Abstract
Introduction
There is a demand for publicly funded specialist obesity services in Australia. A range of factors can impact on patient attendance which can result in poorer health outcomes.
Objective
To identify patient factors that predict ongoing in‐person attendance following initial medical assessment at the Canberra Obesity Management Service.
Methods
Data were collated from two retrospective reviews (July 2016–June 2017 and July 2018–June 2019). Predictive modeling was used to determine the likelihood of ongoing attendance.
Results
A total of 396 patients were identified. Mean age was 45.4 years (SD 13.1), mean weight was 139.5 kg (SD 27.8) and mean Body Mass Index was 49.87 kg/m2 (SD 8.7). Demographics and anthropometrics were not predictive of ongoing attendance. Patients with a higher comorbidity burden were more likely to continue attending (p < 0.001). Patients with obstructive sleep apnea (OSA) were 4.9 times more likely to continue attending (p < 0.001). Hypertension was more common among patients who continued attending (p = 0.005); however, this relationship was no longer significant when using a multi‐adjusted model. Comorbid depression and/or anxiety diagnoses were not predictive of ongoing attendance although the p‐value for anxiety severity classification approached significance.
Conclusions
Findings are consistent with existing evidence linking OSA and attendance at specialist obesity services. Hypertension was predictive of ongoing attendance and warrants further research. Determining if anxiety is a true barrier to attendance at specialist obesity services may have implications in terms of optimizing diagnosis and treatment prior to referral or in the early stages of obesity management.
Keywords: attendance, comorbidities, completion, engagement, management, obesity
1. INTRODUCTION
In Australia, rates of obesity have increased from 19% in 1995 to 31% in 2017–2018. 1 , 2 , 3 , 4 The causes of obesity are multifactorial and involve a complex interplay of genetics, environment, psychosocial and cultural factors. Obesity is linked to poor health outcomes, including reduced life expectancy, higher chronic disease burden and increased healthcare expenditure. Patients living with obesity are subject to stigma and bias from the general population, health professionals and themselves. 1 , 2 , 4 , 5 This has traditionally placed the burden of obesity management on the individual. This chronic disease requires accessible and appropriate treatment programs that can benefit both individuals and the wider community.
While there is an established evidence base for the benefits of specialist obesity services in Australia, 6 , 7 , 8 access and engagement can be limited by a range of factors. 9 , 10 , 11 , 12 , 13 , 14 These factors are important to understand given the high demand for specialist obesity services and the risks associated with delaying obesity care, both from an individual and broader public health perspective.
There is substantial variability in published attendance and attrition rates for specialist obesity services. For example, recent Australian studies reported non‐completion or reduced attendance rates of 28% and 36%. 10 , 11 In comparison, a systematic review from the United Kingdom (UK) and Ireland reported drop‐out rates ranging from 13% to 89%. 6 This variability is complicated by differences in service provision and how attendance and attrition are defined or measured.
Some factors that predict attendance at specialist obesity services are already well known. These include patient demographics such as age and locality, and clinical characteristics such as increased comorbidity burden. 6 , 10 However, the impact of mental illnesses such as depression and anxiety appears to be less clear, with some studies reporting no difference 10 , 12 and others suggesting that these diagnoses can increase either attendance or attrition.
This study aimed to determine if certain patient demographics or characteristics predicted ongoing in‐person attendance at a Canberra‐based specialist obesity service. The findings may add to the existing knowledge base or help to identify new factors that lead to early disengagement so that service design can be better targeted toward increasing patient attendance. The findings may also provide further insight into addressing attrition in a highly sought after but scarce resource, which could be applicable to other specialist obesity services.
2. METHODS
2.1. Setting
The Canberra Obesity Management Service (COMS) is a publicly funded, multidisciplinary in‐person program for patients with class III obesity (Body Mass Index [BMI] ≥40 kg/m2) and at least one medical comorbidity. 15 The COMS model of care is in keeping with an expert Australian consensus statement on how to optimize obesity management. 16
The program is medically led with case management, group education and individualized allied health support. There is an initial focus on lifestyle optimization prior to considering intensive interventions such as pharmacotherapy and very low energy diets. There is also limited access to publicly funded sleeve gastrectomy. 15 The COMS offers in‐house sleep and psychiatry clinics (facilitated by a Sleep Physician and Psychiatrist, respectively) for select patients as clinically indicated.
