Activities that have been carried out as part of the evaluation of the Renewed intervention are reported in more detail in the publications. The initial plan was for a feasibility trial followed by a main trial but we anticipated that there might be no changes from the initial feasibility phase which meant that the feasibility study became in effect an internal pilot for the main trial, and the approach was documented in advance in the ISRCTN registration and agreed with PGfAR.
Objectives
This section describes the evaluation of the Renewed intervention during a 30 month RCT. Aims, methods, results, and implications are described for each piece of research as follows:
- RCT to assess clinical and cost-effectiveness.
- Process evaluation exploring how patients and healthcare professionals (HCPs) experienced and implemented the intervention in practice:
- Patient qualitative process study: perceptions of using the intervention for patients.
- Patient quantitative process study: engagement and usage of the Renewed intervention by patients.
- HCPs mixed methods process study: exploration of HCPs’ experiences of and adherence to delivering the intervention within the supported arm of the trial.
The section finishes with a conclusions section which draws the findings together.
Randomised controlled trial to assess clinical and cost-effectiveness
Aims
To assess the effectiveness and cost-effectiveness of a digital intervention to support cancer survivors (‘Renewed’).
Methods
Overview
We evaluated an internet-based intervention (Renewed) to help support healthy behaviour changes and improve quality of life, psychological well-being and symptom management for breast, colorectal and prostate cancer survivors. Participants were randomly allocated to one of three arms: generic advice, access to Renewed or access to Renewed with brief support. All groups were assessed at follow up with patient reported outcome measures.
Inclusion criteria: we chose three contrasting common survivor groups to deal with the likely varying issues in needs and preferences across the gender and age spectrums: breast cancer survivors (younger and older women); prostate cancer survivors (predominantly older men); and colorectal cancer (a range of age and gender). Patients were: identifiable from GP case records; had completed primary treatment (e.g. chemotherapy/radiotherapy treatment) within the last 10 years; not receiving palliative care; had internet access; had poorer quality of life (scoring < 85 on the EORTCQLQ). Patients could be recruited opportunistically but mainly by written invitation.
Exclusion criteria: advanced cancer receiving palliative care; currently active cancer unless on active surveillance/watchful waiting for prostate cancer; another type of cancer (i.e. other than breast, prostate or colorectal) in the last 5 years; currently receiving any cancer treatment other than hormone therapy or finished treatment during the last month; expecting to start any cancer treatment during the study period; severe mental health problems and/or major uncontrolled depression, schizophrenia or dementia; sarcoma or lymphoma of the breast; living in the same household as another participant of the study.
Randomisation
Baseline measures were completed online immediately following eligibility screening and online consent. Automated randomisation (using a 1 : 1 allocation ratio) was then implemented using LifeGuide software (www.lifeguideonline.org; accessed 6 March 2025). Participant randomisation was stratified by:
- cancer type: breast/prostate/colorectal and
- EORTCQLQ-C30 score (64 or less/65 or more30).
Randomisation and follow-up were automated and conducted by software (i.e. keeping the research team blinded). Blinding of patients to the intervention was not possible. Participants were informed online as to which group they had been allocated to immediately via e-mail.
Trial arms:
- Control: Generic advice and follow-up. All patients in the control and intervention groups had access to usual care (GPs and PNs for routine appointments), to manage any new symptoms, make further referrals and so on as appropriate. In addition, participants in this group received a link to the NHS LiveWell website to access evidence-based advice about living a healthier lifestyle, including detailed advice to support patients in their mental health, weight control, exercise, sleep, eating well, alcohol and smoking. All participants had the same structured questionnaire follow-up as in other trial groups, with telephone follow-up should questionnaires not be returned.
- Renewed: web only intervention. This group had access to usual care as in (1) but were also given the Renewed intervention.
The initial core content of Renewed explained how key behaviour changes (physical activity, healthy eating, engagement with mood management strategies and weight loss support) would help reduce symptoms experienced after treatment for cancer. The answers previously supplied for the screening process generated personalised suggestions for modules to access; for example, if a participant indicated that they had been feeling tense or worried, a suggestion to look at the mood management section (Healthy Paths) would appear. In the first session of Renewed, there was also advice and information about active surveillance/watchful waiting for prostate cancer which was only available to participants who indicated that they were in that group. The core content also provided links to additional information and support available for topics not covered by Renewed (such as finances).
After participants completed the core content, they could access four additional optional sections of Renewed: Getting Active for increasing physical activity, Healthy Paths for stress reduction, Eat for Health for diet improvement and POWeR+ for weight loss. Once a participant had accessed each module, they were sent brief automated e-mails as reminders and for motivation with tips, interesting facts and brief information and advice. Participants were able to cancel e-mails at any time and there was a link in each e-mail with that provision.
- Renewed plus support: web-based intervention with additional guidance and support. This group had access to usual care and the Renewed intervention as in (2) but were also given access to brief human support. This support was provided by a nurse or HCA at their surgery where possible, or else a research nurse employed by the NIHR CRN. The ‘Supporters’ received online training (approximately 15 minutes) to learn how to implement the previously validated CARE model28 in order to provide support to their patients in using Renewed.
One to 2 weeks after initially logging on to Renewed, participants could have a face-to-face or phone appointment with their supporter. Where central CRN nurses provided support, sessions were held by phone. Follow-up support by phone or e-mail was then offered at 4 and 8 weeks, so three support sessions could be offered in total. Support sessions were brief (a maximum of 10 minutes) and comprised encouragement to use the website and patient-led discussion around the patient’s goals, how patients had got on with the changes that they tried and how they overcame any barriers. Should further support be needed, participants could e-mail their supporter using the contact form on the Renewed website. All contact and support session details were logged on the ‘Supporter’s Log’ kept in the surgery site-file.
