Stepped approach to improving sexual function after gynaecological cancer: the SAFFRON feasibility RCT

Prevalence of sexual difficulties after cancer treatment

Many cancer patients are reported to have sexual difficulties. 1,2 Gynaecological oncology patients are particularly vulnerable to changes in sexual activity and lack of sexual desire,3–5 with sexual difficulty rates estimated between 40% and 100%. 6 Women undergo a range of treatments for ovarian, cervical, womb and vulval cancer, with different combinations of surgery, chemotherapy and radiation. Some of these treatments have a detrimental effect on women’s internal and external sex organs, surrounding tissues and nerves, and render some menopausal. Following such treatments women report a wide range of difficulties including loss of libido, dyspareunia, vaginal dryness and orgasmic difficulty. In addition, the symptom burden of gynaecological cancers is heavy, with many women reporting pain, fatigue, changes in bowel function, urinary symptoms including leakage, and depression and anxiety,7 which interact with menopausal and sexual difficulties. 8 Hazewinkel et al. 9 and Carpenter et al. 10 both note the major effect of physical well-being on sexual function, as well as the lack of relationship between extent of treatment and formal scores of sexual function.

Other factors contributing to sexual problems

It is unsurprising that women treated for a gynaecological cancer are at high risk of emotional distress. One prevalence study11 found that 23% satisfied criteria for major depressive disorder, and Parker et al. 12 found greater depressive symptoms in gynaelogical oncology patients than in those with breast, urology or gastrointestinal cancers.

Carpenter et al. 10 suggest that some of this greater distress is related to the very high levels of sexual difficulty experienced after treatment. They argue that sexual self-schema is an important moderator of response. Self-schema is the set of beliefs and ideas that people have about themselves based on their life experiences and, when this relates to the sexual domain, positive sexual self-schema (belief in and expectation of themselves to be pleasurably sexual) is associated with more frequent sexual activity, better sexual responsiveness and higher global sexual satisfaction across all disease sites,13,14 suggesting that it makes women more resilient to the adverse sexual impacts of gynaecological cancer. Despite their sexual difficulties, many gynaecological cancer survivors resume intercourse. 13,15 Andersen et al. 16 found that frequency of intercourse in their sample was comparable with available norms for similarly aged women, but these and other longitudinal data have shown sexual satisfaction17,18 and responsiveness16–20 to be significantly impaired following treatment.

Professionals’ response to sexual difficulties

Patients report that sexuality is rarely addressed by physicians. 3 Lindau et al. 18 found that conversation with a physician about the sexual effects of cancer was associated with significantly lower likelihood of complex sexual morbidity among very long-term survivors; however, 62% of 221 participants reported that their physician had never initiated a discussion about sexuality after cancer. In their study of sexuality in a palliative care setting, Vitrano et al. 21 found that patients considered it important to talk about sexuality and to face such an issue with an experienced professional, even though their life expectancy was short. Patients in their study had not had this opportunity. Moreover, some patients were still able to maintain a sufficient sexual activity, in terms of quality and quantity. Faithfull and White22 found that cancer nurses were more likely to focus on the technical aspects of sexual recovery post treatment, for example vaginal dilatation, and offered minimal advice or opportunities for disclosure of sexual dysfunctions, dissatisfaction with partner relationships or mood and other psychological difficulties. Clinical nurse specialists (CNSs) in gynaecological cancer acknowledge that they have an important role in this aspect of care but do not always feel confident or competent to assess or manage patients’ psychosexual needs, and appropriate referral is then problematic. 23,24 Recently, a national psychosexual group of expert nurse ‘champions’ has been formed, and its work to date includes a psychosexual assessment guideline document for nurses.

Diagnosing and treating sexual dysfunction in this setting

Sexual dysfunctions are recognised as difficulties affecting both sexual desire and sexual response25 and six apply to women, as defined by the Diagnostic and Statistical Manual of Mental Disorders:25–27

• sexual desire disorder or decreased libido – hypoactive sexual desire (low interest in sex)

• sexual aversion disorder – objections to having the genitals touched

• sexual arousal disorder

• orgasmic disorder – premature, delayed or absent orgasm following a normal sexual excitement phase

• sexual pain disorder of vaginismus – involuntary spasms of the muscles of the outer third of the vagina that interfere with intercourse

• sexual pain disorder of dyspareunia – pain during intercourse.

