Re: Open source tools for population health epidemiology and public health
On Thu, 2004-12-23 at 23:02, David Forslund wrote: I know a number of folks who would be interested, but the inability to run the software on Win platforms removes them from consideration at this time. Dave Maybe this will be enough of a trigger to get them to try out some linux / unix servers then. :-) Surely they aren't that closed minded? Just think of it...the possibility of getting some great software free and all you have to do to try it out is come up with an old x86 box that you probably have lying around. I applaud their efforts in not spending resources supporting legacy operating systems such as Windows. ;-) Cheers, Tim -- Sincerely, Tim Cook Key ID 9ACDB673 @ http://www.keyserver.net/en/ signature.asc Description: This is a digitally signed message part
Re: Open source tools for population health epidemiology and public health
Original Message From: Tim Cook [EMAIL PROTECTED] To: OpenHealth List openhealth-list@minoru-development.com Cc: Tim Churches [EMAIL PROTECTED] Date: Fri, Dec-24-2004 6:35 AM Subject: Re: Open source tools for population health epidemiology and public health On Thu, 2004-12-23 at 23:02, David Forslund wrote: I know a number of folks who would be interested, but the inability to run the software on Win platforms removes them from consideration at this time. Dave Maybe this will be enough of a trigger to get them to try out some linux / unix servers then. :-) Surely they aren't that closed minded? It isn't an issue of their closed minds. It is an issue of the IT department of the State setting requirements for systems. As you know I'm all for open source but don't think one should be limited by the OS platform one is running on. There are enough battles to be fought with standards and interoperability. It is fairly easy to make the OS disappear. Dave Just think of it...the possibility of getting some great software free and all you have to do to try it out is come up with an old x86 box that you probably have lying around. I applaud their efforts in not spending resources supporting legacy operating systems such as Windows. ;-) Cheers, Tim -- Sincerely, Tim Cook Key ID 9ACDB673 @ http://www.keyserver.net/en/
Re: Open source tools for population health epidemiology and public health
David Forslund wrote: I know a number of folks who would be interested, but the inability to run the software on Win platforms removes them from consideration at this time. All of the tools and infrastructure used are cross-platform, with the exception of PostgreSQL - but that will soon be also with the imminent release of Version 8.0, which will run natively under MS-Windows as well. It's just that MS-Windows is not our primary development platform, and we haven't had time to recompile, test and package various components for it - but we will, in due course (or someone else might volunteer to take on that task). Tim C
Re: Open source tools for population health epidemiology and public health
Tim Cook wrote: On Thu, 2004-12-23 at 23:02, David Forslund wrote: I know a number of folks who would be interested, but the inability to run the software on Win platforms removes them from consideration at this time. Dave Maybe this will be enough of a trigger to get them to try out some linux / unix servers then. :-) Surely they aren't that closed minded? Just think of it...the possibility of getting some great software free and all you have to do to try it out is come up with an old x86 box that you probably have lying around. I applaud their efforts in not spending resources supporting legacy operating systems such as Windows. ;-) We certainly recommend Linux or Unix as a server platform, but we do want the tools to be able to run on Windows laptops for field and mobile use. Its just a matter of POSIX (both Linux and Mac OS X) being our development platform, and thus ports to Windows will always lag some time behind. Tim C
Re: Open source tools for population health epidemiology and public health
This sounds reasonable and certainly is, but there are some more complications. I try to be database independent, too, letting the deployment of a particular database to be site specific. The problem I also ran into at our state is the required use of MSSql on an Windows platform. Something like PostgresSQL was beyond their willingness to move. Also, the system would have to meet CDC reporting requirements, for which much of their software is used, not because they find it useful, but because it is required to meet their reporting requirements. They will be using NEDSS because it is supported by CDC and reduces their exposure to responsiblity of software risk. Thus an open source solution also needs to be NEDSS-compliant or PHIN compatible in the US these days. Something better isn't the only driving force these days. Dave Original Message From: Tim Churches [EMAIL PROTECTED] To: openhealth-list@minoru-development.com Date: Fri, Dec-24-2004 8:01 AM Subject: Re: Open source tools for population health epidemiology and public health David Forslund wrote: I know a number of folks who would be interested, but the inability to run the software on Win platforms removes them from consideration at this time. All of the tools and infrastructure used are cross-platform, with the exception of PostgreSQL - but that will soon be also with the imminent release of Version 8.0, which will run natively under MS-Windows as well. It's just that MS-Windows is not our primary development platform, and we haven't had time to recompile, test and package various components for it - but we will, in due course (or someone else might volunteer to take on that task). Tim C
