CLAIRE DUNHAM: Good afternoon. Could I ask everybody to take your seats? This Mic sounds really loud to me. I want to welcome you to this afternoon's plenary session. I'm really excited about the opportunity to moderate this discussion. Can you hear me? Wow, it's really reverbing back to me. Better? Okay, thank you. My name is Claire Dunham, the director of Parents Too Soon at the ounce of Prevention Fund which is a statewide program with many sites that uses primarily home visiting and parent groups to try to influence parent and child outcomes. Some of you are familiar with this program. We use primarily a para-professional audience although we have added programs with nurses and use more of a professional staff as well. I'm very excited about today's presentation. There are some logistics that I need to share with you about this session. Some of you may know that this plenary is being webcast live. So we may get some questions coming in from on the line, listeners or watchers. And if you have a question for the presenters, which we will hold all the questions until the end of the session, if you have a question I would like to ask you to go to the microphone so that people who are in the room and the people who are listening on the webcast can hear the question. And any other questions that come in through the online system, I'll be reading the questions out so that the presenters can respond. Hopefully I've shared those logistics correctly. Again, I wanted to just say how excited I am about the opportunity to hear from two such serious scientists and researchers who are really out there on the cutting edge of looking at the efficacy and the impact of home visiting programs. We have two presenters for this afternoon's plenary. The first is Dr. David Olds. Dr. Olds will be summarizing the results of three randomized controls. Of a program of prenatal and infancy visiting by nurses. I'll let him describe that for you. Let me tell you a little bit about Olds. He's professor of Pediatrics and Preventive Medicine at the Colorado Health Sciences Center. He directs the Prevention Research Center for Family and Child Health. Dr. Olds has devoted his career to investigating methods of preventing health and developmental problems in children and parents from low income families. He is well-known for his rigorous research on the impact of home visiting programs.
Our second presenter this afternoon is Dr. John Landsverk. Dr. Landsverk has a Doctorate in Sociology and currently a professor in the School of Social Work at San Diego State University and Director of a Child and Adolescent Research Center at Children's Hospital in San Diego. In the research area of home visitation strategies of prevention of child abuse, Dr. Landsverk has been principal investigator on several studies in the civilian and military context. He conducted studies of home visitation programs in three branches of the military service. We are lucky to have both of these presenters with us today. I would like to ask you to hold your questions until the end until both presenters have had a chance to speak. You can get a sense of the overall landscape. If you want to jot your questions down, we'll have the opportunity for Q & A at the end. I hope you'll help me welcome Dr. Olds and Dr. Landsverk. (Applause.)
DR. DAVID OLDS: Thanks, Claire. It is really a pleasure and honor to be with you here this afternoon. Give me a second here. (Laughter). They showed me how to work it, but you have to carry it with you. (Chuckles.)
In 1970, I finished up undergraduate school? Baltimore and went to work in the union square day care center on West Lombard Street. In lots of ways I was a product of the 1960s. I -- Head Start was just starting, of course. I had this abiding belief if I could just help poor four year olds get off to a good start in a preschool program that they would succeed? School and have better chances of life success. I soon realized that for many of the children in my classroom, too many, at least, it was too little and too late. Jamie was a four year old with thick glasses, and frightened. He couldn't sleep at nap time. One day another teacher and I took him down to a classroom to find out why. We asked him why he couldn't sleep. He said you know; when I sleep I often wet myself. When I wet myself and go home, my mom beats me. So for Jamie it was simply safer not to sleep at all. Another little boy in our classroom was communicating by gestures and grunting. He had no language at all. He was cared for by a heroic grandmother. His own mother had been heavily involved in the drug culture. We strongly suspected that this boy's profound developmental disabilities were the result of his exposure to alcohol or drugs during pregnancy. So we knew then that much -- we needed to do more in order to make a difference in the lives of the children. We were concerned about. All right? All right. So this is not ... all right. Did I do something here? I'll try once more (shouting from the audience.) At the screen? Oh! (Laughter). You didn't tell me that part. (Laughter).
I realized after working with poor children and families in inner city Baltimore if we were going to make a difference in the lives of the children and families, we needed a program that was powerful. You couldn't do it with a light touch. What did that mean for us? First of all, the program needed to make sense to parents. It needed to resonate with their deepest cultural beliefs about what is important. Not just cultural beliefs, but personal beliefs. The program needed to have solid clinical and research foundations. The program that we ended up developing was, is now referred to as the nurse family partnership. It is a program that begins working with families during pregnancy and follows them through the child's second year of life. The other thing that struck me at this period in the 1970s was that if we were going to make a difference in the lives of poor children and families, we also needed to evaluate the program really rigorously.
The program that we developed needed to be able to stand up under the most intense scientific scrutiny. It is not enough to have a good idea. I didn't want to be in a position of just promoting a program simply because we happened to think it up. We needed to know, I needed to know personally that what we were promoting was really going to work. For that reason, we've tested the program, as you'll see in a moment, in a series of scientifically controlled studies, randomized controlled trials. The program we developed focuses on low income mothers having first babies. By virtue of focusing on that segment of the population, we ended up serving a lot of young mothers who are unmarried, who are teens. The focus on the population of women having first babies is really crucial. First, women having first babies constitute 40 percent of all births in the United States. 40 percent.
Focusing on that segment of the population is crucial also because we need to find a way of leveraging every ounce of influence that we can exert in bringing about adaptive behavioral change. One way of doing that is to capitalize on the natural sense of vulnerability that low income and all women feel when they are going through pregnancy, especially the first time. Women are frankly worried about what labor and delivery are going to be about, what caring for the vulnerable newborn is going to be like. They need a program in my view where they believe that what is being offered to them will reduce the sense of vulnerability. Reduce the risks that they may have for having complications of labor and delivery or compromised child health and development. To the extent that we are able to get good first time families off to good care and managing the care of their first child well, the public health impact of the interventions will be greater because it will carry over to subsequent children, we think. We'll have more about that in just a moment. Those are the three major goals.
