Theory‐based behavior change intervention to increase uptake of risk‐reducing salpingo‐oophorectomy in women with a BRCA1 or BRCA2 pathogenic variant: The PREVENT randomized controlled trial

Abstract Objective To evaluate the effect of a theory‐based behavioral intervention delivered by genetic counselors on the uptake of risk‐reducing salpingo‐oophorectomy (RRSO) at 12 and 24 months by women with a BRCA1 or BRCA2 pathogenic variant (PV) compared to women who received usual care. Methods In this two‐arm, multi‐site randomized controlled trial participants were randomized to receive a theoretically‐guided behavioral telephone intervention or usual care. Outcome data were collected at 12 and 24 months. Participants in the usual care arm were offered the intervention after 12 months. Results Data on 107 participants were included in the analysis. There was no significant difference in the proportion of women who had a RRSO by 1 year (28.6%‐ intervention; 22.9%‐ usual care (p = 0.54)). At 1 year, women who received the intervention had significantly lower mean decisional conflict (p interaction <0.001) and a higher mean knowledge score at one‐year compared to usual care (p interaction <0.001). At 2 years, 53.9% of participants in the intervention arm had RRSO compared to 32.6% in usual care (p = 0.05). Conclusions A theory‐based behavioral intervention delivered by genetic counselors to women with a BRCA PV who chose not to have the recommended RRSO was effective at reducing decisional conflict and increasing knowledge in women with a BRCA1 or BRCA2 PV.


| INTRODUCTION
Women with a pathogenic or likely pathogenic variant (PV) in BRCA1 or BRCA2 have high lifetime risks of developing ovarian cancer.The cumulative risk of ovarian cancer to age 80 years for women with a BRCA1 PV is 44%, and for BRCA2 is 17%. 1 This can be compared to a 1% risk of ovarian cancer in the general population. 2 For women with a BRCA mutation, the 10-year survival rate for clinically detected ovarian cancer is only 35%. 3 As a result, managing the risk of ovarian cancer is critical for women who have a PV in BRCA1 or BRCA2.
Risk-reducing salpingo-oophorectomy (RRSO) is the removal of the ovaries and fallopian tubes.In women with a PV in BRCA1 or BRCA2, this preventive surgery reduces the risk of ovarian/peritoneal/fallopian tube cancer by 80%, and reduces the risk of all-cause mortality by 77%. 3 Ovarian screening with transvaginal ultrasound is not effective in reducing the risk of death from ovarian cancer.When comparing screening ultrasound to RRSO in women with a BRCA1 PV, RRSO offers a 76% reduction in mortality. 4urrently, the National Comprehensive Cancer Network (NCCN) recommends RRSO for management of ovarian cancer risk in women with a BRCA PV.For women with a BRCA1 PV, the recommended age for preventive surgery is between 35 and 40 years, and for BRCA2 is 40-45 years.However, uptake of RRSO in these high-risk women is not ideal.In a large international study of 6223 women with a BRCA1 or BRCA2 PV from 10 countries, we reported a RRSO uptake rate of 65%. 5 This low uptake rate may be a result of our current models of genetic testing, which may or may not include post-test genetic counseling and ongoing follow-up of women with a BRCA1 or BRCA2 PV.It is unclear if follow-up genetic or behavioral counseling would increase uptake of RRSO in this population.Kinney et al. (2014) have previously reported on a randomized controlled trial (TeleCARE) conducted in the United States to evaluate the effect of a remote risk communication and behavior change intervention among individuals at a high-risk of developing colorectal cancer based on family history. 6They compared the efficacy of a theory based telephone psychoeducational intervention delivered by genetic counselors with a mailed educational brochure for improving uptake of colonoscopy screening among at-risk relatives of patients with colorectal cancer.Within 9 months of randomization, 35% of individuals in the behavior change counseling intervention arm had undergone colonoscopy compared to 16% in the usual care arm (OR 2.83, p < 0.001).
In the current study, we adapted TeleCARE 6 to evaluate the effect of a theory-based behavioral intervention delivered by genetic counselors over the telephone on the uptake of prophylactic RRSO at 12 and 24 months by women with a BRCA1 or BRCA2 PV compared to women with a PV who received usual care.

