Primary care prevention project idea: PP4P

Novel Strategies for the prevention of type 2 diabetes: Pay the patient for performance (PP4P)


Eugene Scharf –MS3
3/9/11
Primary care rotation: Scarsdale medical group


Case:

Patient DM is a 36 year-old man with a past medical history of hypertension and obesity. He was last seen in the office two years ago for a check up. The interval history between now and his last visit is remarkable for an increase in bodyweight of 20lbs. He now presents to the clinic for an annual physical.

PMH:
Hypertension – diagnosed 2004, controlled with medication
Obesity –5’10” 223lbs BMI = 32
PSH:
Inguinal hernia repair 1997
Ankle fracture ORIF 1999

Family history:
CAD in father (MI age 55)
Stroke in grandfather (mother’s side)

Social history:
Born in Mumbai, India. Moved to US in 1995, employed as a computer programmer. Single, sexually active, uses protection. Some exercise on weekends. No tobacco, no caffeine, social drinker, occasional marijuana use.

Medications:
HCTZ 25mg
Ayurvedic herb tea
Multivitamin daily

Allergies: NKDA
ROS: +decreased appetite +fatigue +thirst
Physical Exam:

PE:
Gen: obese man in NAD
Vital signs:
T: 36.8 BP 136/80 P 78 RR 14
+central obesity, waist: 43”

Laboratory studies:

BMP: 140 | 110 | 18 / 175 Ca: 9.0 Mg: 2.3 Phos: 3.3
4.0| 25 | 1.3\

CBC: 7.8\ 40 | 14.0 / 350
Liver chemistry: 7.9/4.5/1.3/0.1/50/45/80
Lipids:
Total cholesterol: 250
HDL: 39
LDL: 160
Triglycerides: 185
A1c: 6.0%
TSH: 5.05
Total T4: 7.6
HIV: negative
Hepatitis panel: negative

How should we think about DM’s case? Does the combination of his increased waist circumference, high blood pressure, abnormal lipid panel, high random glucose, and elevated A1c indicate that he is at risk for developing type 2 diabetes? His social history is also remarkable for a sedentary lifestyle (employed as a computer programmer) and too little exercise (weekends only). Finally, DM is also relatively young (age 46) so it is important to know whether there are any interventions that could possibly have long lasting positive consequences on his life quality and mortality. However, what is certain is that DM’s metabolic and physiological health information indicate that he has come to a critical juncture in his life, and his behavioral decisions will now have lasting implications. In this paper I will argue that preventing his and others progression to type 2 diabetes can be significantly reduced by creating cost-saving patient performance financial incentives. Let us first create the context for this discussion.

In 2001, Boyle et al. published an estimate of the prevalence and incidence of type 2 diabetes in America. Based on data from both National Health Information Survey (NHIS) and National Health and Nutrition Examination Survey (NHANES) his group estimated that the prevalence of diagnosed diabetes in America in 2000 was 4.0% or approximately 11 million. His group then went on to predict that by the year 2050, there would be approximately 29 million diagnosed cases of diabetes, with a prevalence of 7.2%1. However, the American Diabetes Association (ADA) reported in 2010 that the current number of diagnosed cases of diabetes was 18.8 million, as well as 7 million cases of diabetes undiagnosed2. This report also indicated that the annual incidence of diabetes was 1.9 million cases/year. Holding that incidence constant and taking only diagnosed cases into consideration, this means that our nation would arrive at Boyle’s et al. 2001 estimate of 29 million cases by 2020, or thirty years ahead of his groups’ original prediction. This 2010 report by the American Diabetes Association also estimated that there are approximately 79,000,000 cases of pre-diabetes in America. Boyle et al followed up with a 2010 report3 that adjusted their prior predictions and re-stated the prevalence of diabetes in America by 2050 based on estimates of incidence and mortality to be between 21 and 33%, with a “middle of the road” estimate of 25%. Combining this estimate with a 2008 Pew Research Group report4 US population estimate of 438 million the result indicates that an estimated 109,000,000 Americans will have diabetes by 2050, up from 11 million at the beginning of the century.

