We use cookies and other similar technologies (Cookies) to enhance your experience and to provide you with relevant content and ads. By using our website, you are agreeing to the use of Cookies. You can change your settings at any time. Cookie Policy.

Blog: Jill Dyché Subscribe to this blog's RSS feed!

Jill Dyché

There you are! What took you so long? This is my blog and it's about YOU.

Yes, you. Or at least it's about your company. Or people you work with in your company. Or people at other companies that are a lot like you. Or people at other companies that you'd rather not resemble at all. Or it's about your competitors and what they're doing, and whether you're doing it better. You get the idea. There's a swarm of swamis, shrinks, and gurus out there already, but I'm just a consultant who works with lots of clients, and the dirty little secret - shhh! - is my clients share a lot of the same challenges around data management, data governance, and data integration. Many of their stories are universal, and that's where you come in.

I'm hoping you'll pour a cup of tea (if this were another Web site, it would be a tumbler of single-malt, but never mind), open the blog, read a little bit and go, "Jeez, that sounds just like me." Or not. Either way, welcome on in. It really is all about you.

About the author >

Jill is a partner co-founder of Baseline Consulting, a technology and management consulting firm specializing in data integration and business analytics. Jill is the author of three acclaimed business books, the latest of which is Customer Data Integration: Reaching a Single Version of the Truth, co-authored with Evan Levy. Her blog, Inside the Biz, focuses on the business value of IT.

Editor's Note: More articles and resources are available in Jill's BeyeNETWORK Expert Channel. Be sure to visit today!

April 2010 Archives

By Carol Newcomb, Senior Consultant


This is the first in a 3-part blog series, discussing the opportunities that cloud computing offers in healthcare. I present futuristic scenarios from each healthcare contingent’s vantage point:   patients, providers and payers.     A myriad of technologies exist today.   It will be up to healthcare organizations of all types to get their data ready to meet the demands for data integration, security, portability, transparency and accountability in this brave new world.   Mature data governance systems and enterprise-wide data integration will be critical in this endeavor.

Introduction: Welcome to the Brave New World

Now that the Healthcare Reform bill has passed (phew!), it is time to get back to the basics of Healthcare Data Infrastructure.     It is such an interesting time to be involved in Healthcare IT!   The technology is exciting.   The resources are available.   Healthcare providers are gaining some traction in the use of electronic health records.   Consumers are developing interest in eHealth, mHealth and other web-enabled resources.   RHIOs faltered but HIEs are picking up steam, and will also use new web technologies.   Could it be that healthcare data systems will deliver on the promise of accessible, portable, transparent data?    

But hold on a second!   If all that technology exists today, why aren’t we already using it?   It’s complicated, but the main reason is: DATA INFRASTRUCTURE.   There is a mad dash among providers (hospitals, physician’s offices) to transfer from paper-based systems to integrated EHRs.   But, systems that traditionally haven’t been integrated (Finance, HR, Education, Credentialing, Inventory, Purchasing) still won’t be integrated.   Insurance companies receive claims from a host of different feeds, and different departments use data in completely different ways.   Coordination of Benefits isn’t (technically) well coordinated.   My personal ability to assemble a complete PHR might require information from many different doctors’ offices.   Mergers and acquisitions challenge even the most sophisticated health systems.   The list goes on.   Data integration and the need for better data governance everywhere is critical.

There are many kinds of cloud computing, ranging from platform, software and application services to storage and raw computing power.   They will offer opportunities for efficiencies of scale and analytic capabilities, housing massive amounts of raw data.   The technologies and capabilities at hand are truly exciting, but they won’t succeed without sustained infrastructure design and mature governance standards to ensure common usage, security and data integrity.  

I: Patients

It’s so different now when I go to the doctor’s office.   When I show up at the front desk, the receptionist, instead of shoving a clipboard and a pen at me, greets me by name.   ”How are you today, Carol?   Please take a seat and use this e-tablet to write down any changes in your insurance, and what you want to see the doctor about today.   Your wait will be about 10 minutes, but in the meantime, have a cup of coffee, and use the e-tablet to write down anything else that you would like to discuss.   You can also look up some health education materials you may want to find.   Our online library is more trustworthy than some popular websites.   If you like, we can include this information with your medical record, which you will get an electronic copy of when you check out.”

Even though I’m feeling crummy, when I get escorted to the exam room, I’m surprised to see that the setup is completely different.   Instead of the tiny stool the doctor used to sit on, hunched over a computer terminal tucked into the farthest corner of the room (so they can type without prying eyes?), there is a full-sized desk with TWO chairs.   The doctor comes in, shakes my hand, and says ”I see your blood pressure has been under control.   But your blood glucose is a little high.   Let’s look at your readings.”   She sits next to me as she pulls up my chart on the computer, and shows me a graph of the data I’ve been sending through my monitor at home.   Twice a day, the monitor reminds me to check my vitals and transmit them to the office, where they are downloaded to my chart.   ”This could be why you’re feeling light-headed.”

