       Document 0961
 DOCN  M94A0961
 TI    Anonymous but linked HIV testing of male prisoners in Connecticut.
 DT    9412
 AU    Mostasharif F; Altice FL; Singh R; Selwyn PA; Yale University AIDS
       Program, New Haven, CT 06510.
 SO    Int Conf AIDS. 1994 Aug 7-12;10(2):241 (abstract no. PC0324). Unique
       Identifier : AIDSLINE ICA10/94371614
 AB    OBJECTIVE: To determine the prevalence of previously not diagnosed (PND)
       HIV+ inmates; the efficiency of voluntary counseling and testing (VC/T)
       in detecting them; and predictors of their status. METHODS: All male
       state prisoners sentenced > 2 yrs underwent routine laboratory (lab) and
       medical screening, and risk assessment upon entry, from 6-12/93. Discard
       sera and intake information were linked for 978/992 (99%) inmates,
       inmate identifiers were removed, and sera anonymously tested for HIV.
       Referral criteria for VC/T in this facility were history of injection
       drug use (IDU), sexually transmitted diseases, transfusions, same sex
       behavior, or tuberculosis exposure. RESULTS: Of 978 inmates, 61 (6.2%)
       were HIV+ including 44 known HIV+s. PNDs were 17/934 (1:55) of all
       inmates not known to be HIV+; 11/363 (1:33) of those who met referral
       criteria; 10/365 (1:37) of those actually referred for VC/T; 8/228
       (1:29) of those counseled, and 7/203 (1:29) of those tested. Risk
       factors for PNDs were: IDU (53% vs 14% of HIV-s, p < .01); high
       self-assessment of risk (12% vs 1%, p < .01); age > 35 (53% vs 19%, p <
       .01); > 7 incarcerations (29% vs 6.5%, p < .01); and gun-shot wounds
       (18% vs 5%, p = .05). A robust multiple regression model revealed that,
       in addition, they were significantly less likely to be Caucasian, or to
       have undergone a drug treatment program. PNDs also had significant (p <
       .05) lab abnormalities: albumin < 4.0 (29% vs 3% of HIV-s); white count
       < 5,000 (41% vs 8%); AST > 31 (59% vs 20%); hematocrit < 40 (29% vs 8%);
       and platelets < 200,000 (53% vs 22%). Lab values alone had greater
       predictive power than all historical and risk behavior variables.
       DISCUSSION AND CONCLUSIONS: The prevalence of PND status was 1.7%. VC/T
       detected 41% (7/17, 95% CI: 18-64%) of the PNDs by testing 22% of the
       inmates. Alternate referral criteria of abnormal white count, albumin,
       or AST, would have included 76% of PNDs and 28% of HIV- inmates in our
       sample (13/272, 1:21). Anonymous but linked serosurveys provide a unique
       opportunity to evaluate the efficiency of VC/T, characterize the PND
       population, and suggest additional criteria for VC/T.
 DE    *AIDS Serodiagnosis  Connecticut  Counseling  Human  HIV Seroprevalence
       Male  *Prisoners  MEETING ABSTRACT

       SOURCE: National Library of Medicine.  NOTICE: This material may be
       protected by Copyright Law (Title 17, U.S.Code).

