Development of predictive models for successful weight loss in people living with obesity

Alabdullah, Lulwa A.A.D. (2023) Development of predictive models for successful weight loss in people living with obesity. PhD thesis, University of Glasgow.

Abstract

Behavioural weight management programmes are effective in helping some, but not all, patients living with obesity to lose weight. With the emerging pharmacological options for weight loss, and a range of dietary interventions available, it may be advantageous to be able to predict successful short and medium term weight loss, so patients can be moved to other therapies at an earlier stage while they are still engaged. The overall aim of this thesis was therefore to identify factors at the baseline level to predict short and medium terms successful weight loss.

In chapter 2, a hypothesis-driven questionnaire was developed after conducting a literature review to test baseline behavioural and psychological factors ability to predict successful short-term weight loss in individuals undertaking behavioural weight management programmes (in NHS GG&C region). This prospective study was not started due to the COVID-19 pandemic, but the questionnaire and study are ready to be implemented in the future.

In chapter 3, a prospective study investigated baseline clinical, sociodemographic and process factors association with weight loss (>5%) in individuals undertaking behavioural weight management programmes (in NHS GG&C region). The only variable that predicted short (16 weeks) and medium-term successful weight loss (3 years) is the early weight loss (4 weeks) in the programme. Weight loss of 0.5% at 4 weeks had sensitivity 90.4%, specificity 53.6%, PPV 32.9%, NPV 95.7% in the short term and sensitivity 89.9%, specificity 49.5%, PPV 19.6%, NPV 97.3% in the medium term.

In chapter 4, a prospective study tested behavioural and psychological factors ability to predict successful medium-term weight loss in using data from the LookAHEAD trial. Moderate predictive utility was obtained from; age, sex, randomised treatment , baseline weight, bodily pain score, diabetes medication and LDL cholesterol (AUC-ROC= 0.649)

In chapter 5, an external validation study was conducted (using the WRAP trial ) to validate predictors of successful weight loss identified in Chapters 3 and 4. Strong evidence was seen that early weight loss in the programme is a strong predictor of medium-term successful weight loss (consistent with chapter 3). In contrast, only baseline weight and age were validated as predictors of successful weight loss.

In chapter 6, UK Biobank was used to test predictors of weight loss over the medium term in a general population with overweight and obesity. A large proportion(19.7%) of people with overweight or obesity lost a significant amount of weight (≥ 5%) over ~4 years even without known dietary interventions. Moderate predictive utility was obtained from sex, age, initial BMI, diastolic blood pressure, triglycerides, and time spent driving (AUC-ROC= 0.618).

In conclusion, socio-demographic, clinical, process, behavioural and psychological variables do not yield sufficient discrimination to allow prediction of successful weight loss, either in a structured weight management programme, or in the general population. However, early weight loss in the first few weeks of starting an intervention is strongly associated with short and medium term successful weight loss. A threshold of failing to achieve 0.5% body weight loss in the first 4 weeks can identify participants who are unlikely to succeed in the programme (>95% of these will not be successful completers). This approach may allow early identification of patients who might benefit more from other interventions while they are still engaged.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Subjects: R Medicine > RA Public aspects of medicine > RA0421 Public health. Hygiene. Preventive Medicine
• Colleges/Schools: College of Medical Veterinary and Life Sciences > School of Cardiovascular & Metabolic Health
• Supervisor's Name: Welsh, Dr. Paul and Logue, Professor Jennifer
• Date of Award: 2023
• Depositing User: Theses Team
• Unique ID: glathesis:2023-83679
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 27 Jun 2023 08:31
• Last Modified: 27 Jun 2023 08:31
• Thesis DOI: 10.5525/gla.thesis.83679
• URI: https://theses.gla.ac.uk/id/eprint/83679

Related URLs

• Article DOI