A recent review28 of specific complaints of all cancer patients referred to the sexual health programme of the Memorial Sloan Kettering Cancer Center found that the most common complaints for which patients sought help were painful intercourse (65%), vaginal dryness (63%), low sexual desire (46%) and orgasmic disorder (7%). The first two of these are partially managed through current best treatment, that is, topical oestrogen, vaginal dilators and lubricants. 2,29 Tsai et al. ,30 using the Female Sexual Function Index (FSFI), found that 66.7% of a Taiwanese sample of women with cervical cancer experienced sexual difficulties. Many other studies report high rates of sexual difficulties, from 83% for overall general sexual difficulty,31 to 66% significant and 46% moderate sexual difficulties. 32 Serati et al. ,33 in a study of women with early-stage cervix cancer treated with radical hysterectomy, reported that 65.8% of patients suffered sexual dysfunction.

Interventions for sexual dysfunction after treatment for cancer

In contrast to the majority of sexual therapy interventions in which anxiety reduction is often key, management of low sexual desire in the context of gynaecological cancer requires an intervention that additionally addresses the wider range of mediating factors, including loss, life-threat, trauma, change of body image, pre-existing psychological outlook, mood, depression and anxiety symptoms, as well as the relationship in which the woman finds herself. 34–37 Twenty-seven studies of sexual therapy interventions after any cancer are reported in a systematic review by Brotto et al. ,34 and Abbot-Anderson and Kwekkeboom35 found only three interventions specific to gynaecological cancer. These studies across a range of interventions only show small effect sizes, despite patient satisfaction with the interventions. Flynn et al. ,29 in their Cochrane Review of randomised control interventions for psychosexual dysfunction in women treated for gynaecological cancer, concluded that, ‘there is insufficient evidence to support or refute the use of any interventions for psychosexual dysfunction after gynaecological cancer’. Furthermore, they suggested that future investigations required multicentre randomised controlled trials (RCTs) with outcome measures validated in gynaecological cancer patients. They added specifically that ‘investigators should focus on interventions that can be delivered by existing members of the multidisciplinary team treating women with gynaecological cancers. It is more likely that such measures, if found effective, will be affordable and capable of being integrated into standard care’.

Conclusions and need for the research

The evidence presented above shows that it is often the case that women affected by gynaecological cancer are not aware of basic information about the sexual consequences of their gynaecological cancer and its treatment, and do not receive appropriate advice or help to recover sexual function and to adapt to their changed body and relationships. It is recognised by two Cochrane Reviews2,29 that new interventions are needed for sexual dysfunction in gynaecological cancer, and these need to be examined in multicentre RCTs with agreed outcome measures. There is a sizeable population with these problems to be addressed, for there are currently 2 million people in England living with and beyond cancer of all types, and 2.5 million across the UK as a whole. 38 This number is likely to grow by > 3% per year, reflecting the increasing incidence of cancer and better survival rates. By 2030, there are likely to be around 3 million cancer survivors in England,39 of whom a proportion will be survivors of gynaecological cancers. As we have the ability to develop suitable treatments, we have a duty to explore them. 39,40 The current acceptance of the worth of well-being and, conversely, the cost of depression, anxiety or unwillingness to engage with the health-care system – all potential long-term effects for the patient group concerned – are drivers of this research. 41 Better awareness of mental health issues and depression in general42 and in cancer patients,43 plus greater acceptance that these symptoms have causes that can be treated or addressed, is also relevant. In addition, our work was planned at a time when there was more awareness of sexual health, and evidence from cancer user groups,44 policy-makers45 and research46 of more openness to discuss these matters as a medical need. The potential of CNSs to deliver interventions to help with the consequences of cancer treatment has been recognised by the Department of Health and Social Care,47 yet little is known about CNSs’ training or supervisory needs to provide interventions for psychosexual dysfunction to work alongside psychologists. If care pathways exist for addressing sexual dysfunction in cancer, they are currently unique to individual units; providing the evidence for pathways that better meet the requirements of the population will facilitate clinical application of more appropriate and consistent practice. What is currently missing from the literature is a phased proposal that develops an intervention and tests this to facilitate best practice in the treatment of sexual dysfunction for all relevant women in gynaecological cancer centres. This study was planned to answer this question for the NHS by developing a stepped care intervention to be delivered within existing NHS gynaelogical oncology services. A gynaelogical oncology CNS or radiographer was to deliver interventions at step 1 and step 2 as the major treatment delivery, and only a small minority of more complex psychological issues were to be treated at step 3 by a level 4 practitioner (a clinical psychologist or a psychiatrist), to whom all patients should have access according to NHS guidance for psychological support in cancer. 47,48

