Impact of specialized obesity management services on the reduction in the use of acute hospital services

Summary

Severe obesity affects 4% of Australians and is associated with increased use of healthcare services and higher healthcare costs. This study evaluates the effect of attending a public tertiary obesity service on acute hospital use. This record‐linkage study included people aged ≥16 years with severe obesity who attended the Nepean Blue Mountains Family Metabolic Health Service (FMHS), New South Wales, Australia between January 2017, and September 2021. Emergency department (ED) presentations and acute hospital admissions and respective costs in the 1‐year and 3‐years pre‐and‐post first FMHS attendance were compared, overall and for adequate attendance (≥5 visits). A total of 640 patients (74% female, 50% <45 years) attended the FMHS, totalling 15 303 occasions of service, average 24 per person. There was a 31.0% and 17.6% reduction in acute admissions and ED presentations, respectively, translating into 34.0% and 23.4% decrease in costs. Adequate engagement was associated with a 48% decreased risk of acute admission (odds ratio 0.52; 95% confidence interval 0.29–0.94). Over 3‐years, there was a 19.8% and 20.7% reduction in acute hospital admissions and ED presentations, respectively. Findings indicate that tertiary obesity services reduce acute hospital use. Improved access to specialized obesity management may offload hospitals and contribute to acute healthcare cost avoidance.

What is already known about the subject

• Severe obesity affects 4% of all Australians. People with severe obesity use healthcare services more and incur a higher cost than those with lower BMIs.

What this study adds

• Specialist obesity services have a significant impact on reducing emergency department presentations and acute hospital admissions for people with severe obesity. As a result, there is significant cost avoidance to the healthcare system, particularly for those with adequate service engagement (≥5 visits).

• Access to obesity services and effective treatments for those with severe obesity should be improved to offload an increasingly burdened hospital system.

• Greater recognition of obesity presence and severity is needed within state and national datasets to provide a true representation of health system usage and to improve targeted health service provision and the subsequent distribution of funding.

1. INTRODUCTION

The World Health Organization defines the highest class of obesity (Class 3) as a Body‐Mass‐Index (BMI) of ≥40 kg/m², ¹ which is coined ‘severe obesity’. The Australian National Health Survey 2017–18 reported that 4% of all Australians aged 18 and over had severe obesity. ² People with obesity use healthcare services at a higher frequency and incur a higher healthcare cost than people with normal a BMI (18.5–24.9 kg/m²). For those with a BMI between 30 and 35 (Class 1 obesity), healthcare expenditure is 19% higher, and for those with BMI >35 (Class 2/3) it is 51% higher, when compared to those with normal BMI. ³

Specialized public tertiary obesity services provide healthcare to people with severe and complicated obesity across Australia, however access is extremely limited. ⁴ Interventions used in these services are usually multidisciplinary and may include metabolic, hormonal, general medical and psychosocial screening and/or management, lifestyle behavioural interventions (i.e., changes to dietary intake and physical activity patterns), advanced psychological strategies and support, anti‐obesity pharmacotherapy and/or bariatric surgery. ⁵ , ⁶ , ⁷ However, the evidence on the effectiveness of specialist obesity services on health impacts is limited to date.

Two studies have focused on the evaluation of tertiary obesity services, with one focusing on lifestyle behaviour change through individual and groups sessions delivered by a multidisciplinary team, along with pharmacotherapy, leading to bariatric surgery ⁸ ; and the other involving a group programme plus individual appointments with a multidisciplinary weight management team. ⁹ These studies were small, including 64 and 180 persons, respectively, and while outcomes were positive, they focused on short‐term physiological outcomes comparing weight loss, number of medications and change in HbA1c levels pre‐and‐post programme only. Similar results were also found following evaluation of the National Health Service multidisciplinary weight management Tier 3 specialist services in the United Kingdom. ⁵ After 6 months of treatment among 11 735 participants, there was a reduction in mean BMI, waist circumference, HbA1c, fasting blood glucose, insulin usage and blood pressure. ⁵ These findings demonstrate that tertiary obesity services are effective, at least acutely, on improving health outcomes. However, there was no information or outcomes assessed on the long‐term impact (>12 months) of these services on health outcomes, healthcare utilization and costs.