Measures and outcomes
At baseline gender, age, marital status, years of education, ethnicity, height and weight were recorded.
Outcomes were patient self-reported and measured online using the LifeGuide system at baseline, 6 and 12 months (unless indicated), with two e-mail reminders, followed by two postal administrations for non-respondents (enclosing a £10 voucher at 6 months to reward participation and encourage completion), and a final telephone follow-up for those not responding to mail. Information from medical records was obtained by a blinded researcher.
Primary outcome
Health-related quality of life using the EORTC (European Organization for Research and Treatment of Cancer) Quality of Life Questionnaire C30 instrument2 (version 3) summary score (https://qol.eortc.org/app/uploads/sites/2/2018/02/scoring_of_the_qlq-c30_summary_score.pdf; accessed 6 March 2025) measured at 6 and 12 months.
Secondary outcomes
- EORTCQLQ-C30 subscales (baseline, 6 and 12 months): Global self-rated Health, Symptom subscales, Functional subscales (e.g. Physical functioning, Social functioning, Emotional well-being).
- Quality of life [EuroQol-5 Dimensions, five-level version (EQ-5D-5L)]31 for health economic analysis (baseline, 6 and 12 months).
- Brief scales for Depression [Hospital Anxiety and Depression Scale (HADS)]32 Fear of Relapse33 (modified to three items) and The Measure Yourself Concerns and Wellbeing questionnaire for Quality of Life34 (baseline, 12 months).
- Modified enablement scale35,36 (12 months).
- A website satisfaction measure modified for this study (12 months).
- Resource use data were collected through reviewing case notes and extracted from the electronic record at the end of the trial. Details of the information were extracted as follows:
- Primary care, specifically consultations by type (face to face or telephone) and with who (GP, PN, HCA). Up to 12 visits were recorded.
- Medications including both baseline and new medications with up to 15 medications, along with dose and if changed.
- Other community-based services, including home visits, NHS calls, drop in centres, occupational therapy and others.
- Inpatients, including for each admission, reason, disease and dates of admission and discharge.
- Outpatients as for inpatients.
- A&E attendance.
- Problematic Experiences of Therapy Scale with single item measures of self-reported adherence to website recommendations for physical activity, mental well-being, diet (12 months).
- Website usage and entries. With informed consent, this was unobtrusively automatically collected by the LifeGuide system, including time spent on each page and entries (e.g. goal setting and goal-related progress in physical activity, healthy eating, weight, fatigue and mood management).
- Adverse outcomes – serious adverse events (SAEs) were reported by both participants and practice staff, though it was anticipated that almost all of the SAEs would be unrelated to the research.
Sample size
Our aim was to encourage behaviour change, enable better management of symptoms, improve HRQoL, and improve self-rated health in each of the cancer types. We anticipated a long-term impact, but the sample size calculation, and therefore the timing of the primary analysis specified in the protocol was at 6 months (since 6 months was the main timescale of outcomes available in systematic reviews37). We anticipated that the web intervention, being very much briefer than the interventions previously described would be less effective than the more intensive face to face interventions used previously (we anticipated a standardised effect size of 0.15 to 0.30). We also wanted to demonstrate the effectiveness in each of the cancer clinical groups where possible. The Bonferroni correction has been criticised as being unduly conservative,37 so in line with the rationale of Cook and Farewell,38 we used α = 0.05 since within each cancer subgroup there were independent pair-wise comparisons for the two intervention groups versus controls. To detect a difference of 0.3 standardised mean difference (SMD) in any pairwise comparison between intervention subgroup and control for 80% power and α = 0.05, we required 176 intervention participants in each intervention subgroup and 176 controls, 1584 in total for the 3 clinical subgroups (breast, colorectal and prostate), or 1980 allowing for 20% loss to follow-up. We estimated that the total sample would allow us to detect overall differences between intervention groups of 0.15 SMDs. Despite the individually randomised design, cluster effects are possible: we assumed 8 patients per intervention group per practice, and an ICC of 0.03 with the inflation factor of 1.21 [1 + (8–1 × 0.03)], required 2396 participants. Allowing for some leeway in our assumptions, we aimed to recruit 2500 individuals.
Statistical analysis
All participant data were analysed (unless the participant/s specifically requested that their data be removed from the data set) on an intention to treat (ITT) basis, that is as randomised using Stata version 17 (StataCorp LP, College Station, TX, USA; www.stata.com/company). The primary analysis was with imputed data, using a chained equation multiple imputation model for missing data, including all outcomes as well as all variables included in the analysis model. A complete-case analysis was a sensitivity analysis. Generalised linear mixed regression models were used for the analysis of continuous variables, controlling for baseline values, stratification variables, and also controlling for potential confounding variables as appropriate, both for the overall trial sample and for each clinical subgroup (breast, prostate, colorectal). All models included a random effect for practice to allow for clustering of participants within practices, as per the sample size calculation. Pre-planned subgroup analyses were set out in the statistical analysis plan for age, gender and comorbidities. We also performed post hoc within-group analyses documenting the changes from baseline.
Workstream 4 health economic analyses
Detailed health economic analyses including design, data collection and methods were provided in Appendix 2.
In summary, we had hoped to perform a societal analysis but the data to do this was very incomplete, with low response rates. The study took an NHS and Personal Social Service (PSS) perspective, including all incurred costs in the cost-effectiveness analyses. However, as the intervention focused on self-management with in person or online support, the most likely to impact would be expected to be on primary care consultations and medication use (hence using primary care costs as our base case).