The first is to help parents improve the outcomes of pregnancy by helping women improve prenatal health, cutting down on substances, identifying obstetric complications and having those conditions treated through the primary health care system. And the second major goal is to help the child's subsequent health and development by providing helping the parents provide more competent care of the baby in the first two years of life and to help the parents become more economically self sufficient by helping them find a vision for the future and find out how pregnancy can affect their ability to stay in school or find work. Through this the nurses involve fathers and grandmothers and other people in the home environment who can reinforce and support the nurse's efforts to bring about adaptive change in these aspects of woman’s lives. Also to respect those other family members because we need to understand their cultural beliefs and connect with them as a way of bringing about the kind of change that not only this program wants but also is consistent with the values of family and the culture. And the nurse systematically link up families with health and human services in the community.
How does this happen? It's important that you understand how the models reinforce one another to the extent -- hmm. To the extent that the nurses are successful in helping women cut down on the use of substances during pregnancy, such as toe tobacco, babies will be born more biologically intact. We know that exposure to tobacco during pregnancy makes babies more irritable and fussy during the new porn period. As school age children, there are seven longitudinal studies conducted around the world that show that the exposure to tobacco during pregnancy increases the risk of severe antisocial behavior by a factor of two. As a matter of fact some of that work is conducted here at the University of Chicago that revealed that literature. If the nurses are successful in cutting down on baby's exposure during pregnancy, these babies will be easier for parents to care for. That makes, that means that the nurses' efforts to improve care giving are going to be met with the newborn who is going to respond in ways that are predictable. Babies are going to be more consolable. Parents will find they are getting natural reinforcement in the care of their babies so that the attachment relationship will be strengthened to the extent that there has been improvement in prenatal health as well. Moreover, to the extent that the nurses have been successful in improving b and helping families plan subsequent pregnancies and increase the interval between the birth of first and second children, the parents will have more time and energy to focus on the care of the first child.
For low income families, that's crucial. It's easier for parents to stay in school and find work. To the extent that families are improving their economic conditions, we believe -- we don't have evidence to fully support this yet, but we have reason to believe that nurse visited families will move into safer neighborhoods. Even when they are less likely to be exposed to negative peer influences. Even if they are exposed to negative peer influences, as adolescents, we think their susceptibility to peer influences will be reduced because the children will have stronger attachments and bonds with their parents. Okay? We've tested the program in three separate randomized control trials over the last 27 years. First in Elmira, New York, with a sample of low income weeks. The we, the first trial, the folks said you have a program that works. You need to make it more widely available. We took the position not to do that. We needed to know whether the program could work with minorities in urban areas before we offered it up for public investment.
We replicated the study in Memphis, Tennessee with a sample of receive hundred low income African Americans following up with 50 or postnataly. In the meantime people were taking the results of the Elmira trial and taking them to all home visiting programs. The position was that home visiting works with issues of child and adolescent neglect. We were concerned about that. I didn't object to people using our evidence to support other types of home visiting programs at first because it wasn't clear to me whether the results of our trial were to apply only to our particular program or whether they could be applied to a variety of home visiting program types. Well, we viewed the literature in the early 90s on the impact of home visiting programs for -- we using other visiting types.
The results were disappointing. When you look carefully at the random controls of other program types, the results were not producing the impacts everyone would like to see, everyone in this room is committed to seeing. What was not clear to us is why the programs were not working as well as we would like. Was it because the programs themselves were not well developed? Or was it because there was something about para-professional visits themselves that would bring about adapted behavior change no matter what kinds of model they were working with. In the Denver trial we compared nurse home visit ores. When both were used the showed to be effective when delivered by home nurses in the previous trials. We equal lies, we create consistency in the home Mohammed and we retested the program and examined the relative impact of the program developed, when delivered by nurses versus para-professionals and in the Denver trial we also had a large sample of Hispanics. Now, from this program of research we've taken the position that the findings that we believed that we had the greatest confidence in are those replicated across more than one trial. So we know from at least two of the three trials that we can produce significant improvements in woman’s prenatal health, especially cutting down on the use of tobacco and reduce hypertensive disorders of pregnancy.
We have seen reductions in children's medical records in two of the three trials. We have seen reduction of the rates of subsequent pregnancy and increases of the interval between the birth of the first and second child. We have seen significant increases in father involvement including marriage in two of the three trials. We have seen significant improvements in woman’s employment in two of the three trials. And reductions in welfare and food stamp use. Now, the reduction in welfare and food stamp use we think probably doesn't apply as much in today's post welfare reform environment as it did in the 70s, '80s, and early '90s. We don't see the same level of impact on welfare use and food stamps in the third trial which has been conducted in the post welfare reform environment. But we have seen significant improvements in children's school readiness when the children are preschoolers.
Improvements in IQs, improvements in the language development, improvements in the executive functioning that give us confidence that, confidence may be too strong a word. Make us optimistic that the children will function better academically in school because of their better cognitive functioning. Now, what we also see in our el Mira study is increased reports of child abuse and neglect. We focused on women who had all three of the characteristics we used for recruitment. Poor, unmarried and the mothers were teens. We saw reduction in the emergency room visits in the first year of life and that's when the children are most vulnerable going to the emergency departments because of injuries.
The significant reduction we see in the rates of child abuse and neglect are really concentrated among women who had little belief in the control over their life circumstances measured at registration. So we see further belief -- because of the pointer, it's hard, if you hadn't noticed, it wasn't working. What this slide is essentially saying is that the treatment control difference, the nurse control difference in the rates of state verified reports of child abuse and neglect in the first two years of life were greater among women who had little belief in control over life circumstances. They did not believe they could manage life's challenges. The same pattern of results, 60 percent treatment difference results in the child's second year of life was concentrated among women who had little belief in the control over their life circumstances.
