About medical software complexity - full size

(c) Rundall Munroe xkcd.com

At least the last 15 years I'm reading about plans of, for example, HL7 implementation in my country. This plans still far away from reality and not because of small amount of computers in the medical facilities. Quite often they have more that enough. But if you'll ask staff concerning mentioned HL7 or ICD-9/10, SNOMED, LOINC, etc. and what kind of software to use to support it, in the most cases you'll never get answer even about a general subject of your question. And this is not about level of regular nurse. Definitely, I have no in mind to hurt our medicine. Opposite, we have many great doctors, nurses and especially, surgeons.

I just dreaming about equal quality of medical service in any establishment and recalling the last 2 years of hard fight for mother's life. She dead recently...

Maya Angelou
         All great achievements require time

Not long ago one very respectable doctor said in the interview that few months enough for top level specialists to teach medical staff of my country the modern world standards. In general, it is not a big deal to organize a such learning process. But let's make one step ahead - how to create strong daily control of final result quality? If you were examining modern tools intended to operate with mentioned above and other standards, you'll probably confirm that they require serious knowledge and constant concentration. Known and widely in use way to proceed will be creating a team of inspectors. Than to create a team of inspectors over that first inspectors, and so on. 'Sweet' hugs of Parkinson's law in which we already living...

I'm thinking that most efforts must be applied to creating of international automated computerized environment, which does not need much attention from staff to learn how to operate with it. Since this is quite a sensitive topic, I inform in advance that pursues only the goal to examine the situation from different sides, and briefly outline own point of view on problem's decision. As USA takes leading positions in the world's healthcare and openly paving the way forward for my country as well, I will refer mostly to its known respectable media sources.

Asked about the reason of my interest to HL7 and other standards I say that, unfortunately, for people of my age is not easy to find a worthy application of their forces and knowledge here. I searched hard an engineering task, the solution of which would help to hold on. Now, after mother's death, I see no reason to continue research in this area and do not planning subsequent publications on medical subjects. Here's a picture of my digital medical library concerning HL7, to confirm that the articles are my personal opinion, based on hundreds of documents that I read before to express own attitude to the subject:

Any opinion, especially in medicine, must be based on recognizible experience and knowledge. So, I'm publishing here the brief explanation of main principles and over 300 pages detail description of own software 'Blood Center Manager' and 'Blood Center Manager Assistant' (further, BCM(A)) in the field of Blood Establishment Computerizing.

You know about change in the modern healthcare industry strategy from fee-for-service to the brand new trend fee-for-quality (or fee-for-value). Our current Minister of Health stating that doctor's fee must directly depend on the number of patients he/she serving. It is hard question about correlation of time, which doctor spent, for example, for patient screening comparing to its results tracing and storing according to the rules of ICD-10 with help of modern software, designed for that purpose. Definitely, a talk about the exceptionally precise calculation of fee needs attention, but final goal of all manipulation with tools, papers and software is healing. So, from the point of view of a regular patient, the definitions of modern Merit-Based Incentive Payment System (MIPS) or Alternative Payment Model System (APMS) is incomplete. Let's take a look to the scary wide range (1 to 100 points) MIPS and find a word about main purpose of healthcare, mentioned above:


Article: “Doc Fix” shifts physicians into value based payment, highlights the importance of ICD-10
By: Ann Chenoweth

... Physician scores will be assessed based on how well they perform in the following four areas: quality, resource use, clinical practice improvement areas and meaningful use of EHRs.

What do you think about the next formula?

Patient payment amount = regular fee for qualified service + REWARD FOR RECOVERY

EHR, HL7, SNOMED, LOINC, CDA, ICD, etc. are good individually and has its own quite long history, but still does not expose applicable clear, inexpensive, protected and simple interoperable joint implementation, as it seems at first glance. Looks like most serious achievements still ahead:


Article: Deaths by medical mistakes hit records
By: Erin McCann

... It's a chilling reality – one often overlooked in annual mortality statistics: Preventable medical errors persist as the No. 3 killer in the U.S. – third only to heart disease and cancer – claiming the lives of some 400,000 people each year.

March, 2015
Video: The next outbreak? We are not ready
By: Bill Gates

In my opinion, for a qualitative solution of the problem is not enough efforts of one healthcare regulating organization or software/hardware manufacturer. It requires synchronous actions of international and national regulators like WHO, Open Government, IHE International, IHTSDO, ONC, FHIM, HITSP, USA FDA and OpenFDA, as well as manufacturers of hardware and soft goods like Baxter, Haemonetic's, Texas Instruments, Freescale, IBM, etc. It is not about the next super technology "toys", but the creation of an international protected trusted environment, where blood derivatives production is a part.

April 21, 2015
By: Jack Welch

... Leadership today is all about two words - it's all about truth and trust.

And since trust is extremely expensive, then such powerful organizations needed to create worldwide approved standards and available automated means to support it at all stages. For example, in my opinion, building of a powerful fractionation plant hardly appropriate now in my country. On the other hand, current level of production control are insufficient for serious international partnership in the field of Blood/Plasma collection.

