A postcard primer prior to mailed fecal immunochemical test among veterans: a randomized controlled trial. J Gen Intern Med. 2023;1-7. Published online June 8, 2023. doi: 10.1007/s11606-023-08248-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Deeds SA, Moore CB, Gunnink EJ, et al. Implementation of a mailed faecal immunochemical test programme for colorectal cancer screening among veterans. BMJ Open Qual. 2022;11(4):e001927. doi: 10.1136/bmjoq-2022-001927 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Hopewell S, Clarke M, Moher D, et al. CONSORT for reporting randomized controlled trials in journal and conference abstracts: explanation and elaboration. PLoS Med. 2008;5(1):e20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.US Department of Veterans Affairs Office of Women’s Health . Breast care registry. Accessed July 26, 2023. https://catalog.data.gov/dataset/breast-care-registry-bcr
- 22.Oeffinger KC, Fontham ETH, Etzioni R, et al. ; American Cancer Society . Breast cancer screening for women at average risk: 2015 guideline update from the American Cancer Society. JAMA. 2015;314(15):1599-1614. doi: 10.1001/jama.2015.12783 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Perlin JB, Kolodner RM, Roswell RH. The Veterans Health Administration: quality, value, accountability, and information as transforming strategies for patient-centered care. Am J Manag Care. 2004;10(11 Pt 2):828-836. [PubMed] [Google Scholar]
- 24.Hernandez SE, Sylling PW, Mor MK, et al. Developing an algorithm for combining race and ethnicity data sources in the Veterans Health Administration. Mil Med. 2020;185(3-4):e495-e500. doi: 10.1093/milmed/usz322 [DOI] [PubMed] [Google Scholar]
- 25.DeSantis CE, Ma J, Goding Sauer A, Newman LA, Jemal A. Breast cancer statistics, 2017, racial disparity in mortality by state. CA Cancer J Clin. 2017;67(6):439-448. doi: 10.3322/caac.21412 [DOI] [PubMed] [Google Scholar]
- 26.Nelson K, Taylor L, Lurie N, Escarce J, McFarland L, Fihn SD. Neighborhood environment and health status and mortality among veterans. J Gen Intern Med. 2011;26(8):862-867. doi: 10.1007/s11606-011-1710-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.US Dept of Agriculture Economic Research Service. Rural-Urban Commuting Area Codes. Accessed March 21, 2023. https://www.ers.usda.gov/data-products/rural-urban-commuting-area-codes.aspx
- 28.Gagne JJ, Glynn RJ, Avorn J, Levin R, Schneeweiss S. A combined comorbidity score predicted mortality in elderly patients better than existing scores. J Clin Epidemiol. 2011;64(7):749-759. doi: 10.1016/j.jclinepi.2010.10.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.O’Brien PC, Fleming TR. A multiple testing procedure for clinical trials. Biometrics. 1979;35(3):549-556. doi: 10.2307/2530245 [DOI] [PubMed] [Google Scholar]
- 30.Li X, Herrmann C, Rauch G. Optimality criteria for futility stopping boundaries for group sequential designs with a continuous endpoint. BMC Med Res Methodol. 2020;20(1):274. doi: 10.1186/s12874-020-01141-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Wassmer G, Pahlke F. rpact: Confirmatory adaptive clinical trial design, simulation, and analysis. Accessed March 21, 2023. https://www.rpact.org
- 32.Holmberg MJ, Andersen LW. Adjustment for baseline characteristics in randomized clinical trials. JAMA. 2022;328(21):2155-2156. doi: 10.1001/jama.2022.21506 [DOI] [PubMed] [Google Scholar]
- 33.Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc B. 1995;57(1):289-300. doi: 10.1111/j.2517-6161.1995.tb02031.x [DOI] [Google Scholar]
- 34.Bucher A, Blazek ES, West AB. Feasibility of a reinforcement learning-enabled digital health intervention to promote mammograms: retrospective, single-arm, observational study. JMIR Form Res. 2022;6(11):e42343. doi: 10.2196/42343 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Keller PA, Harlam B, Loewenstein G, Volpp KG. Enhanced active choice: a new method to motivate behavior change. J Consum Psychol. 2011;21(4):376-383. doi: 10.1016/j.jcps.2011.06.003 [DOI] [Google Scholar]
- 36.Milkman KL, Beshears J, Choi JJ, Laibson D, Madrian BC. Using implementation intentions prompts to enhance influenza vaccination rates. Proc Natl Acad Sci U S A. 2011;108(26):10415-10420. doi: 10.1073/pnas.1103170108 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Trial Protocol and Statistical Analysis Plan
eTable 1. Opt-in Group Responses to Automated Phone Call Outreach
eTable 2. Sensitivity Analysis and Primary and Secondary Outcomes at 180 Days
eTable 3. Sensitivity Analysis and Adjusted Primary and Secondary Outcomes at 180 Days After Intervention
Data Sharing Statement