This study aims to evaluate the impact of attendance at a specialized tertiary obesity service on acute hospital utilization and costs. Assessing the impact on healthcare utilization and costs associated with attendance at multidisciplinary specialist tertiary obesity services in Australia is important to inform health service planning and policy. This study aims to highlight the potential benefits of improved access to specialized obesity management for those with severe obesity.

2. METHODS

2.1. Study population and data sources

The study population included all patients aged ≥16 years who attended the Nepean Blue Mountains Family Metabolic Health Service (FMHS) in New South Wales (NSW), Australia between January 2017 and September 2021. The FMHS is a public, specialist tertiary obesity service in Greater Western Sydney. It is based in the Nepean Blue Mountains Local Health District, an area with a prevalence of obesity (BMI >30 kg/m²) of 28.5% in 2020, compared to the NSW state average of 22.5%. ⁸ The FMHS has four streams: (i) paediatrics, (ii) adolescents and young adults, (iii) non‐pregnant adults, and (iv) pregnancy and postnatal. The entry criteria into the FMHS are very restrictive due to the high demand for services and limited capacity of the unit. As such, only those adults with severe obesity are eligible for the service, which includes a BMI ≥40 kg/m² in all circumstances and one of the following: a severe chronic health condition related to obesity, young age, Australian Aboriginal ethnicity, pregnancy planning, a family member in the service or BMI ≥70 kg/m². Further details outlining the FMHS eligibility criteria can be accessed in Supplementary File S1. Approximately 2.4% of patients in the non‐pregnant adult stream obtain bariatric surgery per year. Details of the adult health service pathway are provided in Supplementary File S1. For this study, the non‐pregnant adult stream was targeted for analysis and thus adult women who had recorded a pregnancy‐related admission within 9 months after their first contact with the FMHS were excluded.

Non‐pregnant adults attending the FMHS were identified from the NSW Non‐Admitted Patient data collection (NAP). The NAP and other data collections for this study were sourced from the NSW Ministry of Health Register of Outcomes, Value and Experience (ROVE) database, which comprises linked records of multiple health administrative datasets including the NAP, the Admitted Patient Data Collection (APDC), the Emergency Department (ED) data collection and Registry of Births, Deaths, and Marriage's death registrations. The NAP data collection consists of all NSW Health non‐admitted patient clinical and/or therapeutic services provided. It includes patient demographic information, the service provider and type. The APDC is a census of all public and private hospital admissions. It collects demographic, clinical and health service information, including a primary diagnosis and up to 50 fields of diagnoses recorded for each hospital admission. Diagnoses are coded according to the 10th revision of the International Classification of Diseases, Australian Modification (ICD‐10‐AM). The ED data collection is a statutory collection of ED presentations across all public hospitals in NSW. Available data included records from July 2012 up until September 2021. Record linkage of the respective datasets was conducted by the NSW Centre for Health Record Linkage and personal identifiers are replaced with a unique person project number, with only deidentified data available for analyses.

2.2. Study outcomes and explanatory variables

Health service utilization and respective costs were defined as ED presentations and acute inpatient hospital admissions (excluding planned and elective admissions) 1‐year and 3 years pre and post initial FMHS attendance. Total presentations were calculated per patient and categorized into none, 1–2 and ≥3 visits for each of the two acute hospital care settings. The overall cost of ED presentations and acute hospital admissions were calculated in Australian dollars (AU$) using the National Weighted Activity Unit (NWAU) assigned to each episode of care and multiplying by the NSW National Efficient Price for each respective year.

Socio‐demographic explanatory variables assessed included age, sex, and engagement with the service. People who had ≥5 clinic visits to the FMHS were predefined prior to any statistical analysis as having ‘adequate engagement’. This number was based upon the expert opinion of clinicians, who felt that by this point an individual would be likely to have an established relationship with the FMHS and been exposed to 3–4 members of the multidisciplinary team. This includes a medical appointment with an endocrinologist, psychologist, dietitian, and physiotherapist. Treatment includes a variety of complementary behavioural lifestyle interventions to address obesity and/or medical management of obesity. Reason for admission to hospital was assessed using the primary diagnosis field and categorized using major ICD10‐AM diagnosis chapters (e.g., Diseases of the circulatory system: ICD10‐AM codes I00–I99; Diseases of the digestive system: K00–K99). We also evaluated the accuracy of the APDC in recording an obesity diagnosis during an inpatient admission using either of the obesity‐related ICD10‐AM codes, U78.1 or E66, in any diagnosis field.

