By Matthew Hile on 11/21/2007 2:30 PM

Again from the Robert Wood Johnson Advancing Recovery conference, David Gustafson did a audio presentation (as he could not attend the meeting in person)  focused on the problems of hand-off, discussing how poor our systems are moving an individual from one level of care to another. In statistics presented earlier in the morning Todd Molfenter reported that out of 10,000 individuals referred for substance abuse treatment, only 10 (no I did not drop any zeros) make it through the end of follow-up care. Yes treatment works but the system leaks like a sieve. 

 

David's presentation discussed a wide variety of hand-off lessons learned from football and medicine. He offered clear, usable, and proven techniques for fixing this problem.  (Indeed the Advancing Recovery project in Kentucky reported reducing their no show rate of individuals moving from a hospital program to out-patient from 70% to a rate ranging between 30 to 0%.)

 

One thing stuck a familiar cord. David emphatically stated that hand off failures are never the fault of the human being. They are always the fault of the system .  What struck me about this was its similarity to a maxim I developed a number of years ago (Hile's modest maxim and corollary - after all if you are going to name a maxim after yourself you should do so modestly.)

 

Information will be accurate only if it is useful to the person collecting it.

If people resist collecting information we have not made it useful.

 

It is clear that unless we carefully develop systems that support their users to move from place to place, to get accomplished what they need to accomplish, our systems will not be successful and that we will loose people who we had hoped to help.

By Matthew Hile on 11/12/2007 7:39 AM

I am attending a conference on the Robert Wood Johnson Advancing Recovery project as part of Missouri's grant team (www.MIMH.edu/AdvancingRecovery-MO). During that meeting Victor Capoccia of NIATx observed that
 

there are no consistent descriptions of treatment or providers for substance abuse services.

His observations were based on addiction treatment but really the problem is also true for mental health and medical related treatment.

Why is consistent important?

o    To allow for consumer choice there must be a way to compare one set of treatments/services with another.

o    To aggregate information across a community this information needs to be available and collectable.

We have done limited approaches to do this - basically a centralized repository of information about treatment programs for mental health, substance abuse, and MR/DD services across a circumscribed region (SAMHI.MIMH.edu). There are problems with this approach. Most importantly it requires a centralized system to collect, organize, and present this information on an ongoing basis. This is expensive and time consuming.

What about instead of a centralized group being responsible for this would be to push this out to the treatment providers themselves. What would this look like

o    An XML standard program description format.

o    A XSLT to allow this XML form to be presented in a consumer friendly format.

o    A search engine could aggregate these descriptions that would automate the ability to compare and contrast services and to match consumer needs to treatment provider capabilities. This search engine could go to the web sites or could subscribe to this information via RSS feeds. This makes provider updated information automatically and continuously aggregatable. 

o    Case managers could subscribe as well so that they can automatically keep up with revised information from the providers in their area or that they commonly use. (For more on the underlying concepts of what is called syndication-oriented architecture see this post by Jon Udell.)

There are clear advantages to consumers for such a system. This as evidenced by the various popular comparison shopping sites for consumer goods. For the providers this would help them promote their services more broadly, and serve the dual purpose of providing information on their site AND in a common aggregation location.

What do you think? Where might this go?

By Matthew Hile on 6/22/2007 5:11 AM

We have all come to rely on sending and receiving information via email. However, working with protected health information it is important to realize that it is neither appropriate, nor HIPPA compliant, to send that data without protecting it so that only the intended recipient can access its content.

The State of Missouri - Department of Mental Health has implemented a new secure email messaging system to comply with their operation regulations (DOR 8.340). Since the various state departments share a consolidated email system through the Office of Administration I assume that other departments also use this approach. Today I got my first secure message and thought it would be worth describing the experience.

Initial registration

The first thing that happens is that you get an official looking email message with a note directing you to click the attachment (an html link) to view your message. This gave me pause. For security reasons you should never click on an attachment that you were not expecting. Rather than calling the sender I opened the header information (in Outlook’s Option section) to assure myself that it was from the department. (It would have been better to call the sender but I did not expect him to be in the office at 5am). This link lead me to a secure web page where my browser warned me that it could not verify the security certificate that was being used. I checked out the certificate and could not tell why it would not verify but it looked OK and, liking to live dangerously, I decided to accept it for a single use. This opened the web page in my browser and informed me that I would now be able to receive secure messages.

The technology for this solution is provided by Proofpoint and uses a PKI (Public Key Infrastructure) identity-based encryption. The major advantage of this is that a user does not need to download/install any software nor do they need to have any skills other than clicking on various links and buttons.

