Clinical data random information

I've become an information hoarder. As I spend more time thinking about Information Management and speeding the move to better technical systems, I am amazed how general the principals of design are between the different industries.

Here is a noobs (i.e. me) "plain spoken" understanding of a key term in managing patient data across hospitials and for predicative analytics and personal health decison making.

Level setting (i.e. in general the definition of Clinical data warehousing) Clinical data warehousing is a patient identifier organized, integrated, historically archived collection of data.

For the most part the purpose of CDW is as a database for hospitals and healthcare workers to analyze and make informed decisions on both individual patient care and forecasting where a hospital’s patient population is going to need greater care (i.e. patient’s are showing up as obese; therefore the need for specific hospital programs to fight diabetes are a good idea).

Data warehousing in healthcare also has use in preparing for both full ICD-10 and meaningful use implementation. For example; McKesson through its Enterprise intelligence module probably has plenty of CDW management capabilities the only interested in meeting the upcoming ICD-10 and meaningful use deadlines. These kinds of worries are only for US hospitals. However since Canada requires ICD-10 compliance for all EMR systems this does present a benefit to Canadian healthcare.

In principal since data warehousing at its core is about building a relational database and should be EMR supplier agnostic. Since McKesson is an ICD-10 and meaningful use- ready supplier, the database itself should conform to standards that would allow general solutions to be used. This article goes through some of the potential benefits and pain points. It is tailored to clinical trials but the underlying message that building a CDW is a ongoing procedure is the same for other uses.

One example of how this may be done is Stanford’s STRIDE; they used HL7 reference information model to combine their Cerner and Epic databases. This is part of a larger opensource project that may be an option if an organization has some development expertise.

https://clinicalinformatics.stanford.edu/research/stride.html

Since the main user of CDWs tends to be the people doing the analysis (current buzzwords for search for analytics include:BI, Predictive analytics, enterprise planning, etc) it is probably useful for Health IT professionals to understand its WHO and WHAT the CDW is for within the organization...i.e. have a full blown Information Governance plan that places a value on information not just a risk assessment. 

Security without usability isn't better healthcare

I spend a lot of my time understanding how information is stored, accessed and protected as part of my role as a IT analyst. I always am astounded at how little of what is standard practice in many industries as not filtered over to health care and/or life sciences (Pharma+Biotech+academia).

The recent hub-bub about ACA (AKA Obamacare) has completely yelled over the real transformation opportunity in healthcare. Up until the recent deadlines and political fights regarding ACA "everyone" was really concerned about meaningful use. The TL;DR version of the MU legislation is this: make information available to care providers and patients.

So what are we really talking about here? It is really pretty simple; it is information management and the processes that guard against mis-use while enabling productivity.

Lets be honest the EHR/EMR solutions implemented at most organizations do not enable productivity or protect information. Doctors hate them because they do not fit their work patterns (see here), hospitals are have significant issues with data protection (see here) and importantly it is not mitigating the biggest risk to patient outcomes (and hospital liability) (see here).  

It is time to re-think the information silos in healthcare.

So if a single poorly accessed EHR is not the answer, what is?

I would argue that we need to think about this based on information flow and how we expect the value to be delivered. In this case patient care.

An interesting model to think about is the Canadian delivery model. For example; Ontario E-health has determined it is neither cost effective nor timely to build a single system for every hospital.  At the moment, 70% of all physician practices and hospitals already have some sort of EHR system in place. So rip and replace is not an option, the reality is we need to make lemonade.

Since Ontario funds the hospitals through direct allocation of tax revenue, it is loathe flush that money down the drain. 

Therefore the best approach is to control the data itself (including digital images, prescription history, surgery, etc) and letting the individual hospitals control how they view and use the data. 

In other words- Make it easier to access information based on who you are and what you need the information for!

Focus on the Information exchange layer

Consolidated Information Management layout for Patient care focus. 

So how do we do this without moving to brand new systems and shiny new toys?

The same way every other industry is doing it; especially low margin high risk industries such as Oil and gas, Insurance and Manufacturing. Keep the clunky but very secure system and take advantage of the new technologies that enable information sharing. Instead of all-in-one solution add an ECM or portal to manage rights, search and presentation. It will be more cost-effective than doing nothing or rip and replace.