Itemised resource usage data were weighted by their corresponding unit costs hence accumulated costs were calculated for each patient. EQ-5D-5L scores were translated into utility scores based on the UK tariff, and quality-adjusted life-years (QALYs) were calculated using an area under the curve approach. Bootstrap methods based on 1000 replicates with replacement were employed to estimate means costs and QALYs, as well as the differences between groups with confidence intervals (CIs) associated with each estimate. Cost-effectiveness acceptability curves (CEACs) were produced based on the 1000 replicates.
The base-case analysis used NHS costs incurred in primary care. Sensitivity analyses included secondary service use [inpatient, outpatient and accident and emergency (A&E) attendances]. The primary care data are the most reliable, since it excludes those services least likely to have been influenced by the intervention, including inpatient services which were (except for emergencies) likely to have been planned before the intervention. Similarly, any emergency admissions were also unlikely to have been influenced by the intervention. The same arguments apply to outpatient attendances.
Role of funder: the funder, NIHR, had no role in data collection, analysis, interpretation, writing of the manuscript nor the decision to submit.
Results
Recruitment: 58,295 ‘cold calling’ invitations were sent by mail from 494 GP practices. Most people who did not wish to take part did not reply, but 10,840 people returned a reply slip to the study team. Of the 10,840, 143 did not give a reason, and the 10,697 gave one or more reasons, the most common being not wanting any help (5584), followed by not having or not wanting to use internet access (2649) (see Table 1 for a full list of reasons). A total of 7883 individuals expressed interest in participating and were assessed for eligibility (Figure 2); 2732 patients were recruited from 494 practices, between 12 October 2017 and 2 April 2020). The main reason eligible participants were not recruited was that they already had good quality of life (see CONsolidated Standards Of Reporting Trials diagram; Figure 1). By 6 months 83% (2260/2712) and by 12 months 83% (2247/2712) of participants were followed up with complete primary outcome data. The Programme Steering Committee advised stopping recruitment once the overall target was achieved and sufficient sample in prostate and breast subgroups.
| I don’t have any problems that I want help with | 5584 |
|---|---|
| I don’t feel well enough to take part at the moment | 355 |
| I don’t have an e-mail address | 2130 |
| I’m not interested in taking part | 875 |
| I don’t have time | 323 |
| I don’t have internet access | 2649 |
| I don’t want to use a website | 1821 |
| Other | 2873 |
| Total | 16,610 |
FIGURE 2 CONsolidated Standards Of Reporting Trials diagram. This figure has been reproduced from Little et al.1 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) licence, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original text. (Figure not reproduced here; see the original article.)
Baseline characteristics. These were well balanced between groups (Table 2) and between cancers (Table 3) with the obvious exception of gender differences between prostate and breast. Only 430/2712 (16%) of participants were lost to follow-up at 6 months and 454/2712 (17%) at 12 months.
| Generic advice (n = 906) | Renewed plus support (n = 903) | Renewed (n = 903) | Total (n = 2712) | |
|---|---|---|---|---|
| Age (years) | ||||
| Mean (SD) | 64.5 (10.9) | 64.5 (11.2) | 64.5 (10.7) | 64.5 (10.9) |
| Baselinea EORTCQLQ-C30 score | ||||
| Mean (SD) | 72.1 (12.2) | 72.5 (11.8) | 72.7 (11.7) | 72.4 (11.9) |
| BMI | ||||
| Mean (SD) | 28.0 (5.5) | 28. 2 (5.5) | 28.0 (5.4) | 28.1 (5.5) |
| Ethnicity (non-White) | ||||
| 2/906 (2.4) | 19/902 (2.1) | 20/902 (2.2) | 61/2710 (2.3) | |
| Cancer group | ||||
| Bowel/colorectal | 143/906 (15.8%) | 143/903 (15.8%) | 146/903 (16.2%) | 432/2712 (15.9%) |
| Breast | 474/906 (52.3%) | 471/903 (52.3%) | 471/903 (52.2%) | 1416/2712 (52.2%) |
| Prostate | 289/906 (31.9%) | 289/903 (32.0%) | 286/903 (31.7%) | 864/2712 (31.9%) |
| Comorbidities | ||||
| Cardio | 291/779 (37.4%) | 325/787 (41.3%) | 329/789 (41.7%) | 945/2355 (40.1%) |
| Lung | 130/779 (16.7%) | 126/787 (16.0%) | 149/789 (18.8%) | 405/2355 (17.2%) |
| Other | 548/779 (70.4%) | 545/787 (69.3%) | 563/789 (71.4%) | 1656/2355 (70.3%) |
| Gender | ||||
| Male | 379/906 (41.8%) | 380/903 (42.1%) | 368/903 (40.8%) | 1127/2712 (41.6%) |
| Female | 527/906 (58.2%) | 523/903 (57.9%) | 535/903 (59.3%) | 1585/2712 (58.4%) |
| Time since last cancer treatment (years) | ||||
| Mean (SD) | 4.0 (3.0) | 4.1 (3.2) | 3.9 (3.3) | 4.0 (3.1) |
| Bowel/colorectal (n = 432) | Breast (n = 1416) | Prostate (n = 864) | Total (n = 2712) | |
|---|---|---|---|---|
| Age (years) | ||||
| Mean (SD) | 67.0 (10.6) | 59.4 (10.1) | 71.6 (7.4) | 64.5 (10.9) |
| Baselinea EORTCQLQ-C30 score | ||||
| Mean (SD) | 72.1 (12.1) | 71.6 (12.3) | 73.8 (10.9) | 72.4 (11.9) |
| BMI | ||||
| Mean (SD) | 28.4 (5.7) | 28.0 (6.0) | 28.0 (4.3) | 28.1 (5.5) |
| Comorbidities | ||||
| Cardio | 181/382 (47.4%) | 349/1248 (28.0%) | 415/725 (57.2%) | 945/2355 (40.1%) |
| Lung | 80/382 (20.9%) | 206/1248 (16.5%) | 119/725 (16.4%) | 405/2355 (17.2%) |
| Other | 280/382 (73.3%) | 858/1248 (68.8%) | 518/725 (71.5%) | 1656/2355 (70.3%) |
| Gender | ||||
| Male | 246/432 (56.9%) | 17/1416 (1.2%) | 864/864 (100.0%) | 1127/2712 (41.6%) |
| Female | 186/432 (43.1%) | 1399/1416 (98.8%) | 0 (0.0%) | 1585/2712 (58.4%) |
| Time since last cancer treatment (years) | ||||
| Mean (SD) | 4.1 (3.2) | 4.0 (3.1) | 3.8 (3.1) | 4.0 (3.1) |
EORTCQLQ-30
There was improvement in both the generic advice groups and both the intervention groups (Figure 3 and Table 4), with no statistically significant differences between groups overall at the primary time point of 6 months for the total score. However, the Renewed plus support group did experience higher quality of life at 6 months than the usual-care group in the prostate cancer subgroup by 2.03 points (95% CI 0.25 to 3.80). At 12 months, there was a statistically significant improvement in the Renewed plus support arm with QoL higher in this group than usual care by 1.42 points (95% CI 0.33 to 2.51). All cancer groups showed improved QoL over usual care but the results were only significant for prostate cancer (Table 5).
