PDF File: 

HCT3_wp_gressens.pdf

SAS

Business intelligence software helps healthcare organizations to integrate clinical data from multiplesources and allows that information to be easily viewed, shared and analyzed.

What Exactly Is an Electronic Health Record?

The National Health Information Infrastructure initiative targets that every American should have an electronic health record (EHR) within 10 years. But what an EHR actually is, and what that means for healthcare organizations, remains unclear. Most commonly an EHR is thought of as an electronic tool used by physicians and clinicians to record their notes, order tests and drugs and store their patients’ information. Most modern EHR applications do all those things and much more, but the use of computing tools by the clinician while the patient is in the exam room is only part of the whole EHR picture.

It’s actually more accurate to say that patients don’t have one record, they have several. These records comprise a mix of physical (paper notes, X-ray film) and electronic files (images, digital wave outputs, text notes). There are varying amounts of administrative, financial and clinical data specific to each patient on each of these records and, of course, these data sets are created by many different healthcare organizations and the patients themselves. These organizations are far from being on the same page as to how they share that information. There is universal agreement, from Newt Gingrich to Sen. Hillary Clinton, that getting all that data into an EHR accessible by physicians and patients will have two major benefits.

Two Big Benefits for an EHR

The first rationale for the EHR is to improve the quality and efficiency of the care process as patients experience it. The focus has been on getting tools and applications into the clinicians’ hands at the point of care. These tools are intended to deliver the patients’ medical history, medications, lab results and other pertinent information to the clinician, while at the same time being able to record the clinicians’ notes, take their orders for lab tests, diagnostic imaging, prescriptions and referrals to other clinicians. In addition the tools are often designed to suggest care protocols to the clinician while warning against potential hazards such as medication errors. Clinician use of the EHR should make the patient’s experience safer, more efficient and more effective.

The second reason for an EHR is to analyze overall patterns of care within a population, using the data recorded in the data repository that is the foundation of an EHR. Those interested in the population level include health plan and hospital medical directors checking whether their patients are being treated with the most appropriate care guidelines, private and government payers looking at high-cost “outliers” or the cost-effectiveness of treatments and clinical researchers assessing post-approval drug safety. As more clinical data is captured electronically and analyzed, more innovation will occur in how it’s used.

So Why Care About Business Intelligence?

Two trends are intersecting that increase the potential for this analysis while healthcare organizations struggle to reduce costs and improve performance. The first is that even though clinician adoption of the EHR is only increasing slowly, more and more clinical information is being transmitted and captured electronically. That information includes not only basic but useful diagnosis and procedure information on claims, but also lab results, radiology reports, medication histories and clinical notes transcribed into XML or other searchable formats. The second is that the tools with which significant value can be extracted from this data are becoming more powerful and easier to use. In fact it doesn’t really matter whether a clinician is using an EHR at the point of care; there is now enough data available for medical directors, researchers and clinicians to make meaningful interventions in the overall management of the care process, including in the care of individual patients, without having to wait for days or weeks for special IT reports.What they need is the ability to easily collate and manipulate this data from multiple sources: This is the lair of an emerging application called business intelligence (BI).

Business Intelligence Using the EHR

There are many potential uses of BI using clinical data, often created using the EHR and stored in clinical data repositories. Below are some brief examples of how the use of BI has changed clinical care processes, and has improved patient care as well as financial performance.

Inpatient care: Hospital profit margins are extremely sensitive to variations in inpatient care such as patient length of stay, staff and equipment use. Yet a recent study published in Health Affairs showed significant variation in resource use in the care of similar patients even between prestigious academic medical centers[1]. The information collected in EHRs can be used to determine the resources used in patient care. Using a data mart storing information from both the EHR and various departmental clinical and financial systems, BI systems allow the assessment of the use of supplies, staff and procedures, and equate that back to decisions made by individual clinicians. The analysis can drill down to individual patients, such as types of supplies used in surgery or imaging, and cross-match that with length of stay, frequency of re-admissions, outcomes and overall costs — all of which might have significant impact on overall profit margins per admission, as well as being a key to improving patient outcomes. That information can then be communicated back to clinicians.

Population care management: Gathering information from many divergent clinical and financial databases, health plan medical directors have been attempting to profile the overall care delivered by contracted physicians to their populations for over a decade. But the effectiveness has been limited in part because the data used has been out of date and difficult to analyze and communicate. Despite these limitations, looking at macro-level claims data has enabled the discovery of outliers that may have serious consequences for payers, providers and patients. For example, two physicians were responsible for a dramatic increase in cardiac procedures in Redding, Calif., in the late 1990s — clearly visible in the Dartmouth Health Atlas analysis of surgery rates. This had severe consequences for the hospital’s parent, Tenet Healthcare. But that information was hard to find. BI promises to deliver this type of analysis much closer to real time, and make it far more available to analysts within hospitals, health plans and government agencies responsible for looking at physician practice patterns, population health or even fraud.

