Healthcare utilization and expenditures among adults with type 2 diabetes mellitus and comorbid psychological distress

TL;DR

This study finds that psychological distress in adults with type 2 diabetes in rural China is linked to increased healthcare use and costs.

Contribution

This is the first study to evaluate the association between psychological health and healthcare expenditures in rural Chinese adults with T2DM.

Findings

Higher diabetes distress scores are associated with increased outpatient visits, inpatient use, and outpatient costs.

Depressive symptoms are linked to higher outpatient visits, inpatient use, and total medical expenditures.

Psychological screening and care interventions are suggested to reduce healthcare burdens in this population.

Abstract

Depressive symptoms and diabetes distress are associated with adverse outcomes in adults with type 2 diabetes mellitus (T2DM). However, no prior studies evaluated the association between healthcare expenditures, utilization, and psychological health in adults with T2DM in rural areas of China. The aim of this study was to explore the association between psychological health—specifically depressive symptoms and diabetes distress—and healthcare utilization and expenditures in this population.

This cross-sectional study was conducted in 15 rural health clinics in Jiangsu Province, China, and involved 843 adults with T2DM. Psychological health was assessed using the 17-item Diabetes Distress Scale (DDS-17) and the 10-item Center for Epidemiologic Studies Depression Scale (CESD-10). Healthcare utilization and expenditures were evaluated with negative binomial, logistic, and ordinary least squares (OLS) regression models.

Of 843 participants, a total of 32.62% of the sample were comorbid with distress, depression, or both. Each 1-point increase in diabetes distress score was associated with a 44% increase in the expected number of annual outpatient visits (IRR = 1.44, P<0.05), a 52% increase in the odds of inpatient service utilization (OR = 1.52, P <0.05), and a 43.3% increase in outpatient costs (eβ-1 = 0.433, P<0.05). For each point increase in depressive symptom score, the expected number of annual outpatient visits increased by 4% (IRR = 1.04, P<0.01), the odds of inpatient service utilization increased by 4% (OR = 1.04, P <0.01), and total medical expenditures increased by 2% (eβ-1 = 0.02, P<0.05).

Our study observed that Diabetes distress is associated with higher outpatient costs, while diabetes with depressive symptoms is associated with higher total medical expenditures. These findings suggest that psychological screening and care interventions for adults with diabetes may be essential.

Tables

Table 1. Basic characteristics in adults with type 2 diabetes mellitus (N = 843).

Variable	Mean ± SD/M± QR/n (%)	Diabetes distress score	P	Depressive symptoms score	p
Gender
Male	282 (33.45)	1.41 ± 0.40	0.001†	5.15 ± 5.79	< 0.001‡
Female	561 (66.55)	1.51 ± 0.45		6.98 ± 5.88
Age (year)	66.08 ± 9.10
<60	199 (23.61)	1.50 ± 0.44	0.019*	5.53 ± 5.52	0.005†
60-70	328 (38.91)	1.52 ± 0.47		7.15 ± 6.26
> 70	316 (37.94)	1.42 ± 0.41		6.09 ± 5.70
Education
Illiterate	332 (39.38)	1.53 ± 0.45	0.005†	7.37 ± 5.99	< 0.001‡
Primary	252 (29.89)	1.48 ± 0.43		6.95 ± 6.21
Secondary	190 (22.54)	1.43 ± 0.45		4.59 ± 5.00
Post-Secondary	69 (8.19)	1.35 ± 0.34		4.33 ± 5.21
Marital status
Married	695 (82.44)	1.48 ± 0.44	0.947	7.17 ± 6.17	0.070
Unmarrieda	148 (17.56)	1.48 ± 0.44		6.2 ± 5.85
Comorbiditiesb
Cardiovascular diseases	135 (16.01)	1.50 ± 0.43	0.505	6.98 ± 5.72	1.192
Hypertension	490 (58.13)	1.49 ± 0.43	0.444	6.57 ± 5.91	0.259
Cerebrovascular diseases	72 (8.54)	1.58 ± 0.41	0.030*	8.82 ± 6.62	< 0.001‡
Other diseases	432 (51.25)	1.53 ± 0.47	0.001†	6.98 ± 6.16	0.002†
Baseline FPG (mmol/L)	8.16 ± 2.51
FPG≥7.0	524 (62.16)	1.47 ± 0.43	0.7.09	6.39 ± 6.06	0.904
FPG<7.0	319 (37.84)	1.48 ± 0.45		6.34 ± 5.67
Duration of diabetes	8.56 ± 5.76
Diabetes distress scores	1.48 ± 0.44
Moderate or severe distress (DDS≥2)	105 (12.46)	-	-	-	-
Depression scores	6.37 ± 5.91
High Depressive Symptoms(CESD≥10)	212 (25.15)	-	-	-	-
Utilization of medical services
Number of outpatient visits per year	0 ± 2	-	-	-	-
Utilization of inpatient services	249 (29.54)	-	-	-	-
Healthcare expenditures(CNY)
Outpatient costs	700 ± 1700	-	-	-	-
Drug costs	1000 ± 2100	-	-	-	-
Inpatient costs	6000 ± 10500	-	-	-	-
Total medical expenditures	3500 ± 7950	-	-	-	-