We have our first indication that this is a program that has its greatest impact on families who are most vulnerable. And if we have limited community resources, we believe that you need to focus on the families who need it the most. We also saw in a 15 year follow-up of the original study that the impact of the program on the rates of child abuse and neglect endured over the 15 year period when we concentrated the analysis on women who were b low income and unmarried at registration. There was 80 percent control over that 15 year period and also for the mother there was a 44 percent treatment control difference in their own behavioral problems due to the use of alcohol and drugs p and 69 percent reduction of the mother's own are by the time the first child was 15. How is it that a program that begins during pregnancy might have these kind of outcomes, especially when you look at women 's own involvement with substances and are? The answer is that women are making critical decisions in their own lives that will alter their life course trajectories.
One young mother told her nurse: You know, I'm going to break off my relationship with Tony. And the nurse said how come? And she said well, you don't know this, but Tony is involved with the drug trade between Rochester, New York, and New York City. I've decided it's not good for me or my baby to have Tony around anymore. So the nurse said well, that sounds like a smart thing to do. Soon after that this mother told Tony that she was going to break off the relationship. And tope's response was, well, that's fine. Go ahead and break off the relationship. Now I'm going to have to kill you. This mother had to go into hiding, domestic violence shelters; there were none at the time, to avoid being murdered. But she, the nurse was able to connect with her about the things that were most important to her. And that was the protection of her baby and protecting herself and help her make a decision that would have long, life altering consequences. We know that the major source of influence on women 's involvement with criminal activity and substance abuse are men in their lives involved in those activities. While we want fathers involved, we also want to help women make the right kinds of decisions for themselves and their babies. It's in that connection with women and their families around protecting themselves and their children that can have long lasting effects on outcomes like that.
For the children by the time they were 15, when compared to counterparts in the control group, there was a 50 percent treatment control difference in the number of arrests by the time the children were 15. And 69 percent treatment control difference in the number of convictions or proceed bags violations by probation violations by the time the children were 15. Significant reduction in number of partners, and reduction in tobacco and alcohol use. That is particularly important because children who begin using alcohol by the time they are 4 time more likely to become alcoholics than those who refrain until after they are 21.
The Rand Corporation conducted an evaluation that showed the impact of the program on government spending leads to a four-dollar -- the savings of four dollars for every dollar invested. These are conservative, careful economic evaluations conducted in the context of randomized control trial which is very unusual. On the other hand, these findings do not apply in today's environment because over half of the savings that you see in this slide came in the form of reduced welfare expenditures. In today's post welfare reform environment, we aren't going to see these kinds of savings. Okay? Memphis. Because so many of the beneficial effects of the program were concentrated on women who were low income and unmarried, in Memphis we recruited women who were all unmarried and all very low income. What we see, we knew we would not see program effects on the State verified reports of abuse and neglect because the rates in that community were so low, there was no room for reduction. Only three to 4 percent in children in that population, age two, birth to two. So we hypothesized we would see effects on injuries in the children's medical records that would reflect the prevention of child abuse and neglect. In fact we saw there were 23 percent treatment control of all types of health accounts for injuries and 8 percent in the number of days that children are hospitalized with injuries.
This slide shows the three children in the nurse visited condition who were hospitalized with injuries in the first two years of life. Notice that these children were all 12 months of age or older. And two of these three children were hospitalized with ingestions. Now, the denominator here is only 204 because we -- I won't go into the details. Disproportionately assigned women to the control group and followed only 750 post 98ally. Look at the corresponding rates for the control group. 58 percent of these children were hospitalized -- 44 percent of these children were hospitalized prior to six months of age. This, these are the control group children. These children are not mobile. They were not creating risk because of their increased mobility.
Something else was going on. 58 percent of these children are hospitalized with severe trauma. Broken limbs, subdural hematoma as, fractured skulls, serious injuries that are reflective of seriously deregulated care. While we were unable to corroborate the impact of the program on child abuse and neglect, we saw what was going on when we opened up the medical records. Impacts on health care encounters of all types were concentrated on mothers even in this homogeneously socioeconomicly homogeneous group of mothers, unmarried. The impact is further concentrated on mothers who have little belief in the few psychological resources. We extended the consent in the limited belief in the control over circumstances -- higher levels, also that psychological resource is on the X axis. What this is saying, the treatment control difference again is concentrated in the more vulnerable children, more vulnerable more or less. Same thing applies when we look at the number of days children were hospitalized with injuries, concentrated again on the most vulnerable segment of this at-risk population. When we followed the children in Memphis to two and a half years after the program ended we see that there are corresponding improvements in enduring effects on maternal life course. We saw closely, reduction in closely spaced following pregnancy, admission for subsequent children, increases in welfare use, increases in father involvement and increases in marriage. We were also seeing, this is -- these results are in a paper that is in press and pediatrics. Children in Memphis also have correspondingly higher IQs, better language development and fewer mental health problems when you look at the total problem scale by the time the children are -- by the time the children are six.
