Clinical outcomes and adverse events of bariatric surgery in adults with severe obesity in Scotland: the SCOTS observational cohort study

Article history

The research reported in this issue of the journal was funded by the HTA programme as project number 10/42/02. The contractual start date was in January 2012. The draft report began editorial review in March 2022 and was accepted for publication in August 2022. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HTA editors and publisher have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this report.

Copyright statement

Copyright © 2024 Mackenzie et al. This work was produced by Mackenzie et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This is an Open Access publication distributed under the terms of the Creative Commons Attribution CC BY 4.0 licence, which permits unrestricted use, distribution, reproduction and adaptation in any medium and for any purpose provided that it is properly attributed. See: https://creativecommons.org/licenses/by/4.0/. For attribution the title, original author(s), the publication source – NIHR Journals Library, and the DOI of the publication must be cited.

2024 Mackenzie et al.

Obesity

Body mass index (BMI) is an indicator of body fat calculated by dividing a person’s weight in kilograms by their height in metres squared. Obesity is defined as BMI ≥ 30 kg/m² and severe obesity as BMI ≥ 40 kg/m². 1

Obesity and, in particular, severe obesity are associated with a variety of negative health outcomes, including increased risk of most major chronic diseases and premature death. 2 Obesity-related comorbidities are defined as conditions either directly caused by obesity or known to have their presence or severity affected by obesity. Consequently, it is anticipated that these comorbid conditions will improve or enter remission in the presence of effective and sustained weight loss. 3 A non-exhaustive list of known obesity-related comorbidities includes:

• premature mortality

• cardiovascular conditions

  • hypertension

  • atherosclerosis

  • myocardial infarction (MI)

  • stroke

  • congestive heart failure

  • cardiac arrhythmias

• metabolic conditions

  • type 2 diabetes mellitus (T2DM)

  • prediabetes

  • dyslipidaemia

  • non-alcoholic fatty liver disease

• pulmonary conditions

  • obstructive sleep apnoea

  • asthma

• musculoskeletal conditions

  • degenerative arthritis

  • immobility

  • pain

• reproductive conditions

  • polycystic ovary syndrome

  • infertility

  • sexual dysfunction

• genito urinary conditions

  • impaired renal function

  • kidney stones (nephrolithiasis)

  • stress urinary incontinence

• central nervous system conditions

  • impaired cognition

  • headache

  • idiopathic intracranial hypertension (pseudotumour cerebri)

• psychosocial conditions

  • impaired health-related quality of life (HRQoL)

  • depression

  • anxiety

  • other psychopathy

• cancers.

People with severe obesity are at increased risk of several of these comorbidities, including T2DM, cardiovascular disease (CVD) and depression. This can lead to the development of multiple morbidities, often at a young age, which, in turn, causes reduced HRQoL, increased healthcare costs and heightened risk of mortality. 4

In recent years, severe obesity has emerged as a major public health concern, with rates increasing rapidly in a number of countries across the world, including the United States of America (USA) where the prevalence of BMI > 40 kg/m² among adults rose by 96% between 2000 and 2018, and around 9.2% of the adult population is now considered to have severe obesity. 5 Similarly, levels of severe obesity have risen in the United Kingdom (UK), with 3.3% of all adults in England6 and 3% of all adults in Scotland now estimated to have a BMI ≥ 40 kg/m². 7

Obesity treatment

As the prevalence of severe obesity rises, effective treatment is a priority. Treatments for severe obesity may be surgical or non-surgical. Bariatric surgery is the collective term for a number of surgical interventions with the primary purpose of achieving large-scale weight loss. Non-surgical treatment usually involves a multicomponent approach comprising behavioural therapy, dietary change and increased physical activity, and can also involve pharmacotherapy. 8,9

Clinical effectiveness and cost-effectiveness of bariatric surgery

A 2014 Cochrane systematic review of 22 randomised controlled trials (RCTs) for bariatric surgery found it to be more clinically and economically effective for the treatment of severe obesity than non-surgical measures after 2 years. 8 This finding was supported by data from trials with longer term follow-up, and cohort and modelling studies. 15,16 Indeed, in their 2019 systematic review and meta-analysis of 33 datasets reporting long-term (≥10 years) outcomes following bariatric surgery, O’Brien and colleagues reported that surgical procedures resulted in a weight loss effect three to four times greater than that of non-surgical therapy. 15 Similarly, Borisenko et al. found bariatric surgery to be cost-effective at 10 years post surgery16 and results from the Swedish Obese Subjects (SOS) study demonstrate cost-effectiveness of surgical procedures over 15 years, particularly in those with diabetes. 17

UK bariatric surgery guidelines

UK NICE guidelines63 currently indicate that bariatric surgery is a treatment option for those with a BMI ≥ 40 kg/m², or between 35 kg/m² and 40 kg/m² in the presence of other significant diseases (T2DM or hypertension, for example), which could be improved if they lost weight. People with a BMI of 30–34.9 kg/m² with onset of T2DM within 10 years and people of Asian ethnicity with onset of T2DM at a lower BMI are also considered for assessment, as are any adults with a BMI > 50 kg/m². In all cases, non-surgical weight management must have been attempted but not resulted in clinically beneficial weight loss before bariatric surgery is indicated. 53,63

In Scotland, Scottish Intercollegiate Guidelines Network guidance64 is followed. These guidelines are broadly similar to those of NICE, stipulating that bariatric surgery should be considered for those with a BMI ≥35 kg/m² and one or more severe comorbidities which are expected to improve significantly with weight reduction, such as mobility problems, arthritis and T2DM. As per NICE guidelines, there is a requirement for evidence of completion of a multicomponent, structured weight-management programme that has not resulted in significant and sustained improvement in comorbidities. 64

Bariatric surgery rates

Despite evidence that bariatric surgery is a more clinically and economically effective intervention for the treatment of obesity than non-surgical options, for the UK NHS, like many other health systems, the volume of bariatric surgery procedures commissioned is very low. In fact, despite obesity prevalence being among the highest in the European Union, the UK performs only nine bariatric surgery procedures per 100,000 people,4,65 while Sweden, a country with a similar health service but lower obesity prevalence, performs 70–80 procedures per 100,000 people. 66 In North America, the rate of surgery is around 40–50 per 100,000 people, with the majority of these operations performed in the USA. 65 Moreover, despite escalating levels of severe obesity in the UK, numbers of NHS bariatric procedures are falling, with a reduction of 31% between 2011/2012 and 2014/2015. 4,6 Rates of surgery have also been observed to vary between countries in the UK; no NHS bariatric surgery is performed in Northern Ireland, while few NHS weight-loss operations take place in Wales and Scotland as compared to England. 4

It has been suggested that one reason for low bariatric surgery rates in the UK is that rather than general practitioners (GPs) referring directly to surgical services, patients must follow a complex pre-surgical tiered pathway and barriers are often encountered (see Table 1). 4 Other barriers include the perception among patients and clinicians that bariatric surgery is high risk, and the fact that commissioners restrict funding for bariatric operations, despite evidence of cost-effectiveness. 4 The latter may be due to the initial high cost of bariatric surgery, with savings recouped in subsequent years.

The low prioritisation of bariatric surgery within the UK and the strict criteria for access to surgery, including complex pre-surgical pathways and pre-surgical weight-loss requirements,67 results in low numbers of individuals with severe obesity actually receiving surgery. Those receiving surgery generally do so after many years of alternative conservative interventions, at a point when their mean BMI is extremely high, at around 45 kg/m², and they are at a median age of 47 years. 68 To date, it is unclear how this delay in treatment impacts on health, physical functioning and HRQoL.

UK bariatric surgery care pathways

Pre- and postop care is a major component of the total cost of bariatric surgery. 67 However, anecdotal evidence suggests that bariatric surgery care pathways vary considerably, including clinical psychology provision. International bariatric guidance, while based on best practice, does not specify the optimal model of care69,70 and there is little evidence of whether intensive pre- and postop care improves outcomes and is cost-effective compared to less intensive care. Further investigation is therefore required as to the extent the intensity of pre- and postop bariatric surgical care is a factor affecting patient outcomes after surgery.

