Healthcare Data Quality On A Shoestring: 10 Tips from Stuart Brown Interview – Data Quality Pro

Stuart Brown provides practical tips for healthcare data quality

Stuart Brown provides practical tips for healthcare data quality

How can you create data quality success and industry awards in Healthcare with limited funds and resources?

In this interview Stuart Brown of Zenith Solutions in Australia explains how the data quality changes he made within a public sector health organisation provided a range of benefits that culminated in the organisation winning an award for data quality.

Stuart is a prime example of how the correct deployment of data quality techniques can overcome limitations of technology and resources.


  

 

There was a situation of each silo pointing at another silo saying – “the errors must be coming from them!” so sitting down and working towards some consolidation was necessary.

— Stuart Brown

 

Data Quality Pro: It’s been a while since we last spoke, I believed you had direct involvement in winning a Data Quality award recently, what was that for?

Stuart Brown: Yes! 2011 was a busy year full of good fortune. When we last spoke, I was working with ACT Health as their sole “Data Quality Officer”.

While the organisation had a reporting team that ran validation reports, there was no great expertise focused on reviewing the rules, innovation within the organisation, and engaging the business. I took a less technical angle angle and focused more on soft approaches to softer issues. No budget, no fancy tools, just analysis and communication. The project and work had some great results and ended up earning the organisation the IAIDQ Data Quality Award for the Asia Pacific region.

Quite an honour for a small state government department! It’s about that time of the year again with IAIDQ and the DQ Congress taking new submissions for the award.

Data Quality Pro: That’s incredible, congratulations Stuart. Can you describe what life was like before the initiative? What were some of the issues you witnessed that were created by poor data quality?

Stuart Brown: I think the main issues that were occurring prior to the initiative were due to silos. There were some good intentions throughout the organisation for good quality data, and support from national bodies, however it just lacked some form of glue.

The national reporting area reported on data out of the operational system that had been transformed to fit national reporting specifications. The operational system had some exception reporting on business rules, but there was no reporting to monitor whether these were being run and actioned.

A lot of what the initiative did was to tie some governance to committees already dealing with data, and try to bridge that gap between operational and national reporting. In doing so, we found gaps in address validations. We then created a report that audited execution of quality reports and brought across a few business rules from one report to another.
It all sounds very straight forward, but the impact was around 1,500% improvement in our key national data set and 2-300% improvement in two other data sets.

There were more positive impacts, but those are obvious standouts. When you look at that trend over four or five years, it paints a pretty picture to justify your position!

I was speaking at user groups, meeting with staff in every area, and generally raising awareness of data quality and the WIIFM (What’s In It For Me) factor. With healthcare, that’s quite easy, everyone ultimately martyrs themselves for patient care. It’s a gift and a curse for healthcare professionals.

Once you can tie different data elements to patient outcomes and case studies it is easier to get people on board. There was a situation of each silo pointing at another silo saying – “the errors must be coming from them!” so sitting down and working towards some consolidation was necessary. All sorts of knowledge was gained by every side and the reporting took such a surprising leap.

The idea was then to build a more regular review process to share validations and business rules between the silos, and even from the federal government. They supplied a lot of the specifications for their own reports, however sometimes old rules would be left in reports and people told to ignore them. These things were making the data look much worse than it was and just removing some obsolete rules suddenly started reflecting everything in a more positive light.

Data Quality Pro: You said “…everyone ultimately martyrs themselves for patient care. It’s a gift and a curse for healthcare professionals…” – why is it a gift and a curse?

Stuart Brown: It’s a gift for the patient and a curse for the work/life balance and mental state of the professional! An expectation also begins to build that people will go a bit further than they should be as an employee as it is in the patient’s best interest.

Everyone is there to help the patient, and it takes a special kind of self-less human being to work in healthcare service delivery. I should note that in this instance, Ward Clerks were in charge of fixing errors in the admin system and they are very much at the coalface and working with clinicians. They get to see the frustration of patients, as wrong data means they did not get the follow up letters or reminders for certain appointments and so forth. There is also a real dollar value for the organisation in relation to wasted time for clinicians.