The Denver study included a large portion of Hispanics. And again we randomly assigned families to a nurse or para-professional. We wanted the para-professional program to work. We gave the para-professionals twice the level of supervision. We paid them well. We gave them a lot of detailed guidance in working with the families that they were working with. Nevertheless what we saw was that the nurses in the Denver study were producing effects that corresponded to what we had seen in our previous two trials with the para-professionals producing relatively small effects. So just -- I'm not reviewing all the findings, but hitting the highlights. This is particularly important. This is changes in -- over the course of pregnancy. The most visited women are reducing the number of coat nene, the number of cigarettes. That's a nicotine metabolite. This is a slide showing the timing of the first -- survival analysis of the timing until the first subsequent pregnancy. The most control difference is statistically significant. The para-professional control different is not statistically significant. I argue that that difference is clinically significant. We look at the children's -- (music playing) I apologize. This is the children's language development for, at 21 months. We see that the nurses are producing effects that are clinically significant in our -- the para-professionals are not. And these findings are concentrated on children born to low resource mothers like we saw in Elmira and membership sis study.
Beneficial effects on mothers who initially were more vulnerable. We look at the children's executive functioning, concentrated sustained attention, plan or actions. Again an effect that is concentrated, you know, in the low resource, subgroup. But is significant for the most group, marginally significant for the nurse visited group. Statistically significant for the para-professional visiting group.
In 1997 we were invited by the Justice Department to set up the program in high crime neighborhoods. We resisted those invitations for a long time. We felt we didn't -- for two reasons. We didn't know that the evidence was going to go here and really support public investment of a program. We wanted to make sure we had replicated findings. Secondly, I didn't feel we had developed the program well enough, articulated it well enough to be able to reproduce the clinical elements of the program reliably. We were concerned in the process of setting up the program in new communities, we would be compromised or watered down as we scaled up. We spent a lot of time making sure that any new community who develops a program is going to be well prepared both in the community and the organization to develop it well. We provide a lot of training, technical assistance to those sites. And we've developed detailed visit by visit guidelines to guide the nurses with their families. We developed an electronic clinical information system that allows us to -- it oh he web based, that allows us to monitor the program on a visit by visit basis for every one of the families in the program around the country. Stripping out identifiers, so we can continuously assess the program and use the data from that clinical information system to work with sites, to help them continuously -- so we can learn from them and also continuously improve the program. These are the typical sources of funding for the program.
The communities have to raise the money for the program in developing it in their local community or states. What I -- I'm not sure that this is going to work. I think that, I had a little video clip. It doesn't look like it's going to work. So thank you very much. (Applause.)
JOHN LANDSVERK: While Chris is loading the slides for the next show, the work that you've just seen has encompassed the last 27 years. Obviously it is a careful program of research dating from the 1970s. In the 1980s, there was another movement beginning which was concentrated on home visitation and but with the use of para-professionals as opposed to nurse home visitors. It began from a statewide perspective in Hawaii. And migrated to the United States and is now known as Healthy Families America with its headquarters located here in Chicago. In the mid-'90s there had been very little rigorous research testing of that para-professional model and my colleagues and I in San Diego were fortunate enough to compete for requests for proposals from the State Department of child welfare in their prevention branch, to replicate the Hawaii healthy families program, but to also enhance it and to test it in the California environment. At the same time the request for proposals -- and this is very important for the overall set of studies that I will talk about. The request for proposals specified that the trial was not only to be randomized, but it was to replicate a randomized trial that had already gotten off the ground in Hawaii, led by Ann Dugan from Johns Hopkins. We replicated the measurement techniques, the timing for the measurement, replicating the screening and assessment taking place in hospitals as per what was being done in Hawaii. And I'm going to talk about that clinical trial and its results in San Diego. And then refer back to the implications of that and reflexes on those results, looking at what we now know very clearly from the Hawaii experience with the randomized clinical trial. And I also think can be branch.
And I'll speak something about Dr. Olds' work testing para-professionals in Denver. Just want to acknowledge the funding. But what is important here is the middle statement. This was to test the Hawaii model which was becoming really the model in the United States for what has now become a social movement called Healthy Families America, but within the State of California. So direct replication. Screening and assessment procedures were used to identify the one in four families who were at high risk for child mall treatment. Same assessments were used. They were not as in the Olds model done in the prenatal period. They were picked up at birth and used not as demographic markers, but as list of markers for high risk of child mall treatment that had been developed in Denver sometime ago. Both used a randomized design with yearly outcomes measured to the third birthday. Data collection was independent of the program with the same measures, with the kind of blinding that is possible under that circumstance.
These were the study objectives and they are really not much different from the study objective that Dr. Olds' work has had over the years to decrease the risk of child mall treatment to increase functioning of the parents, increased developmental capacity of the children. Intervention was weekly home visits for up to 36 months. Actually, the Hawaii model, if any of you will remember, stated optimistically that those visits should come right up to the school House door at five years. We never, of course, have seen it implemented at that level. But here the clinical trial was up to 36 months. Limited case loads, no more than 25 per home visitor. There were enhancements of the Hawaii program. San Diego had the benefit of learning from that experience. So team case management principles were put in place to a much greater degree than was evident in Hawaii. There was a child development specialist role that was created, thinking that would enhance in particular the ability of para-professional visitors to deal with developmental issues as they came up. And there were also structured groups, support groups that these families could participate in with all the accouterments, childcare, transportation, et cetera.
Pre-service and wrap around training and the usual kinds of focuses. This is a structure. You see the team leader which was an LCSW social worker, child development specialist. Five home visitors, which is an extremely low supervision load. Each one of them then having no more than 25 cases to carry. Study performance was very successful in both San Diego and Hawaii. There had been no significant differences between the groups detected at baseline or follow-up years. So there was excellent retention as well as no differential attrition. Substantial cohort retention in both sides. In San Diego there were 247. Intervention group and 241 families that had been randomized into the control condition. In both Hawaii and San Diego, the retention was roughly at 85 percent retained for measurement. Not for program, but for measurements over the three years. I think this accounts for the kind of resources we had to follow these families. And it would indicate a much higher confidence than in the outcomes. Total home visits within San Diego were in year one 3800 on down to year four. Most important, the bottom you slides, across 36 months there was a mean number of home visits of 43, meeds yup of 40. The usual kind of range, zero to 144. What is important here is that in San Diego, if you practice rate these for the two and a half years that you see in Dr. Olds' studies, this is the level of dosage that is a very high dosage. I think it's typical of what Dr. Olds has found. This was not true in Hawaii where by the end of the first year half of the families were no longer in contact or receiving any services by the home visitors.
