The burden of physical activity on type 2 diabetes public healthcare expenditures among adults: a retrospective study

During the last decades, the prevalence of some chronic diseases increased significantly in the worldwide population [1‐3]. The burden associated with the outcomes of such diseases generates a substantial impact in public health costs. In particular, Type 2 Diabetes Mellitus (T2DM) may be considered one of the main chronic diseases related to increased healthcare costs, due to the set of co-morbidities derived from its onset.

The economic costs from T2DM complications were estimated to be higher than US$130 billion in the United States during 2002 [4], a situation similar to developing countries. In Latin America, costs due to diabetes were estimated to be US$65 billion during 2000 [5]. Specifically in Brazil, T2DM represents a significant public health cost in hospitalizations [6], especially considering the accelerated process of demographic transition.

Habitual physical activity presents a protective effect against the onset of some diseases, such as overweight and T2DM [7‐10]. Diloreto et al. [11] in a longitudinal design carried out with Italian T2DM subjects, identified that increased physical activities during leisure time may result in cost savings in healthcare. Beltoldi et al. [12] also observed a significant negative association between physical activity and the need to use medication.

Despite these results there is still a lack of studies on the association between habitual physical activity and public healthcare costs due to T2DM in developing nations, therefore, there is no evidence that the promotion of physical activity by government could be useful to reduce T2DM treatment costs. Thus, the purpose of this study was to analyze the relation between the level of physical activity and the expenditures in public healthcare with T2DM treatment.

The mean age was 60.1 ± 8.9 years, ranging from 32.7 to 75.0 years. The mean BMI and fasting glucose were 30.6 ± 6.5 kg/m² and 164.9 ± 72.1 mg/dl, respectively. The number of medical consultations per year was 1.83, ranging from 0 to a maximum of 6. Table 1 shows the general characteristics of the sample grouped, according to gender. Mean values of public health expenditures and physical activity score were similar in men and women (p > 5%), but women presented higher expenditures on medication for other diseases (p = 0.006). The sample included mainly white individuals, presenting poor glycemic control and high percentage of subjects with increased adiposity. Women presented higher prevalence of overweight and obesity (p = 0.008) and elevated WHR (p = 0.003). Physical activity level was not associated with gender (p = 0.586).

In our sample, the total expenditures were approximately US$ 18,500 and medication for T2DM represented 58.4% of these overall expenditures (medication for other diseases 4.7% and medication for T2DM treatment 53.7%). Nursing consultations, medical consultations and laboratory tests represented 24.5%, 7.6% and 9.5% of the overall cost, respectively.

Diabetic individuals aged over 50 years presented higher expenditures on medication for the treatment of diseases other than T2DM (p = 0.041). Smoke habit was not associated with any expenditure indicator (Table 2). As could be expected, insulin use was associated with increased expenditures in general medication (p = 0.001) and specifically for T2DM treatment (p = 0.001). Additionally, insulin use was also associated with higher expenditures in medical consultations (p = 0.005), nursing consultations (p = 0.001) and overall public healthcare expenditures (p = 0.001).

WHR was positively associated with expenditures on medication for diseases other than T2DM (p = 0.045). Diabetic neuropathy was not associated with increased expenditures in healthcare (Table 2). There was an association between physical inactivity and public expenditures in healthcare for medication to treat diseases other than diabetes (p = 0.006) and medical consultations (p = 0.026). Sedentary and moderately active diabetics presented similar results (p > 5% in all comparisons).

The logistic regression results (Table 3) show that physically active diabetic individuals were 81% less likely to be in the highest tertile of expenditures due to treatment of other diseases (p = 0.007) and 74% less likely to be in the highest tertile of expenditures on medical consultations (p = 0.029) when compared to sedentary diabetic individuals. Expenditures on laboratory tests presented only a marginally significant association (decreased odds of 62%; p = 0.092).

Age (r = 0.20; p = 0.023), BMI (r = 0.33; p = 0.001) and WHR (r = 0.20; p = 0.025) also were related to medication for other diseases. Physical activity score was negatively related to expenditures on medical consultations (r = - 0.23; p = 0.011) and medication for other diseases (r = - 0.23; p = 0.010).

Older diabetic individuals presented higher expenditures on medication for the treatment of diseases other than T2DM. Similarly, Plotnikoff et al. [20] identified that age was positively related to healthcare utilization in Canadian persons with both types of diabetes. In fact, ageing is frequently pointed out as an independent risk factor for development of several chronic diseases, including T2DM [1, 3, 8, 21]. Additionally, the period of disease onset during the lifespan may be a determinant of the magnitude in healthcare expenditures, due to the co-morbidities associated to prolonged diabetes [22, 23].

