Population estimates and characteristics of Australians potentially eligible for bariatric surgery: findings from the 2011–13 Australian Health Survey

Bariatric surgery is more effective than conservative interventions to treat resistant obesity and is considered cost-effective.^1–3 Generally, bariatric surgery is recommended for those with resistant Class 3 obesity (body mass index (BMI) ≥40 kg m^–2) or resistant Class 2 obesity (BMI 35–39.9 kg m^–2) and obesity-related comorbidities.^2,4,5 In recent national guidelines for obesity management,^2,5 bariatric surgery has also been recommended for consideration for those with resistant Class 1 obesity (BMI 30–34.9 kg m^–2) and type 2 diabetes mellitus (T2DM). This is because of accumulating evidence that metabolic health improves after surgery.^3,6

As in many other countries, significant numbers of Australians are overweight or obese. Four million adult Australians, or 27.2% of the adult population, were estimated to be obese in 2011–12, up from 19.1% in 1995.⁷ Although obesity is more prevalent in areas of socioeconomic disadvantage⁷ and surgical outcomes appear comparable by funding type,^8–10 >90% of bariatric surgery in Australia is privately funded (16 650 primary privately funded procedures were performed in 2015¹¹), a funding pattern that appears similar to that in other countries, such as Mexico and the United Arab Emirates.¹² Not all Australian jurisdictions provide publicly funded bariatric surgery and where it is available the waiting period can be prolonged.^13–15 Of additional concern is that the wait for bariatric surgery may be associated with declining health.¹⁶ The extent to which supply is falling short of potential demand in Australia and in many countries is unknown.¹⁷ In one Canadian study, 23% of patients in a publicly funded weight management program and who were deemed eligible for bariatric surgery expressed interest in pursuing a surgical pathway.¹⁸ Many individual, social and environmental factors (e.g. a patient’s health status, recommendations made by health professionals, exposure to other recipients of bariatric surgery) can influence a preference for surgery.¹⁹

The objectives of the present study were to use national population survey data from the 2011–13 Australian Health Survey (AHS)²⁰ to: (1) estimate the number of Australians potentially eligible for bariatric surgery; (2) describe their demographic characteristics, health status and health service use; and (3) estimate the potential demand for surgery in the public and private health systems.

Data were extracted from the cross-sectional 2011–13 AHS conducted by the Australian Bureau of Statistics (n = 31 837).²⁰ The survey used a stratified multistage area sample of private dwellings to ensure a nationally representative sample. The AHS comprised two main surveys: the National Health Survey (NHS) and the National Nutrition and Physical Activity Survey (NNPAS). Participants completed only one survey. Common to both surveys was a core component that included questions and measures of sociodemographic characteristics (e.g. sex, age, geographical location) and physical and health characteristics (e.g. measured height, weight, blood pressure, self-rated health, health conditions, smoking status).

Participants from either the NHS or NNPAS were invited to complete the National Health Measures Survey (NHMS). The NHMS collected blood and urine samples and tested for chronic disease biomarkers, including fasting plasma glucose, blood lipids, albumin, creatinine and alanine aminotransferase (ALT). The sample for the present study was drawn from the NHMS.²⁰ In addition, we conducted a subanalysis using a sample that had completed both the NHMS and NHS, enabling analysis of private health insurance status and health service use. Only those with complete measured height and weight data were included in our analyses. The structure, response rates and sample sizes of the AHS are summarised in Fig. 1. Further details on the AHS can be found in the user’s guide.²¹

Fig. 1.  Design of and response rates for the 2011–13 Australian Health Survey (AHS). AHS participants completed the National Health Survey (NHS) or the National Nutrition and Physical Activity Survey (NNPAS). The AHS core component was common to both surveys. Response proportions represent adequately or fully responding households, except for the National Health Measures Survey (NHMS) where the response proportion reflects the total number of participants relative to total number of participants in the core component of the AHS. The sample used in the present study was drawn from the NHMS. Adapted from the AHS users’ guide.²¹

Participants were classified as potentially eligible for bariatric surgery based on survey data that best approximated the 2013 Australian criteria for considering bariatric surgery,² that is for adults (aged 18–65 years) with resistant Class 3 obesity (BMI ≥40 kg/m²), or Class 2 obesity (BMI 35–39.9 kg/m²) with at least one obesity-related comorbidity (at risk of a cardiovascular (CV) event or mortality, or experiencing hypertension, T2DM, chronic kidney disease, non-alcoholic steatohepatitis (NASH) or gastro-oesophageal reflux disease (GORD)), or Class 1 obesity with poorly controlled T2DM and increased CV risk. A summary of the variables and classification criteria is given in Table 1. Our classifications were limited by the data available within the AHS and did not cover the range of factors considered when making a clinical judgement about eligibility for surgery (e.g. classifying resistant obesity, patient preference). Consequently, we make reference to potential eligibility only. Bariatric surgery may be recommended for those outside of the 18–65 years age range,⁹ but our analysis was based on Australian guidelines only.

