MCHB/EPI Miami Conference — December 7 - 9, 2005

Innovative Research on Teen and Unintended Pregnancy Prevention — Transcript

BETH BARNET: But as I'm going to tell you, this is a lot of preliminary data. I'm going to talk to you about our Randomized Trial to Reduce Repeat Pregnancy in Adolescent Mothers. And I want to acknowledge my project staff and collaborators and of course my funders.

So just to give you a bit of background, Healthy people 2010 calls for reductions in rapid repeat birth. And they define a rapid repeat birth as a birth occurring within two years of an index birth, or any repeat birth during adolescence. Now the prevalence in teens of a rapid repeat pregnancy is between 25 and 50 percent, and between 15 and 25 percent have a rapid repeat birth. And while overall adolescent pregnancy rates are declining, there's actually been increases in the proportion who are having rapid repeat pregnancies and births.

So why does this matter? Well, the negative consequences of adolescent pregnancy are really magnified by having a rapid repeat birth. And those are, as you all know, decreased school achievement, more parenting stress, worse parenting behaviors, on and on.

So the goals of this project were to conduct an intervention that was aimed at reducing adolescent repeat pregnancy that was grounded in theory, that had an explicit focus on motivation, that attempted to address some of the malleable approximate risk factors and also that we were really able to study the processes of doing the intervention itself so that we could look at the fidelity with which we implemented this intervention and of course to evaluate it rigorously.

So our study objectives were to evaluate the effectiveness of computer assisted motivational interviewing to reduce repeat pregnancy, and I'm going to describe what the CAMI is, and also to compare the effectiveness of computer assisted motivational interviewing in reducing repeat pregnancy when it's delivered in a context of a large or home visitation program, which we had ongoing, versus as a stand alone intervention.

So what is the CAMI? Well, the CAMI is a computer program that measures contraceptive and sexually transmitted infection, attitudes and behaviors. So the teenager sits at this computer and there's an algorithm built into this computer program that's based on stages of change theory. And it stages the teenager with respect to her readiness and commitment and motivation to remain non-pregnant. And also assesses her risk for having a rapid repeat pregnancy and for getting STDs. And then what happens is there's a printout that's generated from her inputted responses that a trained counselor then performs stage matched motivational interviewing.

So just to give you some sense of what it is that the teen sees, this is one of the screens. So a screen would say something like, "Please click on the button that best describes who you've had sex with in the last six months." "I had sex only with my baby's father." It assesses other partners as well. And then based on what choice she makes, the questions lead down a road to assessing her use of condoms, other contraceptives and various kinds of risky behaviors.

So this is what a printout would look like and with this particular printout the teenager is in pre-contemplation, not interested in using condoms, with the baby's father. On the other hand, she's in maintenance for using oral contraceptives, birth control pills. Meaning that she has said that over the past six months she's been using birth control pills.

Now we also have a risk assessment that says this same teen is at high-risk for repeat pregnancy and at high risk for sexually transmitted infections. Why? Well, it's because throughout her responses she's said, "Well, yeah, I sometimes miss taking my pills." So she's, even though she's using an effective method of birth control, she's not using it effectively. The teen will also see if she's high-risk this baby will cry and a siren will go off just to sort of reinforce that that's what the assessment, her own responses, have derived.

So this is sort of a schematic of how the intervention and study goes. We recruit teenagers in their third trimester of pregnancy. We get informed consent from both the adolescent and her parent or guardian if they're less than 18. They undergo a very extensive baseline measurement that assesses all sorts of things: demographics and other characteristics that we think are likely to influence the outcome. And then they're randomly assigned to one of two intervention arms and a usual care control arm.

The first intervention arm is a home visited plus a CAMI group in which the teen receives biweekly home visits and then the CAMI assessment and counseling every three months until the child is two years old. And then the CAMI only group in which they don't receive this enhanced biweekly home visiting but they do receive the CAMI assessment and counseling every three months. And then we do follow-up assessments at one year and two years post-partum.