Surgical Obesity Treatment Study (SCOTS)

Patient involvement

Patients identified via bariatric surgery peer support groups in Scotland were involved in the design and conduct of this research. During the protocol development stage, patients provided input with regard to data collection and defining research questions. Two focus groups were held with patients to discuss methods of recruitment; they contributed to recruitment procedures and development of materials. Patients were included in a promotional recruitment video, with consent. A patient member was included on the independent study steering committee and patients were invited to a meeting to discuss plans for dissemination of study results.

Study design

SCOTS was a national, prospective, observational cohort study of adults aged over 16 years eligible for primary bariatric surgery in Scotland. Participants were recruited from 3 December 2013 to 28 February 2017. A mean of 3 years postop follow-up continued until October 2020. A detailed protocol for the SCOTS study was published71 but was amended in 2016 with follow-up reduced from 10 to 3 years.

Study setting

The study was conducted in 10 NHS-funded and 4 private hospitals in Scotland. Centres were eligible to participate if they performed bariatric surgery. Bariatric procedures included gastric banding, gastric bypass and SG. All patients’ healthcare interactions in Scotland are recorded by use of a single patient identification number. Information technology systems are common across all 14 health board areas, and a single government-funded department (Information Services Division) collates information for research purposes. This has allowed these systems to be utilised for post-surgical patient follow-up in SCOTS.

Study participants

Adult patients (aged 16 years and over) scheduled to undergo a primary bariatric procedure at any hospital in Scotland were eligible for invitation to the study.

Participant screening

Eligibility checks were undertaken by clinical bariatric surgery teams or research nurses while patients attended preop bariatric assessment clinics. Patients were approached at least 4 weeks prior to surgery. Patients were consented in clinic or referred to the SCOTS research team, who provided further study information and then obtained fully informed consent. Patient information sheets were provided at least 24 hours before patients consented to the study. In the case of those patients who were to be recruited by the research team, a patient information sheet was provided when permission was sought to hand over contact details to the research team. An independent contact was provided on the patient information leaflet so that patients could discuss participation in research studies with someone independent of the study team, should they wish to do so.

Patients consented for clinical data linkage (part one), postal, electronic and/or telephone follow-up (part two) and whether they were interested in future research. Those requiring a translator were consented for clinical data linkage only.

Withdrawal of subjects

Participants could withdraw from the study at any point for any reason. Level of withdrawal was recorded. Data were retained unless complete withdrawal was requested.

Study procedures / data collection

Part two: outcome data collection

Recruited participants completed questionnaires preoperatively and 3 years postoperatively. All participants were asked if they would be willing to be contacted about other research in the future.

Outcome measures

Patient-reported outcome measures (PROMs) questionnaires collected health-related information, including weight, medical history, smoking status, alcohol use, GI symptoms, urological health, depression, anxiety, HRQoL and obesity-specific QoL (O-QoL), life optimism, physical activity, health-care utilisation, employment and social security.

Questionnaires/instruments utilised for PROMs:

• Comorbidity was assessed by self-report using a questionnaire designed specifically for this study (Report Supplementary Material 3).

• GI reflux symptoms were evaluated using a questionnaire developed for the REFLUX trial. 73

• Urological health was assessed using the International Prostate Symptom Score (IPSS),74 where a score ≥ 8 indicates moderate to severe symptoms, and the International Consultation on Incontinence Questionnaire-Urinary Incontinence Short Form (ICIQ-UI SF), wherein a score ≥ 6 indicates moderate incontinence. 75

• Female reproductive health data were obtained using a modified version of the questionnaire developed for the Longitudinal Assessment of Bariatric Surgery study. 76

• Information on male erectile dysfunction was obtained using a modified version of the questionnaire developed for the Massachusetts Male Ageing Study. 77

• Anxiety and depression were assessed using the Generalised Anxiety Disorder Assessment (GAD-7)78 and Patient Health Questionnaire (PHQ-9)79 instruments, respectively. A PHQ-9 score ≥ 10 is indicative of moderate to severe depression, while a GAD-7 score ≥ 6 is reflective of moderate to severe anxiety. 78,79

• Smoking status was ascertained using a questionnaire specifically developed for this study (Report Supplementary Material 3).

• Alcohol use was determined using a modified version of the Alcohol Use Disorders Identification Test (AUDIT). 80

• HRQoL was assessed using the Rand 12-item Short Form Survey (SF-12)81 and Euro QoL 5-level EQ-5D version (EQ-5D-5L)82,83 instruments.

• O-QoL was assessed using the Impact of Weight on Quality of Life-Lite (IWQOL-Lite) questionnaire. 84 Standardised scoring was used when interpreting IWQOL-Lite questionnaires. 85

• Life optimism was determined using a modified version of the Life Orientation Test (LOT), wherein a score range of 0–13 reflects low optimism (high pessimism). 86

• Data on physical activity were obtained using the International Physical Activity Questionnaire (IPAQ) – Short Form. 87

• Information on participants’ employment, social security status and healthcare utilisation was obtained using questionnaires specifically developed for this study (Report Supplementary Material 3).

• Questions relating skin excess following bariatric surgery were specifically developed for this study (Report Supplementary Material 3).

• Information on postop plastic surgery was obtained using questions adapted from Ertelt et al. 88

Each participant’s quintile of the Scottish Index of Multiple Deprivation (SIMD), an area-based measure of socioeconomic status,89 was derived from their postcode. Combining a number of indicators of socioeconomic status across seven domains, the SIMD provides a relative measure of deprivation which can be used to compare data zones by ranking them from most to least deprived. The seven domains include income, employment, health, education, skills and training, housing, geographic access and crime. 89

Clinical data

Height and weight at the start of the weight-management programme were reported by clinical staff at the time of recruitment, allowing BMI to be calculated. Date of surgery, operation type, weight at operation and American Society of Anesthesiologists (ASA) grade were reported by the clinical teams. Weight at routine clinical follow-up visits and any revisional bariatric surgery procedures were also recorded.

Outcomes and data sources are summarised in Table 2. Figure 1 summarises the patient journey through the study.

FIGURE 1.

SCOTS flow chart summarising patient journey through the study.

Procedures for data collection

Patient-reported outcome measures

Completion of questionnaires could be either by post or electronically via a secure link sent by e-mail. Two reminders were sent by the participant’s chosen method and a third reminder, if required, was sent by post to all participants. No further strategy was used after three reminders.

Where patients did not complete PROMs and there was no reliable clinical weight record, they were contacted after 3 years from their date of bariatric surgery requesting completion of a weight questionnaire (simply asking their current weight). The patients were offered an incentive (£30 high-street voucher) for completing both the year-2 and year-3 questionnaires. For those no longer in clinical follow-up / not completing PROMs, an incentive (£10 high-street voucher) for completing the year-3 weight questionnaire was offered.

Clinical data

A bespoke electronic data-collection system / web-based portal was developed for SCOTS. This was secure, password-protected and used to collect clinical data from participating sites. It also allowed patients to complete questionnaires online. Information from written questionnaires was entered into the database manually by the research team, as required.

Following their operation, patients had their weight-loss surgery details entered into the SCOTS electronic data collection system / web-based portal. The clinical teams then used the electronic data-collection system to include follow-up weights, gastric-band adjustments and reoperations (including reasons for reoperations).

A detailed breakdown of patient contact within the study is described in Table 3.

TABLE 3 - SCOTS patient visit schedule, describing each interaction with a patient if they consented to part one and part two of the study

SCI Diabetes, Scottish Care Information – Diabetes; SMR01, Scottish Morbidity Record 01.