Table 1.  Description of variables used to determine potential eligibility for bariatric surgery
ALT, alanine aminotransferase; ACR, albumin : creatinine ratio; AHS, Australian Health Survey; BMI, body mass index; BP, blood pressure; CKD, chronic kidney disease; Class 1 obesity, BMI 30–34.9 kg m^–2; Class 2 obesity, BMI 35–39.9 kg m^–2; Class 3 obesity, BMI ≥40 kg m^–2; CV, cardiovascular; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; HbA1c, Haemoglobin A1c; HDL-C, high-density lipoprotein cholesterol; HF, heart failure; N/A, not applicable; NASH, non-alcoholic steatohepatitis; NHS, National Health Survey; SBP, systolic blood pressure; T1DM, Type 1 diabetes mellitus; T2DM, Type 2 diabetes mellitus

Other variables included in the analysis were: (1) the index of relative socioeconomic disadvantage, which ranks geographical areas of residence according to their social and economic status; (2) remoteness area category based on the location of a participant’s residence and classified as major city, inner regional or outer regional; (3) private health insurance status reported by participants; (4) self-rated health reported by participants as excellent, very good, good, fair or poor; and (5) health service use reported by participants, including consultation with a general practitioner or specialist, having been admitted to hospital as an in-patient or visits to an emergency department or as an out-patient during the previous 2 weeks. Health service use and private health insurance status data were extracted from the NHS, whereas data for the remaining variables were extracted from the core component of the AHS (Fig. 1).

Summary data are presented as the mean for continuous variables and percentages for categorical variables. A weighted Poisson regression model was used to estimate associations with factors influencing health service use, including age, sex, socioeconomic status, remoteness area category and private health insurance. In all analyses, estimates were weighted with sampling weights provided by the ABS within the recommended survey weight module svr (N. Winter, Boston College Department of Economics, Chestnut Hill, MA, USA), and 95% confidence intervals (CIs) derived using replicate weights.²¹

Population estimates were calculated based on a sample of 6804 adults (aged 18–65 years) with complete height and weight data who had completed the NHMS. Of the 3 352 037 Australians aged 18–65 years estimated to be obese, 882 441 (26.3%) were estimated to be potentially eligible for bariatric surgery. This was comprised mostly of those with Class 2 or 3 obesity (Table 2). There was variation between the states and territories: Queensland had the highest percentage of the population potentially eligible for bariatric surgery (7.5%) and Western Australia the lowest (5.1%; Table 3).

Table 2.  Population estimates of adult Australians aged 18–65 years potentially eligible for bariatric surgery by obesity class; findings from the 2011–13 Australian Health Survey (AHS)
Eligibility criteria for bariatric surgery are as per National Health and Medical Research Council of Australia guidelines.² All estimates are for adults aged 18–65 years, with 95% confidence intervals in parentheses. Weights used for population estimates were determined by the Australian Bureau of Statistics at time of the survey in 2011–13.²⁰ The sample number refers to the size of the sample from which the estimates were made. Refer to Fig. 1 for the design of the AHS. Class 1 obesity, body mass index (BMI) 30–34.9 kg m^–2; Class 2 obesity, BMI 35–39.9 kg m^–2; Class 3 obesity, BMI ≥40 kg m^–2

Table 3.  Population estimates of adult Australians aged 18–65 years potentially eligible for bariatric surgery for each Australian jurisdiction; findings from the 2011–13 Australian Health Survey
Eligibility criteria for bariatric surgery are as per National Health and Medical Research Council of Australia guidelines.² All estimates are for adults aged 18–65 years with 95% confidence intervals (CIs) in parentheses. The sample number refers to the size of the sample from which the estimates were made

Table 4 compares the characteristics of those potentially eligible for bariatric surgery with those classified as ineligible for surgery, except for those potentially eligible with Class 1 obesity because of the small sample size (n = 17). Slightly more females were potentially eligible for surgery due to their higher prevalence of Class 3 obesity. Compared with the obese ineligible population, those in the potentially eligible population were more likely to be female, reside outside a major city, be of low socioeconomic position and rate their health as ‘poor’. As a consequence of the selection criteria being dependent on comorbidity, those with Class 2 obesity potentially eligible for surgery had, on average, poorer obesity-related health and were older (by 4.9 years) than those with Class 3 obesity. Hypertension was the most common reason an individual with Class 2 obesity became potentially eligible for bariatric surgery. Potential eligibility for bariatric surgery was associated with more health service use independent of age, sex, remoteness area category, private health insurance and socioeconomic status (Table 5). As expected (due to the selection criteria), being potentially eligible for surgery with Class 2 obesity was associated with more medical appointments in the previous 2 weeks.