2.2. Study design
This retrospective study used predictive modeling to identify factors that may have impacted on the likelihood of ongoing attendance following initial medical assessment at the COMS.
Ongoing attendance was defined as patient attendance at any COMS appointments, including group education, individual allied health support, or medical reviews following their initial medical assessment. Non‐ongoing attendance was defined as a patient having no further engagement with COMS following their initial medical assessment.
2.3. Participants
Existing data from two previous COMS projects were included in the study (project A from July 2016 to June 2017 and project B from July 2018 to June 2019). The original selection criteria for both projects comprised all newly referred patients who underwent an initial COMS medical assessment. Exclusion criteria were applied to existing COMS patients and newly referred patients who did not attend their initial medical assessment.
2.4. Data collection
Information was sourced from the electronic medical records of 396 patients. Data were collated, de‐identified and entered into Microsoft Excel prior to being exported into a statistical program. Data were securely stored and analyzed by an ACT Health Biostatistician.
Variables of interest comprised baseline patient demographics at the time of initial COMS medical assessment including age (in years), sex at birth (male vs. female) and measured anthropometric data (weight in kg and BMI kg/m2).
Clinical characteristics included the total number of medical comorbidities, as well as the number and type of the top 10 comorbidities (pre‐determined as per previous COMS research). 17 , 18 The top 10 comorbidities were hypertension, depression and/or anxiety, obstructive sleep apnea (OSA), type 2 diabetes mellitus, osteoarthritis, gastro‐esophageal reflux disease, asthma, dyslipidaemia, non‐steatohepatitis, and polycystic ovarian syndrome.
Self‐reported screening tools at the time of initial COMS medical assessment were analyzed and included numeric scores for the Epworth Sleepiness Scale (ESS) and the Depression, Anxiety and Stress Scale (DASS)‐21, as well as severity classifiers of normal, mild, moderate, severe and extremely severe for DASS‐21. 19 , 20
2.5. Statistical analysis
Analysis was performed on a valid data basis and data imputation did not occur. Normality of continuous variables was tested graphically, using histograms and probability‐probability plots. Normally distributed continuous variables are presented as mean (standard deviation) and median (interquartile range) when normality was not met. Categorical variables are presented as frequencies (relative frequencies). Associations between categorical variables were assessed using Pearson's chi‐square test (or Fisher's exact test when needed), and correlation between continuous variables was assessed using Pearson's r correlation coefficient (or Spearman's rho correlation coefficient when needed). Means comparisons between groups were performed using Student's t‐test (or ANOVA when needed) and Mann–Whitney U‐test (or Kruskal–Wallis test) was used when normality was not met. Binary Logistic regression models were used to explore the associations between independent predictors and the likelihood of attendance beyond the initial medical assessment. The analysis was adjusted for age, sex at birth, BMI, personal medical history of hypertension, depression and/or anxiety, OSA, type 2 diabetes mellitus, DASS‐21 numeric scores and symptom severity classification. Statistical analysis was performed in STATA 18.0. Statistical significance was set at alpha = 0.05.
2.6. Ethics
An ethics application was assessed by the ACT Health Research and Ethics Governance Office as a Quality Assurance/Improvement (QAI) activity (2023.LRE.00125). This study was conducted according to the National Health and Medical Research Council (NHMRC) National Statement on Ethical Conduct in Human Research and the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement. 21 , 22
3. RESULTS
A total of 396 patients were identified. Mean age was 45.4 years (SD 13.1) and the majority of patients were female (n = 288, 73%). Mean weight was 139.5 kg (SD 27.8) and mean BMI was 49.87 kg/m2 (SD 8.7). Of the 396 patients, 320 (81%) continued to attend following their initial medical assessment. No age, sex or anthropometric differences were observed based on attendance status (Table 1).
TABLE 1.