Receiving secure messages

Once you have been verified it is easy to read and respond to any secure messages that you receive. A message will arrive telling you to open the attachment. In Outlook you will be warned that it is a potentially unsafe attachment and told to open it only if it is from a trustworthy source. Saying OK at that point will start your browser and open a secure web page (this time the browser was able to verify the certificate) with a button to click to read the message.

The email message interface is simple and quite easy to use giving you options to reply and forward the message in a secure manner. Files can also be attached and they will also be secured for transmission.

Usability

Positives: From an end user perspective this is the easiest PKI implementation I have ever seen. There is no software to load, no user registration, no passwords to remember, really nothing required from the user other than a few links to click. Given the usual complexities of PKI this is a great leap forward. Secure email can be sent to people on different email systems who can reply and forward securely as well.

Negatives: 1) Users’ email systems must accept attachments to get to the secure site. There are some corporate environments that strip out attachments to increase their level of security. 2) Because it takes at least three clicks to read an email there is an added level of frustration/irritation. 3) The user has to use a different interface to enter email and does not have the features that they take for granted (e.g., checking email addresses, spell checking). 4) Unless they go through the multi-step process of copying and saving the content in some other form, email messages are not searchable or quickly available. Say I get a series of secure messages with the same title (e.g., “RE: [encrypt] This is how to encrypt email”). To find a particular message in that set I need to click on each message three times to read it. That would get old fast. 5) While it is easy to reply and forward a secure message it is difficult for a user from outside of the system to initially create a secure message to send to the Department. As far as I can tell to do this you would open a previously received secure message, click reply, delete the current information (e.g., to, subject, message) then send.

Final thoughts

There are a lot of usability issues with this system. However, its ability to use PKI to send secure emails without the traditional difficulties associated with that infrastructure make it a useful approach to solving the problem of securing protected health data as it is sent over the internet.

By Matthew Hile on 6/7/2007 1:43 PM

Recently I received a set of questions about one of the web sites we have created to support a Robert Woods Johnson funded project to implement a mental health system change process to implement evidence based practices in substance abuse treatment providers in Missouri. Because these were “lessons learned” questions I thought others might find these responses useful.

Q - Is the entire website using DotNetNuke? Are you doing it alone?

R – The site, supporting Missouri’s Advancing Recovery project, was built entirely with DotNetNuke (DNN) an open source web application framework based on Microsoft’s .NET platform. It uses many of the standard modules (e.g., Photo Gallery, HTML pages, Documents) and a couple of independently create ones (viz., Ktomics WWWiki, DNNstuff SQLView).

Q - How much time does it take you to initially develop the website?

R – Using DNN the initial site creation is quite easy to accomplish. Since the Institute has been using it for a number of sites (including our home site MIMH.EDU) once we have the content we can get a basic site up and onto the web in a day. My general approach is to get a site posted quickly then continue to add information, features, and organizational structures as time goes on.

Q - How much time does it take you to maintain the website?

R - Maintenance includes a variety of tasks. Because much of the site is private (available only to the project’s participants) I need to evaluate all new registration requests and decide it they are to have access. Actually approving them takes only a moment but often I need to check with others to see if they are valid participants. Also, I have gotten a number of requests from folks who are not part of our specific project (e.g., evaluators and RWJ staff). To accommodate their legitimate requests to view the site I needed to modify our initial security model granting these users “read only access” to many of the private pages and denying them access to one (viz., the contact information of all of the participants). This fine grained security tailoring is supported by DNN but it a couple of days to conceptualize, setup, and test.

Another maintenance task is to keep the base DNN software up-to-date. Because we support a number of sites, using modules developed by others and custom modules created by our staff it is a time consuming task to make major changes and to test all of the applications. This is not unique to DNN but is something that needs to occur with all software upgrades.

Q - How much time does it take you to update the website?

R - Updating the site refers to adding new content. Again, because of the DNN framework the mechanics of this are relatively simple. I receive, find, or create the content, decide where it should be placed, format it appropriately and add it to the site. For some information, like a RWJ project newsletter, this takes 10 minutes. Creating other materials take longer depending on what needs to occur. For example, there was a recent article about medication assisted treatments for substance abusers in the St. Louis paper that I wanted to site. Because I was afraid that the link to the original article could disappear I wanted a copy of the article on one of our pages. It took about an hour from start to final formatted and linked web page.