Aims

• To develop a stepped care psychosexual intervention [a Stepped Approach Intervention to Improve Sexual Function after Gynaecological Cancer (SAFFRON)] on the Increasing Access to Psychological Therapies (IAPT) model together with a treatment algorithm for assigning women to levels of intervention.

• To establish whether or not women treated for gynaecological cancer with moderate to severe sexual dysfunction are willing to participate in a randomised trial and adhere to treatment.

• To indicate likely rates of recruitment to a future evaluation of the SAFFRON intervention.

• To pilot a stepped care psychosexual intervention (SAFFRON) on the IAPT model. 49

• To establish whether or not the SAFFRON intervention is acceptable to patients.

• To establish whether or not SAFFRON is deliverable by a gynaelogical oncology cancer centre multidisciplinary team.

• To indicate the most appropriate outcome measures for use in a larger trial.

• To inform estimates of the likely effect size, which will assist sample size calculations for a larger trial.

Research questions

• Are women treated for gynaecological cancer who develop moderate to severe sexual dysfunction willing to participate in a randomised trial of treatment and adhere to that treatment?

• Will women agree to be randomised to an intervention to treat sexual dysfunction?

• Are different tumour sites, treatments and cancer stages at approach associated with different rates of participation in the trial and uptake of the treatment?

• Is the stepped care system operable within the NHS system as it stands?

• What is the likely effect of the three levels of intervention on sexual function, mood and self-esteem as measured by standard measures?

• What is the rate of attrition from each treatment modality?

Purpose of research

• Is it possible to design and pilot a RCT that can potentially answer the question ‘Is SAFFRON a clinically and cost-effective treatment for sexual dysfunction after treatment for gynaecological cancer in the NHS?’?

• Can the SAFFRON intervention be evaluated in a feasibility randomised trial?

• Is a stepped approach acceptable and practical?

• Can it be done within NHS settings?

Embedded qualitative study

The embedded qualitative study was to explore the attitudes towards participating in the study of both staff and patient participants, and to examine potential barriers to and concerns about participation.

Interviews were planned with:

• trial participants in the intervention arm

• trial participants in the control arm

• participants who withdrew from the study

• staff who delivered the intervention

• staff in the clinics who were not involved in delivering the intervention.

Location and access

All intervention documents are archived at University College London (UCL). Level 1 contains elements derived from a document for which the American Cancer Society holds copyright, and these are unable to be reproduced in any form outside the study, including in this report and online.

Level 2 manual and worksheets can be obtained from either Dr Sue Gessler or Dr Lori Brotto on application (this permission level is set by Dr Lori Brotto as the copyright holder).

Level 3, following the opinions of the Trial Steering Committee (TSC) and the project team, is not available outside the prospect of an IPT training setting on ethical and intellectual property grounds and a prospective study for its use in the gynaelogical oncology setting by appropriately trained individuals. Any such approach should be made to Dr Sue Gessler or Professor Alessandra Lemma.

Inclusion criteria

• Women aged > 18 years (with partners at their choice) treated for any gynaecological malignancy with surgery and/or chemotherapy and/or radiation at UCLH Gynaecological Cancer Centre or University Hospitals Bristol Gynaecological Cancer Centre.

• ≥ 3 months post end of treatment.

• Any sexual orientation.

• With sexual function difficulties identified by initial screen (three clinical questions within clinical interview posed by doctor or nurse).

Exclusion criteria

• Poor English.

• Current drug or alcohol abuse.

• Current sexual therapy or psychotherapy.

Justification for not limiting the study to one gynaecological cancer site

The HTA programme call asked for all gynaecological cancer types.