Re: Open source tools for population health epidemiology and public health
Side note re limited to non-Windows by postgresql. I've often run cygwin postgresql and, while it takes a few minutes to setup, has run fine for me. I've read elsewhere of people having stability problems with that env but have not encountered any on my end. I don't know if the stability problems with cygin/postgres have been resolved or if my db apps just haven't stressed postgres enough to hit a cygin/postgres instability. I ack that the native flavor postgres should be much faster than cygin indirection flavor, but whether that's a problem or not is dependent on amount of data and desired response time. Just a note. And, yes, native postgres 8.x should be out shortly. Cheers, Heitzso
Re: Open source tools for population health epidemiology and public health
David Forslund wrote: This sounds reasonable and certainly is, but there are some more complications. I try to be database independent, too, letting the deployment of a particular database to be site specific. The problem I also ran into at our state is the required use of MSSql on an Windows platform. Something like PostgresSQL was beyond their willingness to move. Sure, but again, there is nothing fundamental which ties NetEpi Case Manager to PostgreSQL - it was just that PostgreSQL was (and is) our database server of choice for development purposes. We use the Python DB-API 2.0, which abstracts away most of the database server-specific characteristics, but of course, every SQL database has its own set of extensions, and there are a few Postgresisms which we have used out of expediency (we were under considerable self-imposed pressure to get something working in the midst of the SARS outbreak of 2003)- this use of Postgresisms also needs to be abstracted out in order to support a wider range of back-ends (including embedded databases such as SQLite) - that is a task we plan to do, but again it is not the highest priority (of course if someone wants to sponsor the work or do it themselves, the priority would be changed). Also, the system would have to meet CDC reporting requirements, for which much of their software is used, not because they find it useful, but because it is required to meet their reporting requirements. Yes, if use of the tools in the US is a goal, then meeting CDC requirements is a necessity. However, in developing the tools, the US did not even enter our thoughts - simply because so much money has been poured into public health informatics over the last two or three years in the US that it doesn't need any help from us. Frankly, when we started developing NetEpi Case Manager, the data collection tool, in early 2003, we had Australian needs primarily in mind, but were also mindful of its potential utility in developing and transitional countries, especially those to our north and north-west (where, incidentally, Linux and other open source infrastructure tends to have a much greater mindshare within government departments and NGOs than it does in richer countries, although MS Windows is still probably the dominant OS - and hence does need to be supported by the NetEpi tools, in due course). They will be using NEDSS because it is supported by CDC and reduces their exposure to responsiblity of software risk. Thus an open source solution also needs to be NEDSS-compliant or PHIN compatible in the US these days. Yes, I agree. However, we currently have no plans to make the NetEpi tools NEDSS or PHINS compatible, primarily because the US is already well-provided for in publc health informatics. We are more interested in maximising the utility of the NetEpi tools for use in Australia and other parts of the world. That said, we have borrowed (and will continue to borrow) ideas from NEDSS and PHINS where they are valuable and not overly complex (the NEDSS data model in particular tends towards the baroque, in my opinion, but it is trying to address a very wide range of problems, far more ambitious in scope than what teh NetEpi tools hope to address) - and of course not encumbered by patents, but as far as I know, none of the material in the NEDSS or PHINS design documents is patented or likely to be patented. Something better isn't the only driving force these days. Again, I agree entirely. But we don't claim that the NetEpi tools are better than anything else - we only strive to make them as useful as possible to the target audience, which is public health practictioners in countries which do not already have well-funded and strongly organised public health informatics programmes (like most developed countries, there is some funding for public health informatics in Australia, but compared to the US on a per-capita basis, it is still an order of magnitude less, and on an absolute basis, absolutely miniscule - the same is true of most OECD countries, I think, with teh exception perhaps of Canada in the last 18 months, for obvious reasons). Nor do we hope that the NetEpi tools drive out competitors - if they do no more than catalyse other open source development projects in public health informatics, then they will have served their purpose (although we do of course hope that they are used). Cheers, Tim C Original Message From: Tim Churches [EMAIL PROTECTED] To: openhealth-list@minoru-development.com Date: Fri, Dec-24-2004 8:01 AM Subject: Re: Open source tools for population health epidemiology and public health David Forslund wrote: I know a number of folks who would be interested, but the inability to run the software on Win platforms removes them from consideration at this time. All of the tools and infrastructure used are cross-platform, with the exception of PostgreSQL - but that will soon be also
