[ Scientific Activities - Actividades Científicas ]
Honorary Committee Lecture
Cost and Cost effectiveness of Hypertension
Impact on Care
Istituto di Clinica Generale e Terapia Medica and
Centro di Fisiologia Clinica e Ipertensione,
Università di Milano,
Ospedale Maggiore and
Istituto Auxologico Italiano,
The impact of cardiovascular disease on health and health care costs
For nearly three decades, in the United States and in most (though not all) industrialized countries, mortality from coronary heart disease and stroke has been continuously declining. There is data (Source National Center for Health Statistics, adjusted for age, race and sex) compiled for the United States by the National Heart, Lung and Blood Institute showing that between 1972 and 1995 age-adjusted mortality from coronary heart disease declined by 54% and mortality from stroke by 59% (1). The same data, however, shows some disturbing trend: over the last 5 years the rates of decline have apparently leveled off and the rate of stroke may be actually increasing.
Furthermore, the death rates and the increasing prevalence of elderly people in most industrialized countries makes the absolute decline in cardiovascular death less impressive.
Finally, the prolongation of life and the changes in lifestyle seen in developing countries has made the World Health Organization calculate that cardiovascular disease will still be the first cause of death and disability in the next 20 years, at least (2).
In 1993, cardiovascular disease cost the United States an estimated sum of US$ 210 billion (3). Hypertension is undoubtedly one of the most important causes of this cardiovascular morbility, and not only because its well known role in determining stroke and coronary disease. Recent data indicate that hypertension precedes approximately 90% of all cases of heart failure (4), a condition which is continuously increasing (465,000 new cases diagnosed each year in the United States, 5), and is particularly expensive because of frequent and prolonged hospitalization (heart failure is the most common hospital discharge diagnosis for patients older than 65 years, 5).
End-stage renal disease has also been remarkably increasing: between 1982 and 1995, there was a 2.5 increase in its incidence in the United States (6), and similar increases have been reported from Europe. Hypertension ranks second, after diabetes, among the causes of end-stage renal disease. The very high cost of management of renal failure by dialysis and transplantation need not be underlined.
On the whole, health care expenditures in the United States have increased from 15% of the gross domestic product in 1993 to an estimate of 20% in the year 2000 (7).
The importance of cost-effectiveness analysis
The huge cost that disease imposes to societies and the amount of financial resources that societies must devote to disease prevention and care justify the increasing interest placed on the assessment of the cost-effectiveness of diagnostic and therapeutic interventions. Doctors often feel uneasy with this analysis, not only because of their ignorance of economics, but mostly because their mentality and education make them more susceptible to the appeal of being effective and beneficial than to the concerns of cost. However, the availability of costly diagnostic procedures and therapeutic interventions is increasing while the amount of resources that each society can devote to health is not unlimited. Therefore, the issue of assessing the cost-effectiveness of medical care is a reasonable and realistic one (8).
On the other hand, advocating the use and the usefulness of cost-effectiveness analysis should not conceal the numerous uncertainties, approximations and arbitrary criteria inherent in cost-effectiveness analysis and in the decisions based on them (8). These uncertainties and approximations are particularly large when we are dealing with preventive interventions, as is the case for antihypertensive therapy, as in prevention the costs of today must be compared to the benefits in a less or more distant future. There is no doubt, however, that, if we want to limit the continuing escalation of the cost of disease and health care, this can only be done through increasing investment on prevention.
A discussion of cost-effectiveness of the management of hypertension should consider, in succession, the value and limitations of the assessment of cardiovascular risk (that antihypertensive treatment intends to prevent), the assessment of the costs and that of the benefits of antihypertensive treatment.