Patient contact	Time	Routine care	SCOTS	Part 1	Part 2
1	1 year to 6 weeks before surgery	Agree bariatric surgery	Patient given/sent invitation-to-participate letter and patient information sheet (and asked if they are willing to be contacted by SCOTS research team for further information and informed consent if no local recruitment).	1	1
2	1 year to 6 weeks before surgery	Patient pre-surgical clinic visit	Patient asked if they are willing to consent to participate in SCOTS. If so, patient signs consent forms. Clinical team and patient complete contact details and baseline height and weight.	1	1
3	At least 4 weeks before surgery		Patient contacted by SCOTS team to complete preop questionnaire.	1	1
4	Date of surgery	Patient has bariatric surgery	Clinical team enters details of bariatric surgery on SCOTS electronic data-collection system.	1	1
5	Patient admitted to hospital		If patient hospital admission is possibly related to bariatric surgery, identified by record linkage to SMR01.	1	1
6	At routine follow-up visits	Patient attends routine clinical visits	Clinical teams enter weight and reoperation details.		1
7	2 years post surgery	Patient has routine annual diabetes care (if has diabetes)	Patient completes 2-year post-surgical questionnaire. Blood results available via SCI Diabetes.		1
8	3 years post surgery	Patient has routine annual diabetes care (if has diabetes)	Patient completes 3-year post-surgical questionnaire. Blood results available via SCI Diabetes.		1
9	3 years onward	Patient continues to have routine diabetes and post-bariatric surgery care	Patient contacted to thank them for completing PROMs. Record linkage continues. Patient informed about future SCOTS publications.		1
Total				5	9

Bariatric surgery care pathway site survey

Statistical analyses

Data were analysed as available, without any imputation for missing data. All analyses were performed using SAS (version 9.3). Continuous data are reported as means and standard deviations (SD) or medians and lower (Q1) and upper (Q3) quartiles depending on data distribution, and counts and percentages are reported for categorical data. Comparisons between groups were made by Kruskal–Wallis test for continuous variables and Fisher’s exact test for categorial variables. Paired t-tests and Wilcoxon signed-rank test were used for change outcomes for continuous variables, McNemar test for dichotomous categorical variables and Bowker test for agreement for grouped categorical outcomes.

Binary logistic regression models were used for non-progression to surgery, admission to intensive-therapy unit (ITU) / high-dependency unit (HDU), readmission, <10% weight loss, reduction in diabetes medication, ‘need for specialist aids’ and ‘equipment in the home to assist with daily living’ outcomes. Length-of-stay outcomes were modelled with negative binomial regressions. Linear regression models were used for change in weight, change in HbA1c and change in QoL outcomes.

Regression model effect estimates, incidence rate ratios (IRRs) or odds ratios (ORs), and corresponding 95% CIs and associated p-values, are provided. The value of p < 0.05 is considered statistically significant.

A complete case analysis was performed with numbers of participants with available data listed. All results using health record data were subject to Public Health Scotland’s disclosure control protocol and outcomes affecting fewer than five participants cannot be reported (shown as xxx in tables).

Sample size

At the time of study development, 230 operations were funded in NHS Scotland each year (of which approximately 60 were bypass). Bariatric surgeons performed an additional 270 private procedures per year and they were willing to commit to entering data (approximately 80 bypass). Therefore, 500 procedures per year were expected to be entered into the database with a belief that as numbers of people with severe obesity (BMI > 40) are rising rapidly, this number will increase despite financial constraints.

From previous studies,92,93 we expected a 10-year mortality of around 5% (100 deaths). This sample size would allow the mortality rate to be estimated with 95% CI to within ± 1% (i.e. for a 5% 10-year death rate, the 95% CI will be between 4% and 6%). We planned to compare this mortality rate with an age-sex-matched healthy population from the Registrar General of Scotland’s life tables (assumed known with no sampling error). One hundred deaths were sufficient to allow us to build a predictive model for death post surgery (conventionally one requires around 10 events per prognostic covariate considered). An initial sample size of 2000 was proposed as it would easily provide adequate power for the original outcomes under investigation.

However, there were a number of unforeseen recruitment issues that have impacted on the numbers stated above and in 2016 the sample size was revisited. In order to explore whether the sample at that time was likely to show meaningful results, the available statistical power for detecting 3-year difference in HbA1c and QoL (physical and mental components) was calculated using the numbers of participants available as of 29 July 2016. The majority of data to inform sample size were taken from papers where bariatric procedure was LAGB as this is recognised as the least effective of the bariatric procedures so will give a conservative estimate of measure of effect. This shows that there is >99% power to show differences in these outcomes at 3 years with the current sample size.

In order to explore the likely event rate for cardiovascular events and deaths, we performed health record linkage for currently recruited participants, linking with inpatient care and death records. Follow-up is from the date of surgery, so for those patients who went on to have surgery details entered (n = 180), there are 272 separate admissions. Of those 272, there were 72 ‘emergency’ admissions, and 4 of these, in three patients, are ‘circulatory disease’ using the main condition only. The codes included for these hospital admissions are:

• angina pectoris (x1)

• acute MI (x1)

• pulmonary embolism without mention of acute cor pulmonale (x1)

• orthostatic hypotension (x1).

Using date of operation as the starting point, we have a total (crude) follow-up time of 203.52 years (the mean is 1.13 years and the median is 1.04 years). As the number of cardiovascular events is so low, it is impossible to extrapolate this to a future event rate at this time. It should be noted that the participants have been cleared as healthy for elective surgery, meaning that it is unlikely that there would be many cardiovascular events in early follow-up.

Following discussions with the NIHR, it was agreed that recruitment will stop at approximately 400 patients and these numbers will be sufficient to answer the majority of objectives initially set.

Ethics, regulatory and reporting requirements

The study was performed according to the Research Governance Framework for Health and Community Care (second edition, 2006)94 and was registered prospectively at the International Standard Randomised Controlled Trials Number (ISRCTN) registry: ISRCTN47072588. A favourable ethical opinion for the study was obtained from the West of Scotland Research Ethics Committee 4 on 7 February 2013 (13/WS/0005).

Permission for linkage and access to data from participants’ electronic health records was granted by the Public Benefit and Privacy Panel for Health and Social Care on 11 October 2019.

Introduction

With bariatric surgery care pathways known to vary considerably, the first step in obtaining better evidence of what works is to establish what is currently delivered. To this end, a survey of NHS-funded SCOTS study sites was undertaken in order to describe current services, to estimate their costs and explore differences in financial impact. This was necessary to facilitate further investigation as to what extent the intensity of preop and postop bariatric surgical care is a factor which may affect patient outcomes after surgery.

Results

Discussion

Bariatric surgery care pathways are widely regarded as varying considerably and international bariatric guidance is not specific with regard to the optimal model of care. 70 The results described in this chapter illustrate the large nationwide variability in preop and postop care, a likely consequence of widespread uncertainty regarding best practice and a lack of more detailed guidance with respect to service delivery. There is little evidence as to whether intensive preop and postop care improves outcomes and is cost-effective compared to less intensive care. This is likely to be more complicated than one standard pathway for all, with patient preferences also paramount in terms of type of provision (one-to-one or group sessions, for example). Furthermore, pre-surgery targets vary widely95 but are often low-cost group interventions and funded from a separate budget to surgery. Maximum cost is around £100–£200 per patient. However, these targets do add to the complexity of the pathway for the patients and variation in time and access to surgery, and therefore the usefulness of these targets is currently a subject of debate. 96–99

Impacts resulting from the benefits of dietician and psychological support prior to bariatric surgery have been published. Livhits et al. 100 undertook a systematic review which found that preop weight loss appears to be associated with greater weight loss postoperatively. In a more recent review, Gerber et al. 101 found the same beneficial effects from preop weight loss. On the other hand, it has been shown that psychological support before and after bariatric surgery had no impact on weight loss. 102 This study recommends further research to evaluate the longer-term implications for both weight loss and psychological support, and thereby the most effective timing for delivery of these interventions. As to why some sites offer more comprehensive services than others, decisions on staff resourcing are possibly being made on the basis of cost and availability of specialists, as there is currently no evidence as to whether these different models of care pathways improve outcomes. Indeed, this study illustrates how variable these costs are, even across health centres within the same country context, and this difference alone is worth highlighting. Therefore, it is important to evidence outcomes of these services.

Furthermore, there is a concern that bariatric surgery cost-effectiveness models may either omit pre-surgery and post-surgery care costs as part of their economic analyses or treat patients and the delivery of these services homogeneously by applying average costs. In a systematic review of a critical appraisal of economic evaluations of bariatric surgery,103 the considerable heterogeneity of what costs are included in economic studies and the frequent omission of different types of healthcare resource use were highlighted. Despite the identification of preoperative and postoperative costs, there was no detail reported on care pathways explicitly as an important cost component of an economic evaluation of bariatric surgery. A recent study by Gulliford et al. ,104 estimating the costs of bariatric surgery drawn from UK NHS tariffs, included preoperative weight management as part of the cost of the surgical procedure but only referred to the cost of medical weight-management services. There was no reference to bariatric surgery care pathway costs being included. 104 In the same model, a flat rate of £875 was also included for postoperative reviews. Procedure costs are not captured here and are assumed to be relatively standardised given the clear guidance on surgical procedures and, in Scotland, there is national procurement so device costs would also be standard across all sites. In their systematic review, Picot et al. 105 found the costs of bariatric surgery generally to be presented as standard unit costs with aggregate costs differing dependent on what is included in the total costs of surgery rather than any differences due to site variation. One study106 did find variation by gender but offered no explanation as to why.

