HMO project update -development required

innocuous wrote on Tuesday, May 02, 2006:

Hi all,
The HMO project I mentioned earlier in my posts is underway and taking shape. I would like help with the OpenEMR community to help us use the oss for our project.
In short the trial HMO will be of 100 doctors managing healthcare of 50000 families (4 members per family) spread over a WAN. The area is a slum area of India, but a perfect testbed for diseases and illness.
We will have a central server which will be accessed by the GP’s. Links to pharmacy and hospital will be also there.
Some of the shortfalls/development required are:

1) Doctors can view only their own patients and not all patients. Secondly if referral is given with access code, they can view other GP’s patient also.

2)Calender to be modified so that only the calender of logged in doctor is shown.

3)Patient data to be exportable. We plan to give rewritable CD’s to each patient so that they have their medical data with them wherever they are. This data can then be invaluable incase of emergencies where their primary provider is not available. The data should include all the data of the patient including demographics, encounters etc. Obviously, the data should be updated with each encounter

4)The prescription module needs considerable addons which are:
a)there is no way to delete or delist medications?
b)there is no time frame for taking a medicine i.e. start date and end date
c)DRUG DRUG interaction check is CRUCIAL for us. This should be part of the database on the central server which is updated constantly. An alert should popup if any adverse drug reaction is present.
d)CLINICAL DECISION SUPPORT SYSTEM to be built into the system. We have access to a wealth of information from research institutes, medical sources, colleges etc to build the database.
This should be as an alert system. For example if a person above 65yrs with xyz disease is being prescribed abc drug and the preferred drug is zzz then an alert should popup that the preferred drug for this patient for this ailment is zzz.
I am sure everyone will agree that this will be invaluable to healthcare

5)One major hurdle we see is that GPs’ are not comfortable using a computer and will be reluctant to attend ‘computer classes’. Therefore if possible the encounter screens can be made more graphical with point and click showing images instead of text where needed (for example if chest are is targetted, then a chest icon brings up relavant forms and information.

6)With the same computer literacy concern, the initial medical history capturing becomes a concern. What we are trying is to build an instant medical history program wherein a series of question (in vernacular) are asked the patients and they reply with simple yes or no. Intelligent branching can be used to choose only relevant questions. There are sources available for this, the links of which I will post shortly.

7)Upgrade ICD9 with ICD10 and the ability to add custom codes.

8) Since we are working on a family system, the patient database to be modified to include a family system with 2 parents and two children and the ability to link all the members of the family ie If I click on the male parent, then links to his wife, children are provided.

9)Security and encryption needs to be implemented.

I request all of the developers out there to help us develop these features. Everything will be released as GPL and I am sure that this can be an important landmark in medical history with OpenEMR. I understand that some development will require funding. Please send me your quotations for these features and also which feature will someone be willing to do for free.
Ours is a not for profit HMO and targeting the poor. THe main purpose of this HMO is to bring good healthcare to the poor, who are in dire needs of good healthcare but cannot afford it.
We have bookmarked appx 3000 US dollars for the software for the trial.

Any advise and help will be highly appreciated.
Thanks

sankar1234 wrote on Tuesday, May 02, 2006:

innocuous :
Nice to see your project is on trial.  Did you try some kind of UN funding or grants from your local government.  That is the best bet I believe.  WRT Decision support and Drug interactions, it is not an easy task. In US companies charge around $50.00 per month per provider.  The reason is physicians are working full time 24/7 maintaining these databases as they become obsolete everyday with drug recalls and technology.

WRT export patient data on a daily basis,  So you can write a simple query yourself and export the data to a PDF file.

#1 and #2: Calendar of the physician can be shown. Just modify the query with extra conditions.

Security and encryption: are all part of certificates. Just buy the certificates yourself from India.  OPenEMR should work with the certificates.

innocuous wrote on Wednesday, May 03, 2006:

Hi,
Can anyone please quote for doing these developments?
Thanks

drbowen wrote on Wednesday, May 03, 2006:

Dear Innucuous,

The features that you are discussing will help us with the HIPPA law requirements inside the US.  We all need the drug-drug interaction section and clinical decision support.

I am glad to hear that your project is proceeding.    You will need concrete proposals from the developers for the different modules that you need to be able to apply for grant money.

If you can find some potential sources of grant money (such as those proposed by sankar above), I would be glad to collaborate with you to get these grants into the appropriate agencies.

Sam Bowen, MD
drbowen at openmedsoftware dot org

innocuous wrote on Thursday, May 04, 2006:

Thanks for the help drbowen.
Will any of the developers out there, quote for the customisation and modules required?
Thanks

sunsetsystems wrote on Friday, May 05, 2006:

Your items #2, 3, 7 and 9 are already implemented or would be easy to implement.

#1, 4, 6, 7, and 8 are moderately to very difficult and even designing them to the extent that they can be quoted on would take a great deal of effort.  I think a good way to move this forward would be for you to be much more explicit about how you would like to see things work.  This would surely generate some interesting discussion and help to converge on good designs.