Assessment of cardiovascular risk
Cardiovascular risk is traditionally expressed in relative terms, namely how many times the risk of suffering a cardiovascular event is increased in the presence of a given risk factor, compared with the average risk of the normal population. In taking decisions about treatment, it is probably more correct, or at least more informative, not to consider any risk factor in isolation, but calculate the global risk of cardiovascular disease resulting from the constellation of risk factors often coexisting in the same individual. Global risk is more appropriately expressed in absolute, rather than relative, terms, as the percentage of subjects with a given risk profile that are expected to develop a cardiovascular event within 5 or 10 years. The recent guidelines on the management of hypertension prepared by the World Health Organization and the International Society of Hypertension (9) have provided a simple table, by which hypertensive patients can be stratified in four classes of overall cardiovascular risk, and calculated that hypertensives at low global risk have a probability of less than 15% of suffering a major cardiovascular event in 10 years; those at medium risk a probability of between 15 and 20%; those at high risk a probability between 20 and 30%; and those at very high risk a probability of over 30%. A limitation intrinsic to this table, as well as other tables for risk calculation, is that it uses an equation derived from the extensive experience of the Framingham Heart Study (10), and caution should be taken before taking for granted that the data obtained from this study can be extrapolated to the rest of the westernized world. However, that the risk figures given in this study are a reasonable approximation to be used in risk prediction is supported not only by the Framingham Heart Study data, but also by the event rates occurred, for comparable risk profiles, in the placebo-treated or less actively treated groups of previous trials, such as the British Medical Research Council Trial on mild hypertension and the United States Hypertension Detection and Follow-up Program (11).
Assessment of the costs of antihypertensive therapies
Both direct and indirect costs must be assessed. Direct costs (medical supervision, hospital admissions, drugs) can be calculated with reasonable precision. However, this precision is lower longer is the prospective duration of treatment, and this applies to all primary prevention interventions, such as antihypertensive treatment (8). In this cases, several arbitrary criteria must be set and uncertain forecasts be made: for example, how much inflation will affect costs, how much drug costs will change (the cost of currently used drugs will certainly decrease, that of new drugs will be higher), and so on.
Prospective indirect costs are even more difficult to assess, and their calculation is often the result of arbitrary decisions. The costs of possible events can be calculated, to some extent at least: the cost of supporting for 30 years a man aged 40 years who has suffered an incapacitating stroke can be estimated approximately. How much a prolonged absence from work because of a major cardiovascular event may cost, in terms of decreased wealth production, can also be estimated, though much more approximately. Finally, it is certainly very difficult to translate sufferings and death into money, although insurance companies, lawyers and judges must often exercise their judgment in this matter (8).
Restricting assessment of costs to the easier exercise of calculating direct costs only, in 1996 we have calculated that the cost of one year of antihypertensive treatment may be up to about US$ 550 (US$ 1/day for drugs and US$ 200/year for medical visits and diagnostic examinations (12)). According to estimations made in 1997 by the American Heart Association (13), about US$ 14 billions were going to be spent in the United States during 1998 for medications and healthcare providers for the management of hypertension. These figures satisfactorily correspond to our calculations of costs of medicines and medical care, if provision is made of the fact that no more than 50% of hypertensives in the United States receive treatment, and no more than 50% of them have their blood pressure well controlled by treatment. It is likely, therefore, that good and generalized treatment of hypertension in the United States may cost twice as much as the figure spent in 1998, although, prospectively, good and generalized prevention may markedly reduce the direct costs for hospitals and nursing homes, as compared to those incurred in 1998, and the indirect costs due to lost productivity.
Assessment of the benefits of antihypertensive treatment
The assumption is often made that the benefit of antihypertensive therapy has been correctly measured by intervention trials, and that data obtained by trials can be accurately entered in cost-effectiveness calculations. There is no doubt that randomized trials of antihypertensive therapy have been an invaluable instrument for establishing that treatment of hypertension has indeed been beneficial. However, as also recognized by the major guidelines (9,14), trials have been unable to provide a correct quantitative calculation of these benefits, for at least two major reasons (8).
The first reason is that the average treatment benefit calculated from meta-analysis of available trials (15) is that produced by a modest decrease in blood pressure (5-6 mmHg in diastolic pressure), and it is arbitrary and incorrect to take it as the benefit of antihypertensive treatment tout court. Table 2 from the 1999 WHO/ISH guidelines indicates that a reduction of diastolic blood pressure twice as large (10 mmHg) as that obtained in trials is expected to produce benefits at least 1.5 times as large. Recent results of the Hypertension Optimal Treatment (HOT) trial have shown that intensive lowering of blood pressure is associated with rates of cardiovascular mortality and morbidity much lower than those observed in trials (16), and it has been calculated that the benefits of intensive blood pressure reduction demonstrated by the HOT study are even greater than those predicted by WHO/ISH guidelines for reductions of systolic blood pressure of 20 mmHg and of diastolic blood pressure of 10 mmHg (11).