Table 2. Regression analysis of psychological health on healthcare utilization in adults with type 2 diabetes mellitus.

Variable	Group	Model 1	Model 2	Model 3	Model 4
		IRR/OR (95%CI)	IRR/OR (95%CI)	IRR/OR (95%CI)	IRR/OR (95%CI)
Annual number of outpatient visits
DDS scores		1.52 (1.10-2.10)*	1.44 (1.05-1.98)*
CES-D scores				1.04 (1.01-1.06)†	1.04 (1.02-1.07)†
Age			0.98 (0.96-0.99)*		0.98 (0.96-1.00)*
Gender (ref: Male)	Female		0.87 (0.60-1.25)		0.84 (0.58-1.20)
Education (ref: illiterate)	Primary		0.85 (0.57-1.26)		0.82 (0.56-1.21)
	Secondary		0.91 (0.58-1.43)		1.05 (0.67-1.65)
	Post-Secondary		0.94 (0.48-1.84)		1.01 (0.51-1.99)
Marital status (ref: Unmarried)	Married		1.42 (0.93-2.17)		1.43 (0.94-2.18)
Duration of diabetes			1.02 (1.00-1.05)		1.02 (00.99-1.05)
Cardiovascular disease (ref: Without)	With		1.68 (1.13-2.50)*		1.70 (1.15-2.52)*
Hypertension (ref: Without)	With		1.26 (0.92-1.72)		1.28 (0.94-1.76)
cerebrovascular disease (ref: Without)	With		1.88 (1.12-3.18)*		1.95 (1.16-3.27)*
With other chronic diseases (ref: Without)	With		1.61 (1.20-2.17)†		1.70 (1.27-2.29)‡
Baseline FPG			0.98 (0.93-1.04)		0.97 (0.92-1.03)
Utilization of inpatient services
DDS scores		1.61 (1.16-2.23)†	1.52 (1.08-2.15)*
CES-D scores				1.05 (1.03-1.08)‡	1.04 (1.01-1.07)†
Age			1.02 (1.00-1.04)		1.02 (1.00-1.34)
Gender (ref: Male)	Female		0.71 (0.49-1.02)		0.70 (0.48-1.01)
Education (ref: illiterate)	Primary		1.07 (0.73-1.57)		1.06 (0.72-1.55)
	Secondary		0.88 (0.55-1.41)		0.93 (0.58-1.49)
	Post-Secondary		0.71 (0.35-1.44)		0.72 (0.36-1.45)
Marital status (ref: Unmarried)	Married		1.14 (0.75-1.73)		1.17 (0.77-1.78)
Duration of diabetes			1.04 (1.01-1.07)†		1.04 (1.01-1.07)†
Cardiovascular disease (ref: Without)	With		1.19 (0.79-1.79)		1.19 (0.79-1.80)
Hypertension (ref: Without)	With		1.11 (0.80-1.54)		1.12 (0.81-1.55)
cerebrovascular disease (ref: Without)	With		2.03 (1.22-3.37)†		1.94 (1.16-3.24)*
With other chronic diseases (ref: Without)	With		1.69 (1.23-2.33)†		1.70 (1.23-2.34)†
Baseline FPG			0.99 (0.93-1.05)		0.98 (0.92-1.05)

Table 3. Regression analysis of distress and depressive symptoms on healthcare expenditures in adults with type 2 diabetes mellitus.