Another important feature of updating the site is to let user know what has changed. Since our user base is not sophisticated what I do is to send them a weekly email newsletter (through the DNN application) that lists what has been added and changed. Obviously the actual time I spend doing this varies based on how much material is added. It has ranged from 1-16 hours a week but is most often only a few hours.

Q - Is anyone else in addition to you adding information to the wiki or the forum or any other tab? Are any providers adding directly?

R – DNN supports participant contributions in many of its modules. In my mind that is one of the most compelling aspects of the application. However, our site participants are not accustomed to adding information to a web site and, to date, I have been unsuccessful in getting that to happen. What I have done it to continually solicit information from the participants. They email me the information and I put it onto the site. This has not become habitual on that part – I still need to ask people to send me this or that for the site – but when asked the information is readily provided.

Q - Has the wiki caught on/been used as much as you would like? Why or why not?

R – I am a huge fan of wiki supported collaborative writing. Basically a wiki allows any authorized user to easily edit current pages and to create new pages. However, for the reasons described above, our users are not actively using this to share their information.

The other problem with a wiki is organization. While it is easy to create additional pages it is difficult to keep up with an organically growing structure. And, because users need to have mental models about a site that they use, this makes the totally free form approach we adopted difficult is multiple individuals are making structural changes. Wikis that have an inherent structure (e.g., Wikipedia), or a commonly agreed upon structure, may be a better bet for this technology.

Q - I see you use both the wiki and a repository page to place documents. How do you decide which to use?

R – I have experimented with a number of ways to provide documents to our users and have yet to hit on one that I think really works. The basic issue is discoverability; users must be able to find what is on the site. Both the wiki and repository are searchable so that users can enter appropriate search terms and find pages containing that information. However, with the repository users have the additional advantage of going to a single page to scan the available documents which have names, categories, and descriptions. As time has gone on I have gravitated toward adding files to the repository pages.

Q - Has the forum caught on? Why or why not?

R – The forums have not. I have added a few questions and a high percentage of our users have viewed those posts but there have been essentially no comments or discussion. I think there are at least two reasons. First, we have not created a culture within our project for using this as a way to ask questions and receive support. Second, as mentioned above, I think that providing contributions on a web site is a new concept that has yet to attract our users. It might be possible to jump start this process by getting a few participants to use this area extensively and by example demonstrate to others its value. To date, we have not done this.

Q - Can you tell who is accessing the website and how they are using it? What do you see?

R - For the private sections of the site we have a total of 27 users. In the last two months 18 of those folks accessed the site. The most frequently visited pages are our wiki (304 requests), news page (144 requests), forum (109 requests), and medication assisted treatment documents (105 requests). For our publicly available pages we had the most hits on the home page (380 requests) and consumer/family information page (105 requests).

Q - What do you think is the primary value of the website to providers? To the state?

R - For both I think it is the one shared place where all of the project materials are available. Additionally, it provides a public face for the project which is often not done for many grant efforts.

Q - The providers identify the website as a strength of the project. What are three of the reasons?

R - The providers have liked having one place where all of the relevant information is stored and shared, the weekly updates on what is new, and the information created specifically for consumers and their families. They may be able to give additional reasons but these are things that I have heard.

Q - Is there anything you would do differently, either at the start or now?

R – Work more aggressively to develop a sense of a web based shared community. The participants seem to have a strong bond with each other in this project learning and sharing freely of their knowledge and experience. But that has not been successfully translated to their direct contributions to the web site.

By Matthew Hile on 5/8/2007 9:25 AM

Amazon allows individuals to write public comments and reviews for all of the products they sell and the businesses use Amazon to sell their own merchandise. I, like most folks, religiously read these reviews and use them to inform my purchasing decisions. When I have a particularly strong opinion, I will even contribute a review of my own.

For some time I have been trying to quietly promote the idea that users of mental health and substance abuse services should have a similar ability. People who use these services should have an open forum that allows them to publically tell the world what they liked – or did not like – about a particular provider or service. Ideally, like Amazon, these comments should be available within the context of a search for a specific treatment service as in MIMH’s  Substance Abuse and Mental Health Information Online system or SAMHSA’s Substance abuse treatment facility locator. For the consumer the advantages of this are easy to see. They could select a provider based on the shared experiences of others rather than simply figuring out which provider is close to home. Like asking your friends to recommend a restaurant, you would benefit from others’ experiences.

Consumer concerns

What are the down sides of this? From the consumer perspective I think the major difficult would be authority – that is “can I trust this information.”