Patient demographics, baseline anthropometrics and medical comorbidities according to attendance status (n = 396).
| Total n = 396 (100.0%) | Ongoing attendance n = 320 (80.8%) | Non‐ongoing attendance n = 76 (19.2%) | p‐value | |
|---|---|---|---|---|
| Mean age, years (SD) | 45.4 (13.1) | 45.7 (12.9) | 44.1 (13.9) | 0.331 |
| Sex at birth | 0.938 | |||
| Female, n (%) | 288 (72.7) | 233 (72.8) | 55 (72.4) | |
| Male, n (%) | 108 (27.3) | 87 (27.2) | 21 (27.6) | |
| Mean weight, kg (SD) | 139.7 (27.8) | 139.4 (26.8) | 141.0 (32.5) | 0.669 |
| Mean BMI, kg/m2 (SD) | 49.87 (8.7) | 50.02 (8.3) | 49.24 (10.1) | 0.478 |
| Median comorbidities, n (1st, 3rd quartile) | 3 (2, 4) | 3 (2, 4) | 3 (1,4) | *0.124 |
| Median top 10 comorbidities, n (1st, 3rd quartile) | 3 (1, 4) | 3 (1,4) | 1 (0.3, 3) | *<0.001 |
| Hypertension | 160 (40.4) | 0.005 | ||
| Yes, n (%) | 160 (55.4) | 140 (43.8) | 20 (26.3) | |
| No, n (%) | 129 (44.6) | 180 (56.2) | 56 (73.7) | |
| Depression/anxiety | 109 (48.7) | 0.123 | ||
| Yes, n (%) | 193 (66.6) | 162 (50.6) | 31 (40.8) | |
| No, n (%) | 97 (33.4) | 158 (49.4) | 45 (59.2) | |
| Obstructive sleep apnea | 157 (39.6) | <0.001 | ||
| Yes, n (%) | 157 (56.9) | 143 (44.7) | 14 (18.4) | |
| No, n (%) | 119 (43.1) | 177 (55.3) | 62 (81.6) | |
| Type 2 diabetes mellitus | 78 (19.7) | 0.111 | ||
| Yes, n (%) | 78 (29.9) | 68 (21.3) | 10 (13.2) | |
| No, n (%) | 183 (70.1) | 252 (78.8) | 66 (86.8) | |
| Osteoarthritis | 92 (23.5%) | 0.078 | ||
| Yes, n (%) | 93 (36.9) | 81 (25.3) | 12 (15.8) | |
| No, n (%) | 159 (63.1) | 239 (74.7) | 64 (84.2) | |
| Gastro‐esophageal reflux disease | 75 (18.9) | 0.269 | ||
| Yes, n (%) | 75 (28.7) | 64 (20.0) | 11 (14.5) | |
| No, n (%) | 186 (71.3) | 256 (80.0) | 65 (85.5) | |
| Asthma | 86 (21.7) | 0.163 | ||
| Yes, n (%) | 86 (32.7) | 74 (23.1) | 12 (15.8) | |
| No, n (%) | 177 (67.3) | 246 (76.9) | 64 (84.2) | |
| Dyslipidaemia | 92 (23.2) | 0.087 | ||
| Yes, n (%) | 92 (35.2) | 80 (25.0) | 12 (15.8) | |
| No, n (%) | 169 (64.8) | 240 (75.0) | 64 (84.2) | |
| Non‐alcoholic steatohepatitis | 48 (12.1) | 0.636 | ||
| Yes, n (%) | 48 (18.9) | 40 (12.5) | 8 (10.5) | |
| No, n (%) | 206 (81.1) | 280 (87.5) | 68 (89.5) | |
| Polycystic ovarian syndrome | 36 (12.5) | 0.955 | ||
| Yes, n (%) | 44 (17.2) | 29 (12.4) | 7 (12.7) | |
| No, n (%) | 212 (82.8) | 204 (87.6) | 48 (87.3) |
Abbreviation: SD, standard deviation.
Note: all percentages are valid; *Mann–Whitney U test; polycystic ovarian syndrome data for only women.
The median number of comorbidities was three (3) (1, 4). Patients who continued to attend had a statistically significant higher number of the 10 most common comorbidities compared with those who disengaged (p < 0.001). Comorbid OSA and hypertension were more common among patients who continued attending COMS after their initial medical assessment. These findings were statistically significant (p < 0.001 and p = 0.005, respectively). See Table 1. However, hypertension as a predictor of attendance lost its significance when a multi‐adjusted model was used (Table 2).