The aim of this research was to understand whether differences in these care pathways are predictors of health outcomes, and thus influence cost-effectiveness from the benefit side. This study underlines the need to better understand the cost-effectiveness of bariatric surgery care pathways, and whether the varying level of intensity of services offered is an important factor in influencing outcomes. The SCOTS study provides the follow-up data required to assess whether this classification of preop and postop care pathways is a predictor of health outcomes. Classification of the intensity of preop and postop bariatric surgical care can now be considered for investigation as a factor which may affect patient outcomes after surgery. If further findings do demonstrate that more intensive (and expensive) services lead to better outcomes, it is not envisaged that this will change bariatric surgery from being cost-effective at the usual willingness-to-pay thresholds for reimbursement on the NHS given the modelled ICER of £10,126 per quality-adjusted life year. 55 However, budgetary impact is an important consideration and it is acknowledged that these costs do matter for payers, hospital resource use and more local-level decision-making. Should these pathways be found to be predictors of better health outcomes, the case for investment in these care pathways would be self-evident.

Conclusions

This study, focusing on preop costs and the first 12 months following surgery in which the majority of costs will occur, has illustrated the large nationwide variability in preop and postop care pathways across Scotland, and the subsequent financial impact on the provision of bariatric surgery services. This is a likely consequence of widespread uncertainty regarding best practice and a lack of more detailed guidance regarding service delivery. Health economic analyses do not always capture these costs103 or apply a flat rate. 104 There is a lack of evidence base and a clear requirement for the evaluation of bariatric surgical services to identify the care pathways preceding and following surgery which lead to the largest improvements in health outcomes and remain cost-effective to the health provider.

Introduction

There is a lack of evidence to inform the delivery and follow-up of bariatric surgery for people with severe obesity. SCOTS is the first national epidemiological study established to investigate long-term outcomes following bariatric surgery. In addition, SCOTS collected clinical and patient-reported health outcomes from treatment-seeking individuals from across Scotland with severe obesity before they underwent bariatric surgery. This chapter describes the health-related characteristics of the recruited SCOTS cohort and examines relationships between age, preop BMI and other health-related factors.

Results

Recruitment

Participants were recruited over an approximate 3-year period from December 2013 to February 2017 with follow-up continuing until October 2020.

Over the recruitment period, a total of 548 patients were approached and screened for eligibility to participate. Of these, 103/548 (19%) were excluded or declined to participate (see Figure 2). We recruited 445/548 (81%) participants but one participant withdrew consent, leaving a recruited sample of 444 (81%). Of the recruited sample, 413/444 (93%) consented to data linkage and questionnaire follow-up, while 31/444 (7%) consented to data linkage only. Of these 413 participants, a total of 164/413 (40%) were not included in the subsequent analysis: 129 did not return a baseline questionnaire and 35 had bariatric surgery before their baseline PROMs questionnaires were completed. Of the 129 who did not return baseline questionnaires, 84/129 (65%) progressed to surgery, 43/129 (33%) did not progress to surgery and the status of 2/129 (2%) was unknown. Completed preop baseline PROMs data for 249/413 participants (60% of those consented) were available for analysis (see Figure 2).

FIGURE 2.

Screening, consent and follow-up.

Discussion

Despite escalating levels of severe obesity in Western society and the concomitant increase in bariatric surgical procedures being performed in some countries,107 there is a dearth of information on the health status of people living with severe obesity. This national Scottish cohort study of people seeking surgical treatment for severe obesity recruited 444 adults from 14 centres across Scotland over a 3-year period, including all NHS centres and major private hospitals undertaking bariatric surgery. We found that higher BMI and older age were associated with decreased physical QoL, increased use of specialist aids and equipment in the home, and a high prevalence of comorbidities.

There has been a significant increase in the prevalence of BMI ≥ 40 kg/m² in recent decades; in Scotland, obesity prevalence has trebled in women age 16–64 years since 1995. However, it is hard to assess the global increase due to lack of reporting of BMI ≥ 40 kg/m² in national health survey data. 108 While it is known that healthcare resource use increases in people with a BMI ≥ 30 kg/m², with service use estimated to be over 25% higher than for those with a BMI in the normal weight range,109 few data exist for those with BMI 40 kg/m² and above. In 2016, the Global BMI Mortality Collaboration110 conducted an individual-participant-data meta-analysis of 239 prospective studies and found a 2.8 times increased risk of all-cause mortality for people with a BMI of 40–60 kg/m². Greive et al. 111 conducted a systematic review which focused on the economic cost of severe obesity (BMI ≥ 40 kg/m²) and found limited literature describing increased prescribing, outpatient utilisation and intensive care admission and hospital length of stays during critical illness. However, in neither study was there disaggregation of BMI beyond >40 kg/m², meaning that the health consequences of severe obesity are not yet fully described.

There has been extensive research on the relationship between HRQoL and obesity. 112 Ul-Haq et al. 113 performed a meta-analysis of eight studies (43,086 participants) and found physical QoL, measured by the Rand 36-item Short Form Health Survey (SF-36), was reduced by 9.7 points in those with BMI 40 kg/m² compared to those with a BMI in the normal range, although, again, there was no disaggregation above BMI 40 kg/m². Van Nunen et al. 114 performed a meta-analysis to compare the general, non-treatment-seeking population to patients within weight-management programmes and those seeking bariatric surgery. They found that those seeking surgical treatment reported the most severely reduced HRQoL, perhaps reflecting their reasons for seeking definitive surgical treatment. Our cohort of treatment-seeking individuals, who completed a rich battery of patient-reported measures, provides data to show that HRQoL and O-QoL of those with the highest body mass is extremely poor and this is compounded by increasing age. QoL scores of SCOTS participants in both the upper BMI (≥55 kg/m²) and older age (≥55 years) groups included physical scores comparable with those reported by cancer patients receiving palliative care,115 patients with chronic heart failure expressing end-of-life preferences116 and patients with end-stage kidney disease. 117 Furthermore, patients with severe chronic obstructive pulmonary disease (COPD) report higher QoL scores, indicating a better QoL, than our cohort of treatment-seeking obese participants. 118 As far as we are aware, this is the first study to investigate physical and mental health in patients with severe obesity awaiting bariatric surgery with finer-level consideration of BMI up to ≥55 kg/m².

As previously outlined, UK guidelines63 currently indicate that bariatric surgery is a treatment option for those with BMI ≥ 40 kg/m², or between 35 kg/m² and 40 kg/m² in the presence of other significant diseases which could be improved if they lost weight. Non-surgical weight management must have been attempted but not resulted in clinically beneficial weight loss before surgery is specified. However, baseline SCOTS data appear to suggest that the low prioritisation of bariatric surgery and a lengthy preop pathway in the UK is associated with surgical treatment being reserved for individuals at an older age with very high BMI. Indeed, in 2018, the Global Registry initiative of the International Federation for the Surgery of Obesity and Metabolic Disorders (IFSO) reported a global median pre-bariatric surgery BMI of 41.7 kg/m²,119 as compared to 47.6 kg/m² in the SCOTS cohort. Similarly, IFSO reported a median patient age of 42 years at the time of bariatric surgery,119 as compared to a median age of 47 years for SCOTS participants. This combination of higher BMI and older age means that, at the time of surgery, Scottish patients have high levels of comorbidity and poor physical functioning.

Bariatric surgery is considered a highly cost-effective intervention. 55 However, the health economic models rely on data primarily from US and Scandinavian studies,16,17,120–122 where BMI and age at the time of surgery are lower than in the UK. Higher BMI and older age are risk factors for postop complications123,124 and also associated with lower total weight loss. 125–127 T2DM remission rates are negatively correlated with age. 128 As such, focusing bariatric surgery provision on those with older age and higher BMI may result in higher costs of surgery with increased length of hospital stay, higher rates of postop complications, lower overall weight loss and lower rates of disease remission. Consequently, the impressive health benefits and resultant cost-savings of bariatric surgery observed in clinical trials and observational cohorts from other countries may not be fully realised for the UK/Scottish population.

