Working Paper: NBER ID: w2278
Authors: Jeffrey E. Harris
Abstract: As of March 31, 1987, the U.S. Centers for Disease Control had reported 33,350 cases of acquired immune deficiency syndrome. Yet by that date, physicians had actually diagnosed 42,670 cases. The difference arises from significant delays in the reporting of AIDS cases to public health authorities. An estimated 70% of cases are reported two or more months after diagnosis; about 23% are reported seven or more months later; and about 5% take more than three years to come in. Moreover, the probability distribution of delays has been shifting to the right, with the median delay increasing by 0.6 months since mid-1986. From the data on reported cases and the estimated probability distribution of reporting delays, I reconstruct the actual incidence of AIDS from January 1982 through March 1987. The doubling time of the epidemic fell from about 6 months in 1982 to 15-16 months in 1986.
Keywords: AIDS; reporting delays; public health
JEL Codes: I18
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| reporting delays (C22) | undercounting of AIDS cases (C80) |
| increased reporting delays (C41) | increased median delay (C41) |
| increased reporting delays (C41) | slower rate of new diagnoses (I12) |
| shifting distribution of reporting delays (D39) | increasing proportion of unreported cases (K42) |
| increasing proportion of unreported cases (K42) | deceleration of the epidemic (F44) |