Variable	Group	Outpatient costs		Drug costs		Total medical expenditures
		Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
		β (95%CI)	β (95%CI)	β (95%CI)	β (95%CI)	β (95%CI)	β (95%CI)
Diabetes distress
DDS scores		0.35 (0.02-0.67)*	0.36 (0.03-0.68)*	0.13 (-0.28-0.54)	0.19 (-0.24-0.62)	0.23 (-0.06-0.52)	0.26 (-0.04-0.56)
CES-D scores
Age			0.01 (-0.02-0.02)		0.01 (-0.02-0.03)		0.01 (-0.01-0.03)
Gender (ref: Male)	Female		-0.02 (0.39-0.35)		0.14 (-0.32-0.60)		-0.15 (-0.48-0.18)
Education (ref: illiterate)	Primary		-0.10 (-0.49-0.29)		0.12 (-0.36-0.59)		0.21 (-0.14-0.56)
	Secondary		0.42 (-0.03-0.86)		0.22 (-0.32-0.77)		0.04 -0.38-0.46)
	Post-Secondary		0.16 (-0.48-0.80)		0.11 (-0.64-0.86)		0.08 (-0.54-0.69)
Marital status (ref: Unmarried)	Married		0.02 (-0.41-0.44)		0.19 (-0.34-0.73)		0.08 (-0.30-0.45)
Duration of diabetes			0.03 (-0.00-0.15)		0.03 (0.01-0.06)*		0.02 (-0.01-0.04)
Cardiovascular disease (ref: Without)	With		0.21 (-0.17-0.59)		-0.25 (-0.72-0.23)		0.07 (-0.29-0.43)
Hypertension (ref: Without)	With		0.06 (-0.26-0.38)		0.20 (-0.19-0.59)		0.02 (-0.27-0.32)
cerebrovascular disease (ref: Without)	With		0.08 (-0.39-0.55)		-0.23 (-0.79-0.33)		0.24 (-0.20-0.69)
With other chronic diseases (ref: Without)	With		0.03 (-0.29-0.35)		-0.08 (-0.46-0.31)		-0.01 (-0.29-0.28)
Baseline FPG			-0.05 (-0.11-0.01)		-0.01 (-0.09-0.06)		-0.02 (-0.07-0.04)
Depressive symptoms
DDS scores
CES-D scores		0.02 (-0.00-0.04)	0.02 (-0.01-0.05)	0.02 (-0.01-0.05)	0.02 (-0.01-0.05)	0.03 (0.01-0.05)*	0.02 (0.01-0.05)*
Age			0.01 (-0.02-0.02)		0.01 (-0.02-0.03)		0.01 (-0.01-0.03)
Gender (ref: Male)	Female		-0.02 (-0.39-0.35)		0.14 (-0.32-0.60)		-0.15 (-0.48-0.18)
Education (ref: illiterate)	Primary		-0.09 (-0.48-0.29)		0.15 (-0.33-0.62)		0.19 (-0.16-0.54)
	Secondary		0.46 (0.01-0.91)*		0.29 (-0.27-0.84)		0.08 (-0.34-0.51)
	Post-Secondary		0.15 (-0.49-0.79)		0.14 (-0.61-0.89)		0.09 (-0.52-0.70)
Marital status (ref: Unmarried)	Married		0.04 (-0.39-0.46)		0.19 (-0.34-0.72)		0.09 (-0.29-0.46)
Duration of diabetes			0.02 (-0.01-0.05)		0.03 (-0.01-0.06)		0.02 (-0.01-0.04)
Cardiovascular disease (ref: Without)	With		0.21 (-0.16-0.59)		-0.24 (-0.72-0.24)		0.08 (-0.28-0.44)
Hypertension (ref: Without)	With		0.07 (-0.26-0.39)		0.19 (-0.20-0.58)		0.03 (-0.27-0.32)
DDS scores
cerebrovascular disease (ref: Without)	With		0.09 (-0.38-0.56)		-0.22 (-0.77-0.34)		0.23 (-0.22-0.67)
With other chronic diseases (ref: Without)	With		0.06 (-0.26-0.37)		-0.07 (-0.45-0.31)		0.01 (-0.27-0.30)
Baseline FPG			-0.06 (-0.12-0.01)		-0.02 (-0.09-0.06)		-0.02 (-0.07-0.04)