A recent systematic review35 advocated comprehensive and systematic assessment of sexual concerns using reliable and valid measures in large representative samples that include all gynaecological cancer diagnoses, stages of illness and types of treatment. Rees56 suggests a broader definition of sexual satisfaction to be included in our level 2 and 3 interventions.

We used a person-centred, problem-based approach, as there is evidence that extent of treatment does not necessarily correlate with sexual dysfunction. 9,10

Participant recruitment

Recruitment rate

We aimed to recruit two or three participants per week for 9 months across both sites (Figure 1).

FIGURE 1.

Flow chart for intended study. ETAU, enhanced treatment as usual.

All women attending a gynaecological oncology clinic following primary treatment for cancer in UCLH or Bristol were potentially eligible.

Providing information on the trial was to take place during the clinical follow-up visit and to form an integral part of the consultation.

Trial interventions

Process through levels of intervention

Figure 2 shows the flow chart for the intended study.

• All women who consented to be part of the study were to be assessed on the entry criteria.

• Entry to study required a FSFI score of ≤ 26 and self-rated sexual difficulty.

• All eligible women received level 1 booklet from RA (ETAU: justified in Description and justification of the duration of treatment, subject participation and trial follow-up).

• All women were to be assessed at 4 weeks by RA following receipt of level 1 booklet.

• If the FSFI score was ≤ 26 after level 1, then intervention arm women progressed to level 2 (CNS/radiographer intervention).

• At completion of level 2 (three sessions attended), reassess. If the FSFI score was ≤ 26, offer level 3 intervention. If women failed to attend, RAs were to follow up if women had agreed during the consent process.

FIGURE 2.

Intervention flow chart for intended study.

Data

Description and justification of the duration of treatment, subject participation and trial follow-up

Outcome measure collection time points were set independent of the intervention delivery to ensure standardised collection of data and consistency between the intervention group and controls:

• T1 – baseline assessment and level 1 booklet given.

• T2 – 4 weeks (to allow use of the booklet and reassessment).

• T3 – 10 weeks from baseline. All women in intervention group should have completed level 2 by this time. Two to five fortnightly sessions. Length of treatment derived from existing manual and evidence.

• T4 – 25 weeks from baseline (primary end point).

• T5 – Follow-up data collected 8 months after baseline (follow-up).

The willingness of women to participate at each level was to be measured by counting the uptake of each intervention when offered and attendance at sessions offered. Their preference for any particular therapy was to be measured at baseline with the preference for treatment measure.

Trial follow-up was planned to be 8 months from baseline to evaluate whether or not any advantage of treatment endured over time, and to allow for, and to compare against, natural recovery rates in the ETAU group.

The follow-up time of 8 months, as opposed to 12 months as originally planned in the protocol submitted to NIHR, was proposed by the TSC when it first met the project team in August 2015. It was suggested that this would be a meaningful follow-up period while extending the time period available to recruit participants who could (theoretically) progress sequentially through all three interventions within the time period of the study. Unfortunately, this was not discussed with the funders at the time and was held to be a breach of contract as well as being considered an inappropriately short follow-up period.

Qualitative investigation

Semistructured interviews were to be conducted at 9 months with a purposive sample of 18 patients randomised to the stepped care arm of the trial to explore their experience of the process of recruitment to the trial and of receiving the interventions. Good and poor responders were to be chosen, as were those who did not take up the offer of interventions. Women were to be asked about their experience of the ‘stepped care’ model, and whether or not they liked the interventions that were offered to them. Women who dropped out or withdrew were to be asked for their reasons at point of leaving (if women had previously consented to do this at randomisation). We also intended to interview clinical staff to understand their experiences of recruiting participants to the trial, and senior clinicians and managers about the operation of the SAFFRON intervention and stepped care model. These data were to be analysed thematically using the framework approach71,72 with the aid of QSR NVivo qualitative data analysis software (QSR International, Warrington, UK) to provide important information on acceptability and potential for harm, as well as potential obstacles to and facilitating elements of the intervention.

At follow-up, the RA was to establish and record any additional treatments or information women had accessed, and record timing of drop-out from treatment or follow-up and reasons for drop-out, when ascertainable. They could telephone the women, e-mail them (with permission from women if given in the initial consent form) or visit them in their homes if they were too ill or fatigued to attend clinic.