This peculiar slide here shows two lines: The blue upper line which is number of service days. In other words, within a month, was there any contact? Was there any services being provided to these families across 36 months or three years? The red line is the more critical, is were there any home visits on a month-by-month basis? What you can see is that there was actually very good retention of a high proportion of these families. Contrary to what was observed in Hawaii which was a clinical trial within an ongoing service system, in San Diego where we had only this trial to do and we didn't have to worry about more families coming in, what you see is on into the third year, way deep into the third year, you see that home visits were being provided to families in the 45 to 50 percent range. In other words, almost half of the families were still receiving an in-home home visitation by the third year. Again, this supports that notion that the dosage was not just concentrated at the front end but really across.
These, by the way are intent to treat statistics. There were a number of families that by this time had migrated out-of-state and obviously could not receive or had migrated to Baja, California. Of course, they would not be able to receive home visitations. We always maintain intent to treat analysis. I am now going to show you a series of simple bar graphs which are, which show some separation between those families that that received the home visitation in blue on the right, and those families that were in the control group which are signified by red. Repeat pregnancy, you see slight differentiation at 24 months and 36 months. Between the two conditions, but it was not statistically significant. When we look at, however, -- intriguing kind of finding that we don't fully understand in any way is that almost all of that differentiation between crop and intervention group was concentrated in the Angelo population. We do not understand that. We do not have the money to randomize within strata. It's really an exploratory finding. Let me again say to you that repeat pregnancies, which has been a hallmark of outcomes in the nurse home visitation was not observed here in the San Diego trial. If we look at ends live birth, again you see the small differentiation at six months, 22 percent of the intervention group had had a live birth during that 36 months subsequent births versus 27 percent, 5 percent give. Is not statistically significant. Maternal depressive symptoms looked promising at year two.
There is a statistically significant difference in depressive symptoms for the mothers. This is somewhat related, I think, to maternal resources that are involved. So you can see at year two some widening. Unfortunately, it had disappeared by year three. So there was a, across the three years no been fish I am beneficial impact on this fact that we thought should have widened by the third year. You do see that there was the good news, there was reduction across the three years. The tough news is, there was really little differentiation between the two conditions. Any neglect just really wasn't much difference whatsoever. We used the conflict tactic scale, version 2 that Murray Straus has developed, a standardized measure, probably the best we have at the present time from a survey point of view.
Use of corporal punishment. You had a 78 to 73 at year three. It was not significantly significant. There was a trend there. What we did find is also true on physical assault. Again, though, note that on each one of the slides, the blue bar is slightly lower than the re bar. So there is a consistent finding on the major markers and this is where we finds significant difference renovation -- I have a slide I'll show you. We didn't always find it. It is not statistically significant. It is difficult to treat it with much confidence in the intervention, but it is in the direction one would want it to be in. Any psychological aggression at year two we had a probability of .04. We suggested that we were starting to see some widening. Just like maternal depressive symptoms, by year three on the right-hand side the bars had closed. You have really a wash-out of what looked like it might be an effect. We did find at years two and three that among those women who reported using psychological aggression, talking very undercutting to their children, abusive to their children, among those who reported using that technique to control their children, there was a statistically significant difference in the direction of the blue bar in terms of, between the two conditions. So that the controls were showing higher frequency among users.
Frequency, corporal punishment at year three, we see that among those who used corporal punishment by year three, there was some differentiation. It was statistically significant. So there are these little snippets, if you will, of differentiation always in the direction of benefit from the intervention group, but none of them statistically significant that you could attribute therefore to the work of the para-professional program. At years two -- one and two, we used daily. We did observe up to about two points differentiation on the bail in favor of the intervention group. Unfortunately when we used the Stanford Binet and one other adaptive functioning measure at the year three, there was absolutely no difference between the groups in terms of developmental level of the children.
Number of well child visits. This was a s consistent finding in favor of the home visitation program. There were significantly more well child visits that were being observed in the intervention group a compared to the control group. This was, by the way, a trial and a program that was based in a children's hospital. Now, most all those -- let's go to all. There's a whole list of outcomes with no significant difference.
Physical aggression, neglect. Mother-child interaction as observed on the end task, as observed by the home standardized measurement. Subsequent mental health of the perment. Intimate partner violence, maternal life course, use of social mental health and health services. This was a particularly disappointing finding. San Diego is a fairly robust social and health network area. We found that those in the control group without the benefit of home visitor were finding their way to services, add juvenile services at roughly the same way as those visited by para-professionals in the home. We found know differences in child adaptive functioning at years two and three, Stanford-Binet an the use of safety measures in the home which one would think would be a benefit that one should be able to see an effect of. So overall, where there were some differences in the very few measures that we saw, there were very small magnitude. They were consistently in the direction of better outcomes for the intervention group. But almost always they were nonsignificant. Now, we've come some time from the mid '90s and on. And what I want to do is reflect in the last two slides on three studies. And what I think we now know, something more than we did before those studies about para-professionals. Three randomized studies -- by the way, I'm not including here a review of the other studies that have been done either experimental or non-experimental in the HFA portfolio of studies across this country. In San Diego where abdomen stand alone randomized trial, it wasn't ongoing, 247 families came in. That's all those para-professionals dealt with across the three years.
