
==== Front
Prev Med Rep
Preventive Medicine Reports
2211-3355

S2211-3355(24)00294-8
10.1016/j.pmedr.2024.102879
102879
Infectious Disease
Factors associated with receiving an obesity diagnosis and obesity-related treatment for patients with obesity class II and III within a single integrated health system
Szymanski Raphael a
Abraham Megha a
Childs William a
Le Kristina a
Velez Christopher a
Vaughn Ivana bc
Lamerato Lois bc
Budzynska Katarzyna kbudzyn1@hfhs.org
ac⁎
a Department of Family Medicine, Henry Ford Hospital, Detroit, MI, USA
b Department of Public Health Sciences, Henry Ford Health, Detroit, MI, USA
c Henry Ford Health + Michigan State University Health Sciences, Detroit, MI, USA
⁎ Corresponding author at: Department of Family Medicine, Henry Ford Medical Center - Harbortown, 337 E Jefferson Ave, Detroit, MI 48207, USA. kbudzyn1@hfhs.org
05 9 2024
10 2024
05 9 2024
46 1028798 5 2024
29 8 2024
30 8 2024
© 2024 The Authors
2024
https://creativecommons.org/licenses/by-nc-nd/4.0/ This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Highlights

• This cross-sectional study characterized diagnostic and prescribing trends within a single health system by evaluating the rates of obesity diagnoses (per HCC coding) and obesity-specific treatments for patients with the physiological features of class II and III obesity.

• Of over 67,000 patients with the clinical characteristics of class II or III obesity, only 40.7% were clinically diagnosed with obesity, and 23.5% received an obesity-related intervention.

• Patients with BMI≥40 kg/m2 were significantly more likely to be diagnosed with severe obesity than those with BMI 35–39.9 kg/m2.

• Patients with a diagnosis of severe obesity and Black and female patients were significantly more likely to be offered obesity-related treatment.

Objectives

The prevalence and associated adverse effects of obesity on health and healthcare cost make it a primary public health concern. However, individuals with the physiological features of obesity may be underdiagnosed and undertreated. We aimed to determine the prevalence of obesity diagnoses and obesity-related treatments in an integrated health system and determine the factors associated with receiving an obesity diagnosis and treatment for this indication.

Methods

This retrospective cross-sectional study of data from the Henry Ford Health electronic health record included adult patients with a body mass index (BMI) indicating clinical evidence of class II and III (severe) obesity in 2017 and who received treatment through 2019. The primary outcome was prevalence of obesity diagnosis and obesity-related treatment. Logistic regression evaluated the patient-level factors associated with odds of having obesity diagnosis and treatment.

Results

Among 64,741 patients meeting the clinical definition of definition of severe obesity, only 40.7 % were clinically diagnosed with obesity, and 23.5 % received an obesity-related intervention. Patients with BMI≥40 kg/m2 (class III) were more likely to be diagnosed with obesity than those with BMI 35–39.9 kg/m2 (class II) (odds ratio [OR] 5.84; 95 % CI, 5.62–6.07). Patients with a diagnosis of obesity (OR 2.92; 95 % CI, 2.80–3.05), Black patients (OR 1.46; 95 % CI, 1.40–1.53), and female patients (OR 1.47; 95 % CI, 1.41–1.54) were more likely to be offered obesity-related treatment.

Conclusions

Severe obesity may be underdiagnosed in patients who have BMI 35–39.9 kg/m2 and 1 comorbidity.

Keywords

Obesity class II and III
Diagnosis
Treatment
Abbreviations

BMI body mass index

ICD-10CM International Statistical Classification of Diseases and Related Health Problems, 10th revision, Clinical Modification

OR odds ratio

US United States

GLP-1 glucagon-like peptide 1
==== Body
pmc1 Introduction

Obesity is a major public health concern in the United States (US) and throughout most of the world due to its high prevalence, direct and indirect adverse effects on health and quality of life, and cost to the healthcare system (Klein et al., 2002). Obesity was defined as a disease by the World Health Organization and Centers for Medicare and Medicaid Services in 2013 (World Health Organization, 2021), and by 2014, 66.2 % of women and 73.0 % of men worldwide met the clinical criteria for having overweight or obesity (National Center for Health Statistics, 2016). In 2016, 1.9 billion adults worldwide met the criteria for having obesity, accounting for 39 % of the global population (World Health Organization, 2021). Obesity is estimated to cost $173 billion a year within the US (Centers for Disease Control and Prevention, 2022), and multiple studies have suggested that obesity leads to frailty, poor health-related quality of life, and disease clustering (Agborsangaya et al., 2013, Fortin et al., 2004, Garin et al., 2016, Jackson et al., 2015, Kivimäki et al., 2017, Strandberg et al., 2012). It has also been linked to poorer mental health and serious comorbid conditions, including cardiovascular disease, diabetes, and cancer (Tucker et al., 2021). The influence of obesity on some of these chronic conditions has been identified as one of the leading causes of death in the US and worldwide (D’Souza et al., 2018). Large observational studies have shown a three- to eight-year reduction in disease-free life in individuals with obesity (Nyberg et al., 2018), as well as a an up to three-fold increased risk of all-cause premature death (GBD 2015 Obesity Collaborators; Global BMI Mortality Collaboration, 2017). Importantly, individuals with obesity face damaging levels of societal stigma, which can have an additional impact on mental health and result in disparities in health care (Harwood et al., 2022, Lewis et al., 2016, Puhl and Heuer, 2010).