Re: Open source tools for population health epidemiology and public health
I have no problem with your comments, with one exception. The state of IT in public health in the US, despite the efforts of the CDC, NEDSS, and PHIN is pretty abysmal. I can't compare it to Australia, but on any scale, they are in the dark ages here in the US and need all the help they can get. It certainly isn't your job to do so, but it would be nice to provide them better capability (and more efforts and true interoperability). They keep asking for better case management tools, and this isn't really in the purview of the CDC. Thanks, Dave Original Message From: Tim Churches [EMAIL PROTECTED] To: David Forslund [EMAIL PROTECTED], Andrew McNamara [EMAIL PROTECTED] Cc: openhealth-list@minoru-development.com Date: Fri, Dec-24-2004 1:40 PM Subject: Re: Open source tools for population health epidemiology and public health David Forslund wrote: This sounds reasonable and certainly is, but there are some more complications. I try to be database independent, too, letting the deployment of a particular database to be site specific. The problem I also ran into at our state is the required use of MSSql on an Windows platform. Something like PostgresSQL was beyond their willingness to move. Sure, but again, there is nothing fundamental which ties NetEpi Case Manager to PostgreSQL - it was just that PostgreSQL was (and is) our database server of choice for development purposes. We use the Python DB-API 2.0, which abstracts away most of the database server-specific characteristics, but of course, every SQL database has its own set of extensions, and there are a few Postgresisms which we have used out of expediency (we were under considerable self-imposed pressure to get something working in the midst of the SARS outbreak of 2003)- this use of Postgresisms also needs to be abstracted out in order to support a wider range of back-ends (including embedded databases such as SQLite) - that is a task we plan to do, but again it is not the highest priority (of course if someone wants to sponsor the work or do it themselves, the priority would be changed). Also, the system would have to meet CDC reporting requirements, for which much of their software is used, not because they find it useful, but because it is required to meet their reporting requirements. Yes, if use of the tools in the US is a goal, then meeting CDC requirements is a necessity. However, in developing the tools, the US did not even enter our thoughts - simply because so much money has been poured into public health informatics over the last two or three years in the US that it doesn't need any help from us. Frankly, when we started developing NetEpi Case Manager, the data collection tool, in early 2003, we had Australian needs primarily in mind, but were also mindful of its potential utility in developing and transitional countries, especially those to our north and north-west (where, incidentally, Linux and other open source infrastructure tends to have a much greater mindshare within government departments and NGOs than it does in richer countries, although MS Windows is still probably the dominant OS - and hence does need to be supported by the NetEpi tools, in due course). They will be using NEDSS because it is supported by CDC and reduces their exposure to responsiblity of software risk. Thus an open source solution also needs to be NEDSS-compliant or PHIN compatible in the US these days. Yes, I agree. However, we currently have no plans to make the NetEpi tools NEDSS or PHINS compatible, primarily because the US is already well-provided for in publc health informatics. We are more interested in maximising the utility of the NetEpi tools for use in Australia and other parts of the world. That said, we have borrowed (and will continue to borrow) ideas from NEDSS and PHINS where they are valuable and not overly complex (the NEDSS data model in particular tends towards the baroque, in my opinion, but it is trying to address a very wide range of problems, far more ambitious in scope than what teh NetEpi tools hope to address) - and of course not encumbered by patents, but as far as I know, none of the material in the NEDSS or PHINS design documents is patented or likely to be patented. Something better isn't the only driving force these days. Again, I agree entirely. But we don't claim that the NetEpi tools are better than anything else - we only strive to make them as useful as possible to the target audience, which is public health practictioners in countries which do not already have well-funded and strongly organised public health informatics programmes (like most developed countries, there is some funding for public health informatics in Australia, but compared to the US on a per-capita basis, it is still an order of magnitude less
Re: Open source tools for population health epidemiology and public health
