A randomised control trial assessing the impact of an investment-based intervention on weight-loss, glycated haemoglobin and psychological outcomes 1 year post bariatric surgery

Patient and Public Involvement

The idea for this study originated from a series of qualitative interviews with patients about their experiences of weight loss through BS and medication^{13,16,17,20,21}. From these interviews it became apparent that investment in a weight management treatment approach could result in greater success of that approach. Discussions were then had with patients from two weight management clinics (University College London; St Richards Hospital, Kent) and with patients from the National Support Group, WLSinfo. These discussions addressed the content of the investment intervention to explore what aspects of BS involve investment and are important to patients and resulted in the focus on financial, social, personal and physical domains. The choice of outcome measures was informed through previous qualitative interviews with patients who had had BS which highlighted a key role for changes in eating behaviour, quality of life and perceptions of control^16,17. The choice of an RCT as the design of the study was independent of PPI and based upon the need to evaluate the effectiveness of investment as an intervention. The findings of the study have been disseminated through the support of WLSinfo (now Obesity UK) and will be further disseminated through patient networks on social media.

Participants

University College Hospital (UCH) in London, UK, offers an NHS standardised bariatric service for patients living with obesity with a BMI over 40 (or 35 with serious comorbidities). All potential patients are assessed by the multidisciplinary bariatric team (physician, anaesthetist, dietician, psychologist and surgeon). If they are approved for surgery, they then attend the hospital’s bariatric surgery pre-operative assessment clinic where they receive final routine tests and information prior to their surgery. At this clinic, consecutive adult patients were approached by the research fellow (AH), and recruited those who consented. Recruitment took place over a 15-month period from November 2013 to January 2015. Inclusion criteria were that the patient had consented to the trial, was aged 18 or over, was attending the bariatric clinic at UCH, had been accepted for surgery and had funding from their CCG in place. Those who were deemed not to be able to read English effectively following initial discussions with the research fellow were excluded as this would pose a difficulty in implementing the intervention.

Sample size calculation

A feasibility study of the impact of an investment-based intervention²⁴ showed a significant effect on weight loss by three months (mean difference = 7 kg; n= 98; d = 0.3). Assuming an attrition rate of 40% by 12 months we aimed to invite 200 participants (100 in each condition) to ensure final data from at least 60 in each group. This would provide power to detect a small / medium difference in weight (controlling for baseline weight) by 12 months follow up.

Design

The study used an open-randomised parallel group control trial design with patients allocated to receive either usual care or the low cost and low effort investment-based intervention pre and post bariatric surgery. Weight loss and glycated haemoglobin (HbA1c) were assessed in the clinic, with the psychological outcomes assessed at baseline in the clinic and by post, at 1-year follow-up. This study received favourable ethical opinions from the Bloomsbury Research Ethics Committee, National Research Ethics Service (REC 13/LO/1056; IRAS ID 126012; Date 15/07/2013) and the University of Surrey Ethics Committee (EC/2014/21/FAHS FAST TRACK; Date 12/02/2014). The structure and reporting of this trial was guided by the CONSORT statement for clinical trials. Trial registration was ClinicalTrials.gov identifier NCT02045628; December 2, 2013.

Randomisation

Patients were randomised using third-party-blinded randomization provided by the clinical trial unit at the University of Surrey. Information sheets were sent out in advance of the patients’ preoperative appointment, which they attended two weeks prior to their operation to have routine tests. At this appointment, patients saw the research fellow who explained the trial, obtained informed consent and randomised them to one of the two arms of the trial using the predetermined randomisation. Whilst randomisation was blinded, neither the research fellow nor the patient were blinded to the arms of the trial. Baseline measures were completed at this time in the clinic.

Consenting procedures

Patients who have been approved for surgery by the MDT and have funding in place for their surgery are sent a letter giving them a pre-operative assessment date to attend the clinic about 2 weeks prior to their surgery. Participant information sheets about the trial and consent forms were sent to patients along with this letter. At this clinic they were approached by the research fellow (AH), given another copy of the information sheet and a consent form. They were also advised that they could withdraw from the study at any time before the study end date by contacting the research team and that their data would be removed from the study.

Intervention

Patients allocated to the usual care group (control) received preoperative tests and a dietician consultation. In this they were given a standard diet sheet informing them about their diet preoperatively and the postoperative stages of food progression from only consuming liquids to soft food then back to all foods. Patients returned for surgery approximately 2 weeks later, and after a median post-surgical stay of two nights, they were discharged home. They then returned to the clinic at 6 weeks, 3, 6 and 12 months to see the dietician and/or specialist nurse.