Primary care physicians encounter many patients with obesity, (Tucker et al., 2021) and the prevalence and associated disease burden are increasing worldwide (Dai et al., 2020, GBD, 2019 Risk Factors Collaborators, 2020) to global epidemic proportions (Noria and Grantcharov, 2013). The clinical classification of obesity is important for medical evaluation and treatment, since being within a higher obesity class is associated with increased mortality (Berrington de Gonzalez et al., 2010, Flegal et al., 2013, Prospective Studies Collaboration, 2009). Weight and obesity categories defined in the World Health Organization recommendations are widely used and include a scale that goes from normal to overweight, obese, and severely obese with increasing body mass index (BMI) (Suissa et al., 2021, World Health Organization, 2000). Overweight and obesity are also specifically classified by the US Preventive Services Taskforce and Centers for Disease Control and Prevention as overweight, class I, II, and III obesity, also defined by increasing BMI ranges (Ackermann et al., 2008). Importantly, individuals with class II (BMI 35.0–39.9 kg/m2) and class III obesity (BMI≥40 kg/m2) are at a higher risk of developing metabolic and cardiovascular disease and disability than those who are overweight or who have class I obesity (Dixon, 2010, Haslam and James, 2005, Walls et al., 2010). Also, evidence suggests that the prevalence of individuals approaching class III obesity is growing relative to those who are overweight or in class I (Walls et al., 2010, Sturm, 2007, Pasco et al., 2013). While obesity-related comorbidities are frequently addressed in primary care, obesity itself is chronically undertreated (Ciciurkaite et al., 2019, de Heer et al., 2019, Fitzpatrick and Stevens, 2017, Valero-Elizondo et al., 2019), and few studies have examined how obesity is diagnosed (Ciemins EL, Joshi V, Cuddeback JK, Kushner RF, Horn DB, Garvey WT. Diagnosing Obesity as a First Step to Weight Loss: An Observational Study. Obesity (Silver Spring)., 2020, Ciemins et al., 2021) or what clinical features are most important for the proper diagnosis of class II or III obesity in the US, especially in urban and medically underserved areas. A recent study within the US showed that documentation of an obesity diagnosis (BMI above 30) was associated with weight loss, even when controlling for confounding factors. (Ciemins et al.,2020).

Because obesity is associated with an increasing number of diverse disease burdens, it may be an important target for multimorbidity prevention that avoids the complexities of multitarget preventive regimens (Kivimäki et al., 2022). By undertreating obesity in the primary care setting, clinicians may be missing valuable opportunities to markedly improve the overall health of their patients at both the personal and population level (Tucker et al., 2021). Additionally, in the US, severe obesity is a hierarchical condition category code (HCC) required for risk adjustment; thus, HCC coding relies on International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) coding to assign risk scores to patients. Each HCC is mapped to an ICD-10-CM code, and severe obesity is scored high. Therefore, not adequately documenting the diagnosis and treatment of patients within the highest obesity range may be leaving millions of dollars in potential reimbursement unused, which could in turn be employed for providing meaningful interventions to this patient population.

The aim of this retrospective cross-sectional study was to assess the level at which patients with class II and class III obesity (herein collectively referred to as “severe obesity”) were diagnosed with obesity and treated at our institution and to determine the parameters associated with both diagnosis and initiation of treatment specifically for obesity. Our goal was to determine to what level severe obesity may be underdiagnosed or undertreated at our institution so that policies and practices may be developed to better serve this patient population and prevent unnecessary future morbidity.

2 Materials and Methods

2.1 Study population

This was a cross-sectional retrospective study of patients who had the clinical parameters indicative of class II and III obesity and had two or more outpatient primary care encounters that took place between January 1, 2017, and January 31, 2017, within a large metropolitan health system in the US in the state of Michigan. Data were collected for the year 2017 to clearly delineate diagnostic and treatment trends from any that may have been affected by the COVID-19 pandemic that started in 2020. This study was approved by the Institutional Review Board at Henry Ford Health on March 24,2023 (IRB#16302). Due to the retrospective nature of the study, the need for consent was waived. This study aligned with all guidelines for the protection of human subjects.

2.2 Inclusion and exclusion criteria

Inclusion criteria included age 18–75 years, record of two or more clinical visits in 2017 in primary care clinics, and clinical evidence of class II or class III obesity. For the purposes of this study and to align with current recommendations of the Obesity Medicine Association, we use the term “severe obesity” instead of the now non-recommended term “morbid obesity.” Based on Centers for Medicare and Medicaid Services criteria, we defined patients as having “severe obesity” if they had class II obesity defined as BMI 35.0–39.9 kg/m2 with the presence of at least one comorbidity or class III obesity defined as BMI≥40 kg/m2 (Centers for Medicare and Medicaid Services, 2012). Diagnoses found on problem lists were not included. Comorbidities included the following per Medicare guidelines for surgical management of severe obesity (Centers for Medicare and Medicaid Services, 2019): type 2 diabetes, essential hypertension, hyperlipidemia/dyslipidemia, obstructive sleep apnea, obesity-hypoventilation syndrome, arthropathy of spine or weight-bearing joints, metabolic dysfunction–associated steatotic liver disease/steatohepatitis, and pseudotumor cerebri. Exclusion criteria included patients who had only one clinical visit during the study period. This criterion was used to assure that patients had a continuity of care and a primary care connection. (Fig. 1).Fig. 1 Criteria to determine final study population in adult patients (2017–2019) within a large metropolitan health system in the US in the state of Michigan. Abbreviations: BMI, body mass index.

2.3 Dataset source and study variables

Data were obtained from the Henry Ford Health electronic health record databases and filtered based on inclusion/exclusion criteria. Comorbidities were defined as the presence of specific International Statistical Classification of Diseases and Related Health Problems, 10th revision, Clinical Modification (ICD-10-CM) codes corresponding to the aforementioned comorbidities. All possible ICD-10-CM codes were individually reviewed to assure correspondence with the comorbid diagnosis.

