Correspondence

Expenditures for the Care of Patients with HIV

N Engl J Med 2001; 344:1948-1949June 21, 2001

Article

To the Editor:

In Figure 1 of their important report on expenditures for the care of patients infected with the human immunodeficiency virus (HIV) (March 15 issue),1 Bozzette et al. show the trends in average monthly spending for the care of HIV-infected American adults. Spending on prescription drugs rises over time, and spending on inpatient care falls; both of these are consistent with the trends in clinical care and in other studies. However, the drop in spending on outpatient care by about one third is surprising, given the need for close monitoring of the use of antiretroviral therapy and evidence of a 20 to 38 percent increase in utilization rates for outpatient care since the introduction of highly active antiretroviral therapy.2,3

An alternative explanation is regression to the mean. The HIV Cost and Services Utilization Study sample was drawn from those receiving regular care during a two-month period in 1996. Patients with a high rate of use of outpatient services during that period were more likely to be selected. Some of these patients no doubt consistently use outpatient care at a high rate. However, use of outpatient services varies over time; if patients were at a point of high use in early 1996 (and thus preferentially sampled), their early utilization rates for outpatient care would not have been representative. Thus, the apparent downward trend in the use of outpatient services may actually represent a statistical rather than a clinical phenomenon. Adjustment for regression to the mean is critical for identifying the true trends in spending on outpatient care.

James G. Kahn, M.D., M.P.H.
University of California, San Francisco, San Francisco, CA 94143-0936

3 References

1. 1

   Bozzette SA, Joyce G, McCaffrey DF, et al. Expenditures for the care of HIV-infected patients in the era of highly active antiretroviral therapy. N Engl J Med 2001;344:817-823
   Full Text | Web of Science | Medline

2. 2

   Torres RA, Barr M. Impact of combination therapy for HIV infection on inpatient census. N Engl J Med 1997;336:1531-1532
   Full Text | Web of Science | Medline

3. 3

   Rahman A, Deyton LR, Goetz MB, Rimland D, Simberkoff MS. Inversion of inpatient/outpatient HIV service utilization: impact of improved therapies, clinician education and case management in the U.S. Department of Veterans Affairs. Presented at the 12th World AIDS Conference, Geneva, 1998. abstract.
   

To the Editor:

Bozzette et al. do not discuss the fact that the economic costs of health care are highly dependent on the health care systems in different countries. Recent European estimates show that the use of combination antiretroviral therapy is associated with reduced health expenditures for patients with advanced HIV infection but with increased costs for the total population infected with HIV.1,2 However, if the figures for health costs of Bozzette et al. were applied to the clinical data of our own HIV cohort in Spain,2 the use of combination antiretroviral therapy would be associated with a dramatic drop of about 40 percent in the total health expenditures related to HIV (as compared with 24 percent unadjusted and 10 percent adjusted in the article by Bozzette et al.), with savings of up to 60 percent in patients with AIDS.

Two main reasons underlie this apparent discrepancy: first, the costs of medical care (as compared with pharmaceuticals) are clearly lower in Europe; second, universal health coverage ensures that antiretroviral therapy is available to patients with severe immunosuppression — who benefit the most from such treatment — regardless of their socioeconomic status. In the United States, higher expenditures for medications are not related to having more advanced disease but to being white, male, and homosexual, and to having higher socioeconomic status. If combination antiretroviral therapy were universally available to all patients with advanced HIV disease, even greater economic savings would have been achieved, in addition to the substantial clinical improvement of patients. We wonder, as Schroeder does (March 15 issue),3 why “every citizen of the most prosperous nation” does not “have basic health insurance.”

Maria Velasco, M.D., Ph.D.
Hospital Clinic, 08019 Barcelona, Spain

Carlos Guijarro, M.D., Ph.D.
Fundación Hospital Alcorcon, 28922 Madrid, Spain

3 References

1. 1

   Sendi PP, Bucher HC, Harr T, et al. Cost effectiveness of highly active antiretroviral therapy in HIV-infected patients. AIDS 1999;13:1115-1122
   CrossRef | Web of Science | Medline

2. 2

   Velasco M, Gómez A, Fernández C, et al. Economic impact of HIV protease inhibitor in the global use of health-care resources. In: HIV medicine. Vol. 1. Oxford: Blackwell Science, 2000:246-51.
   

3. 3

   Schroeder SA. Prospects for expanding health insurance coverage. N Engl J Med 2001;344:847-852
   Full Text | Web of Science | Medline

Author/Editor Response

The authors reply:

To the Editor: The unexpected decline in expenditures for outpatient services that rightly concerns Kahn is an artifact of our decision to combine routine outpatient and emergency department care in our report. Presenting the categories separately would have shown that the net overall decline was the combined result of the expected increase in expenditures for routine outpatient care and a decrease in expenditures for emergency department care. Nonetheless, we agree that our study is potentially vulnerable to bias resulting from regression to the mean for the reasons cited by Kahn.

As we suggested in our paper, split-sample analyses were not consistent with regression to the mean. Specifically, we separately examined the estimated expenditures for patients who were interviewed less than six months after the end of the sampling period and the estimated expenditures for patients who were interviewed later. Estimated base-line expenditures for the patients who were interviewed earlier included costs for the care that led them to be sampled, whereas base-line estimates for the patients who were interviewed later did not. If regression to the mean were operating, the estimated declines in expenditures for the patients who were interviewed earlier should have been greater than those for the patients who were interviewed later. This was not the case, since the patterns of decreases were similar in the two groups. For example, the average declines in expenditures at six months were $291 and $297 per patient per month, respectively. Several other lines of reasoning also argue against regression to the mean as an explanation for our results. For example, declines in hospital use were not uniform and were lower for subgroups with poor access to newer antiretroviral agents than for subgroups with good access.

We agree with Velasco and Guijarro that the implications of our work for net expenditures in a given system will depend on the relative value and the underlying propensity for the use of services and pharmaceuticals. However, for every subgroup in every setting, the economics of treatment for HIV are tightly coupled to the clinical reality that drives it. Because of this, in both developed and underdeveloped countries, one should expect that increasing the resources devoted to delivering technologically advanced antiretroviral chemotherapy will free up resources that would otherwise have gone to acute care.

Samuel A. Bozzette, M.D., Ph.D.
Geoffrey Joyce, Ph.D.
Daniel F. McCaffrey, Ph.D.
RAND Health, Santa Monica, CA 90407-2318

Citing Articles (2)

Citing Articles

1. 1

   Curtis Handford, Anne-Marie Tynan, Julia M Rackal, Richard Glazier, Richard Glazier. 2006. Setting and organization of care for persons living with HIV/AIDS. .
   CrossRef

2. 2

   Herminia Palacio, Xiuhong Li, Tracey E Wilson, Henry Sacks, Mardge H Cohen, Jean Richardson, Mary Young, Alvaro Mu??oz. (2004) Healthcare use by varied highly active antiretroviral therapy (HAART) strata. AIDS 18:4, 621-630
   CrossRef