Full text

Introduction

1

Diabetes mellitus (DM) is a complex chronic disease that is one of the most important public health problems due to its high prevalence and cost of healthcare (1). DM and its complications represent a huge burden on healthcare service utilization and expenditure, resulting in global healthcare expenditures of 966 billion USD in 2021 (1–3). Diabetes often combines with other complications, such as diabetic nephropathy (4), cardiomyopathy (5), retinopathy (6), osteoporosis (7), and psychiatric symptoms (8). Depression and diabetes distress are two major forms of psychological distress frequently experienced by adults with type 2 diabetes mellitus (T2DM), increasing the risk of mortality, diabetes-related complications, and poor self-management and quality of life (9, 10). According to systematic review studies, the overall estimated prevalence of depression in people with T2DM is 19%-28% (11, 12). Studies in the U.S., Canada, Denmark, and Bangladesh reported that diabetes distress affects about one-third of adults with T2DM (13–16). In addition, depression and diabetes distress can also overlap, with approximately 4.5% of adults with diabetes screening positive for both (17). Depression and diabetes intersect in pathogenesis, such as inflammation (18, 19), oxidative stress (20, 21), metabolic disorders (18, 22), and others. Moreover, the relationship between psychological health and T2DM can be bidirectional (23–25). Therefore, depression and diabetes distress may lead to poorer glycemic control. The economic burden of diabetes, especially comorbid psychological and emotional disorders, deserves further study.

The increased prevalence rates of comorbid psychological distress raise the issue of the association of psychological distress and healthcare expenditures and utilization (26). Studies based on the National Health Insurance Research Database (NHIRD) from Taiwan revealed that healthcare utilization and expenditures for persons with DM comorbid mental disorders were significantly higher than non-comorbid individuals (8, 27). Additionally, one study with 400,495 individuals with DM in the United States found that major depressive disorder contributed to an increase in the excess medical expenditures among persons with diabetes by increasing inpatient care, outpatient care, and medication expenditures (28). Another recent population-based cohort study in Canada suggested that comorbid mental disorders can lead to higher healthcare costs for persons with chronic diseases through increased hospitalization, emergency department visits, and length of hospital stay (29). However, studies on diabetes, mental health, and health care spending and utilization are inconsistent. A previous cross-sectional study in the German National Health Interview and Examination Survey (GNHIES) indicated no increase in the length of hospitalization or visits to mental health specialists among adults with DM comorbid mental disorders (30). Meanwhile, fewer prior studies have examined the healthcare costs associated with psychological distress in Chinese adults with T2DM. Additionally, less attention has been paid to differences in various psychological distresses. Hence, the association of psychological distress with healthcare expenditures and utilization in adults with T2DM in the context of the increasing prevalence of co-existing diabetes and psychological distress was worthy of further study.

Structured interviews for the clinical diagnosis of depression are impractical in large-scale surveys (31). The definition of depression is based on symptoms forming a syndrome and causing functional impairment (32). This study aimed to estimate the healthcare expenditures and healthcare utilization associated with depressive symptoms (measured by the CESD-10) and diabetes distress (measured by the DDS-17) in Chinese adults with T2DM. Specifically, the following research questions will be addressed: (I) To what extent are comorbid depressive symptoms associated with increased healthcare expenditures and utilization in adults with T2DM? (II) To what extent is comorbid diabetes distress associated with increased healthcare expenditures and utilization in persons with T2DM? (III) How do the magnitude and pattern of associations with healthcare utilization and expenditures differ between diabetes distress and depressive symptoms? (IV) Do these associations persist after adjusting for key demographic characteristics?