The following baseline patient characteristics were extracted: age, sex, race/ethnicity as reported in the medical record, BMI, median household income, obesity-related comorbidities, and number of obesity comorbidities. Age was also categorized as 18–35, 36–50, 51–64, and 65–75 years old. Sex was categorized as male and female. Race and ethnicity were categorized as self-reported in the medical record as Asian, Black/African American, Hispanic, White, and other/unknown. Median household income was obtained by mapping patients’ addresses to census block group level median household income. Primary outcomes included the presence of a coded severe obesity diagnosis within the study period (“severe obesity diagnosis”) and whether a medical intervention for severe obesity was ordered within the study period (“obesity-related intervention”), which includes both pharmacological therapy and referrals to specialists with the intent to directly treat severe obesity. Patients were considered to have a severe obesity diagnosis if a corresponding ICD-10-CM code (E66.01 or E66.20) for “morbid obesity” was entered in the medical chart during the three-year study period. Likewise, patients were considered to have received a severe obesity-related intervention if a corresponding treatment was present in the chart within the same period. Anti-obesity medications included lorcaserin, phentermine, phentermine/topiramate, bupropion/naltrexone, orlistat, metformin, liraglutide, and semaglutide (brand name Ozempic-Wegovy was not approved for treating obesity during the study period). Of note, medications to treat diabetes that were given to patients who had a diabetes diagnosis were not considered as a severe obesity treatment in our model, as the treatment of diabetes with diabetes medication does not constitute a primary obesity treatment. However, medications indicated for treatment of diabetes that were prescribed for patients who did not have a diabetes diagnosis were presumed to be given with the intent to treat severe obesity and were included as an obesity-related intervention.

2.4 Data analysis

Baseline characteristics were used as the main predictors in all models. Descriptive analysis included frequencies with percentages or means with standard deviations as applicable. Bivariate analysis of patient characteristics and outcomes to assess the patient characteristics associated with receiving a diagnosis of severe obesity or receiving obesity-related interventions was done with chi-square analysis. Logistic regression models assessed the association of the aforementioned variables with a diagnosis of severe obesity and with obesity-related intervention. Data were analyzed in SPSS version 26 (IBM, Armonk, NY).

3 Results

Of 405,540 medical records screened, a total of 64,741 patients (∼16 %) had the clinical characteristics of severe obesity during the study period (63.2 % women and 36.8 % men). Of these patients, 37,733 (58.5 %) had BMI 35.0–39.1 kg/m2 with at least one comorbidity and 27,008 (41.7 %) had BMI≥40 kg/m2. The mean ± standard deviation age was 53 ± 13.6 years. Patients had a mean of 1.8 ± 1.2 obesity-related comorbidities, and most patients (72.5 %) had a median household income between $25,000 and $75,000 (Table 1). All demographic information and patient characteristics are summarized in Table 1.Table 1 Baseline demographic characteristics of adult patients 18–75 years old with clinical definition of class II or class III obesity in 2017 within a large metropolitan health system in the US in the state of Michigan.

	Number (%)
N=64,741	
Sex		
Female	40,937 (63.2)	
Male	23,804 (36.8)	
Age, years, mean ± SD	53 ± 13.6	
Age group in years		
18–35	8,340 (12.9)	
36–50	17,316 (26.7)	
51–64	24,586 (38.0)	
65–75	14,449 (22.4)	
Race and Ethnicity		
Asian	361 (0.6)	
Black/African American	20,536 (31.7)	
Hispanic	1,676 (2.5)	
White	38,875 (60.0)	
Other/Unknown	3,293 (5.1)	
Severe obesity category		
Class II: BMI 35.0–39.9 kg/m2 with obesity-related comorbidities	37,733 (58.3)	
Class III: BMI≥40 kg/m2	27,008 (41.7)	
Obesity-related comorbidities		
Hyperlipidemia	28,234 (43.6)	
Hypertension	42,445 (65.6)	
Nonalcoholic fatty liver disease	1,184 (1.8)	
Obesity-related hypoventilation	293 (0.5)	
Obstructive sleep apnea	10,034 (15.5)	
Osteoarthritis of weight-bearing joints	11,090 (17.1)	
Pseudotumor cerebri	177 (0.3)	
Type 2 diabetes	22,342 (34.5)	
Number of obesity-related comorbidities, all, mean ± SD	1.8 ± 1.2	
Number of obesity-related comorbidities		
0	10,380 (16.0)	
1	18,870 (29.1)	
2	16,447(25.4)	
3	13,249 (20.5)	
4	4,771 (7.4)	
5	943 (1.5)	
6	78 (0.1)	
7	3 (0.005)	
Median household income overall, US dollars, mean ± SD	$58,202 ± $25,212	
Median household income category, US dollars		
<$25,000	3,744 (5.8)	
$25,000 to <$50,000	21,372 (33.0)	
$50,000 to <$75,000	25,604 (39.5)	
$75,000 to <$100,000	9,220 (14.2)	
≥$100,000	4,039 (6.2)	
Income data missing	762 (1.2)	
BMI, body mass index; SD, standard deviation; US, United States.

*Data are reported as number (%) unless otherwise indicated.

