 |
|
Cost-effectiveness of the Lower Treatment Goal (of JNC VI) for Diabetic Hypertensive Patients
William J. Elliott, MD, PhD;
David R. Weir, PhD;
Henry R. Black, MD
Arch Intern Med. 2000;160:1277-1283.
ABSTRACT
| |
Background The recommendation of the Sixth Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC VI) to lower blood pressure (BP) in diabetic patients to less than 130/85 mm Hg may have negative economic consequences. A formal cost-effectiveness analysis was therefore performed, comparing the costs and potential benefits of a BP goal of less than 140/90 mm Hg (as recommended by JNC V) vs less than 130/85 mm Hg (as in JNC VI).
Methods A 24-cell computer model was populated with costs (1996 dollars), relative risks, and age-specific baseline rates for death and 4 nonfatal adverse events (stroke, myocardial infarction, heart failure, and end-stage renal disease), derived from published data. Costs and benefits were discounted at 3%.
Results For 60-year-old diabetic persons with hypertension, treating to the lower BP goal increases life expectancy by 0.48 (discounted) years and lowers (discounted) lifetime medical costs by $1450 compared with treating BP to less than 140/90 mm Hg. The lower treatment BP goal results in an overall cost savings over a wide range of initial conditions, and for nearly all analyses for patients older than 60 years.
Conclusions Any incremental treatment for 60-year-olds that costs less than $414 annually and successfully lowers BP from below 140/90 to below 130/85 mm Hg would be cost saving in the long term, due to the reduction in attendant costs of future morbidity. The lower treatment goal recommended for high-risk hypertensive patients compares favorably in cost-effectiveness with many other frequently recommended treatment strategies, and saves money overall for patients aged 60 years and older.
INTRODUCTION
HYPERTENSION is an easily diagnosed and eminently modifiable risk factor for adverse cardiovascular and renal outcomes, including dialysis, stroke, myocardial infarction (MI), and death.1 For many years, the treatment blood pressure (BP) goal for most persons with hypertension in the United States has been less than 140/90 mm Hg. In fact, lowering diastolic BP in particular has been controversial since Stewart,2 Cruickshank et al,3 and others4 drew attention to the "J-shaped" curve: in hypertensive persons with preexisting coronary artery disease, lowering diastolic BP below 85 mm Hg may increase risk. Despite this concern, several authoritative bodies (including the National High Blood Pressure Education Program's Panel on Diabetes and Hypertension, the National Kidney Foundation, and the American Diabetes Association) have suggested that patients with diabetes and those with renal impairment should have a lower target BP than other individuals.5-7 In 1997, the Sixth Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC VI) recommended a lower treatment BP goal of less than 130/85 mm Hg for diabetic persons or those with renal impairment.1 The JNC VI also recommended that the BP goal for patients with hypertension with renal disease and at least 1 g of proteinuria over 24 hours be lower still (<125/75 mm Hg). Neither recommendation had supporting evidence from clinical trials available at that time.
Since the publication of JNC VI in November 1997,1 two studies have reported that more intensive treatment of hypertensive diabetic patients is not only safe, but also reduces morbidity and mortality. In the Hypertension Optimal Treatment (HOT) Study,8 the best results among diabetic patients were seen in those randomized to a diastolic BP goal of 80 mm Hg or less; in the United Kingdom Prospective Diabetes Study (UKPDS) 38,9 the lower BP goal of 150/85 mm Hg or less was associated with improved prognosis. In both studies, regimens consisting of multiple drugs were necessary to achieve these lower goals.
According to authorities responsible for health care financing, there remains great concern (if not doubt) about the economic wisdom of implementing this lower BP treatment goal. In the short term, increased expenditures will be necessary to purchase more pharmaceuticals, pay for the treatment of inevitable additional adverse drug reactions, and fund the extra visits to health care providers. Yet we could expect considerable savings in the longer term if fewer of the treated patients developed heart failure, strokes, or MI, or needed treatment for end-stage renal disease (ESRD). We therefore undertook a formal incremental cost-effectiveness analysis of the new recommended target BP for diabetic patients to estimate the time-based costs and/or benefits of this strategy.
