DONNA STROBINO: Thank you, Jerry. I’m actually delighted to be here to talk about Healthy Steps. One of my partners in crime is Cynthia Minkovitz who’s actually the P. I. on the five and a half year follow up that we’re currently going to give you some data on today, of Healthy Steps. And I also want to acknowledge, sort of, the role of several other people in the project. In fact, our evaluation team, at one point, I think, was somewhere between 25 and 30. But a couple of other key contributors to that were Bernie Geyer who was the original principal investigator on the first component of the Healthy Steps evaluation.
Holly Grayson, who’s here, who has been the director of the Women’s and Children’s Health Policy Center, the Women’s and Children’s Health Policy Center actually on their Web site will have all the documents that you need to look at if you want to know more about the science and you want to know more about several of the other publications that have been published with regard to the Healthy Steps evaluation. So I’m delighted to be here.
As Lisa Simpson showed us this morning, she provided us with some Commonwealth Fund data suggesting that there were some deficiencies in the amount of information that parents were getting about developmental topics. That was one of the key motivations for the Healthy Steps Project. Cynthia had also done some work indicating that there were actually practice barriers to delivering services, particularly developmental friendly services. And there’s also been a renewed interest in looking at healthcare professionals and their relationship with families through the Future of Pediatrics Education, Number 2 that was published in 2000. I know there’s been some other publications subsequently that focus on the family.
Let me tell you very briefly about the Healthy Steps Program. Mike Barth is going to tell you and Anita Berry are going to tell you a lot more about it this afternoon, but basically it’s a program that was jointly developed through Boston University Department of Pediatrics, Barry Zuckerman as the lead there, and then the Commonwealth Fund with--this is basically the last major project that Margaret Mahoney who had been the Director of the Commonwealth Fund was involved with and that was jointly developed.
It is a physician developmental specialist partnership. It has several components to it including: enhanced well-child care and enhanced with regards to developmental issues, home visits, a warm telephone line, so basically a telephone line that you can call the developmental specialists if you had questions about development; child developments and family checkups; written materials that basically involved some link letters, educational information that involved bringing the family up to date with what their next visit was going to be and what they needed to know to with that visit; parent groups and then linkages with community resources. And the project involved two Healthy Steps specialists per site. And as part of the original project there was to be one Healthy Steps specialist to 100 families and I know that has been modified to some extent in the subsequent projects.
So the program goals were several. The first was to promote the knowledge skills and competence of mothers and fathers in their child-rearing abilities. I’m going to give you some date related to that today. To promote the healthy development of young children, I’m not going to give you information on that. We decided from the beginning that we were not going to look at child development as an outcome, although we do look at parent's perceptions of their child’s development, particularly social development, and their behavior. And then to promote the clinical capacity and effectiveness of pediatric primary care to meet the needs, particularly developmental needs, of families with young children. I’m not going to talk about that today, but you will see on the center’s Web site that, in fact, there are some papers related to the changes that we saw in pediatric practice.
Basically, sort of, to boil down the approach as to what makes Healthy Steps unique, it’s basically expanding pediatric services in the home practice of the families by adding developmentally oriented services that are both in the home and in the office and that are coordinated between the home and the office with the same person providing that care. And it’s also a universal intervention. From its very beginning it was seen as an intervention that was not just for disadvantaged families, not just for low birth weight babies, but for all families.
Our design, which is basically you can go to the Web site of pediatrics in 2000 and we have a very detailed article on the design for those of you who may want to fall asleep or may be of interest to it.
But basically there were 15 sites across the country and at six sites we were actually able to implement a randomized control design where we actually allocated, in blocks of four, families to intervention or as a control. Control in both instances for the quasi-experimental and randomization was usual care. So that intervention was those elements that I described before that were added on to usual care and there nine quasi-experimental sites. These were across the country. For the nine quasi-experimental sites we had nine control sites, so we actually, as part of the evaluation, worked with 25 pediatric practices--I mean 24.
