Skip Navigation U.S. Department of Health and Human Services
Agency for Healthcare Research Quality
Archive print banner

Preventable Hospitalization Costs and Mapping Tool: Transcript

This information is for reference purposes only. It was current when produced and may now be outdated. Archive material is no longer maintained, and some links may not work. Persons with disabilities having difficulty accessing this information should contact us at: Let us know the nature of the problem, the Web address of what you want, and your contact information.

Please go to for current information.

On June 16, 2008, Melanie Chansky presented the AHRQ Preventable Hospitalization Costs and Mapping Tool on a Web conference. This is the transcript of the event's presentation.

June 16, 2008
1:00-2:15 p.m. ET

Jeff Gallagher: Good morning and good afternoon, ladies and gentlemen. Thank you so much for taking time to join us on the call: Preventable Hospitalization Costs and Mapping Tool. My name is Jeff, and I will be the host for this event. I will be in the background answering any general and technical questions.

Today's event is being recorded. However, we will make two options available to you to be able to ask questions. You can type a question in the lower right-hand corner of the screen where you see the Q&A box; type your question in the smaller rectangular square and hit the Send button. Also, we are going to pause a few times during the call to have you ask a live question. To ask a live question, please press star 1 on the telephone keypad, say your name clearly, and press pound. You will then be introduced into the call where you will be able to ask the live question.

Also today, we are going to have a few polling questions that will appear on the right-hand side of the screen. Without further ado, it is my pleasure to turn the call over to Margie Shofer. Margie, the call is yours.

Margie Shofer: Thank you, Jeff. Hello, I am Margie Shofer in the Office of Communications and Knowledge Transfer at the Agency for Healthcare Research and Quality. Thank you for joining us for this Webinar, Preventable Hospitalization Costs and Mapping Tool. This event is the final of three followup events featuring AHRQ tools that were shared at two workshops held this past December and January. We focused on the State Snapshots at the end of April and on the Asthma Return on Investment Calculator last month. We see all of these events as the first step in what we hope to be a series of follow-on technical assistance opportunities. If, after learning more about the mapping tool today, you are interested in further assistance from AHRQ in using this tool, please let us know.

We are hosting this Webinar in response to interest in the County-Level Mapping Tool expressed by participants from both workshops. We know some of you attended one of the workshops, whereas others may have had less time to interact with the tool. As such, we will spend some time reviewing the basics of the tool and will then move on to a demonstration of using the tool. We also will address questions about data raised at the workshop and discuss interpreting results for a variety of policy applications. We will conclude with a discussion of future plans.

We would appreciate your active participation, as the primary purposes of today's Webinar are to: further probe into underlying data sources and other technical issues you may encounter in using the tool, explore practical applications for the tool, collect suggestions for future tool enhancements or modifications, and learn how you envision using the tool and for what purposes. Again, related to this last point, we really hope you'll tell us about the types of technical assistance you might need in order to make better or full use of AHRQ Preventable Hospitalization Cost: A County-Level Mapping Tool.

Throughout the Webinar, as Jeff mentioned, we will conduct a few polls. Shortly, you will see the first poll asking whether you have used the county-level mapping tool. We will mention the polling questions as we move along the presentation and would greatly appreciate if you would answer all polling questions. This is a neat feature, so we really love having it and it would be great if you could all answer it.

Today's presentation will be given by Melanie Chansky, a health research scientist at the Battelle Centers for Public Health Research and Evaluation. Melanie is an anthropologist with a focus on medical anthropology and community health. She is a project manager for the team that supports and develops the AHRQ Quality Indicators.

The other person's name on the introduction slide, Marybeth Farquhar, is my colleague here at AHRQ. She is a QI Senior Advisor in the Center for Delivery, Outcomes, and Markets. Her office oversees the county-level mapping tool. Unfortunately, Marybeth was unable to attend at the last minute, but I wanted to acknowledge her hard work on the tool.

Without further delay, I will hand this over to Melanie.

Melanie Chansky: Thanks, Margie. Hi, everyone. I am Melanie Chansky, and I am going to be conducting most of the presentation today. I really appreciate you all being here. To get started, I wanted to let you have an overview of what we are going to be talking about today.