METHODS
OVERVIEW
The incremental cost-effectiveness of moving from the traditional BP target (recommended by JNC V10) to the lower BP goal proposed in JNC VI was calculated using a computer model, by comparing the discounted lifetime costs and years lived by a cohort of 60-year-old diabetic patients, initially free of cardiovascular or ESRD, and following them through the development of morbidity until death (Figure 1). Four major morbid states are incorporated in the Markov model: stroke, MI, heart failure, and ESRD. Strokes are divided into major or minor varieties, because of differences in costs generated. That yields a total of 24 possible comorbid states. Baseline incidence rates for each outcome (Table 1) vary according to decade of age and comorbid status (eg, the risk of MI and heart failure is higher for patients with a previous MI). Age-specific death rates were derived from standard life tables, augmented by excess mortality associated with each of the morbid states. A fixed fraction of strokes and MIs are fatal. Event costs are generated by each stroke and MI, initial hospitalization for heart failure, or institution of ESRD therapy, each of which also generates an annual cost of follow-up, until death supervenes.
|
|
|
|
Figure 1. Schematic representation of the Markov model, having 4 morbid states (stroke, myocardial infarction, heart failure, and end-stage renal disease), and death, used to generate costs and events for moving from the older blood pressure goal (<140/90 mm Hg) to that recommended by the Sixth Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC VI) (<130/85 mm Hg) for diabetic hypertensive patients. For simplicity, 19 states comprising comorbidity and categories of stroke are not shown.
|
|
|
|
|
|
|
Table 1. Parameters Used in Markov Model*
|
|
|
In general, the model calculates, for each year of observation, the proportion of individuals who incur each of the morbid events, the number who die, and the costs associated with new morbid events and the follow-up of prior events. The calculations are continued until all patients have died, after which an estimate of the average life span and total cost of medical care for the cohort can be generated. The effect of moving to the lower BP goal of JNC VI is modeled as a reduction in the relative incidence of the 4 morbid events compared with rates estimated for the "old" BP goal of less than 140/90 mm Hg. As with all cost-effectiveness estimates, results are expressed as (discounted) cost per (discounted) year of life saved.25
BASELINE INCIDENCE RATES FOR MORBID EVENTS
Incidence rates for stroke, MI, heart failure, and ESRD in a high-risk hypertensive population (eg, persons with non-insulin–dependent [type 2] diabetes mellitus) managed under the old guidelines (target BP <140/90 mm Hg) were taken from a variety of population-based studies. We began with a MEDLINE search, using the MESH heading of "diabetes mellitus" and the subheadings "complications, epidemiology, and prevention and control"; the bibliography of each reference was also reviewed. Incidence rates for stroke in the general population were taken from a meta-analysis reported by Taylor et al,20 and multiplied by 2.0, the increase in relative stroke risk due to diabetes in several published studies.14, 26-29 Myocardial infarction rates by age were calculated using prediction equations based on Framingham data, with systolic BP at 140 mm Hg, and including diabetes as a risk factor.30 Incidence rates for heart failure were also based on Framingham data, combining normotensive and "mild" hypertensive rates, and adjusting for the relative increase in risk due to diabetes.12, 22 The incidence of ESRD depends more on the duration of diabetes, rather than age. We used the cumulative incidence of ESRD by duration of type 2 diabetes reported in a population-based survey of Rochester, Minn, and used age 50 years as the average age at diagnosis31 to calculate incidence rates by age.32
In this model (as in clinical practice), having an event or entering a morbid state changes the incidence rates for future events. Survivors of minor stroke have an elevated future risk of all types of stroke. The risk of fatal stroke is increased for survivors of severe stroke. A minor stroke or second severe stroke would have little effect on costs for survivors of severe stroke, so the cost and risk of any nonfatal stroke following major stroke were ignored. Survivors of MI are at increased risk of subsequent MI and for the development of heart failure. Prior MI and heart failure may increase the risk for ESRD, but these effects were not included in the model, since the relative risks (RRs) are not well quantified, and the likelihood of that outcome occurring is low.
SEVERITY, FATALITY, AND LONG-TERM SURVIVAL IN MORBID STATES
The model apportions strokes according to 2 recent studies33-34: severe (25%), minor (50%), or fatal (25%). Long-term survival rates after stroke are taken from 2 recent sources.35-36 The proportion of fatal MIs has been stable at 35% for several years, according to the American Heart Association.23 Because our model already includes recurrent events and an age-specific risk of dying from unrelated causes (discussed below), we adjusted reported overall long-term survival rates for MI and stroke survivors to obtain the increment to mortality rates that is attributable to prior cardiovascular events, but not due to recurrent events.