The SCOTS dataset represents a unique and rich resource. A major strength of the study is its representativeness. Indeed, every clinical team providing publicly funded bariatric surgery in Scotland approached their patients for recruitment to the study, rendering it highly representative of the population in comparison to other studies undertaken in the field. However, the number of participants with valid baseline questionnaires was lower than anticipated. In many cases, this could be attributed to the participant undergoing surgery before completing the questionnaire, or the participant leaving the bariatric surgery pathway before surgery. The overall length of the questionnaire may have also played a role. Participants living in the most deprived areas were well represented in our cohort and the mean QoL findings were broadly similar to those of bariatric surgery cohorts from across the world. 129–131 A further strength of the study is that questionnaires were externally validated and wide-ranging, containing a number of unique questions covering medical, social, psychological and physical functioning domains. This wide range of self-reported health measures will allow us to account for a range of potentially mediating and confounding factors in future analyses. In addition, we have revealed the extent of comorbidities, including musculoskeletal, urinary and mental health problems affecting people with severe obesity. Low numbers of some comorbidities meant that this could not be a focus of this analysis.

A limitation of this study is that selection for bariatric surgery is often based on the presence of comorbidity so these results, while applicable to a treatment-seeking population, may not be directly applicable to the whole population with severe obesity in the wider society. While we will have access to medical records via electronic health record data linkage in follow-up, the current analyses are based on self-report of selected comorbidities. It is well known that self-reported weights are underreported, particularly by people with very high BMI. 132 However, we are confident of the accuracy of weight and height as these data were collected in clinic during the recruitment visit.

Conclusion

Obesity is a multisystem disease which affects every facet of a person’s life. Our data have shown that higher BMI combined with older age is associated with very poor physical functioning, and HRQoL and O-QoL. Indeed, QoL scores for those living with severe obesity in Scotland are akin to those seen in the end stage of diseases such as cancer and heart failure. The health consequences of severe obesity and the extent to which treatments such as bariatric surgery can improve these are not yet known. Researchers should ensure that they include people with severe obesity in population cohorts and treatment studies, and study the impact of severe obesity in more detail; there are substantial differences in the health status of those with a BMI >50 kg/m² and those whose BMI is around 40 kg/m². Policy-makers should consider the health and care needs of the growing numbers of individuals living with obesity. There will be considerable future demand for health care and services must be designed to accommodate the physical needs of the individuals. While primary prevention of obesity is clearly paramount to avoid more people developing such a debilitating, chronic condition, investment is urgently needed, both in the UK and globally, to provide increased access to bariatric surgery and other forms of effective weight management, directly targeting patient groups who will benefit from surgical intervention as early in the disease course as possible.

Introduction

As outlined in previous chapters, SCOTS obtained full data on pre-surgical and post-surgical pathways, criteria for progression to surgery, staffing and frequency of visits. This allowed services to be outlined by intensity and cost. Herein, we compare these factors, along with patient-related factors such as age, BMI and comorbidity, to patient outcomes at 1 year postoperatively, including weight loss, length of hospital stay, readmission and the need for ITU/HDU admission postoperatively.

Results

Progression to surgery

Of the recruited sample of SCOTS participants (recruited sample), 336/444 (76%) progressed to surgery (operated sample), 92/444 (21%) did not progress to surgery (non-progression to surgery sample) and 14/444 (3%) were still awaiting surgery at the end of the SCOTS study (awaiting surgery sample, see Figure 3). Baseline characteristics of these four samples are shown in Table 9. The cohort that did not progress to surgery had a higher proportion of males, a higher proportion of participants aged 55 years or older, a higher proportion of participants in the lowest SIMD quintile and a higher median BMI at the start of the weight-management programme than those who progressed to surgery.

FIGURE 3.

Screening, consent and follow-up to 1 year post initial bariatric surgery.

TABLE 9 - Baseline characteristics of recruited, all operated, non-progression to surgery and awaiting surgery samples of SCOTS participants

		Recruited sample N = 444	All operated sample N = 336	Non-progression to surgery sample N = 92	Awaiting surgery sample N = 14
Sex, N (%)	Male	123 (27.7)	85 (25.3)	36 (39.1)	1 (7.1)
	Female	321 (72.3)	251 (74.7)	56 (60.9)	13 (92.9)
	Missing	0	0	0	0
Age (years)	Mean (SD)	46.2 (9.1)	46.01 (9.15)	47.10 (8.76)	45.41 (9.06)
	Missing	0	0	0	0
Age group, N (%)	<35 years	61 (13.7)	47 (14.0)	11 (12.0)	2 (14.3)
	35–44 years	116 (26.1)	86 (25.6)	25 (27.2)	4 (28.6)
	45–49 years	109 (24.5)	86 (25.6)	20 (21.7)	3 (21.4)
	50–54 years	79 (17.8)	61 (18.2)	15 (16.3)	3 (21.4)
	55 + years	79 (17.8)	56 (16.7)	21 (22.8)	2 (14.3)
	Missing	0	0	0	0
SIMD quintile, N (%)	Quintile 1 (most deprived)	135 (30.5)	98 (29.3)	32 (34.8)	4 (28.6)
	Quintile 2	108 (24.4)	83 (24.9)	25 (27.2)	0 (0.0)
	Quintile 3	84 (19.0)	56 (16.8)	24 (26.1)	4 (28.6)
	Quintile 4	68 (15.4)	57 (17.1.)	7 (7.6)	3 (21.4)
	Quintile 5 (least deprived)	47 (10.6)	40 (12.0)	4 (4.3)	3 (21.4)
	Missing	2	2	0	0
BMI at start of weight-management programme (kg/m2)	Median (Q1; Q3)	47.2 (42.7; 53.6)	46.4 (42.4; 52.0)	50.2 (45.5; 59.4)	42.3 (39.5; 45.2)
	Missing	1	1	0	0
BMI group at start of weight-management programme, N (%)	BMI < 40	52 (11.7)	44 (13.1)	4 (4.3)	4 (28.6)
	BMI 40–44	115 (26.0)	91 (27.2)	18 (19.6)	6 (42.9)
	BMI 45-49	116 (26.2)	92 (27.5)	22 (23.9)	1 (7.1)
	BMI 50–54	71 (16.0)	55 (16.4)	14 (15.2)	2 (14.3)
	BMI 55+	89 (20.1)	53 (15.8)	34 (37.0)	1 (7.1)
	Missing	1	1	0	0
Weight at start of weight-management programme (kg)	Median (Q1; Q3)	130 (117; 151)	129 (115; 146)	145 (126; 168)	115 (98.6; 126)
	Missing	0	1	0	0
BMI at date of initial bariatric surgery (kg/m2)	Median (Q1; Q3)	—	43.2 (39.7; 48.7)	—	—
	Missing		56
BMI group at date of initial bariatric surgery, N (%)	BMI < 40	—	80 (28.6)	—	—
	BMI 40–44		85 (30.4%)
	BMI 45-49		59 (21.1%)
	BMI 50–54		34 (12.1%)
	BMI 55+		22 (7.9%)
	Missing		56
Weight at date of initial bariatric surgery (kg)	Median (Q1; Q3)	—	121 (107; 136)	—	—
	Missing		56
Change in weight from start of weight-management programme to date of initial bariatric surgery (kg)	Median (Q1; Q3)	—	−7.4 (−14.0; −2.0)	—	—
	Missing		57

The main reasons reported by sites for non-progression to surgery for SCOTS participants are shown in Table 10, with patient decision (for reasons other than weight loss/stress) being most frequently given (37% of participants with non-progression) followed by failure to achieve pre-surgical goals (31.5% of participants with non-progression).

TABLE 10 - Reasons for non-progression to surgery for the non-progression sample of SCOTS participants

More than one reason could be cited per participant.