Margie did go over that a little bit. I am going to start off by giving a very broad overview of the mapping tool, just some background on the tool. Then, I am going to also do an actual demonstration of the tool, which is pretty cool. I am going to let you guys see it on my desktop. I will work with it using a few data sets that I have for testing. Then, for the rest of the presentation, I am going to focus on the data, both the data that are underlying the tool as well as the data that are produced by the tool, and the data that you need to work with the tool. I am going to talk a little bit about how to interpret results and how to use the results, just given our own experiences in testing and developing the tool. Finally, I am going to let you have a bit of an introduction to the plans that we have and I hope to get some suggestions from you. If any of you have used the tool and already have ideas, this will be a great time to speak up. If you haven't used the tool, and anything occurs to you from what I present today, it will be really great to let us know. I can take the suggestions back to our development team.

As Margie said, you will have an opportunity to ask questions during the Webinar. I am going to try my best to answer all of the questions that you have. If there is anything that I can't answer today, you will be given information you need to get questions to me in other ways. And Marybeth Farquhar, who was going to be here, would have also had a perspective I would have liked, so that's something that might be missing. Regardless, we'll make sure all your questions get answered.

I am going to move forward. To start off, I am going to give the overview of the mapping tool. I assume that most of you are probably familiar with AHRQ's Quality Indicators, which is the overall project that the mapping tool is part of. As Margie said, I am part of the team that develops and maintains the Quality Indicators (QIs).

This slide gives you a basic overview that the QIs use existing hospital discharge data and have several severity adjustment methods that we actually incorporate into creating the rates. There are now five modules: inpatient, patient safety, prevention, pediatric, and neonatal indicators. If you're not familiar with the QIs, and you want more detail than I just gave, you can go to our Web site:

Now, let's go to the point of today's presentation, the actual mapping tool itself. This is a relatively new software tool to support the QI project. This one focuses on helping users to better understand the geographical patterns of potentially preventable hospital admission rates for selected health problems. "Selected health problem" is just a way we refer to the indicators you can use with the tool. There is also a cost component as part of this tool, as you can tell from the name. So we also hope that users can use this tool to allocate resources more effectively by calculating the potential cost savings that would occur if admission rates were actually reduced.

The mapping tool has three main functions. The first, which I believe is fairly obvious, is it creates maps. They are State-level maps that show the rates of hospital admissions for the various QIs that you select on a county-by-county basis within your State. Another primary feature is the calculation of potential cost savings that may occur if the number of hospital admissions for those health problems are reduced in each county. Finally, there is the ability to place additional information about your local populations onto maps, which will help you to indicate the number of people who are at greatest risk for those health problems in each county.

I will go into more detail about all of these aspects of the tool in just a little bit, as well as showing you what this means in the real world, in the actual use of the tool.

I mentioned all of the different modules of the QIs before. However, the mapping tool does not actually process all of the QIs at this time. As I said before, this is a relatively new tool. We only released it on October 31st of last year. Currently, what we have are all of the prevention quality indicators, as well as all of the area-level pediatric indicators. Right now, we do not have the inpatient quality indicators, the patient safety indicators, or the neonatal indicators.

That was my overview of the tool. The next thing I want to do is go into a demonstration of the mapping tool. Before I start that, Jeff, is there a way for me to see the results of the poll?

Jeff Gallagher: Absolutely. Let me share them with you right now.

Melanie Chansky: Great. Oh, great, o.k. The results here show that 78 percent of you have not reviewed or used the mapping tool in the past, so I think that this will be great. This will be a first opportunity for you to see what the tool is like and what it can do. It is very simple, so I am going to go ahead and share my desktop with all of you.

Jeff Gallagher: Folks, you should see the screen changing just a little bit. That is the normal application of the WebEx product. We are seeing your desktop. Please proceed.

Melanie Chansky: Thanks, Jeff. I have set aside a couple of icons here that are for me to use for the demonstration. I will open up the mapping tool. It has an AHRQ icon, like any other software tool would have for the QI project. What you should be seeing now is the basic Overview screen that the tool starts out with whenever you open it up. This is a very simple tool. Right now it is especially very simple now that we are in the early stages of developing it.

This Overview screen doesn't have any of the functions of the tool included, but it is kind of like an in-program help screen. You can scroll down and get some basic information about the tool and what is going to happen; it even shows sample maps. I will not go over that because that is why I am doing the demonstration today in addition to showing you how to use the tool. But this is a resource for you. As you're using the tool, you can come back to the screen at any time. We also have several help documents located on the QI Web site, and you will get the link later to where you can download this tool for free.

The way you navigate through the tool is by using the series of four buttons you should see on the left side of the screen. We are currently on the screen called "Overview." If you click the next button down, it says "Specified Discharge Data Set." You will come to a screen where this is where you are going to actually put in the data that you are going to use to run the tool. Just like the last screen, there is, at the top, some in-program help about the file format you need to use to make the tool work, the variables that are required, etc. Since that's another thing I am going to be going into, I will not show you that too specifically at this point.

