XI International Conference on AIDS(internet-link: "http://www.interchg.ubc.ca/aids11/homepage.html") Abstracts, Vancouver, CA, 1 :159 (Mo.C.1608)

A New Objective Method for Computing Risk Indices Applied to Cohorts of HIV-Discordant Heterosexual Couples

DESVARIEUX, Moïse,* WITTKOWSKI, Knut M. DEVINCENZI, Isabelle,*** SUSSER, Ezra(internet-link: "http://cait.cpmc.columbia.edu/research/faculty/facu0330.html")

* HIV Center for Clinical and Behavioral Studies, New York, N.Y, and Columbia University, New York, N.Y.
** Eberhard-Karls-University, Tübingen, Germany,
*** European Centre for the Epidemiological Monitoring of AIDS, St. Maurice, France

Keywords: HIV, AIDS, discordant couples; behavior, measurement; sex behavior; heterosexual; transmission efficacy

Objective: Evaluation of HIV intervention programs requires objective criteria for assessing changes in risk behavior. Previous studies have concentrated on a single variable only, have relied heavily on postulates about the relative importance of risk factors, or have assumed that the risk of transmission is linear over time or number of contacts. We propose a new method for evaluating the risk of HIV transmission associated with specific activities in serodiscordant couples.

Methods: The proposed method is based on the MrgL-procedure[1] and entails 4 steps.

  1. Individuals are partially ordered according to indisputable a priori knowledge (e.g. 'more of the same' is 'worse'). If person X is worse in all variables than person Y, it is assumed to be at higher risk.
  2. All permutations of rankings among the individuals in accord with these partial orderings are selected.
  3. These rankings are weighted by their correlation with a ranking based on the a posteriori observations, if available.
  4. The MrgL score is computed as the average of the (weighted) selected rankings.

Lattice of Observations
A-posteriori
Set of A-priori Rankings
Scores
Ranking
Unw...
Wghtd
(1|4|w||+) 4 1 2 3 2 1 4 4 3 0.000 -0.323
(4|3|w||-) 2 4 4 4 4 3 3 3 2 -0.833 -0.309
(1|6|m||-)
|
 2 2 1 1 3 4 1 2 4 0.000 -0.108
(2|0|m||-) 2 3 2 1 1 2 1 1 1 -0.524
Weights:
.1 .4 .1 .6 .4 .6 .1 .9 .4 .9 .6 .9
---
^

Results: We have applied the method to data from different cohorts of discordant heterosexual couples, including the 500 couples of the European Study Group on Heterosexual Transmission of HIV.[2] By means of the unweighted scores alone, we were able to adjust for correlation among the variables. In addition, the weighted scores allow to break ties by utilizing the additional information of the observed time until seroconversion.

Conclusions: The new procedure is objective, based on minimal assumptions and easily applicable. We propose it as a standard for reporting of changes in risk behavior. Such a standard is essential for cross-study comparison of risk.

  1. WITTKOWSKI, K.M. (1982) An extension to Wittkowski. Journal of the American Statistical Association 87
  2. DEVINCENZI, I. (1994) A longitudinal study of Human Immunodeficiency Virus transmission by heterosexual partners. The New England Journal of Medicine 331

MOÏSE DESVARIEUX
Columbia University, Div. Epidemiology, 600 W 168 St PH 18, New York, N.Y.
Fax: 1(212)305-9413; email: md108@columbia.edu(md108@columbia.edu)

Related presentations at the XI International Conference on AIDS(internet-link: "http://www.interchg.ubc.ca/aids11/homepage.html")

Other publications on HIV and AIDS

Additional publications and presentation of Knut M. Wittkowski



diese Seite enthΣlt

Internet-Links



XI International Conference on AIDS(http://www.interchg.ubc.ca/aids11/homepage.html)

SUSSER, Ezra(http://cait.cpmc.columbia.edu/research/faculty/facu0330.html)

XI International Conference on AIDS(http://www.interchg.ubc.ca/aids11/homepage.html)