Health Insight Technologies recently announced that it has rebranded as Humedica, Inc., a venture-backed firm in Boston that plans to offer a software-as-a-service approach to clinical intelligence. Their model intends to reduce the burden of BI implementations by eliminating local infrastructure. Using ETL services, Humedica will store healthcare data at a national-level using a centralized clinical data warehouse; then offer access to the data for quality reporting and research, presumably for a fee. The article “Humedica Wants to Dose U.S. Healthcare Crisis with Clinical Analytics” by Ryan McBride offers a great overview of Humedica’s impressive endeavor.

Dan Housman
Managing Director, Analytical Applications

Labels: Clinical Intelligence, Data Warehousing, SaaS, Translational Research

Recombinant is scheduled to co-present with Amazon Web Services at the upcoming Oracle OpenWorld conference on October 14th in San Francisco, CA. Joseph Adler, Solutions Architect will explore the cost, performance, and operational advantages of healthcare and life sciences data warehousing in the cloud. For more information about the conference, visit the OpenWorld website.

Labels: Cloud Computing, Data Warehousing

The i2b2 and related Crimson team recently published an article titled "Instrumenting the health care enterprise for discovery research in the genomic era". The article describes the potential to acquire large-scale samples needed for genetic studies by integrating data warehousing with lab systems.

Hospital network infrastructure for phenotyping and collecting biospecimens for biorepositories and omics data acquisition has so much more scale than current methods. New ideas and pioneering work in this area are likely to make a substantial impact in how translational research evolves over the next few years.

Recombinant expects to see both consented and non-consented/discarded sample models for high-scale phenotype-genotype matching. Ideally these models will integrate with translational research stacks such as i2b2, caBIG (caTissue/caGRID), and frameworks like GenePattern for analytics.

Recombinant is working on multiple projects to bring some of these ideas into an open source tool set that can be implemented across sites via the same model as i2b2.

Dan Housman
Managing Director, Analytical Applications

Labels: i2b2

John Halamka published a blog post titled “The Draft FY10 IS Clinical Systems Plan” regarding Beth Israel Deaconess Medical Center last week. CIOs at health systems always have a long list of projects and a finite amount of resources. It was sobering to read that BIDMC won’t have a big bull’s eye on data warehousing, however there was funding for decision support in the plan for the following areas:

• Implement Performance Manager reports and dashboards as needed to support
  organizational needs.
  

  Implement clinical data marts as needed to enable quality measurement, pay
  for performance goals, and other decision support needs.
  

  Enhance the Community Provider Index to better support Health Information
  Exchange via NEHEN gateways.
  

  Implement enhancements to the Patient Activity Profile to support enhanced
  reviews required by JCAHO.
  

  Enhance SOAR (Accounts Receivable workflow) to support denial tracking and
  appeals workflow
  

  Explore the introduction of new Business Intelligence tools as funding
  permits
  

  Support Cactus and NEHEN Express users

Dan Housman
Managing Director, Analytical Applications

Labels: Quality Reporting

Marcia Gulesian wrote a blog post titled “Functional Design of an Ontology” which provides a helpful introduction and analysis of i2b2.

Dan Housman
Managing Director, Analytical Applications

Labels: i2b2

There is an interesting blog post about healthcare data warehousing titled “What do copy editors and data miners have in common?” by Blake Zenger. His comparison of data miners and warehouse developers to the relationship between editors and authors can help folks improve their understanding of the roles on a clinical data warehouse team.

Dan Housman
Managing Director, Analytical Applications

Labels: Data Warehousing

It is relatively easy to analyze healthcare data for activities that have occurred, but difficult to confirm when an activity has not. If a patient is diagnosed with a disease, we know about it because it is indicated in their medical record. However, a medical record without a diagnosis does not guarantee that the patient is free of the disease. Many of the activities that are important to researchers relate to things that didn’t occur, or are at least closely related to the “dark matter” in the data. A vast amount of interesting information is hidden in the murky area of missing data.

A recent example involved the analysis of a population of diabetes patients. We segmented the data set into periods to identify patients who either performed or did not perform an HbA1c test every six months. The algorithm identified a surprisingly large population of non-performers from the 2008 data; approximately three times larger than expected. The query attempted to pull all of the patients in the database who met the criteria for our definition of diabetes. This definition, for simplicity sake, meant that the patient was diagnosed at some point and was currently alive. This definition was flawed; just having data about a diagnosis did not guarantee that we also had data about the non-event of missing an HbA1c test within the last six months.

The first problem we identified was a substantial number of patients from the population of about 100,000 stopped receiving care in the health system after they were diagnosed within a 15-year period of data. This meant that the non-event of skipping an HbA1c required validation against a second non-event of not receiving any care at all. It was relatively easy to resolve this problem by eliminating patients from the query who had no facts in the data warehouse for the previous five years. Patients with a chronic disease should have had healthcare visits within this time frame, however the decision was arbitrary. Had we instead chosen to ignore patients without a visit during just one previous year, our results might have been suspicious.