Such developments are not for singles because the strategy of operating with data must be changed at the international level, and only description of idea, original definitions of data and its relations will require several months. In addition, the only software does not solves the problem and specialized hardware must be created, also. And in any case, the physical implementation of projects of this complexity should not start without approval at the level of above-mentioned organizations, which is quite complex process itself.

(c) Rundall Munroe xkcd.com

But otherwise, we will get the next set of regular software, which is living self private life, exposing bunch of screens with a bunch of fields all together actively 'eating' doctor's valuable time.

Jun 12, 2015
Article: EHR Software BuyerView | 2015
By: Gaby Loria

... 59 Percent More Buyers Replacing Existing EHR Software in 2015

Tens of thousands pages of already created standards might be explained in a different way, and as a result implemented in the unexpected style. Is the next is about the same idea and similar approach? Maybe the question is just who first began using of XML?

FHIM Information Modelling Process Guide (Version 1.0, 09-01-2014)
... The Federal Health Information Model exists to support the exchange of health information among federal agencies and their

Introduction to HL7 Standards
        HL7 and its members provide a framework (and related standards) for the exchange, integration, sharing, and retrieval of electronic health information.

God bless you to avoid, for example, 'catch-22' paradox on your way of trying to 'sew' all information together. 
flow charts
(c) Rundall Munroe xkcd.com

Ones, long time ago I was thinking to apply for USA FDA 510K concerning BCM(A) software. I was expecting definitions of code instrumentation requirements or, for example, some similar to JTAG regulations, which is widely in use in the process of automated digital hardware testing. Unfortunatelly, I got an answer that approach of such purpose medical device self-testing based on approved by FDA data patterns does not exists. In result, I had to refuse when understood a real procedure and its approximate cost.

Roughly the same time I had chance to talk in person with VP and engineer of Haemonetic's Corp., and was asking is it possible to change RS-232 port in the PCS-2 machines to more usable for customers Ethernet, or at least RS-485. In a short words, I was thinking to connect all PCS-2 machines to the facility local network and use information of its scales, the number of pumps cycles (which depends on donor's plasma fat) and some other data to improve Center's productivity by staff work-load and hardware service scheduling. It is simple task up to the moment we came to the theme of 510K approval ...

It was about 15 years ago and maybe something changed in current, but I'm not sure:

Article: FDA must make smarter use of big data
By: Jack Beaudoin

"Our inefficient, less-than-modern, drug discovery and device approval process drives up cost and delays treatment," said the initiative's co-chair, former Senator Bill Frist. "We must accelerate the process of getting safe and effective drug and medical devices to patients."

Let's recall how Ebola started and pay attention to the next warning:

May 27, 2015
Article: The most predictable disaster in the history of the human race
By: Ezra Klein

'Gates's model showed that a Spanish flu–like disease unleashed on the modern world would kill more than 33 million people in 250 days.'

I'm completely agree with the header of the next article, but unfortunately not so much optimistic concerning some of its expressions:

April 16, 2015
Article: DeSalvo at HIMSS15: 'True interoperability, not just exchange'
By: Mike Miliard

In her keynote Thursday morning at HIMSS15, National Coordinator Karen DeSalvo, MD, said it's now time to "focus beyond adoption" of health IT and create an interoperable, learning health system "upon the strong foundation we all have built"...

By my opinion, the next continuation means serious changes or even complete rebuilding of foundation, mentioned above:

... To get there "as quickly as possible," three big things must happen, she added.

First, we need to "standardize standards, including APIs."
Second, we need "clarity about the trust environment" around data security and privacy.
Third, it's critical to "incentivize, in a durable and sustainable way, interoperability and
the appropriate use of health information" said DeSalvo.

Getting back to interoperability, HL7 or FHIM does not have much possibilities to care about what the software on both sides of line will do with the data of message. So, in general we may have 3 different and quite complex software means, where two intended to support activity of message transmitter and receiver facilities and, for example, HL7's protocol in between. What is "interesting", that message receiver usually operating with data in its sole discretion, that the result may be one reason for the following:

May 27, 2015
Article: Hackers hit health system, swipe data on 220K
By: Mike Miliard

Cyberattackers were able to swipe the personal and protected health information of both employees and patients, including patient names, ID numbers, Social Security numbers, dates of birth, medical diagnoses, treatment data, drivers' license information and other medical-related information.

HL7 exposes a set of rules based on RIM, which is, accoringly to Gunther Schadow and other in the article 'The HL7 Reference Information Model Under Scrutiny', built 'in a form in which Entities (e.g., people places and things, nouns) are related in Roles (relators) to other Entities, and through their Participations (prepositions) interact in Acts (verbs)'. Entity here is a key source, which creates a complex list of relations with other members of the chain. To reduce complexity of HL7 implementation the FHIR appeared, but again with the next statements:

May 2015
HL7 Policy Summit - Paris
By: Ewout Kramer

If we want FHIR to be part of the future of heathcare, we need to change our attitude...