2.3. Statistical analysis

We used descriptive statistics and contingency tables to explore socio‐demographic characteristics of people attending the FMHS and differences in ED presentations and acute admissions in the 1‐year and 3‐years pre and post clinic attendance. We also evaluated differences in total and median (with interquartile range) number, cost of acute admissions and ED presentations and reason for acute admission before and after initial contact with the FMHS. We applied univariable logistic regression models to assess the association between main explanatory variables of adequate attendance (≥5 visits), age, sex, and previous presentation with subsequent admission to hospital or presentation to ED in the 1‐year following clinic attendance and controlled for these factors using multivariable models. Given the large proportion of people without hospitalizations, the effect of adequate attendance on the number of episodes (among those just presenting to hospital) in the subsequent 1‐year period was assessed using zero‐inflated Poisson regression models after controlling for age, sex, and previous episodes. To assess the recording of obesity, we evaluated the overall and annual proportion of inpatient admissions that had a recorded obesity diagnosis. All analyses were conducted using SAS v9.4. This study was approved by the Nepean Blue Mountains Human Research and Ethics Committee, HREC No. 2022/ETHOO540.

3. RESULTS

There were 640 patients who attended the FMHS, totalling 15 303 occasions of service (average of 24 occasions of service per person). The socio‐demographic characteristics are presented in Table 1. Three‐quarters of adults attending the clinic were female, half were aged under 45 years and 11.4% were born overseas.

TABLE 1.

Socio‐demographic characteristics of the adult non‐pregnant cohort of people with severe obesity attending the Nepean Blue Mountains Family Metabolic Health Service.

	Total n = 640
Year of first visit	n (%)
2017	144 (22.5)
2018	163 (25.5)
2019	135 (21.1)
2020	83 (13.0)
2021	115 (18.0)
Sex
Male	166 (25.9)
Female	474 (74.1)
Country of birth
Australia or New Zealand	567 (88.6)
Other	73 (11.4)
Age
16–29 years	130 (20.3)
30–44 years	190 (29.7)
45–59 years	203 (31.7)
≥60 years	117 (18.3)

The number, proportion and value of acute admissions and ED presentations for 1 and 3 years pre‐and‐post the initial FMHS visit are presented in Figure 1 and Table 2, respectively. On average, only 19.8% of patients had an acute admission to hospital, with most having only 1–2 admissions and 3.5% having ≥3. The proportion of patients admitted to hospital reduced from 23.3% pre‐FMHS attendance to 17.0% at 1‐year post‐FMHS attendance; and from 42.6% to 38.6% in the 3 years post, when compared to the 3 years prior to, attendance (Figure 1).

FIGURE 1.

Rates of acute hospital admissions and ED presentations in people attending the Nepean Blue Mountains Family Metabolic Health Service.

TABLE 2.

Number and cost of acute admissions and emergency department presentations of people attending the Nepean Blue Mountains Family Metabolic Health Service.

	People	Number of episodes (pre‐post)	Difference in episodes (% change)	Cost (AU$) (pre‐post)	Difference in costs (% change)
Acute admissions
One year total	511	210–145	−65 (−31.0)	1 029 077–679 011	−350 066 (−34.0)
Attendance
<5 visits	84	51–44	−7 (−13.7)	355 397–244 988	−110 409 (−31.1)
5+ visits	427	159–101	−58 (−36.5)	673 680–434 023	−239 657 (−35.6)
Three years total	249	239–199	−40 (−16.7)	1 407 000–1 428 710	21 710 (1.5)
Attendance
<5 visits	33	22–24	2 (9.1)	207 120–231 214	24 094 (11.6)
5+ visits	216	217–175	−42 (−19.4)	1 199 880–1 197 496	−2384 (−0.2)
ED presentations
One year total	511	636–524	−112 (−17.6)	427 064–327 233	−99 831 (−23.4)
Attendance
<5 visits	84	164–147	−17 (−10.4)	108 470–94 608	−13 862 (−12.8)
5+ visits	427	472–377	−95 (−20.1)	318 594–232 625	−85 969 (−27.0)
Three years total	249	725–596	−129 (−17.8)	464 065–397 283	−66 782 (−14.4)
Attendance
<5 visits	33	56–63	7 (12.5)	36 942–41 518	4576 (12.4)
5+ visits	216.0	669–533	−136 (−20.3)	427 123–355 765	−71 358 (−16.7)