One consideration was to analyze enrollment data from insurance payer files, however it wasn’t readily accessible from the EMR, and the data was incomplete given that it omitted patients without insurance coverage from major commercial payers. The insurance payer files would have had PCP data, but likely nothing about patients who stopped receiving care from one health system while having a PCP in another.

There was also a significant challenge to capture the patients that would come in and out of the system only when they were extremely sick. Those patients deemed themselves to be healthy enough to avoid having a doctor’s visit or a lab test for a couple of years, but then suddenly showed up for one and then disappeared.

Another non-event occurred among patients who frequently relocated. There was a field in the data warehouse to track each address a patient had in their history, however it came from EMPI registration data, and most patients did not update their registration, especially when departing the system.

Despite having a field for vital status in their records, we could barely verify with certainty whether a patient was alive or not. While a morbid thought--at what visit to the physician was a patient to report if they had passed? Death was only consistently verifiable when it occurred in a hospital or if information was received from social security records. Inevitably there were families with a financial motive to conceal mortality because social security checks arrived by mail for living elderly citizens with limited assets.

Regardless of the challenges, these were the patients that needed to be identified for important testing and monitoring of their chronic disease state. How were we to know who was a patient when they were not consistently in the health system as facts? We had to identify patterns among both the visits that occurred as well as the non-events. These were linked pieces of information operating in opposite worlds. A lack of visits could be dispelled by a visit. The patterns were not universal for interpreting quality. A patient with a certain age, gender, health, race, and economic background could be expected to have a very different level of activity in the system, thus a universal rule couldn’t apply to all patients and all measures of performance.

The nature of clinical intelligence work often deals with unknown non-events that are never registered as facts in data sets. This creates doubt about the quality of the analyses and recommendations.

This is just a few of many reasons for chart abstraction in core measures. When structured medical record fields are missing event data to comply with a measure, it doesn’t guarantee that the event never occurred nor does it eliminate the possibility of it being documented in a clinical note. The only way to eliminate doubt from a small set of cases is to analyze the full chart of each patient according to CMS guidelines.

I encounter instances of non-events on a daily basis. In many cases, it is a factor of whether or not we have up-to-date data. It can often take up to 30 days to receive information from a source system through a monthly load. Many source systems aren’t capable of providing instant data feeds for analysis or it is too cost prohibitive to do so. This requires us to run reports against information that is near real-time, meaning it is not up-to-the-minute but relevant for the state. An example would be to determine whether or not a patient received a lab test that was ordered two months ago. We are unable to provide 100% confidence for non-events without up-to-the-minute data. However, if we have a record of a test being performed, we are 100% certain that it occurred. In many instances, we need to account for latency in the reporting and provide tolerances for the lack of information. Grace periods are required to account for these uncertainties.

While it is challenging enough to access the data we have, it is critical to think carefully about the data we don’t, and the impact it has on our results. Maybe we need more data? I certainly find it odd that in a world where many people tweet ten times a day, we have tremendous uncertainty about individual health.

Dan Housman
Managing Director, Analytical Applications

Labels: Clinical Intelligence

Dr. Charles M. Watts, senior vice president of medical affairs at Northwestern Memorial Foundation, presented an overview of Karl Weick’s high-reliability organization principles in relation to healthcare. Among his key points was an importance to focus on failure in order to ensure safety. He stated that "Chronic wariness is the tone in a safe environment...hubris is the enemy."

Dr. Watts simplified data warehouse quality as "data quality equals completeness multiplied by validity." He provided an example of an average newborn baby weighing 32 kilograms or about 71 lbs in a healthcare system. The occurrence was attributed to inconsistent data entry with staff using kilograms and grams interchangeably. The solution was to transform the data into one unit of measurement and to ultimately correct the consistency in data entry.

Dr. Watts also demonstrated two cases of applying improvement to increase safety and reliability. The first case involved shoulder dystocia, a tremendous liability risk that occurs when a newborn is stuck in the birth canal. By instituting a simulator, a standard protocol, and a training program, the existing $20 million annual liability was successfully eliminated. The second case involved a decrease in severe adverse events even though the total number of reported adverse events actually increased. This was attributed to improvements in both safety and visibility. Dr. Watts stated "I don't think mistakes went up--reporting of mistakes became more acceptable and we should celebrate that."

Deb Batson, clinical research data warehouse architect at Children’s Hospital Denver, mentioned an example of finding married 6-year-olds. This mistake was attributed to a registration system that was prone to data entry error during hospital admissions.

The reports from Memorial Sloan-Kettering, National Institutes of Mental Health, Duke, and Ottawa differed tremendously. It would be helpful to find a better way to execute technology transfer of reports between organizations.

The folks at Intermountain built an amazing tool for improving labor costs using Cognos as well as a meta-report search engine. The engine allowed users to browse and launch reports from multiple BI tools using just one portal. It appeared to be a good solution for groups with more than one reporting tool.

Dan Housman
Managing Director, Analytical Applications

Labels: Data Warehousing

There were a handful of interesting presentations from Northwestern Medical Faculty Foundation (NMFF), Ottawa Hospital, Duke, and MD Anderson on the first day of the HDWA conference.