What I’m going to talk about today are really the results that look at whether or not Healthy Steps enhances quality of care and parenting practices. We started with the sample. As I said we recruited up to 200 intervention families. We started with almost 5,600 children. We had close to 3,000 in the intervention group and 2,600 in the control group. We recruited them between September 1996 and November 1998. And we did that because we needed to stagger enrollment across the 25 sites with which we were working because we actually went into each site and worked with them in developing the evaluation and how it was going to be implemented in each site. And we needed to stagger that because we had just so many people who were qualified to be able to do that.
In order to eligible, you had to be able to make a visit to the practice within the first month of life and you were recruited--to be well enough so that you can make a visit--so we eliminated those sick children who could not make a visit during the first month of life and you were recruited either at the office or the hospital. And in the practice they were followed up to three years and we’ve actually now followed the families up to five and a half years.
We had several sources of data. We did a newborn registration form on all of the families, which included quite a bit of information about their demographic characteristics, about how they learned about the practice and some other information about services they got during pregnancy. We then have done a series of parent interviews of which at least 98% of them are to mothers. We did them at two to four months, 30 to 33 months and followed them up at five to five and a half years. The first two interviews were supported by Commonwealth and all the local funders that funded the program, as well as the evaluation and the five and a half year follow up was actually funded by ARC. And we did medical records reviews. And what you can basically see is we have close to a 90% response rate at two to four months, two thirds at 30 to 33 months and about 60% at five and a half years. And I think we got--I think it was 87% of the kids we followed with medical records.
Our families are very diverse. While they don’t exactly reflect the exact percentages in the U.S. , they reflect the diversity of the U.S. population, so we have a significant number of teenage moms, but also not seen on this graph, a significant number of moms who were over 30. We have both low-educated and high-educated moms. We have several minority groups represented.
The mother’s employment about 40% worked during pregnancy. Dad’s employment about 90% of them were employed. About two thirds of them were married with the father living with them. We had about 6% low birth weight babies and we had about a little bit over 30% whose babies were insured by Medicaid.
So we have a diverse sample. It does vary across the sites in terms of the characteristics of the samples at the sites. I want to talk about two types of outcomes today.
One is quality of care, where we use the IOM report to look at those aspects in quality of care that were important, including effectiveness. The services actually worked. They’re good services. Family centeredness, timeliness, getting care at the appropriate age, efficiency, staying with the same provider and then equity making sure that we see the same kind of effects across different family types.
I’m going to talk a little bit about some of the parenting outcomes that we looked at. Parenting practices that promote development were some of them. Parenting practices related to safety and then finally parents perceptions of their children’s behavior, their health and their social development.
We did an intention to treat analysis, which meant that you stayed in the same group even if you left the practice you were still considered to be part of the practice. Our analysis that I’m going to present today are mostly logistic regression analysis that are adjusted for baseline characteristics. The table I showed you about the family we adjusted for those characteristics and a few others. We also adjusted for site. The families that went to a given site tended to be more alike than they were with two families that went to other sites and so we needed to take that into account.
We did, for those of you who are interested, we used general estimating equations to estimate the outcomes and I’m reporting most of the results as odds ratios and when I’m not I’m going to let you know. Okay.
So basically what our hypothesis was or our conceptual framework was that Healthy Steps would result in better quality care and that would, in turn, result in better parent outcomes and child outcomes. The first aspect of quality care I’m going to look at is effective care and here what we’re doing is we’re assessing a lot of the services that were offered through Healthy Steps. And I want to explain this table because I know when you first look at this you probably went, “Ah, what is that?” But anyway--I don’t know if I’m getting this to work or not. I guess I’m not.
Well, anyway. Let’s look at the middle two bars, okay? And what we’re looking at here is we’re comparing intervention, who are yellow, with control. Okay, so we’re comparing intervention, here, with control families. These are the raw percentages. These are the percentages that we get when we look at unadjusted results. These are the odds-ratio estimates that we get when we adjust for the covariates and site and we look at the odds of a family getting one or more home visit for intervention versus control families.
