AMCHP 2005 ANNUAL CONFERENCE
DELIVERING RESULTS, IMPROVING PREGNANCY & BIRTH
February 19-23, 2005
BETSEY ELTONHEAD: Thank you Lisa. Good morning everyone. And thank you for the opportunity to talk about this. HealthMetrix is a healthcare consulting company located outside of Boston but we do what we call best practice projects throughout the country. And we’ve done it on prenatal care, on cervical cancer screening, on women’s reproductive health, and the list goes on. And we always find them fascinating and so we were very happy to be approached by New Hampshire again because we’ve done it, best practice projects there on WIC services and also on family planning. So it was great to go back again and do home visiting. We have a strategy in terms or doing best practice projects that first you define the service that you’re doing the best practice project on. And for home visiting we wanted to look at not just the episode of care, which as Lisa talked about, started from the enrollment in the program to when the women graduated with their baby, which was the child’s first birthday. But also the visits that made up the episode of care. And how do we do this is we will systematically measure across the sites and there were six sites that we, chosen, I asked to do.
Clinical quality, unit costs, participant satisfaction and staff satisfaction because you’re looking for that magic combination of all those things that everything is working well, and then you’re trying to find out, what is it that they do, that in there process and their structure that then create those wonderful outcomes. I’ll give you the ending first so you know what happened is that what we found overall was this project had very high participant satisfaction, the highest I’ve seen in all the best practice projects that we’ve done in HealthMetrix and I think speaks towards how valued the service was by the participants. Good clinical outcomes, more variation there. High staff satisfaction and the opportunity in looking at the costs and how the different programs structured things, there’s an opportunity to reduce unit costs on an average by 19 percent. You can also take those savings and then in fact, increase your access 23 percent, while maintaining or enhancing the clinical outcomes and the satisfaction outcomes. Now how is this possible?
Those of you who’ve been around the block like I have and don’t like to necessarily admit it is that if you look at the same service being provided in different settings, you find there’s a lot of differences in how that service is being provided. And the name of the game is variation. That’s one of the reasons I love my job. I’m kind of a variation geek. You just kind of go out there and you see, oh you do that? How does that happen? And you do the measurement of the variation. See a little shaking of the heads, yeah. You know that you’ve gone to different sites and you see same service provided differently. Now, there’s good, reasons often, for why that’s provided differently. There’s often questions though about some other reasons why it’s provided differently. Could be the population that you serve. But it could be the training, just how people got trained. It could be the structure of your physical setting. It could be various different reasons. And here is just to kind of tease you a little bit.
We saw a lot more variation than this, but you can see the variation that we saw looking at home visiting. The number of visits per episode, you can see the range of 24 to 50. Now this is for the people that graduated from the program and this could reflect not only the differences of when they entered the program in their pregnancy, the average is about the second trimester, but also, the frequency of visits that the participants got. The time here for the prep to the visit, the preparation to the visit, the length of the actual visit time in the home, and then the follow up time after the visit to do the charting and the referrals and everything, this is per visit. You can see again the range, and this is on average. So you compare averages per site. And you can see the wide variation. Just look at the length of the visit time, 25 minutes to 86 minutes. And the prep time really ranges too. Now as Lisa said also, this is a program that uses both registered nurses and usually Bachelor’s level home visitors, trained home visitors.
What we found is the percentage of visit in an episode of care by home visitors if you compare across the sites, ranged anywhere from 49 percent to 88 percent with an average of 77 percent of the visits typically performed by the home visitors. No show rate. Actually kind of low compared to a lot of other sites I’ve been to in terms of looking at clinics, but two to thirteen percent and what we call non-direct clinical time. So this does not include the visit prep, the actual visit, the follow up, the travel, the playgroups that they might do, it’s the things like meetings. What else, the time spent on no shows. So you get ready to go to a visit, they’re not there, and other time, things for training, things for special events or whatever. And then the management and supervision FTE and what was interesting about that one is it didn’t necessarily relate to the size of the program. So you go into then the measurement of the outcomes. So we saw the variation. Now we’re measuring outcomes.
The first thing we did was measure the participant’s satisfaction through a one page double-sided survey, fifth grade level. And there was one question on there that this bar chart reflects, there are a lot of other questions but this one was rate your overall satisfaction with the services received to date. Seven was completely satisfied, one completely dissatisfied, an average of 6.78 across the sites, which as I said, was the highest I’ve ever seen in all the best practice projects that we’ve done. What’s also interesting though if you think about seven being completely satisfied. How often do you go to a healthcare provider and say you’re completely satisfied with what’s going on there? But six would be very satisfied. There really is not significant statistical difference than between the averages of the sites. All you know is that the participants are pretty happy and I think that speaks towards the services that the New Hampshire home visitors and nurses were providing. Went a little further and we looked at the drivers of satisfaction.