Data Quality Pro: Managing rules in a unified manner so that everyone knows which “hymn sheet” they are working towards must be quite a challenge in a huge, distributed health care organisation – how did you support or improve this process?

Stuart Brown: There is no revolutionary principle to this. It’s really basic, find your champions and harness them. Find the right committees or groups and slot in on the agenda. Tie everything back to the WIIFM (What’s in it for me) principle!

If you go in to a meeting with someone and straight away have the WIIFM factor sorted, everything will go a lot smoother. Know the audience, why they will care about DQ, how it may be able to make their job easier, what benefits will they see etc. It’s all about being there to help them, not about what you are trying to achieve. “This will take fifteen minutes a day” is a less interesting proposition than “this half an hour a day will shave off a few hours of work at the end of the week”.

It’s not just about time, its about how it will make their time at work easier and more enjoyable. Good data means less headaches for you and the patient, or client as it may be. Fixing data at the source means that those who are stuck with fixing it often are far removed from those who read reports and understand the impact on reporting.
Having management huff and puff doesn’t always work in the public service, people have to care themselves. You need a groundswell to coincide with the executive push. That being said, I think the organisation still has a lot of work to do and they really do need that full time position of someone to hold their hand and keep things going.

I was lucky in that there were some quality guru’s that really wanted to improve there area and we were able to use their area for surveys and a pilot. Without those champions, everything is an uphill battle. If people resist, follow the path of least resistance and publish the benefits. Eventually when they realise they had that opportunity to cash in on the benefits, they’ll come around. Organisational cultures never change overnight, it can take time and patience!

I may have digressed slightly there, back to the point of managing rules. Basically, when someone is dealing with business rules they are at the data level and it’s a much easier sell. The more appropriate question to ask them is ‘why wouldn’t you work together?’ Doesn’t that make all of their jobs easier? Aren’t they going to make both their bosses look better and get better outcomes for themselves in their role? Resistance in these situations can be attributed to agendas, politics, empire building, job security, or key person syndrome. There’s no one size fits all or easy solution, this is why having someone dedicated to, or at least tasked with, bridging the gap is such a necessary approach.

Data Quality Pro: Are there any kinds of benefits that you helped to attain as a result of your improvement drive?

Stuart Brown: I’ll give one small example where I was able to quantify a real dollar impact.
There were some particular address issues that were compounding monthly and never seemed to be getting fixed. The validation reports were inconsistent between a particular system and the national reports, and the disparity was growing.

Through the analysis we found out there were different validations going on, and were able to fix the validation in one system. It was along the lines of not validating every element of the address, and only certain aspects. Every element really needed to be validated to ensure it was an existing address.

Once this issue was fixed, the reports were reflecting the same numbers, however the number was still growing! I dug deeper with the systems support team and had them run some audits over who was last to tamper with all these addresses and 90% of data quality issues were originating from one business unit.

I had a phone discussion with their business information manager and immediately found a failing in a business process from a new manager that had skipped the systems training. After a bit more analysis and a meeting or two we had everything back under control and they were courteous enough to supply support to fix the issues originating from their area (thankfully we drummed the correct process back into the staff members and manager). Within a week the addresses were all manually fixed at the source with visual confirmation against physical medical records and it was all done.

That one week recouped $140,000 for the organisation in relation to cross-border care given to patients from another state that received care in our state. Without those correct addresses that would have been a direct loss worn by the organisation.
Now, this is small in the big scheme of things, but that was much more than the yearly salary for the Data Quality Officer position (double in fact) and a good business case for keeping the role around.

There are other figures that can be used but I believe the fact this one week had a strong impact.

Data Quality Pro: That’s really impressive and it just illustrates how much low-hanging fruit there is in healthcare organisations that are maturing their stance on data quality management. So there must be many other thousands of healthcare organisations with “no budget or fancy tools” – how did you succeed on such minimal resourcing? Can you talk us through your processes?

Stuart Brown: Yes, particularly the government organisations that spend all their money on keeping a well stocked stationary cabinet!
The fact I had the one focus was a big factor. We had people running validation reports and sending them off into the ether to be corrected, but those staff members didn’t have the time to chase down people to correct the errors let alone go out into the organisation and identify stewards.