GENERAL MORTALITY
Mortality from the 4 modeled morbid states enters directly through the fatality of events or the increment to death rates associated with each morbid state. We introduce mortality from all other causes at age-specific rates taken from the 1992 US life table, recalculated after elimination of deaths due to the 4 states already in the model.37 Numerous studies38-40 suggest that persons with type 2 diabetes have approximately twice the risk of dying at any age as the general population, or equivalently, about 7 years' less life expectancy in middle age. To reach this figure for overall mortality in the baseline cohort, we assigned an RR of 1.5 for death from all other causes. The resulting mortality from causes unrelated to the model is assumed to be unaffected by treatment for hypertension.
EFFECTS OF FURTHER REDUCTIONS IN BP
We modeled the effect of more intensive treatment as a lower RR of having each of the hypertension-related morbid events. This method does not distinguish between either beginning antihypertensive drug therapy or intensifying drug treatment, but instead models the effect of reducing BP to 1 of 2 targets. In each case, we used literature data specific to reduction of BP in the 140/90 vs 130/85 mm Hg range. For stroke and MI we incorporated the results of a meta-analysis of epidemiological studies and clinical trials,11 and converted the reported RR for both stroke and MI for an average 6-mm Hg drop in diastolic BP to an RR for the increment of 5 mm Hg recommended by JNC VI. For heart failure, our primary model used an RR of 0.75 for a 10-mm Hg reduction in systolic BP as found in Framingham12; this is somewhat more pessimistic than the RR of 0.59 associated with a 10-mm Hg reduction in systolic BP seen in the Systolic Hypertension in the Elderly Program (SHEP).17 For ESRD, the RR attributed to more intensive treatment was set at 0.80, which is a conservative value reported by Perry et al13 for the difference between 90 and 85 mm Hg diastolic BP in their large Veterans Affairs clinic experience. A more optimistic estimate of 0.70 was reported in a recent meta-analysis19 of trials using angiotensin-converting enzyme inhibitors providing an average reduction of only 5 mm Hg in systolic BP among patients already diagnosed with nondiabetic renal disease.
COSTS
We considered only direct medical costs. Reported costs were converted to 1996 US dollars using the medical component of the consumer price index, and discounted at 3% per year. The initial estimate of the incremental annual cost of intensified treatment was $300 ($200 for additional medications: 0.78 x the average wholesale price for a branded antihypertensive agent41-42 and $100 for 2 "extra" visits to health care providers).43-44 These estimates are very close to average values seen in the HOT and UKPDS studies.8, 44 We included no additional costs for hospitalizations and/or injuries due to adverse reactions, since these are very unusual with current antihypertensive drugs. The average first-year cost of stroke events is given by Taylor et al.20 We assumed that the one third of stroke survivors with severe neurological deficits account for two thirds of the costs, which implies that their costs are twice the average and 4 times higher than the costs of survivors of minor stroke, as suggested by Gage et al.34 The same allocation was applied to annual long-term follow-up costs.34 Nursing home costs following stroke were considered separately. In the model, Taylor et al's estimate of discounted lifetime nursing home costs per stroke was presumed to occur in the first year following a severe stroke. For both stroke and MI, fatal events were assumed to incur half the cost of severe events (excluding nursing home costs), and long-term follow-up costs excluded the costs of recurrent events because those events enter the model directly. The first-year cost of MI was estimated by Lightwood and Glantz21 and includes the expected costs of revascularization procedures under recent utilization patterns. Annual follow-up costs for MI come from the same source. We assume that all new incident cases of heart failure are associated with a hospitalization, generating event costs. In contrast to stroke and MI, we do not explicitly model recurrent heart failure events. Instead, the costs of hospitalization weighted by the fraction of prevalent cases hospitalized each year are included as part of the annual maintenance costs. Annual costs for treating ESRD are provided by the Health Care Financing Administration, which covers approximately 92% of patients with ESRD.24 We used the annual cost of dialysis; transplantation has lower maintenance costs but when the cost of surgery is amortized over expected lifetime, the annual cost is not very different.
The target population is high risk; it is also high cost even after excluding the costs of the 4 disease states modeled explicitly. We included annual costs of $2129 per person for managing type 2 diabetes to represent the costs associated with the underlying conditions that put the population at high risk for cardiovascular and renal diseases.45 We assumed that these costs are not altered by antihypertensive treatment; including them reduces estimated cost-effectiveness because extending life extends the period over which they are incurred.