Reason for surgery not proceedinga	Number of participants (%)
Medical/surgical/anaesthetic reason	16 (17.4)
Psychological contraindication	7 (7.6)
Failure to achieve pre-surgical goals	29 (31.5)
Achieved weight loss via other means	7 (7.6)
Major life event or stressor	1 (1.1)
Patient decision (for reasons other than weight loss/stress)	34 (37.0)
Sought private surgery	0
Other	7 (7.6)
Not known	3 (3.3)
>1 of reasons listed above	11 (12.0)

As outlined in Table 11, BMI ≥ 55 kg/m² and male sex were both associated with around twice the odds of non-progression to surgery as compared to BMI 45–49 kg/m² and female sex, respectively.

TABLE 11 - Odds of non-progression to surgery – multivariable logistic regression

Variable	Explanatory variable groups	OR (95% CI)	Overall p-value
Age group (years)	<35	0.95 (0.40, 2.25)
	35–44	1.49 (0.74, 3.00)
	45–49	1.00 (–)	0.59
	50–54	1.22 (0.56, 2.69)
	55+	1.67 (0.78, 3.54)
BMI group at recruitment (kg/m2)	<40	0.34 (0.11, 1.08)
	40–44	0.79 (0.39, 1.60)
	45–49	1.00 (–)	< 0.001
	50–54	0.91 (0.42, 1.95)
	55+	2.81 (1.45, 5.47)
Sex	Male	1.96 (1.16, 3.32)
	Female	1.00 (–)	0.01
SIMD quintile	SIMD Q1 (most deprived)	1.00 (–)	0.03
	SIMD Q2	1.05 (0.55, 2.00)
	SIMD Q3	1.54 (0.79, 2.98)
	SIMD Q4	0.41 (0.17, 1.02)
	SIMD Q5 (least deprived)	0.36 (0.12, 1.14)

Baseline characteristics of all operated sample of SCOTS participants by bariatric operation type

Baseline characteristics of the operated sample, by operation type, are summarised in Table 12. Mean age of the all operated sample was 46 years (± 9.2 years), with a higher proportion of women than men (75% vs. 25%). Approximately half of the all operated sample of participants were aged 35 to 49 years, with one-third being 50 years or older. There was a median weight change of −7.4 kg (Q1 −14.0; Q3 −2.0) for the all operated sample from the start of the weight-management programme until the date of initial bariatric surgery. The median BMI at date of initial bariatric surgery for the all operated sample was 43.2 kg/m² (Q1 39.7; Q3 48.7), with almost 8% having a BMI ≥ 55 kg/m². Over half of the participants in the all operated group (54%) lived in areas of high socioeconomic deprivation (SIMD quintiles 1 and 2). Of the all operated sample of SCOTS participants, 42/336 (12.5%) had LAGB surgery, 128/336 (38.1%) had RYGB surgery, 165/336 (49.1%) had SG surgery and for 1/336 (0.3%) operation type data were missing (see Table 14).

TABLE 12 - Baseline characteristics of all operated sample of SCOTS participants by operation type

One participant has missing operation type data and therefore numbers in operation type columns add up to 335.

		All operated sample N = 336a	LAGB N = 42	RYGB N = 128	SG N = 165
Sex, N (%)	Male	85 (25.3)	13 (31.0)	33 (25.8)	39 (23.6)
	Female	251 (74.7)	29 (69.0)	95 (74.2)	126 (76.4)
	Missing	0	0	0	0
Age (years)	Mean (SD)	46.01 (9.15)	46.95 (8.78)	45.42 (8.35)	46.18 (9.84)
	Missing	0	0	0	0
Age group, N (%)	<35 years	47 (14.0)	5 (11.9)	18 (14.1)	24 (14.5)
	35–44 years	86 (25.6)	10 (23.8)	34 (26.6)	42 (25.5)
	45–49 years	86 (25.6)	9 (21.4)	39 (30.5)	38 (23.0)
	50–54 years	61 (18.2)	11 (26.2)	21 (16.4)	28 (17.0)
	55+ years	56 (16.7)	7 (16.7)	16 (12.5)	33 (20.0)
	Missing	0	0	0	0
SIMD quintile, N (%)	Quintile 1 (most deprived)	98 (29.3)	15 (35.7)	29 (22.8)	54 (32.9)
	Quintile 2	83 (24.9)	9 (21.4)	33 (26.0)	41 (25.0)
	Quintile 3	56 (16.8)	8 (19.0)	25 (19.7)	23 (14.0)
	Quintile 4	57 (17.1)	10 (23.8)	21 (16.5)	26 (15.9)
	Quintile 5 (least deprived)	40 (12.0)	0 (0.0)	19 (15.0)	20 (12.2)
	Missing	2	0	1	1
BMI at start of weight-management programme (kg/m2)	Median (Q1; Q3)	46.4 (42.4; 52.0)	45.7 (42.2; 49.5)	46.5 (41.9; 51.9)	46.4 (42.7; 52.9)
	Missing	1	0	1	0
BMI group at start of weight-management programme, N (%)	BMI < 40	44 (13.1)	4 (9.5)	23 (18.1)	16 (9.7)
	BMI 40–44	91 (27.2)	15 (35.7)	28 (22.0)	48 (29.1)
	BMI 45–49	92 (27.5)	15 (35.7)	35 (27.6)	42 (25.5)
	BMI 50–54	55 (16.4)	6 (14.3)	23 (18.1)	26 (15.8)
	BMI 55+	53 (15.8)	2 (4.8)	18 (14.2)	33 (20.0)
	Missing	1	0	1	0
Weight at start of weight-management programme (kg)	Median (Q1; Q3)	129 (115; 146)	130 (115; 143)	127 (114; 146)	129 (116; 148)
	Missing	1	0	1	0
BMI at date of initial bariatric surgery (kg/m2)	Median (Q1; Q3)	43.2 (39.7; 48.7)	42.8 (40.5; 47.9)	42.6 (38.6; 48.9)	43.9 (39.8; 48.8)
	Missing	55	1	32	22
BMI group at date of initial bariatric surgery, N (%)	BMI < 40	80 (28.6)	9 (22.0)	33 (34.4)	38 (26.6)
	BMI 40–44	85 (30.4)	16 (39.0)	26 (27.1)	43 (30.1)
	BMI 45–49	59 (21.1)	11 (26.8)	16 (16.7)	32 (22.4)
	BMI 50–54	34 (12.1)	4 (9.8)	14 (14.6)	16 (11.2)
	BMI 55+	22 (7.9)	1 (2.4)	7 (7.3)	14 (9.8)
	Missing	55	1	32	22
Weight at date of initial bariatric surgery (kg)	Median (Q1; Q3)	121 (107; 136)	122 (111; 134)	120 (106; 134)	121 (107; 137)
	Missing	55	1	32	22
Change in weight from start of weight-management programme to date of initial bariatric surgery (kg)	Median (Q1; Q3)	−7.4 (−14.0; −2.0)	−3.6 (−15.0; 0.0)	−9.4 (−16.0; −2.6)	−7.4 (−12.0; −2.5)
		56	1	33	22

LAGB surgery had the largest proportion of participants aged 50 years or older (42.9%) and the highest proportion of participants (35.7%) living in areas of the highest level of socioeconomic deprivation (SIMD quintile 1). SG surgery had the highest proportion of participants with a BMI ≥ 55 kg/m² at the date of initial bariatric surgery (9.8%). RYGB surgery had the greatest median weight change from the start of weight management until initial bariatric surgery; −9.4 kg (Q1 −16.0; Q3 −2.6).

The proportion of each surgery type by year of operation is shown in Figure 4. LAGB decreases as a proportion of operations performed and SG increases over the period of the study, with 65% of operations being SG and only 7% LAGB by 2017.

FIGURE 4.

Number and type of initial bariatric procedure by year of operation.

Discussion

We have described outcomes up to 1 year post bariatric surgery, including associations with baseline descriptive characteristics and patient-reported outcomes, in a contemporary cohort of patients from Scotland, UK.

Of the 444 recruited participants, 92 (21%) did not progress to bariatric surgery, with the most frequent reasons given being patient decision or failure to achieve pre-surgery goals. There is limited evidence as to the effect of pre-surgery goals on post-surgery outcomes133–135 and, as these goals are often specific to the individual rather than common across a programme, it is unclear what these goals were for each individual SCOTS participant. Often, goals are to achieve a 5% or 10% pre-surgery weight loss, or are related to maintaining a certain attendance rate at the pre-surgical programme.