This may look very high to you and higher than what you see here as the relative difference and that’s because we’re comparing the odds of getting one or more home visit for families in the intervention group versus the odds in control group. So it’s not a risk ratio for those of you who are used to thinking that way. Almost all the rest of the tables are going to look like this, so is everybody okay with what I just said? All right.
So what you basically see is if we look at four or more services that were the Healthy Steps related services plus one or more visits, or discussion of six or more developmental topics, basically the intervention families were much more likely to get those services than the control families.
Here what we’re looking at are developmental services. So do they get a developmental assessment? Do they get books? Reach Out and Read was part of the intervention, and do they get information about community resources. These are all at 30 to 33 months. I’m not going to show you two to four month data here, but basically again we can see how much greater the odds was of getting any of these services for intervention versus control families. Okay.
Patient center. In patient center we talked mainly--I’m sorry--about satisfaction with care. And I’m not going to show--we actually developed a scale which I’m going to talk about later for the five and a half year data but when we looked at this at 30 to 33 months, we did it as dissatisfied versus satisfied and that’s confusing. So I thought I would show you this slide and this is a question we ask, “Did someone go out of the way for you at the practice?” And you can see again that the odds is two times greater for intervention than control families reporting that someone went out of the way for them and the vast majority of the families reported it was the Healthy Steps specialist in the intervention group that went out of the way for them. In the other group it was the pediatricians, which I thought was interesting.
Do they get timely care? Okay. We’ve done two analysis and these are data from the medical records and what we’re doing here is we’re following age appropriate well-child visits in conformity with the AAP recommendations for when those should happen. Basically if you look at the bottom here we’re looking at the odds and it’s close to two in most instances, and basically what you can see is that the intervention families had a greater odds of getting age appropriate well-child care than the control families at all of these age points.
We also looked at on time vaccinations, DPT 1, DPT 3, MMR and then up-to-date by 24 months and that includes four DPT’s, one MMR and three polio vaccines. And, again, what you can see is that the intervention families have a greater odd of getting up to date immunizations, as well as timely immunizations than the control families. We didn't expect to see this, what we believe to be very important and fairly large effects on utilization of care. Okay.
Are the services efficient? And what we’re looking at here an odds ratio of close to two. Did they have their last visit after 20 months, so were the kids still in the practice, you know, at the time they were about two years of age? And, again, the intervention families had a greater odd of still being in the practice at 20 months.
And then finally, equitable. I’m not going to show you these data because we’ve done a lot of analysis, none of which I think you want to see today. But basically what we found is that all of these results with regard to services and quality of care persisted across subgroups. So the persisted across parity, so they weren’t different for first time moms or second or greater time moms. They weren’t different by income and we had a nice spread of income that we looked at in *TER-tiles and they weren’t different for young versus average age versus older moms. So we had the same effects. So this is just a summary that we saw enhanced quality of care related to effectiveness, patient centeredness, timeliness, efficiency and equitable care.
Now I’m going to show you a little bit about parent practices and I’m going--here I’m going to show you both two to four month and 30 to 33 month data. And, again, we have the same kind of table interpreted the same way and here we’re looking at whether or not the parents place the baby in the correct sleep position at both bedtime and at naptime. And the intervention group had a greater odds of doing so than the control group.
Here we’re looking at book sharing and playing and the sharing of books everyday really refers to did they show the baby picture books each day and playing everyday with the child. And here, what we see is that intervention families had a greater odds of engaging in those activities with their babies than did the control families.
I’m going to show you now the data for the toddlers. And probably the most important finding that we have for the toddlers is with regard to discipline strategies. And I might add the very last slide on the hand out actually describes the questions that were asked to get at the discipline practices. So they’re there for you to look at if you have a question about that.