TABLE 2.
Multivariable binary logistic regression model predicting the likelihood of continuing with COMS following initial medical assessment (N = 396).
| Characteristic | Odds Ratio (95% CI) | p‐value |
|---|---|---|
| Age (per 1 year) | 0.984 (0.956, 1.013) | 0.281 |
| Sex (male vs. female) | 0.592 (0.267, 1.312) | 0.197 |
| BMI (per 1 kg/m2) | 0.979 (0.944, 1.016) | 0.259 |
| Diagnosed osteoarthritis (yes vs. no) | 1.519 (0.606, 3.810) | 0.372 |
| Diagnosed obstructive sleep apnea (yes vs. no) | 4.931 (2.008, 12.112) | <0.001 |
| Diagnosed type 2 diabetes mellitus (yes vs. no) | 1.266 (0.471, 3.407) | 0.640 |
| Diagnosed hypertension (yes vs. no) | 1.368 (0.630, 2.972) | 0.429 |
| Depression severity | 0.478 | |
| Mild versus normal | 0.450 (0.167, 1.210) | 0.114 |
| Moderate versus normal | 0.556 (0.221, 1.399) | 0.212 |
| Severe versus normal | 0.557 (0.120, 2.580) | 0.454 |
| Extremely versus normal | 1.010 (0.184, 5.527) | 0.991 |
| Anxiety severity | 0.058 | |
| Mild versus normal | 1.948 (0.581, 6.533) | 0.280 |
| Moderate versus normal | 0.403 (0.149, 1.090) | 0.073 |
| Severe versus normal | 0.293 (0.090, 0.953) | 0.041 |
| Extremely versus normal | 0.555 (0.132, 2.338) | 0.422 |
| Stress severity | 0.578 | |
| Mild versus NORMAL | 2.165 (0.606, 7.736) | 0.235 |
| Moderate versus normal | 2.254 (0.673, 7.551) | 0.188 |
| Severe versus normal | 2.679 (0.569, 12.609) | 0.212 |
| Extremely versus normal | 1.810 (0.224, 14.631) | 0.578 |
There was no difference in ESS or DASS‐21 numeric sub‐scale scores between patients who continued to attend COMS and those who did not. No difference was observed in DASS‐21 depression and stress classifiers; however, the p‐value approached significance (0.057) for the anxiety severity classification (Table 3).
TABLE 3.
Self‐reported screening tool scores at initial COMS medical assessment according to attendance status (n = 396).
| Total n = 396 (100.0%) | Ongoing attendance n = 320 (80.8%) | Non‐ongoing attendance n = 76 (19.2%) | p‐value | |
|---|---|---|---|---|
| DASS‐21 depression score, median (1st, 3rd quartile) | 6 (3, 10) | 6 (3, 10) | 6 (3, 8) | *0.842 |
| DASS‐21 depression severity | +0.461 | |||
| Normal, n (%) | 133 (38.8) | 117 (39.5) | 16 (34.0) | |
| Mild, n (%) | 54 (15.7) | 45 (15.2) | 9 (19.1) | |
| Moderate, n (%) | 84 (24.5) | 69 (23.3) | 15 (31.9) | |
| Severe, n (%) | 29 (8.5) | 25 (8.4) | 4 (8.5) | |
| Extremely severe, n (%) | 43 (12.5) | 40 (13.5) | 3 (6.4) | |
| DASS‐21 anxiety score, median (1st, 3rd quartile) | 5 (2, 9) | 5 (2, 9) | 6 (3, 9) | *0.531 |
| DASS‐21 anxiety severity, n (%) | +0.057 | |||
| Normal, n (%) | 126 (37.0) | 109 (36.9) | 17 (37.0) | |
| Mild, n (%) | 62 (18.2) | 58 (19.7) | 4 (8.7) | |
| Moderate, n (%) | 49 (14.4) | 39 (13.2) | 10 (21.7) | |
| Severe, (%) | 38 (11.1) | 29 (9.8) | 9 (19.6) | |
| Extremely severe, (%) | 66 (19.4) | 60 (20.3) | 6 (13.0) | |
| DASS‐21 stress score, median (1st, 3rd quartile) | 7 (4, 2) | 7 (4, 11) | 6 (2.5, 11) | *0.303 |
| DASS‐21 stress severity | +0.890 | |||
| Normal, n (%) | 191 (56.0) | 162 (54.9) | 29 (63.0) | |
| Mild, n (%) | 34 (10.0) | 30 (10.2) | 4 (8.7) | |
| Moderate, n (%) | 57 (16.7) | 50 (16.9) | 7 (15.2) | |
| Severe, n (%) | 44 (12.9) | 40 (13.6) | 4 (8.7) | |
| Extremely severe, n (%) | 15 (4.4) | 13 (4.4) | 2 (4.3) | |
| ESS, median (1st, 3rd quartile) | 6 (3, 11) | 7 (4, 11) | 5 (3, 9) | *0.316 |
Note: DASS‐21 scores; *Mann–Whitney U test; +chi square.