Your $3000 budget is not enough, but your needs are widely shared and perhaps some collaboration will make it happen.

– Rod
www.sunsetsystems.com

sunsetsystems wrote on Friday, May 05, 2006:

> #1, 4, 6, 7, and 8 are moderately to very difficult

Sorry, I meant #1, 4, 5, 6 and 8.

– Rod
www.sunsetsystems.com

drbowen wrote on Friday, May 05, 2006:

I will need to know more about your project and if possible interview Dr. Lele.  I assume Dr. Lele speaks, reads and writes English?  Would he be amenable to me emailing him?  Does he have a curriculum vitae that I could see.  How did he win the Padmabhusan award?  How long has serving been serving as the head of nuclear medicine and research at the Lilavati Hospital.
Tell me more about Mumbai, its size, location, province, demographics?

How many annual admissions does Lilavati Hospital serve. Where does Lilavati Hospital get its funding?  Is it privately owned or owned by the state?  Does it have outpatient clinics that it operates?

How many IT people are in your office? What are their current responsibilities?

I will need to ask for more information to "flesh out" a description of your project.

Sam Bowen, MD

drbowen wrote on Sunday, May 07, 2006:

1. Doctors can view only their own patients and not all patients. Secondly if referral is given with access code, they can view other GP’s patient also.

I would think that the easiest and most secure way to implement this is to set up each physician as a separate database.  If the project is viewed as a ASP provider at the central server then each individual physician becomes a separate practice.  This would give very clean differentiation between the separate client databases.  This would also solve the calendering problem at the same time.

2. Calender to be modified so that only the calender of logged in doctor is shown.

3. Patient data to be exportable. We plan to give rewritable CD’s to each patient so that they have their medical data with them wherever they are. This data can then be invaluable in case of emergencies where their primary provider is not available. The data should include all the data of the patient including demographics, encounters etc. Obviously, the data should be updated with each encounter.

Exporting the data in pdf format would be easy to accomplish.  When a physician needs access to another physicians data this could be released to that physician as a compressed pdf file.

4. The prescription module needs considerable addons which are:

a)there is no way to delete or delist medications?

The “Prescriptions list” is intended to provide a running log of the prescriptions that have been prescribed.  This type of log does not need a “deletion” or “de-list” function.  The prescriptions module should list a duration or discontinue date for temporary medications (such as antibiotics).

b)there is no time frame for taking a medicine i.e. start date and end date

The chronic medication list does have a start date listed.  It does not have a stop date, or a reason for discontinuation.  These should be added.

There is no interaction between the prescriptions list and the chronic medications list requiring a double entry.

c)DRUG DRUG interaction check is CRUCIAL for us. This should be part of the database on the central server which is updated constantly. An alert should popup if any adverse drug reaction is present.

A drug – drug interaction check is indeed crucial and does not yet exist in OpenEMR.  There should be a interaction audit at the time the practitioner attempts to add a new medication.  A pop – up or other warning message should occur for known interactions. 

d)CLINICAL DECISION SUPPORT SYSTEM to be built into the system. We have access to a wealth of information from research institutes, medical sources, colleges etc to build the database.
This should be as an alert system. For example if a person above 65yrs with xyz disease is being prescribed abc drug and the preferred drug is zzz then an alert should popup that the preferred drug for this patient for this ailment is zzz.
I am sure everyone will agree that this will be invaluable to healthcare

Yes, this would invaluable.  But what exactly do you mean by a decision support system?  I have been reading some material and interviews about Dr. Lele.  It is my understanding that Dr. Lele is familiar with and is a big proponent of expert systems where the software assists in making “most probable” diagnosis decisions.  Writing such an “expert system” from scratch is a big undertaking.  While desirable, this is likely going to be outside the scope of this project for some time.  But I don’t see this as critical to the initiation of OpenEMR in your project.

5. One major hurdle we see is that GPs’ are not comfortable using a computer and will be reluctant to attend ‘computer classes’. Therefore if possible the encounter screens can be made more graphical with point and click showing images instead of text where needed (for example if chest are is targeted, then a chest icon brings up relevant forms and information.

Yes, the answer is again, this is a big addition to the existing project.  While desirable,  I don’t see this as critical to the initiation of OpenEMR in your project.

6. With the same computer literacy concern, the initial medical history capturing becomes a concern. What we are trying is to build an instant medical history program wherein a series of question (in vernacular) are asked the patients and they reply with simple yes or no. Intelligent branching can be used to choose only relevant questions. There are sources available for this, the links of which I will post shortly.

Yes, the answer is again, this is a big addition to the existing project.  While desirable,  I don’t see this as critical to the initiation of OpenEMR in your project.

7. Upgrade ICD9 with ICD10 and the ability to add custom codes.

8. Since we are working on a family system, the patient database to be modified to include a family system with 2 parents and two children and the ability to link all the members of the family ie If I click on the male parent, then links to his wife, children are provided.

The difficulty of this I would leave up to the more experienced developers. I doesn’t sound as hard or involved as the above listed improvements

9)Security and encryption needs to be implemented.