| All participants | |||
|---|---|---|---|
| Generic advice | Renewed plus support | Renewed | |
| 6 months | |||
| Mean (SD) | 76.0 (14.31) | 76.7 (14.41) | 76.1 (13.99) |
| Complete cases | 806 | 705 | 749 |
| Complete cases | REF | 0.50 (−0.67, 1.66) | −0.42 (−1.57, 0.72) |
| Imputed (100 imputations) | REF | 0.52 (−0.53, 1.57) | −0.20 (−1.23, 0.84) |
| 12 months | |||
| Mean (SD) | 75.7 (15.13) | 77.2 (14.07) | 77.0 (14.42) |
| Complete cases | 803 | 702 | 742 |
| Complete cases | REF | 1.11 (−0.10, 2.31) | 0.72 (−0.46, 1.91) |
| Imputed (100 imputations) | REF | 1.42 (0.33, 2.51)a | 0.94 (−0.13, 2.01) |
| Colorectal/bowel cancer | Breast cancer | Prostate cancer | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Generic advice | Renewed + support | Renewed | Generic advice | Renewed + support | Renewed | Generic advice | Renewed + support | Renewed | |
| 6 months | |||||||||
| Mean (SD)a | 75.4 (14.35) | 75.1 (14.36) | 75.2 (13.31) | 75.8 (14.70) | 76.3 (14.97) | 75.9 (14.18) | 76.4 (13.60) | 78.3 (13.31) | 77.0 (13.97) |
| Compete cases | REF | −2.23 (−4.97, 0.50) | −2.29 (−4.88, 0.31) | REF | −0.48 (−2.37, 1.41) | −1.01 (−2.85, 0.82) | REF | 1.99 (−0.01, 3.99) | −0.06 (−2.09, 1.97) |
| Imputed (100 imputations) | REF | −1.56 (−1.21, 1.10) | −1.29 (−3.81, 1.22) | REF | 0.27 (−1.21, 1.74) | −0.13 (−1.61, 1.36) | REF | 2.03 (0.25, 3.80) | 0.42 (−1.32, 2.16) |
| 12 months | |||||||||
| Mean (SD) | 74.1 (14.67) | 76.4 (13.61) | 77.0 (14.64) | 76.0 (15.60) | 76.9 (14.54) | 77.2 (14.26) | 75.8 (14.49) | 78.3 (13.41) | 76.7 (14.57) |
| Complete cases | REF | 0.76 (−2.12, 3.64) | 1.53 (−1.25, 4.33) | REF | −0.02 (−1.91, 1.88) | 0.22 (−1.59, 2.04) | REF | 2.00 (−0.13, 4.14) | −0.24 (−2.41, 1.94) |
| Imputed (100 imputations) | REF | 1.04 (−1.68, 3.76) | 1.74 (−1.03, 4.51) | REF | 0.75 (−0.76, 2.27) | 0.99 (−0.50, 2.47) | REF | 2.76 (0.87, 4.65) | 0.58 (−1.29, 2.45) |
FIGURE 3 Graphical representation of improvement in EORTC QLQ-C30 over time. This figure has been reproduced from Little et al.1 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) licence, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure above includes minor additions and formatting changes to the original text. (Figure not reproduced here; see the original article.)