Abbreviations: DASS‐21, depression, anxiety and stress scale (21) scores; ESS, Epworth sleepiness scale.
Patients with an existing diagnosis of OSA were 4.9 times more likely to continue attending the COMS following their initial medical assessment. This was irrespective of other patient demographics or clinical characteristics (Table 2).
4. DISCUSSION
Rates of non‐attendance following initial medical assessment at COMS (19%) were less than the non‐completion rates observed in a case‐control study of a Sydney‐based metabolic rehabilitation service for severe obesity (36%). 10 This difference could be due to COMS capturing disengagement after first contact compared to the Sydney‐based study, which reported drop‐out rates over a 12‐month period. 10
An observational cohort study reported on attendance at a metabolic obesity center. The Adelaide‐based study noted that 28% of patients attended half or less than half of their scheduled follow‐up appointments. 11 While this study provides insight into their service, direct comparisons are limited due to differences in COMS flow of service and attendance definitions. This highlights the importance of consistent care models and data collection across specialist obesity sites for meaningful comparison. 16 , 23
Despite not being able to directly compare attendance rates with other services due to differences in service provision, attrition measures and length of program, this study found similarities with other specialist obesity services in terms of how demographics, anthropometrics and comorbidities impact on attendance.
Similar to other Australian studies, patient factors such as sex, weight and BMI were not found to be predictive of ongoing attendance at COMS. 10 , 11 While no relationship was found between age and attendance at COMS, younger age has been identified in numerous studies as a predictor of non‐completion of specialist obesity programs. 10 , 24 , 25 , 26
Interestingly, the mean age of the COMS cohort was lower than that observed in these other studies. 10 , 24 , 25 , 26 This is despite COMS BMI eligibility criteria and mean BMI being higher than many other services and studies, and that obesity has historically been more prevalent in older age groups. This likely reflects the growing rates of obesity among all ages, including younger people. 1 , 4
The COMS comorbidity data were collected from referral letters and patient disclosure at the time of initial medical assessment. Due to prolonged wait times to access the COMS, additional comorbidities may not have been formally diagnosed, leading to potential for under‐reporting in this cohort. While active screening for obesity‐related comorbidity does occur at the initial COMS medical assessment, 17 , 18 subsequent diagnostic rates were not the focus of this study.
A recent systematic review and meta‐analysis revealed that obesity assessments and weight management discussions were more likely to occur in a primary care setting if a patient had obesity‐related comorbidities. 9 It is possible that an awareness of obesity and its complications has led to patients with a higher comorbidity burden prioritizing their health and attending the COMS.
This study revealed that patients were more likely to continue with COMS if they had a higher number of the 10 most common comorbidities (p < 0.001). While not all studies have specifically looked at the cumulative number of comorbidities, they have examined the impact of individual comorbidities on attendance and completion at specialist obesity services. Our findings are similar to other studies, which demonstrate earlier drop‐out rates in patients with less obesity‐related comorbidities. 27
The strongest predictor of ongoing attendance was an existing diagnosis of OSA, with patients being 4.9 times more likely to continue beyond the initial medical assessment (p < 0.001). This supports findings from a UK and Ireland‐based systemic review, 6 and an Australian metabolic rehabilitation service, which showed that a diagnosis of OSA was more likely to result in completion of their program. 10 It is possible that patients with an existing OSA diagnosis have increased awareness of the risks of OSA and the benefits of weight loss, including tangible symptomatic improvement.