David Forslund wrote: I have no problem with your comments, with one exception. The state of IT in public health in the US, despite the efforts of the CDC, NEDSS, and PHIN is pretty abysmal. I can't compare it to Australia, but on any scale, they are in the dark ages here in the US and need all the help they can get. I suspect that reflects decades of very low levels of investment in public health infrastructure (especially personnel) in the US prior to 2001. No matter how much money is thrown at the problem, it just takes time to build infrastructure, especially organisational infrastructure and trained, experienced personnel. It certainly isn't your job to do so, but it would be nice to provide them better capability (and more efforts and true interoperability). They keep asking for better case management tools, and this isn't really in the purview of the CDC. CDC has something called the Outbreak Management System (OMS), I understand, which I think has recently been re-written in Java? Not really sure - very hard for people outside US public health circles to find out what is going on inside CDC, beyond is published on the CDC web site, and all our email enquiries go unanswered. CDC is so big that I don't think people inside teh organisation know what everyone else is working on, either, particularly in teh last few years with all the funds flowing in for public health informatics. Tim C Original Message From: Tim Churches [EMAIL PROTECTED] To: David Forslund [EMAIL PROTECTED], Andrew McNamara [EMAIL PROTECTED] Cc: openhealth-list@minoru-development.com Date: Fri, Dec-24-2004 1:40 PM Subject: Re: Open source tools for population health epidemiology and public health David Forslund wrote: This sounds reasonable and certainly is, but there are some more complications. I try to be database independent, too, letting the deployment of a particular database to be site specific. The problem I also ran into at our state is the required use of MSSql on an Windows platform. Something like PostgresSQL was beyond their willingness to move. Sure, but again, there is nothing fundamental which ties NetEpi Case Manager to PostgreSQL - it was just that PostgreSQL was (and is) our database server of choice for development purposes. We use the Python DB-API 2.0, which abstracts away most of the database server-specific characteristics, but of course, every SQL database has its own set of extensions, and there are a few Postgresisms which we have used out of expediency (we were under considerable self-imposed pressure to get something working in the midst of the SARS outbreak of 2003)- this use of Postgresisms also needs to be abstracted out in order to support a wider range of back-ends (including embedded databases such as SQLite) - that is a task we plan to do, but again it is not the highest priority (of course if someone wants to sponsor the work or do it themselves, the priority would be changed). Also, the system would have to meet CDC reporting requirements, for which much of their software is used, not because they find it useful, but because it is required to meet their reporting requirements. Yes, if use of the tools in the US is a goal, then meeting CDC requirements is a necessity. However, in developing the tools, the US did not even enter our thoughts - simply because so much money has been poured into public health informatics over the last two or three years in the US that it doesn't need any help from us. Frankly, when we started developing NetEpi Case Manager, the data collection tool, in early 2003, we had Australian needs primarily in mind, but were also mindful of its potential utility in developing and transitional countries, especially those to our north and north-west (where, incidentally, Linux and other open source infrastructure tends to have a much greater mindshare within government departments and NGOs than it does in richer countries, although MS Windows is still probably the dominant OS - and hence does need to be supported by the NetEpi tools, in due course). They will be using NEDSS because it is supported by CDC and reduces their exposure to responsiblity of software risk. Thus an open source solution also needs to be NEDSS-compliant or PHIN compatible in the US these days. Yes, I agree. However, we currently have no plans to make the NetEpi tools NEDSS or PHINS compatible, primarily because the US is already well-provided for in publc health informatics. We are more interested in maximising the utility of the NetEpi tools for use in Australia and other parts of the world. That said, we have borrowed (and will continue to borrow) ideas from NEDSS and PHINS where they are valuable and not overly complex (the NEDSS data model in particular tends towards the baroque, in my opinion, but it is trying to address a very wide range of problems, far more ambitious in scope than what teh
Re: Open source tools for population health epidemiology and public health
Hi all, Here is something from Herman Tolentino who is in CDC public health informatics: The problems you stated are all symptomatic of a cultural problem in public health in general. For a very long time, health informatics has been in the fringes of public health where information technology has traditionally been handled by epidemiologists. Public health informatics is relatively new, even if for example, the CDC public informatics fellowship program where I am in now is already six years old. Change is forthcoming but very slow. It's evolving though with the reorganization in CDC which is interesting to watch as it happens. Within CDC there are also institutional barriers to rapid change arising from being a federal agency that has to comply with numerous security and administrative legal requirements, which will make working with it and working within it quite a challenge now and in the near future. Tim is right, capacity building takes time particularly for public health, and anything we do in that direction greatly helps the effort. The most important thing is, I agree, to get things done and make useful applications. This is the context of systems development in CDC. Cheers, Herman D Tolentino