You’ll need to be able to configure your central server with SSL and use certificates to authenticate your clients.

Innocuous,

I have to ask, while many of these improvements would indeed be wonderful to have, some of them are projects as big as rewriting OpenEMR from scratch.  I wonder, how many of these features are in the category of “they would nice to have” and how many of them are really critical to get your project started?

Would you consider the following as a workable solution:

1 & 2 – implement each physician as an independent practice with a separate OpenEMR instance and separate MySQL database.

Export the entire patient record as a “comprehensive report” into a pdf for writing to a CDROM.

Improvement of the chronic medications to include a discontinuation date and a reason for discontinuing the medication shouldn’t be too hard.

In my mind the most critical of these is the drug - drug – allergy  interaction cross checking and alert system. 

A decision support system or “expert system” is a large project.  The OpenEMR guidance committee have discussed how to accomplish this and are restructuring the database to make this more realizable.  However this is a long term project that is going to take a lot of man hours.

A graphical point and click system that is user friendly is usable on hand held palmtops or PDAs is also being discussed and development has already been started on this.  This is another long term committee project.

The history taking module is interesting and I have had similar thoughts.  Again this is in the future both not yet available.  Clearly for your location in Mombai (formerly Bombay for the geographically challenged among us)  we would need a lot of help with the vernacular.

The eighth item I would have leave up to Rod Roark or other with  more experience with database development.

It seems that the biggest module that would be critical for your project is the drug – drug – allergy cross checking?

Sincerely,

Sam Bowen, MD

sunsetsystems wrote on Monday, May 08, 2006:

Agreeing with Dr. Bowen, but with a couple of minor notes and corrections:

Medications as currently implemented do have both start and end dates.  Like other “issues”, they can also be associated with encounters and with a diagnosis.  As Sam notes, what’s missing is a direct association with prescriptions and so double entry is required, however this would not be difficult to fix.

There is also currently a feature to export patient data to an XML format.  This is currently just demographics but could be easily enhanced.

ICD10 codes are easily adapted, as I implemented changes last year to abstract the billing code types.  See openemr/custom/code_types.inc.php.

Family associations could be supported by adding a SQL table for that purpose, and of course some associated PHP code.  Should not be difficult.

– Rod
www.sunsetsystems.com

drbowen wrote on Monday, May 08, 2006:

Innocuous,

What are you thinking in terms of hardware?

Laptops or hardwired PCs?

How many units will you need?

I’m thinking that if you want to got with laptops you will likely enough 120-130 to allow for breakage from rough handling.

Our local police force uses the Panasonic CF-50 Toughbook in their patrol cars.  The IT administrator had one involved in a car crash during a high speed chase that cracked the case.  He was still able to download all the data after the crash.

Sam Bowen, MD

innocuous wrote on Monday, May 08, 2006:

We are working on providing Laptops.
We feel the GP’s will be more comfortable with laptops instead of destop units.
We still have to decide whether to go in for a WAN or the internet. However I feel that a WAN will be required to prevent delayed access or any bottlenecks as this is healthcare.
One laptop per GP is sufficient. Yes, around 120 laptops is what we are targetting for the initial project.
-Suvinay

innocuous wrote on Monday, May 08, 2006:

Dr.Bowen,
The most important requirement is drug drug interaction check among the other minor to medium modications required.
The big ones, like you correctly pointed out are:
1)Graphical interface
2)Instant medical history capturing
3)Decision Support System (MAJOR)
We know that these will take time, but would like to make a start so that we get there in the future and do not just spend time ‘talking’ or ‘wishing’ as happens in many cases.
Thanks all,
Suvienay

drbowen wrote on Monday, May 08, 2006:

You sould like you already have a definite plan for the hardware.  Do you already have funding for the laptops?

drbowen wrote on Monday, May 08, 2006:

sould = sound

innocuous wrote on Tuesday, May 09, 2006:

Yes, we have a plan in mind when it comes to preparing the GP’s for EHR adoption.
We plan to give them one laptop plus printer cum scanner so that they can access information and give prescriptions to the patients (in vernacular) and also additional information which educates the patients.
What we plan to do is provide the hardware and charge easy installments to the GPs’ so that they own the hardware over a period of a year. This will ensure that they will take care of the hardware as we have seen that anything provided ‘free’ is misused and not taken care of.
The funding for the hardware will be taken care of when the need arises as it is part of the overall HMO proposal. Prior to that we need to have a workable EMR ready. Time frame 2-3 months.
I will get additional information about the project and mail you personally.
Thanks
-Suvienay

drbowen wrote on Tuesday, May 16, 2006:

Ballards has suggested looking at OpenCyc

http://www.opencyc.org/

For the decision support.

drbowen wrote on Sunday, May 21, 2006:

I have started new page on the http://www.oemr.org/ wiki to discuss this request in more detail.

I have included many of hte comments from this thread.

http://www.oemr.org/modules/wiwimod/index.php?page=MumbaiLilavatiHospital&back=OpenEmrFinancing