Pediatric AIDS in New York City: Estimating the Distributions of Infection, Latency, and Reporting Delay and Projecting Future Incidence

Advances in medical statistics arising from the AIDS epidemic

Many statisticians have contributed to studies of the HIV epidemic and progression to AIDS. They have developed new statistical methodology, where needed, to address HIV-related issues. The transfer of methods from one area to another often involves a substantial delay. This paper points to methods that were developed in the HIV context and have either already found applications in other areas of medical research or have the potential for such applications, with the hope that this will promote a speedier transfer of the research methods. Among the new tools that HIV studies have placed firmly into the pool of statistical methods for medical research are the methods of back-calculation, methods for the analysis of retrospective ascertainment data and methods of analysis for the combined data from clinical trials and associated longitudinal studies. Notions that have been stimulated substantially are use of surrogate endpoints in clinical trials and screening blood products by the use of pooled serum samples. Research activity in many other areas has been boosted substantially through contributions motivated by HIV/AIDS studies. Noteworthy examples are analyses for doubly-censored lifetime data and methods for assessing vaccines for transmissible diseases.

Estimating AIDS incidence and jack-knife variance from a continuous delay distribution and incomplete data

Usual methods for estimating AIDS incidences are based on the inflation of a discrete reporting delay distribution, which often results in very imprecise estimates of the incidence in the most recent past. In this paper, we propose an alternative approach to estimate the AIDS incidence by inflating a continuous reporting delay distribution for each reported case. Covariate effects on reporting delays are evaluated by a proportional hazards model for the reverse time hazard function. A jack-knife variance for the estimated AIDS incidence is given. Study results showed that precision of estimates is improved by using the continuous time model as compared with those estimates given by its discrete counterpart. This feature is useful in assessing current trends in AIDS incidence.

Modelling the costs of paediatric HIV infection and AIDS: comparison of infected children born to screened and unscreened mothers

OBJECTIVE

To estimate the cost of managing children with HIV/AIDS in the UK from a health service perspective.

DESIGN AND METHODS

Epidemiological, resource use and unit cost data were combined within a decision analytic model. A Markov model was developed to predict the prognoses of HIV-infected children under different assumptions about natural history, treatment efficacy and the timing of antiretroviral therapy. Resource use estimates for various stages of HIV/AIDS were based on published data relating to the UK health service and clinical judgement; unit cost data were taken from a London centre.

RESULTS

The base-case results suggest that the cost of caring for an HIV-infected child is higher if the mother's infection was known about at or before the child's birth (antenatal screened cohort): lifetime costs ranged from 46 427 pound sterling to 119 502 pound sterling per child in the screened cohort and from 38 691 pound sterling to 86 014 pound sterling in the unscreened cohort. However, the screened cohort benefited from longer life expectancy (base-case, 11.66 versus 10.09 years) and AIDS-free life expectancy (base-case, 7.13 years versus 6.22 years). Results are sensitive to assumptions about natural history and treatment efficacy: for example, if antiretroviral therapy was initiated at birth, and assuming optimistic natural history parameters, discounted costs could increase to 215 077 pound sterling and the additional lifetime cost of a child born to a screened mother could be 72 491 pound sterling.

CONCLUSIONS

Results reflect the marked uncertainty regarding the cost of, and prognosis for, children with HIV/AIDS in the new era of more potent antiretroviral combination therapies. These results are part of an assessment of the relative cost-effectiveness of alternative antenatal HIV testing strategies.

Distribution of the latency period for perinatally acquired AIDS

In the United States, over 86 per cent of paediatric cases of acquired immunodeficiency syndrome (AIDS) were infected through maternal or perinatal transmission and current estimates suggest that approximately 6000 children are born to infected women each year. In industrialized countries, less than 25 per cent of infants maternally exposed to the human immunodeficiency virus (HIV) are infected. The exact time of infection is unknown. However an endpoint, birth, exists from which the latency period or the time to diagnosis with clinical AIDS can be measured. In New York City, 7 years of reliable surveillance data were available for children with perinatally acquired AIDS. Using these data, the only estimable aspect of the latency is a conditional distribution, which conditions on diagnosis occurring before age 7. Because newborns carry maternal antibodies, a positive antibody test at birth is only an indication of maternal infection. Screening all newborns for HIV antibodies, while maintaining the anonymity of the mother and child, provides data on the number of children maternally exposed. The above conditioning can be removed by combining the surveillance and newborn screening data to obtain the unconditional latency distribution. We estimate a median diagnosis age of 4.1 years, somewhat longer than previously believed.