An ounce of prevention? (Health Care Technology).

The pervasive Increase in the cost of medical care has stimulated myriad responses that can be broadly subdivided into two categories: (1) decreasing the cost of medical care for symptomatic diseases and (2) decreasing the Incidence of disease. Decreasing the cost of care can involve medical management strategies deployed by managed care organizations--implementing practice guidelines, monitoring hospital length of stays, physician profiling, and medical necessity criteria. In contrast, prevention focuses on modifying risk factors, either through patient education and lifestyle changes or primarily drug interventions when risk factors are considered unmodifiable.

While preventive health care is intuitively attractive, both from a disease morbidity and cost of care aspect, it is most effective when the natural history of a disease can be precisely predicted and when there is effective therapy to modify the risk factor. The evolving treatment of HIV infection is such an example. The natural history of HIV infection is closely tied to the viral load, which represents the risk factor for opportunistic Infections and other causes of HIV morbidity and mortality. Titers of the viral load can be monitored and antiviral therapy Implemented or changed if the viral load exceeds certain thresholds. The ability to monitor treatment response through serial viral load measurements and the availability of a growing menu of antiviral therapy, including most prominently a variety of protease inhibitors, is credited with the decreasing morbidity and mortality of HIV.

In contrast, if the natural history is uncertain, perhaps due to its multifactorial nature, and treatment not entirely effective, there will inevitably be inefficiency in preventive treatment. An unpredictable natural history will result in many patients who are treated who were not destined to develop the disease. When the preventive intervention is less than 100 percent effective, there will be patients treated who nevertheless develop the disease. The inefficiency of treatment must then be balanced against the effectiveness of preventing versus treating the targeted disease, and the comparative cost of each approach.

Complicated balancing act

In this complicated balancing act, the inefficiencies of preventive therapy may be buried beneath the surface, due in part to the method of reporting and perhaps to the enthusiasm for preventive medicine in general. For example, results of preventive trials are often most prominently reported in terms of the relative risk reduction, for example, the percentage decrease in the incidence of a certain outcome between the treatment and placebo group.

However, the absolute reduction in outcomes may provide a better measure of the treatment efficiency.

In a hypothetical example of cholesterol lowering therapy, drug therapy may be associated with a reduction of cardiac mortality from 10 percent in the placebo group to 5 percent in the treatment group. While this may be expressed as a highly significant 50 percent relative risk reduction for mortality, the absolute risk reduction is only 5 percentage points. An interesting way of interpreting this data is to consider how many patients must be treated for one to benefit, referring to the number needed to treat (NNT), in order to have one additional patient alive at the end of the trial than if all patients were given a placebo. Essentially, if only 5 percent of the group benefited from the treatment, 20 patients must be treated to benefit one patient.

The absolute risk reduction and the number needed to treat will vary according to the predictive value of the risk factor, while the relative risk reduction may not. For example, in another hypothetical group of patients with higher underlying risk of cardiovascular mortality, the same cholesterol lowering therapy may reduce the incidence of cardiovascular death from 50 percent in the placebo group to 25 percent in the treatment group. While the relative risk reduction is still 50 percent the absolute risk reduction is 25 percentage points and the number needed to treat drops down to four. More dramatic absolute risk reductions can occur as the incidence of disease rises in the placebo group, due perhaps to an improved predictive value of the risk factor used for patient selection.

Preventive health measures

To provide perspective, the following discussion focuses on the absolute risk reduction and the numbers needed to treat associated with two preventive health measures that have been widely researched and endorsed: (1) preventing coronary artery disease by treating hypercholesterolemia and (2) preventing fractures by treating osteoporosis.

1. Hypercholesterolemia

Guidelines for treating hypercholesterolemia have been published by the National Cholesterol Education Program of the National Heart, Lung and Blood Institute. (1) These guidelines established an LDL level of less than 130 mg/dL as desirable for those without preexisting coronary heart disease, and less than 100 mg/dL for those with preexisting coronary heart disease.

While dietary therapy and lifestyle alterations are always the first line of treatment, many patients will require additional drug therapy to achieve these LDL goals. Throughout the 1990s a series of trials examined the treatment effectiveness of HMG CoA reductase inhibitors, a class of cholesterol lowering agents also referred to as statins. (These trials are often referred to as the "statin trials.") Several examples of the statin trials are profiled in Table 1. While a number of different outcomes were studied, the risk reduction in fatal events is profiled. since by definition, treatment therapy was ineffective.

As can be seen from Table 1, the relative risk reduction remains within a relatively narrow range, while the numbers needed to treat exhibits a broader range. The WOSCOPS trial focused on secondary prevention; for example, modifying a risk factor in the absence of symptomatic disease. Here, not unexpectedly the number needed to treat is the highest.

2. Osteoporosis

Detecting and treating osteoporosis is a preventive health measure designed to reduce the incidence of fractures, typically of the hip or vertebrae. Low bone mass is the most predictive risk factor, but fracture risk is multifactorial in nature. Recently, the National Osteoporosis Foundation recommended treatment for those whose bone mineral density fell below a certain threshold. (6) While exercise, adequate diet, and smoking cessation are important primary prevention measures, many women with low bone mass will require drug therapy.

