Hospital admissions of HIV-infected patients from 1988 to 1992 in Maryland

To determine how the patterns of inpatient hospital care for HIV-infected patients have evolved in recent years, we analyzed data obtained from a statewide hospital discharge database from Maryland for the years 1988, 1990, and 1992. For each of these years, we compared demography, diagnoses, lengths of stay, use of the intensive care unit, third-party payer, and hospital charges (inflation-adjusted to 1992 dollars). HIV-infected patients accounted for 0.42% of all Maryland's hospital admissions in 1988, 0.68% in 1990, and 1.1% in 1992, with progressively more women and African-Americans hospitalized. Average lengths of stay fell from 11.7 days (1988) to 10.7 days (1990) and 9.5 days (1992) (p < 0.0001). Average charges per admission fell from $11,634 (1988) to $9,938 (1990) and $8,618 (1992) (p < 0.0001). Medicare or Medicaid paid for 50.9% of hospital admissions in 1988, 56.8% in 1990, and 66.8% in 1992 (p < 0.001). In-hospital mortality rates (7.8% in 1988, 7.9% in 1990, and 7.7% in 1992; p = 0.783) were stable, as was severity of illness. P. carinii pneumonia (PCP) was the most common principal diagnosis, but it declined in prevalence from 13.6% in 1988 to 9.1% in 1992 (p < 0.0001). Principal diagnoses of other opportunistic infections remained stable (8.0% in 1988, 9.9% in 1990, 8.6% in 1992; p = 0.90), as did other nonopportunistic infections (32.8% in 1988, 27.2% in 1990, and 30.0% in 1992; p = 0.16). Non- PCP pneumonias increased from 7.6% (1988) to 10.2% (1992) (p < 0.0001). Substance abuse as a principal or secondary diagnosis increased from 30.9% (1988) to 34.3% (1992) (p < 0.001). HIV-infected patients comprised an increasing percentage of statewide hospital admissions from 1988 to 1992, with African-Americans, women, and substance abusers having increasing rates of admission. Although severity of illness and mortality rates remained unchanged, lengths of stay and hospital charges have declined, partly offsetting economic consequences of the higher incidence of hospital use.