In his inaugural address, President Obama called on the use of technology to "raise health care's quality and lower its cost." The Defence Health Services Systems' Clinical Data Mart (CDM) is one such project, delivering functionality and benefits across the  US Military Health System, which serves more than 9,200,000 people worldwide. Anecdotal evidence points to CDM as likely being the largest automated clinical reporting tool in existence. Empirical evidence proves that the CDM is being hugely effective in improving the quality, safety and efficiency of care.

To understand CDM and its effectiveness, it is necessary to understand the underlying need. The US Department of Defense captures more computable outpatient data than any other organization in the world. The data is stored in AHLTA, the Military Health System's electronic health record that provides 24/7 global visibility and access to medical professionals around the globe. For example, a doctor can see a patient in Germany and document the care delivered in AHLTA. When the patient returns to the US, doctors can leverage the records in AHLTA to continue the care. Patient data is captured in AHLTA in discrete data fields and defined by structured terms, so that the data is computable and easily queried.  In other words: a potential gold mine, indeed, the mother lode, for clinical data analysis.

In the past, the challenge was to provide analytical access to this data without consuming large amounts of processing time that would slow overall system performance. Due to AHLTA's sheer size, this is a daunting task: a single day of patient care in AHLTA equates to more than 112,000 patient encounters and more than 3,000,000 computable terms. Hence the need for a powerful clinical reporting system to better manage and leverage the available data in order to drive informed decisions, perform outcomes analysis, and measure, analyze and manage direct patient care, wellness and disease prevention.

Enter CDM: built and maintained by the Military Health System's office of Defense Health Service Systems (DHSS) and the only clinical reporting tool for AHLTA in military treatment centers today. AHLTA data has been collected for more than seven years, but the surface has only been scratched in mining that data to transform military medicine. CDM is the key that unlocks that data and turns it into transformative information capable of raising health care's quality and lowering its costs.

Providing timely, automated clinical data analysis across the Military Health Systems' almost ten million beneficiaries, CDM's potential to make a difference in peoples' lives is enormous. Just a few examples include:
	Using CDM's near real-time data, clinicians are able to quickly identify patients at risk and track chronic diseases such as asthma and diabetes.
	Practice methods are able to be systematically reviewed using the latest clinical evidence to relate tests and treatments to patient outcomes.
	Clinical blind spots can be uncovered that, if left unchecked, could result in unintentional harm to patients.

CDM's worldwide launch in February 2008 had an immediate impact on the military health community across all branches of service. For example, the Army has used CDM to identify  56,000 patients worldwide at risk for chronic kidney disease (80% had not yet been diagnosed!). The Air Force used it to report the latest corneal refractive surgery results worldwide. And the Navy used the system in a national vaccine recall, identifying at-risk patent in hours instead of weeks. 

Hence, CDM presents boundless opportunities to not just make demonstrable differences in peoples' lives, but to actually save those lives.

Using commercial off-the-shelf technologies, Defence Health Services Systems has created a successful global clinical reporting system in CDM, providing robust data warehousing and business intelligence capabilities. CDM was developed to meet all Military Health System security requirements and has been expanded from a single server system to a multi-tiered, clustered solution for performance and scalability. 

Operational requirements for the system included ad-hoc reporting of selected AHLTA data and single sign-on with AHLTA for Web-based reporting. Design goals included being as loosely coupled as possible to AHLTA, with zero impact on the performance of AHLTA's 50 terabyte clinical data repository (which store's transactional data from AHLTA). 

CDM's DBMS is Oracle 9.2, its analytic front end is Business Objects XI, and its underlying data integration platform is Informatica PowerCenter 8. Investments were made purposefully into the latest versions of such robust tools to provide headroom for future growth into additional data areas. These platforms and toolsets were also selected for their depth of capabilities. The use of scalable, standard technology platforms is congruent with best practices for business intelligence (to this end, the Defense Health Services System, which maintains more than 29  products used throughout the Military Health System, is expanding its use of the Informatica platform from the CDM implementation to become its enterprise standard for data integration).

Keeping CDM provisioned with near real-time data, the Informatica platform manages the direct and staged workflows of accessing data from AHLTA's repository and integrating it into CDM in a form optimized for reporting. This includes handling complex data transforms. A high performance data integration platform is required as the CDM  loads 27,000,000 records three times a week, and grows weekly at approximately one GB. Currently, it holds 500 GB of data, 200 GB of indices, and has 250 tables logically stored in as normalized entity relationship data model. The platform's codeless development enables the rapid implementation of changes to meet evolving reporting demands.