The mean number of primary care visits within the study period for the entire cohort was 7.7 ± 6.2 visits, and most patients (42.1 %) had 5 to 9 visits. A total of 26,353 patients were clinically diagnosed with severe obesity during a clinical visit within the study period, revealing a 40.7 % prevalence of severe obesity diagnoses. There were 15,218 (23.5 %) patients who had an obesity-related intervention, and 8,744 (13.5 %) patients had a medication for treating obesity prescribed. A summary of healthcare and interventions for the study population is in Table 2.Table 2 Follow-up characteristics of obesity diagnosis and obesity-related medical interventions in adult patients 18–75 years old (2017–2019) within a large metropolitan health system in the US in the state of Michigan.

	Number (%)*
N=64,741	
Number of primary care visits from 2017 to 2019, mean ± SD	7.7 ± 6.2	
Number of primary care visits from 2017 to 2019		
2 to 4	21,725 (33.6)	
5 to 9	27,242 (42.1)	
10 or more	15,774 (24.4)	
Obesity diagnosed at visit	26,353 (40.7)	
Obesity-related intervention (all)	15,218 (23.5)	
Referral for obesity intervention	8,429 (13.0)	
Type of referral for obesity intervention		
Bariatric surgery	2,522 (3.9)	
Center for weight management	2,150 (3.3)	
Diabetes prevention program	352 (0.2)	
Nutrition services	1,894 (2.9)	
Preventive cardiology	2,825 (4.4)	
Preventive cardiology with weight management program	2,825 (4.4)	
Obesity medication prescribed†	8,744 (13.5)	
SD, standard deviation.

* Data are reported as number (%) unless otherwise indicated.

† Includes diabetes medications for patients without diabetes diagnosis. Anti-obesity medications included lorcaserin, phentermine, phentermine-topiramate, bupropion, naltrexone, orlistat, metformin, liraglutide, and semaglutide.

Notably, of the 37,733 patients with physiological features of class II obesity, only 5,351 (19.8 %) were diagnosed with severe obesity and 4,298 (15.9 %) had an obesity-related intervention. Of the 27,008 patients with physiological features of class III obesity, 21,002 (55.7 %) were diagnosed with severe obesity, and 10,920 (28.9 %) had an obesity-related intervention (Table 3). Bivariate analyses showed that all variables were significantly associated with receiving a diagnosis of severe obesity or having an obesity-related intervention, including sex, age strata, race/ethnicity, obesity category, number of obesity-related comorbidities, median household income, and number of primary care visits (all p < 0.001) (Table 3).Table 3 Bivariate analysis of the association between patient characteristics and receiving a diagnosis of severe obesity diagnosis or an obesity-related intervention in adult patients 18–75 years old with severe obesity within a large metropolitan health system in the US in the state of Michigan (2017–2019).

	All Patients with Severe Obesity
N=64,741	Severe Obesity Diagnosis
n = 26,353	Obesity-Related Intervention
n = 15,218	
	No.	No. and row %	P value	No. and row %	P value	
Sex			<0.001		<0.001	
Female	40,937	17,417 (42.5)		11,019 (26.9)		
Male	23,804	8,936 (37.5)		4,199 (17.6)		
Age group, years			<0.001		<0.001	
18–35	8,340	3,761 (45.1)		2,729 (32.7)		
35–50	17,316	7,630 (44.1)		5,327 (30.8)		
51–64	24,586	9,586 (39.0)		5,347 (21.7)		
65–75	14,449	5,376 (37.1)		1,815 (12.5)		
Race and Ethnicity			<0.001		<0.001	
Asian	361	119 (32.9)		109 (30.2)		
Black/African American	20,536	10,107 (49.2)		6,059 (29.5)		
Hispanic	1,676	680 (40.6)		486 (29.0)		
White	38,875	14,098 (36.3)		7,785 (20.0)		
Other/Unknown	3,293	1,349 (41.0)		779 (23.7)		
Severe obesity category			<0.001		<0.001	
Class II: BMI 35–39.9 kg/m2 with obesity-related comorbidities	37,733	5,351 (19.8)		4,298 (15.9)		
Class III: BMI≥40 kg/m2	27,008	21,002 (55.7)		10,920 (28.9)		
Number of obesity-related comorbidities			<0.001		<0.001	
0	10,380	4,498 (43.3)		3,263 (31.4)		
1	18,870	6,283 (33.3)		4,731 (25.1)		
2	16,447	6,425 (39.1)		3,827 (23.3)		
3	13,249	5,739 (21.8)		2,273 (17.2)		
4	4,771	2,866 (56.3)		891 (18.7)		
5	943	658 (69.8)		210 (22.3)		
6	78	61 (78.2)		21 (26.9)		
7	3	3 (1 0 0)		2 (66.7)		
Median household income category (US dollars)			<0.001		<0.001	
<$25,000	3,744	1,704 (45.5)		987 (26.5)		
$25,000 to <$50,000	21,372	9,512 (44.5)		5,354 (25.1)		
$50,000 to <$75,000	25,604	9,752 (38.1)		5,798 (22.6)		
$75,000 to <$100,000	9,220	3,509 (38.1)		2,100 (22.8)		
≥$100,000	4,039	1,557 (38.5)		947 (23.4)		
Number of primary care visits 2017–2019			<0.001		<0.001	
2 to 4	21,725	7,643 (35.2)		4,217 (19.4)		
5 to 9	27,242	11,689 (42.9)		6,825 (25.1)		
10 or more	15,774	7,021 (44.5)		4,176 (26.6)		
BMI, body mass index; US, United States.

*Study groups are described as number and row % with the All Patient row data as denominator.

Statistical significance p < 0.05 based on chi-square analysis.