Results for subscales at 6 months
There was a significant difference in self-rated global health in both Renewed groups (Table 6). The Renewed plus support group also showed improvement in the physical function and cognitive function subscales.
| Generic advice | Renewed support | Renewed | |||
|---|---|---|---|---|---|
| Mean (SD) | Mean (SD) | Imputed (100 imputations) | Mean (SD) | Imputed (100 imputations) | |
| Global healtha | 64.4 (19.89) | 66.3 (18.54) | 1.82 (0.14, 3.52) | 65.9 (19.11) | 1.88 (0.18, 3.58) |
| Functional subscales a | |||||
| Physical function | 77.7 (20.68) | 78.7 (19.88) | 2.00 (0.64, 3.36) | 78.2 (19.88) | 0.50 (−0.82, 1.82) |
| Role function | 71.9 (28.36) | 71.3 (28.14) | −1.01 (−3.45, 1.42) | 71.0 (27.49) | −1.06 (−3.44, 1.31) |
| Emotional function | 69.8 (22.72) | 70.5 (22.15) | 0.79 (−1.01, 2.60) | 70.0 (21.93) | −0.05 (−1.87, 1.76) |
| Cognitive function | 73.5 (22.10) | 76.5 (20.86) | 2.32 (0.52, 4.12) | 75.5 (21.33) | 0.69 (−1.10, 2.48) |
| Social function | 71.8 (27.55) | 73.6 (27.46) | 1.46 (−0.83, 3.75) | 72.7 (26.68) | 0.70 (−1.58, 2.97) |
| Symptom subscales b | |||||
| Fatigue | 38.6 (22.60) | 37.2 (22.75) | −1.18 (−3.08, 0.73) | 38.7 (22.51) | 0.20 (−1.72, 2.12) |
| Nausea and vomiting | 5.5 (11.62) | 6.0 (12.67) | 0.55 (−0.59, 1.70) | 5.8 (11.78) | 0.26 (−0.85, 1.36) |
| Pain | 31.5 (28.07) | 31.8 (28.19) | 0.43 (−1.89, 2.75) | 32.2 (28.33) | 0.81 (−1.49, 3.12) |
| Dyspnoea | 20.8 (25.81) | 19.0 (24.65) | −1.47 (−3.64, 0.70) | 20.11 (26.41) | −0.92 (−3.00, 1.17) |
| Insomnia | 41.5 (31.22) | 41.5 (32.43) | 0.001 (−2.81, 2.81) | 42.0 (31.15) | 0.39 (−2.38, 3.16) |
| Appetite loss | 11.4 (21.37) | 10.8 (20.83) | −0.36 (−2.28, 1.56) | 11.1 (20.65) | 0.07 (−1.81, 1.95) |
| Constipation | 17.0 (25.79) | 15.6 (24.86) | −1.51 (−3.75, 0.73) | 15.8 (25.50) | −0.68 (−2.88, 1.52) |
| Diarrhoea | 11.0 (20.62) | 11.2 (22.66) | 0.55 (−1.49, 2.58) | 12.2 (23.06) | 0.89 (−1.15, 2.92) |
| Financial difficulties | 10.4 (23.42) | 11.8 (23.82) | 1.03 (−0.83, 2.89) | 11.44 (23.02) | 1.02 (−0.85, 2.89) |
Results for subscales at 12 months
The generic advice group remained stable by 12 months, but there were further improvements in both Renewed groups, with significant differences compared with generic advice in patients’ perception of global heath, dyspnoea and constipation (Table 7). The Renewed plus support group also showed improvement in the physical, emotional, cognitive fatigue and dyspnoea subscales. All the other scales in the support group apart from financial difficulties improved compared to generic advice and this was significant for seven scales. A similar pattern was seen for the Renewed group without support: with significant changes in global health and improvement in dyspnoea and constipation. All but two of the sub-scales improved compared to generic advice.
| Generic advice | Renewed support | Renewed group | |||
|---|---|---|---|---|---|
| Mean (SD) | Mean (SD) | Imputed (100 imputations) | Mean (SD) | Imputed (100 imputations) | |
| Global health | 63.8 (20.46) | 66.6 (19.01) | 2.78 (1.08, 4.48) | 66.6 (18.68) | 3.06 (1.39, 4.74) |
| Functional subscales 2 | |||||
| Physical function | 78.3 (21.12) | 79.6 (19.57) | 2.25 (0.88, 3.62) | 78.5 (20.54) | 0.14 (−1.22, 1.50) |
| Role function | 71.8 (28.63) | 72.5 (27.57) | 0.02 (−2.36,2.41) | 73.2 (27.63) | 0.89 (−1.50, 3.27) |
| Emotional function | 68.9 (23.13) | 71.5 (22.35) | 2.72 (0.84, 4.61) | 70.2 (22.82) | 0.99 (−0.87, 2.84) |
| Cognitive function | 73.6 (22.43) | 76.3 (20.59) | 1.92 (0.13, 3.71) | 76.0 (21.33) | 1.12 (−0.65, 2.89) |
| Social function | 72.9 (28.45) | 75.6 (27.62) | 2.22 (−0.20, 4.64) | 74.7 (27.83) | 1.31 (−1.10, 3.72) |
| Symptom subscales 2 | |||||
| Fatigue | 38.4 (22.91) | 35.6 (22.50) | −2.67 (−4.58, −0.75) | 37.0 (22.78) | −1.25 (−3.15, 0.66) |
| Nausea and vomiting | 6.1 (12.58) | 5.9 (12.58) | −0.24 (−1.41, 0.93) | 5.2 (12.02) | −0.96 (−2.11, 0.20) |
| Pain | 31.2 (28.32) | 30.9 (27.08) | −0.22 (−2.53, 2.10) | 31.8 (28.79) | 0.86 (−1.48, 3.20) |
| Dyspnoea | 22.2 (19.26) | 19.3 (25.45) | −2.73 (−4.92, −0.55) | 19.7 (25.31) | −2.78 (−4.91, −0.64) |
| Insomnia | 42.5 (32.56) | 41.2 (32.55) | −1.30 (−4.14, 1.55) | 40.1 (30.84) | −2.27 (−5.08, 0.54) |
| Appetite loss | 11.0 (20.46) | 10.3 (20.15) | −0.68 (−2.53, 1.18) | 10.5 (20.88) | −0.13 (−1.96, 1.69) |
| Constipation | 18.9 (26.29) | 16.6 (25.22) | −2.36 (−4.62, −0.11) | 15.5 (24.95) | −2.77 (−4.99, −0.55) |
| Diarrhoea | 11.8 (21.31) | 11.2 (21.60) | −0.57 (−2.51, 1.37) | 11.9 (21.40) | −0.42 (−2.38, 1.53) |
| Financial difficulties | 9.2 (21.60) | 9.8 (21.49) | 0.28 (−1.49, 2.04) | 9.5 (21.61) | 0.36 (−1.40, 2.13) |
Within-group analysis: improvement from baseline
The minimal clinically important difference for the EORTC questionnaire has not been robustly determined in our target population, but assuming a 6-point difference could be important,39,40 more than 40% in each group achieved a 6-point improvement at 6 months with no difference between the groups. By 12 months there was continued improvement in both intervention groups (Table 8), but flattening off in the generic advice group.