Statistical plan

The aim of the study was to establish the feasibility of a larger trial of the proposed stepped care intervention.

The critical parameters that were to be used to quantitatively assess feasibility were the:

• consent rate

• proportion in the intervention group who moved up at least one step on the intervention

• proportion of all randomised subjects who had a useable (non-missing) score for total FSFI (the proposed primary outcome for any subsequent study) at 8 months.

To inform whether or not a main trial might be viable, the study planned to examine whether or not it is possible to achieve:

• a consent rate of ≥ 40%

• ≥ 70% of randomised subjects having a useable (non-missing) score for total FSFI (our primary outcome) at 8-month follow-up.

We also intended to use the feasibility data to provide an estimate of standard deviation of the FSFI score that would be required for the sample size calculation of the main trial.

A sample size of 100 patients, randomised equally to the two treatment groups, was chosen for this feasibility work. This sample size is shown below to be adequate to ensure sufficient precision to exclude the minimum acceptable values for the three critical parameters based on exact, one-sided 95% confidence intervals around assumed values for each parameter:

• Consent rate. We expected that 50% of women would consent to randomisation. With a sample size of 100 women, this consent rate could have been estimated with a lower 95% confidence interval of 44%.

• Proportion in the intervention group who move up at least one step on the intervention. We expected that 80% of the 50 women allocated to the intervention group would step up at least one level on the intervention during their treatment period. We planned to estimate our expected proportion of 80% with a lower 95% confidence bound of 68%.

• Proportion of all randomised subjects who had a useable score for total FSFI at 8 months. We estimated that 80% of the 100 women randomised in the trial would provide useable data to score the FSFI at 8 months. With 100 women we would have been able to estimate 80% with a lower 95% confidence bound of 72%.

Data from 100 patients would also be adequate to provide a precise estimate of the standard deviation of the primary outcome (FSFI) score to inform the sample size calculation for the main trial. 73

Economic evaluation plan

The aim of the economic evaluation was to optimise the methods of conducting an economic evaluation of the stepped care system compared with ETAU as part of a full trial. This would have included an estimate of the cost of stepped care and an evaluation of using a gynaelogical oncology-specific version of the CSRI to collect resource use data. A preliminary cost–utility analysis would have been conducted, reporting the incremental cost per QALY gained of the stepped care system compared with ETAU. The key aim of this would have been to evaluate the effect of using the EQ-5D-5L to calculate QALYs in this patient group versus a sexual function measure (the SQOL questionnaire) to calculate QALYs.

Recruitment rate required by the study

We aimed to recruit two or three women per week for 9 months between both sites.

Changes to methods before planned study commencement

1. The REC review changed the original study design. Concerns about women in the control arm being identified as having clinically significant sexual difficulties, and the awareness that there is no consistently available TAU for this, led the REC to ask for the level 1 booklet to be given to participants in both arms of the study. This point had been raised by patient advocates during the intervention development and was accepted by the research team.

2. When the TSC was convened in August 2015, its members were concerned that the full study would not be achievable with 1 full year of follow-up. The research team had hoped that there could be a no-cost extension to the study but this was not possible. To make the study more achievable, they proposed (1) reducing the follow-up period to 8 months and (2) introducing the SQOL as a within-treatment measure allowing the FSFI to be the outcome measure for sexual function. This required a substantial amendment and resubmission to the REC and engendered further delay. Convening the TSC earlier would have helped mitigate this delay.

Closure of study

The study was closed by the funder, NIHR, in November 2015 because of slow progression to recruitment as well as changes to the protocol made after its original filing with NIHR, which led to a substantial REC amendment. There are, therefore, no results of the stepped care intervention with patients and the questions posed in the objectives cannot be answered here.

Set-up and site opening delay

Administrative delay in the legal and financial set-up of the study, and hence funding of access to clinicians’ time to contribute to the intervention development, posed severe challenges to the timely progress of the study in the first instance. The set-up of the study was hindered by the chief investigator (CI) being a clinician without previous experience of running a trial. Administrative support from within the NHS was a secondment from an inexperienced, albeit enthusiastic, staff member who was unaware of mechanisms to raise the profile of the study locally within the local bureaucracy.