What you have to do, you just need to find your test data set, which I set a couple aside. I will use test data that I have from California 2001. You have to actually go and select the State and the year before you move forward. That is really all you do here. You're just telling the program where to find your data. The real "meat" of the tool comes in the last two buttons.

I will move to the third button and select the QIs to process. Here, you can see there is not really a need for much help at the top of the screen because this is fairly self-explanatory. You can see that it lists all of the PQIs, prevention quality indicators, in this tab. It lists the PDIs on this tab. You can select one, you can select all of them, whatever you want to do. It can run all of them at once if that is what you want it to do.

I am going to just select one. I am going to select one PDI, asthma admission rate. At this point I am going to stop here and I am going to run the tool. The fourth button is optional. I am not going to do that yet. I will click Submit. The bottom of the screen goes through a few progress bars, telling you what it is doing at that point. The last step writes the data to Excel, and it puts up a box that tells you that the data processing is complete. It also tells you where to find the outputs for the tool. I click o.k. I am going to close down the tool and show you what came out of it.

I selected one PDI, asthma admission. The map is called PDI 14, named after the actual indicator that you selected. Let me increase the size on this a little. At the top of the screen, you will see the name of the indicator. You will see the counties within the State; they are actually labeled. The numbers you see on the map are the labels for the counties. Those are the county FIPS codes that come from the U.S. Census. It is just a way to link to the actual counties.

Down in the lower left-hand corner, you will see the risk-adjusted rate for 10,000 people. This map presents the rates, divided into quintiles, not based on any national benchmark, just based on the data within your State. It goes from the green, which is always the best rates, whatever that is for that particular indicator, to the dark red, which always represents the worst rates for that particular indicator. You can see on the map which ones are doing relatively well for asthma admission and which ones are doing relatively poorly for asthma admission.

This is the basic map. I will close this down because I think we have seen what there is to see there. The data that were in that map were also made into this Excel spreadsheet called PDI. What you see in this is a list of all of the counties in the State, as well as the State rate at the top. It is the only benchmark that is really included in the tool. This is what you could use for doing further analyses of the data.

Besides the Excel spreadsheet, the mapping tool produces a CSV (comma-separated values) data file, which might be something you would rather use for actually running analyses. But the Excel spreadsheet looks nicer. It is formatted really nicely. I am going to go into a little bit more detail about what is in all of these columns later. There is nothing very earth shattering here. You have the numerator (the number of cases), the denominator, the rate per person, the risk-adjusted rate per person, the standard error, the difference in the overall risk-adjusted rates—this column actually compares back to the State rate, and it only says something in this column if you are significantly higher or significantly lower than the State rate—and finally, the cost savings given a 10 percent reduction in numerator cases.

You can see the State rate, which is a lot higher than all of the county rates. If any of you see numbers that seem quite low or just out of whack in some way, it is because this is a test data set and these are not real numbers for any of the counties. You can use those cost savings in presentations or however you would like.

Let me close this down because I would also like to show you one other function of the tool: adding in the fourth box. I am going to go back and do the exact same indicator so that you can see the comparison. I am going to focus on the map this time rather than the Excel outputs: California 2001. I will do PDI 14, asthma admission rates, and then map display options. This adds a feature to the map that shows you the population at risk in each county for each indicator. I think that will be a lot easier to explain when I show it to you. However, in order to do this, you have to have a separate data set. I will give you more details about that later.

This in-program help screen will tell you everything you need to know to set up the population data set. I am going to click Plot Secondary Population Onto Map. Then it also prompts you to tell where your data set is. After I do that, I click Submit, and it runs through again. It is really doing the same things, and it tells you the same thing as it did before, where everything is saved. It always saves your files in the folder where your data set was located, so that was in the test data folder.

Here is my map. As you can see, this one adds a feature onto the map. It looks a little busy here because there are a lot of counties in California. It adds a stick figure onto each map that, in relative size, shows you the population at risk within a county. In the lower left-hand corner, you can see that the population at risk is defined for this indicator because it is a pediatric indicator. The population at risk is persons aged 0 to 17. Unfortunately, you can't define it beyond that in the current use of the mapping tool. There are certain defined age groups that are considered populations at risk for certain indicators. That is an intrinsic part of the way the tool is set up right now. So this map is showing you pediatric populations within these counties.