The majority of moms use multiple strategies, primarily non-physical discipline strategies. This is at 30 to 33 months. Half of them did spank with their hand or slapped on the hand, but 5% or less of them actually were involved in severe physical discipline where they spanked with an object or they slapped in the face. But what we found was that intervention families had a lower odds of engaging what we think of as harsh discipline practices than control families. And we compared two different types of harsh discipline practices. One that had to do with threatening, yelling, slapping on the hand or spanking with the hand and then also slapping in the face or spanking the child with an object. The second one as we think of as much more harsh discipline and, again, we have differences, about a 25% reduction in the odds for intervention versus control families.
We also looked at problem behaviors. We gave the families the child behavior checklist scales for anxious and depressed behavior, aggressive behavior and sleep problems. And what we basically found was when we looked at the data and compared our data against national data we found that the mean scores on those problem behaviors were similar to other data that had to be reported on none referred populations, but they were substantially lower than the clinical cut offs. That said, there were more Healthy Steps mothers who reported aggressive behaviors and sleep problems.
And these show you the means and actually when I show you the five and a half year data I’m going to actually show you cut offs for clinical or potentially borderline problems. But basically, the intervention families reported more aggressive behavior, more problems sleeping. We believe that this is a result of the learning that they have from the Healthy Steps specialists to be more vigilant about what’s going on with their children and also to be more willing to discuss behavior issues with the Healthy Steps specialists or with anyone at the practice.
So this is the big question. Actually, you don’t have this on your handout because these are fresh off the press and we’re actually going to be writing them up. So my colleagues felt that we should probably not share them publicly until we’re going to publish them but are there sustained or emerging effects? You know, sometimes in these projects you see sustained affects and sometimes you actually see findings that you didn’t expect to emerge later on. So we followed the families to the children being five and five and a half years.
So we looked at quality of care, parenting practices and perceptions of the child, and, again, we used the quality of care chasm and what I’m showing you here is information on anticipatory guidance and these are for families who remained in the practice at five to five and half years of age. I can tell you the findings are very similar if we look at the full sample for those who actually left the practice. But the intervention families had a greater odds of discussing four of six different developmental topics than control families. This is where the project stopped.
They report that and what they report is also that anticipatory guidance measure is--I think it’s out of CAPS, but I can’t tell you for sure on that but basically it’s a series of questions that are asked about, "Did you discuss this topic with your provider?" If you answer yes, then you’re asked, "Well, did you get what you wanted?" If you answer no you then are asked, "Well, do you know something about the topic already?" And if you did then you’re not asked further, but if you didn’t then you’re asked, "Well, did you want to discuss that topic?" And it’s, sort of, a scaled measure here, which basically shows that, again, the intervention families had a greater odds of getting what we would call the needed anticipatory guidance that they wanted, relative to control families.
Satisfaction with care, now this is the measure that we use. It actually had to do with a series of items that we asked about related to whether or not providers at the site supported the families with regard to developmental and health related issues. And here we’re looking at satisfaction with care and this is that they actually reported that they were highly satisfied with care. You can see we have high percentages baseline here, that’s often seen in these in these survey’s, but we do have a greater odds of the intervention family saying that they were more satisfied with the providers support they got than the control families. Remember, again, this is after the program is still over but they’re still at the same practice.
Retention of practice, this is, sort of, a similar measure to the measure that we looked at after 20 months but here we asked them at five and a half years were they still in the practice. And what we found is that there was still a greater odds of the intervention families still being in the practice than the control families. And this is an issue that’s often very important to HMO’s, in terms of retaining the families in their practice.
We also looked at the discipline practices, and we have similar findings to what we had at 30 to 33 months. And that is, again, that intervention families have a lower odds of reporting ever slapping in the face or spanking the child with an object than control families. And they’re also at five and a half years have a greater odds of reporting that they negotiate often or more with the families than control families, so they use less harsh discipline practices and more what we might think of as appropriate discipline practices.