Another important limitations of trials is that trials can only assess short-term benefits, and therefore are an excellent measure of the benefit of antihypertensive treatment in the elderly, such as in subjects older than 70 years whose life expectancy is not much longer than the average duration of an interventional trial (8,12). In younger patients with mild to moderate hypertension, the goal of treatment are long-term ones and benefits cannot correctly be assessed by relative short-term trials (8,12).
The actuarial versus the intervention trial approach in evaluating cost-effectiveness of antihypertensive therapy
We have previously suggested that a more faithful assessment of the benefits of antihypertensive therapy to use in cost-effectiveness calculations may be provided by actuarial data indicating the reductions of life expectancy caused by a given increase in blood pressure at any given age. Useful actuarial data were provided in 1961 by the Statistical Bureau of the Metropolitan Life Insurance Company, based on experience accumulated during the period 1935-1954 when no antihypertensive therapy was available (17).
Here we include the results of some calculations we have made of the direct costs (medical supervision and drugs) of antihypertensive management per additional year of life gained based on actuarial data (12). The estimated costs are about US$ 2000 in 45-year-old men with blood pressures of 150/100 mmHg and about US$ 3000 in women of the same age and blood pressures; even when blood pressure is only 145/95 mmHg, the estimated costs per year of life gained are approximately US$ 4000 and US$ 6000 in 45-year-old men and women, respectively, which is more than 20-fold less than the costs calculated from the short-term outcomes of the trials.
These results also shows another striking difference between the results of the actuarial and the trial approach: the cost per year of life gained increases with increasing age, as the additional years of life to be gained are expected to be fewer. This disproves the paradox, suggested by calculations based on trials, that costs per year of life gained are very high in the young and quite low in the old. On the contrary, the actuarial approach indicates antihypertensive treatment to be more cost-effective in the young than in the old, although, of course, the total life cost of treatment is greater in the young.
The benefits of lowering blood pressure calculated from the actuarial data estimated by the Statistical Bureau of the Metropolitan Life Insurance Company (17) are in line with observational data of secular trends in hypertension control between 1950 and 1990, reported by the Framingham Heart Study, showing that long-term treatment of long-term sustained hypertension is associated with a striking 60% reduction in 10-year cardiovascular mortality (18). The data of the HOT study, though obtained after a relatively short-term treatment, also confirm that intensive lowering of blood pressure can reduce the expected risk by more than 50% (11).
The effect of a more correct and optimistic assessment of benefit for antihypertensive treatment decisions is obvious: a frequent expression of cost-effectiveness is the calculation of the number of patients to be treated for 10 years to avoid a cardiovascular event, and a more optimistic assessment of benefit would certainly improve cost-effectiveness estimations, and favour more liberal decisions on treatment initiation or continuation. A further warning is about using the absolute benefit of treatment as the only guideline for treatment decisions: it is true that higher risk patients are likely to have a higher absolute benefit, but data from several trials, including the recent HOT study (11), have shown that higher is the risk of the patients at initiation of treatment higher remains the residual risk of cardiovascular morbidity and mortality the patients remain exposed despite antihypertensive treatment.
In conclusion, as we have previously remarked (8,12), there is still a large amount of uncertainty concerning how to assess benefits and calculate costs, and well-balanced information on all possible ways of assessment should be made available to patients, doctors and health providing agencies. It is understandable, but not commendable, that third parties, particularly when they are national health services, choose the most conservative cost-benefit analysis and translate them into prescriptive guidelines. On the other hand, the doctors task is to prioritise the patients interests and, whenever treatment is safe and well-tolerated (as is most often the case with antihypertensive treatment), to prefer a reasonable overestimation of benefits to strategies based on short-term benefits that are at best the minimum ascertained ones.