The numbers are divided into three groups, as you can see in the legend in the lower left-hand corner. The numbers are very low—they are actually meant to be population numbers, which would be the population of people aged 0 to 17 within that county. The reason the numbers are very low is that the data are not real. Otherwise, it would be a lot bigger because California is a big State. You can see what the gist of it is. Essentially, if you saw a county that was coated in bright red and had a large stick figure in it, you would know right off the bat that county had a lot of people at risk for this and had bad outcomes for this indicator. That might be a county you want to look at further. On the flip side, you might see a county that is coated bright green that does really well with the indicator and has a large stick figure in it. That can lead to other questions such as: What are they doing right? What are they doing to make the rates so low in comparison?

This map is a nice graphical representation of relative populations at risk, which could lead you to ask more questions and maybe want to go back and look in more detail at certain counties and the exact reasons certain counties may see different results. I think we're coming up to a Q&A period, but I want to do one more thing with the tool to show you what will happen if you have something wrong with your data set.

This data set relies on having your variables named correctly. We get a lot of questions about things going wrong with the tool, and it often has to do with variables not being correctly named. This time, I am going to use a file that I purposely have an error in, this New York data set. I am going to select New York 2003 and pick any indicator. It immediately knows that my data set is wrong. As it turns out, it could not find one of the variables. It comes up with an error message on the screen that tells you exactly which variable it couldn't find. You know to look in your data set and either find what was meant to be that variable but might have a different name in what you have or add that in if it needs to be added. I found that to be helpful because I have done demos of this before and had that exact error in my data. This was very helpful in finding that. I am going to stop there with the demo.

I believe we go to a Q&A session right now. If there are any questions—

Margie Shofer: This is Margie. As Melanie said, we are going to stop and take some questions that you might have on the presentation thus far. You can submit your questions two ways. We can either hear from you on the phone, and we love hearing your voice, so we encourage you to do that. If you would like to do that, you can press star 1, record your name, and hit pound. You also have the option of submitting a question via the Web by going to the Q&A section that you see on the right-hand side and submitting a question that way. While we're waiting for questions, I will ask you a question, Melanie.

Melanie Chansky: Sure.

Margie Shofer: From what year are the underlying data and how often are they updated?

Melanie Chansky: For the underlying data, we update this tool once a year. It is updated around the exact same time that we update all of the quality indicators. Currently, that takes place each year in February and March. When this tool is updated, every part of it is brought up to date to the current indicator specifications and current census data (if there are any new census data to include). Right now, for instance, all of the QIs are the March 2008 versions of the QIs. I don't remember if there were any census data changes we needed to make this year. The main part of the data that changes is really the QI specifications, so the data are updated yearly. We may change our schedule for updating the QIs to fall at the beginning of fiscal years, so you may see another update to this in October of '08.

Margie Shofer: I have two questions from Eric on the Web. First he asks: Please tell us what format and software are needed for data input.

Melanie Chansky: I haven't tested all of the possibilities, but any comma-separated-values file should work. The ones I have been using for testing are in Excel, but if you have something in a simple text editor, that should work as well. It just needs to be a.csv file. Also, I am going to ask about this later with a polling question, but if you have ever used our Windows Quality Indicator software, you have files that are right there ready to go into this.

Margie Shofer: The second question from Eric says: Our data analysis staff worries that our county populations are too small to make this tool useful. Can we lump county data together to get numbers with statistical significance?

Melanie Chansky: I definitely hear what you are saying. Right now, the tool is very simple. It is exactly what was presented to you. At this point, there is not a way for us to do that. It has a lot to do with the cost of the mapping part of this for us and what sort of programming sophistication it would take to allow that. However, that's not to say that you couldn't take the outputs of this tool and aggregate them yourself. It is just that we don't currently have that offered as a feature of the tool. It is definitely a concern that I could see happening in a lot of States, particularly given how rare some of these indicators are and, for some States, how many counties they have and how small the counties are.

Margie Shofer: I am going to ask the operator if we have any questions via the phone.

Operator: No, Ma'am, we have no questions at this time.

Margie Shofer: I will go back to a couple of questions via the Web. Wei asks: Are the cost savings data based on charges or reimbursement?

Melanie Chansky: It is based on total charges. I will be giving a little bit more information about exactly what that will mean for you in the next segment of the presentation.

Margie Shofer: Teresa asks: Can the tool only be used by HCUP partners?