RESULTS
If one begins in the model with a cohort of 60-year-olds (the average age of patients with type 2 diabetes in the United States31), reducing BP in high-risk individuals from 140/90 to 130/85 mm Hg is associated with a further increase in average life span of 0.48 discounted years; the average 60-year-old person would see actual life expectancy increase from 16.5 to 17.4 years. The total discounted lifetime medical costs (including those associated with additional treatment) decline by $1450 from $59,495 with a goal BP of 140/90 mm Hg, to $58,045 with a goal BP of 130/85 mm Hg. The cost-effectiveness ratio for 60-year-olds (calculated as the incremental lifetime medical costs divided by the change in life expectancy) for the lowered BP goal of JNC VI is therefore negative, because there is overall cost savings, as well as life extension. The discounted per-person anticipated expenditures over time are shown in Figure 2, where the upper curve represents savings from avoiding morbid events and the lower curve indicates cost of incremental medical therapy to achieve the lower BP goal. During the first few years of follow-up, the curves are nearly overlapping, but after age 65 years, the anticipated savings from avoided morbid events rises more rapidly than does the cost of maintaining the lower BP. The breakdown of lifetime costs for each of the types of morbid events is shown in Figure 3, for the BP goals of 140/90 mm Hg and 130/85 mm Hg. At $414 per year, the incremental direct costs of treatment for 60-year-olds are offset by the discounted lifetime costs of avoided morbid events; this parameter ("breakeven" cost of treatment) is of great interest in the sensitivity analyses discussed below.
|
|
|
|
Figure 2. Time course of expenditures (thin line) and savings from avoided medical costs (thick line) for a cohort of 60-year-old high-risk diabetic patients treated to a blood pressure goal below 130/85 mm Hg.
|
|
|
|
|
|
|
Figure 3. Lifetime allocation of resources to a cohort of 60-year-old diabetic hypertensive patients (under baseline conditions) for each type of medical care outcomes under 2 blood pressure goals: less than 140/90 mm Hg (as previously recommended) and less than 130/85 mm Hg (as recommended by the Sixth Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure [JNC VI]). MI indicates myocardial infarction; HF, heart failure; and ESRD, end-stage renal disease.
|
|
|
These calculations were repeated for a cohort of diabetic patients beginning antihypertensive therapy at other ages, with results summarized in the top of Table 2. For 50-year-olds, a BP goal of 130/85 mm Hg would be expected to increase life expectancy from 23.0 to 24.0 years (or 0.48 years, discounted), and increase lifetime medical costs from $61,827 to $62,628. This translates to a cost-effectiveness ratio of $1664 per life year gained, which is still considerably lower than the internationally accepted standard of $100,000 per life year gained. The breakeven annual cost for achieving the lower treatment goal BP for a 50-year-old cohort is $251, or 39% lower than for the 60-year-olds, reflecting the 31% to 89% lower absolute rates of adverse cardiovascular events in the younger group. For a cohort of 70-year-olds, the lowered BP goal of 130/85 mm Hg is associated with an increase in life expectancy of 0.39 years (discounted) and a per-person lowering of lifetime medical costs by $3212, which (like the 60-year-old cohort) results in a net cost savings, compared with the BP treatment goal of 140/90 mm Hg. The breakeven annual cost for 70-year-olds is even higher (at $634) than in 60-year-olds, again because of the greater absolute risk of all adverse events in the older group.
|
|
|
|
Table 2. Sensitivity Analyses Varying Baseline Parameters*
|
|
|
Sensitivity analyses were also performed which varied the baseline assumptions rather widely. The results are shown in Table 2. Despite systematic alterations in all of the important conditions, there were few changes in the estimates of years of life saved, lifetime medical costs, or breakeven annual costs, when lowering BP from less than 140/90 to less than 130/85 mm Hg. Perhaps the most important of these analyses reduces the RR of the lower BP treatment goal ("treatment effectiveness" in the penultimate line of Table 2) to the lower limit of the 95% confidence intervals listed in Table 1. This method of estimating the economic consequences of intensifying BP treatment shows that, even with expectations of benefit at the lower limit of current estimates, lowering BP from below 140/90 to below 130/85 mm Hg should increase discounted life years by 0.38, decrease discounted lifetime medical costs by $661, and decrease the "breakeven cost for treatment" by only 15%. All of these sensitivity analyses resulted in alterations in the expected direction from the baseline conditions: there was an increase in cost per year of life saved with an increase in likelihood of an adverse event, reduction in mortality from any competing risk, or increase in the ultimate cost of adverse events. Changes that increased the cost of funds (discount rate) or decreased the effectiveness of adverse event reduction due to BP lowering treatment similarly increased the cost per year of life saved.