So as I said, there were a bunch of questions in the survey and we then did multi regression analysis to find out what are those things that really impact the overall satisfaction of the participants. And you can see how the different sites line up there. What was interesting were these ones.
The staff were organized and knew what they were doing. That’s how the participants rate the perception of the staff being organized and knew what they were doing. And the rest of the ones had to do with the participants rating on how very clearly they understood the various things such as pregnancy warning signs or why breast-feeding is good for you, taking care of the baby after delivery and so forth. What was, we went a little further and we thought about, well what are the, are there any relationships? And we looked at but did not find a relationship between visit length and the participant’s satisfaction with the visit length, or the overall satisfaction with the program. Because some people would have hypothesis that the longer you spend with them, the happier they’d be or the shorter the time you spend with them the happier they’d be. We didn’t find it either way. We also did not find any significant differences between the staff type providing the visits.
So participants were just as happy with home visitors as they were with nurses providing the visits and visa versa. So that was, just, there’s a lot more in this but I have, unfortunately with the time, I’ll just have to tease you a little bit and tantalize you. We’ll go on to staff satisfaction. Again there was a one page double-sided survey given to all the staff that provided services in home visiting. This bar chart reflects that magic overall satisfaction question. Again the same scale, with seven being completely satisfied and you see that a little lower satisfaction but generally still pretty satisfied, a 5.9 with six being very satisfied, five being somewhat satisfied. So this is not unusual in best practice projects that you often will see the staff slightly less satisfied than the participants. Here you see the drivers, caseload per staff member, this is again through the multi regression analysis to find out what are those particular things that really drive the overall satisfaction with staff? Peer support, written policies and procedures, and amount of information provided to participants.
We then looked at and tried to find relationships but did not find relationships between the actual time spent on paperwork and staff satisfaction, because we thought maybe if they spend a long time, you know, they had a lot of time to spend on the paperwork they’d be happier or visa versa, we didn’t see that. We looked at the hours spent in meetings and the staff satisfaction with peer support, there was no correlation there. We did look at, and we also looked at the actual caseload per FTE and the staff satisfaction with the caseload and at first glance it also looks likes there’s no relationship there. We’re going to look at it a little bit more in depth because we have to make adjustments for travel. But what we find in the best practice projects that we’ve done is when you find a best practice site, it’s the site that has setup its structure and processees that allow the staff to fly, that they can do their job and they can do it well and they can do a lot of it and still be happy because the way everything is structured, is to bring out the best in the staff.
And so that was not actually a surprise to find at first glance it doesn’t look like there’s a relationship there. So you have very happy participants, pretty happy staff, a little bit more variation and then we get into clinical indicators. The way we did this is we, we did a retrospective chart review and looked at various things, demographics and various clinical indicators that were important. The first thing we looked at, were delivery outcomes. You can see the various sites across there and the charts were reviewed and we separated out high risk from low risk to come up with the various different percentages of adverse outcomes and they’re defined below, things like, you know, gestational age below 37 weeks and the weight below 2,500 grams or above 4,500 grams, kind of classic, the traditional adverse outcomes that you look for in delivery. And again you can see the range here. What was interesting here is when we look at the impact on the demographics on the outcomes, one thing that was surprising that came up was that the site, if a site served, was in a rural area, and served a rural population that they were more likely to have adverse outcomes than the sites in the urban settings, which when you think about it makes sense, because you think about availability and access to services, and it clearly came out here that if you knew where these sites were located you can go, five and six are kind of out in the middle of no where and it goes on from there.
The other things that we looked at for clinical indicators, we had a lot more tables but I won’t show them to you today because of time. WIC enrollment, health insurance for mom and baby, healthcare provider for mom and baby, whether the mom was screened for depression throughout being under the service, the smoking, whether the mom decreased or increased or stayed the same. And also whether they breast-fed. What we did was we looked at the status of these things when the mom entered the program, when she was pregnant. We looked at them again right after delivery and they we looked at them again right before the exit from the program. Then what we did was we ranked them and this is what you can see here in terms of the programs. And here it’s like a golf score. If you look at the very bottom, the lower is the better score. And you can see the best practice came out the best of comparatively of well, all the sites. But why we do the clinical is that if we’re finding there is a site that has very happy participants and very happy staff and the costs are lower, and there’s things that they’re doing that produce those lower costs and the happy participants and staff, we want to make sure that the clinical outcomes are good before we start making recommendations to other sites about you might want to try this, you might want to try that. And this then gives us confidence in giving those recommendations to the other sites to consider that we’ll say at least you’ll maintain your current clinical outcomes and hopefully you’ll improve them too.