Alendronate is a drug that was approved by the FDA in 1993 to treat osteoporosis. A variety of large. randomized trials focusing on alendronate were published throughout the 1990s. Only two trials focused on the incidence of fractures, and even then on vertebral rather than hip fractures. While vertebral fractures can be associated with significant morbidity, hip fractures are thought to be the most morbid of osteoporosis complications. However, due to the low incidence of fractures in the relatively short timeframe, many studies focused instead on the intermediate outcome of bone mineral density. Therefore, the number needed to treat to avoid one incidence of hip fracture is not known, but is certainly considerably higher than those reported in Table 2.

Conclusion

In general, compared to the relative risk reduction, the number needed to treat is not prominently reported and in some cases this statistic must be independently calculated from the data presented. Until we are able to Identify more predictive risk factors, there will always be inefficiencies in preventive medicine, and many must be treated for one to benefit. However, the absolute risk reduction and the number needed to treat are useful measures to highlight what can actually be achieved with preventive therapy.

ABLE 1

PROFILE OF REDUCTION IN FATAL EVENTS FOR PATIENTS TREATED FOR
HYPERCHOLESTEROLEMIA

Trial                  Outcome            Relative Risk
                       Studied            Reduction (%)


4S Group (2)           Total mortality         29%
Pts with history
of CHD and
elevated cholesterol

CARE Trial (3)         Fatal CHD               24%
Pts with history of
CHD but with
average LDL levels

LIPID Trial (4)        Fatal CHD               24%
Pts with history of
CHD, broad range
of cholesterol levels

WOSCOPS Trial (5)      Nonfatal MI             31%
Pts with elevated      or fatal coronary
cholesterol, no CHD    event

Trial                  Absolute Risk                      Number
                       Reduction                          Needed
                       (percentage points)               to Treat

4S Group (2)           6.7 percentage points, from          15
Pts with history       22.6% in placebo group to 15.9%
of CHD and             in treatment group
elevated cholesterol

CARE Trial (3)         3 percentage points, from 13.2%      33
Pts with history of    in placebo to 10.2%
CHD but with           in treatment group
average LDL levels

LIPID Trial (4)        1.9 percentage points, from 8.3%     52
Pts with history of    in the placebo group to 6.4%
CHD, broad range       in the treatment group
of cholesterol levels

WOSCOPS Trial (5)      2.4 percentage points, from 7.9%
Pts with elevated      in placebo group to 5.5%
cholesterol, no CHD    in treatment group 96

ABLE 2

TRIALS FOCUSING ON PATIENTS TREATED WITH ALENDRONATE FOR OSTEOPOROSIS
AND THE INCIDENCE OF FRACTURES

Trial                Outcome        Relative Risk
                     Studied          Reduction

Black et al (7)      New vertebral       53%
Women with existing  fractures
vertebral fractures

Cummings (8)         New fractures       36%
Women with low
bone mass
but no fractures

Trial                Absolute Risk Reduction            Number Needed
                                                          to Treat

Black et al (7)      7 percentage points, from 15%           14
Women with existing  in the placebo group to 8%
vertebral fractures  in the alendronate group

Cummings (8)         6.5 percentage points, from 19.6%       15
Women with low       in the placebo group to 13.1%
bone mass            in the alendronate group
but no fractures

Note

The Blue Cross and Blue Shield Association is an association of independent, locally operated Blue Cross and Blue Shield Plans. The information presented in this column does not necessarily represent the policy or views of either the Association or any of the Plans.

References

(1.) Expert Panel on Detection. Evaluation and Treatment of High Blood Cholesterol In Adults, JAMA 1993;269:3015-23.

(2.) The Scandinavian Simvastatin Survival Study Group. Randomized trial of cholesterol lowering In 4,444 patients with coronary heart disease. The Scandinavian Simvastatin Survival Study (4S), Lancet 1994;344:633-8.

(3.) Sacks, F.M., Pfeffer, M.A., Moye, L.A. et al. The effect of pravastatin on coronary events after myocardial Infarction In patients with average cholesterol levels. N Engl J Med 1996;335:1001-9.

(4.) The long-term intervention with pravastatin in ischaemic disease study group. Prevention of cardiovascular events and death with pravastatin in patients with coronary heart disease and a broad range of initial cholesterol levels, N Engl J Med 1998;339:1349-57.

(5.) Shepherd. J., Cobbe, S.M., Ford, I. et al. Prevention of coronary heart disease with pravastatin in men with hyperchoiesterolemia, N Engl J Med 1995;333:1301-7.

(6.) National Osteoporosis Foundation, Physicians Guide to Prevention and Treatment of Osteoporosis. Washington, DC. National Osteoporosis Foundation, 1998.

(7.) Black, D.M., Cummings. S.R., Karpf. D.B. et al. Randomized trial of effect of alendronate on risk of fracture in women with existing vertebral fractures, Lancet 1996;348:1535-41.

(8.) Cummings. S.R., Black. D.M.. Thompson, D.E. et al. Effect of alendronate on risk of fracture In women with low bone density but without vertebral fractures. JAMA 1998;280:2077-82.

Elizabeth Brown, MD, is National Medical Consultant for the Blue Cross and Blue Shield Association in Chicago. Illinois. She can be reached by calling 312/297-6186 or via email at Elizabeth.Browngsbcbsa.com.