Logistic regression analysis showed that individuals who had class III obesity were almost six times more likely to have received a diagnosis of severe obesity than those who had class II obesity (odds ratio [OR], 5.84; 95 % CI, 5.62–6.07). Other significant factors associated with having a diagnosis of severe obesity included Black/African American race (OR, 1.81; 95 % CI, 1.74–1.89) and number of comorbidities (OR, 1.37; 95 % CI, 1.35–1.39). Other factors such as female sex, age stratum, median household income, and number of primary encounters were significantly but modestly associated with a diagnosis of severe obesity (Table 4).Table 4 Adjusted multivariable logistic regression analysis of characteristics associated with diagnosis of severe obesity in adult patients 18–75 years old (2017–2019) within a large metropolitan health system in the US in the state of Michigan.

Covariate	Odds Ratio (95 % CI)	P value	
Class III obesity (Ref: Class II)	5.84 (5.62–6.07)	<0.001	
Female sex (Ref: Male)	1.04 (1.00–1.08)	<0.05	
Age stratum (Ref: 18–35 years)	0.87 (0.86–0.89)	<0.001	
Black/African American race (Ref: White)	1.81 (1.74–1.89)	<0.001	
Number of obesity-related comorbidities	1.37 (1.35–1.38)	<0.001	
Median household income stratum (Ref: <$25,000)	1.06 (1.04–1.08)	<0.001	
Number of primary care encounters stratum (Ref: 2–4 visits)	1.21 (1.18–1.24)	<0.001	
BMI, body mass index; Ref, reference; Class III Obesity, BMI≥40 kg/m2; Class II Obesity, BMI 35–39.9 kg/m2 with obesity-related comorbidities; Age Strata (13–35, 35–50, 51–64, ≥65 years); Median household income strata in US$ (<$25,000; $25,000 to <$50,000; $50,000 to <$75,000; $75,000 to <$100,000; ≥$100,000); Number of primary care encounters (2–4; 5–9; 10 or more).

Relative to individuals with BMI 30–39.9 kg/m2 and one comorbidity.

Odds ratios (OR) and 95 % confidence intervals derived from logistic regression. Fully adjusted models included obesity class, sex, age, race, number of comorbidities, median household income, and number of primary care encounters. Statistical significance p ≤ 0.05.

A second logistic regression analysis showed that patients who had been diagnosed with severe obesity during a primary care visit during the study period were almost three times more likely to receive an obesity-related intervention than those who did not (OR, 2.92; 95 % CI, 2.80–3.05). Black/African American (OR, 1.46; 95 % CI, 1.40–1.53) and female patients (OR, 1.47; 95 % CI, 1.41–1.54) were over 40 % more likely to be offered a medical intervention for obesity. All other demographic factors were significantly but modestly associated with having a medical intervention for obesity (Table 5). Of note, the number of comorbidities was negatively associated with having an obesity-related intervention (OR, 0.88; 95 % CI, 0.86–0.90).Table 5 Adjusted multivariable logistic regression analysis of characteristics associated with receiving a severe obesity intervention strategy in adult patients 18–75 years old (2017–2019) within a large metropolitan health system in the US in the state of Michigan.

Covariate	Odds Ratio (95 % CI)	P value	
Obesity diagnosis at a visit (Ref: No)	2.92 (2.80–3.05)	<0.001	
Class III obesity (Ref: Class II)	1.17 (1.12–1.22)	<0.001	
Female sex (Ref: Male)	1.47 (1.41–1.54)	<0.001	
Age stratum (Ref: 18–35 years)	0.72 (0.71–0.74)	<0.001	
Black/African American race/ethnicity	1.46 (1.40–1.53)	<0.001	
Number of obesity-related comorbidities	0.88 (0.86–0.90)	<0.001	
Median household income stratum (Ref: <$25,000)	1.10 (1.08–1.12)	<0.001	
Number of primary care encounters stratum (Ref: 2–4 visits)	1.37 (1.33–1.41)	<0.001	
BMI, body mass index; Ref, reference; Class III Obesity, BMI≥40 kg/m2; Class II Obesity, BMI 35–39.9 kg/m2 with obesity-related comorbidities; Age Strata (13–35; 35–50; 51–64; ≥65 years); Median household income strata in US$ (<$25,000; $25,000 to <$50,000; $50,000 to <$75,000; $75,000 to <$100,000; ≥$100,000); Number of primary care encounters (2–4; 5–9; 10 or more).

Relative to individuals with BMI 30–39.9 kg/m2 and one comorbidity.

Odds ratios (OR) and 95 % confidence intervals derived from logistic regression. Fully adjusted models included obesity diagnosis at visit, obesity class, sex, age, race, number of comorbidities, median household income, and number of primary care encounters. Statistical significance p ≤ 0.05.

Ms. No.: PMEDR-24–436.

4 Discussion

In this retrospective cross-sectional study of how patients were diagnosed with obesity and treated for obesity, over 50% of patients who had clinical characteristics indicative of obesity did not receive an obesity diagnosis during a medical visit within the study period. In our study, patients with class III obesity were much more likely to receive a diagnosis of severe obesity than individuals with class II obesity, suggesting that those with class II obesity may be particularly likely to be underdiagnosed. This finding is important, since we also saw that the presence of a severe obesity diagnosis was a key factor associated with whether or not a patient was offered an obesity-related medical intervention or treatment.

Extensive literature on severe obesity in relation to surgical procedures exists; however, limited studies have been done regarding the frequency or accuracy of diagnostic and treatment patterns for individuals with severe obesity (Centers for Medicare and Medicaid Services, 2019). To our knowledge, our study is one of only a few that have focused on the frequency and accuracy of receiving a diagnosis of and treatment for obesity per the HCC code. Previous studies have shown that the sensitivity of diagnostic coding for obesity is low, and coding errors might be a factor influencing results of studies that implemented code-based obesity diagnosis. However, whether diagnosis of severe obesity is prone to similar coding inaccuracies is not clear (George et al., 2017, Golinvaux et al., 2014).