For those participants who had bariatric surgery, SG was the predominant type and very few LAGB procedures were performed, in keeping with the international norm. 136 Those having SG surgery were older, had higher BMIs and resided in areas with higher levels of socioeconomic deprivation than those undergoing RYGB surgery. This is likely to reflect a choice made by the surgical team based on risks associated with each surgery type; RYGB surgery is known to have a slightly higher complication rate than other bariatric surgery types and is, therefore, less suitable for those who may have a greater severity of underlying comorbidity. 137

Weight change of SCOTS participants at 1 year post bariatric surgery was consistent with 1-year weight change data published from trials and other cohort studies, including registries where RYGB surgery was shown to result in the greatest weight loss. 138 Weight-loss failure, defined as a loss of less than 10% at 1 year post bariatric surgery, was an infrequent outcome for SCOTS participants, with the majority of cases seen in those who had LAGB surgery. While those SCOTS participants with the highest pre-surgical BMI had the greatest postop weight loss, higher preop BMI has not been shown to be a predictor of greater postop weight loss in other studies. 139

The role of preop weight loss within the bariatric surgery pathway is controversial, with many studies concluding that mandatory preop weight loss does not result in greater weight-loss outcomes. 140 Certainly, our data suggest that those with greater weight loss preoperatively have less weight loss postoperatively. However, for those SCOTS participants with greater preop weight loss, total weight loss from the start of the preop programme was greater and preop weight loss was associated with a far lower odds of having less than 10% weight loss at 1 year post surgery. Contrary to what may have been expected, greater pre-surgical weight loss was associated with increased odds of ITU/HDU admission post bariatric surgery. As this was an observational study, it is impossible to make a causal link but, rather than pre-surgical weight loss somehow necessitating ITU/HDU admission postoperatively, it may simply be a marker for patients whom the surgical team perceived as being at increased risk of post-surgical complications due to high body mass or multiple comorbidities and, therefore, for whom higher preop weight-loss targets were set in an attempt to reduce risk.

For SCOTS participants, the median length of stay for the initial bariatric procedure was 3 days, which is high compared to the international mean of 2 days. 136 Numbers of participants admitted to ITU/HDU postoperatively were also high, although it was noted that the median length of stay in ITU/HDU was only 1 day and very few of those admitted had a second operative procedure during their admission. Mortality was minimal, with no deaths within 30 days of bariatric surgery and very low numbers up to 1 year post surgery. Taken together, this suggests sites may be overly cautious when it comes to the postop management of patients who have undergone bariatric surgery.

Up to 1 year post bariatric surgery, readmissions for SCOTS participants were higher than the international mean. 136 However, very low numbers of operative procedures were performed during these admissions despite ‘surgical’ reasons for the readmissions being most common. The reasons for high rates of readmission without the need for surgery are not known. It may reflect the high baseline comorbidity of those having surgery or be a symptom of an inefficiency in the pathways for specialist care. We know that the volume of bariatric surgery procedures is low in Scottish hospitals compared to international practice and that may mean that there is an increased number of referrals and admissions by non-specialists.

SCOTS site intensity was defined by the high, medium or low cost of the preop programme. Lower site pathway intensity was associated with increased odds of admission to ITU/HDU during the initial bariatric surgery procedure for those SCOTS participants undergoing bariatric surgery on the NHS. Within the operated NHS-funded cohort, participants being treated at sites with a high-intensity preop surgical pathway had a lower median BMI at the start of the pathway and a far larger weight loss during the pre-surgical pathway, resulting in a significantly lower BMI at the time of initial bariatric surgery. This could certainly explain their lower admission rates to ITU/HDU postoperatively. However, the characteristics of a site with a high-intensity pre-surgical pathway should also be considered; implementation of a multidisciplinary preop pathway is suggestive of a site with a greater number of dedicated bariatric surgery staff and, potentially, a higher volume of bariatric surgery patients. It may, therefore, be this experience within the team that results in the difference in outcomes. The lower initial BMI in the high-intensity pathway cohort may reflect a higher volume of patients and a more transparent pathway for referral from primary care.

These results raise questions as to the efficiency of bariatric surgery in Scotland, with high numbers of ITU/HDU admissions and hospital readmissions but very limited reoperations suggesting that a cautious approach may be being adopted. The fact that weight loss is in keeping with international means and mortality rate is low serves to reassure that despite this apparent over-caution, the outcome aims of surgery are being achieved. However, there are many questions on bariatric surgery care that this study cannot answer and that is a major limitation. While we know the typical pathway of care in each site, we do not know what happened to each individual participant and the basis for each decision. An ethnographic study would add depth to our understanding of pre-surgical pathways and identify areas where decisions are made and the level of uncertainty and evidence on which those are based. The complexity of interpreting observational data means that it is impossible to make conclusions on the benefits (or otherwise) of preop weight loss, its use as a predictor of longer-term weight loss outcomes, and how it should be used in practice. Studies to date have either looked at those achieving significant weight loss preoperatively or have compared those with and without insurance-mandated weight loss. 97,141,142 What is required is a large randomised trial of weight loss versus weight stability in the population seeking bariatric surgery with follow-up for several years postoperatively.

Our results indicate that 21% of those starting a bariatric surgery pathway do not, ultimately, undergo surgery. This represents not only wasted resources and wasted opportunity to provide alternative treatments but a potential psychological cost to the individual associated with ‘failure’. Male sex and higher BMI were associated with increased odds of non-progression. This could offer some indication as to the population that may benefit from early review by an anaesthetist and/or other measures to inform decision-making by the individual patient earlier in the pathway. However, ‘failure to achieve preop goals’ was also a major reason for non-progression and this may reflect inequalities with regard to the preop pathway; weight-management programmes often require regular attendance at face-to-face classes and consultations which can be during working hours and, with men more likely to work full-time, attendance may have been an issue for them. We know that in our cohort, those with higher BMI had far poorer physical function, meaning that repeat attendance and, consequently, achieving pre-surgical weight loss goals may have proved difficult. The effects of multiple long-term comorbidity medications may have also contributed to lower pre-surgical weight loss. In addition, poorer mental health may have made regular class/consultation attendance and non-surgical weight loss difficult for participants. Poor attendance would have led to exclusion of individuals with complex comorbidities when evidence suggests that they would benefit significantly from bariatric surgery. 143

Conclusion

In conclusion, surgical outcomes for weight and mortality for the SCOTS cohort are in keeping with international results, but there are high numbers of ITU/HDU admissions and hospital readmissions that may be unnecessary. Questions about the optimal pre-surgery pathway to maximise safety and weight loss remain unanswered as current pathways may reflect volume and therefore specialist experience of staff and impact on patient selection.

Introduction

In this chapter we compare pre- and post-surgical pathway intensity and patient-related factors such as age, BMI and comorbidity, to patient outcomes at 3 years postoperatively, including weight loss, diabetes outcomes, length of hospital stay, readmission, reoperation and the need for ITU/HDU admission postoperatively.

Results

Follow-up to 3 years post surgery

Of the recruited sample of SCOTS participants (recruited sample), 336/444 (76%) progressed to surgery (operated sample), 92/444 (21%) did not progress to surgery (non-progression to surgery sample) and 14/444 (3%) were still awaiting surgery at the end of the SCOTS study (awaiting surgery sample, see Figure 3). Baseline characteristics of the all operated sample (n = 336) are shown in Table 9 and all operated with available preop PROMs in Table 15. Baseline characteristics for the all operated with available 3-year PROMs (n = 85) and all operated with T2DM (n = 192) are shown in Tables 25–27 and Report Supplementary Material 11, Table c, respectively. The sample with T2DM had a higher proportion of participants who were male, higher mean age and a higher proportion of participants aged ≥55 years, than the all operated cohort. Figure 5 outlines the follow-up to 3 years post bariatric surgery of the all-operated sample.