Further in that slideshow Mr. Kramer confirms, that FHIR is not a standard and anyone can create appropriate FHIR-profile for own needs. The next is not directly about FHIR, but concerning point of view to the standards itself. Imagine the situation when some busy doctor assigning specific injections for a patient, based on the list of previous prescriptions. According to CDA the documents looks good, but a little detail - doctor just changed working place, and in previous facility same data usually where ordered by dates in opposite direction. It is not a fact that doctor will catch mentioned difference. So, the standard approach is preferable, respectable and much trusty for patient, by my opinion.

Jun 9, 2015

Blog: The role of a patient with respect to their chart
By: Keith W. Boone

... The RIM also allows me to relate the organization to a patient chart via the maintainer role, so I think there's some duplication here. The best that I can do is show that the patient role has indirect authority (a role relationship) over the health chart role. But I need a direct role relationship between the patient and their chart is because roles confer rights and responsibilities.

... This gap is interesting because it shows the disconnect in thinking about patients and their rights and responsibilities with respect to their chart.

... RIM Harmonization is a pretty challenging process, and I've already got a lot of irons on the fire (or is that FHIR?).

Trying to use RIM as a Swiss Army Knife, which can do a lot, but not intended to do anything perfect, we have to keep in mind its mediatory role by definition. But mediator can't responsible care of data on both sides (transmitter and receiver).

The next collage I made from:

Mar 2015
Video: The surprisingly logical minds of babies
By: Laura Schulz

Baby learning

Notice how already in the second year of life, the child's action dramatically varies considerably in relation to the subject matter that does not meet expectation. Of cource, to a large extend it depends on the fact that children do not have a clue about business relations.

Who is the head in this house?

In the last about 20 years an incredible efforts and amount of money spent to design and implement of HL7 standard, which main purpose is interoperability support. In the days when the standard was conceived, there was no other solution. But in novadays its general idea looks exhausted, especially taking into account that its development still in progress.

By my opinion the next must be done:

- change the point of view to the data with which the software, staff and regular users operating
- change approach from health information (and not only) exchange to its availability
- put Data onto the key position, instead of Entity, which operating with it
- separate Data owners, providers (holders) and Entities, which wishes to use it

Let's take a look to brand new IBM's Identity Mixer. Certainly, I do not know how wide IBM Corp. thinking to expand and implement that idea. To me it is good example of strategy, where data can be used as the key, with respect to the difference of its owner and provider. Data, which someone or something (robot) want to use, may check by itself that someone's eligibility (Entity and its Status, Action, Location, Time, etc.) accordingly to standard dependency declaration manifests, because data is not just a big file-cabinet in some cloud, and the way of the same data use and presentation may be vary, depending on many parameters. Concerning manifests I want to add, that to be understandable and easy operable by responsible designers, its models might be designed by UML, BPMN or other similar purpose means. But widely in use XML does not fit to computers and manifest's final realization must be transformed to the values, applicable for fast processing.

Quite often we can read about researches in the field of self-learning computing systems and special frameworks for manipulation of clinical data for decision support in healthcare like, for example, GELLO. Let's imagine a patient, who has strong insomnia and constipation. The GELLO's description of automatic decisions in each separate case is clear and can be even simple. But it may create 'funny' effect in this particular case, if applied simultaneously, because at least as I understand, a regular 'drug-drug interaction alert' does not exists in this case. Probably some special 'clinical reminder' must be applied, or the descriptor of decision will get more complex and its complexity limited just by its creator knowledge and laziness.  

Please, take into account a drastic changes in the climate picture of the worldmodern genetic experiments, including the key elements in the our food chain, dramatic shortening of the period of effectiveness of modern antibiotics, etc. All this taken together, directly or indirectly affect the decision on the treatment of diseases and of course, some important role should be assigned to  computers. But not in the sense of a final decision. So, it is good idea to leave data collecting and primary processing to computers, making final decisions and treatment to physicians, which must use recomendations of some International Association of Doctors (control of this is a task for computers), and take serious attention to the next warning: 

Article: Stephen Hawking warns computers will overtake humans within 100 years
By: Sam Shead

... Speaking at the Zeitgeist 2015 conference in London, the internationally renowned cosmologist and Cambridge University professor, said: “Computers will overtake humans with AI at some within the next 100 years. When that happens, we need to make sure the computers have goals aligned with ours.”

By the way, if you'll read with attention the descriptions of BCM(A) software, you'll probably see why I had to expand the name of BCM with "Assistant", while developing project for Windows. BCMA comparing to BCM has different strategy, but not just operating systems, for which it was made.

SuiteYou know, in the times of USSR we had great stand-up comedian, humorist and actor Arkady Raikin. The next is from the TV film "People and mannequins"

- You think I'm on the wrong button buttoned? If I fasten on that, it will be even worse ... Guys who sewed this suit?

It turned out 100 people and say: "We!"

- Who are we?
- We have a narrow specialization - one sews a pocket, the other - the hand, I personally sew on buttons. There are claims to the buttons?
- No - sewn to death, will not tear off. And who sewed a costume? ...

It is a time for the world's key organizations to join efforts in the one responsible solid international hard working team, which will operate as a one wise man:

For the transgression of a land many are the princes thereof: but by a man of understanding and knowledge the state thereof shall be prolonged.
Holy Bible, Proverbs 28:2