There was a 31.1% and 17.6% decrease in acute admissions and ED presentations, respectively, between the year pre and post FMHS attendance (Table 2). This reduction was two to three times higher for those with ≥5 (36.5% for acute admissions and 20.1% for ED presentations), compared with those with <5 visits (13.7% and 10.5%) (Table 2). After considering patient age, sex and previous health contacts, for those with ≥5 visits, there was a 35% reduction in risk of acute admissions (adjusted relative risk (aRR) 0.65; 95% confidence interval [CI] 0.39, 1.09) and 66% fall in ED presentations (aRR 0.44; 95% CI 0.35, 0.54) (Supplementary Table S1). This translated into an overall 34.0% and AU$350 066 reduction in the total cost of acute admissions and 23.4% and AU$99 831 decrease in the cost of ED presentations. Reduction in cost was greater for patients having ≥5 visits compared with those with <5 visits (Table 2). Slightly lower, but similar rates of hospital admissions and ED presentations were observed for the period 3‐years after the first FMHS visit when compared to the 3 years prior (Table 2, Figure 1).

Compared to patients with <5 visits to FMHS, those attending with ≥5 visits had a 48% reduction in odds of hospital admission (adjusted odds ratio, aOR, 0.52; 95% CI 0.29, 0.94), while there was no there no association between obesity service attendance and ED presentations in the 1 year following the initial clinic visit (Table 3). Previous ED presentation or hospital admission was associated with a 2.7‐fold (95% CI 1.87, 3.90) and 3.8‐fold (95% CI 2.32, 6.30) increased odds of admission or ED presentation (Table 3).

TABLE 3.

Association between patient characteristics and admissions to hospital and ED presentations at 1 year following initial clinic visit.

	Acute admissions			ED presentations
	n (%)	Crude OR (95% CI)	Adjusted OR (95% CI)	n (%)	Crude OR (95% CI)	Adjusted OR (95% CI)
Age
16–29 years	10 (12.2)	0.67 (0.31, 1.45)	0.64 (0.29, 1.41)	31 (37.8)	0.92 (0.54, 1.57)	0.82 (0.47, 1.43)
30–44 years	26 (18.4)	1.09 (0.62, 1.94)	0.93 (0.51, 1.70)	65 (46.1)	1.30 (0.83, 2.02)	1.23 (0.77, 1.94)
45–59 years	31 (17.1)	Ref	Ref	72 (39.8)	Ref	Ref
60+ years	20 (18.7)	1.11 (0.60, 2.07)	0.98 (0.51, 1.88)	39 (36.4)	0.87 (0.53, 1.42)	0.81 (0.49, 1.35)
Sex
Male	29 (19.9)	Ref	Ref	59 (40.4)	Ref	Ref
Female	58 (15.9)	0.76 (0.47, 1.25)	0.98 (0.58, 1.66)	148 (40.5)	1.01 (0.68, 1.49)	1.04 (0.69, 1.57)
Attendance
<5 visits	22 (26.2)	Ref	Ref	35 (41.7)	Ref	Ref
5+ visits	65 (15.2)	0.51 (0.29, 0.88)	0.52 (0.29, 0.94)	172 (40.3)	0.94 (0.59, 1.52)	1.00 (0.61, 1.64)
Previous admissions or ED
No	46 (11.7)	Ref	Ref	83 (29.9)	Ref	Ref
Yes	41 (34.5)	3.95 (2.43, 6.44)	3.83 (2.32, 6.30)	124 (53.2)	2.03 (1.34, 3.07)	2.71 (1.87, 3.90)

Abbreviations: CI, confidence interval; OR, odds ratio; Ref, reference group.

In the 1‐year following initial FMHS visit, there was a decrease in the proportion of people admitted to hospital for diseases of the circulatory (14.3%–8.3%) and respiratory systems (10.5%–5.5%) and for endocrine, nutritional, and metabolic diseases (5.2%–2.1%). However, there was an increase in admissions for diseases of the genitourinary (9.0%–10.3%) and digestive systems (7.6%–9.7%) and for injuries (7.6%–9.7%) (Figure 2). For all admissions to hospital, we found the diagnosis of obesity to be recorded in only 30.8% of all acute hospital admissions. This proportion increased from 17.2% in 2015 to 46.6% in 2021.