TABLE 25 - Baseline sociodemographic characteristics of the all operated who consented to PROMs with available 3-year PROMs outcome data sample by operation type

		All operated with PROMs N = 85	LAGB N = 13	RYGB N = 33	SG N = 39
Sex, N (%)	Male	25 (29.4%)	3 (23.1%)	14 (42.4%)	8 (20.5%)
	Female	60 (70.6%)	10 (76.9%)	19 (57.6%)	31 (79.5%)
	Missing	0	0	0	0
Age (years)	Mean (SD)	46.86 (8.46)	46.16 (8.42)	46.11 (7.84)	47.74 (9.08)
	Missing	0	0	0	0
Age group, N (%)	<35 years	9 (10.6%)	2 (15.4%)	4 (12.1%)	3 (7.7%)
	35–44 years	23 (27.1%)	3 (23.1%)	9 (27.3%)	11 (28.2%)
	45–49 years	18 (21.2%)	3 (23.1%)	7 (21.2%)	8 (20.5%)
	50–54 years	21 (24.7%)	4 (30.8%)	8 (24.2%)	9 (23.1%)
	55+ years	14 (16.5%)	1 (7.7%)	5 (15.2%)	8 (20.5%)
	Missing	0	0	0	0
SIMD deprivation quintile, N (%)	Quintile 1 (most)	17 (20.0%)	1 (7.7%)	4 (12.1%)	12 (30.8%)
	Quintile 2	25 (29.4%)	4 (30.8%)	13 (39.4%)	8 (20.5%)
	Quintile 3	13 (15.3%)	4 (30.8%)	4 (12.1%)	5 (12.8%)
	Quintile 4	18 (21.2%)	4 (30.8%)	7 (21.2%)	7 (17.9%)
	Quintile 5 (least)	12 (14.1%)	0 (0.0%)	5 (15.2%)	7 (17.9%)
	Missing	0	0	0	0
Weight at start of weight-management programme (kg)	Median (Q1; Q3)	129 (115; 142)	122 (115; 130)	127 (120; 145)	130 (113; 142)
	Missing	0	0	0	0
Weight at date of initial bariatric surgery (kg)	Median (Q1; Q3)	122 (106; 134)	113 (102; 124)	123 (108; 134)	121 (106; 139)
	Missing	16	1	10	4
Change in weight from start of weight-management programme to date of initial bariatric surgery (kg)	Median (Q1; Q3)	−7.1 (−13; −0.8)	−3.0 (−24; 0.0)	−7.7 (−17; −2.0)	−7.6 (−12; −0.3)
	Missing	15	1	10	4
BMI at date of initial bariatric surgery (kg/m2)	Median (Q1; Q3)	42.6 (39.1; 48.6)	41.1 (39.4; 47.4)	40.8 (36.1; 48.3)	44.2 (40.2; 49.8)
	Missing	15	1	10	4
BMI group at date of initial bariatric surgery, N (%)	BMI < 40	22 (31.4%)	4 (33.3%)	10 (43.5%)	8 (22.9%)
	BMI 40–44	19 (27.1%)	3 (25.0%)	4 (17.4%)	12 (34.3%)
	BMI 45–49	17 (24.3%)	5 (41.7%)	5 (21.7%)	7 (20.0%)
	BMI 50–54	7 (10.0%)	0 (0.0%)	2 (8.7%)	5 (14.3%)
	BMI 55+	5 (7.1%)	0 (0.0%)	2 (8.7%)	3 (8.6%)
	Missing	16	1	10	4
Marital status, N (%)	Married/civil partnership/living as married	53 (63.1%)	8 (61.5%)	24 (68.6%)	23 (71.8%)
	Other (single/separated/divorced widowed)	31 (36.9%)	5 (38.5%)	11 (31.4%)	9 (28.1%)
	Missing	1	0	0	1
Ethnic group, N (%)	White	81 (95.3%)	12 (92.3%)	30 (90.9%)	39 (100.0%)
	Mixed	2 (2.4%)	1 (7.7%)	1 (3.0%)	0 (0.0%)
	Asian/Asian Scottish/Asian British	1 (1.2%)	0 (0.0%)	1 (3.0%)	0 (0.0%)
	African Caribbean/black	1 (1.2%)	0 (0.0%)	1 (3.0%)	0 (0.0%)
	Other	0	0	0	0
	Missing	0	0	0	0

TABLE 26 - Baseline PROMs (comorbidities) of the all operated who consented to PROMs with available 3-year PROMs outcome data sample by operation type

		All operated PROMs N = 85	LAGB N = 13	RYGB N = 33	SG N = 39
Comorbidities	Diabetes N (%)	43 (50.6%)	5 (38.5%)	22 (66.7%)	16 (41.0%)
	Missing	0	0	0	0
	N (%) self-reported comorbidities:
	None	4 (4.7%)	0 (0.0%)	1 (3.0%)	3 (7.7%)
	1–2	32 (37.6%)	8 (61.5%)	12 (36.4%)	12 (30.8%)
	≥3	49 (57.6%)	5 (38.5%)	20 (60.6%)	24 (61.5%)
	Missing	0
Depression	Median (Q1; Q3) preop PHQ-9 score	8.0 (4.0; 12.0)	8.0 (5.0; 11.0)	8.0 (4.0; 12.0)	9.0 (4.0; 13.0)
	N (%) PHQ-9 score ≥ 10 (moderate to severe depression)	34 (40.1%)	5 (38.5%)	11 (33.3%)	18 (46.2%)
	Missing	0	0	0	0
Anxiety	Median (Q1; Q3) preop GAD-7 score	4.0 (1.0; 10.0)	6.5 (3.5; 11.0)	4.0 (1.0; 10.0)	4.0 (1.0; 10.0)
	N (%) GAD-7 score ≥ 6	35 (41.7%)	8 (66.6%)	10 (30.3%)	17 (43.6%)
	Missing	1	1	0	0
QoL	Preoperative mean (SD) SF-12 PCS score	40.27 (12.27)	45.40 (11.04)	38.81 (12.11)	39.61 (12.64)
	Missing	6	0	4	2
	Preoperative mean (SD) SF-12 MCS score	45.53 (9.47)	43.41 (7.12)	47.20 (10.41)	44.96 (9.43)
	Missing	6	0	4	2
	Preoperative median (Q1; Q3) EQ-5D-5L score	0.7 (0.4; 0.8)	0.8 (0.6; 0.9)	0.7 (0.4; 0.8)	0.6 (0.2; 0.8)
	Missing	1	0	1	0
	Preoperative mean (SD) EQ-5D-5L VAS	70.0 (40.0; 80.0)	0.8 (0.6; 0.9)	0.7 (0.4; 0.8)	0.6 (0.2; 0.8)
	Missing	1	0	1	0
	Preoperative mean (SD) IWQOL-Lite Physical Function score (standardised scoring)	48.28 (26.26)	48.78 (28.33)	48.42 (27.37)	47.09 (24.95)
	Missing	0	0	0	0
	Preoperative mean (SD) IWQOL-Lite Total score (standardised scoring)	45.50 (22.49)	43.98 (23.24)	48.22 (22.47)	42.99 (22.39)
	Missing	0	0	0	0

TABLE 27 - Baseline PROMs (health behaviours) of the all operated who consented to PROMs with available 3-year PROMs outcome data sample by operation type

		All operated PROMs N = 85	LAGB N = 13	RYGB N = 33	SG N = 39
Life optimism	Preoperative mean (SD) LOT score	13.24 (4.54)	14.38 (4.91)	13.41 (4.96)	12.70 (4.05)
	Missing	3	0	1	2
Smoking status	Current n (%)	13.24 (4.54)	14.38 (4.91)	13.41 (4.96)	12.70 (4.05)
	Former n (%)	13.24 (4.54)	14.38 (4.91)	13.41 (4.96)	12.70 (4.05)
	Never n (%)	13.24 (4.54)	14.38 (4.91)	13.41 (4.96)	12.70 (4.05)
	Missing n (%)	0	0	0	0
Alcohol use	Preoperative median (Q1; Q3) AUDIT score	3.0 (1.0; 5.0)	4.0 (1.0; 4.0)	2.0 (1.0; 5.0)	3.0 (1.0; 5.5)
	Missing	6	0	3	3
Physical activity	N (%) ≥1 walking, moderate or vigorous activity in last 7 days	79 (92.9%)	13 (100.0%)	29 (87.9%)	37 (94.9%)
	Missing	0	0	0	0
	Preoperative median (Q1; Q3) IPAQ score (MET minutes/week)	1039.5 (462.0; 2479.5)	1836.0 (462.0; 2622.0)	1381.5 (462.0; 2820.0)	808.5 (346.5; 1947.5)
	Missing	28	2	11	15

FIGURE 5.

Follow-up to 3 years post initial bariatric surgery.