Our study presents an examination of temporal trends in hospitalizations and hospital charges affected by diabetes for age-specific and sex-specific strata across the entire U.S. population. We report significant increases in hospitalizations associated with diabetes, in particular for younger adults, over a recent 14-year period despite the fact that overall hospitalizations for young adults have decreased or remained relatively stable over the same period. According to the Centers for Disease Control and Prevention (CDC), people with diabetes are much less likely to be hospitalized compared with a decade ago.11
Therefore, the increases in population-adjusted rates of hospitalizations may reflect increasing diabetes prevalence in the adult population. Consistent with our hypothesis, we found that the largest increases in hospitalizations associated with diabetes occurred among adults 30–39 years of age, which may well be related to increases in the prevalence of diabetes among this age group. One previously published study based on nationally representative data from the Behavioral Risk Factor Surveillance System (BRFSS) found that the largest increase in diabetes (70% increase) occurred among people aged 30–39 years over a similar time period (1990–1998).2
The larger increases in hospitalizations that we saw among young women aged 20–39 years compared with men, even after excluding hospitalizations associated with pregnancy and childbirth, could potentially be due to increases in diabetes prevalence and may be greater for women vs. men in this age group, particularly given higher rates of obesity for women vs. men documented in these age groups.20
However, we are unaware of any studies that have evaluated both sex-specific and age-specific trends in diabetes prevalence in the United States over this period. One study using National Health and Nutrition Examination Survey (NHANES) data from a similar time period (1988–2002) reported greater increases in diabetes prevalence among men (7.9% to 10.2%) than in women (7.8% to 8.5%), but because this study combined adults of all ages, we speculate that larger increases in diabetes prevalence among young adult women compared with young adult men may have been masked by opposite trends in the older age groups.21
This hypothesis is supported by data from Hillier and Pedula,22
who studied patients with new-onset type 2 diabetes in a health maintenance organization between 1996 and 2000 and found that patients were more likely to be female if they had diabetes onset at a younger age (<45 years) but were more likely to be male if they had diabetes onset at older ages (≥45 years).
Another possibility is that the greater rate of hospitalizations among women compared with men is due to greater morbidity among women than men with diabetes. Studies have shown that women with diabetes have low use of diabetes-related preventative care, are less likely to receive aggressive medical management, and experience worse outcomes after hospitalization for CVD.23–27
Furthermore, for those aged 20–39 years, affective diagnoses for women and schizophrenia for men were among the top five primary diagnoses associated with diabetes hospitalizations. The high prevalence of affective disorders among younger women may be a reflection of the fact that both younger age and female gender are associated with increased risk of depression among patients with diabetes,28,29
which can be associated with increased healthcare use and expenditures for people with diabetes.30
Despite the higher rates of hospitalizations among women, men were more likely to have complications of diabetes mellitus, particularly between the ages of 20 and 39 years, although differences were not large. This finding contrasts with previous reports that have noted equivalent31
or poorer control of diabetes among young women compared with men.32
The high rates of schizophrenia diagnoses among younger men is of particular interest, given the recently publicized association of atypical neuroleptics with incident type 2 diabetes.33
We found that population-adjusted hospitalizations were greater in elderly men than in elderly women with diabetes in 2006. We are unaware of previous studies that examined hospitalizations in elderly men vs. women among adults with diabetes. A cross-sectional examination of NHANES data from 1999 to 2002 found that elderly people with diabetes were more likely to be female than male,34
suggesting that greater rates of comorbid disease among men, rather than greater diabetes prevalence in men, caused more frequent hospitalizations in men with diabetes. Consistent with this explanation, male sex has been associated with higher comorbidity and all-cause mortality among older managed care enrollees with diabetes.35
Particularly among the elderly, greater hospitalization rates in men may reflect other factors less directly related to illness that may favor more frequent hospitalization in elderly men, such as low socioeconomic status.36
Despite these findings, the percent increase in hospitalizations between 1993 and 2006 was greater in elderly women than in men, suggesting that the burden of disease may be shifting.
We note that population-adjusted hospitalization rates were higher for men compared with women in most age strata, although overall population-adjusted hospitalization rates were higher for women across all years. This was accounted for by the fact that fewer men survive into their 70s and 80s, resulting in a lower Census population estimate for the denominator.
Finally, the dramatic increase in hospital charges related to diabetes over the 14-year study period is significant. Although the financial burden of diabetes hospitalizations is shared equally by private and public payers among younger people,10
Medicare bears a disproportionate burden of the costs for older people with diabetes. As rates of diabetes continue to increase, particularly among younger cohorts of individuals, the future economic burden on Medicare will only escalate as these people age. In aggregate, these trends may serve as a compelling economic rationale for third-party payers in the public and private sectors to focus on diabetes prevention, particularly among younger adults in the population.
We elected to perform our analysis starting in 1993, given that hospital data were drawn from a much larger sample of states for the NIS starting that year. We, however, speculate that increases in hospitalizations associated with diabetes were occurring before 1993, as the CDC reported increases in diabetes between 1988–1994 and 1999–2002.21
Therefore, the increases that we report over this period likely represent a trend that has been occurring over a much longer period of time.
We note that the trends we report may be overestimated because of multiple factors. First, discharges from the NIS can include multiple hospitalizations for a single person; therefore, increases in population-adjusted rates of diabetes hospitalizations could well be related to increases in the severity of illness associated with diabetes (resulting in multiple hospitalizations for individuals) as well as an increased number of individuals with diabetes. Second, as a result of increasing awareness of type 2 diabetes during the period, providers may have been more likely to record a diagnosis of diabetes with hospitalizations. Finally, the definition of diabetes changed from having a fasting blood glucose of 140
mg/dL to having a fasting blood glucose of 126
mg/dL, in 1997, and the increased hospitalization rates could reflect the greater number of people subsequently diagnosed with diabetes according to the new criterion.37
Although we previously conducted a study of diabetes hospitalizations between 1993 and 2004 using the NIS, that study was exclusively focused on a subset of individuals, namely, children and young adults. We were unable to evaluate trends by diabetes type, particularly in children, as administrative coding by diabetes type may be subject to greater variability. Finally, because obesity as a comorbid diagnosis is likely undercoded, we were unable to quantify the extent to which obesity-related trends are affecting trends in type 2 diabetes prevalence.
Additional population-based studies are needed to better understand whether the differential increase in diabetes-related hospitalizations among young women of reproductive age and among young adults more broadly represents increasing incidence and prevalence of diabetes vs. an increasing burden of comorbid disease. The findings of these studies will have significant policy implications for care delivery and healthcare financing as these populations age.