As we look at the graph, we discuss my recent surgery and other medications I am taking, and we come up with a plan to make some modifications.   ”I talked to your surgeon and he says things went well.   But I see from your e-tablet that you’ve also been researching Alzheimer’s.   Is there someone in your family who is having symptoms?”   We discuss my Dad for a while; then she adds the note to my family history, and makes some changes to my prescription.   She then offers to review everything we’ve discussed today on the computer, to make sure she has noted everything correctly.     ”When can I see you again?”   I’m startled because usually the front-desk clerk makes all the appointments, but she finds an opening in her calendar, and automatically sends me an email reminder.

When I’m ready to leave, the front desk clerk hands me a printed copy of the summary from today’s visit, and she swipes my health card to update my medical record and file the bill at the same time.   I now carry this card in my wallet, along with my insurance cards and credit cards because I want to have it with me.   I can also access this information through the web at home, or even at work, through my personal login.   When I go next week to see the dentist, he will be able to access all of my latest drug and medical history information, in case he also has concerns about my blood glucose or blood pressure.  
That coffee was really good.   Maybe I’ll have another cup on the way out!

This benefit of having integrated data available through remote exchanges to this patient will enable her to have a data-informed discussion with her practitioner and to be a better informed patient.   Her office visit is more personalized and streamlined as a result of integrated scheduling and billing.   The time spent is more educational because she can get information for her condition as well as her relatives, from a trusted source.   And through sharing this information, important updates to her family history become part of her own health record.   This is the profile of the new healthcare consumer as a result of Cloud technology.

photo by kevinpoh via Flickr (Creative Commons license)

CarolNewcomb_thumb Carol Newcomb is a Senior Consultant with Baseline Consulting. She specializes in developing BI and data governance programs to drive competitive advantage and fact-based decision making. Carol has consulted for a variety of health care organizations, including Rush Health Associates, Kaiser Permanente, OSF Healthcare, the Blue Cross Blue Shield Association and more. While working at the Joint Commission and Northwestern Memorial Hospital, she designed and conducted scientific research projects and contributed to statistical analyses.

Posted April 29, 2010 6:00 AM
Permalink | No Comments |

By Stephen Putman, Senior Consultant

Hazard - photo by mandolux via Flickr

Data Profiling is the act of extracting descriptive   and statistical data about what resides in a database in order to facilitate data quality and design efforts, among other things. Tools for performing profiling range from open-source software (DataCleaner, Talend) to database system tools (Microsoft SQL Server, Oracle) to specific and purpose-built products (vendors Trillium and DataFlux both offer data profiling solutions). All of these tools automate and standardize the data profiling task, and integrate with other areas of the design effort to varying degrees. However, most organizations do not use them, relying instead on SQL queries, run manually. Why does this happen?

The main reason is that data profiling is not seen as a high-priority activity. Therefore, the investment in licenses, hardware resources, and training for a profiling system is seen to be an unnecessary expense. Consequently, data and business analysts are forced to use existing tools such as SQL query clients to perform basic profiling tasks.

Many times, this is due to a basic misunderstanding of the power of a data profiling tool, and the data it analyzes. Profiling data is useful in logical and physical data modeling, functional system requirement production, and test script generation. Profiling gives the overall system scope and context, and ensures that the system produced satisfies the requirements of the business by characterizing the values present in source systems.

Manual SQL data profiling is fraught with risk. It relies on the skill and knowledge of the analysts producing the queries for accuracy and effectiveness, and there is no guarantee of consistency of results or standardization.. Modern profiling tools have optimized data gathering technology that is faster than manual SQL and scenario repositories that can be shared between analysts and projects to promote reuse.

Implementing a robust data profiling system is an essential part of an effective data management environment. The cost of putting such a system into place is more than offset by the benefits of deep analysis of data sources in both the transactional and analytic spaces. The essential activity is to discover just how powerful such a system can be in your organization.

Photo by mandolux via Flickr (Creative Commons license)

Posted April 22, 2010 6:00 AM
Permalink | No Comments |

By Carol Newcomb, Senior Consultant

Checklist by mykpwedding via Flickr

Having worked at least half of my professional life either with consultants or as a consultant myself, I need to offer some practical advice on increasing the probability of success on a consulting engagement.

Remember, a consultant is just another tool in your management toolbox that provides either the leverage (expertise/experience) or the hard skills that you and your staff are missing today.