It is also possible that enhanced attendance at the COMS among patients with OSA reflects the long wait list for the local public Sleep Medicine service. The fact that the COMS has had at least one Sleep Physician permanently on staff since 2015 may have led referrers and patients alike to be cognisant of the benefits of having both obesity and OSA managed in the same setting.
The presence of comorbid hypertension was demonstrated to increase the chance of ongoing patient attendance at COMS following the initial medical assessment (p = 0.005). This is unexpected given that other local research has not found a similar link, 10 , 11 and a large Canadian study from 2016 noted that male patients with hypertension were more likely to disengage early from an obesity program. 25 It is worth noting; however, that hypertension failed to remain a significant predictor of attendance in the multi‐adjusted model and that other predictors likely had a confounding effect.
This study found that comorbid depression and/or anxiety was not a predictor of ongoing attendance at COMS. While other local studies have not found a link between depression or anxiety and attendance, 10 , 11 one international study suggested that a diagnosis of depression can lead to early attrition among females. 25 In contrast, Inelman et al. (2005) noted that depression was more common among patients who completed a dietary intervention for obesity. 27
The DASS‐21 numeric scores and severity classifications did not predict the likelihood of ongoing attendance at COMS. Although there has been more research on depressive symptom severity predicting non‐completion of weight management programs, 24 , 26 , 27 , 28 , 29 there is growing evidence linking higher self‐reported anxiety scores at baseline with early non‐attendance and non‐completion. 24 , 28 , 30
It is possible that the DASS‐21 anxiety severity classification in this study (regardless of a formal diagnosis of anxiety) may predict ongoing attendance; however, the p‐value was marginal. While further investigation is warranted to determine if a true difference exists, there may be more immediate scope to optimize diagnosis and management of anxiety among COMS patients as part of the in‐house Psychiatry clinic. This could lead to improved attendance and overall health outcomes.
Atlantis et al. (2019) reported that Continuous Positive Airway Pressure (CPAP) use was associated with an increased likelihood of completing a metabolic obesity program. 10 Although this COMS study did not analyze CPAP use, it did consider subjective sleepiness as measured on the ESS. 21 The absence of a difference in ESS scores in this study could suggest that patients who continued to attend COMS had good OSA symptom control, resulting in less daytime sleepiness and improved attendance.
4.1. Strengths and limitations
While the COMS has previously collected qualitative data on the reasons for non‐attendance, 17 , 18 this is the first attempt to perform predictive modeling on certain patient demographics and clinical characteristics. This study adds to the knowledge of factors that may impact on attendance at specialist obesity services.
While the retrospective nature of this study is a limitation, it does serve as a consecutive series of adult patients attending COMS and therefore has high general applicability to similar publicly funded specialist obesity services. Participant numbers were sufficient to enable meaningful statistical analysis and to result in statistically significant findings.
Combining data from two projects at different time points could be considered a limitation. However, identical inclusion criteria were used and no exclusions were applied to those who attended their initial medical assessment. Variables of interest were the same for both projects and COMS service provision remained unchanged over the course of each project. Furthermore, both projects occurred prior to the onset of COVID‐19, removing the impact of lockdowns on attendance.
The grouping of depression and/or anxiety as one comorbidity prevented the ability to determine whether these diagnoses could individually predict the likelihood of ongoing attendance at COMS. Given that depression and anxiety have been noted to impact on attendance in some international studies, further research is warranted at a local level with a focus on anxiety symptom severity.
In keeping with local expert guidance, 16 this study is part of an effort to promote research capacity within the COMS and to enable inter‐service comparisons. Although collected several years ago, much of the data used in this study is in line with that identified in a Delphi study as standard baseline data to collect in Australian specialist obesity services. 23 Moving forward, the COMS will strive to collect data on additional factors that could impact attendance, including socioeconomic, geographic, educational and culturally diverse profiles of patients.
5. CONCLUSION
This study provided insight into patient demographics and clinical factors that predict ongoing attendance at COMS. The findings were consistent with the growing evidence base and support the link between a diagnosis of OSA and attendance at specialist obesity services. To the authors' knowledge, comorbid hypertension as a predictor of ongoing attendance is not well established and warrants further research. Similarly, determining if anxiety is a true barrier to attendance at specialist obesity services may have implications in terms of earlier diagnosis and treatment, which could optimize attendance and management of obesity and its related comorbidities.