| Study design and hypotheses
In this two-arm, multi-site randomized controlled trial (RCT), participants were randomized to the intervention arm or the usual care arm.Outcomes were measured at 12 months.For those in the control arm (usual care), the intervention was offered after completion of the 12 month outcome measures.All participants who had not undergone RRSO were contacted at 24 months to collect information on RRSO scheduling and/or completion.
We hypothesized that women who received the follow-up theoretically-guided behavioral telephone genetic counseling intervention would have higher uptake of RRSO at 1 year compared to women who received usual care.We also hypothesized that at 1 year after receiving the intervention, women would have lower decisional conflict and cancer-related distress, and higher knowledge levels compared to women who received usual care.We also evaluated an exploratory aim recognizing that booking a gynecological consult, electing for surgery, and booking surgery may take longer than 1 year.For this exploratory aim, we hypothesized that 2 years after randomization, those who received the intervention would have higher uptake of RRSO compared to those who received usual care.The trial was approved by Research Ethics Boards at all participating centres.

| Study participants
Eligible study participants included women enrolled in a multicenter, longitudinal study of BRCA1 and BRCA2 mutation carriers at 5 sites (Women's College Hospital; Grandriver Hospital; London Health Sciences Centre; Princess Margaret Hospital; Moffitt Cancer Center) in addition to self-referrals through advertisements with commercial genetic testing laboratories and an online support group (Facing Our Risk of Cancer Empowered-FORCE).
Eligibility criteria included the following: ( (4) at least 12 months since genetic testing or most recent contact by parent follow-up study; (5) provided written consent to be re-contacted for future research; and (6) could speak and understand English.
Women were excluded if: (1) currently receiving treatment for another cancer diagnosis (including breast cancer); (2) pregnant; (3) given birth in the last 6 months; or (4) were scheduled for or already had a surgical date for RRSO.

| Screening and random assignment
Eligible participants were identified in the clinic databases at the participating centres and contacted by telephone.Written consent was obtained, and following the online collection of baseline data, all eligible, consented women were randomly allocated to either the intervention group (follow-up telephone genetic counseling (FTGC) intervention) or the usual care group.Randomization was centrally controlled using a web-based randomization service (www.randomize.net) with stratification based on age, gene (BRCA1 versus BRCA2), and site of recruitment.

| Data collection and primary outcome assessment
A research assistant blinded to group allocation contacted all participants by telephone at 12 months postrandomization to collect outcome data (primary and secondary outcomes).After completion of follow-up questionnaires, women in the usual care group were offered the follow-up telephone genetic counseling intervention.At 2 years, all participants who had not undergone RRSO at year one were contacted over the telephone by the research assistant to collect data on uptake or scheduling of RRSO (self-reported by participant).

| Primary outcome
Uptake of RRSO.The primary outcome was a yes/no response to uptake of RRSO (scheduled or completed) as identified by the Follow-up Questionnaire administered at 12 months post-randomization.This questionnaire included questions regarding any new diagnoses of cancer, uptake of risk-reducing surgery, and intent for risk-reducing surgery.If a participant answered no to having completed a RRSO, then they were asked if they had a RRSO scheduled.If yes, then they were classified as positive for uptake of RRSO.

Decisional conflict
Decisional conflict was assessed using the decisional conflict scale (DCS). 7The DCS measures a person's perception of the difficulty of making a decision (in this case about RRSO).The questionnaire consists of 16 items with response options ranging from 0 (strongly agree) to 4 (strongly disagree).Item responses are summed and scaled to create a total score ranging from 0 (no decisional conflict) to 100 (extremely high level of decisional conflict).Scores lower than 25 are associated with implementing decisions and scores exceeding 37.5 are associated with delay or feeling unsure about implementation. 8n addition to a total score, five subscale scores can be calculated representing uncertainty, information, values clarification, support and effective decision.In the trial sample, the alpha for the total scale was 0.95 or higher at all time points.

Cancer risk and prevention knowledge
Cancer risk and prevention knowledge was assessed using a knowledge index that was developed and tested by our group and used in previous studies involving women with BRCA pathogenic variants. 9,10This index included items regarding risks of breast and ovarian cancer associated with having a BRCA1/2 mutation, risk reductions associated with preventive options, and knowledge questions regarding each of the preventive options (risk-reducing mastectomy, RRSO and chemoprevention).Scores were calculated by summing the number of correct answers, and total scores could range between 0 and 8.