NMFF presented an open source SQL server integration services extension for regular expression extractions from free text, a de-identification utility, and TaskMaster, a system for processing ad-hoc data requests. One of the clever features in their workflow management approach was an integration that pulls data from the eIRB into the data warehouse to display the details of the eIRB process. It also had the capability to link out from the tool to execute tasks such as creating a report from a SQL query. Another interesting component of their model was the use of distributed analysts to query the database. The analysts operated within their own groups, but this required segmented hospital data to prevent inappropriate queries.

The Ottawa Hospital presented a dashboard view of their hospital-acquired infections graphically overlaid onto the hospital floor plan to identify infection hotspots.

Duke presented a poster on an open source framework that supports patient recruitment for clinical trials using Mirth.

MD Anderson created a new group called the Office of Performance Improvement. The organization utilized Minitab, QI Analyst, and ultimately Statit to effectively generate control charts in order to rectify challenges found among common BI tools such as Cognos and Business Objects. One of the challenges was an inflated length of stay measurement from last year. This was due to Hurricane Ike, as it was inappropriate to discharge patients in the midst of 110 MPH winds.

One of the common trends among the HDWA presentations was an initial model for data warehouses and delivery systems that provided free access in order to drive adoption, but eventually transitioned to a fee-for-service model for sustainability. Only a few of the organizations were successful thus far in making that transition.

Dan Housman
Managing Director, Analytical Applications

Labels: Data Warehousing

Imagine for a moment that the government issued a mandate that required consumers to pay a significant portion (75% or higher) of all healthcare costs out-of-pocket.

This is a serious reality for many tourists who arrive in the United States from foreign countries without insurance, but with a reasonable bank account, and a need for care.

When this occurs, a provider must not only explain the options available, but also the cost of each to the patient. Procedures are then chosen based on a balance between cost and a desire for the highest quality outcome.

Among the challenges providers struggle with is how to explain the cost of a procedure in marketing materials or during a consultation visit. This isn't quite as simple as one might think, e.g. determining the cost of a colonoscopy. The total cost to the patient has a real chance of including the removal of polyps and a need for a secondary colonoscopy. The patient visit also includes underlying services beyond the core procedure; and the cost of these additional activities must be accounted for.

This situation warrants the need for procedure pricing that is averaged for high and low costs based upon potential episodes. Packages need to be established with pricing and options clearly communicated to allow for consumer choice. A handful of health systems are now attempting to cater to this market using simple analytics to support pricing models.

In contrast, the decision-making of an insured patient differs tremendously. They either do not worry about the cost of a procedure or they choose the opposite of the lowest cost. And why not? If a patient is fully covered for a colonoscopy with the option of either a $3,000 or an $8,000 procedure, they will likely choose the more expensive option under the assumption that it is superior care.

At the moment I am scheduling my own dental care for 4 crowns to minimize out-of-pocket expenses. My dental insurance plan has a maximum annual coverage limit, thus I make medical decisions based upon what I must pay. It appears that dental insurance is more efficient than health insurance in this way. It encourages consumer behavior, but generally covers maintenance and prevention activities.

Now imagine that competition on price became a leading healthcare issue, created by consumers forced into out-of-pocket buying decisions. Perhaps a price and quality comparison website would appear in the market. Prospective patients would select from a list of procedures and then nearby providers would list their pricing. Some providers would likely advertise to get to the top of the list, while others would simply rise to the top through competition. Patients might even rate or comment on the outcome of a previous procedure using such a tool.

Medicine at times has taken a track opposite to the one governed by Moore’s law in the semiconductor industry. The evolution of new technologies in medicine serves to increase cost. No one that I am familiar with is working on the assumption that we will reduce the cost of colonoscopies by 50% every 5 years.

How do we encourage patients to act as consumers? Consumers are needed to improve quality and minimize costs.

One place to start is to evaluate quality through patient reported outcomes and let the mix between insurers and providers figure out how it drives the market. Perhaps insurers should reward the healthy with financial incentives.

It appears that capitalism has failed, but I don't know why.

I’ve read a handful of research studies that mentioned inactive patients were uncomfortable with consumer-driven insurance spending because it required effort on the patients’ behalf to reap the benefits. This suggests that such a plan would encourage patients to work toward becoming healthy. Could these plans become more cost-effective for employers over time? Will self-insured employers or HMOs--folks who bear the full cost of the risk do some pioneering in this area?

Dan Housman
Managing Director, Analytical Applications
Recombinant Data Corp.

Labels: Quality Reporting

I’m not an iPhone user, but I’d like to be one today. HealthMap, an application created by Clark Freifeld and John Brownstein from the Children’s Hospital Informatics Program, uses GPS to determine when a disease outbreak occurs near an iPhone user; and thus alerts them of the danger.

I’d like to see HealthMap ported to BlackBerry devices- that way I too could avoid the next dinner party in a suburb with a major H1N1 outbreak, or drive around an Ebola or Marburg traffic jam!

HealthMap is a free download from the iTunes store.

Dan Housman
Managing Director, Analytical Applications
Recombinant Data Corp.