For reporting, CDM comprises more than 400 data elements with the ability to correlate the data across the enterprise, with drill-down from the enterprise to the patient and provider levels. Key front analytic and reporting end features include ad hoc and pre-configured reporting functionality to query essential clinical data such as reports on immunizations, diabetes, asthma, hypertension, obesity, smoking cessation, high cholesterol and more. It also provides multi-dimensional analysis, data mining and forecasting/trend analysis capabilities, and is the only tool within the Military Health System to provide this range of capabilities.

In day-to-day practice, a typical Web-based report leveraging the above technologies might involve correlating key data related to diabetes management, including vital signs, lab results and medications  with the key information all displayed on a single screen with trending graphs. Users can click the graphs to drill down to various levels of detail, while alerts can be highlighted for rapid provider review. Thus, for example, a patient can be pinpointed as overdue for important preventive tests.

Another example would be a report providing an enterprise view enabling comparison between treatment facilities. For instance, are clinicians consistently counseling members to stop smoking? 
Another enterprise example is the ability to do near real-time reporting of actual clinical diagnoses on potentially life threatening conditions.

None of this was possible prior to the launch of CDM with its current technology foundation.

Yes

Yes

The prime purpose of CDM is to enable the medical community to get out ahead of health issues through the use of analytics on timely, accurate data  including taking a longer, broader, deeper and more accurate view of health risks, responding more quickly and effectively to developing situations and trends, making better informed decisions, and developing prevention and care scenarios that reduce risks and improve outcomes.

CDM positively impacts a wide range of stakeholders including senior leadership, clinic-level providers, population health experts, military readiness personnel, business analysts and researchers. These stakeholders are empowered through access to accurate, relevant clinical data and rapid response to queries (minutes, not hours or days), and the ability to identify deficiencies for further investigation.  They are discovering previously undiscovered relationships in the Military Health Systems' data. And they are identifying populations at risk because they have not accessed care or because they are particularly exposed to risk via adverse health events. 

Users in the medical community benefit by being able to perform tasks faster, more efficiently and more effectively  and also by being able to actually expand the definition of effective care. Patients benefit from faster, safer and more effective care. And the larger tax-paying population benefits from increased value from their tax-fueled investments, and through significant cost savings through the reduction of life-support requirements.

In specific current use cases:

	The Army is using CDM to reinforce its policy for ensuring preventive care.  Many beneficiaries receive care outside of the military including procedures such as mammograms and colonoscopies.  Since most of the nation has yet to adopt an electronic record, the challenge is to capture that care electronically to guarantee accurate reporting. The Army initiated a Pneumovax vaccine incentive program to improve vaccine compliance in patients over 65.  By using AHLTA data to document preventive health counseling, the Army was able to show a 100% increase in vaccine compliance and exceeded national clinical quality benchmarks.

	The Air Force shows the power CDM for data mining, surveillance and reporting clinical outcomes.  In the military, corneal refractive surgery such as lasik is performed at regional medical centers throughout the country. Patients travel for treatment and return home for follow-up care.  The Air Force created a form in AHLTA to collect detailed post operative information to report and track surgical outcomes in CDM. The surgeons now get very detailed reports on their results allowing them to improve their standard of care.

	The Navy successfully leveraged CDM when a pediatric vaccine was recalled. In less than 3 hours, officials quickly identified nearly 5,000 patients potentially affected by the recall and notified nearly 500 patients who had received the vaccine.  A previous recall took 6 weeks to manually review the paper medical records and notify affected patients.

Quoting Cheryl Ann Kraft, Regional Program Manager for Immunizations at Portsmouth Navy Medical Center, "This is a tool that many different levels of providers and managers can use to glean information from AHLTA in an automated manner while liberating man hours back to patient care. It is a tool that facilitates provider coding buy-in because it shows providers that what they key in really makes a difference in the reports of their patient population. We can use that data to work better and to make a difference to our patients. Thanks to the reporting available in CDM, we are getting a true overview of our entire practice, not just a mere case of percentage reporting. The possibilities are endless."