Working across a range of hospitals and a university proved difficult for set-up, and some hospitals asked to release a clinical staff member to write the intervention. Professor Alessandra Lemma had never undertaken such a financial or legal arrangement before. It seemed very clear that setting up a psychosocial study within a medical setting was very complex and unfamiliar to those who were highly competent at setting up a drug study.

The development of the interventions took substantially longer than the time estimated in the application. The lack of prepared interventions had been signalled as a risk by one reviewer in response to the original response to the commissioned call from the HTA programme.

The study was initially suspended by NIHR in October 2014 for financial reasons, with the assumption that it would be at least 1 year before it was restarted, and many involved assumed that it was permanently halted. Momentum was lost at this point, which was, in retrospect, key. Once funding restarted in February 2015, it was problematic to re-engage clinicians and the services, as they had made plans that excluded the study. Employment processes within the university for RAs needed to be started from scratch.

Once the interventions were ready and the study was able to progress, nursing teams felt unable to participate as trainees and deliverers of a new psychoeducational intervention, causing the need to seek other staff to deliver level 2. This was because of maternity leave at one site, a prospective hospital merger at the other and the need to have two CNSs at each site, one study trained and the other not. By contrast, clinical psychologists at both sites were fully trained in the level 3 intervention and were piloting it without difficulty. They were unconcerned about delivering the level 3 intervention, largely because it offered new resources in terms of skills, and the intervention fell within the professional demands of their role.

A substantial ethics amendment was required after the meeting of the TSC in August 2015, in which the committee recommended the addition of two further measures and, in addition, suggested that in order to keep the study within its time limits, the follow-up period should be shortened. All three changes formed part of the amendment.

There were differences between the two sites with respect to trust research and development approval. One was flexible and responsive. The other had still not approved research and development at the point of the study closure despite the application being over 2 months in the system.

Achievements of study and responses to obstacles

This section describes that which was completed and changes that occurred in the pre-recruitment phase to address obstacles that arose.

The development of the stepped intervention was completed. The study adapted to input from the REC that asked it to change the TAU arm to give participating women a better clinical experience.

Training of CNSs (as originally envisaged) at both sites was delayed by issues of:

• maternity leave of one study CNS, with locum cover inappropriate for the study in terms of both training and in length of contract

• lack of funding for backfill of CNS time while delivering level 2 intervention

• work demands at one site changing substantially after the project commenced because of a projected merger of hospitals within the time frame of the study.

Teams in both sites were keen that the study should proceed but neither team of CNSs was able, at the time the study went live, to offer CNSs for training and participation. The study therefore approached supernumerary staff at each site to train in, and deliver, the level 2 intervention. At UCLH this staff member was a former nurse consultant, currently running the UCLH Cancer Centre Macmillan Support and Information Service, and at University Hospitals Bristol the staff member was a consultant specialist radiographer who had academic time in her job plan that enabled her to participate.

Training

Initial training of level 2 therapists for Bristol and UCLH was written and delivered using a Microsoft PowerPoint^® (Microsoft Corporation, Redmond, WA, USA) summary of the background of the study [see section 4 PDF; URL: www.journalslibrary.nihr.ac.uk/programmes/hta/1111102/#/ (accessed December 2018)], and by working through each of the required sessions. Teaching involved some didactic elements and substantial role play and group discussions. Although highly experienced in general counselling techniques, feedback from both therapists revealed that they were not yet confident about carrying out specific psychological techniques, such as taking a behavioural history, or ‘gently challenging beliefs’ as required in one session. It appeared that each session required about 3 hours’ teaching time. At the point the study was halted, both had requested, in addition to supervision on pilot cases, top-up training. This was arranged but could not occur as the study was halted. This suggests that cross-disciplinary training with psychological interventions requires more time than was envisaged in this study.

Training of three clinical psychologists in IPT-GO was delivered by Professor Lemma in a full day. This used the national IPT training as its foundation, and additional material was included to show how IPT-GO differed from usual IPT for depression. All three psychologists felt confident to deliver IPT-GO and initial pilot cases completed before formal opening of recruitment. Professor Lemma judged them capable of carrying out IPT-GO within the confines of the study (i.e. with follow-up supervision on the first cases). The difference in confidence and capacity may derive from the fact that Professor Lemma is highly experienced in training for IPT. It may also be related to the fact that clinical psychologists are, by definition, used to delivering a range of different therapies. Although newly formulated, IPT-GO was closely related to their core skills and discipline, and they were able to develop quickly variations on their technique as required.