Materials and methods

2

Study design and participants

2.1

We designed a cross-sectional study at baseline in 15 rural health clinics of Jiangsu Province in southeastern China. The study sample consisted of 5 health clinics located in rural communities in Tongzhou district, Nantong city, and 10 health clinics in Dafeng district, Yancheng city. Participants with type2 diabetes were randomly drawn from physician-managed registries. Data were collected through face-to-face, interviewer -assisted questionnaires administered on-site at the community health centers by trained local surveyors. From January to March 2019, we identified from recruited adults who (I) had been diagnosed with T2DM for at least 1 year by physicians; (II) were aged 45 years or above; and (III) were able to communicate in Mandarin. We excluded adults with severe diabetes complications or functional deficits, schizophrenia, severe dementia, or bipolar disorder, and a traumatic experience in the past 6 months. Eventually, 900 individuals from 15 clinics participated in the study, and 843 valid questionnaires were collected, with a response rate of 93.67%.

Measures

2.2

Predictor variables

2.2.1

Diabetes distress

2.2.1.1

Diabetes distress was assessed with the 17-item Diabetes Distress Scale (DDS-17) (33). Participants rated the degree to which they felt the social and medical care distress related to diabetes and its management. Item responses were based on a Likert scale ranging from 1 (not a problem) to 6 (a severe problem), with higher total scores reflecting greater diabetes distress. A mean score ≥2 serves as the cut-point for moderate or greater diabetes distress, indicating a level of distress worthy of clinical attention (34). The DDS-17 was shown to have good psychometric properties for internal consistency reliability (α=0.90) and test-retest reliability coefficient (α=0.74) in Chinese adults with type 2 diabetes mellitus (35). The DDS-17 demonstrated good reliability in the present study (α = 0.87).

Depressive symptoms

2.2.1.2

Depressive symptoms were assessed with the short form 10-item Center for Epidemiologic Studies Depression Scale (CESD-10) (36). It shows the extent to which individuals experienced depressive symptoms during the prior week and has been validated as a tool to identify individuals at high risk of depression with wide applicability in the general population. The CESD-10 was self-rated by T2DM adults on a 4-point Likert scale, where 0 = “rarely or none during the time (<1 day)” to 3 = “most of the time (5–7 days).” The scores of the CESD-10 range from 1 to 30, with higher scores indicating more severe depressive symptoms, and individuals who had scores≥10 were classified as having high depressive symptoms. The CESD-10 has high reliability and validity in Chinese middle-aged and elderly populations (α = 0.82) (37). Cronbach’s ɑ=0.85 indicated that the scale was reliable.

Dependent variables

2.2.2

Healthcare utilization

2.2.2.1

Healthcare service utilization was measured as the annual number of outpatient visits (preventive, acute care, and emergency department) and the utilization of inpatient services and length of stay per year by a self-designed questionnaire. Preventive visits included general check-ups and visits for immunizations. Acute visits included all other outpatient visits not categorized as preventive. The participants were asked to answer three questions regarding their status of healthcare utilization: “Have you had a doctor visit or other healthcare professional in the past year?” “Have you had any outpatient visits or emergency department visits in the past year, including preventive and acute care?” and “Have you been hospitalized for any length of time in the past year, and if so, report your length of stay?”

Healthcare expenditures

2.2.2.2

The healthcare expenditure refers to all direct payments made for healthcare services and products within a calendar year, including costs for outpatient services, inpatient services, and drugs. Three expenditure outcomes were assessed: outpatient costs, drug costs, and total medical expenditures. Total medical expenditures summed outpatient, inpatient, and drug costs incurred for the calendar year. The costs associated with primary, specialty, and ancillary care visits were included in the outpatient costs. Inpatient costs comprised those related to admissions with overnight stays, admissions without overnight stays, and emergency department visits that occurred right before an admission. Drug costs included prescription medicines, over-the-counter medicines, medical devices, and dietary supplements. Expenditure data were based primarily on participant self-reports. For costs that were difficult to recall precisely, some participants supplemented their recall by presenting electronic or paper fee statements from healthcare providers. Participants were instructed to report the total amount paid for healthcare, which included both third-party-payer payments and out-of-pocket payments.

Control variables

2.2.3

Covariates in this study include demographic characteristics: gender (male and female), age group (<60, 60-70, and >70), education (illiterate, primary school, secondary school, post-secondary school), marital status (married/unmarried), and features involving clinical diabetes: duration of diabetes and comorbidities. Comorbidities were based on self-reported diagnoses in the past 12 months, including cardiovascular diseases, hypertension, cerebrovascular diseases, and other chronic conditions.