MD Public Health Informatics Fellow Centers for Disease Control - Original Message From: openhealth-list@minoru-development.com To: David Forslund [EMAIL PROTECTED] Cc: openhealth-list@minoru-development.com Subject: Re: Open source tools for population health epidemiology and public health Date: Sat 12/25/04 07:36 AM David Forslund wrote: I have no problem with your comments, with one exception. The state of IT in public health in the US, despite the efforts of the CDC, NEDSS, and PHIN is pretty abysmal. I can't compare it to Australia, but on any scale, they are in the dark ages here in the US and need all the help they can get. I suspect that reflects decades of very low levels of investment in public health infrastructure (especially personnel) in the US prior to 2001. No matter how much money is thrown at the problem, it just takes time to build infrastructure, especially organisational infrastructure and trained, experienced personnel. It certainly isn't your job to do so, but it would be nice to provide them better capability (and more efforts and true interoperability). They keep asking for better case management tools, and this isn't really in the purview of the CDC. CDC has something called the Outbreak Management System (OMS), I understand, which I think has recently been re-written in Java? Not really sure - very hard for people outside US public health circles to find out what is going on inside CDC, beyond is published on the CDC web site, and all our email enquiries go unanswered. CDC is so big that I don't think people inside teh organisation know what everyone else is working on, either, particularly in teh last few years with all the funds flowing in for public health informatics. Tim C Original Message From: Tim Churches To: David Forslund , Andrew McNamara Cc: openhealth-list@minoru-development.com Date: Fri, Dec-24-2004 1:40 PM Subject: Re: Open source tools for population health epidemiology and public health David Forslund wrote: This sounds reasonable and certainly is, but there are some more complications. I try to be database independent, too, letting the deployment of a particular database to be site specific. The problem I also ran into at our state is the required use of MSSql on an Windows platform. Something like PostgresSQL was beyond their willingness to move. Sure, but again, there is nothing fundamental which ties NetEpi Case Manager to PostgreSQL - it was just that PostgreSQL was (and is) our database server of choice for development purposes. We use the Python DB-API 2.0, which abstracts away most of the database server-specific characteristics, but of course, every SQL database has its own set of extensions, and there are a few Postgresisms which we have used out of expediency (we were under considerable self-imposed pressure to get something working in the midst of the SARS outbreak of 2003)- this use of Postgresisms also needs to be abstracted out in order to support a wider range of back-ends (including embedded databases such as SQLite) - that is a task we plan to do, but again it is not the highest priority (of course if someone wants to sponsor the work or do it themselves, the priority would be changed). Also, the system would have to meet CDC reporting requirements, for which much of their software is used, not because they find it useful, but because it is required to meet their reporting requirements. Yes, if use of the tools in the US is a goal, then meeting CDC requirements is a necessity. However, in developing the tools, the US did not even enter our thoughts - simply because so much money
Open source tools for population health epidemiology and public health
I am pleased to announce that developmental versions of some tools for population health epidemiology and public health are now available under a free, open source software license - see http://www.netepi.org (please note that the release notes for the NetEpi Analysis tool can be found in the documentation section of the project web page on SourceForge). The current development team (comprising two members: Andrew McNamara and myself) will be working on NetEpi fairly intensively over the first half of 2005, with a view to a Version 1.0 release of the tools by mid-year. We would be very happy to hear from anyone wishing to contribute to development in any way, including assistance with informal and/or formal testing of each new version. Wishing everyone a safe and happy Festive Season and a free and open source New Year, Tim C Sydney, Australia
Re: Open source tools for population health epidemiology and public health
I know a number of folks who would be interested, but the inability to run the software on Win platforms removes them from consideration at this time. Dave Tim Churches wrote: I am pleased to announce that developmental versions of some tools for population health epidemiology and public health are now available under a free, open source software license - see http://www.netepi.org (please note that the release notes for the NetEpi Analysis tool can be found in the documentation section of the project web page on SourceForge). The current development team (comprising two members: Andrew McNamara and myself) will be working on NetEpi fairly intensively over the first half of 2005, with a view to a Version 1.0 release of the tools by mid-year. We would be very happy to hear from anyone wishing to contribute to development in any way, including assistance with informal and/or formal testing of each new version. Wishing everyone a safe and happy Festive Season and a free and open source New Year, Tim C Sydney, Australia