This structure controls movement and access to patient data, allowing for quick access to the appropriate information based on job and location.  It provides a structure that takes advantage of the current investment in a secure database yet provides a flexible layer that is designed to convey information in context for end users. 

This may not be the best system or the system that you would design from scratch with an unlimited budget, but it gives a long term flexibility AND doesn't require a rip and replace of your current EMR/EHR. It should provide very good, highly usable healthcare at a reasonable cost.

The way they are going about the change may not be splashy but it will work for both patients and doctors- that’s a great thing. The one thing it won’t fix is the doctors who refuse to use it-and that is a bad thing.

There is additional cost involved in this model but if teh doctors and nurses do not use what you have now.....would salvaging that investment be better?

Love any comments or critique of the model.

Twenty skills that I -or any Ph.D- has that are in demand

A while ago Christopher Buddle posted a blog on SciLogs about what you needed to know before becoming a professor. Many of those skills are the ones in demand outside of academia. 

It got me thinking generally what skills I have amassed over a Ph.D, Post-doc and faculty position. For any other "recovering scientists" reading this please feel free to steal this list, add to it or perfect it. Any comments or critique would be welcome. 

1. Project management- over my academic career I managed to publish several papers in top journals. Some required precise planning of tasks and experiments on a short deadlines against competition. This requires ensuring that each set of experiments is finishes with a high quality deliverable.
2. Human resource- as a professor I had to hire, fire and develop staff. This included students and early career professionals where you are balancing what they are capableof today, with their career goals. I picked projects for them that they matched their skills.
3. Project planning- a PhD is a set of projects, that need to be planned out, with a full timeline, deliverables and costs set out. In addition a key part of a successful PhD or post-doc is knowing when to kill a project.
4. Stakeholder relationship- each stage of a PhD requires you to set out goals with your faculty advisory committee. These people will provide guidance and advice for where you should spend your time. Part of success is ensuring that you cogent show progress toward each of the members ideas ofyour success. The stakes get higher as you move to a post-doc where you are expected to manage the project and manage the expectations of your boss.
5. Budget building- as a professor I needed to build RFPs, prioritize purchases based on project needs-as well as the long term strategy of the lab, source infrastructure, mange vendors and raise funds.
6. Publications- part of a scientists job is to communicate results to the community. This includes typical writing skills but also graphic design, matching the presentation visualizations to the message and audience.
7. Data management- all aspects of data management including ensuring high quality data recording metadata, designing database considerations. Build database querying, integrating public and owned data into a complete set.
8. Analytics- a key part of my PhD was defying how to quantitate behavior and images. This requires a clear analytic method that allows reproducibility through clear, logical rubric for scoring purposes.
9. Web based research-not just the query but also the decision on good sources and bad ones.
10. Public speaking- I have given hundreds of lectures to all sizes of groups both lay groups and expert groups. This gives me a large set of tools to fall back on for presentation design
11. Individual drive- to do a PhD you need to an internal drive to do what must be done.
12. Intellectual flexibility- as part of my PhD I learned at least 12 different technical skills at a high enough level to use them in peer reviewed publications and teach them to others. I learned these through reading and just dpingi didn't need to be walk through them multiple times.
13. Records management- my laboratory work in a high demand, high competition environment. We needed to have all experiments documented in a way that would stand up to legal review and could be used as part of a patent process.
14. Understanding of several healthcare related regulations- part of my work was related to drug discovery and some of it was in collaboration with clinicians. Meaning that we ensured that all documents and protocols met the required standards.
15. Graphic design- genetics is a hard area to explain without pictures. I designed many successful visualizations using Photoshop, powepoint and old matte photography techniqies.
16. Process design- my laboratory was at the bleeding edge of genetics. This meant that we were constantly building new processes and testing resources that would be best for that process.
17. Process optimization- due to the unique methods we constantly needed to set production standards and build analytics that allowed us to evaluate and optimize process and make changes that reduced cost and increased reproducibility and accuracy.
18. Contract negotiations-as part of my job, I have negotiated service contracts, terms of employment 
19. Fund raising- academic labs are also look for new sources of funding and interacting with potential investors/funders
20. Strategic product planning -a key part of success is understanding where government priorities are now and the next five years to develop a funding strategy. Successful scientists also have a understanding of the competitive landscape and position their employees and infrastructure to keep up.