Regular use of exogenous insulin increased healthcare expenditures significantly with T2DM treatment among all patients, an expected result, since exogenous insulin is one of the most expensive medications financed by the Brazilian public health system. Higher expenditures on medical consultations and nursing consultations among exogenous insulin users, however, were relatively unexpected results. Nevertheless, the major part of the difference identified was related to consultations for medication dispensing, that is, insulin-dependent patients would require more frequent consultations than non-insulin-dependent patients, given that insulin demand special caution in storage and transportation. Furthermore, exogenous insulin is a pharmacological tool used by diabetic subjects with poor glycemic control and, in turn, with more complications of the disease. Khattab et al. [23] evidenced that diabetics with combined use of insulin and oral antidiabetic agents have lower glycemic control than diabetics with only oral antidiabetic agents. Our hypothesis is strengthened, because 69% of the diabetics presented low effectiveness in glycemic control. Thus, the presented results indicated that there is high proportion of diabetic patients with poor glycemic control and this event strongly increases the healthcare costs.

In obese persons, evidence indicates that adipose tissue (mainly visceral) releases a variety of adipokines associated with insulin resistance, dyslipidemia and arterial hypertension [24], suggesting that obesity has central role in the genesis of T2DM. Worldwide being overweight and obese have been associated with increased healthcare expenditures among adult populations [6, 25]. The results presented in this paper reinforce evidence previously published, since higher BMI and WHR were significantly related to higher expenditures on medication for chronic diseases other than diabetes. Moreover, our findings call attention to the high occurrence of general and central obesity among individuals with T2DM.

Diabetic neuropathy was not related to increased costs, disagreeing with previous data, in which the presence of diabetic neuropathy significantly increased healthcare costs [26]. Even with the presence of foot ulceration in some patients (n = 8), none of these had either previous diagnosis of diabetic neuropathy or received medication for the treatment of neuropathy (e.g. pain medication). This lack of preventive care related to diabetic neuropathy would justify the absence of differences and constitutes a concern due to the increased risk of ulceration.

Diabetic individuals engaged in physical exercise for at least 30 minutes per day present better glycemic control and lower medication use than sedentary ones [12, 23] and this relationship appears stronger in T2DM than type 1 [20]. The present study observed the same patterns in the association between habitual physical activity and expenditures on medication for treatment of diseases other than diabetes (e.g. antihypertensive and hypolipidemic agents). There is a well-documented beneficial effect of physical activity in metabolic and cardiovascular outcomes. Regular participation in physical activities acts on different biologic pathways increasing nitric oxide release, clearing of fat free acids and decreasing in oxidative stress, physiological processes associated with a decrease in the risk for arterial hypertension, dyslipidemia and T2DM [27‐29]. Among type 2 diabetic individuals, higher physical fitness is inversely related to blood pressure, body fatness, uric acid and total cholesterol [30]. Therefore it is possible that such multiple beneficial effects from habitual physical activities may be the reason for the decrease identified in public healthcare expenditures due to medication and laboratory tests.

In fact, physically active diabetic individuals are less susceptible to disease complications, resulting in lower expenditures on general medical consultations. On the other hand, the lower number of general medical consultations could also be associated to a better self-perception of health, since physical activity has also been significantly associated with a better quality of self-perception of health [31].

It is noteworthy that our study only calculated expenditures related to preventive medicine, while expenditures generated by complications of the T2DM were not (hospitalization and emergency room). Our results suggests that the impact of large preventive actions targeting diabetic patients attended by public health care system can have relevant contribution in the reducing both the occurrence and the costs of these secondary outcomes. Thus, it will also be necessary that health professionals orientate diabetic patients in leading a healthy lifestyle by means of physical activity and other healthy behaviors.

Limitations must be recognized. First, the cross-sectional design does not offer support for causality statements in the relationship between healthcare expenditures and physical activity level. Second, metabolic equivalents, steps, counts and time of activity are commonly used to describe physical activity and, therefore, there are guidelines based on these variables. On the other hand, the Baecke questionnaire provides a specific score without previously proposed cutoff, limiting the generalization of the results. Third, the burden of both hospitalizations and other diabetic complications/co-morbidities in overall expenditures had not been calculated and, therefore, should also be identified as both a limitation and a possible focus for future studies analyzing the relationship between physical activity and healthcare expenditures among diabetic patients.

In summary, the results indicate that type 2 diabetic patients attending public healthcare units with higher engagement in physical activity presented consistently lower healthcare expenditures for the public healthcare system. The decrease observed in expenditures to treat the diabetic patients was mainly due to the reduction in medicines used to treat diseases other than T2DM and medical consultations, which represent approximately 12% of the overall cost related to preventive services. Therefore, habitual physical activity promotion could be an important strategy for healthcare systems to decrease costs associated with the treatment of diabetes and its co-morbidities.