Recently updated statistics show that costs related to diabetes in the US population in 2007 reached $174 billion.1 It is estimated that the care of an individual with diabetes now costs 2.3 times as much as medical care for an individual without diabetes. The overall prevalence of diabetes in the US has now increased to 7.8%, a clear increase from an estimated 5–6% less than a decade ago.
Recently updated statistics show that costs related to diabetes in the US population in 2007 reached $174 billion.1 It is estimated that the care of an individual with diabetes now costs 2.3 times as much as medical care for an individual without diabetes. The overall prevalence of diabetes in the US has now increased to 7.8%, a clear increase from an estimated 5–6% less than a decade ago. This means that about one in five US healthcare dollars are being spent on the entire care of someone with diabetes, with one in 10 US healthcare dollars being spent on diabetes-specific needs in 2007.sup>2
The costs associated with diabetes have been increasing since the late 1970s, with significant rises noted in the late 1990s and early 2000s, and these costs are now escalating even more rapidly than predicted. In 2003, the American Diabetes Association (ADA) estimated the total cost of diabetes in 2002 to be around $132 billion, with $91.8 billion in direct expenses and $39.8 billion in indirect expenses. These numbers were used to project an estimated total cost in the US of $200 billion per year by 2020.3 Currently, we are already spending $178 billion per year, and this is a conservative figure. With one in three US citizens born after 2000 predicted to develop diabetes,4 costs could rapidly consume the entire healthcare budget of even a wealthy nation such as the US.
To reduce costs, one must first look at the distribution of diabetes-related expenses. These mostly come from hospital inpatient care, where an estimated $58,344,000,000 was spent in 2007.2 Outpatient medications and supplies were estimated to be $27,684,000,000, with physician office costs totalling $9,897,000,000. Costs related to diabetes education other than those that might fall under the category of physician office visit were not included in these estimates, as visits with diabetes educators, registered dietitians, and pharm Ds were not designated separately in the 2007 analysis. What evidence is there that diabetes education costs might offset overall diabetes-related expenses?
Published reviews of diabetes self-management training have shown somewhat unclear results with regard to the benefits attained. Some studies have been hampered by limited cost analysis, others by sample size or duration of targeted intervention. Dijkstra and colleagues5 reported one-year results from an intervention involving 764 people in The Netherlands with type 2 diabetes randomized to usual care, a professional-centered group in which physicians and diabetes specialist nurses were coached by a diabetes specialist, or a patient-centered group in which coaching occurred with the addition of patient provided printed targets for diabetes management. Glycated hemoglobin (HbA1c) decreased by 0.2% in the professional-centered group and by 0.3% in the patient-centered group. An estimated improved life expectancy of 0.34 years was seen in the professional-centered group and 0.63 years in the patient-centered group compared with usualcare patients. Quality-adjusted life-years (QALY) increased by 0.29 in the professional-centered group and 0.59 in the patient-centered group. Incremental costs per QALY were based on epidemiological data and were felt to be cost-effective at $43,303 and $21,700 for the professional- and patient-centered groups, respectively.
Trento and colleagues6 reported results from a three-year randomized controlled trial with 62 patients in Italy. All patients had type 1 diabetes. Half of the cohort met with a team of physicians and psychologists in 15 group visits over the three years. Topics at the group visits were pre-determined by focus groups identifying educational needs. Brief post-group meeting individualized appointments were offered to those asking for them. The usual care group received one-on-one visits with a physician every two to three months. Health behaviors, knowledge of diabetes, and quality of life (QOL) measures significantly improved in the care patients group (p<0.001). Cost was estimated to be $26.31 per point gained in QOL. Knowledge gained through the group care model influenced behaviors (p=0.004), while QOL changed independently of either knowledge or behavior (p<0.001). Among controls, QOL worsened (p<0.001) and knowledge and behavior remained unchanged. High-density lipoprotein (HDL) cholesterol increased among patients randomized to group care (p=0.027) and total cholesterol decreased in the controls (p<0.05). HbA1c decreased in both groups, although not significantly.
Trento has also reported a five-year intervention in type 2 diabetes patients comparing group visits with one or two physicians and diabetes educators with one-to-one visits with a physician.7 One hundred and twelve patients were randomized to either of the two study arms, and 84 completed the study (42 from each arm). Diabetes concepts were provided in an interactive forum. Four sessions were repeated in years one and two, spread out into seven sessions in years three and four, and repeated again in year five. Topics included weight management, food choice, self-glucose monitoring, smoking cessation, and understanding diabetes complications. HbA1c remained unchanged in the group visit cohort, while it increased by 1.7% (1.1–2.2; p<0.001) in the usual care group. Bodyweight decreased in the group visit cohort by an average of 3.5kg, but remained the same in the usual care group (p<0.015). Knowledge of diabetes and problem-solving ability (as tested by yearly questionnaires) relating to areas such as preventing acute and chronic complications, weight management, food choices, and smoking cessation improved continuously throughout the five years in the group visit cohort, but worsened in the usual care group from year three (p<0.005). QOL improved in the group visit cohort after two years, but worsened for the usual visit cohort after two years (p<0.001). While cost analyses were not included, HbA1c levels did not change in either group, but this was in the setting of an increase in oral hypoglycemic agent dosage in the usual care cohort, with doses either remaining unchanged or decreasing in the group visit cohort.