The intervention mothers also reported that their children looked at books daily. A greater odds for the intervention children than the control children. Again, Reach Out and Read was an important intervention for the program and from the very beginning the Healthy Steps specialists encourage families or parents to actually read to their children and show them picture books.
And then finally these are the data from the child behavior checklist. We use the one from two to three years of age for the 30-month children. This one is the one that captures behaviors for children at six to 11 years of age. There’s not one--yeah, this is the six to 11 years of age because we decided that the four to five year old instrument was no longer appropriate. But basically what we see here is, again, that intervention families are more likely to report problem behaviors in their children that are of either borderline or of clinical concern than control families and, again, we think that this is probably a result of the early education they got to become more aware of their children’s behavior and to talk about it more. Okay.
We saw no program effect at five and half years on safety practices. When we did see them on at least sleep position at two to four months. And I also didn’t comment on the fact that at two to four months we also saw some effects on feeding patterns. Intervention families had a lower odds of giving cereal to their baby by two to four months or giving water to the baby than the control families.
We didn’t see any program impact on perception of the child’s health or the child’s development and we had a series of questions that we asked about the child’s health. The series of questions we used the PEDS, which is Francis Glasco’s instrument that asked about concerns that parents have. We also asked about some delays and other problems and we found no difference there. And we didn’t find any differences in terms of the perception of the child’s social skills and we asked a lot of information about the social skills. Okay.
So what are the limitations? Well, we have parent report of the receipt of the services in parenting practices. Now, with regard to the utilization of well-child care and immunizations, those data do come from medical records. The data, though, about the developmental services about home visits all came from the parents. It would be very hard to believe, from my standpoint, given the kinds of differences that we saw, that they did not reflect real differences because there were huge differences in terms of the receipt of developmental services for intervention versus control families.
We also asked about parenting practices. And while we might be concerned that there might be reasons why intervention families might describe more parenting practices than control families, the fact that we had consistency with regard to the discipline strategies from 30 to 33 months to five to five and a half years suggests to us that those are probably real findings. We also had at two to four months some differences in favor of the intervention group with regard to reading books and we saw that finding again at five and a half years. So we have reason and we didn’t have some of the other findings that I noted at two to four months or 30 to 33 months which makes us believe that we probably do have some real findings here.
One of the other things that we had to do was we needed to do a modified intention to treat for the vaccine analysis. I don’t know if any of you have ever done it but it’s fairly complicated but you have to be able to have a visit after when you would get the vaccine to be counted into the sample that actually got the vaccine. So it’s not intention to treat because all we’re doing is picking up people who had a visit later on and we’re getting rid of some who dropped out and in intention to treat you’re supposed to keep those who dropped out in. When we play around with it we still found pretty much the same findings when we did it on a different denominator. But in fact, the people who do vaccination research would have not been happy with the way did the analysis because this is the way it’s usually done and to make it comparable we needed to do it that way.
We have select pediatric practices. We know we had, in some instances, the cream of the crop and that we really started off with a high ceiling. That said, the fact that we have findings suggests to us that there may be something real about this intervention.
We also have attrition and the attrition is not in the unexpected direction. As we follow the families more and more we got higher or less disadvantaged families in our sample. We had the more advantaged families in our sample. While that’s the case, we did not have selective attrition. So the characteristic of the sample that we followed-up for intervention families was similar to control families, so that we didn’t lose anybody differently in intervention than control families. So we don’t think it biases our results. If we did have selective attrition we would be concerned about bias, but we don’t think that’s the case. Okay.
So what can we say about the Healthy Steps intervention? Well, Healthy Step intervention family received care which was more effective, family centered, timely, and efficient. Healthy Steps lead to improvements in some of the parenting outcomes, and Healthy Steps had sustained, although small effects, for quality of care, discipline practices and perceptions of child behavior.