Comparing professionals with yearly outcomes, replication of design and measures from Hawaii. Hawaii was an embedded service system. I believe that that substantially accounts for why we saw such low dosage in Hawaii. They basically had more families continuing to come in and that is where their investments had to go. Denver, a stand alone randomized trial comparing nurses, para-professionals, controls ... what we have here are three carefully implemented randomized trials that we're looking at issues related to para-professionals' delivery of home visitation services. San Diego and Hawaii, differences observed. Note, maternal life course. Parenting behavior and attitudes. No differences. Child abuse and neglect, really no differences except that one among users, some differentiation by year four. Year three, frequency of use. Child preventive health care. Not observed in Hawaii. Or San Diego. Home environment. Cognitive development, problem behaviors, subsequent birth or spacing. None of these were observed to differentiate the comparison groups, control groups and the intervention groups in Hawaii or San Diego.
Positive outcomes, number of well child visits was observed in San Diego, not in Hawaii. Mother's school attendance. I didn't show but there was a slight increase in that. Developmental index at years one and two in San Diego, not observed in Hawaii, but washed out in terms of developmental measures that were appropriate by year three. Frequency and mother to child psychological aggression, corporal, among years, et cetera. Denver, you see that work coming out of Denver and Dr. Olds' work. No reason to go over this. Now, this is the last slide. Let me take you through the logic of this because what we've going here are not three studies that were well linked to each other. Two of them were in terms of design and measures, but different research teams.
And then a third trial in Denver which was quite different. Different research team. Different model. One, will greater dosage result in better outcomes? Maybe the reason that para-professionals aren't getting the results that we see in here is because they are not keeping the families in long enough and they are not giving enough of the active ingredients. San Diego equaled the Denver professional trial in terms of dosage and both of them were greater than Hawaii. As far as I can see, there were comparable results across. In other words, we really didn't see clear benefit of para-professionals. Second issue is maybe greater structure, maybe very -- greater detail to the program. So para-professionals know exactly what they are to do in the home at years one, year and a half, et cetera.
Now, it seems to me that the Denver trial is clearly a program with high structure. And if any program should show benefit from para-professionals, I believe it should be in Denver because you had more program structure in a sense making up for the less educational achievement that we see in yours. But again, comparable results across the threes trials. Longer program duration result in better outcomes? Denver at 2.5 years, somewhat better than Hawaii at three years. In Hawaii, there were very few families being visited at three years, however. Again, comparable results. Keeping families in longer does not seem to be result in better been benefit. Will different context, research teams lead to better results? There was lots of difference here. Now, if we had seen consistent benefit from the para-professionals, then different context, different study teams, different program models, that would have been a terrifically robust fining. Finding. Unfortunately what we have here is a not so terrific robust finding.
Finally what I would want to say is that overall, the differences between the para-professionals and control condition as I read what came out of Hawaii and what has come out of Denver and now that I know much more that has come out of San Diego, is the differences were few in number and they were small when observed. Thank you. (Applause.)
CLAIRE DUNHAM: Is this microphone on? There's a couple of questions that came in from our online listeners. The first two came in for Dr. Olds. One was from Julie Ronoper. She is asking, does it matter if the nurses were LPNs versus RNs?
DR. OLDS: We have had very few LPNs ever work in the program. In the Elmira study, I think there was 1 L.P.N. in the study. In Denver and Memphis, we always used bachelor prepared nurses. We think there is an advantage in having Bachelor's prepared nurses especially when you are -- because there's so much independent decision making that needs to be made in the field that we think there's a real value in using Bachelor's prepared nurses. Now, in some rural areas in sites around the country there are places where associate degree nurses are being employed using the program. They almost always have had lots of experience and are working towards bachelor degrees. They have other members of their team who have high levels of preparation and we think that that works.
CLAIRE DUNHAM: There was a follow-up question that came in from Penny Selby asking, Dr. Olds, have you done any research on using other professionals such as early child and adolescent specialists.
DR. OLDS: We have not and we think those studies need to be done. There are programs that programs that use child adolescent specialists and they show little impact. One of the reasons that nurses produce more powerful effects in this program, I believe, is that we are dealing with pregnancy. We are dealing with newborns. We are dealing with women and families concerned about physical health issues and because of the tremendous salience of physical health issues during pregnancy and the early months after delivery, families want and view nurses as those who are going to be able to reduce that sense of vulnerability more significantly and that increases the persuasive power. The real knowledge and competence with dealing with the challenges that women, babies and families experience during that time.
CLAIRE DUNHAM: I have a third question from myself. And that is, actually for Dr. Olds and Dr. Landsverk. I'm interested in whether or not you have, in your analysis of your data, looked at levels or methods of determining real engagement of families. What you look at the number of home visits, there's another way sometimes that researchers will use to say of those families who are really hooked with the program or who identified through some definition, say they had a certain number of home visits in the first three months of service, for example, is the analysis of the outcomes different for those families?
JOHN LANDSVERK: It's interesting. When we look at the quantitative measures of implementation in the Denver study in particular, in, this follows up on what Dr. Landsverk was saying. We don't see that increased numbers of completed home visits or increased contact with the program is associated with better outcomes. And partly because the visitors have flexibility in their way in which the program is delivered. What has happened in all our trials is that families who need the program more, get more visits. So the ability to really sort that out in the way that you are asking, would really require a separate randomized control trial where we actually vary the frequency or duration of contact. It's just not possible to do that.
DR. LANDSVERK: Women, I certainly want to certainly underscore that what you see in the trials is what David said. The more challenged stressful families will show the most dosage. We did try to tease that out. What you'll find with the para-professionals is that substance abuse and other family challenging conditions did not go against engagement. There is good news there. We anticipated that para-professionals might have had trouble engaging high risk families coming out of the one in four births, but that did not appear there.