Cancer-related Distress
Cancer-related Distress was assessed using the Impact of Event Scale (IES), 11 a self-report measure designed to measure current subjective distress about a specific stressor, in this case, risk of cancer.The scale consists of 15 items: 7 intrusion items and 8 avoidance items.Participants rate the frequency of intrusive and avoidant behaviors using a four-point frequency scale (0 = not at all, 1 = rarely, 3 = sometimes, 5 = often).Scores range from 0 to 75, with separate intrusion and avoidance sub-scales scores (with a possible range of 0-35 for intrusion, and 0-40 for avoidance).Lower scores depict lower levels of distress.In the trial sample, alpha was at least 0.91 and 0.88 for intrusion and avoidance, respectively, at all time points.

Uptake of RRSO at 2 years
An exploratory outcome was a yes/no response to uptake of RRSO (scheduled or completed) as identified by the follow-up questionnaire administered at 2 years postrandomization.Only participants who had not completed a RRSO at 12 months were contacted at 2 years.Similar to as described above for the primary outcome measure, women were classified as positive for uptake of RRSO if having undertaken or scheduled an appointment for the procedure.

| Study intervention
Participants assigned to the intervention group received an intervention that was based on the TeleCARE intervention, that was shown to be effective at increasing uptake of colorectal screening in high-risk relatives. 6,12This intervention integrates the Extended Parallel Process Model (EPPM) and motivational interviewing.It is based on theoretical constructs that are hypothesized to motivate an individual with a BRCA mutation to undergo prophylactic RRSO (i.e., threat and efficacy perceptions and behavioral intentions).
This theory-based behavioral telephone follow-up genetic counseling (TFGC) intervention consists of several components: All women randomized to the intervention completed the Risk Behavior Diagnosis Scale (RBDS) 13 to assess perceptions about cancer risk and severity, response efficacy and self-efficacy related to uptake of RRSO.The participant was mailed an educational brochure and tailored visual aids to be used during the telephone counseling session, which were tailored to each participant based on RBDS scores.• Part 2: Telephone Follow-up Genetic Counseling (TFGC) Intervention Delivery.
The intervention delivered by genetic counselors was based on the EPPM, implementation intentions and motivational interviewing principles.To standardize the counseling sessions, all genetic counselors were trained and supervised by a motivational interviewing and EPPM expert.Each counselor completed 20 h of training time before delivering the intervention.In the month following the training, counselors submitted four practice tapes for coding and quality control, and participated in four weekly supervision sessions with the trainer.Tapes were coded using the Motivational Interviewing Treatment Integrity (MITI) coding system, a widely used measure of MI adherence. 14To ensure treatment fidelity, during the first 6 months of the intervention period, a random sample of tapes (one per counselor, per month) were coded, with summary scores provided to counselors on intervention delivery.
One trained genetic counselor from each participating cancer genetics clinic delivered the intervention to ensure that participants continued to have ongoing relationships with their original cancer genetics clinic.
The intervention incorporated risk communication and behavior change approaches designed to: (1) raise awareness of threats about BRCA-associated breast and ovarian cancers; (2) raise awareness of the benefits of RRSO; and (3) address barriers to help increase motivation and self-efficacy to elect for the preventive surgery.The counselor tailored the counseling session based on the participants' RBDS scores.As a component of the counseling, participants were asked if they would like a referral to a local gynecological surgeon to discuss RRSO.A tailored letter was sent to the participant describing the current recommendations and details of the referral if appropriate.

| Usual care
Women allocated to the usual care arm had access to standard post-genetic testing care provided to all women identified with a BRCA PV (which includes referral to appropriate specialists).After the assessment at 12-month post-randomization, women in usual care were offered the intervention.