Statistical analyses

2.3

Statistical analyses were conducted with STATA software version 16. In the main analyses, the psychological variables were treated as continuous. Descriptive statistics with t-tests and χ2 tests were used to compare differences in the study sample of psychological variables in continuous forms (Table 1). Then, multivariable regression models with negative binomial distributions were used to examine the association of psychological variables with the annual number of outpatient visits, and logistic regression models were used to explore the utilization of inpatient services in adults with T2DM (Table 2). Model 1,3 only explored the relationship between psychological variable scores and the annual number of outpatient visits and utilization of inpatient services. Model 2,4 further examined the relationship between psychological variables scores and outpatient and inpatient utilization by adding demographic characteristics. The ordinary least squares (OLS) regression models and log links were used to estimate the association of psychological variables with healthcare expenditures, including outpatient costs, drug costs, inpatient costs, and total medical expenditures (Table 3). Model 1,3,5 only explored the relationship between psychological variable scores and healthcare expenditures. Model 2,4,6 further examined the relationship between psychological variables scores and healthcare expenditures by adding demographic characteristics. Age, gender, education, marital status, duration of diabetes, chronic medical conditions, and baseline fasting plasma glucose level are covariates, and they are controlled for in the models. All statistical tests were two-tailed, with α set to 0.05. We tested the collinearity issues, and VIF showed that there was no such issue (VIF<5) (38).

In additional analyses, we re-coded the psychological scale scores into categorical variables based on established clinical cut-offs. This created four mutually exclusive groups: (1) Diabetes mellitus (DDS<2, CESD<10), (2) Comorbid diabetes distress (DDS<2, CESD<10), (3) Comorbid depressive symptoms (DDS<2, CESD≥10), and (4) Comorbid diabetes distress and depressive symptoms (DDS≥2, CESD≥10). The same models were applied to the categorical variables to estimate the association of each group relative to the reference group with the outcomes.

Results

3

Participants characteristics

3.1

Of 843 participants with T2DM, the mean age was 66.8 ± 9.10 years, and 66.55% were women, and the majority(82.44%) were married (Table 1). 60.62% had primary education or higher. 74.97% had T2DM for more than 5 years. Hypertension (58.13%) was the most common comorbidity. The average scores of diabetes distress and depressive symptoms were 1.48 (SD 0.44) and 6.37 (SD 5.91), respectively. The rate of inpatient service utilization was 29.54%. The median outpatient cost, median drug cost, median inpatient cost, and median total expenditure were ￥700, ￥1000, ￥6000, and ￥3500, respectively.

Association of psychological distress with healthcare utilization

3.2

Table 2 presented the results of the regression analysis of psychological health on annual number of outpatient visits and utilization of inpatient services. Model 1 showed that for each 1-point increase in the diabetes distress score, the expected number of annual outpatient visits increased by 52% (incidence rate ratio [IRR]=1.52, 95%CI: 1.10-2.10), and the odds of inpatient service utilization increased by 61%(OR = 1.61, 95%CI:1.16-2.23). Model 2, after adjusting for demographic characteristics, showed that the significant association persisted: for every 1-point increase in the diabetes distress score, the expected number of annual outpatient visits increased by 44%(IRR = 1.44, 95%CI: 1.05-1.98), and the odds of inpatient service utilization increased by 52%(OR = 1.52, 95%CI: 1.08-2.15). In Model 3, for each 1-point increase in depressive symptom score, the expected number of annual outpatient visits increased by 4% (IRR = 1.04, 95%CI: 1.01-1.06), and the odds of inpatient service utilization increased by 5% (OR = 1.05, 95%CI: 1.03-1.08). In Model 4, after adjusting for demographic characteristics, the significant association persisted: for each 1-point increase in the depressive symptoms score, the expected number of annual outpatient visits increased by 4% (IRR = 1.04, 95%CI: 1.02-1.07), and the odds of inpatient service utilization increased by 4% (OR = 1.04, 95%CI: 1.01-1.07). Age, comorbid cardiovascular disease, cerebrovascular disease, and other chronic conditions in patients with T2DM were significantly associated with the number of outpatient visits. The duration of disease, comorbid cerebrovascular disease, and other chronic conditions in adults with T2DM were significantly associated with their utilization of inpatient services.