In our population, Black patients and female patients were more likely to receive an obesity-specific treatment. However, we note that the OR describing the association of female sex with treatment was 1.04, and this finding may not be clinically relevant. But Black patients were almost twice as likely to receive an obesity-specific treatment in our study. Obesity disparities are known to exist for Black and Latino adults, and clinical trials for weight loss drugs often do not recruit representative populations (Herbozo et al., 2023). However, a study looking at anti-obesity medication use trends from 2010 to 2019 also showed that African American patients were more likely to receive anti-obesity pharmacotherapy, and the authors suggest that health care provider practice may be a major cause (Elangovan et al., 2021). Considering the growing popularity of newly approved GLP-1 anti-obesity drugs, understanding how both patient and provider knowledge and beliefs play a role in prescribing behaviors will be important.

4.1 Limitations

Multiple limitations are inherent to large-scale studies using retrospective data from patient databases. The use of ICD-10-CM codes in lieu of actual diagnoses may underestimate rates of diagnosis. Likewise, orders such as treatments and referrals often require association with an ICD-coded diagnosis, which could partially explain why diagnosis of obesity was such a strong potential predictor of treatment. However, the lack of such diagnoses also suggests a lack of treatment for severe obesity, highlighting the importance of a diagnosis. Interventions such as motivational interviewing, which may only be captured by visit diagnoses such as exercise counseling or dietary counseling, were also not captured. Other variables that may be associated with obesity diagnosis and treatment could not be obtained from our database, which may have influenced our results. These include but are not limited to provider specialty, provider level of training, and patients’ preferred language. While considering these factors may or may not have altered our results, they may be important to include in future studies to provide insight into the specific types of providers and patients who may disproportionately benefit from education or a focused intervention. Also, contraindications for treatment were not considered in our models, which may have led to an underestimation of the treatments that were provided. For medications, these would include absolute and relative contraindications to treatment, such as use of other medications or presence of other diseases. For referrals, these would include absolute and relative contraindications for surgery, intensive diet, or intensive exercise regimen. This would have produced innumerable potential parameters to be considered in the models, many of which were unavailable in the database or may have been subject to a provider’s interpretation of the clinical context. As such, these were not included in the final models.

Finally, in our study and in the literature at large, analyzing treatment of severe obesity with pharmaceutical interventions for patients with diabetes presents a logistical challenge. Medical treatments for severe obesity in our study were largely divided into two categories: obesity-specific medications and diabetes medications that promote weight loss and may be prescribed for patients without diabetes. By removing treatment with a diabetes medication for patients with diabetes as a variable, we may have underestimated the number of pharmaceutical interventions that were intended to treat obesity. For example, treating patients who have diabetes with metformin, a first-line diabetes medication, could also be considered as intent to treat severe obesity. Although inclusion of this treatment and group was considered, we felt that it might markedly overestimate the rate of treatment for severe obesity. While we could have removed all patients with diabetes from the analysis, we felt that this would have removed a large segment of the study population, as diabetes is not only highly prevalent in patients with clinical characteristics of obesity, but also is one of the risk factors in the definition of class II obesity. Importantly, patients included in our study had been treated before the most popular of the latest GLP-1 drugs had been approved for weight loss, and these drugs now represent a paradigm shift in the medical approach to treating obesity and diabetes. Thus, prospective studies on rates of obesity diagnosis and treatment in this new GLP-1 era are needed to determine how these rates may change.

Lastly, the translation of electronic medical record data into meaningful clinical definitions relies heavily on the clinical team’s perspectives on common clinical practices. Therefore, bias is inherent when using visit diagnoses and orders as variables and outcomes.

5 Conclusions

Our findings suggest that a potentially high-yield intervention for treating patients with severe obesity would be to increase the rate of severe obesity diagnoses, especially for individuals with class II obesity. This might be achieved through policies that emphasize screening for severe obesity within this population, as the diagnosis requires the presence of at least one comorbidity, and therefore extends it beyond a 1-step, simple definition. Within the context of the electronic medical record, an electronic prompt suggesting an obesity diagnosis and potential treatments might encourage clinicians to address the possibility of diagnosing severe obesity more frequently. System-wide registries similar to those used for diabetes, tobacco use, and hypertension would allow changes to be instituted at the system level. Finally, patient-driven efforts towards treatment may increase if patients are educated about the criteria for diagnosis, associated morbidity and mortality, and potential treatment options.

Funding

This research did not receive any specific grant from funding agents in the public, commercial, or not-for profit sectors.

CRediT authorship contribution statement

Raphael Szymanski: Writing – original draft, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Megha Abraham: Writing – original draft, Project administration, Methodology, Formal analysis, Data curation, Conceptualization. William Childs: Writing – original draft, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Kristina Le: Writing – original draft, Project administration, Methodology, Formal analysis, Data curation, Conceptualization. Christopher Velez: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Ivana Vaughn: Writing – review & editing, Validation, Supervision, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation. Lois Lamerato: Writing – review & editing, Writing – original draft, Supervision, Software, Resources, Project administration, Methodology, Formal analysis, Data curation, Conceptualization. Katarzyna Budzynska: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The data that has been used is confidential.