COMMENT
In this attempt to estimate the lifetime costs associated with adverse sequelae of hypertension in a 60-year-old cohort of diabetic hypertensive patients, the economic benefits of achieving the recently recommended lower BP goal of 130/85 mm Hg (vs the older goal of 140/90 mm Hg) actually far outweigh the increased costs of treatment, primarily because of the number of clinical events (heart failure, stroke, MI, or ESRD) that are prevented. Although a similar overall economic benefit was not seen in 50-year-olds (presumably because of their lower absolute risk of these events), the increased prevalence of diabetes in older adults makes it likely that, on a population basis, the recommended strategy to decrease BP to below 130/85 mm Hg appears to be sound economic, as well as therapeutic, policy. Intensive lowering of BP in older high-risk individuals appears to be one of the very few chronic medical treatments that is overall cost saving.
The cost-effectiveness and economics of BP control have been the subjects of a few reports over the years, but few have compared the costs and economic benefits of achieving a lower BP goal in large numbers of patients. One of the best-known examples of estimating the cost-effectiveness of antihypertensive drug therapy was based on costs at Boston, Mass, retail pharmacies. Using an early version of the Coronary Drug Project computer model, Edelson et al46 concluded that the cost per year of life saved increased across 4 different antihypertensive drugs: propranolol, hydrochlorothiazide, prazosin, and captopril. The cost-effectiveness of BP control with captopril in the Captopril Cooperative Study Group's trial in young persons with type 1 diabetes with proteinuria and renal impairment18 showed an overall lifetime savings in direct medical costs of $32,550 per patient with captopril (even before its price decreased due to availability of the generic formulation).47
Our model estimates the cost savings from reduced morbidity independently from the costs of treatment. Therefore, we can estimate a breakeven cost of treatment from the estimates of morbidity savings alone, independent of the cost of any specific drug and its economic sequelae (eg, extra visits to health care providers to treat side effects or hospitalizations for drug-related serious adverse effects). This breakeven cost provides an upper limit on how much additional treatment can be provided each year, before the therapeutic approach begins to lose money overall.
The most recent report of the economic consequences of improving BP control in diabetic patients comes from the UKPDS 38,9 in which a comparison was made between 2 BP goals: 180/105 or lower or 150/85 mm Hg or lower. In the clinical trial, the lower BP goal was associated with a 24% reduction in diabetes-related end points, 32% reduction in diabetes-related deaths, 44% reduction in strokes, and 37% reduction in microvascular complications. All of the estimates of the effectiveness of lowering BP in UKPDS are similar to, but more optimistic than, those used in our analyses: 0.56 (95% confidence interval [CI], 0.35-0.89) for stroke, 0.79 (95% CI, 0.59-1.07) for MI, 0.44 (95% CI, 0.20-0.94) for heart failure, and 0.58 (95% CI, 0.15-2.21) for renal failure. The accompanying economic analysis, performed simultaneously with, and based on the results of UKPDS 38, indicates that the lower BP target was cost saving over the 8.4 years of therapy, since the costs of the more intensive treatment strategy were overcome by the avoidance of expensive clinical events.44 These investigators did not consider the potential implications of a still lower BP goal (eg, <130/85 mm Hg), as we have done. Nonetheless, even when pessimistic discount rates were applied only to the beneficial effects (and not to the costs) of the more intensive treatment regimen in the UKPDS, there was overall savings compared with allowing the diabetic patients to maintain a sustained BP as high as 180/105 mm Hg. These results from a randomized clinical trial support the basic conclusion seen in our analyses (which are based on epidemiological estimates and meta-analyses), that more intensive lowering of BP is associated with an overall cost savings in high-risk hypertensive patients.