FIGURE 2.

Reasons (primary diagnosis) for acute admissions to hospitals 1‐year pre and post initial clinic visit.

4. DISCUSSION

We found that people with severe obesity attending a specialized tertiary public obesity service reduced their acute hospital admissions and ED presentations by 34%, resulting in AU$350 066 cost avoidance in the first year following initial clinic visit. The greatest cost avoidance was among patients with adequate engagement (≥5 visits). Adequate engagement with the FMHS resulted in a reduction of just below 37% in acute hospital admissions and a 20% decrease in ED presentations in the year after first attendance when compared to the year prior. This decrease was sustained 3 years after initial engagement in those with adequate engagement, with an approximate reduction of 20% in both ED presentations and acute admissions when compared to the 3 years prior. We also found that patients who attended the FMHS at least 5 times had a 50% lower risk of being hospitalized in the first year post initial visit when compared to those who attended <5 times, highlighting the effect of engagement and interventions received. Diseases of the circulatory and respiratory systems were some of the most prevalent indications for acute hospital presentations and they were also reduced after contact with the FMHS. We found that only one in three patients confirmed to have class 3 obesity had a diagnosis of obesity recorded in their hospital record during acute presentations to hospital.

Results show a reduction in health service use and costs are potentially attributable to the continuity of care provided by FMHS clinicians, which includes frequent monitoring and timely treatment of medical complications, holistic management of obesity and facilitated engagement with the patient's General Practitioner (GP). ¹⁰ , ¹¹ While needing investigation, further declines in hospital presentations observed are unlikely to be due to weight changes alone, or bariatric surgery (with only ≈2% in this cohort having surgery per year). Other factors, such as changes to patient reported measures (e.g., patient experience, quality of life and patient activation), may be contributing factors ⁹ , ¹² , ¹³ and should be investigated in future studies.

To our knowledge, there are no prior studies that have looked at the impact of specialized tertiary obesity services on healthcare utilization and costs for adult patients with obesity. However, previous studies suggest an association between increasing BMI and increased rates of hospitalizations. A population‐based prospective cohort study of 246 361 individuals in the 45 and Up study from New South Wales, Australia, demonstrated that for every 1 kg/m² increase in BMI, there is a 4% increase in hospitalizations. Previous studies have also confirmed that weight loss is a predictor of reduced healthcare utilizations and cost in the long term. ¹⁴ Thus it could be inferred that, if services can assist a patient to maintain a stable body weight or achieve weight loss, this may affect downstream healthcare utilization and costs.

The evidence demonstrates that specialized tertiary obesity services in Australia are effective in achieving weight loss, ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ however other interventions have shown a reduction in disease risk factors through a variety of treatment methods, including meal replacement diets, pharmacotherapy and bariatric surgery. ⁵ , ¹⁷ , ²⁰ Other patient reported measures of such services should be explored in future studies for their impact on health service use. There are many challenges experienced when trying to evaluate the impact of specialized tertiary obesity services on health outcomes, health utilization and costs. Within services themselves there is a lack of resources and agreement on outcomes to conduct research on comprehensive health and wellbeing impact of interventions received. This may explain the lack of evidence found on this topic.

Low rates of formal coding of obesity in state and federal datasets is another challenge when trying to understand the impact of obesity on acute health services. In the acute hospital setting there is a lack of routine body weight and height measurements taken to enable BMI calculation to identify obesity when a person enters the system and during admissions. Furthermore, obesity is rarely included in diagnosis lists for coding by clinicians. We found that only one in three hospitalizations in our cohort of patients with severe and complex obesity had been coded for it, consistent with what has previously been reported in the literature. ²¹

Potential factors that could be contributing to low rates of obesity coding include: (i) the lack of recognition by clinicians of its role in hospital presentations due to the high rates in the background population and inadequate education related to the condition; (ii) limited weight recording capacity due to human resourcing and physical infrastructure constraints; (iii) lack of standardization in the documentation of obesity in clinical notes; and (iv) lack of resourcing in hospital systems to address obesity, reducing the impetus to identify it. Incomplete coding of obesity, as demonstrated in this study, may lead to a significant underestimation of health service use attributable to obesity, which has flow on effects for national and state health policy and clinical and research funding. This contributes to a lack of access to evidence‐based therapies for people with obesity and a lack of appreciation of obesity as a chronic disease. ²² Inadequacies in coding for obesity in clinical data sets should be proactively addressed as a matter of urgency to enable a full appreciation of the cost of obesity to our healthcare services.