A big part of my job was focusing on those reports, trending them, analysing them, and finding the right people or solutions to rectify them. I opened up communication in the organisation and offered a point of call for support in relation to data quality beyond getting a developer. I was no developer, I was a conduit for the developers and helped re-build the bridges that organisations burn as they turn into silos and focus on their competing priorities.

The public health care sector in Australia is fortunate with a great deal of information governance, however I’m unsure they are fully aware of it. There are peak bodies such as the Australian Institute of Health and Welfare providing data set specifications and a national health data dictionary. There are common code sets and terminologies such as ICD-10-AM and SNOMED-CT. At a state level, there are certainly representatives being sent to national committees and relaying that back to their jurisdictions.

These governance committees exist, often with a focus for standardising or improving data on their charter. Great outcomes can be found by just researching what committees already exist around these organisations, finding the right fit, and perhaps broadening or redefining their charter slightly. At the end of the day, they are governing information and trying to ensure they have correct data, just without the buzzwords of ‘Data Governance’ and ‘Data Quality’.

Perhaps they don’t totally fit with every person’s purpose for data quality but you can leverage them as a starting point.

In the last year or so I have really started learning a lot more about the tools and marketplace. There are more mature open source solutions out there and affordable technologies, so smaller organisations can certainly get into toolsets. I think the important part here is getting your foundations and basics down. Learn how to drive before you walk into the car yard, and don’t buy the Porsche if you are in a country town with dirt roads.

There are sites such as yours providing a great deal of insight into Data Quality Frameworks and methodologies. Really, the greatest port of call for researching approaches was through Data Quality Pro and Google! I found out that the peak statistical organisation of our country, theAustralian Bureau of Statistics (ABS), and Canadian Institute of Health Informatics (CIHI) both built their frameworks from the same Statistics Canada framework.

CIHI gave us a healthcare spin and ABS had a methodology department more than happy to meet up with us and help proliferate their framework through the country. There was a great deal of knowledge out there to leverage, and federal government departments actually encouraging us to open up the lines of communication with them.

Data Quality Pro: You’ve been selected to talk about “Silent Weapons for Quiet Wars: Data Governance by Stealth” – can you tell us more about what you aim to discuss in this presentation?
Stuart Brown: The talk is for the Health Informatics Society of Australia (HISA) Data Governance Conference. The fact that an organisation such as HISA is holding specific conferences around Data Governance, and now for the 2nd year, shows that the Health industry is getting serious. They’ve got a great deal of exciting e-health initiatives happening in Australia and, of course, the undercurrent of everything will be ‘how do we manage the data?’

The brief that I’ll be speaking to is all about what can be achieved through some perseverance and limited budgets. Everyone would like to know what they can do to change the status quo or what they can achieve while being one person in a big organisation. I’ll be taking my past experience as an example to show people how I was able to gain some great outcomes for the organisation by not trying to move mountains.

The Health industry in Australia, and indeed worldwide, have a great level of governance around terminology and so many other aspects of their work. It’s all about harnessing that environment around them.

Data Quality Pro: Finally, what is the data quality community like in Australia? Is it maturing? Are there plenty of opportunities or is it very much a fledgling marketplace and community?
Stuart Brown: I think it is ever-growing. My personal belief, particularly with Data Governance, is that the local organisations are quite behind Europe and the USA. In most instances these practices are born out of necessity, and Australia just hasn’t had the sort of drivers that the finance industry in the USA had that drove development of the concepts and helped build the community. It’s a much smaller marketplace over here, we don’t have the local presence of as many of the tools. The big guys are around, and I think they are seeing the market start to react and pick up a bit. As a country, though, we are catching up. Now we have some niche service providers to help bring skills and expertise to Australian organisations.

It also seems auditors and regulators have started calling out organisations for data specific audit items or regulations that previously didn’t exist or were broadly defined as ‘information collections’ and so forth. There have been more jobs opening up at both public and private organisations with Data Governance Manager or Data Quality Analyst in their titles. It’s great to see.

It’s really an exciting time over here as eyes open and businesses awaken. I’ve been fortunate enough to start working with a niche Data Quality & Data Governance service provider, Zenith Solutions. Of course, there are also some great conferences running and some University interest being shown in the fields that have been contributing on that academic level.