I next sought to investigate the cost burden of diabetes. To this end the ADA published a cost analysis in 20075. This analysis reported the annual national burden of diabetes to be at least $174 billion in 2007 dollars (a figure >5.5 times the 2010 NIH total budget). This cost was broken down into medical expenditures ($116 billion or roughly 2/3 of the cost) and lost productivity ($58 billion, 1/3 of the cost). This analysis concluded that half of all annual medical expenditures in diabetes related to inpatient medical care. Also, considering lost productivity, 82% of the total $58 billion of this cost came from reduced work productivity (35% or $20 billion) and early mortality (47% or $27 billion). The report concludes that persons diagnosed with diabetes on average incur healthcare costs of $11,744/person year, of which $6,649 are costs strictly related to diabetes. Also this cost analysis did not estimate lost productivity of disease burden on well persons (e.g. – well person missing work to care for relative with diabetes), the annual burden of Medicare diabetes administration costs, and assumed the annual cost of persons with pre-diabetes to be $0. Thus the actual costs were determined to be higher than the calculated estimate of $174 billion. Therefore, given the predicted incidence and high future prevalence of diabetes combined with the high annual financial burden of the disease I next sought to investigate whether preventive data exist, and if so what measures are either cost effective or cost saving.

Modern diabetes prevention studies began in 1986 with the China Da Qing study, where 577 Chinese men with impaired fasting glucose were randomized to control, diet, exercise and diet & exercise experimental wings. At a 6 year follow-up, the proportion of individuals with diabetes was significantly lower in the diet & exercise group compared to the control arm (mean 44% vs 66%). The first large scale trial to support the benefit of glycemic control was published 1998, the United Kingdom Prospective Diabetes Study (UKPDS) showed that tight glycemic control in patients with newly diagnosed type 2 diabetes prevented microvascular endpoints (nephropathy, retinopathy, neuropathy) by 25%7. In short 3,867 newly diagnosed type 2 diabetics were randomized either a control group of conventional therapy with the aim of keeping blood glucose levels <270mg/dL versus experimental arms of insulin or sulfonylureas with the aim to keep blood glucose levels <110mg/dL. The UKPDS had a confounding variable worth reporting; in the experimental group sufficient glycemic separation (<110mg/dL) could not be achieved by just one agent thus metformin was added to the experimental group. As well the conventional group (control) had difficulty maintaining consistent blood glucoses of <270mg/dL, thus 80% of the control group received metformin as well. In fact 58% of the total control group patient*years of UKPDS were confounded by metformin use. Mean A1c’s were 7.0% vs 7.9% in experimental versus control, and at 15 year follow up there were no differences in myocardial infarction, stroke, or diabetes related mortality. A trial published in the New England Journal of Medicine (NEJM) 20018, approximately 85,000 female nurses taken from the nurses health study begun in 1976 were followed for 16 years for incident cases of diabetes (of which there were 3300). The study’s finding showed that the RR of diabetes increased sharply depending on BMI at the time of enrollment. BMI’s of >30 had a relative risk of 20.1 (16.6 -24.4), and BMI’s of >35 had RR of 38.8 for acquiring diabetes. This study clearly demonstrated that BMI and bodyweight greatly predicted disease risk. Another NEJM trial published in 20039, by the Diabetes Prevention Program Research Group demonstrated the superiority of diet and exercise compared to metformin or placebo in preventing diabetes. In this trial (DPP), the experimental group consisted of diet and exercise and was formally defined as losing 7% of bodyweight, physical activity >150 minutes per week, and dietary fat constituting <25% of total caloric intake. Both the metformin and placebo group were given information on the importance of exercise, but no requirements. Interestingly, on 50% of those in the diet and exercise group made the weight goal, average follow up was 2.8 years. The trial results clearly showed that diet and exercise prevented new cases of diabetes significantly better than either metformin or placebo. Cumulative incidence rates over follow up were for this study 14% vs. 22% vs. 29% at 3 years. It was shown that the number needed to treat (NNT) to prevent one incident case of diabetes in the diet and lifestyle group was 6.9. Finally, consider that diet, exercise, and weight loss interventions have been shown to persist past the treatment duration. For example, 20-year follow-up of China da Qing study published in the Lancet10 showed that the relative risk of acquiring diabetes of those members of the diet & exercise treatment arm compared to the control arm was 0.57 at 20 years follow up. This finding supports the notion that early diet and exercise interventions have protective effects that outlast their treatment periods. This finding has enormous implications for diabetes prevention. My next step in understanding current methods for diabetes prevention was learning whether or not models exist for predicting absolute risk of acquiring diabetes. One such model from the Framingham Offspring Study11 attempted to assess this risk and created a model for a person’s 8-year risk of acquiring diabetes. This predictor model includes the clinical measurements fasting glucose, BMI, HDL-C, triglycerides, family history, and blood pressure together to create a weighted score with which to predict a person’s risk of diabetes. If patient DM’s clinical data are entered into this model, we arrive at a score of 10 or 20, depending on whether or not his fasting glucose is more or less than 100mg/dL. The uncertainty about this metric creates uncertainty about his true 8-year risk; as it is either <3% or 18%. However, the advent of hemoglobin A1c as a point of care diagnostic tool for diabetes