| Statistical analysis
The analyses used an intent to treat (ITT) approach and included all eligible participants who were randomly assigned to a treatment arm.Balance in background clinical and demographic covariates was assessed across arms using t-tests and Chi-squared tests or Fisher's Exact tests as required.The primary outcome of the proportion of participants who received or scheduled a RRSO by the 1-year follow-up was assessed using logistic regression.For those with an unknown outcome due to loss to follow-up, negative and multiple imputations methods were undertaken and compared to the known outcomes analysis.Negative imputation assumes that if no record of the outcome is observed then the procedure was not scheduled or undertaken.Multiple imputation (MI) used information on verified RRSO by year 1 and background variables, including age, education level, income level, presence of children, breast cancer diagnosis, mother had ovarian cancer (yes/ no), annual ovarian screening (yes/no) to replace missing outcome values.Using the fully conditional method (FCS), which implements chained equations, 20 models were imputed.
The differences by treatment arm in secondary psychosocial outcomes were assessed using Linear Mixed Models (LMM), which upholds the ITT principle by using all available data.This method is comparable to ANCOVA with MI but provides more power, especially as the amount of missing data increases. 15An unstructured covariance matrix was used with the Kenward-Rodger approximation method; the latter tends of perform better under small sample conditions. 16Least squares means (LSM) derived from the LMM are presented with 95% confidence intervals.Raw pre-and post-test scores by treatment arm are presented for descriptive purposes.
A post hoc exploratory analysis was undertaken to determine and compare across arms the proportion of participants who received or scheduled a RRSO by 2 years post-randomization and the incidence of received or scheduled RRSO between 1 and 2 years postrandomization.For the latter calculation, only those who continued to be "at risk" were included in the analysis (i.e., those who had/scheduled a RRSO or who dropped out by year 1 were removed).Similar methods for modeling and dealing with missing data were used as for the primary outcome analysis.
All analyses were undertaken using SAS (version 9.4; SAS Institute).To evaluate the primary outcomes, alpha was set at 0.05.For secondary outcomes, to accommodate multiple testing, the alpha was set to 0.005 using a Bonferroni correction.

| Sample size
In the previous follow-up telephone-based genetic counseling intervention by Kinney and colleagues, uptake of colonoscopy was 15% in the usual care group and 35% in the intervention group.Since uptake of RRSO was not expected to be as high as colonoscopy (which is a screening intervention rather than a prophylactic surgical intervention), we based our sample size calculation on the ability to detect a slightly smaller difference of approximately 15%.Sample sizes of 121 in the intervention arm and 121 in the usual care arm were required to detect this difference with 80% power, using a two-sided Z test with pooled variance and alpha set at 0.05.To account for 20% loss to follow-up, a total sample size of 304 (152 per group) was targeted.Due to recruitment challenges (including lower than expected number of eligible participants at the study sites), the required sample size was not achieved.With the sample size recruited, we had 46% power to detect a difference of 15% at 1 year of follow-up.

| Study cohort
The CONSORT flowchart is shown in Figure 1.A total of 167 women were deemed eligible to participate in the study, of which 116 were enrolled.Nine individuals were lost between enrollment and baseline due to withdrawal, loss to follow up or death, resulting in a total of 107 participants who were included in the analysis in one of the two treatment arms: 53 in the intervention arm and 54 in the usual care arm.Overall, 94% of those in the intervention arm received the behavioral intervention delivered by genetic counselors.
A total of 107 participants were included in the outcome analysis.For 17 of these participants, no outcome data were available at 1 year, resulting in a retention rate of 84.1%; there was no significant difference in retention by treatment arm (p = 0.46).Those with follow-up data at 1 year and those without did not differ significantly on any key demographic or clinical background variables.
Table 1 shows the baseline demographic and clinical characteristics by treatment arm.The mean age of participants was 45.2 (SD = 8.5) years.Overall, 57% of participants had a BRCA1 mutation, and 43% had a BRCA2 mutation.A total of 63% of participants had previously consulted with a surgeon about RRSO and 17% had a mother with ovarian cancer.
There were no differences in clinical characteristics between arms.A single demographic variable was different across arms: a higher proportion of individuals in the intervention arm had children compared to the usual care arm (p = 0.024); however, there was no difference across arms in intention to have future children (p = 0.323).Because having children versus not having children may affect decisions regarding undertaking a RRSO, this variable was adjusted for in the modeling.As this was not a planned adjustment, both unadjusted and adjusted models are presented.

| Primary outcome
There was no difference in the proportion of women who scheduled or underwent a RRSO by the 1 year follow-up.For those with a known outcome, the raw proportions of RRSO were 28.6% in the intervention arm and 22.9% in the usual care arm (p = 0.54).The expected difference of 15% between arms was not found at 1 year.Similar results were observed across imputation methods and whether adjustment for baseline imbalances were undertaken (Table 2).

| Secondary outcomes
Figure 2 shows the raw scores of the secondary measures by treatment arm at pre-and post-test for descriptive purposes.Table 3 shows the ITT analyses, which are the focus of the results presented below.