Table 4.  Characteristics of adult Australians aged 18–65 years potentially eligible for bariatric surgery; findings from the 2011–13 Australian Health Survey (AHS)
Eligibility criteria for bariatric surgery are as per National Health and Medical Research Council of Australia guidelines.² All estimates are for adults aged 18–65 years and, unless stated otherwise, show mean values with 95% confidence intervals in parentheses. Individuals with missing height or weight data are not included. The sample number refers to the size of the sample from which the estimates were made. Population estimates for Class 1 obesity were not reported because of low sample size. ALT, alanine aminotransferase; BMI, body mass index; CKD, chronic kidney disease; CVD, cardiovascular disease; FPG, fasting plasma glucose; HbA1c, haemoglobin AIc

Table 5.  Risk and relative risk of using health services in the previous 2-week period for adult Australians aged 18–65 years by obesity category and potential eligibility for bariatric surgery; findings from the 2011–13 Australian Health Survey
The outcome ‘medical appointments’ refers to whether a participant had an appointment in the previous 2 weeks with a general practitioner or specialist or at a hospital out-patient facility or day clinic. Hospital visit refers to whether a participant visited hospital as an in-patient or attended an emergency facility. Model 1 was adjusted for age and sex. Model 2 was also adjusted for remoteness area category and socioeconomic and private health insurance status. The sample number refers to the size of the sample from which the estimates were made. RR, relative risk; CI, confidence interval; Class 1 obesity, body mass index (BMI) 30–34.9 kg m^–2; Class 2 obesity, BMI 35–39.9 kg m^–2; Class 3 obesity, BMI ≥40 kg m^–2

Of the total 882 441 Australians estimated to be potentially eligible for bariatric surgery, 45.8% (405 594) were without private health insurance (sample n = 165), of whom 54.5% (95% CI 42.5–66.6) were female. More of those potentially eligible for surgery (78.5%; 95% CI 69.5–87.4) reported that private health insurance was unaffordable than did the ineligible obese population (67.3%; 95% CI 60.0–74.5). Overall, the proportion of females and males potentially eligible for bariatric surgery with private health insurance was similar (48.9% female; 95% CI 38.6–59.1). However, there were differences between the sexes within the Class 2 and 3 obesity categories, in which the proportion of females with insurance was 39.3% (95% CI 25.4–53.2) and 67.3% (95% CI 54.5–80.0) respectively. Of those potentially eligible for bariatric surgery with private health insurance (sample n = 192), 36.7% (95% CI 25.4–47.9) were of low socioeconomic position (at or below Quintile 2 (Q2), the most disadvantaged).

We assessed how sensitive our prevalence estimates were to possible misclassification of comorbidities for those with Class 2 obesity (see Table 1 for comorbidity definitions and their limitations). We considered our definition of NASH (elevated Alanine Aminotransferase (ALT)) to be the most error prone and removing this comorbidity reduced the total number estimated to be potentially eligible for surgery by 107 023 to 775 418. The number of individuals who became potentially eligible for bariatric surgery based solely on other specific criteria was as follows: risk of a CV event or mortality, n = 4274; existing chronic kidney disease, n = 18 550; existing hypertension, n = 125 472; and existing T2DM, n = 16 247. A more conservative HbA1c cut-off point of ≥7.0 mmol/mol to indicate the presence of diabetes, instead of a cut-off point of ≥6.5 mmol/mol (as used in the AHS), reduced the population estimate by 5791.

The findings of the present study provide compelling evidence that the potential need for bariatric surgery in Australia far outweighs availability, especially through the public health system, a situation also seen elsewhere (e.g. Canada^22,23). Even if only 5% of those estimated to be potentially eligible for bariatric surgery in Australia sought this pathway (44 122/882 441), demand would still exceed current capacity. Further, approximately half those Australian adults potentially eligible for bariatric surgery would likely need to access this service through the public health system if they chose to seek this intervention. This finding highlights the immediate need for guidelines on the prioritisation of eligible patients for publicly funded bariatric surgery. In addition, the total number of individuals estimated to be potentially eligible for surgery provides a strong signal that more funding of public surgery and other effective interventions to assist this population group are necessary. Given the limitations in the supply of publicly funded bariatric surgery in Australia, health economic modelling is needed to determine prioritisation for the allocation of this limited resource.