Our methods and conclusions have a number of inherent limitations. Most of the baseline assumptions for our computer model are based not on direct observation or clinical trial data but instead on epidemiological studies and meta-analyses. Our analyses do not consider the costs and benefits of treatment for hypertensive diabetic patients whose BP is not controlled, even to less than 140/90 mm Hg; instead, we focused on a comparison of the 2 goals of treatment. Our model does not explicitly identify diabetic patients who, at baseline, already have manifestations of disease that increase the risk of a subsequent morbid event (eg, transient ischemic attack or stable angina pectoris). If we incorporate in the computer model a baseline prevalence of preexisting conditions found in 55- to 64-year-old American diabetics in the Third National Health and Nutrition Examination Survey (NHANES III) (heart failure, 17%; MI, 16%; stroke, 14%; ESRD, 1.8%), our major conclusions change minimally: the lower BP goal of JNC VI would be expected to add 0.41 discounted years of life and reduce lifetime health care costs by $874.
Our model may underestimate the cost-effectiveness of the recently recommended BP goal of less than 130/85 mm Hg for several reasons. In general, we preferred to use conservative estimates whenever a choice was possible. Our model is somewhat more conservative than that of the UKPDS, in that our comparison is with a group whose BP is maintained at less than 140/90 mm Hg (vs <180/105 mm Hg in the UKPDS). We used complication rates for high-risk patients, which are more typical of Americans, and discounted both costs and effects at 3% per annum, as recently recommended by the Panel on Cost-effectiveness in Health and Medicine.25 Our model neglects the prevention of second and subsequent MIs (despite the frequent occurrence of recurrent MI in diabetic patients), because clinical trials typically report only on the first occurrence of an end point. Heart failure is modeled as a finite risk (and cost) per year of hospitalization, and therefore ignores both nonhospitalized heart failure and additional treatment costs for the several effective therapies becoming more widely used for this disease. End-stage renal disease is modeled in a similar fashion, and the additional costs of both treatment (short of dialysis or transplantation) and more frequent monitoring of renal function are omitted. The rather large benefit of the lower initial BP in reducing the rate of progression of hypertension to higher stages16 is also neglected in these calculations. Potential benefits of lowered BP in treatment of retinopathy in diabetic patients,9 improved quality of life,8 and cognitive function8 seen among more intensively treated hypertensive patients in recent studies were also ignored, which would contribute to a more optimistic cost-effectiveness and cost utility of the lower BP goal. Last, in the sensitivity analyses, even when the most pessimistic estimates of the reduced risk associated with adverse outcomes for the lower BP goal are used to estimate outcomes for patients older than 60 years, there is overall cost savings.
There are, of course, perspectives from which these putative benefits are not seen. For government, and perhaps society as a whole, it may be important to consider the greater life expectancy for individuals older than 65 years with the lower BP goal as a negative, since such persons would usually be entitled to Social Security and other pension benefits, and lengthening life would be likely to increase Medicare costs. The perspective for our calculations was limited to health care costs,48 which appear to be affected positively for individuals older than 60 years.
The actual annual cost of intensifying antihypertensive drug therapy in all high-risk patients is difficult to estimate, partly because of uncertainty regarding the frequency of "extra" visits to health care providers. It is reasonable to expect that more frequent visits early on to achieve goal BP will result in fewer visits (and other adverse events) later. Most of the cost data about intensified treatment focus on the cost of additional medications necessary to achieve the lower BP goal. Integration of Figure 3 from the UKPDS 38 report indicates that, on average, each patient in the tight BP control group received (at least) 0.78 additional antihypertensive medications compared with the control group9; preliminary reports from the HOT Study provide a very similar figure when comparing the groups treated to diastolic BPs of 90 and 85 mm Hg. Although cost estimates have not yet been reported for the HOT Study, antihypertensive drugs contributed only 7% or 14% of the total costs in the less or more intensively treated diabetics, respectively, in UKPDS 38.44 The cost of additional antihypertensive medications in the more intensively treated group was £73 (or $118) per year, which was associated with an overall savings of £210 ($340) per patient due to reduced cost of morbidity.44 The vast majority of the marketed antihypertensive agents in the US cost less than the breakeven cost calculated by our model ($414 per year) for 60-year-olds.41-42 The cost of antihypertensive medications currently amounts to only 24% of the total cost of care for hypertension in the United States23; our analyses (and those of the UKPDS44) conclude that the marginal cost of adding 1 or more BP-lowering medications to the regimen of high-risk patients is small, especially when compared with the discounted long-term benefits.