Association of psychological distress with healthcare expenditures

3.3

Results from the ordinary least squares regression were presented in Table 3. Outpatient costs were positively associated with diabetes distress scores (p<0.05). Model 1 indicated that each 1-point increase in the diabetes distress score was associated with a 41.9% increase in outpatient costs (e^β^-1 = 0.419, 95%CI: 0.02-0.67). Model 2 indicated that each 1-point increase in diabetes distress score was associated with a 43.3% increase in outpatient costs (e^β^-1 = 0.433, 95%CI: 0.03-0.68). Drug costs, inpatient costs, and total medical expenditures were not related to psychological distress scores (P>0.05), and the duration of disease in adults with T2DM was a factor affecting drug costs.

The total medical expenditures were positively associated with depressive symptom scores (p<0.05). In Model 5, a 3% increase in outpatient costs (e^β^-1 = 0.03, 95%CI: 0.01-0.05) was observed for each 1-point increase in a depressive symptom score. In Model 6, each 1-point increase in depressive symptom score was associated with a 2% increase in total medical expenditures (e^β^-1 = 0.02, 95%CI: 0.01-0.05). Drug costs, inpatient costs, and total medical expenditures were not associated with depressive symptom scores (P>0.05). The education level was one of the factors that influenced outpatient costs.

Additional analysis

3.4

Results from sensitivity analyses further verified the robustness of the main analysis results (Additional Tables 3, 4). Sensitivity analysis showed that there were significant differences in age, sex, education level, marital status, duration of diabetes, distress score, depressive symptom score, annual number of outpatient visits, inpatient service utilization, outpatient costs, and total medical expenditures (P<0.05) among the four groups (Additional Tables 1, 2). The total medical expenditures were higher in adults with psychological health issues compared to the group of adults with T2DM (Additional Table 4). A total of 32.62% of the participants were comorbid with distress, depressive symptoms, or both(CESD ≥10 and/or DDS ≥2) (Additional Figure 1).

Discussion

4

To our knowledge, this is the first study of the association of T2DM healthcare expenditures with diabetes distress and depressive symptoms in rural communities in China. Our study has several findings. First, diabetes distress and depressive symptoms were associated with an increased expected number of annual outpatient visits and inpatient service utilization. Second, diabetes distress was associated with higher outpatient costs, and depressive symptoms were associated with higher total medical expenditures. Third, the duration of the disease factor was associated with drug costs. The education level factor was associated with outpatient costs. Substantial evidence has demonstrated that comorbid depression and depressive symptoms among adults with diabetes was associated with poor diabetes outcomes (39). Mental health issues in adults with diabetes deserve further attention.

The main findings of our study illustrate that poor mental health status was associated with higher expected annual outpatient visits and healthcare utilization in adults with T2DM. Previous studies have also shown that diabetic comorbidities of depressive symptoms or mental distress were associated with higher health service utilization, particularly inpatient care, outpatient visits, and prescriptions (8, 41). Notably, the strength of this association differs between diabetes distress and depressive symptoms. While they overlap to some extent, yet the two concepts are not interchangeable and should be treated as distinct conditions (42). Diabetes distress, as assessed by the DDS-17, refers specifically to the emotional burdens directly related to diabetes management (43, 44). Each one-point increase in its score often has been associated with greater concerns regarding complications, self-management failures, and related issues (34), which may coincide with more frequent healthcare-seeking for diabetes-specific problems. In contrast, depressive symptoms, measured by the CESD-10, encompass a broader range of affective and somatic symptoms; their relatively smaller effect sizes could also be explained if they shape decisions for care indirectly, for instance, by reducing treatment satisfaction (45–47). Similar to previous studies (8, 48), In our study, cardiovascular disease, cerebrovascular disease, and other chronic diseases co-morbidities were positively associated with higher annual outpatient visits and inpatient service utilization. In addition, the duration of T2DM was positively associated with the utilization of inpatient services. People with long-term diabetes are more concerned about their health and more cautious about changes in their condition, resulting in more use of inpatient services. Therefore, further classification for diabetic population for early screening and intervention is needed (49), such as T2DM patients with longer disease duration, multiple comorbidities,and co-existing mental health conditions.