So these are the baseline characteristics of our sample. On average the teens are 17 years old but they have a range of 12 to 19. It's predominantly a African-American sample, on Medicaid. At baseline more than 42 percent of these kids have already dropped out of school. This is not necessarily the first pregnancy for these teens and in fact 11 percent have already had a prior birth. Many of them, almost 40 percent, have had a past sexually transmitted infection. But for some disconnect reason only three percent actually perceive themselves at-risk for getting an STD. This is actually a very high-risk population. They've experienced sexual abuse. They have high rates of violence exposure in their households. At baseline, 37 percent of them scored depressed and in looking at the baby's father characteristics they are on average older, significantly older, as much as--and they are adult men sometimes-as old as 39. At baseline three-quarters of the teens say that they are in a relationship with their baby's father but interestingly enough by two years post-partum that declines to about 39 percent. And at baseline almost 10 percent of their baby's fathers are in jail.

So I'm going to just present some very preliminary outcomes because we're still in the intervention phase and in the data collection phase. But we have 114 completed one year follow-ups and 43 completed two year follow-ups. And for this kind of population, which is an incredibly mobile, difficult to reach population, we're actually achieving very high rates of evaluation follow-up.

So repeat pregnancy at one year. Overall this cohort has a repeat pregnancy rate of 26 percent. That means that one year after the delivery of the index child for this study, 26 percent of them have had another pregnancy. And as you can see there is unfortunately no differences by group. When we start looking at the two year outcomes, we're now seeing 50 percent of them, of this group, is having a repeat pregnancy. And we are beginning to see some differences by group. Actually significant differences by group, that in the control group we're seeing 73 percent of them having a repeat pregnancy, in the home visited plus CAMI group, 50 percent, and in the CAMI only, only 27 percent have had a repeat pregnancy.

We also want to look at some of the other factors associated with having a repeat pregnancy. And so let me just go over this slide. This is looking at associations of repeat pregnancy at the two year mark with some baseline looking at associations with various characteristics. And what we see, and I've highlighted both because it's a small sample. I've highlighted both significant differences and also trends towards the significance. Older age at baseline is associated with having a rapid repeat pregnancy. Being dropped out of school at baseline is also associated with a rapid repeat pregnancy. Being older at your first pregnancy. Being in a household with--actually, this is measured at the one-year follow-up, so this persistent living in an environment with persistent violence appears to be associated with having a rapid repeat pregnancy. Persistence of poor mental health, that is, if the teen was depressed or had poor mental health at baseline and continues to have poor mental health at the one-year follow-up, that appears to be associated with a repeat pregnancy. And I've included this social connectedness variable, which is taken from actually the Longitudinal Study of Adolescent Health. It's a variable that measures how well the teen feels connected and cared for by a parent, by either their mother or their father. And while there doesn't seem to be any association with mother connectedness score, there appears to be some protective effect for the teen reporting that she has a close and caring relationship with her father.

Now of course this doesn't tell the whole story. It's important when you do these kinds of community-based interventions to also look at the processes of conducting the intervention itself. And these are tough kinds of interventions to do, particularly because of the kind of population, being adolescent and being very mobile. So that when we actually look at participation rates in the program, more than 70 percent participate in some respect during the first year. But we also have a kind of a strict definition for what we consider being engaged and retained in the program, which we define as receiving more than 75 percent of the expected visits, CAMI sessions, et cetera. And as you can see, participation is actually, I'm not going to say terrible, but in the first year the home visited CAMI group actually gets the intervention sort of as we feel that it should be given. About 57 percent of them do, versus only 38 percent of the CAMI only group. And this falls off by year two. But we are achieving very high evaluation follow-up rates and so we'll be able to sort of tease out what does all this mean. I hope we'll be able to tease that out.