The economic analysis presented here shows that, compared with the prior BP goal of 140/90 mm Hg, the newly recommended treatment goal for BP of less than 130/85 mm Hg in persons with diabetes and those with renal impairment is associated with an overall cost saving for individuals older than 60 years under a wide variety of initial assumptions. Even in younger individuals, the cost per year of life saved is smaller than many other frequently recommended treatment modalities, including advice to stop smoking cigarettes ($258749), thrombolytic therapy for acute MI ($32,67850), or drug treatment of high blood cholesterol levels for primary ($11,040-$52,46351) or secondary prevention ($3800-$13,30052). It appears, therefore, that the lower BP goal recommended by JNC VI for high-risk patients is not only effective in preventing expensive cardiovascular events but also (for many patients and payers) cost saving in the long term.
AUTHOR INFORMATION
Accepted for publication August 3, 1999.
This work was supported in part by an unrestricted educational grant from Hoechst Marion Roussel Inc, Kansas City, Mo, to Rush-Presbyterian-St Luke's Medical Center on behalf of the authors, and the National Institute on Aging, Bethesda, Md (K01 AG 00703).
Presented in part at the 13th Annual Scientific Meeting of the American Society of Hypertension, New York, NY, May 14, 1998.
Reprints: William J. Elliott, MD, PhD, Department of Preventive Medicine, Rush-Presbyterian-St Luke's Medical Center, 1700 W Van Buren St, Suite 470, Chicago, IL 60612-3624.
From the Department of Preventive Medicine, Rush Medical College of Rush University at Rush-Presbyterian-St Luke's Medical Center (Drs Elliott and Black), and the Irving B. Harris Graduate School of Public Policy Studies and the Department of Economics, The University of Chicago (Dr Weir), Chicago, Ill.
REFERENCES
| |
1. The Sixth Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC VI). Arch Intern Med. 1997;157:2413-2446.
ABSTRACT
2. Stewart IMDG. Relation of reduction in pressure to first myocardial infarction in patients receiving treatment for severe hypertension. Lancet. 1979;1:861-865.
ISI
| PUBMED
3. Cruickshank JM, Thorpe JM, Zacharias FJ. Benefits and potential harm of lowering high blood pressure. Lancet. 1987;1:581-584.
ISI
| PUBMED
4. Farnett L, Mulrow CD, Linn WD, Lucey CR, Tuley MR. The J-curve phenomenon and the treatment of hypertension: is there a point beyond which pressure reduction is dangerous? JAMA. 1991;265:489-495.
ABSTRACT
5. National High Blood Pressure Education Program Working Group report on hypertension in diabetes. Hypertension. 1994;23:145-158.
ABSTRACT
6. Jacobson HR, Striker GE. Report on a workshop to develop management recommendations for the prevention of progression in chronic renal disease. Am J Kidney Dis. 1995;25:153-156.
7. American Diabetes Association. Standards of medical care for patients with diabetes mellitus. Diabetes Care. 1998;21(suppl):S23-S31.
8. Hansson L, Zanchetti A, Julius S, et al on behalf of the HOT Study Group. Effects of intensive blood pressure lowering and low-dose aspirin in patients with hypertension: principal results of the Hypertension Optimal Treatment (HOT) randomised trial. Lancet. 1998;351:1755-1762.
FULL TEXT
|
ISI
| PUBMED
9. Turner R, Holman R, Stratton I, et al for the United Kingdom Prospective Diabetes Study Group. Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes: UKPDS 38. BMJ. 1998;317:707-713.
10. The Fifth Report of the Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure (JNC V). Arch Intern Med. 1993;153:154-186.
FULL TEXT
|
ISI
| PUBMED
11. Collins R, Peto R, MacMahon S, et al. Blood pressure, stroke, and coronary heart disease, part 2: short-term reductions in blood pressure: overview of randomised drug trials in their epidemiological context. Lancet. 1990;335:827-838.
FULL TEXT
|
ISI
| PUBMED
12. Levy D, Larson MG, Ramchandran SV, Kannel WB, Kalon KL. The progression from hypertension to congestive heart failure. JAMA. 1996;275:1557-1562.
ABSTRACT
13. Perry HM, Miller JP, Fornoff JR, et al. Early predictors of 15-year end-stage renal disease in hypertensive patients. Hypertension. 1995;25:587-594.
FREE FULL TEXT
14. Stamler J, Vaccaro O, Neaton JD, Wentworth D for the MRFIT Research Group. Diabetes, other risk factors, and 12-year cardiovascular mortality for men screened in the Multiple Risk Factor Intervention Trial. Diabetes Care. 1996;16:434-444.