Melanie Chansky: Well, no. Anyone can use the tool as long as they have the basic required data elements, which I will go over in the next segment. The one really important piece is that currently the cost part can only be used by HCUP partners, which is probably the worst part about the tool. The reason is the cost-to-charge ratios that are used that underlie how the tool calculates the cost. For those of you not from HCUP States, that is something we are working on. We are looking at fiscal year 2009 for a fix for that. You can use all of the other features of the tool right now, just not the cost savings part yet.

Margie Shofer: Hanten asks: Do you have the same tool for the Inpatient Quality Indicators?

Melanie Chansky: No, we do not. I think that's a good suggestion and we should try to do more to incorporate the other modules in the future. Right now, as I said, we only have the Prevention Quality Indicators and the pediatric ones. The reason for that is all the PQIs are area based, and the PDIs we included are also area based. Right now, we do not have the capability to map the provider-level or hospital-level quality indicators. However, there are a couple of important points. There are inpatient indicators, and there are patient safety indicators that are area based. If we were going to expand to anything, I think you would see us go there first to the other area-based indicators. We are trying to improve the tool and make it more sophisticated. I think we are all thinking in that direction in the future. We have only had the tool out for a short period of time, so we are still working on improving it.

Margie Shofer: Hanten also wants to know if you need to buy extra mapping software to use the tool.

Melanie Chansky: Not to use the tool, no. If you want to do anything more sophisticated, I suppose you could buy extra mapping software. However, I do not have any on my computer at all, and it works just fine.

Margie Shofer: Do we have time for one more question?

Melanie Chansky: Why don't we take one more and I will move on.

Margie Shofer: Bill says that he used the PQI software but receives the following error when trying to use the cost savings tool: Index and length must refer to a location within the string.

Melanie Chansky: O.k.

Margie Shofer: Any suggestions on that or should we save that for later?

Melanie Chansky: That is a question that probably needs to go to our tech support line because we can troubleshoot more specifically what your problem is.

Margie Shofer: O.k.

Melanie Chansky: I am going to move forward with the presentation now. We will have another chance for people to ask questions later. I will also provide you with the information you need to get more questions to the QI team in the future. So let me move forward.

The next question will be about an overview of the data that go into the tool. Before I move into it, we have a polling question that I would like you to respond to. It is about whether you have ever reviewed or used the Windows Quality Indicators software. I would definitely appreciate it if you could look at that question and respond to it. I will come back to that in a minute. Now you can see the question on your screen.

Moving forward (I did get into this a little bit before), I want to tell you about what data are underlying the tool and what the tool can produce. The main thing this tool really needs to run is the current indicator specifications about the year that the data come from (as Margie asked me to address this question from one of the participants). We currently have the tool populated with 2008 indicator specifications. You could actually find all of these in the technical specifications document that is located on the QI Web site. It has things such as the ICD-9 codes that go into the numerator and things like that. The really detailed definitions are found on the QI Web site.

The next important piece is the cost-to-charge ratio. This is for the cost part of the tool only. We use the variable total charges to do our cost calculations. The cost-to-charge ratio allows us to translate the charges into actual costs. This is why only HCUP partners can currently do the cost calculations. We have the cost-to-charge ratio set up where it is linked to an HCUP hospital ID. So if your State does not have HCUP hospital IDs, you cannot use this right now. However, the fix that we are working on actually involves using data from Medicare cost reports. I do not have many details right now because this is something where we are just getting into learning about the implications. That fix is coming in fiscal year '09.

The next piece of underlying data is census data, which is how we have those maps and county codes in the software. If there is new census data, we update it at that time. However, the kind of census data we have in the tool is pretty stable. The FIPS codes do not change very often, but if they do, we make a point of looking for that.

What kinds of data do you need to provide to use the tool? Beyond the underlying data I just mentioned, you need to put in your own data set. There are certain required data elements and then there are some optional data elements related to the cost savings and the population data. You can see from the way that I went through the tool earlier that the population data were not required to run it. I went back in and ran it a second time adding that, but it is not necessary at all.

I can see the results of the poll right now. It looks like most of you have not used the Windows Quality Indicators software. That is our more sophisticated software tool. It is the one where you can do analyses with your own data sets using all of the quality indicators. One of the reasons that I wanted to ask if you used it before is the variables that I am going to be going into on the next slide. The Windows Quality Indicators software has this very nice wizard that runs through your data set and helps you to map whatever variables you have to the variable names that we look for in the software. For the 11 percent of you who have used the Windows Quality Indicators software, you are probably familiar with the wizard. Unfortunately, we do not have this in the mapping tool right now, so you really have to have all of your variables named correctly for this to work. But that is one of the easiest things to do. It is just nice to have that option of leaving your data set the way you have it and telling the software how to map it to the appropriate variables.