Decisional conflict
As shown by the linear mixed modeling, the intervention arm had a significantly lower mean decisional conflict at year 1 (LSM =17.5; 95% CI = 11.2-23.8)compared to the usual care arm (LSM = 30.6;95% CI 25.0-36.2;p interaction <0.001).Using the adjusted alpha of 0.005, three decisional conflict subscales were also significantly lower in the intervention arm at follow up: uncertainty, support and effective decision (Table 3).The unadjusted estimates were highly consistent with the adjusted estimates and can be found in the Supplemental File.

Cancer risk & prevention knowledge
The raw baseline knowledge score was 6.3 (SD = 1.3) for the intervention arm and 6.3 (SD = 1.2) for the usual care arm.The linear mixed modeling found that the intervention arm had a significantly higher mean knowledge score at one-year (LSM = 7.0, 95% CI = 6.6-7.5)compared to the usual care arm (LSM = 6.2; 95% CI = 5.8-6.6;p interaction <0.001).

Cancer-related distress
The raw baseline IES score was 18.4 (SD = 15.8) for the intervention arm and 17.1 (SD = 17.9) for the usual care arm.Based on the linear mixed modeling, the intervention arm had a non-significantly lower mean total IES score at year 1 (LSM = 15.1, 95% CI = 9.5-20.7)compared to the usual care arm (LSM = 17.2, 95% CI = 12.4-22.1;p interaction = 0.32).Neither IES subscale was significantly different by treatment arm (Table 3).

| Exploratory outcome
Based on the original arm allocation, the proportion of women who scheduled or underwent a RRSO by the year 2 follow-up was higher in the intervention arm compared to the usual care arm based on analysis of known outcomes (p = 0.05).All participants who reported that they had scheduled a RRSO at year 1 had undergone RRSO by year 2. The raw proportions of RRSO at year 2 were 53.9% in the intervention arm and 32.6% in the

F I G U R E 2
Comparison of raw scores of secondary measures between treatment arms at pre-and post-test.p value is derived ANOVCA.
T A B L E 3 Differences in secondary outcomes at follow up by treatment arm using all available data in linear mixed modeling.control group, exceeding the 15% expected difference (Table 4).Results were similar across methods of dealing with missing data.When looking at the incidence of RRSO between year 1 and year 2, there was a significant difference between arms, with the intervention arm having a higher proportion.The raw proportions of RRSO between year 1 and year 2 were 33.3% in the intervention arm and 9.4% in the usual care arm (p = 0.03).The significant difference was found with adjustment and across both methods of dealing with missing data.Of those in the usual care arm with a known outcome who received the intervention after year 1 (and did not have a RRSO by year 1), 2 received/scheduled a RRSO out of 22.This is compared to 1 out of 10 with a known outcome in the usual care arm who never received the intervention after year 1 (p = 0.94).

| DISCUSSION
For women with a BRCA1 or BRCA2 PV, risk-reducing salpingo-oophorectomy (RRSO) between the ages of 35 and 45 years is recommended in international guidelines.However, uptake is not ideal. 5In the current study we evaluated a theory-based behavioral intervention delivered by genetic counselors that integrated the Extended Parallel Process Model and motivational interviewing focusing on RRSO for women with a BRCA PV who had not undergone the recommended preventive surgery.After 1 year, there was no significant difference in uptake of RRSO in women who received the intervention compared to those who received usual care.After 1 year, women who received the intervention had significantly lower levels of decisional conflict and higher knowledge levels compared to those who received usual care.In an exploratory analysis, women who received the intervention had significantly greater uptake of RRSO at 2 years compared to those who received usual care.8][19][20][21][22] Physical and mental health-related quality of life is similar to that of population controls before and after surgery. 23As a result of these possible side-effects, some women struggle with the decision to have RRSO.Uncertainty about the course of action given competing priorities such as risk, loss, regret or challenges to personal life values is referred to as decisional conflict. 24In lay terms, decisional conflict is indicative of the level of comfort an individual feels in making a decision. 25Decisional conflict scores are highly predictive of the extent to which individuals make healthrelated decisions. 26A B L E 4 Intervention effect on BSO booking/uptake using logistic regression between year 1 and year 2 follow-up based on different methods of dealing with missing data.In the current study, women had elevated levels of decisional conflict related to RRSO at baseline.These elevated levels of decisional conflict were observed again at 1 year for women randomized to usual care.However, for women who received the intervention, decisional conflict was significantly lower at 1 year.Women in the intervention arm went from scores that suggested that a decision about RRSO would not be made, to a mean score suggesting that a decision would be made.This suggests that the behavioral intervention may have targeted domains specific to decisional conflict, including knowledge, harms and benefits of options, and support. 25nowledge is a modifiable decisional need that contributes to decisional conflict. 24In the current study, knowledge was significantly higher at one-year for women who received the intervention compared to those who received usual care.The intervention included educational components about cancer risk and severity.Key components of the intervention were based on the Extended Parallel Process Model (EPPM). 27According to the EPPM, people who believe that the recommended action (e.g., RRSO) is effective at preventing the threat (e.g., breast and ovarian cancer, death) and have confidence in their ability to implement the action (high self-efficacy) are better able to deal with the threat by using danger control strategies.In the current study, this was accomplished through acknowledging and following through with the recommended action (RRSO).