Authority is present if you know the individual commenter and trust their opinions. For the most part on the web it is unlikely. However, we can look at the individual’s other comments – a history of their behavior – and get a sense of whether or not we trust them. Amazon supports this sort of authority by letting you look at all an individual’s comments. If they all say “this is the best thing since sliced bread” or “this is the worst product ever foisted on the human race” you have little trust in their ability to discriminate between a good and bad product. If however, you read a set of balanced comments, some positive some negative, your level of trust in their opinions will increase. For the current topic, comments relating to mental health and substance abuse providers, I suspect that there would be relatively few comments that cross a range of provider/service combinations that authority based on user history would be relatively weak.

Authority can also be demonstrated if there are a lot of comments about a provider/service. Here the consumer perceives the general direction of most of the comments getting a sense of authority from the convergence in this wider community participation. Rather than trusting the authority of an individual commenter they trust the authority of the community. The more comments, the more authority.

Provider concerns

Providers, on the other hand, have different concerns. For example, “What if someone who was never in our care trashes our services?” Who would do this? Perhaps a unscrupulous competitor hoping to drive individuals to their own services. Could this happen – yup. Would it be a problem – well that would depend but at least two factors mitigate the potential deleterious effects of such an attack.

First, if it were only one of many comments AND if it contrasted with the trends of those comments it is unlikely that anyone would give it much authority. That is this negative voice would be overshadowed by the more numerous positive and balanced comments. Even if there was only one comment, this negative one, as described above it would have little authority on its own and therefore little impact.

Second, perhaps a more useful response toward this sort of attack, as well as the legitimate negative comments of a dissatisfied consumer, is the notion of transparency. We all know that every business and every service will have unhappy users. In the past business have tried to hide this fact and their errors but the new radical transparency movement has turned that notion on its head. Rather than hiding negatives business talk openly about them  – perhaps apologizing for a mistake or talking about practices that address, resolve, or prevent the issue. As a consumer I can learn more useful information by seeing how a problem is addressed then from any other individual comment. Have a problem – that is to be expected – deal with it in a forthright, reasonable, and ethical manner – then you are someone with whom I will do business.

So yes there are negatives in allowing individual to comment on their experiences with mental health and substance abuse providers (or really any business). However, these negatives cannot only be managed but can serve to enhance the provider in the eyes of the consumer.

By Matthew Hile on 4/16/2007 7:41 AM

I was listening to the .NET Rocks podcast over the weekend and heard two shows relevant to DNN developers. The first was a discussion with Dan Appleman who briefly mentioned using DNN and his thoughts that with this sort of framework available folks would be foolish to code their own sites (unless there was a really good reason to do so).

The second podcast focused entirely on DNN and was a second interview with Shaun Walker the head of the DotNetNuke project.

Nice to hear DNN discussed on this popular show.

By Matthew Hile on 4/16/2007 6:18 AM

http://searchdotnet.com/

This custom Google search site was created by Dan Appleman, a well known and highly respected developer, to solve some of his own difficulties in finding high quality .NET code and answers (.NET Rocks podcast where he talks about this site). I used it this morning and it returned the authoritative results I needed without a lot of the extra stuff. If you do .NET development this is a great resource!

Note that to make it really handy you can easily add it to your list of search providers in Firefox or IE .

By Matthew Hile on 4/13/2007 10:01 AM

There has been a lot of discussion about the relative merits of Google's and Microsoft's saved maps. As this erupted I had just completed a saved Google Map for Missouri's Advancing Recovery project and thought that I would try the same thing on Microsoft's Live Maps.

The purpose of my map was to show the locations of all of the participants in Missouri's Advancing Recovery project. To do this I wanted to have the state highlighted and marks for the location of each of the participants. I had first thought to do this with an image in a paint program but thought a more interactive experience would be better. I had never used either of these sites to created a custom map so this is the review of a neophyte map maker.

Step 1: Highlight Missouri

Google

Microsoft

No obvious way to start this process. However, being an inveterate right-clicker I right clicked on the map and got the context sensitive menu. A started by Drawing a Shape. At click I placed a small square box and as I went along a helpful message "Double click to end this shape" followed my cursor. As a nice touch, after completing the shape I could click on any of the boxes and move them to a different location or right click to get to another context sensitive men to delete or add points. I was able to complete this step in one go.

Here the options were displayed on the screen and selecting the mark and area option allowed me to start clicking on the Missouri's border to create the highlighted shape. To finish a drawing you need to right click on the last location and select the "Finish drawing" option. Since my last click had been on top of my first one I moved off of that location before selecting the finish option. This added a final mark which ended up loosing Missouri a large v shaped chunk of territory to Iowa. It took me three tries to mark the map to display Missouri to my satisfaction. There is no apparent way to edit a shape once it is created or even to undo the last clicked location.