So our planned analysis, we are going to do an intention to treat analysis. We will create multi-variable models looking at group impact on repeat pregnancy. We'll adjust for baseline differences. And we're also going to do sub-set analysis on teens who sort of receive high dose of the intervention so that we can sort of look at, well, who does this kind of intervention work for, et cetera. We'll do some survival analysis and also some path analysis which will help us to identify the contribution of demographics, psychosocial variables, stages of change, and program participation and variation in program participation, to our outcome of repeat pregnancy.

So I just want to also end with just talking about some of the challenges that we have encountered doing this kind of community-based intervention research. We have a lot of what we call no-shows. So the home visitor, the CAMI counselor, makes an appointment, confirms the appointment, goes out to the house and even confirms at 15 minutes. "I'm in my car, I'll be there in 15 minutes." "Oh, yeah, yeah, yeah. I'll be there.' And then they get there and either they've fallen asleep or something better came up. They've gone to the mall, or there's too much drug trafficking for the home visitors to go into the house. Now some of this I would call sort of developmentally appropriate. It is developmentally appropriate for teenagers to be sort of irresponsible. But so that has to factor in to, you know, what we can expect with, you know, reaching and engaging these teens. Also that they're very mobile. The phones go on and off, so reaching them is also a challenge.

And there's also something that I call sort of the irony of success that we really didn't anticipate as an outcome, is that one of our goals is to actually get the teenagers back in school. And we have actually been very successful at getting teens back into school. However, when we get them back into school they're too busy, which is, I guess, a good thing. But they're too busy to really be engaged and meet with the home visitor and the CAMI counselor as frequently as we want them to.

I also want to say, and this sort of echoes I guess the planetary sessions that we heard, that the issue of a medical home, having teens engaged in primary care. We have found that that is really important to our retention. And that's sort of one of the messages that I got from Ruth's presentation is that you have such great follow-up because you have that access also in the medical setting, whereas the teens in our program are not necessarily connected certainly to our health care setting. There it's much more diffuse in the community of Baltimore . And okay, I'm finishing up.

So what are we learning from this? That, you know, these kinds of interventions, and we all want to move all this information that we get at these meetings, the epidemiological, risk, all sorts of information, we want to move it into action, we want to move it into interventions, and how to do that is a real challenge. And I think one of the things that's really important is that we monitor what it is that we're doing. Because many of us who participate in programs think that we're delivering programs a certain way but unless we actually monitor and measure how we're delivering the services, you know, very carefully measuring that, we don't always know how successful we're being in doing what we think that we're doing. I think that rigorous evaluation of these kinds of interventions is really critical. Very hard to do, our CTs are expensive so there's program costs, then there's research costs, but I think that that's how we're going to really help move this kind of research and knowledge forward, if we can try to be as rigorous about evaluating these interventions as possible.

And then there's also issues on, you know, when you use paraprofessionals how do you make sure that they acquire the skills that you want them to and that there's ongoing training and adherence to protocols?

So just to end up here, you know, it's really too soon for us to draw any conclusions. And these kinds of community-based interventions are very challenging, but they're absolutely feasible. They can be done. And I think that the challenges that we find that we're facing are very much related to what I call real-world clinical challenges. These are the kinds of challenges that we face all the time in a clinical setting, trying to get people to change their behaviors, adhere to medication, do the things that we think are going to help make them healthier.

Other things, what I've presented here, but also from other findings, other analyses that we've done with this work, I think that a key issue in all of this may be the issue of poor mental health and that poor mental health is associated with increased repeat pregnancy. And it's possible that that's a mediator so that we really need to address mental health issues while we're talking about contraceptive use and things like that.

And again I've talked a little bit about the science of these kinds of behavioral interventions and I think that we need to develop our methodology of this science, move it forward, so that these interventions, models, can be delivered, that the training is done in a more effective way, participants receive the training better. And so we're hoping to evaluate CAMI within this framework and if we're successful in reducing repeat pregnancy, and even if we're not successful, we may know more about the behavior change process involved in this important issue.