The use of other members of the treatment team to complement the care of physicians has also been shown to be effective. Ragucci8 reported a one-year observational study of pharmacist-/certified diabetes educator-provided education and medication co-management in type 2 diabetes. Although the study had no control group and was limited to university-based primary care clinics, 191 patients were seen and followed, and the average HbA1c level dropped from 9.5 to 7.8% (p<0.05). In 155 patients, average blood pressure decreased from 141/79 to 135/75mmHg (p=0.007), and in 98 patients low-density lipoprotein (LDL)-cholesterol decreased from 114 to 112mg/dl (not statistically significant). Based on an estimated cost savings of $820 for every 1% decrease in HbA1c, savings of $61,500 were calculated. Leal et al.9 reported results from a disease management program run by pharmacists in a population of mostly indigent, Spanish-speaking, somewhat transient patients at a community health center. Between four and 36 months, 199 patients were followed. Collaborative practice agreements between medical staff and pharm Ds allowed for medication management, dose adjustment, and teaching with a focus on targeting glycemic, blood pressure, and lipid standards of care. Statistically significant differences (p<0.001) were seen in pre-treatment and follow-up measures of total cholesterol, triglycerides, LDL-cholesterol, HbA1c, and systolic and diastolic blood pressure. The HbA1c level alone dropped by an average of 2% (95% confidence interval [CI] 1.8–2.3). Cost analyses were not performed, but the decrease in HbA1c alone would translate into significant savings. Cranor and colleagues10 reported a five-year study in a mixed type 1 and 2 diabetes population in which community pharmacists followed patients after a certified diabetes educator provided education. Although the study included no control group and not all data on all patients were reported, mean direct medical cost from decrease in HbA1c was calculated to drop by $1,200–1,872 per patient per year from baseline for one employer group. Additionally, work productivity gained was estimated to be $18,000 annually.
Registered dietitians (RDs) have also been shown to provide patients with information that, over time, allows for better diabetes self-management. In a very short study over only six months, Franz11 showed that patients with type 2 diabetes who spent 151 minutes with an RD using practice guidelines versus just 65 minutes with an RD and no practice guidelines had an improved cost-effectiveness ratio. This ratio compared relative cost per unit change in glucose as $5.32 for usual care with $4.20 for the practice guideline intervention. Wolf and colleagues12 also reported a study of interventions by RDs versus usual care accompanied by written materials. The group receiving the lifestyle case management intervention by RDs achieved greater weight loss (3kg at four months; p=0.006), reduced HbA1c (p=0.02), and decreased prescriptive medication use (although only 0.08 medications less, this was statistically significant at p=0.03) compared with usual care.
Case management in inpatient and outpatient settings with regard to economic benefit was reviewed by Klonoff et al.13 As these types of study have not been performed in a randomized controlled fashion, the extrapolation of results is limited. It is also unclear whether specific subsets of patients may benefit more than others. Levitan et al.14 reported results from an observational comparison of length of stay and in-hospital day costs, comparing a team diabetes consultation approach using endocrinologists, nurse educators, and RDs in 34 patients with an endocrinologist consultation alone in 43 patients and no endocrine consultation in 27 patients. The time of evaluation overlapped, but was not the same for all observational groups. However, the case management approach resulted in a shorter average length of stay of 3.6 days versus 5.5 days for the endocrinologist-alone group (p<0.05). With no endocrine consultation, the average length of stay was 8.2 days (p<0.0001). Overall hospital costs decreased by $120,000 for those seen by a team. Each day’s delay in consultation resulted in a one-day increase in length of stay. In outpatient settings, the results are mixed with regard to using a decrease in hospital admissions as an end-point. A program aimed at children did show some benefit,15 but other programs for adults were not as positive.16
Recent reports have suggested that bariatric surgery may be another approach in the diabetes treatment armamentarium. A meta-analysis including 22,094 patients—15.3% of whom had type 2 diabetes— undergoing bariatric surgery showed substantial improvement (if not resolution) of the markers of diabetes.17 Of those individuals undergoing lap banding, 48% showed complete remission of hyperglycemia. This was true for 84% undergoing Roux-Y gastric banding and >95% undergoing biliopancreatic diversion, and this remission appears to last for more than 10 years post-surgery.18 Whether this is due to a hormonal signal that is not yet understood, as suggested by the very often immediate post-operative effect on glycemic control or weight loss, or a combination of both, is unclear.19 Also unclear are the long-term consequences of bariatric surgery. In a Swedish study that reported over 10 years of follow-up,20 mortality was decreased in a very overweight population undergoing bariatric surgery, but diabetes was limited to only 7–8% of the cohort studied. Additionally, some long-term studies show recidivism as defined by weight regain if patients are followed over time. In his Finnish study, Tolonen21 showed that if success is defined as an estimated weight loss of >50%, 60% achieved this at three years postsurgery, but at eight to nine years, only 35% achieved the goal. Further clinical studies are clearly needed that are targeted to diabetes populations with longterm follow-up to understand where bariatric surgery may fit into the treatment algorithm of diabetes, and perhaps more specifically which subpopulations will benefit the most.