DR. OLDS: If I may follow up a little bit on the issue, one of the things that was striking to us is that -- in fact, in the Denver trial the para-professionals were considered fewer visits than the nurses did. It wasn't for lack of trying. The para-professionals were out trying to complete visits more frequently, but families were simply not answering the door when they were visited by para-professionals compared to nursing. They were answering the door for nurses but not answering the door as much when visited by para-professionals.
AUDIENCE: I have a question about pre-sources. It seems from any view of the literature the evidence is stronger from what you are suggesting for nurse home visitors, but when we talk in discussion? Illinois and I'm sure in other places we are talking about family case management programs. How can we get more people out there? The answer is nurses are too expensive. The second answer is, nurses don't want these jobs because they don't pay enough. There's those kind of responses that you get, especially that nurses are expensive an we will be using para-professionals to sort of as extensions in some ways because we can't afford the in yours to do this. My question is, how can we somehow marry these two models, is there a way to bring the nursing piece and the para-professional piece together in any effective way and has anyone been looking at that?
DAVID OLDS: There is a trial that has been conducted right here in Chicago that has attempted to bring nurses and para-professionals together. The impact of that intervention, I think, is promising, but I would not say spectacular, from what I've seen. I think that the question about resources is an important one. We need to acknowledge that nurses are in short supply. One of the things that strikes me -- I want to make two points. One of them is that this is not a job for all nurses. First of all. But for those nurses who are committed to this kind of work, it is a job that they love. They love because it gives them a sense of deep gratification. And we have seen that nurses have moved to communities where the program is operating so they can get a job in the program. And our data indicate that once nurses get involved in this program, they are less likely to drop out. Actually, we think that some dropping out is probably a good thing. Maybe the nurses didn't fully appreciate what this job is really all about. Nurses have said to me over and over again, and they'll say this publicly. This is not a job. In many ways this is a calling. And they view it -- I think that the impact of the program that you see here comes from that deep sense of commitment and sense of calling. And I don't think you can achieve these kinds of results if nurses are treating it as just a job. I think this comes from something much deeper inside.
DR. LANDSVERK: I won't comment about the nurses or para-professionals except to note that from the beginning in dealing with child welfare at the State level in California, there was just a very clear notion that nurses were too expensive and too rare to use. And that was just the way it was. So the question then becomes, how could we ever afford to use the nurse home visitation program? I think the question probably needs a view of now these findings across these three studies, the question needs to be posted: How can we afford to continue public investment in programs for which there is no observable benefit? (Applause.)
AUDIENCE: I'm the only Olds model program in Illinois and we started four years ago. I personally went to the training. I'll make a few comments. The structure is the curriculum. It is just magnificent. I have been a nurse for 42 years. What is it that we can do with a family for three years? I was in home health for a lot of years. When you see the curriculum and you take the training. And I did that along with some of the county board people, king county, about 40 miles out. It is a positive, it creates all of the things. Recruitment. Absolutely no problem. If I have to recruit for case management or the Olds model, I will always have openings in case management. Case management is a tracking system, it is not public health in nursing. This one is.
Twenty-five cases, we have a lot of support within the faculty at Aurora University. I have gone from (inaudible) nurses to six and I'm full. Every time I turn around, we have more. We have other projects in the families for para-professionals. The cost is almost the same, folks, when I ask the question. Mother's day, our first graduate said I'm married I'm a high school graduate. I'm going to college. But mostly I am a great mother! Okay? And they move back, okay? My counterparts around when they move into the para-professional model, they come back. Why? Wouldn't we all love to have had a nurse for three years almost (Chuckles.) And asked the question: How am I going to get through this? And all of those outcomes are just so significant. We've got to move, you know, towards more and more of these programs. We saw Dr. Olds with Barbara boxer and Hillary Clinton in DC a few weeks ago and other funding. There is going to be other funding. The board of education now adopted the Olds model here in Illinois. We have to make that change. But nursing recruitment? You're wrong. They come flooding. They come flooding! (Chuckles.)
AUDIENCE: I just have one follow-up on a question that came to Dr. Olds about early childhood development specialist and having been a part of healthy steps, which I'm sure that healthy steps for young children? Where early childhood development specialists as well as social workers and occupational therapists and others have worked as healthy steps specialists. I'm aware that the results are not as exciting. But it was not a risk based model. The healthy steps models was for all families, all children. So again, the outcomes, we wouldn't expect that they would be as exciting. I'm interested in your comments about that.
DR. OLDS: I'm glad you raised the issue. Because the healthy steps program is a nice example of the way in which other provider types can be integrated into primary care and produce been fish effects. When beneficial effects. When healthy steps began because the group came to talk to me, I said if you want greater effects, it seems to me you should concentrate this on families who need it the most. One of the questions we have to ask, what is the return on the investment in terms of functional improvements? In terms of cost savings? And I think, I'm glad that Healthy Steps has been done. It shows some beneficial effects. I just wish that we had, you know, better data on the extent to which that program really works with the most vulnerable families. I think that the most vulnerable families, we have some tremendous challenges in really helping them pull their lives together for, both as parents and as providers. I think that with that population, we need -- my sense of things, it's just not based on good evidence. But my sense is that nurses are likely to produce more beneficial effects than say a child development specialist who begins visiting during pregnancy. It doesn't resonate, I don't think, to the parents' perceived needs as much. That's my guess.
AUDIENCE: I have a question also for Dr. Olds. I have been a nurse home visitor and I have worked in a team of a nurse and a community, in those days it was called community aid home visitation program. Now have worked with the community based Doula model, which is para-professional home visiting, but within the context of a team that includes a nurse. And in my experience, the roles that we played were different. So I'm wondering particularly in the Denver study where you compared the role of the para-professional and the nurse, if you would unpack the black box a little bit and talk a little bit about the roles they played and whether they were set up in terms of training, role definition, super vision, and outcomes that you were looking at to be either the same or different?