Time for some convergent evolution in knowledge management

As I move from the ivory tower of Neuroscience to the practical, business related advice that Info-Tech gives clients on their IT environment I'm amazed at how many parallels I see in the needs and the solutions in all kinds of human endeavours.

For example, I just finished talking to a vendor about how Enterprises can manage and maximize their content (images, documents, blogs, etc). Much like my own thinking on this, @OpenText is convinced the core issue is about information movement not what it is stored in (i.e. a word doc VS a excel).

For me this comes back to a practical problem that I had as a graduate student. My Ph.D was on how gene expression relates to brain development. The brain develops in a fascinating manner; it starts out as a tube that grows outwards at specific points to build the multi-lobed broccoli-esque structure that allows all vertebrates but particularly mammals to have diverse behaviours and life long learning.These complex behaviours rely on an the immensely diverse set of brain cell types. Not only is their great diversity of cells but each cell needs to get to the right place at the right time.

Think of a commute on the subway; not only do you need the right line but if you don't get to the station at the right time you won't get to work on time. This could lead to you getting fired. For brain cells this could lead to death. For the organism it could mean sub-normal brain function-and potentially death. The fact that the process works is a testament to the astounding flexibility and exception management built into cells by their epigenetic programming.

There is however one big problem with the type of brain development: the skull. The skull limits the number of cells that can be created at any given time. Practically this means that the level of control that must be exerted on the number of any one cell type is very tight.The control comes from coordinating which genes are expressed in each cell type to allow cells to make decisions on the fly. Usually it starts by the brain cells take off in a random direction that then informs them of what type of brain cell they will end of being when they arrive. The cells then proliferate as they move based on the contextual information that they receive about how many more cells are needed. This all happens through cell to cell communication and rapidly changing patterns of gene expression.

(Wait for it Ill get back the parallel problems honest....)

As you can imagine this was (and still is) a daunting problem to investigate. My research involved a variety of time staged images; reams of excel workbooks on cell counts, brain size; word docs on behaviour and whole genome expression sets. It was the a big data problem before the phrase existed. (Business parallel No.1). In reality I had no problem keeping track of all this data and looking at each piece and doing the analysis on each piece. I had very good notes(metadata) and file naming conventions (classification) to ensure that I could easy find the file I needed. I was in effect a content management system (Business parallel No.2).  The problem was synthesizing the separate analysis into a cogent piece of information i.e. something that can be shared with others in a common language that allows other to build their own actionable plan. (Business parallel No.3).

Any scientist reading my dilemma from 15 years ago can probably relate-and so can anyone else that uses and presents information as part of their job. The reality is that technology can only solve the problem if people recognize the problem and WANT to be systematic in their habits.......the will power to be repetitive in their approach to work is sorely lacking from most knowledge based workers. Ironically a lack of structure kills creativity by allowing the mind too much space to move within. The advent of the online databases by NIH from genomic, chemical and ontological data has given a framework for scientists to work within to quickly get up to speed in new areas of investigation. Unfortunately this has not trickled down to individual labs (again more proof that trickle down anything doesn't work effectively-its just not part of human nature).

This lack of shared framework across multiple laboratories is becoming a real problem for both Pharma and academia (and everyone else). The lack of system has led to reams of lost data and the nuggets of insight that could provide real solutions to clinical problems (Business parallel No. 4). This also leads to duplication of effort and missed opportunities for revenue(grant) generation.(Business parallel No.5).

From a health perspective, if we knew more about what "failed drugs" targeted, what gene patterns they changed and what cell types they had been tested on we could very quickly build a database. From a Rare disease perspective the cost of medical treatment is partially due to the lack of shared knowledge. How many failed drugs could be of use on rare diseases? We will never know.

This is a situation where scientists can learn from the business community for the technical tools to really allow long term shareable frameworks. These technical controls are available at any price. Conversely the frameworks and logic that scientists use to classify pieces of content to link them have lessons for any knowledge worker.

Its time for some open-mindedness on both sides, the needs for all kinds of organizations and workers are converging-too much data, too many types of data, not enough analysis. Evolution is about taking those "things" that work and modify them for the new environment.