BI is also useful for direct care management. For example, Active Health Management (recently acquired by Aetna) has been using close to real-time analysis of medical, pharmaceutical and lab claims to directly contact doctors and patients if their system flags a potentially adverse event. Similarly, BCBS of Kansas City has been using BI to determine whether their members with diabetes or heart conditions are seeing their doctors on a regular basis, receiving recommended tests or filling the correct prescriptions. In these cases, contacting the doctors and patients concerned allows for better and less expensive care.

As more clinical data is collected via EHRs and various other clinical and financial databases, the use of BI for this analysis and consequent intervention in care management will become more common and more effective. Kaiser Permanente is making a huge investment in its HealthConnect EHR and associated data warehouse and BI tools. John Mattison, CMIO of Kaiser Permanente of Southern California, described the goal of the BI function as giving the ability to perform analysis to anyone, given their role and appropriate security level, in order to come up with more creative ways to answer problems. “The goal is to have best practices where they are known and create a good model to discover best practices where they are not yet known,” and then to feed that clinical content to clinicians at the point of care. Mattison believes that within five to seven years provider organizations that don’t find a way to combine the EHR with this BI and overall process improvement function will be unable to compete, as Kaiser will be able to demonstrably provide better value at a lower cost. As more end payers start to “pay for performance,” these competitive pressures will increase.

Public health and bioterrorism: As information is collected in the EHR in public health clinics and emergency rooms, clinicians and public health officials are beginning to use BI to track potential disease outbreaks. For example, physician group Emergency Medical Associates uses BI in conjunction with its encounter records to track disease outbreaks, such as TB, or possible bioterrorism attacks. Based on key words used by patients in articulating their symptoms, patients are placed into disease groupings, and the number of patients in each grouping is then compared with statistical norms. Unusual levels or trends of symptoms could indicate the early stages of an outbreak.

Patient safety. Patient safety, particularly around medication administration errors, has become a critical issue for healthcare providers. Intermountain Healthcare in Utah has been perhaps the leading proponent of collecting and analyzing data on medication use and associated adverse drug events. Using the connection between BI and their EHR systems, Intermountain has created a robust and complete error reporting system, and after feeding that back to clinicians, it has seen reductions in error rates that significantly exceed comparable provider systems. This has a direct impact on patient safety, patient outcomes and lives saved.

The Technology Behind Business Intelligence

These examples of BI use the data that’s collected as part of the wider EHR in the patient’s journey around the healthcare system. Even without direct clinician use of an EHR, a significant amount of clinical data is available already in electronic format, albeit in different silos. The goal of BI is to access information, regardless of where it resides, giving all the relevant decision makers a single view of the true data. A good BI application performs several related tasks:

Data integration — BI applications extract, transfer and load from multiple transaction database types into special databases called data marts.

Data integrity — Sophisticated BI applications have internal controls to check that the data integrated is correct and conforms to proper business rules.

Prebuilt datamarts and reports — Modern BI applications provide prebuilt data marts and “canned” analysis, often presented in a “dashboard.” This gives immediate value to users, without requiring a learning curve or dedicated IT resources.

The Impact of Business Intelligence In the EHR World

As EHRs help create much more clinical, financial and administrative information about patients, BI will change how healthcare organizations use their data.While some effects are obvious, other innovations are as yet unknowable. But they will share common threads.

The first will be increased speed of reporting. Frequently the data in transaction systems can only be reported on after several days or weeks, requiring extensive intervention from IT staff. BI allows that data to become actionable information within days or hours.

The second is “democratization” of critical information within organizations, achieved by the much wider distribution of BI tools and data, and consequently a greater ability for different teams to analyze key business and clinical processes. This data can also be shared securely online between different stakeholders such as hospitals and affiliated physicians, health plans and even patients. Overall the increased use of this information will enable more sophisticated cross-stakeholder collaboration for processes like disease management.

The third thread is the increased ability for critical decisions to be made by operational and strategic executives, without needing special IT or analysis skills. BI applications will increasingly appear on the executive information system dashboard. For healthcare organizations this will allow greater oversight of financial, clinical and operational issues.

The overall implication is that the management of healthcare organizations including hospitals, large physician organizations and private and government payers will be able to examine their clinical and business processes much more closely and see changes in those processes much more quickly. BI will enable organizations to view the types of activity and innovation that are valuable from both a clinical and cost-effective perspective in the complex process of delivering patient care.

Endnote

1. Fisher, et al.; “Variations In The Longitudinal Efficiency Of Academic Medical Centers,” Health Affairs Web exclusive, October 2004. http://content.healthaffairs.org/cgi/content/ abstract/hlthaff.var.19v1.