Then you get to costs. This reflects what we call adjusted costs. We do give unadjusted costs to sites too, but then we have to scrub them a lot, so that you can compare apples to apples and you can kind of see then is this, you want it to reflect the structure in the process of the site, not necessarily that they have a higher risk population or maybe they pay their nurses more compared to other sites. So we make a lot of adjustments in terms of salary, assuming travel time is the same for each site for a visit, assuming that only English visits are done because there was one site that had a higher percentage of Spanish speaking population and in fact, those visits did take longer because interpretation issues. We took out indirect costs. So there’s a bunch of scrubbing that we did and what this reflects then is a range of cost for the episode of car for home visiting from 3,170 to $10,710 per episode. So you see the wide variation for costs. The average is around 5,600, 5,700.
Now what’s different here is you can see the best practice model that’s identified that’s different from the best practice. What was wonderful about this project is that the best practice that cause, that’s in there that was about $3,500. They had the majority of indicators that made them the best practice. But there were certain best sub processees that we identified at other sites that even the best practice site could learn from. And then if you take those little processees and stick them in the best practice model, or best practice and model it, you actually come up with the best practice model. In terms of then, well what, what are some of the things that made up the best practice model that we discovered? They’re closely aligned then to the drivers of cost. The first one that had the biggest impact was what we call the percentage of non-direct clinical time. That again is the things like how much time the staff spent on meetings, on events, on other things that you can’t account for, on no shows, things like that.
And we found the best practice model, that percentage was under 20 percent. Once it starts going over there it starts to really increase your costs in terms of providing the service. The other thing that we found was for the best practice, the time spent on the visit and the associated function. So again you have the prep, and the visit, and the follow up, and you can see the various times that we found that the best practice was able to produce excellent outcomes in terms of satisfaction and clinical but do it within that amount time, which was very efficient. Now what we did then, and I’ll go on to the other ones, but each of these recommend, each of these issues we then provided recommendations to the sites, which was often just having the sites talk with each other that, of ways, of how could they make them selves more efficient. To give you an example, a lot of staff spend time copying forms before they do the visit.
One site, it was the best practice sub processes actually had an administrative person or a volunteer do mass photo copying and file it up, so that when you’re then a very expensive nurse compared to the home visitor, all you do is go to the file and you pick out the form. Much better use of the nurse’s time or even the home visitor’s time then spending the time for each visit trying to figure out what they need to copy and take with them. So it’s kind of very common sensical things that we then are able to identify through the best practice and then inform the various sites about this is what this site does. The other thing that you should be aware of, we do these projects with blinded information, so that the sites getting this information will get their own information, they’ll know who they are in terms of their outcomes. But all the information on all the other sites is blinded and yet the state, gets all blinded information.
All states always try to guess which side is which, but we’re pretty good at kind of mixing things around and the idea is to do a project that allows for a feeling of safety so that the sites can really share with you what’s happening and you can provide them the actual information back and then it’s usable. Another driver of costs and also important to the best practice was the staff mix. This is what typically happened with the best practice at 88 percent of the visits were done by the home visitor for an episode of care and 12 percent by the nurses and I, we got very specific about how they did those visits. What’s important here is that nurse’s salaries on average compared to the home visitors were 88 percent higher. So when you think about that, you really want to leverage your nurses. You want to make sure you use the nurses in a way that nurses have been trained to be used, not necessarily having to make referrals to the Medicaid Program because you could have a home visitor doing that. Not maybe doing, you know, calls to various places or filling out forms, you know, that doesn’t need the nurse’s specific knowledge.
So one of the things about using the nurses is be very clear about when they go out to visit the people, it doesn’t need a nurse. Now, this is what typically happens, you want to make sure that who goes out to visit is customized to what the needs of that particular individual are. So this is just more guidelines rather than it’s a rule. But then you have to have it in your head about what’s important to be done here and who’s the best person to do it. And the last thing that we made a recommendation on was on supervision. This is for the sites, the sites that we were looking at that the best practice was a pretty big program and yet they were able to do it within about a FTE for management and supervision, which included the clinical supervision too. And as the person that I was working with in New Hampshire said that one of the reasons they think this program was the best practices because it was a child and family health program and that they’ve been doing it for over a hundred years, so they’ve kind of got it down, they’ve got their systems down for other programs and they probably then used that for the home visiting program too.
So that’s it. I’d be happy to take questions at the end of the, everyone discussing, but I think it gives you an idea then about how you can take an approach, a best practice strategy approach to systematically measure the outcomes the you’re trying to achieve, compare across sites in a way that’s fair, and also learn from that in terms of what are the processees and structures that produce those outcomes and then the sites can learn from each other and then everyone can raise up to a higher outcome measurement. Thank you very much.