Colonel Kevin Abbot, MD provides a dramatic example of CDM at work in an application that is benefiting the entire Military Health System and its beneficiaries  and with the potential to benefit the entire population. 

A widely published and highly respected kidney specialist, Dr. Abbott is a Professor of Medicine at the Uniformed Services University of the Health Sciences and serves as staff officer for the U.S. Army Office of the Surgeon General's Information Management Division. Previously he served as Chief of Nephrology and Nephrology Consultant to the Surgeon General at Walter Reed Army Medical Center.  

Says Dr. Abbott, "As a nephrologist, I know that unrecognized chronic kidney disease is a major risk factor for avoidable adverse patient events.  People with undiagnosed chronic kidney disease can be inadvertently harmed by the inappropriate use medications,  x-ray contrast material and certain invasive procedures."

Among other issues, a 20 year old with chronic kidney disease has same heart disease risk as an 80 year old. And even some over-the-counter pain medicines can cause complications.

Continues Dr. Abbott, "When I first saw the CDM prototype in 2007, I felt certain it could be used to enhance our automated reporting for chronic kidney disease."

Dr. Abbott worked with the CDM team to develop a cohort, or analysis routine, that would spot individuals across the AHLTA system at risk for chronic kidney disease (CKD). Stages of CKD are explicitly defined by a formula that includes a serum creatinine level from lab results, patient age, gender, and race to define the kidneys' estimated glomerular filtration rate, or eGFR.  

What the work revealed was that 56,000 individuals in the system were at risk, and that only roughly 20% of them were already diagnosed. Thus over 44,000 patients in the system potentially possessed undiagnosed CKD. 

Explains Abbott: "Using this CDM report, we have demonstrated to our enterprise that chronic kidney disease is under recognized and we will use this method to run another report in six to 12 months to access our improvement in tracking and identifying chronic kidney disease.  You might ask why would almost 80% of patients not be recognized or coded for CKD?  It's because out facilities did not automatically report the eGFR in a standard manner.  Our hypothesis is if we generate this number and integrate point of care alert systems we can prompt them to better recognition and coding of this condition.  We thought global and got to act local and got them to change and integrate  the lab and pathology communities with providers so that we can report this and take the appropriate action and assess our coding with eGFR reporting." 

Concludes Abbott, "These are vulnerable patients and automated clinical reporting is in its infancy. We're only at the beginning of what we can do in electronic reporting that will improve patient outcomes and help control costs.  On average, a person with chronic kidney disease who requires dialysis will spend $100,000 a year for that treatment.  Our goal is to have healthy patients and avoid unnecessary costs. Automated clinical reporting helps providers identify patients truly at risk of harm.  Rapidly reporting those results to providers truly is priceless and is another example of the value of CDM and AHLTA in improving the health of our nation."

Defense Health Services Systems believes that CDM is likely the largest and most effective automated clinical reporting system in the Military Health System, in the country, and in the world. Private sector health organizations that have seen the project have shown keen interest in having the application and many are very interested in being able to share in the data. 

The project's exceptional aspects include its scope, timeliness of data, rich functionality, and opportunities for use. 

In scope, it draws off the immense size of the underlying AHLTA system and its 50 TB transactional database, which captures data on almost 10,000,000 individuals  the largest population to be investigated by this type of automatic clinical reporting. 

Every week, AHLTA supports 2,200,000 prescriptions, 642,400 outpatient encounters, 102,900 dental procedures and 2,100 births.

In timeliness, it extracts and integrates data from AHLTA for near real-time reporting. And it is available around the clock.

In functionality, it provides a single, rich tool for measuring, monitoring analyzing and managing direct patient care, and identifying  opportunities for improvement  using actual clinical results combined with demographic data to power analysis, prove assumptions and uncover opportunities. Built on a scalable, state-of-the-art, commercial-off-the-shelf architecture for data integration and business intelligence, it is capable of supporting clinical reporting demands now and in the future.

Regarding opportunities for use, the system is virtually open ended. For example, in addition to the several use cases cited previously, CDM is being used to improve the effectiveness of smoking cessation, generating disease management reports, improving clinical documentation, and tracking medical readiness among many other uses. It supports reporting on the Joint Commission for Accreditation of Healthcare Organizations and provides credentialing, preventative care statistics, and population health research. And its uses span the Military Health System corporate hierarchy from the enterprise level down to patient-level detail, leveraging the latest automated health technology available.