ABSTRACT
15. Rose G. Epidemiology. In: Marshall AJ, Barritt DW, eds. The Hypertensive Patient. Kent, England: Pitman Medical Publishers; 1980:1-21.
16. Moser M, Hebert PR. Prevention of disease progression, left ventricular hypertrophy and congestive heart failure in hypertension treatment trials. J Am Coll Cardiol. 1996;27:1214-1218.
ABSTRACT
17. Kostis JB, Davis BR, Cutler J, et al for the SHEP Cooperative Research Group. Prevention of heart failure by antihypertensive drug treatment in older persons with isolated systolic hypertension. JAMA. 1997;278:212-216.
ABSTRACT
18. Lewis EJ, Hunsicker LG, Bain RP, Rohde RD. The effect of angiotensin converting enzyme inhibition in diabetic nephropathy. N Engl J Med. 1993;329:1456-1462.
FREE FULL TEXT
19. Giatras I, Lau J, Levey AS for the Angiotensin Converting-Enzyme Inhibition and Progressive Renal Disease Study Group. Meta-analysis of the effectiveness of ACE-inhibitors in preventing the progression of renal disease. Ann Intern Med. 1997;127:337-345.
FREE FULL TEXT
20. Taylor TN, Davis PH, Torner JC, Holmes J, Meyer J, Jacobson MF. Lifetime cost of stroke in the United States. Stroke. 1996;27:1459-1466.
FREE FULL TEXT
21. Lightwood JM, Glantz SA. Short-term economic and health benefits of smoking cessation: myocardial infarction and stroke. Circulation. 1997;96:1089-1096.
ISI
| PUBMED
22. Kannel WB, Belanger AJ. Epidemiology of heart failure. Am Heart J. 1991;121:951-957.
FULL TEXT
|
ISI
| PUBMED
23. American Heart Association. 1999 Heart and Stroke Facts, Statistical Update. Dallas, Tex: American Heart Association; 1999.
24. Department of Health and Human Services. Health Care Financing Research Report: End Stage Renal Disease, 1991. Baltimore, Md: Health Care Financing Administration; 1991.
25. Weinstein MC, Seigel JE, Gold MR, Kamlet MS, Russell LB. Recommendations of the Panel on Cost-effectiveness in Health and Medicine. JAMA. 1996;276:1253-1258.
ABSTRACT
26. Manson JE, Colditz GA, Stampfer MJ, et al. A prospective study of maturity-onset diabetes mellitus and risk of coronary heart disease and stroke in women. Arch Intern Med. 1991;151:1141-1147.
ABSTRACT
27. Kittner SJ, White LR, Losonczy KG, Wolf PA, Hebel JR. Black-white differences in stroke incidence in a national sample: the contribution of hypertension and diabetes mellitus. JAMA. 1990;264:1267-1270.
ABSTRACT
28. Wolf PA, D'Agostino RB, Belanger AJ, Kannel WB. Probability of stroke: a risk profile from the Framingham Study. Stroke. 1991;22:312-318.
ABSTRACT
29. CDC Diabetes Cost-effectiveness Study Group. The cost-effectiveness of screening for type 2 diabetes. JAMA. 1998;280:1757-1763.
FREE FULL TEXT
30. Anderson KM, Odell PM, Wilson PWF, Kannel WB. Cardiovascular disease risk profiles. Am Heart J. 1990;121:293-298.
31. American Diabetes Association's Diabetes 1997 Vital Statistics. Alexandria, Va: American Diabetes Association; 1997.
32. Humphrey LL, Ballard DJ, Frohnert PP, Chu CP, O'Fallon WM, Palumbo PJ. Chronic renal failure in non-insulin dependent diabetes mellitus: a population-based study in Rochester, MN. Ann Intern Med. 1989;111:788-796.
33. Jorgensen HS, Nakayama H, Raaschou HO, Olsen TS. Acute stroke: prognosis and a prediction of the effect of medical treatment on outcome and healthcare utilization: The Copenhagen Stroke Study. Neurology. 1997;49:1335-1342.
FREE FULL TEXT
34. Gage BF, Cardinalli AB, Albers GW, Owens DK. Cost-effectiveness of warfarin and aspirin for prophylaxis of stroke in patients with nonvalvular atrial fibrillation. JAMA. 1995;274:1839-1845.
|