So what are the implications of this? Basically, this is a universal practice-based intervention, which suggests that by doing this we can actually enhance the quality of care for families of young children. All of us would agree in this room that that maybe a goal unto itself, but it isn’t necessarily going to buy us the political clout we need to think about paying money and intervening or developing this intervention in a practice where we may not have reimbursement or other costs covered for it. It, however, is a new strategy that appears to meet families needs.
So it goes beyond just quality of care and that can benefit all families not just those who are poor or are high-risk outcomes and almost all the other work that’s been done in this area is either Head Start or it’s the ECLC. It’s not the ECLC. It’s the newborn low birth weight follow-up that’s been done by Jeannie *Roksan and her colleagues where I think the kids are now passed eight years of age. But those were high-risk children and they were low birth weight. Okay.
So what about implications for early childhood intervention? You know, we’re involved as a healthcare system and you as health providers or individuals who are doing research, or who are in the practice community, thinking about what to do on our early intervention services. And one of the things that's very nice about this intervention is it coordinates and integrates services for families and it does that in very important ways. It does it in the practice and links it in the home. Okay? So we have the same person who’s seeing the child both in the practice and in the home, and it also brings resources together from the community that the Healthy Steps specialist or people in the practice can mobilize for that particular family.
It uses a non-physician provider, so most of the Healthy Step specialist were nurses, early childhood educators, or a few of them were social workers. So it uses a non-physician provider where the cost of that provider is less than a physician that frees up some of the physicians to be involved in other activities besides giving developmental content to the parents. And therefore on average it should cost less money in the long run if what you’re really interested in is implementing development and services and making a commitment to doing that.
And then finally, it’s for everybody and I think the thing although I don’t have the findings here, but one of the interesting findings here, but one of the interesting findings for us was when we’ve looked at some subgroups, particularly at five and a half years, we’ve actually seen although there maybe differences across all income groups, it looks like the middle income group may be benefiting more than either the low- or the high-income group.
And one of the things that we found early on in looking at the evaluation of Healthy Steps was that the amount of services that low-income families got was much more than the middle- and high-income families and that, in fact, what we were doing in terms of the home visit in particular was we were increasing it for low-income families, but we were brining middle-income and high-income families up to the same level in terms of home visiting. So in the beginning we saw that there was a different baseline.
More recently when we looked at the data the middle-income families at five and half years looked like they may have a greater effect. So we know that a lot of the low-income families are being provided services. We know that the high-income families can afford to get them. It’s the middle-income family that’s often the one that gets left sided. Sort of, like the average child in school. This was encouraging for us. We haven’t gone ahead and done additional analysis so that’s, sort of, just a preliminary look but it suggests to us that there may be groups who don’t normally get these kinds of services who could truly benefit from them.
What I’d like to do is, first of all acknowledge the Healthy Steps families and practices and there are several folks here who either were--Jerry being one of them--Anita--were either involved in Healthy Steps or are currently involved in Healthy Steps and I want to acknowledge the wonderful cooperation that we had in doing this evaluation. It really was a labor of love for probably hundreds of people besides us because we had people working in the sites and helping us in the sites. I want to acknowledge my team, and we still have folks at work who are involved in it. We have a programmer who’s been with us since the very beginning. And if you didn’t know this has really been a labor of love for the team because we’ve been involved in this since 1995, so it’s like Barth was commenting, we’ve been involved in it for 10 years and we’re still hoping maybe to follow-up the moms.
I want to acknowledge the Commonwealth Fund, too. And the Commonwealth Fund in part because of not only the fact that they were generous in their support to us but they were so helpful in their direction. They provided us so much support, particularly their leadership in doing the project and the local funders. There were some sites where there were half dozen local funders that funded the program. And they also, when they funded the program, funded part of the evaluation. And so, again, when I saw there were hundreds people involved in this evaluation, that’s true not only from the people who actually did the work, but the people who funded us and then I want to acknowledge the Agency for Health Care Research and Quality that actually funded the five and a half year follow-up.
And with that--