CONFLICT OF INTEREST STATEMENT
All authors have completed the ICMJE form for disclosure of potential conflicts of interest. Dr Hsin‐Chia Carol Huang has received a major research grant from the CHS Private Practice Fund and ESSA for research into the Long‐Covid Rehabilitation Service. Dr Hsin‐Chia Carol Huang also received an honoraria payment from Astra Zenaca for chairing the Canberra Peer to Peer Meetings for Respiratory and Sleep Medicine.
ACKNOWLEDGMENTS
The authors would like to acknowledge Dr Chiaki Kojima, Dr Ryan Burns, Dr Elise Firman and A/Prof. Paul Dugdale for contributing to previous research efforts. Thank you to Dr Deborah Inman and Dr Ashvini Munindradasa for commenting on the manuscript. The authors would also like to thank all COMS staff and patients.
Brightman L, Huang H‐CC, Georgousopoulou E. Predictors of ongoing attendance at an Australian publicly funded specialist obesity service. Obes Sci Pract. 2024;e70010. 10.1002/osp4.70010
REFERENCES
- 1. World Health Organisation . Obesity and Overweight: Factsheet. World Health Organisation; 2021. [Google Scholar]
- 2. Bluher M. Obesity: global epidemiology and pathogenesis. Nat Rev Endocrinol. 2019;15(5):288‐298. 10.1038/s41574-019-0176-8 [DOI] [PubMed] [Google Scholar]
- 3. World Obesity . Clinical Care for Obesity. World Obesity Federation; 2021. [Google Scholar]
- 4. Australian Institute of Health and Welfare . Australia’s Health: Overweight and Obesity. Australian Government; 2022. [Google Scholar]
- 5. Buchmueller T, Johar M. Obesity and health expenditures: evidence from Australia. Econ Hum Biol. 2015;17:42‐58. 10.1016/j.ehb.2015.01.001 [DOI] [PubMed] [Google Scholar]
- 6. Brown T, O’Malley C, Blackshaw J, et al. Exploring the evidence base for tier 3 weight management interventions for adults: a systematic review. Clinical Obesity. 2017;7(5):260‐272. 10.1111/cob.12204 [DOI] [PubMed] [Google Scholar]
- 7. Khera R, Murad M, Chandar A, et al. Association with pharmacologist treatments for obesity with weight loss and adverse events: a systematic review and meta‐analysis. J Am Med Assoc. 2016;315(22):2424‐2434. 10.1001/jama.2016.7602 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Williams K, Maston G, Schneuer F, Nassar N. Impact of specialised obesity management services on the reduction in the use of acute hospital services. Clinical Obesity. 2023;13(5). 10.1111/cob.12592 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Atlantis E, Chimoriya R, Seifu C, et al. Enablers and barriers to implementing obesity assessments in clinical practice: a rapid mixed‐methods systematic review. British Medical Journal Open. 2022;12(11):e063659. 10.1136/bmjopen-2022-063659 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Atlantis E, Lin F, Fahey P, Kormas N. A predictive model for non‐completion of an intensive specialist obesity service in a public hospital: a case‐control study. Biomedical Central Health Services Research. 2019;19(1):748. 10.1186/s12913-019-4531-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Brook E, Cohen L, Hakendorf P, Wittert G, Thompson C. Predictors of attendance at an obesity clinic and subsequent weight change. Biomedical Central Health Services Research. 2014;14(1):78. 10.1186/1472-6963-14-78 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Brightman L, Dunne A, Huang H. Mental illness as a predictor of patient attendance and anthropometric changes: observations from an Australian publicly funded obesity management service. Australas Psychiatr. 2021;29(3):266‐271. 10.1177/1039856220953713 [DOI] [PubMed] [Google Scholar]
- 13. National Health and Medical Research Council . Clinical Practice Guidelines for the Management of Overweight and Obesity in Adults, Adolescents and Children in Australia. National Health and Medical Research Council; 2013. [Google Scholar]
- 14. Rigas G, Williams K, Sumithran P, Brown WA, Caterson ID. Barriers to progression through Australian obesity management pathways. Australian Journal of General Practice. 2023;52(7):472‐480. 10.31128/ajgp-07-22-6501 [DOI] [PubMed] [Google Scholar]