Advantage Google. The ability to edit a shape AFTER it has been initially created is clearly something that makes the job easier. It is possible that Microsoft has this feature but I was unable to discover it. However, neither system was able to automatically highlight the state. It would seem like a logical feature to be able to automatically highlight some geographic feature (e.g., city, state, country, ...).

Step 2: Add locations

Google

Microsoft

Search for a location on the map the information bubble has a simple like "Save to My Maps." You can then edit title, location add links, and so on. Simple and easy to understand. In my search Google had difficulty locating one location.

Search for a location and a push pin is automatically added to the map. A left click on the pin then selecting "edit" provided this box. For my needs it required more work to use as I had to move the address into the Notes section then add the agency as a title. Also, while I did not need this, unlike Google you can not change the push pin image. In my searching the Microsoft site had difficulty locating two locations.

Both approaches worked well but I found Google's to be a bit easier and more flexible. However, because it allows you to enter html code there may be the possibility for cross-site scripting vulnerabilities.

Step 3: Final touches

Google	Microsoft
Once the map is created it can be easily shared with a url. One difficulty I had was the the items were listed in the order in which they were entered and there was no way, other than deleting and re entering a location which I did, to change that order. Than final map in maps.google.	The order of items was easily adjusted by simply dragging them to a new location. The final map maps.live.

Closing thoughts

As a novice map maker I need to give Google the advantage. It was easier for me to use and had sufficient flexibility to meet my simple needs. In particular the ability to adjust the shape after it was created is a significant advantage if you are doing complex shapes, like the state of Missouri.

By Matthew Hile on 4/10/2007 4:09 AM

I am a lazy programmer - If I can figure out a way for the computer to do something for me I am all over it. Recently I was working on a DNN module and noticed that my class definitions of various objects were remarkably similar. I cut and pasted from one into another. For example these two info classes contain the same PortalID and ModuleID properties.

Public Class EventInfo         Private _PortalId As Integer         Private _ModuleId As Integer         Private _EventId As Integer         Public Property PortalId() As Integer             ...         End Property         Public Property ModuleId() As Integer             ...         End Property         Public Property EventId () As Integer             ...         End Property    End Class

Public Class UserReminderInfo         Private _PortalId As Integer         Private _ModuleId As Integer         Private _UserReminderId As Integer         Public Property PortalId() As Integer             ...         End Property         Public Property ModuleId() As Integer             ...         End Property         Public Property UserReminderId () As Integer             ...         End Property    End Class

It irritated me that I needed to copy and past the duplicate code. I made errors not copying enough of the code. It also worried me that if I changed it in one place I would need to change it in others.

So enter Object Oriented Programming (OOP) and inheritance in the form of .NET Abstract classes. To set this up I created a new base class that would be shared with these (and other classes). Using the MustInherit key word and declaring the common properties.

Public MustInherit Class InfoBase         Private _PortalId As Integer         Private _ModuleId As Integer        Public Property PortalId() As Integer             ...         End Property         Public Property ModuleId() As Integer             ...         End Property     End Class

Now I can simplify my other two classes by having them Inherit the InfoBase information as show below.

Public Class EventInfo         Inherits InfoBase         Private _EventId As Integer         Public Property EventId() As Integer             ...         End Property    End Class

Public Class UserReminderInfo         Inherits InfoBase         Private _UserReminderId As Integer         Public Property UserReminderId() As Integer             ...         End Property    End Class

I save time by not having to retype or cut-and-paste and reduce errors that may creep in if I need to go back and redefine any of the shared properties. A win-win. For more information on abstract classes check out http://www.devx.com/dotnet/Article/28086/1954 or http://www.startvbdotnet.com/oop/abstract.aspx.

By Matthew Hile on 3/28/2007 11:05 AM

In my tech reading I found a description of a brief video that both entertains and describes what XML is and why the Web 2.0 is important. Titled Web 2.0 ... The Machine is Us/ing Us this video by Michael Wesch, a cultural anthropologist from Kansas State University, is five minutes worth spending.

One thing I found ironic. I read about this video while reading David Margulius' column in the second to last print issue of InfoWorld, a tech magazine that I have been reading and learning from for years. Next month they are relying exclusively on their internet based publishing in a way that relies on and use the  technologies Michael so cleverly describes.