Incidence
At one-year post-randomization, there were no differences in RRSO uptake between the intervention and usual care groups (28.6% vs. 22.9%; p = 0.54).However, in an exploratory analysis there was a significant difference between the groups at 2 years post-randomization.The raw proportions of RRSO were 53.9% for those who received the intervention and 32.6% for those who received usual care.This unexpected finding surpassed the 15% expected difference in uptake between the two groups based on the TeleCARE intervention study results.Some women in the control arm did elect to receive the intervention after 12 months, and there was no difference in uptake of RRSO by 24 months between those who received the intervention and those who did not.Although this study was not designed to measure the primary outcome at 2 years, these data suggests that future evaluations of uptake should incorporate longer follow-up to ensure that women have adequate time to be referred for a gynecological consult and book this elective surgery.Wait times for consultations and surgery vary by location and may have contributed to lower than expected uptake rates of RRSO at 1 year.
Most women who learn that they have a BRCA1 or BRCA2 PV do not receive long-term genetic counseling follow-up.This may contribute to the less than ideal uptake rates of RRSO in this population.In the current study, we demonstrated that a follow-up theory-based behavioral intervention delivered by genetic counselors increased RRSO among women who had not elected for the procedures after receiving their initial positive genetic test results.However, this may require additional genetic counseling resources which may be challenging in the current clinical environment.Additional follow-up interventions require evaluation, including provision of the intervention by other healthcare providers including specially trained nurses or other healthcare professionals.
There are several strengths to this study including the rigorous methodology employed in the randomized controlled trial, including stratification by site of referral.We recruited women from throughout North America using a variety of sources, including self-referral which contributed to the generalizability of our findings.We used a theoretically-derived intervention with relatively rigorous supervision and quality control.
There are limitations to this study, including the small sample size that did not meet the planned sample size requirements (not powered for primary outcome).We still observed significant differences between groups in secondary outcomes, including decisional conflict and knowledge.In addition, we did observe significant differences in RRSO uptake after 2 years between those who received the intervention compared to those who received usual care, which could have been due to wait times for consulations and surgery.In addition, the women included in this study self-selected to be involved in a research study and may not be representative of the population of women with a BRCA1 or BRCA2 mutation.Finally, the study population was primarily non-Hispanic White thereby limiting the generalizability of our findings.
Overall, this study provides evidence that a theorybased behavioral intervention delivered by genetic counselors to women with a BRCA PV who had not elected for the recommended RRSO was effective at reducing decisional conflict about having a RRSO, and increasing knowledge about the effectiveness of cancer risk reduction options in women with a BRCA1 or BRCA2 PV.There was no statistical difference in uptake of RRSO after 1 year between those who received the intervention and those who received usual care.Further research should be done to explore our exploratory finding that after 2 years significantly more women had a RRSO in the intervention group compared to those who received usual care.Our results suggest that women with a BRCA1 or BRCA2 PV who have not elected for a recommended RRSO may benefit from theory-based behavioral interventions to increase uptake of RRSO.

F I G U R E 1
CONSORT flow chart.T A B L E 1 Baseline demographic & clinical characteristics by treatment arm for intent-to-treat analysis.
Intervention effect on rRSO booking/uptake using logistic regression at 1 year follow-up based on different methods of dealing with missing data.
bNegative outcome imputation treated unknown PO as no PO.cAverage number of PO/bookings based on twenty imputation sets.

from year 1 to year 2 a Odds of RRSO in intervention v. control 95% CI p value Intervention Control Intervention Control No.
b Adjusted for having children.c Negative outcome imputation treated unknown PO as no PO.d Average number of PO/bookings based on 20 imputation sets.