There are three main reasons to hire consultants:   1) you want coaching from expert advisors to mentor your existing team and provide tactical advice; 2) you don’t have the skill-sets that you need (ETL programmers, software/hardware experts, etc.); or 3) to provide frameworks and processes to build and sustain ongoing programs (like business intelligence or data governance).

Regardless of the type of consulting engagement, your staff need to be on board with the presence of consultants in their midst.   Ask yourself, as the sponsor, some basic questions:
  • Have you explained to your team what the purpose of the consulting engagement is intended to achieve and their role in making it successful?
  • Are you as a sponsor willing and able to provide authority for the consultants on the ground?
  • Are you willing to spend time with your team and the consultants in the same room, to work out issues, questions, strategy and tactics as they evolve over time?   Don’t think you can just punt and run.
  • Do you have committed resources on your team to help with knowledge transfer?   This is not an ”all other duties as assigned” task for your staff.
  • If consultants will replace or supplement existing staff, is your staff aware and comfortable with this?   This will inevitably become a problem if the expectations are not managed upfront.
  • Are IT and Business ready to start working together?
  • Do you have true business users and are they ‘on board’ with your project?
  • Is there a way to prioritize issues as they come up?   One thing all consulting engagements do is bring to light an unsightly set of issues that become obstacles to progress.
  • Can you be honest about your culture? Are their repeated patterns or behaviors that could sabotage a key project?
  • Can you comfortably ”park” issues that are not showstoppers in order to move ahead despite them?

There are probably many more items I could include on this ”preparedness checklist.”   The point is that consultants can often be perceived as disrupting the natural order of things.   Either you, as their sponsor, will need to introduce them and guide them through the organization to ensure they are effective, or you need to clearly delegate to your staff what the expectations are for them to do the same.   If technical knowledge transfer needs to occur, don’t just assume it will happen.   I have seen too many consultants spend the first few weeks twiddling their thumbs either because they don’t have access to key systems, or they don’t understand the environment and haven’t been provided any documentation to work with in the meantime.   How much time will you allow for consultants to become proficient?

If the consulting engagement is to collect facts, conduct interviews, summarize findings, provide best practices, roadmaps or assessments then make sure they have a political lay-of-the-land upfront.   Otherwise, they will attend meetings they have no context for, they will assume things that may be incorrect, and they will be off-target in their recommendations. Also, many consulting firms sell their methodologies as separate and packaged products. If you expect your consulting firm to transfer its intellectual property as part of the engagement then your consulting contract should stipulate that. If you have secured the funding to bring consultants into your organization, you need to ensure that the project is a win-win.   The risks to you are probably larger than to the consultant group, because at the end of their engagement, they go away.   A good consulting firm will have ensured knowledge transfer along the way, but your staff carries some of the responsibility for that, too.  

If you’re considering or starting a consultant engagement, here are a few things to consider:
  • Is your department/division receptive to advice or change?
  • Are department heads, directors and managers on board with the engagement?
  • If ‘best practices’ are recommended, does the department have the appropriate technical tools and agility to put them into practice?
  • Have you done everything within your control to ensure the success of this engagement?

If the answer to any of the above is NO, then reconsider your readiness, regroup with your staff, and discuss alternative strategies with the consulting group you are working with.

If you have an opportunity to change and improve your department, leverage the effort to instill that change.     You may not get more than one chance.

photo by mykpwedding via Flickr (Creative Commons License)

CarolNewcomb_thumb Carol Newcomb is a Senior Consultant with Baseline Consulting. She specializes in developing BI and data governance programs to drive competitive advantage and fact-based decision making. Carol has consulted for a variety of health care organizations, including Rush Health Associates, Kaiser Permanente, OSF Healthcare, the Blue Cross Blue Shield Association and more. While working at the Joint Commission and Northwestern Memorial Hospital, she designed and conducted scientific research projects and contributed to statistical analyses.

Posted April 15, 2010 6:00 AM
Permalink | No Comments |

By Stephen Putman, Senior Consultant

Circuitboard City by Wonderlane (via Flickr)

I am a computer engineer. I have been a computer engineer for over twenty years. I know a lot of other computer engineers. I have seen many techniques for implementing Business Intelligence systems, with varying degrees of effectiveness.

I have a simple request of my fellow engineers:
Please stop designing functions into your analytical applications without input from your business users.
It is a simple request, one that should be easy to understand and implement. Unfortunately, I have seen many companies where this idea is central to the fabric of the IT organization. Non-business managers are confident that they know what the business needs, either through hubris or a soured relationship with their users. Functionality is designed and developed that bears little relationship to how data is consumed and used in the business organization, usually without regard for the cost of development and maintenance of the resultant structures.