| 6 months | 12 months | |||||
|---|---|---|---|---|---|---|
| EORTCQLQ | Generic advice | Renewed plus support | Renewed | Generic advice | Renewed plus support | Renewed |
| Mean at baseline (SD)a | 72.0 (12.19) | 72.0 (12.19) | 72.4 (11.80) | 72.6 (11.71) | 72.4 (11.80) | 72.6 (11.71) |
| Mean at follow-up (SD)a | 76.0 (14.31) | 76.0 (14.31) | 76.7 (14.41) | 75.7 (15.13) | 77.2 (14.07) | 77.0 (14.42) |
| Difference based on paired t-test (95% CI) | 3.91 (3.16 to 4.66) | 4.32 (3.52 to 5.11) | 3.53 (2.76 to 4.29) | 3.61 (2.81 to 4.40) | 4.85 (4.04 to 5.67) | 4.40 (3.60 to 5.19) |
| % achieving MCID improvement of 6 points | 41.95% | 44.68% | 42.67% | 41.97% | 48.15% | 45.46% |
Other secondary outcomes
This analysis relies heavily on imputation since we received only half of the HADS data and other 12-month secondary outcome questionnaire data. Therefore, we have also presented the crude means from the complete-case data. There was an improvement in both anxiety and depression in the Renewed plus support group at 12 months (as well as in the renewed group but this was not statistically significant) (Table 9). However, the improvement was modest at < 0.5 point. The bothersomeness on the MYCAW1 was lower in the support group, and the other MYCAW responses were in the same direction but not significant. The fear of recurrence scale showed no significant difference between the groups. The patient enablement instrument was significantly lower in both groups, which suggests higher enablement (these were coded 1 = strongly agree to 7 = strongly disagree) suggesting that on average one in three people said they agreed they felt enabled to manage their condition compared to those having generic advice (where on average people only slightly agreed). There were not enough events to look at survivorship but deaths across the three arms are summarised below.
| Generic advice | Renewed plus support | Renewed | |||
|---|---|---|---|---|---|
| Mean (SD) | Mean (SD) | Imputed (100 imputations) | Mean (SD) | Imputed (100 imputations) | |
| HADS – Anxiety | 7.54 (4.56) | 7.14 (4.54) | −0.43 (−0.77, −0.08) | 6.95 (4.27) | −0.35 (−0.71, 0.01) |
| HADS – Depression | 5.76 (3.95) | 5.28 (3.74) | −0.40 (−0.70, −0.10) | 5.37 (3.81) | −0.26 (−0.59, 0.08) |
| PEI | 3.22 (1.27) | 2.91 (1.21) | −0.30 (−0.44, −0.16) | 2.86 (1.23) | −0.35 (−0.50, −0.21) |
| MYCAW1 | −0.73 (1.71) | −0.96 (1.73) | −0.24 (−0.43, −0.05) | −0.89 (1.71) | −0.15 (−0.37, 0.07) |
| MYCAW2 | −0.68 (1.81) | −0.84 (1.75) | −0.15 (−0.38, 0.06) | −0.80 (1.78) | −0.11 (−0.35, 0.14) |
| FRRS1 | 3.37 (1.16) | 3.43 (1.18) | 0.05 (−0.07, 0.18) | 3.38 (1.16) | 0.01 (−0.13, 0.15) |
| FRRS2 | 3.42 (1.21) | 3.49 (1.19) | 0.05 (−0.07, 0.18) | 3.43 (1.19) | 0.002 (−0.13, 0.14) |
| FRRS3 | 2.52 (1.30) | 2.58 (1.31) | 0.06 (−0.07, 0.19) | 2.53 (1.30) | 0.02 (−0.12, 0.16) |
Deaths: there were six deaths: one in the generic advice group, three in Renewed plus support and two in Renewed.
Health economics
Mean NHS primary care costs were lower in Renewed and Renewed plus support groups (–£141, –£77, respectively). There were very small differences in estimated quality of life based on the EQ-5D-5L. In addition, a substantial proportion of data in EQ-5D-5L were missing, leading to significant uncertainty in the estimates. Both interventions were likely to be cost-effective (incremental costs per QALY £11,281, –£7542, respectively) (Table 10); and the Renewed intervention is probably more cost-effective than Renewed plus support (see Appendix 2 for the full Health Economic Analysis). Both intervention groups were dominant in term of the effectiveness measures – demonstrating improved outcomes with lower costs.