- 15. Canberra Health Services . Obesity Management Service. ACT Government; 2023. [Google Scholar]
- 16. Atlantis E, Kormas N, Samaras K, et al. Clinical Obesity Services in Public Hospitals in Australia: a position statement based on expert consensus. Clinical Obesity. 2018;13(3):302‐303. 10.1016/j.orcp.2018.11.190 [DOI] [PubMed] [Google Scholar]
- 17. Brightman L, Huang HC.‐C, Dugdale P. Determining patient attendance, access to interventions and clinical outcomes in a publicly‐funded obesity program: results from the Canberra Obesity Management Service. Clinical Obesity. 2019;9(4):e12345. 10.1111/cob.12325 [DOI] [PubMed] [Google Scholar]
- 18. Burns R, Firman E, Huang H.‐CC. Assessing service provision and outcomes at the Canberra Obesity Management Service: a retrospective chart review. Obesity. 2022;30(11):2146‐2155. 10.1002/oby.23575 [DOI] [PubMed] [Google Scholar]
- 19. Johns M. A new method for measuring daytime sleepiness: the Epworth Sleepiness Scale. Sleep. 1991;14(6):540‐545. 10.1093/sleep/14.6.540 [DOI] [PubMed] [Google Scholar]
- 20. Lovibond S, Lovibond P. Depression Anxiety Stress Scales (DASS‐21, DASS‐42). American Psychological Association; 1995. [Google Scholar]
- 21. National Health and Medical Research Council . The National Statement on Ethical Conduct in Human Research (2007 and as Updated). Australian Government; 2018. [Google Scholar]
- 22. von Elm E, Altman D, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP. STROBE initiative. Strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. J Clin Epidemiol. 2008;61(4):344‐349. 10.1016/j.jclinepi.2007.11.008 [DOI] [PubMed] [Google Scholar]
- 23. Ramachandran D, Atlantis E, Markovic T, Hocking S, Gill T. Standard baseline data collections in obesity management clinics: a Delphi study with recommendations from an expert panel. Clinical Obesity. 2019;9(3):e12301. 10.1111/cob.12301 [DOI] [PubMed] [Google Scholar]
- 24. Finucane F, Gibson I, Hughes R, et al. Factors associated with weight loss and health gains in a structured lifestyle modification program for adults with severe obesity: a prospective cohort study. Front Endocrinol. 2023;14. 10.3389/fendo.2023.1257061 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Jiandani D, Wharton S, Rotondi M, Ardern CI, Kuk JL. Predictors of early attrition and successful weight loss in patients attending an obesity management program. Biomedical Central Obesity. 2016;3(1):14. 10.1186/s40608-016-0098-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Rothberg A, McEwen L, Kraftson A, et al. Factors associated with participant retention in a clinical, intensive, behavioural weight management program. Biomedical Central Obesity. 2015;2(1):11. 10.1186/s40608-015-0041-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Inelman E, Toffanello E, Enzi G, et al. Predictors of drop‐out in overweight and obese outpatients. Int J Obes. 2005;29(1):122‐128. 10.1038/sj.ijo.0802846 [DOI] [PubMed] [Google Scholar]
- 28. Jones R, Mueller J, Sharp S, et al. The impact of participant mental health on attendance and engagement in a trial of behavioural weight management programs: secondary analysis of the WRAP randomised controlled trial. Int J Behav Nutr Phys Activ. 2021;18(1):146. 10.1186/s12966-021-01216-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Shell A, Hsueh L, Vrany E, et al. Depressive symptom severity as a predictor of attendance in the HOME behavioural weight loss trial. J Psychosom Res. 2020;131:109970. 10.1016/j.jpsychores.2020.109970 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. McLean R, Morrison D, Shearer R, Boyle S, Logue J. Attrition and weight loss outcomes for patients with complex obesity, anxiety and depression attending a weight management program with targeted psychological treatment. World Obesity. 2016;6(2):133‐142. 10.1111/cob.12136 [DOI] [PubMed] [Google Scholar]