Patients allocated to the investment intervention group received usual care as described above plus carefully framed questions designed to raise the salience of the investment they had made in their procedure at 2 weeks preoperatively, then 3 and 6 months follow-up after surgery. The intervention was based on a pilot study²⁴ and involved participants rating questions relating to the investment they had made in having bariatric surgery to raise the salience of their investment. To this end, the questions and responses were framed in such a way as to emphasise investment in the process of bariatric surgery to optimise the outcomes of the procedure. In particular, the questions encouraged patients to consider how their surgery has impacted upon them in terms of financial, social, personal and physical costs and focuses on factors such as pain, disruption to their family, social and work lives and financial burden.

Primary outcome measures

Patients’ weight was obtained in the clinic 2 weeks preoperatively and postoperatively at 3, 6 and 12 months. The primary outcome was weight loss at 12 months measured in terms of kg, BMI and HbA1c. Participant demographics (age, sex, ethnic group, education and living status) were also assessed at baseline.

Secondary psychological outcome measures

Eating behaviour, health status and control beliefs were assessed using a questionnaire completed in clinic at baseline and by post at 1 year follow up. Eating behaviour was assessed in terms of hunger, intake of snacks, drinks and cooking methods using existing measures that have been used extensively in previous research^16,28–30. Health status was assessed in terms of the General Health Questionnaire (GHQ 16; 31) and self-esteem (e.g. ‘That on the whole you are satisfied with yourself’)^16,30,32 and control beliefs to assess imposed control (e.g. ‘That your stomach is controlling what you eat’), personal control (e.g. ‘That you are controlling what you eat yourself’) and binge eating (e.g. ‘A desire to overeat’) were measured using a measure specifically developed to assessing changes post bariatric surgery^16,30. All unintended harms were recorded. There were no unintended harms reported by the participants in either group.

Statistical analysis

Data was analysed using SPSS version 22. Initially, differences between groups for baseline demographics were assessed using χ2 and t tests. The impact of the intervention on primary outcomes was then analysed using both intention to treat analysis and explanatory analysis. Next, the impact of the intervention on secondary psychological outcomes was assessed in terms of eating behaviour, health status and control. To control for baseline, Analysis of Covariance (ANCOVA) was used.

Flow of participants, follow-up and sample characteristics

At baseline, 225 patients were invited to take part in the study, 13 refused and 212 consented (see Figure 1 for CONSORT flow chart). Of these, 10 subsequently withdrew their participation in the study, 7 were unavailable for follow up and 54 had their operation cancelled, so were excluded from the analysis. Therefore 141 patients were randomised to either the control group (n=68) or intervention group (n=73). Follow up weights at 12 months were obtained from n=119 (usual care n=55; intervention n=64) and HbA1c results from n=80 (usual care n=37; intervention n=40). 24 weights were unavailable as they did not return for their 1-year appointment at the clinic and a further 38 HbA1c results were unavailable due to no test results being recorded at 1 year.

Figure 1. CONSORT flow diagram with flow of participants.

Participants’ baseline demographics (n = 141) for all patients and by trial arm are shown in Table 1. Mean weight at baseline was 130.12 kg (range 88.2-224.8), mean BMI was 47.28 (SD= 8.73) and the majority were white, female, cohabiting, educated to either secondary school or the level of a professional certificate with a mean age of 42 years. 68 had a gastric bypass (49%) and 73 (51%) had a sleeve gastrectomy operation. Patient demographics and baseline measures are shown in Table 1. The results showed no differences between the two arms in terms of demographics.

Primary outcomes

Follow-up weights at 12 months were obtained from n = 119 (usual care n = 55; intervention n = 64) and HbA1c results were obtained for n= 80 (usual care n = 37; intervention n = 43). 24 weights and 63 HbA1c results were unavailable as patients did not return for their 1-year appointment at the clinic.

For the intention-to-treat analysis, final weights and HbA1c were imputed from either their data at baseline, at 3 months follow-up or 6 months follow-up.

Impact of the intervention

Primary outcomes: The primary outcomes were assessed with regard to the impact of the intervention. This was explored using analysis of covariance (ANCOVA) through both exploratory and intention to treat analysis. Outcomes at 12 months follow up are shown in Table 2.

The explanatory analysis indicated no significant difference between groups for weight and BMI change over the 12 months after their bariatric operation. However, there was a significant difference in HbA1c with significantly greater decrease in the control group when compared with the intervention condition. Further analysis was completed with regard to intention to treat for the primary outcomes at 12 months follow up (see Table 3). A similar pattern of results was found with no significant difference in weight and BMI but a greater decrease in HbA1c for the control group compared to those who received the intervention, although this failed to reach significance.

Secondary outcomes: As response rates were low for secondary outcomes (n=73) only explanatory analysis was carried out. Secondary outcomes for eating behaviour, health status and control beliefs are shown in Table 4 and Table 5. The results showed no differences between groups at 12 months follow up for any measures of eating behaviour or health status. Further no differences were found for personal control or binge eating. However, those in the intervention group showed significantly lower beliefs about imposed control than those in the control group.