Another meaningful finding of our study is that diabetes distress and depressive symptoms were also linked to higher healthcare expenditures for adults with T2DM. In our study, diabetes distress was associated with increased outpatient costs; diabetes with depressive symptoms may be linked to higher total medical expenditures. This may be due to the fact that diabetes distress did not reach the level of hospitalization and merely increased the use of outpatient services, thus only being associated with higher outpatient costs. People with T2DM with depressive symptoms would need to treat and cope with more severe psychological conditions, which led to higher total healthcare expenditures. Consistent with other studies, people with diabetes who had depression or distress had higher healthcare costs than those who did not (40, 50–52). A study using a nationally representative sample of the United States noted that among people with diabetes, the total expenditure on health management for adults with depression was 4.5 times that of adults without depression (41). The presence of comorbidity of depression and anxiety may expose more people with diabetes and their families to financial hardship (53). In addition, our research indicated that diabetes duration was one of the factors that affected drug costs. This may be because the longer the disease lasts, the greater the need for drugs to control the condition, and the corresponding increase in drug costs. The cost of diabetes drugs and equipment was a persistent obstacle to achieving glycemic goals (54). Co-occurring depressive symptoms may directly increase the financial burden of people with chronic conditions through psychotherapy, medical services, and the use of antidepressants (55).

Our findings raised the issue of whether necessary screening and care models needed to be developed for people with diabetes and comorbid psychological distress. Given the significant association of diabetes distress and depressive symptoms on the financial burden, specific population screening and monitoring of people with diabetes is critical. Treatment of depression in people with diabetes is both effective and cost-effective, improving overall outcomes (56). A study of a tiered collaborative care plan delivered by a nurse’s Depression Care Manager showed depression outcomes improved in the study intervention group compared with patients in the usual care group, with a 5-year average cost reduction of $3907 (57). Another study from the INDEPENDENT series indicated that multi-component collaborative care model can effectively improve combined psychological and metabolic outcomes of diabetic patients, likely by reducing depressive symptoms, enhancing self-management, and thereby improving the health outcomes (58–60). However, studies have shown that it is still common for adults with diabetes and depression not to receive antidepressants or psychotherapy for depression (61, 62). Therefore, diabetes distress should be monitored routinely (17). If diabetes distress is found, it should be acknowledged and dealt with (63). If necessary, patients should be referred for follow-up treatment (64). Focusing on the clinical predictors of distress and depression in people with diabetes and screening for people with diabetes appears to be a reasonable and feasible way to improve disease management. Adults with a history of depression need ongoing monitoring for relapse of depression in the context of usual care (65). A combination of mental and physical health care can improve medical outcomes for adults with a history of depression (63). A collaborative, person-centered approach to care has been shown to improve depression and medical outcomes (66).

The study has several limitations. First, this study is a cross-sectional study, which limits the inference of causal relationships in the results section, and it is also impossible to clarify whether there is a reverse causal inference between various variables. Future studies can further follow up on the population and adopt a cohort study design to further analyze the causal relationships between mental status, utilization of medical services, and medical service costs. Second, measures of depression symptoms or distress, and comorbidities were answered subjectively by the participants with interviewer assistance. The result may be underestimated, and individuals may choose better response options because of shame. Third, measures of healthcare utilization and expenditure were self-reported over a one-year recall period, which may be subject to recall bias. Fourth, we measured fasting blood glucose only once, without subsequent blood glucose monitoring, so it was impossible to describe changes in psychological conditions associated with changes in diabetes glycemic control, and whether psychiatric disorders lead to higher healthcare costs by exacerbating the development of diabetes. Fifth, the study was conducted in rural health clinics in southeastern China and may not be representative of populations in other regions. The results need to be generalized with caution. Sixth, our study did not consider the availability of health insurance, and future studies could add health insurance to study the impact on diabetes costs.

Conclusion

5

In conclusion, psychological distress (diabetes distress and depressive symptoms) are common in adults with T2DM. Psychological distress is associated with higher healthcare expenditures and more healthcare utilization. The results from our study suggest that diabetes distress is associated with higher outpatient costs and that diabetes with depressive symptoms is associated with higher total medical expenditures. This study reinforces the view that psychological screening and care interventions for adults with diabetes are necessary. Psychological guidance by a qualified mental health professional, preferably with diabetes expertise, is beneficial for improving the psychological status and health outcomes of adults with diabetes.