Acknowledgements

We would like to thank Karla D. Passalacqua, PhD, and Stephanie Stebens, MSIS, at Henry Ford Health for editorial and technical assistance.
==== Refs
References

Ackermann R.T. Finch E.A. Brizendine E. Zhou H. Marrero D.G. Translating the Diabetes Prevention Program into the community. The DEPLOY Pilot Study Am. J. Prev. Med. 35 2008 357 363 10.1016/j.amepre.2008.06.035 18779029
Agborsangaya C.B. Ngwakongnwi E. Lahtinen M. Cooke T. Johnson J.A. Multimorbidity prevalence in the general population: the role of obesity in chronic disease clustering BMC Public Health. 13 2013 1161 10.1186/1471-2458-13-1161 24325303
Berrington de Gonzalez A. Hartge P. Cerhan J.R. Flint A.J. Hannan L. MacInnis R.J. Body-mass index and mortality among 1.46 million white adults N. Engl. J. Med. 363 2010 2211 2219 10.1056/NEJMoa1000367 21121834
Centers for Disease Control and Prevention Overweight and obesity: why it matters accessed 29 April 2024 https://www.cdc.gov/obesity/about-obesity/why-it-matters.html 2022
Centers for Medicare and Medicaid Services Bariatric surgery for the treatment of morbid obesity (CAG-00250R2) accessed 29 April 2024 https://www.cms.gov/medicare-coverage-database/view/ncacal-tracking-sheet.aspx?ncaid=258 2012
Centers for Medicare and Medicaid Services Billing and coding: surgical management of morbid obesity (A57145) for documentation requirements that apply to the reasonable and necessary provisions outlined in this LCD accessed 29 April 2024 https://www.cms.gov/medicare-coverage-database/view/lcd.aspx?lcdId=33411&ver=29 2019
Ciciurkaite G. Moloney M.E. Brown R.L. The incomplete medicalization of obesity: physician office visits, diagnoses, and treatments, 1996–2014 Public Health Rep. 134 2019 141 149 10.1177/0033354918813102 30794761
Ciemins EL, Joshi V, Cuddeback JK, Kushner RF, Horn DB, Garvey WT. Diagnosing Obesity as a First Step to Weight Loss: An Observational Study. Obesity (Silver Spring). 2020 Dec;28(12):2305-2309. doi: 10.1002/oby.22954. Epub 2020 Oct 7. PMID: 33029901; PMCID: PMC7756722.
Ciemins E, Joshi V, Horn D, Nadglowski J, Ramasamy A, Cuddeback J. Measuring What Matters: Beyond Quality Performance Measures in Caring for Adults with Obesity. Popul Health Manag. 2021 Aug;24(4):482-491. doi: 10.1089/pop.2020.0109. Epub 2020 Nov 11. PMID: 33180000; PMCID: PMC8403197.
Dai H. Alsalhe T.A. Chalghaf N. Ricco M. Bragazzi N.L. Wu J. The global burden of disease attributable to high body mass index in 195 countries and territories, 1990–2017: an analysis of the Global Burden of Disease Study PLoS Med. 17 2020 e1003198 32722671
de Heer H. Kinslow B. Lane T. Tuckman R. Warren M. Only 1 in 10 patients told to lose weight seek help from a health professional: a nationally representative sample Am. J. Health Promot. 33 2019 1049 1052 10.1177/0890117119839904 30922060
Dixon J.B. The effect of obesity on health outcomes Mol. Cell. Endocrinol. 316 2010 104 108 10.1016/j.mce.2009.07.008 19628019
D'Souza M.J. Bautista R.C. Wentzien D.E. Data talks: obesity-related influences on US mortality rates Res Health Sci. 3 2018 65 78 10.22158/rhs.v3n3p65 30079383
Elangovan A. Shah R. Smith Z.L. Pharmacotherapy for Obesity-Trends Using a Population Level National Database Obes Surg. 31 3 2021 Mar 1105 1112 10.1007/s11695-020-04987-2 Epub 2020 Sep 28 PMID: 32986169 32986169
Fitzpatrick S.L. Stevens V.J. Adult obesity management in primary care, 2008–2013 Prev. Med. 99 2017 128 133 10.1016/j.ypmed.2017.02.020 28232100
Flegal K.M. Kit B.K. Orpana H. Graubard B.I. Association of all-cause mortality with overweight and obesity using standard body mass index categories: a systematic review and meta-analysis JAMA. 309 2013 71 82 10.1001/jama.2012.113905 23280227
Fortin M. Lapointe L. Hudon C. Vanasse A. Ntetu A.L. Maltais D. Multimorbidity and quality of life in primary care: a systematic review Health Qual. Life Outcomes. 2 2004 51 10.1186/1477-7525-2-51 15380021
Garin N. Koyanagi A. Chatterji S. Tyrovolas S. Olaya B. Leonardi M. Global multimorbidity patterns: a cross-sectional, population-based, multi-country study J. Gerontol. A Biol. Sci. Med. Sci. 71 2016 205 214 10.1093/gerona/glv128 26419978
Gbd Obesity Collaborators, Afshin, A., Forouzanfar, M.H., Reitsma, M.B., Sur, P., Estep, K., et al., 2017. Health effects of overweight and obesity in 195 countries over 25 years N. Engl. J. Med. 377 2015 13 27 10.1056/NEJMoa1614362
Gbd Risk Factors Collaborators, 2020. Global burden of 87 risk factors in 204 countries and territories, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019 Lancet. 396 2019 1223 1249 10.1016/s0140-6736(20)30752-2
George J. Newman J.M. Ramanathan D. Klika A.K. Higuera C.A. Barsoum W.K. Administrative databases can yield false conclusions-an example of obesity in total joint arthroplasty J. Arthroplasty. 32 2017 S86 S90 10.1016/j.arth.2017.01.052 28246011
