|
We started out in1983. We started interviewing women in the 4th
prenatal month of pregnancy. We took a successive number of people, everyone who came into
the clinic, for a period of about eight to ten months. We interviewed 1,360 women, and of
those 1,360 we decided to sub-select children for the alcohol cohort. We took all the
women who drank alcohol at the rate of three or more drinks per week during the first
trimester and then we took a random sample of 1/3 of the women who drank less than that
amount. We wanted to have a full-dose response curve but we didn't want to have a whole
bunch of abstainers and very, very light drinkers. But we didn't have control over it;
it's just not cost-effective. So we ended up with 650 women who we followed into the 7th
prenatal month. Initially the design was to treat these women into the third month, during
the first trimester, and the six month of the second trimester, but we didn't know that
when we asked them about alcohol use currently, nobody did that. They all denied use -
because it's very threatening. They knew that they shouldn't be drinking during pregnancy,
and that drugs were illegal and they shouldn't be using them during pregnancy either. And
they didn't want to run the risk of social workers getting involved and maybe some other
legal problems. When we asked the women about last month's use, they were willing to
answer, and I think they answered quite honestly. For a period of time we had something
that we called a bogus pipeline test, which is a bluff. It's a way of convincing your
subjects that you actually have a measure of abuse but you don't. The interviewer would
pull a test tube out of their pocket and pull a q-tip out of the tube and say "We're
trying to develop a test that would enable us to measure substance use when people come
into the emergency room because it's very important to know if they've been using
substances. We would really like your help. They'd have the q-tip in their hands, ask the
alcohol and drug questions, put the q-tip back in the test tube. It didn't make a whole
lot of difference with alcohol; marijuana went up about five or ten %. What was even more
interesting was after about six months of this, I discovered that we didn't need the test
tube anymore. It was in part an interview effect, which is one of the most difficult
things when you're doing drug research - convincing the researcher that it's okay to ask
all these questions. If it's hard to ask researchers to ask that, it's impossible to ask
clinicians to ask that. That's almost an insurmountable hurdle. So once I convinced my
research assistant that this was okay, people didn't mind and they actually were going to
answer quite honestly. That got transmitted directly from interviewer to subject and we
didn't have any more problems. Our rates stayed up, at exactly what we assumed they should
be. |