| Groups | Mean NHS primary care costs | Difference (compared with Generic advice) | |
|---|---|---|---|
| Generic advice | 265 (254, 275) | ||
| Renewed plus support | 187 (180, 195) | −77 (−90, −65) | |
| Renewed | 124 (118, 130) | −141 (−153, −128) | |
| Renewed vs. Renewed plus support | −63 (−73, −54) | ||
| QALYs (EQ-5D-5L) | Difference (compared with Generic advice) | Incremental cost per QALY gained (ICER) | |
| Generic advice | 0.7 (0.692, 0.709) | ||
| Renewed plus support | 0.699 (0.691, 0.707) | −0.001 (−0.013, 0.011) | 11,281 (−76,949, 97,385) |
| Renewed | 0.705 (0.696, 0.714) | 0.005 (−0.007, 0.018) | −7542 (−116,829, 115,651) |
| Renewed vs. Renewed plus support | 0.006 (−0.006, 0.018) | −6007 (−50,473, 46,389) |
Results for cancer subgroups: There were no significant interaction terms and the subgroup results were generally in line with the main trial results (Table 11).1 After controlling for cancer type, males did significantly better than females on the primary outcome at 6 months. There was also a suggestion that there could be more benefit among longer-term survivors.
| Generic advice | Renewed support | Renewed | ||||
|---|---|---|---|---|---|---|
| N (%) | Mean (SD) | Mean (SD) | Difference (95% CI) | Mean (SD) | Difference (95% CI) | |
| Primary outcome – total EORTCQLQ C30 score at 6 months | ||||||
| Time since end of treatment | ||||||
| Up to 1 year | 427 (16%) | 78.4 (12.87) | 76.6 (14.81) | −0.84 (−3.60 to 1.93) | 76.7 (13.66) | −2.14 (−4.80 to 0.52) |
| More than 1 year | 2231 (84%) | 75.5 (14.51) | 76.8 (14.33) | 0.78 (−0.96 to 1.37) | 76.0 (14.06) | 0.20 (−0.96 to 1.37) |
| Time since end of treatment (post hoc) | ||||||
| Up to 3.5 years (median) | 1311 (50.1%) | 76.8 (13.6) | 75.9 (15.45) | −0.67 (−2.23 to 0.90) | 76.7 (13.90) | −1.43 (−2.95 to 0.09) |
| 3.5 years | 1307 (49.9%) | 75.1 (14.91) | 77.6 (13.26) | 1.75 (0.30 to 3.20) | 75.5 (14.06) | 1.11 (−0.43 to 2.65) |
| Sex | ||||||
| Male | 1126 (42%) | 76.1 (13.60) | 78.3 (13.12) | 1.95 (0.39 to 3.51) | 76.9 (13.86) | 0.48 (−1.03 to 1.99) |
| Female | 1578 (58%) | 75.8 (14.79) | 75.6 (15.18) | −0.49 (−1.89 to 0.91) | 75.6 (14.05) | −0.68 (−2.07 to 0.71) |
| Lung comorbidity | ||||||
| No | 2221 (83%) | 76.9 (13.53) | 77.9 (13.56) | 0.99 (−0.16 to 2.14) | 77.1 (13.65) | −0.13 (−1.26 to1.00) |
| Yes | 452 (17%) | 71.0 (16.90) | 70.5 (16.86) | −1.81 (−4.75 to 1.14) | 72.1 (14.72) | −0.32 (−3.14 to 2.50) |
| Other comorbidity | ||||||
| No | 786 (30%) | 78.9 (12.50) | 79.6 (13.75) | 0.71 (−1.19, 2.61) | 78.3 (13.33) | −0.47 (−2.34, 1.41) |
| Yes | 1875 (70%) | 74.7 (14.83) | 75.5 (14.51) | 0.46 (−0.83, 1.75) | 75.2 (14.16) | −0.06 (−1.33, 1.21) |
Implications
Cancer survivors with lower quality of life given detailed generic online support improve significantly, including those where the primary course of treatment was several years previously – where improvement in quality of life would not be expected. Providing a robustly developed website to address peoples’ support needs, with minimal additional resource implications, is likely to provide modest but significant further longer-term improvement in enablement, symptom management and most importantly self-rated global health, and with lower NHS costs in primary care.
Process evaluations exploring how patients and healthcare professionals experienced and implemented the intervention in practice
Patient qualitative process study: experiences of using a supported digital intervention for cancer survivors in primary care
Please see Appendix 4 for a full summary of this study.
Aims
To explore how cancer survivors used the Renewed intervention and to understand the potential barriers and facilitators to using Renewed and performing the recommended behaviours.
Methods
Forty-two patients took part in the interviews. Data were analysed using inductive thematic analysis.
Results
Three patients did not recall using Renewed and/or being in the study so their data were excluded from the analysis. Thematic analysis identified four themes: (1) The importance of motivation for using Renewed and performing behaviour changes, (2) Minimal usage of Renewed plus supported behaviour changes, (3) Supporter needs to add value and (4) Convenience of Renewed Online.
Practical implications
The results showed that even limited usage of Renewed Online may provide enough information to motivate behaviour change in those with less need for more tailored support. Novel information may need to be presented earlier in the intervention to motivate further engagement with Renewed in those who need more detailed and tailored information to make behaviour changes. This has implications for the implementation of Renewed and similar interventions into clinical practice. Renewed may be implemented as stepped care approach for cancer survivors, where cost-effective support may be offered to many patients with less need for resource intensive and tailored support, and patients only ‘step up’ to more specialist support for more complex needs. In addition, adding support alongside digital interventions may motivate engagement, particularly for people who lack support elsewhere.
Patient quantitative process study: engagement and usage of the Renewed intervention by patients
Please see Appendix 4 for a full summary of this study.
Aims
To evaluate how cancer survivors used Renewed. The specific research questions were: (1) How much did patients use Renewed? Which components of Renewed were most used? (2) Is there a difference in Renewed usage (i.e. frequency) based on patient characteristics? (3) Do patients who accessed human support use Renewed more compared to those who did not access human support or were without access to human support? (4) Do patients who accessed Renewed’s Optional content have greater improvements in quality of life at 12 months?