Here are the variables you have to have for the software to run. If have you this minimum set, the software will run fine. I am not going to run through all of these because you can read them, I am sure. I just want to point out a couple of things.

One is the ageday variable. I should have placed an asterisk next to that because it is not required for everyone; it is required for people under 1 year old, but if you do not have it, the tool will run just fine.

For the primary diagnosis and primary procedure codes, you will probably have many others in your data set, such as secondary diagnoses and procedures. It will not be using those for this tool, but it does not matter if your data set has more than what is listed here. It will ignore those variables.

All of these variables are also required if you used the Windows Quality Indicators tool because these are the basics of what we need to get indicator rates. That's why we really have to have these in the data, in your data set.

The next slide is about optional variables. These are all about the cost savings piece, so there are just two: the total charges variable and the hospital ID. If you include both variables, a cost savings number will be calculated. As I mentioned before, we are going to make a fix for this so any State can use it. Right now, there are also some confidentiality issues associated with HCUP States, in that cost-to-charge ratios are not data that can be publicly available for everyone. There are some fixes in the tool I am not going to go into, but it calculates cost differently depending on whether your particular H CUP State allowed us to use that CCR number. If you want to know more about that, you can talk to me offline or contact our technical support for more of the details about what the calculations specifically are.

Let me share my desktop with you again to show you what an input data file needs to look like. I am going to pull a file I used to run the program before. As you can see, all of the variable names are in the first row. I think that beyond that, this is a CSV file and you are looking at it in an Excel format, so it is in columns. If you had a CSV file, all you would really need between the numbers are commas and it will know how to work with that. As you can see, you can see some of the required variables: PSTCO, patient, county of patient residence, age, ageday, admission source, admission type, etc. (They do have more diagnosis codes than we used, but those would be used if you were using the quality indicators software.) Hospital ID, MDCs, procedure codes, and then the total charge information at the end. This is all the information you will need to make the program work. I think you may be able to set it up in a variety of different ways, but we can always help you troubleshoot if you have issues with that.

Let me bring you back to the presentation now.

For the optional population data set that goes into this, this was the one that lets you make the stick figures on the maps. It has to be a separate file from your main data set. There is just no way to make it one file at this point in time. So you have to have your main data set and then have this separate data set that is just going to have a few variables. You have the county, which is always the State FIPS code followed by the county FIPS code. State is either one or two digits. County codes are three, so you will have either a four- or five-digit number there. Then you have sex, coded one for male, two for female. And age, which is really age groups. You can see how the coding works right there. Then it is population, and it is population by sex and age cell.

I am going to share my desktop again to show you what exactly this means. The data set will look very simple. I think I accidentally stuck that in the recycle bin, so I will pull that out. There we go.

You can see it is a simple file. It is also CSV, so it does not have to look like this in Excel. Each county will essentially have six cells, like you see here. This is County 001, which has six cells. It is basically just because of the age groups. The males are one; the females are two, and then each of the age groups and the population that corresponds with each of those. It is like I said, very simple. And it will allow the stick figures to be placed on your map.

Let me go to the next slide. In terms of data problems that we find from users, most of the problems are typically related to the user data sets that go into the tool. Before this Webinar, I did some review of what kinds of questions people have been asking us. It is really about the kind of data sets that go into the tool. The problems can vary a lot, but we do have a whole team that can provide you with technical assistance on your data set. If the information I just provided does not result in a data set that works with the tool, I think we have a pretty good team who can help you troubleshoot whatever problems you might have.

I am not going to address specific problems because I find they do vary by each user. I could not find one problem that came up enough for me to spend time on it. It is a lot of little things, but we always are here to help.

I think I will speed it up a little because I am a little behind where I want to be.

I will quickly go over the outputs that come out of the tool. All of the outputs that are created by the tool, which I showed you before, are automatically placed in the folder where your data set is located. You get a CSV file of your results, an Excel file of your results, and the maps themselves. I went through what the CSV and Excel files include before, so I think you know that information and I won't read it to you again.

There will be a separate map created for every QI that you select, and it will be named after the QI. In the demonstrations, I only selected one QI at a time so I only got one map. Had I selected multiple indicators, I would have gotten as many maps as indicators. You cannot put more than one indicator on a map. You can open the maps using any graphics program or picture viewer. I opened them using what they were default imported into on my computer, but I have also opened them in other programs depending on what I needed to do with them. You should have some flexibility there.

Now I have come back to where I wanted to be.