It’s easy to determine that this is happening in your organization - you attend meetings where the following phrases are used with increasing frequency:
”I think this is done...”
”They will need this eventually...”
”We didn’t think of that...
We live in an age where the Business Intelligence projects that are completed are frequently over budget, late, and do not serve the needs of business users (too many projects are cancelled or abandoned as well). Much of this failure can be attributed to communication issues between the business users and the system implementation team. The system team means well by trying to fill in the gaps, but this often leads to frustration, uncertainty, and design based on hearsay. They really should try to resist that temptation.

How much more effective will your system be if you ask your users what they want, instead of guessing or assuming you know better?

photo by Wonderlane via Flickr (Creative Commons License)

StevePutman_bw_100Stephen Putman has over 20 years experience supporting client/server and internet-based operations from small offices to major corporations.   He has extensive experience in a variety of front-end development tools, as well as relational database design and administration, and is extremely effective in project management and leadership roles. He is the co-author of The Data Governance eBook, available at information-management.com.

Posted April 8, 2010 6:00 AM
Permalink | No Comments |
By Carol Newcomb, Senior Consultant

One Small Step... photo by Yoni Lerner via Flickr (Creative Commons)

In my last post, I bemoaned the carrot-or-stick approach that has resulted in immature, disintegrated reporting solutions.   In this post, I challenge the notion that business intelligence, as opposed to formal data governance, is a fix to healthcare organizations’ fragmented data.  

Collecting data is easy.   Most electronic health records (EHRs) and personal health records (PHRs) are front-end data acquisition and transfer tools, with sophisticated built-in rules and customized algorithms.   Storing data, metadata, and even ”data exhaust” gets cheaper by the hour.   The trick is turning all this data into meaningful information, and using that information to make smarter decisions.   How hard is it to turn data into reports using your current data warehousing technology?   How frequently do you get reports that just don’t jibe?   Can someone confidently explain how a number on a report got there?

IBM is running a catchy ad that resonates deeply in our data-saturated world.   ”If you can’t believe the data, how can you believe the analytics?”   (Healthcare CFO)   Right on!   Business intelligence is NOT the place to tackle healthcare integrity issues.   Let me pose some fundamental BI questions:

  • What’s your headcount of 3rd year medical residents today?   In each department.
  • How many RNs and LPNs do you have in any single department?   Last Sunday?
  • What is the average patient wait time in your GI lab?   What’s the variability?
  • How many repeat lab tests did you have for patients who are being diagnosed with AIDS?   How many are positive?
  • What’s the dollar amount for lawsuits filed for events that happened on the pediatrics unit?   Were those lawsuits for ‘Never Events’?
  • What is the failure rate of certain brands of hip prostheses used in hip replacement surgery?   How long before they failed?
  • What’s the infection rate on each day of the week in the ICU?   Does it change at shift-change?

These are straightforward administrative and clinical quality questions that most healthcare organizations are hard-pressed to answer without scraping around in spreadsheets, disconnected data systems or standard reports.   Why?   There is a fundamental lack of recognition that data governance and data integration are critical infrastructure elements that continually get no attention in a strategic IT budget, and get zero dollars as a result.   Healthcare data management is stuck in its infancy, crawling on hands and feet.

It is important for healthcare organizations to use the government’s incentive dollars and retool for the 21st century.   However, while incentives are there to facilitate the technology transition from paper to electronic data, there is huge risk in installing yet another front-end system and further complicating your back-end IT infrastructure.     Organizations that have spent years consolidating on SAP, Oracle and other ERP systems have needed to return to fundamental governance practices to resolve the confusion that usually results from inconsistent definitions and non-existent data usage rules even after years of effort.  

The failure to design data governance at the start of an integration project is expensive, and wastes the best features of even the most sophisticated business intelligence tools.    

It is sad to compare healthcare to retail, manufacturing, banking, and just about any other industry in terms of customer-centric capabilities.   Healthcare organizations have failed to take a sustainable approach to data infrastructure because they have ignored the need for data governance.   This churn will only continue until healthcare leaders finally step back and insist that solid infrastructure--including a data governance plan, an MDM plan, a data integration roadmap--are budgeted, staffed and successful.   Otherwise, they had better hope that attestation will continue to work.

photo by Yoni Lerner via Flickr (Creative Commons License)

CarolNewcomb_thumb Carol Newcomb is a Senior Consultant with Baseline Consulting. She specializes in developing BI and data governance programs to drive competitive advantage and fact-based decision making. Carol has consulted for a variety of health care organizations, including Rush Health Associates, Kaiser Permanente, OSF Healthcare, the Blue Cross Blue Shield Association and more. While working at the Joint Commission and Northwestern Memorial Hospital, she designed and conducted scientific research projects and contributed to statistical analyses.

Posted April 1, 2010 6:00 AM
Permalink | No Comments |