Global BMI Mortality Collaboration Body-mass index and all-cause mortality: individual-participant-data meta-analysis of 239 prospective studies in four continents Lancet. 388 2016 776 786 10.1016/s0140-6736(16)30175-1 27423262
Golinvaux N.S. Bohl D.D. Basques B.A. Fu M.C. Gardner E.C. Grauer J.N. Limitations of administrative databases in spine research: a study in obesity Spine J. 14 2014 2923 2928 10.1016/j.spinee.2014.04.025 24780248
Harwood A. Carter D. Eliott J. A public health framework for reducing stigma: the example of weight stigma J. Bioeth. Inq. 19 2022 511 520 10.1007/s11673-022-10199-3 35857213
Haslam D.W. James W.P. Obesity. Lancet. 366 2005 1197 1209 10.1016/S0140-6736(05)67483-1 16198769
Herbozo S, Brown KL, Burke NL, LaRose JG. A Call to Reconceptualize Obesity Treatment in Service of Health Equity: Review of Evidence and Future Directions. Curr Obes Rep. 2023 Mar;12(1):24-35. doi: 10.1007/s13679-023-00493-5. Epub 2023 Feb 2. PMID: 36729299; PMCID: PMC9894524.
Jackson C.A. Dobson A. Tooth L. Mishra G.D. Body mass index and socioeconomic position are associated with 9-year trajectories of multimorbidity: a population-based study Prev. Med. 81 2015 92 98 10.1016/j.ypmed.2015.08.013 26311587
Kivimäki M. Kuosma E. Ferrie J.E. Luukkonen R. Nyberg S.T. Alfredsson L. Overweight, obesity, and risk of cardiometabolic multimorbidity: pooled analysis of individual-level data for 120 813 adults from 16 cohort studies from the USA and Europe Lancet Public Health. 2 2017 e277 e285 10.1016/s2468-2667(17)30074-9 28626830
Kivimaki M. Strandberg T. Pentti J. Nyberg S.T. Frank P. Jokela M. Body-mass index and risk of obesity-related complex multimorbidity: an observational multicohort study Lancet Diabetes Endocrinol. 10 2022 253 263 10.1016/S2213-8587(22)00033-X 35248171
Klein S. Wadden T. Sugerman H.J. AGA technical review on obesity Gastroenterology. 123 2002 882 932 10.1053/gast.2002.35514 12198715
Lewis K.H. Edwards-Hampton S.A. Ard J.D. Disparities in treatment uptake and outcomes of patients with obesity in the USA Curr. Obes. Rep. 5 2016 282 290 https://di.org/10.1007/s13679-016-0211-1 27023070
National Center for Health Statistics Health, United States, 2015: with special feature on racial and ethnic health disparities 2016 National Center for Health Statistics Hyattsville, MD
Noria S.F. Grantcharov T. Biological effects of bariatric surgery on obesity-related comorbidities Can. J. Surg. 56 2013 47 57 10.1503/cjs.036111 23351555
Nyberg S.T. Batty G.D. Pentti J. Virtanen M. Alfredsson L. Fransson E.I. Obesity and loss of disease-free years owing to major non-communicable diseases: a multicohort study Lancet Public Health. 3 2018 e490 e497 10.1016/S2468-2667(18)30139-7 30177479
Pasco J.A. Brennan S.L. Kotowicz M.A. Morbid obesity in women on the rise: an observational, population-based study BMC Public Health. 13 2013 290 10.1186/1471-2458-13-290 23547911
Prospective Studies Collaboration Body-mass index and cause-specific mortality in 900 000 adults: collaborative analyses of 57 prospective studies Lancet. 373 2009 1083 1096 10.1016/S0140-6736(09)60318-4 19299006
Puhl R.M. Heuer C.A. Obesity stigma: important considerations for public health Am. J. Public Health. 100 2010 1019 10128 10.2105/AJPH.2009.159491 20075322
Strandberg T.E. Sirola J. Pitkala K.H. Tilvis R.S. Strandberg A.Y. Stenholm S. Association of midlife obesity and cardiovascular risk with old age frailty: a 26-year follow-up of initially healthy men Int. J. Obes. (lond.) 36 2012 1153 1157 10.1038/ijo.2012.83 22614054
Sturm R. Increases in morbid obesity in the USA: 2000–2005 Public Health. 121 2007 492 496 10.1016/j.puhe.2007.01.006 17399752
Suissa K. Schneeweiss S. Lin K.J. Brill G. Kim S.C. Patorno E. Validation of obesity-related diagnosis codes in claims data Diabetes Obes. Metab. 23 2021 2623 2631 10.1111/dom.14512 34338404
Tucker S. Bramante C. Conroy M. Fitch A. Gilden A. Wittleder S. The most undertreated chronic disease: addressing obesity in primary care settings Curr Obes Rep. 10 2021 396 408 10.1007/s13679-021-00444-y 34297343
Valero-Elizondo J. Aneni E.C. Osondu C.U. Grandhi G.R. Virani S.S. Nasir K. Gaps in provider lifestyle counseling and its adherence among obese adults with prediabetes and diabetes in the United States Prev. Med. 129 2019 105815 10.1016/j.ypmed.2019.105815
Walls H.L. Wolfe R. Haby M.M. Magliano D.J. de Courten M. Reid C.M. Trends in BMI of urban Australian adults, 1980–2000 Public Health Nutr. 13 2010 631 638 10.1017/S1368980009991455 19772687
World Health Organization Obesity: preventing and managing the global epidemic: report of a WHO consultation (WHO Technical Report Series 894) 2000 World Health Organization Geneva, Switzerland
World Health Organization Obesity and overweight accessed 29 April 2024 https://www.who.int/news-room/fact-sheets/detail/obesity-and-overweight 2021
