Page 47 - Special Topic Session (STS) - Volume 1
P. 47
STS353 H. Zhao et al.
Now we apply the proposed estimation procedure to a set of clustered
interval-censored failure time data from a lymphatic filariasis (LF) study
discussed by Williamson et al. (2008) and Zhang and Sun (2013) among
others. The LF is a debilitating parasitic disease in which several worms live
together in several nests and an effective treatment is expected to kill the
worms in all of the nests. The study consists of two treatments, the co-
administration of diethylcarbamazine (DEC) and albendazole (ALB) (new
treatment) versus DEC alone (standard treatment), and the main goal of the
study is to compare their effects on the treatment of LF. The drug ALB is an
anti-parasitic one, manufactured by GlaxoSmithKline, and it is commonly
used to treat interstima worm infections. When coadministered with DEC, it
helps break to the cycle of LF transmission between mosquitoes and humans,
and by using ultrasound, the doctor can detect the movement of the living
adult worms.
The study followed 47 patients, 22 given the new treatment and the other
25 given the standard treatment, for a year since their treatments and they
were periodically examined by ultrasound to see if the worms were still alive.
Thus with respect to the times to the clearance of the worms in each nest,
the variables of interest, only clustered intervalcensored data were observed
with each patient serving as a cluster and the cluster size being the number
of nests of adult filial worms in the patient. In addition, these times to the
clearance may be correlated to the number of nests as pointed out by
Williamson et al. (2008) and can be seen from Table 1, reproduced from
Williamson et al. (2008), which gives the average percentages of the nests
cleared during the one year follow-up for all study individuals. It is apparent
that the time to clear the worms seems to be positively correlated with the
number of nests or the cluster size. In total, 78 adult worm nests were
detected by ultrasound with the cluster size ranging from 1 to 5.
Table 1. Percentages of nests cleared in the LF study during the follow-up
Number of Nests Percentage Cleared
1 81.8
2 62.5
3 50.0
4 or 5 33.3
Table 2. Estimated treatment effects for the LF Data
Model # of within-cluster Estimate SE p-value
resamples
Cox model B=40 -0.5052 0.3148 0.1085
B=80 -0.4940 0.3145 0.1163
B=160 -0.4997 0.3175 0.1154
Probit model B=40 -0.6279 0.4770 0.1805
B=80 -0.6319 0.4836 0.1913
B=160 -0.6202 0.4573 0.1750
36 | I S I W S C 2 0 1 9