helps to remove doubt regarding risk of diabetes when fasting glucose measurements are not available. I assumed that the risk of diabetes could be reasonably predicted and the diagnosis accurately made by using clinical metrics. Thus, in proposing novel diabetes prevention strategies, I next sought to learn whether or not existing diabetes prevention strategies were cost-effective. A 2010 systematic review by Li12 et al of twenty years of cost effectiveness analyses of diabetes prevention strategies indicated that several cost effective strategies exist, and many are cost saving. This review defined an intervention as cost-effective if it resulted in an inflation adjusted cost-effectiveness ratio of less than or equal to $50,000/QALY. One important prevention strategy that was found to be extremely cost effective (<1,200/QALY) in one analysis and cost saving in all others was the use of angiotensin receptor blockers (ARBs) for type 2 diabetes to prevent hypertension and progression to macrobalbuminuria and renal failure versus either placebo or standard therapy. Thus I propose as a first step toward preventing the morbidity of long term diabetes resulting from hypertension and renal failure is the subsidized provision of ARB’s or ACEi to all diagnosed diabetics. However this idea would only help prevent morbidity after a diagnosis and thus not fully address primary prevention. Given that several of the aforementioned diabetes trials have indicated that diet, exercise and reducing bodyweight is superior to all other primary prevention strategies for preventing type 2 diabetes, I next sought to understand if these interventions are cost effective. To this end the same Li et al. review addressed the cost utility of strict glycemic control (with insulin, sulfonylurea, or metfomin), diet, education, or behavioral modifications. In all other healthcare systems outside the U.S. these measures were highly cost-effective or cost-saving. In the US, strict glycemic control as a prevention strategy costs ~41,384/QALY. However in the CDC cost sensitivity analysis that provided this figure13, cost effectiveness for strict glycemic control depends on age. For example; in age groups 25-34 it was 9,600/QALY, and 35-44 was 18,309/QALY, much less than the average for all age groups. But if one refers back to the DPP trial, strict glycemic control was also achieved in a much more cost-effective manner: diet and exercise. Thus we arrive at the central question in diabetes prevention. If diet, exercise and bodyweight reduction have been shown to be superior to all other preventive strategies for type 2 diabetes, why can’t we achieve this goal? People at risk for type 2 diabetes may not control their diet or exercise because there is no incentive to. The morbidity associated with type 2 diabetes is oftentimes decades away, and thus incentive for good behavior now does not exist. However I believe that positive incentives can create cost-saving behavioral modifications resulting in significant prevention type 2 diabetes. Thus I propose that we can reduce the cost burden of diabetes by providing cash incentives to high-risk individuals between the ages of 25-44. Given that Hemoglobin A1c is a reliable marker of blood glucose control for the previous three months, and that glycemic control (fasting glucose <100mg/dL) is associated with reduced risk of diabetes, I propose cash incentives for hemoglobin A1c’s under 6.0% for high risk individuals (impaired fasting glucose, BMI >32, or diagnosed diabetes). For example as stated earlier, in 2007 the economic cost of diabetes was ~6,649/person. If a patient previously diagnosed with type 2 diabetes can show through lifestyle modifications that his/her A1c is now below 6.0% they are eligible for four quarterly payments of $250 (for the quaterly A1c’s) or $1,000 annually. In this scenario, the economic burden of $6,649 is reduced to $1,000, and the patient benefits directly and tangibly from his/her actions with a cash reward. The payments are made from either Medicaid or insurance. Paying patients for outcomes may at first seem counter-intuitive, however the measure would likely be cost saving in light of cost analysis data suggesting the decreased costs related to A1c’s <6.5% (i.e. reduced diabetes incidence). And if a patient is not able to lower his/her A1c to goal, no payments are made and the system absorbs the cost as would have without the program. A patient (similar to our patient DM) with pre-diabetes would be eligible for “Pay for A1c” program thus averting the cost to the system for each year of prevention (as indicated through A1c). One critical component to the program lies in establishing a correlation between A1c and BMI because although A1c defines diabetes, lowering BMI is what can prevent diabetes. To that end I propose a final step, cash incentive for BMI reduction. Thus for every type 2 diabetic patient whose BMI becomes less than 30kg/m2, a one time payment of $1,000 (the patient cannot go back and forth through the set point of 30 to collect multiple payments). As our century progresses the prevalence of diabetes is projected to grow rapidly, such that the annual cost burden of the disease may become unmanageable. In order to prevent this impending liability to say nothing of the human cost, stronger incentives must be created to modify human behaviors to ways congruent with clinical trial proven lifestyle interventions. First, hypertension medications such as ACEi and ARBs have been shown to be cost saving for preventing morbidity associated with hypertension and diabetic nephropathy (the leading cause of renal failure in the US). Thus subsidized provision of these medications for diabetic patients must be policy. Second, proper diet, exercise, and bodyweight control could be attained with financial rewards to those individuals between the ages of 25 – 44, and defined as high risk (impaired fasting glucose, A1c greater than 6.0%, BMI >30, diagnosed diabetic). Two proposals for achieving these measures are “Pay for A1c” cash incentive and a one time cash outlay for diabetics whose BMI’s are lowered to <30 through their efforts of diet and exercise.

