The Alaska Study: A Clinical Trial of Needle Exchange
Dennis G. Fisher, Ph.D. Center for Behavioral Research and Services California State University, Long Beach
This research was funded in part by grant number R01 DA10181 from the National Institute on Drug Abuse. The Needle Exchange Program was funded in part by grant number 96-3-056 from the Alaska Science and Technology Foundation to the Alaskan AIDS Assistance Association.
Dept.s of Labor, HHS... Approp. Act
“…unless the Surgeon General of the United States determines that such programs are effective in preventing the spread of HIV and do not encourage the use of illegal drugs” (U.S. Congress, 1992)
“The ideal study for the empirical evaluation of a needle exchange program would involve a large number of injection drug users enrolled in a needle exchange program who were compared with a suitable control group and followed for HIV seroincidence” (Vlahov & Brookmeyer, 1994)
Clinical Trial Design
Two-arm clinical trial, NEP vs. Pharmacy Sales conducted May 1997 to June 2000. Eligibility of current injectors only using Cagle et al. (2002) procedure for track mark classification. UA for cocaine, morphine, amphetamine. Risk Behavior Assessment (RBA; DowlingGuyer et al. 1994). Phlebotomy for HBV, HCV, & HIV testing.
Participants were randomized to the Needle Exchange condition (n=296), allowing them access to two partner NEPs, or the Pharmacy Sales (n=304) condition, in which participants were instructed in how to buy NS over-the-counter. Follow-up sessions were done at 6-month intervals.
CONSORT Flow Chart
653 Participants Eligible
53 Excluded 3 Refused 50 Did not return
211 Completed 6-month
215 Completed 6-month
184 completed 12-month
185 Completed 12-month
152 completed both
153 completed both
61 Lost to Follow-up
49 Lost to Follow-up
How to avoid crossover
Those in NEP condition had picture taken (stored on server) ID card with bar code only Bar code readers at NEP sites Computer brings up picture, sex, race, DOB. Computer automatically enters date and time of transaction. NEP operator enters numbers and types of syringes, condoms, lube, etc.
Alaska has no state law prohibiting OTC sale of NS Anchorage has a regulation against the sale of drug paraphernalia
– “It is unlawful for any person to sell, or possess with intent to sell, drug paraphernalia, knowing that it will be used to plant, propagate, cultivate,...inject…or otherwise introduce into the human body a controlled substance…” (Anchorage Municipal Charter Code and Regulations, 8.20.020)
Could IDUs purchase syringes in Anchorage, Alaska?
Pharmacy purchase study (Trubatch et al. 2000) modeled after Compton et al. (1992). Alaska state pharmacist telephone survey modeled after Gleghorn et al. (1998; Harbke et al. 2000). Pharmacy purchase questionnaire study (Fisher et al. 2003).
Return Bias by Assignment?
C u t e r pro Ru i g y si n et K l n e r u l i P oi n e r n b A gm ( a a- i ) m v o t a t n s n p M e 1 . 0 P r ay hm a c E hne x ag c 0 . 8
0 . 6
C ulativeProportionRurning um et
0 . 4
0 . 2
1N59=. 5 p . 0 ( , =9) 0 , = 1 7
0 . 0 0
D s o B e e tr i w a f m sln I e e yr a i n v
PROC MIXED approach to repeated measures to model covariance structure. PROC MIXED uses a ridge-stabilized NewtonRaphson algorithm to optimize a residual likelihood function. PROC MIXED uses all data. We used the Akaike Information Criterion to assist in selecting covariance structure. GLIMMIX macro was used for the dichotomous variable of ever shared needles. GLIMMIX allows use of dichotomous dependent variables with PROC MIXED.
Does NEP Reduce Disease?
Pharmacy / NEP RR = 1.5554 (CI=0.5817, 4.1590).
Does NEP Cause IDU Increase?
Pharmacy Sales (n=153) Needle Exchange (n=151)
Mean Number of Injections
0 Baseline 6 Month 12 Month
Results – Proportion of Injections with used NS
PROC MIXED modeled with heterogenous compound symmetric covariance structure. Significant effect of Time on proportion of shared injections F(2, 213) = 4.84, p=.0088. Marginal main effect of Assignment F(1, 556) = 3.06, p=.0809. Non-significant interaction of Time x Assignment F(2, 213) = 1.10, p=.3355.
Proportion of Injections w used NS
Pharmacy Sales Needle Exchange
0.05 0.00 Baseline (N=554) 6 Month (N=133) 12 Month (N=88)
Results – % Participants Share NS in 30 days prior to interview
GLIMMIX macro modeled with compound symmetric covariance structure. Significant effect of Time on ever share NS F(2, 213) = 3.25, p=.0408. Non-significant effect of Assignment F(1, 556)=0.91, p=.3415. Non-significant Time x Assignment interaction F(2, 213) = 0.88, p=.4174.
% Participants Share NS
Pharmacy Sales Needle Exchange
0.00 Baseline (N=554)
6 Month (N=133)
12 Month (N=88)
Sharing: Does Size Matter?
Participants who reported sharing syringes were significantly more likely to prefer 1 ml syringes. Participants who reported not sharing preferred 0.5 ml syringes. BD 2 (1, N 10245) 424.4, p .0001
(1, N 351) 7.9, p .005
Fisher et al. 2001
Condition by NS Source
Source of Sterile NS None NEP only Pharmacy only Both z=2.89, p<.01. NEP 66.8% 16.4% 7.0% 9.8% Pharmacy 79.3% 20.7%
Needle Exchange does not cause an increase in injection drug use. Needle Exchange probably has a small, but positive, effect on reducing disease incidence. If a community wants to maximize the proportion of sterile syringes used by injection drug users, then both needle exchange and pharmacy sales are required.
Compton, W. M., Cottler, L. B., Decker, S. H., Mager, D., & Strongfellow, R. (1992). Legal needle buying in St. Louis. American Journal of Public Health, 82(4), 595-596. Gleghorn, A. A., Gee, G., & Vlahov, D. (1998). Pharmacists’ attitudes about pharmacy sale of needles/syringes and needle exchange programs in a city without needle/syringe prescription laws. Journal of Acquired Immune Deficiency Syndrome and Human Retrovirology, 18(S1), 89-93.
Vlahov, D., & Brookmeyer, R. S. (1994). Editorial: The evaluation of needle exchange programs. American Journal of Public Health, 84(12). 1990-1890.