DR. OLDS: Well, thank you for that question because that is -- it is an important issue. The nurses and the para-professionals in the Denver trial both had essentially the same goals and objectives, the same three goals that I listed here for this program, the para-professionals had for Denver. To improve the outcomes of pregnancy, to improve the child's subsequent health and development and to improve the parents' economic self sufficiency. That's why you see some slight beneficial effect reflected in the data in terms of treatment control differences on some of those outcomes that are assessed in those domains. Where they differed was in two respects. They were both provided with detailed visit by visit guidelines, the para-professionals were provided the same sort of supervision that Dr. Landsverk talked about in his trial in San Diego. The para-professionals had twice the level of clinical supervision that the nurses did. They -- in terms of their roles, the nurses, of course, were inevitably focusing more on issues of physical health because that's their training.
But nevertheless, the para-professionals covered issues of physical health during pregnancy by reviewing some of the checklists for emerging complications appeared Your Honored women when they observed problems or to discuss these with the primary care providers, the para-professionals sent correspondence and spoke on the phone with the physicians and the nurses and sources of primary care, both the mother's care and the child's care. The roles were on the surface similar. The way things differed, though, is that the para-professionals, I think, had more frankly -- it's both a strength and a challenge, I think, because their own personal experience ended up coloring their interaction with the families. The ways in which it was a strength, I think, is that they have been there. They have been through living in the welfare system and had to struggle p and make ends meet. They are able to relate to families on those grounds. That's a strength. But one of the ways in which it became more of a challenge is that many -- an example of where that kind of, drawing on the personal experience might have compromised the ability to work with families, a lot of women in relationships with domestic violence or where the relationship with the boyfriend was not as good as it would be.
The para-professional is more likely to say get rid of the bum. I have been through it. And I'm doing find. What I don't show here is that at age four, the para-professionals visiting were less likely to be married, less likely to be living with the biological father of the child. We thought at first that may be because they were less likely to experience domestic violence. That's not the case. It was the most dissident women who had less domestic violence than the control group at anal four. The business about the personal experience affecting their work with families is both a strength and challenge. Sometimes the boundaries get blurred and I think that sometimes it can be a problem.
AUDIENCE: If I could just follow up on that a little bit. I don't think I was as clear as I could have been. What I have found is that nurses and para-professional home visitors are effective in different ways. And the roles need to be written differently. So that boundary issues and knowledge base and credibility issues can be used in different ways. And also supervised in different ways. So whether or not you have sort of a reflection of that in the program I think makes a big impact on the outcome of those different providers.
DR. OLDS: Well, yes, I think it's interesting. When we actually entered into the Denver trial with a expectation that we would see a different profile for the para-professionals than the nurses visiting families. It wasn't clear to us how it would be reflected. We see that the para-professional families, and I'm recalling -- bear with me on this. There are some hints that they may be using community resources more frequently. But that what we are seeing is that we they are using community resources because that's what the para-professionals know about, in helping the families use that more. But that means that the cost of the program ends up going up as well. I think I'm thrilled that there is a randomized control trial of the program going on here in Chicago. That will give us; in your case you have an opportunity to really do it differently. Maybe with the dual concept, with your kind of supervision, we may see the beneficial effects that we would all hope for here in the Chicago trial. It's just something that I think we have to wait and see.
AUDIENCE: This question is also for Dr. Olds and relates to the cost benefits analysis. The slide that you showed on savings relating to welfare that you said would be not relevant in this era of post welfare reform? Did that slide also include savings resulting from reduced incarcerations? That did follow from some of your results. That certainly is relevant.
DR. OLDS: Yes, it does. That slide, the major sources -- because I don't have time to go into depth on this. That slide reflects primarily 54 percent of the cost savings you sue were due to reduction in welfare expenditures. That means -- and I include in that food stamp expenditures, but other parts reductions in crime related costs for mothers an children. And tax issues related to the mother going into the work force. Small reductions in medical expenditures. Medicare is a double edged sword. We saw reduced costs for injuries in all the trial, but nurses are in the business of identifying health problems and other types of problems that otherwise would go undetected. They are more likely to identify health problems and help them get treatment for those problems, which increases the costs. The same thing applies to child maltreatment, where you observe greater surveillance for child mall treatment in this program. Any reduction you see in child mall treatment has to override the increase in surveillance. I hope that explains that better.
AUDIENCE: One more question.
CLAIRE DUNHAM: One more from the online and this is -- I'm not sure how you will be able to respond, but one of the online participants is asking either speaker what you can tell us about the effects of other visiting programs such as parents of teachers in comparison with the nurse home visiting program. (Chuckles.)
DR. LANDSVERK: We have, as far as I know, one randomized control trial of parents as teachers. And the results were published in 1999 in the journal future of children. If you look at the impact of that program, they are inconsistent, not unlike with the results we saw with the Healthy Families America program published in the same issue of that journal. In both cases, the Healthy Families America and the parents as teachers, when you look at the quasi experimental studies, where you have those who participated in the program compared to non-randomized control groups, there was also better. But unfortunately, we know now after decades of work in the health and social sciences, that unless you have a randomized control trial of the program, replicated randomized controls of the program, you are likely to get the wrong answer. So we just need to keep that in mind.
CLAIRE DUNHAM: Okay. Well, I wanted to thank both of you for an extremely provocative session. Everyone is talking about what to do next. (Applause.)
AUDIENCE: I want to invite everyone to fill out the evaluation forms, and join us upstairs on the 12th floor for a reception. There will be appetizers, cash bar and we hope to see everyone there and tomorrow morning. Tomorrow the plenary begins with workshops. Look at your program book carefully and come upstairs and join us, everyone.