Reference:
1. Boyle JP, Honeycutt AA, Narayan KM, Hoerger TJ, Geiss LS, Chen H, Thompson TJ. Projection of diabetes burden through 2050: impact of changing demography and disease prevalence in the U.S. Diabetes Care. Nov;24(11):1936-40
2. American Diabetes Association
www.diabetes.org/diabetes-basics/diabetes-statistics/
3. Boyle JP, Thompson TJ, Gregg EW, Barker LE, Williamson DF. 2010. Projection of the year 2050 burden of diabetes in the US adult population: dynamic modeling of incidence, mortality, and prediabetes prevalence. Population Health Metrics. 8:29.
4. Pew Research Center 2008. US Population projections: 2005-2050.
5. Dall T, et al. 2008. Economic costs of diabetes in 2007. Diabetes Care Vol31:3, 596-615
6. Pan XR, et al. Effects of diet and exercise in preventing NIDDM in people with impaired glucose tolerance. The Da Qing IGT and Diabetes Study. Diabetes Care 1997;20(4)537-44.
7. UK Prospective Diabetes Study Group. Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes. The Lancet, 1998. Vol. 352, 837-853
8. Hu FB et al. Diet Lifestyle, and the Risk of Type 2 Diabetes Mellitus in Women. NEJM 2001; 345;790-797
9. Tuomilehto J et al. Prevention of type 2 Diabetes Mellitus by Changes in Lifestyle among Subjects with impaired glucose tolerance. NEJM 2001 344:1343-1350
10. Li G et al. The long-term effect of lifestyle interventions to prevent diabetes in the China Da Qing Diabetes Prevention Study: a 20-year follow-up study. The Lancet, 371:1783-1789
11. Wilson P WF et al. Prediction of Incident Diabetes Mellitus in Middle-aged Adults: The Framingham Offspring Study. Arch Intern Med. 2007; 167:1068-1074
12. Li R, Zhang P, Barker L, Chowdhury F, Zhang X. Cost effectiveness of interventions to prevent and control diabetes mellitus: A systematic review. Diabetes Care 2010 33:8
13. The CDC Diabetes Cost Effectiveness Group. Cost-effectiveness of intensive glycemic control, intensified hypertension control, and serum cholesterol level reduction for type 2 diabetes. JAMA 2002; Vol.287:19