Interpretation and use of results. I just wanted to tell you a little bit about what we were thinking about as we created this tool and what sorts of things we were envisioning people might use it for.

There are some possible uses for the mapping tool data. These are some things we brainstormed when we were coming up with this. Since the tool is new, we do not have a lot of examples of what people might have done. You can use it for public reporting because it provides county-level information, if you want to report at a lower level than the State. Intervention targeting is a big thing because you could look at which counties have the worst rates as well as the highest populations of people at risk for certain conditions. If you have to make decisions about where to do an intervention, we can see this tool as being helpful for that.

The next bullet is related: tracking intervention impact. You can use this tool over time to see if the rates are coming down, if the county that was once bright red is moving into a better category. Finally, the identification of best practices is another important thing. What about the counties that are bright green?

Sorry. Excuse me for just a second.

Jeff Gallagher: Folks, we are going to pause for a few short moments so she can take a sip of water or such. Remember to go ahead and continue to type your questions in the Q&A area you see in the lower right-hand corner. Also, when we come up to our live questioning time, you can also press star 1 on the telephone keypad. Remember that you need to speak your name clearly and then press pound, and you will be able to ask your live question. We will be back to the questions and answers in a few minutes. We will see if she is able to continue.

Melanie Chansky: I think I can. I am sorry. This happens whenever I get a cold and talk too much, so I apologize.

We find the data leave you with more questions than answers. Are the rates reasonable? Do they represent significant quality concerns? Then the Excel data produced really need to be manipulated to be more appealing and more dynamic.

To address the first points about the data: Are the rates reasonable? Are they actual concerns? Within the tool, the rates are benchmarked by the State, but that's the only thing you have. If you consult the PQI and PDI user guides, which are provided on the QI Web site, you can get some information about national rates for the indicators. HCUPnet is another source that presents the PQIs at the national level. The National Healthcare Quality Report (NHQR) and the National Healthcare Disparities Reports (NHDR), which are AHRQ products, also present selected PQIs at the national level. These could help you compare what you are seeing in your State with what is going on nationally to see if what you are seeing is a quality concern and if the rates are reasonable.

We really suggest focusing on the maps for presenting the data and focusing on the Excel output to do more analyses or to use them as a source for other graphics, especially if you have more advanced graphics programs than what we provide you with. Another possibility is using the data to create a concise narrative data summary that shows what is going on within your State, or within certain Counties that you are particularly interested in.

Future plans: I only have a little information on this right now. We are looking for suggestions. Right now, we have a couple of thoughts in various stages of reality. One thing that is very likely to happen is that this mapping tool will be incorporated into our Windows Quality Indicators software. You will get some of the benefits I mentioned from the software, such as not needing your data sets to conform exactly to our variable names, and things like that. You will have wizards to help you map your variables onto the names that we want for our program.

Another thought we have is not quite in the programming stage yet. We are still thinking about how to do this. It is allowing for mapping below the county level. For instance, ZIP code level or any other ideas people might have. We were thinking ZIP code and maybe showing hospitals on the map as well. We have Q&A coming up, so if you have any other ideas for ways we can improve the tool, you can go ahead and submit them to the Q&A or any way you want to let us know about that.

I am almost to the end of the presentation. We actually have two versions of the mapping tool available. I showed you the Windows version. We do have a SAS version as well. They are both available for download at the link that is up on the screen right now. If you go there, you will find both versions of the mapping tool, as well as the help documentation that goes along with these tools.

At this point, we want to put up our final polling question for you to answer about what kind of technical assistance you might be interested in receiving with the County-Level Mapping Tool. This is really important. This is one of the main reasons for this Webinar. If you could respond to that, we would really appreciate it. You see Margie's contact information up there. For a specific question about the tool—

Margie Shofer: Actually, Melanie, I will go over that in my closing. If you are ready, we can go to more questions.

Melanie Chansky: I am sorry. That's fine. Sure.

Margie Shofer: We will give folks a little time to answer the polling question. I have some questions that people have been sending over time. If you want to ask questions and you want to ask them over the phone, you can press star 1, and you will record your name and hit pound. I think everybody has grasped how to use the Q&A button. Let me pull up some of the questions.

We had questions from two different people, from Georgette and Diane, both asking questions about if you are either in a low-population State or if you have counties that do not have any hospitals, how can you go about adjusting for that?

Melanie Chansky: I might want more clarification about what adjusting you want. Let me answer for the hospitals. If you have counties without hospitals, it is not a problem with the tool. The tool uses only the county of patient residence. One of the reasons we did this was because if we had this calculated based on hospitals, those counties with hospitals would appear to have artificially high rates for the indicators. We do not use that in the tool at all; we use just the county where the patient actually resides.