Causes of delay

Development of the intervention took substantially longer than initially envisaged. The research team delayed key parts of the study in the attempt to get the interventions right. In particular it was important to acknowledge the patient input, which changed the team’s initial assumptions, in particular interpretation of the commissioned call as attending to mood. Major input adapting mental health materials on mood in the stepped intervention was found pathologising and inappropriate by the patient advocates. The project team found this highly valuable, leading to a more robust and patient-appropriate set of interventions. The research team considered that it was better to delay while getting the interventions right, drawing on wide experience in psychological intervention research. However, these changes led to delays in advertising for RAs and in progress, which contributed to the initial suspension of the study.

The structure of the intervention itself changed during this time. The original design envisaged direct access via the treatment algorithm to levels 2 and 3, as successfully used in IAPT. The REC was concerned for the well-being of women in the control arm, in that they had been identified as being in need of, and wanting, help for sexual difficulties and would be offered nothing. They asked whether we could offer both arms the level 1 booklet, changing the TAU arm to an ETAU arm. This change in design had the advantage of dealing with any prospective contamination between arms by widespread distribution of the booklet within clinics. It also addressed a concern of the patient advocates, who had pointed out as the interventions were developed that much of the level 1 and level 2 materials were unknown to them, and would be a necessary basis for any more psychologically intense intervention. The learning from this element was that mapping directly from psycho-oncology to mainstream clinical psychology interventions is often perceived as unhelpful by both patients and RECs, and such differences should be considered at an early stage.

Administrative delay in the legal and financial set-up of the study, and hence funding of access to clinicians’ time to contribute to the intervention development, posed severe challenges to the timely progress of the study. This appeared to be rooted in institutional barriers to setting up a psychosocial study in cancer services within the NHS; for example, some involved institutions had never participated in this way before and did not understand the processes involved or required. Formal structures between the NHS division and UCL did not work as well as would be expected. Many processes had to be repeated across organisations. The release of clinical time by a full-time clinician to develop the level 3 intervention (AL) required advance agreement, which led to further delay.

Major issues arose as a result of the intention to conduct the trial within the existing clinical teams, including using clinical staff to deliver the interventions. This intention was in line with the recommendation of the 2009 Cochrane review29 aimed at speeding subsequent integration into clinical practice. However, implementation of the study was therefore subject to real-life fluctuations in staff and in the clinical demands on teams the trial necessarily involved.

The study had proposed one CNS to be trained and the other to offer TAU at each site. At one site, the maternity leave of the CNS intended to be trained for the study, together with delays and concerns about locum cover, left the site without a CNS to be trained, and the lead CNS was no longer able to support the study. At the second site, despite initial enthusiasm, at the point of training, the CNS team was unable to offer a staff member, owing to concerns about existing clinical load and a prospective enlargement of their catchment area with a consonant rise in referrals to the centre. Hence searches were made for alternative staff with links to the clinical teams but who were able, owing to the structure of their jobs, to take on an extra task both of training and of clinical input. Although ultimately successful, this also led to delay.

The nursing teams at both sites stated that they were under clinical pressure, and the inability of the study to offer direct backfill for their time was a further obstacle. The original proposal to the HTA programme for this study proposed a trial CNS for 18 months with clerical support to address issues of clinical demand on existing teams, and to examine the effect of the intervention as delivered by CNSs. The team was advised that under the terms of a HTA award, this could not be funded as it would appear as a NHS excess treatment cost. Although the CI secured limited charitable funding to support the CNS teams, terms of the charity prevented direct payment for staff time, and hence this did not fully allay the concerns of the nursing teams. Other psychosocial studies funded differently have tended to establish the intervention with research staff initially, and then proceed with the roll-out into clinical practice. With hindsight, this might have been a preferable approach to the development and testing of the SAFFRON interventions.

Attempting to launch in both sites simultaneously added delay. A pilot phase at UCLH before moving to Bristol could have given opportunities for ironing out logistical issues and service delivery problems.