Multidisciplinary obesity services in Australia often have strict entry criteria and prolonged wait times, severely restricting access to care for those who are affected by obesity. ⁴ Furthermore, services are typically limited to major cities and interventions offered may result in out of pocket costs. ⁴ Public services often have significant gaps in staffing a multidisciplinary team and there is limited access to pharmacotherapy and bariatric surgery. ⁴ To overcome these issues directly related to health equity, there is a need to better understand the capacity for integrated care and virtual care to enhance current services and/or to scale services across the country using a hub‐and‐spoke, ²³ teaching and training approach. Funding is also required to improve the evidence base for current interventions and models of care already funded, and in resourcing of new services through planned evaluation. Furthermore, research is needed to understand the clinical and service factors of the FMHS that drive the reduction in healthcare utilization and costs. The nature of acute hospital care needed by people with severe obesity and the potential impact that specialized tertiary obesity services can have to reduce these admissions (and/or readmissions and length of stay), especially in relation to cardiorespiratory diseases, needs to be explored.

One of the strengths of this study was our ability to identify all patients who attended the FMHS, all of whom are confirmed to have severe obesity. The use of state‐wide linked health data also enabled the capture of all hospital episodes and ED presentations of these patients across all hospitals in NSW, which is otherwise unachievable from a single centre study. Limitations of this study are the lack of clinical data, including anthropometrics such as weight and height, as well as data on use of primary care, medications, and other medical interventions, which limits interpretation of the results seen. In addition, the characteristics of the population attending FMHS may be different to people with severe obesity in the community who did not obtain a referral to the FMHS, due to higher levels of motivation to engage in behaviour change and improve their health. In addition, the current study design does not address whether the difference in ED presentations and hospitalizations between patients who did (or did not) receive the predetermined ‘adequate dose’ of weight management therapy at the FMHS is due to receipt of these services or something inherently different about the sub‐group of patients. There is the possibility that differences observed may be partly due to patient‐level differences, such as in engagement, adherence, motivation and social determinants of health. As such, without measuring outcomes in a control group with the absence of the intervention, pre versus post longitudinal differences in outcomes may be affected by regression to the mean. However, we also found that patients with adequate FMHS engagement had higher reductions in admissions and ED presentations compared to those with low engagement (including post baseline adjusted odds), indicating that it is unlikely that results were solely the effect of the selective nature of the cohort and not the intervention.

In conclusion, we found that attendance at the FMHS was associated with lower use of acute hospital care at 1 and 3 years after the first appointment, when compared to the 1 and 3 years prior, particularly for patients with adequate engagement. Our findings indicate that obesity speciality services are cost‐effective interventions and support scale‐up across different healthcare settings.

AUTHOR CONTRIBUTIONS

Natasha Nassar and Kathryn Williams designed the study. Francisco J. Schneuer analysed the data. Kathryn Williams and Gabrielle Maston wrote the original manuscript. All authors edited and reviewed the final manuscript.

CONFLICT OF INTEREST STATEMENT

Kathryn Williams received payment from Eli Lilly, BI, Pfizer, NovoNordisk and Lilly for participation in a special advisory group/speaker fee. Travel expenses have been provided by Eli Lilly, and she had a leadership role in the Kellion Diabetes Foundation until 2021. Gabrielle Maston has received payment from NovoNordisk in speaker fees. Francisco J. Schneuer has no interested to declare. Natasha Nassar has received grants from the National Health and Medical Research council and Cancer Australia.

Supporting information

Data S1: Supporting information

Williams K, Maston G, Schneuer FJ, Nassar N. Impact of specialized obesity management services on the reduction in the use of acute hospital services. Clinical Obesity. 2023;13(5):e12592. doi: 10.1111/cob.12592