Methods
This was a quantitative process study nested within the RCT of the Renewed intervention. Patients were sampled from all three arms of the Renewed trial: (1) Renewed Online, (2) Renewed Online with brief human support and (3) usual care. Data from patients from the Renewed Online and Renewed Online plus brief human support arms were analysed. This study was carried out 12 months after patients joined the Renewed study.
Patients completed questionnaires at baseline and 12 months post randomisation. The intervention software, LifeGuide, recorded their usage of Renewed. Patients were included in the analysis if they had completed the 12-month follow-up measures. Demographic and clinical characteristics were collected at baseline, and psychological characteristics were collected at 12 months.
Access to human support was characterised in three ways: (1) patients who accessed human support (‘accessed support’), (2) patients who had the option to access human support but did not (‘did not access support’) and (3) patients who were in the Renewed Online only group who did not have access to human support (‘without access to support’).
Results
The sample consisted of the 1760 cancer survivors in the Renewed Online and Renewed Online with brief human support trial arms who completed follow-up measures at 12 months.
How much did participants use Renewed? Which components of Renewed were most used?
Patients accessed Renewed a median of 2 times, ranging from 0 to 268. The majority of patients (97%) accessed the Core content of Renewed, with 84% completing the Core content and reaching the Homepage where they could access the other components of Renewed. Slightly less than half (45%) of patients continued to use Renewed past the Homepage to access the Optional content of Renewed (Getting Active, Eat for Health, Healthy Paths, POWeR). After the Core content, Getting Active was the most used component of Renewed (30%), and least used was Healthy Paths (10%).
Is there a difference in Renewed usage (i.e. frequency) based on participant characteristics?
There were no significant associations between the overall number of times patients accessed Renewed and their characteristics (age, gender, cancer type, age when completed education and years since finishing treatment).
Do those who accessed support use Renewed more compared to those who did not access support or were without access to support?
At 12 months, the results showed a moderate significant difference in usage of Renewed based on access to support (p < 0.001). Those who accessed support used Renewed more than those who did not access support or were without access to support (median number of times Renewed was accessed: 4, 1, 1, respectively). The approximate effect was greatest when comparing those who accessed support to those who chose not to access support (d = 0.37).
Do patients who accessed Renewed’s Optional content have greater improvements in quality of life at 12 months?
At 12 months, those who used Optional content had higher EORTCQLQC30 scores compared to those who only accessed the Core content. The mean score was 2.15 (95% CI, 3.28 to 1.01; p < 0.001) points lower among those who only accessed the Core content compared to those who accessed the Optional content after adjusting for covariates (p ≤ 0.001). There was a small standardised effect (d = 0.2).
Practical implications
These results suggest that motivating patients to engage more with Renewed Online is likely to lead to better outcomes. Renewed was largely used as intended, with usage of Optional content associated with greater improvements in QoL at 12 months. Patients who accessed human support engaged with Renewed more than patients who did not have human support or were without human support, suggesting that the brief support provided by HCPs encouraged engagement. However, as support was optional and had to be requested by patients, the patients who accessed human support may have already been more confident and motivated to engage with Renewed.
Healthcare professionals’ qualitative process study: exploration of healthcare professionals’ experiences of using the intervention
Aims
To explore Supporters’ experiences of providing support to cancer survivors using the Renewed intervention to understand potential barriers and facilitators to the implementation of Renewed.
Methods
Twenty-eight semistructured telephone interviews were conducted with Supporters (practice-based nurses, practice-based HCAs and clinical research nurses). A semistructured interview schedule was developed, which explored: (1) Supporters’ experiences of providing support, (2) perceptions of Supporter training, (3) experiences during support appointments, (4) perceptions of the CARE approach and (5) Supporters’ perceptions of the Renewed programme. Interviews were audio-recorded and transcribed verbatim. Data were analysed using inductive thematic analysis.
Results
The final sample included 28 HCPs consisting of 16 PNs, 6 clinical research nurses, and 6 practice-based HCAs. They provided support for patients at a total of 45 GP practices. Twenty-seven HCPs were female; one was male.
Four themes were developed which reflected factors that Supporters identified as enabling or hindering them to support patients using Renewed. The themes were: confidence to enact the Supporter role, practicalities of delivering support alongside a digital intervention, managing a patient-led approach, and Renewed Online as an acceptable tool with some improvements. The results of this study have been published.41
Practical implications
Our results suggest that HCPs based in primary care found this type of digital intervention with a Support role acceptable and are amendable to contributing to the delivery of support to cancer survivors.
Healthcare professionals are central to providing support for cancer survivors, yet support is limited because of practitioners’ perceived lack of time and expertise in how to help cancer survivors with symptoms after treatment, which are key barriers to providing this support. The results of this study show that these barriers could be overcome with a digital intervention such as Renewed, which provides the majority of support so HCPs only need to provide brief support to maximise engagement, with no need to develop cancer specific expertise or behaviour skills.
Key factors that may support the successful implementation of Renewed in practice include the increasing acceptability of phone support in primary care (particularly in the current climate of the coronavirus disease discovered in 2019 pandemic), and the utility and acceptability of the CARE approach among most HCPs. Ease of the Supporters’ online training and Renewed’s perceived similarity to tools used in current practice also support successful implementation in practice.
Equality, diversity and inclusion
The population were reasonably well balanced for gender, with a good range of educational status, but only 2–3% of the study population compared with 6% from the 2021 census were from ethnic minorities, and which reflects the difficult of recruiting in ethnic minority areas.