CDM faced varying difficulties in all of the above areas, but the major difficulties were with resources, technology and awareness. The project at one time was under-funded and under-powered. 

In government, as in the private sector, multiple projects compete for funds. Defence Health Services Systems inherited the CDM project after it was first introduced and found to hamper the processing power of AHLTA's data repository. There was a risk of CDM not being funded for further development. Leadership needed proof that it would work and that there was broad support. Being a problem-solving organization, Defence Health Services Systems took CDM "on the road" to promote the success of the clinical reporting tool and build support. Defence Health Services Systems was able to demonstrate both, and work on the project was able to continue.

The other major hurdle was technical. An early iteration of the project ran one aging servers and was "too slow" according to early users. It would take minutes to hours to return standard queries. Funding was procured, however, to create a new, more powerful processing environment running on high-end servers, and to implement the latest version of the analytic interface. This yielded a dramatic increase in performance over what had originally been demonstrated  11 seconds versus 3 hours for certain standard queries. 

A final challenge has been awareness  simply getting more people to learn about CDM. A concerted program in this area has continued to build awareness and support across the Military Health Service. For example, as the result of a CDM presentation in April 2007, Dr. Abbott (cited above) realized that CDM would enable him to prove his surmises about CKD.

The merger of organizations to create  Defence Health Services Systems has also helped, as the new group gained some very savvy additional leadership that have been very CDM supportive. People from outside the organization that originally created CDM have come in, and instantly understood the program's value and are supporting  getting out the word.

Resistance was rolling, from proposal to almost the present. The roots of CDM extend back to 2002 and AHLTA's emergence as the military's electronic health record system. AHLTA, with its structured notes, was designed to provide beneficiary health data on demand. But it was not designed as an analytic system. Yet there was awareness of the value of mining the data in AHLTA. Initial attempts were less than satisfactory. The data repository within AHLTA was tapped and data was retrieved and formatted for reporting, but this negatively impacted the repositories transactional performance and gave reporting a bad name in some quarters. The lesson learned was to evaluate impacts, costs and benefits at all levels of the enterprise before proceeding.

Later, in 2003, an attempt was made to create a clinical data warehouse that would extract from the repository and report through a Business Objects interface. But cost/schedule were unacceptable, largely due to aging, low-end hardware donated for the proof of concept. The lesson learned was to make sure the database was fully deployed and functional before demonstrating.

Moving ahead, the project was renamed the CDM, started to leverage an Oracle database and an interface that supported ad hoc queries. Again, speed and the stability of the AHLTA repository was an issue.

By the end of 2007, with new funds and the aim of supporting Wounded Warrior reporting initiatives, new equipment was deployed that significantly improved performance. And the awareness campaign began to build steam. 

Overall lesson learned: with clinical reporting, the application can be problematic, but the capacity is priceless, much coveted, and thus well worth having and continually refining.

Exceeded

Currently, CDM is used exclusively by users of AHLTA. But automated clinical reporting is something of a holy grail across the health industry and is vital to identifying clinical quality factors and preventive health opportunities. Health organizations in the private sector that have seen CDM are enthusiastic about having a similar capability and in sharing in the data. Plans are underway to accomplish this to some degree in the future. 

We are indeed still in the infancy of automated clinical reporting. Most of the nation has yet to adopt an electronic record, much less accurate clinical reporting. But CDM and its achievements can be seen as a model for the entire global health care community, helping organizations inside and outside of government to get started on a successful track.

The uptake of CDM since its launch in its present form has been rapid and steady. Education and awareness campaigns continue, and success always breeds success. Users are constantly finding new applications for CDM. For example, the Portsmouth Navy Medical Center that uses CDM for vaccine recall management is also planning to use it to identify trends and the impact of patient outcomes. And they will soon be doing CDM reports on annual screenings for mammography and diabetes to discover how to continue improving the quality of care. 

Defense Health Services Systems develops and maintains IT products across 70+ military hospitals and 450+ clinics worldwide to support the Military Health System. Hence there are innumerable other applications and a huge population of potential users for CDM just within the military.

The clincher in any discussion of CDM tends to be: imagine having the ability to transform health data into actionable information and insights that improve care and changes the way you practice medicine. Potential users quickly grasp the transformative power of CDM when applied to their particular health care challenges, which netted down to common denominators are: improve quality, reduce costs, save lives.