Currently, there is inequitable access to bariatric surgery in Australia favouring those who can access this service through the private health system.^11,24 Further, recipients of bariatric surgery in Australia are more likely to be aged between 35 and 54 years, of middle socioeconomic status, living in a major city and female.¹¹ These characteristics are similar to those collectively identified in a recent systematic review of 12 retrospective cohort studies conducted in the US, UK, Canada and Australia.¹⁷ That review reported that the average proportion of those eligible who received bariatric surgery ranged between 1% and 5%.¹⁷ During the period July 2011 to June 2012, 11586 privately funded bariatric surgeries (excluding revisions and reversals) were performed in Australia (no published data were available for publicly funded bariatric surgery during this period). According to our results, this represents 1.3% of the population potentially eligible for surgery. Australian-based modelling estimated that increasing the provision of bariatric surgery in Australia through public funds by 30% per year over the 10-year period from 2015 to 2025 has the potential to reduce the number of people with obesity by 4400, resulting in a societal saving of A$170 million.²⁵

The finding in the present study that significantly more females had Class 3 obesity (a pattern seen in other countries, such as Canada and the US²⁶) and that females were more likely to have private health insurance within the same obesity class may explain, in part, why more females are having bariatric surgery.^11,12 However, more research is needed to understand sex differences in the uptake of surgery.¹²

We found that 36.7% (175 356/477 847) of those potentially eligible for surgery with private health insurance were of low socioeconomic position (≤Q2). This has potential implications for the public health system because of the relatively common need for reoperative bariatric surgery. A recent systematic review demonstrated that, on average, 2.5–18.4% of bariatric surgery recipients required a reoperation and 13–25.2% required a subsequent reoperation.²⁷ Patients should be encouraged to maintain their health insurance, which may be more challenging for those experiencing socioeconomic disadvantage.

Ensuring equitable access to publicly funded bariatric surgery and determining the optimal number of surgeries to perform and who should get priority is difficult. Health economic modelling is needed to determine who should be prioritised for this limited resource, a process that will be aided by the recently initiated Australian Bariatric Surgery Registry. This registry will fill important knowledge gaps needed to inform an improved prioritisation system if sufficient numbers of surgeons and patients participate.²⁸

The present study has several limitations, which may have introduced error in the estimates of the numbers and characteristics of those potentially eligible for bariatric surgery, with most summarised in Table 1. In addition, resistant obesity could not be classified from the AHS data, although we expect the effect of this limitation to be small because sustained weight loss is unlikely in this population group.²⁹ In the absence of other data, we used elevated ALT as a surrogate marker of NASH (gold standard for diagnosis is liver biopsy³⁰), which may have misclassified some with Class 2 obesity as potentially eligible. ALT has been found to be an independent predictive marker of NASH, at least in those with a BMI ≥40 kg/m²,³¹ and a systematic review reported the estimated prevalence of NASH in the obese population to be between 10% and 56% (median 33%).³⁰ Of the total population we estimated to be potentially eligible for bariatric surgery, 12.1% had Class 2 obesity and became eligible only through high ALT levels. This figure appears reasonable given the NASH prevalence data available.³⁰ Elevated ALT can arise as a result of excess alcohol intake and, in a subanalysis using the NHS, we estimated that 25% of this group had alcohol intake exceeding 20 mL day^–1. However, this estimate was determined from a small sample (n = 24) using 3-day self-report alcohol consumption data. Furthermore, there were no direct measures or surrogate indicators of GORD in the AHS. Using self-reported medication data in the NHS for the collective category ‘GORD or peptic ulcers’, we found that the sample size increased by only eight individuals or 2.2%. Therefore, the effect on our estimates from the AHS data is likely to be small. There was also the possibility of error in our estimates because of self-report inaccuracies related to disclosure of comorbidities and smoking status. Further, some participants in our sample may have already had bariatric surgery and, of those classified as potentially eligible, not all would want surgery and some may be unsuitable for reasons undetectable through the AHS (e.g. due to clinical contraindications).² Finally, although the AHS was a high-quality national health survey, there are limitations specific to the survey design that have been comprehensively described in the user’s guide, for example those relating to sampling variability and non-sampling error.²¹

The findings of the present study were drawn from a large (n = 31 837), comprehensive and high-quality national health survey that included measured physical and biomedical characteristics. For the first time, population estimates and the characteristics of those potentially eligible for bariatric surgery in Australia have been quantified and described based on the best available evidence, using categories that best approximate the national recommended eligibility criteria.² The findings of the present study have important implications for health service planning, especially now that the inclusion of bariatric surgery in the treatment algorithm of T2DM has been widely endorsed.³²

Key findings relevant to health service planners are summarised in Box 1.

Potential demand for bariatric surgery in Australia, particularly in the public health system and outside major cities, far exceeds current capacity, highlighting an immediate need for improved prioritisation guidelines for eligible patients. Further, the large number potentially eligible for bariatric surgery (n = 882441) provides a strong signal that more funding for public surgery and other effective interventions are urgently needed for this population group.

Martin Hensher is an employee of the Department of Health and Human Services Tasmania and, as such, does not receive direct personal funding from any of the sources declared below. Martin Hensher has been involved in making policy decisions and funding allocations for the provision of bariatric surgery in Tasmanian public hospitals. All other authors declare they have no competing interests.