Increased physical activity and dietary modification are well-known standard non-pharmacological treatments for diabetes. These pre-dated the pharmacological treatments of type 2 diabetes, and are still adjuncts to the treatment program of type 1 diabetes, although evidence to support specific physical activity or its duration or a specific dietary program is still sparse other than as targeted to promote weight loss. There is information as to the value of lifestyle modification in the prevention of diabetes, with recent information as to the durability of lifestyle modification in the long term. The Diabetes Prevention Program (DPP)22 set a goal of losing 7% of initial bodyweight through daily calorie goals of 500–1,000 calories below weight maintenance, 25% of calories from fat, and an activity goal of 150 minutes per week of moderate intensity (similar to a brisk walk). There was a 58% reduction in the progression to type 2 diabetes in those at high risk for diabetes development compared with those in the control group. Whether lifestyle modification will be beneficial to those already diagnosed with type 2 diabetes may be answered by the Look AHEAD study, currently in progress.23 This study also targets a weight loss of 7% of initial weight, although the individual goal is a 10% loss from initial weight. Activity goal is 175 minutes per week. Initial results have been similar to the DPP with regard to weight loss,23 but the development of cardiovascular disease is being tracked in the Look AHEAD study and as an extension of the DPP. This is important, as the DPP lifestyle intervention was shown to have an effect on cardiovascular markers such as C-reactive protein and fibrinogen compared with the control group, although this was shown to a lesser extent than in the metformin group in the DPP.23
Lifestyle modification has been shown to have an effect in other diabetes prevention studies reporting long-term results. In the China Da Qing study,24 in over 20 years of follow-up the decreased incidence of progression to diabetes persisted in the lifestyle intervention group, but the 51% lower incidence of diabetes over six years in the intervention group decreased to a 43% lower incidence over 20 years. Cardiovascular disease events, cardiovascular-disease-associated mortality, and all-cause mortality were not statistically different in intervention versus control groups.
Finally, the importance of community involvement and the leadership of a champion in a community in the treatment of diabetes should be noted. The power of the non-traditional (outside clinic) approach has been shown to be instrumental in many communities, specific to educational messages relating to the need for diabetes control, tools of diet and physical activity, and complications. For example, eastside and southwest Detroit family focus groups were organized by community, age, gender, and language to plan community-based participatory interventions to reduce the prevalence and impact of diabetes and its risk factors. These community residents participated in planning, implementing, and analyzing data from the focus groups and subsequent planning meetings. Themes included: diabetes is widespread and risk begins in childhood, with severe consequences for African-Americans and Latinos; denial and inadequate healthcare contribute to lack of public awareness about pre-symptomatic diabetes; diabetes risks include heredity, high sugar, fat, and alcohol intake, overweight, lack of exercise, and stress; and cultural traditions, lack of motivation, and a lack of affordable, accessible stores, restaurants, recreation facilities, and programs are barriers to adopting preventive lifestyles. Participants identified community assets and made recommendations that resulted in a multilevel intervention program. These included the development of: family-oriented interventions to support lifestyle change at all ages; culturally relevant community and health provider education and materials; social support group activities promoting diabetes self-management, exercise, and healthy eating; and community resource development and advocacy.25 The identification of barriers to the management of diabetes allows an endocrinologist to be able to work on a community level to address these issues and allows an opportunity for him or her to reach more individuals through collaborative work. Kipke et al.26 identified a number of challenges to adolescent health that were specific to concerns regarding weight management and risk for metabolic syndrome in east Los Angeles. These challenges related to lack of quality food produce and an abundance of calorie-dense but nutrient-poor food options. However, the need to not just identify but also engage communities and individual champions in making change, such as described by Boltori et al.,27 can also be translated to office or clinic delivery settings. The group visit, often championed by a patient leader, needs further exploration in type 2 diabetes as a tool to encourage individuals with diabetes to make a long-term commitment to dealing with their chronic illness by non-pharmacological means, or, at the very least, to complement pharmacological management with other strategies.
In summary, non-pharmacological approaches are available and, although cost-effectiveness and long-term benefit data are still sparse, the escalating costs of long-term pharmacological intervention necessitate the continued exploration of these different approaches, from both an individual and a societal point of view.■