In terms of adjusting for low populations, I am probably not the best person to answer this. I could take a stab at it, but I think you want to talk to somebody who is much more comfortable with the details of the calculations in the tool. I do not want to lead you in the wrong direction. It is something we could definitely discuss further offline. This is the kind of question we routinely get in our technical support, and we have many staff who can help.

Margie Shofer: I have two people who are asking questions about DRGs (diagnosis-related groups). If your hospital billing isn't DRG based, is that a problem and can you still use this tool?

Melanie Chansky: It is a problem right now because DRG is one of the required variables. At the moment, if you do not have DRGs, you cannot run the tool. However, it is a good point to make. Maybe that is something that needs to be incorporated in the future, some flexibility in whether that needs to be required. I am trying to take notes as we talk today, so these questions will help us in the future; we can make those kinds of changes. And I think that is a really important thing to point out.

Margie Shofer: O.k. I am going to ask a question: Does the tool allow for mapping at ZIP code or medical service area level?

Melanie Chansky: Right now, no. Right now, you can only map within one State at the county level, so ZIP code and medical service areas are exactly the kinds of things we want to move into to try to make this tool more useful to users. We are looking into that actively, but we are not to the point where I could say that it is going to be implemented.

Margie Shofer: I am going to ask the operator, do we have questions by phone?

Operator: Yes, Ma'am. You have one question coming from the line of Julia Brown. You may proceed.

Julia Brown: Hi. I have two questions actually. One question, you mentioned how the tool generates quartiles for one State based on your State's data. So my assumption would be, if I wanted to compare to another State, the quartile amounts may be different, and I would have to explain that if I were demonstrating that?

Melanie Chansky: Yes, unfortunately that is the way the tool is set up right now.

Julia Brown: O.k. The second question that I have, if you have mapping software of your own, can we simply overlay your data and use our own tool?

Melanie Chansky: This is a question I have gotten in the past, especially from people who have invested in more sophisticated mapping software. The Excel outputs contain all of the rates used in our map. So you could take the outputs that are produced in Excel and CSV format and simply use that to make more sophisticated maps.

Julia Brown: O.k. Thank you.

Melanie Chansky: Thank you.

Margie Shofer: A few of you have asked for us to e-mail you the URL to the downloads. We will do that after the call. All of the things we mentioned and if you asked questions, we will make sure we get back to you after the call. I have a question from Hanten. I am not quite sure I understand exactly, but she asks, can you modify this to display IQIs by hospital locations just for display purposes? I know you said you are not using IQIs with this tool right now.

Melanie Chansky: Can you repeat that one more time?

Margie Shofer: She asks, "Can you modify the mapping tool to display IQIs by hospital locations?" And then she wrote in parentheses, "just for display purposes."

Melanie Chansky: Right now, no. However, if I am interpreting the question correctly, I think that is one of the things we want the tool to be able to do in the future. So look for something that incorporates the hospital locations in the future. Right now, it does not have the capability.

Margie Shofer: Looks like I have another question from Diane, who wants to know how the software handles missing data in the file.

Melanie Chansky: That is a good question. It has been a while since I tested that. Off the top of my head, I don't remember exactly how it is handled. I do not want to say that it handles it well and then be wrong about that. That is something I would suggest you ask offline so I can test it to see what happens. We did all of this when we were testing the tool prior to release, but I have not done it recently enough to remember how we finally addressed that. I apologize.

Margie Shofer: I think we are about out of time for questions. I want to thank you all for your thoughtful questions and your participation in today's event. We hope the discussion was helpful to you. If you have any questions about follow-on technical assistance opportunities, please do not hesitate to contact me, Margie Shofer. As you see, my contact information is on the last slide.

If you have any questions or comments about the tool, or would like to request a copy of the mapping tool, please send an e-mail, listed in the second bullet. For more information about the suite of tools we developed for this project, you can use the link listed on the last bullet. Again, Melanie did ask if you have any thoughts or suggestions about ways to improve the tool, so feel free to e-mail those to the e-mail listed in the second bullet. Thanks again, and this concludes our Webinar. We look forward to hearing from you.

Jeff Gallagher: Ladies and gentlemen, thank you so much for taking time to join us today. You now may disconnect. Please stop closed captioning now.

Return to Contents

The information on this page is archived and provided for reference purposes only.


AHRQ Advancing Excellence in Health Care