Information is power, but sometimes it can be too much of a good thing. Information overload affects workers in every industry – this is particularly true for data-reliant and intensive industries like healthcare. And, it’s not getting any better – by 2020 the healthcare industry alone is estimated to have 25,000 petabytes of data – that’s equivalent to 500 billion four-drawer filing cabinets. With its complexity and breadth, and not least with its impact on our lives, healthcare has to be the poster child for how efficient management of data can improve productivity and help providers make better, more informed decisions.
A first trick, however, is getting at the relevant pieces of information. Today, a lot of manual work goes into accessing and cleaning up data that is siloed and unstructured. Too much information is still on paper, but even where it has been migrated to electronic medical records (EMR), practice management systems, or lab diagnostic systems, much of it is still unstructured. The majority of hospitals are finally implementing EMRs, and many of us are working to provide advanced analytics based on the information going online, but, at a recent Xerox healthcare client council meeting, several CIOs emphasized that there still remains a huge challenge in cleaning, standardizing, and integrating data before it can be used for decision making.
Fortunately, powerful methods are becoming available that can extract relevant events from physician narratives, intelligently aggregate data, customize information for clinicians based on context, and visualize information. For instance, in France, a number of hospitals are testing an emerging application based on Natural Language Processing technology developed at the Xerox Research Centre Europe in Grenoble, France. Researchers designed the solution to help prevent the spread of hospital-acquired infections by finding, extracting, and combining key information in physician narratives distributed in medical records. As another example, our Midas+ Live product accesses and integrates information from diverse hospital systems and puts them on a single dashboard, multiple patients at a time, hugely simplifying a physician’s task to monitor all of his or her patients.
Another challenge is using this abundance of information to make better decisions. Despite selectively whittling down the field to the most pertinent pieces, the amount can still be overwhelming. This is a problem for clinical staff, and we are working on decision support systems for experts, but here I’d like to point to another area, an area where most of the cost in our healthcare system goes: patients with chronic diseases, in particular the elderly and disabled. As relatives and friends, we are increasingly asked to help in the role of caregivers, but the available information does not often help us to make decisions and take actions in that role. In a recent project, we have been looking at the use of social networking to bring in the brain power of volunteers. With this approach, we can harness the expertise and experience of volunteers to personalize the available information to participants’ needs. With a focus on long-term care, and working closely with healthcare providers and government agencies, we’re working to bring this technology to care at home, helping patients to manage their conditions and to closely adhere to care programs in the face of an often bewildering array of medications and care needs.
Another problem Xerox helps partners tackle is the automation of non-core tasks. For example, more than half of a nurse’s work goes toward documentation and task coordination. While important, those don’t add to and often detract from the nurse’s core work. We are developing a solution that uses situational awareness and mobile technology to deliver and update pertinent information, almost as a side effect of the nurse’s work. This technology is still being developed in the lab, but it has the potential to greatly simplify typical workflows that are adjacent to the central task of caring for patients.
Finally, we should remember that optimizing an old process is not always the best solution. We should also rethink healthcare processes that are outdated in the face of new technologies and business models. For example, while we have been automating the healthcare audit business for years, Xerox is also working to provide real-time (pre-pay) fraud detection to payers. There are a number of critical back-office processes that are undergoing similar transformations.
These are just a few ways Xerox is going to task on simplifying healthcare. What are some ways you would suggest?
This article brought to you by Xerox Corporation.
Markus Fromherz is the chief innovation officer of healthcare for Xerox Corporation.
Categories: Uncategorized
WOW just what I was searching for. Came here by searching for ford
transit connect refrigerated van
My web site Inexpensive second hand refrigerated automotive dealership
AHJ is seeking for partner sites in the health care niche. American Health Journal is a medicine content site which owns a significant library of high quality health care videos. We are seeking blog owners who are interested in contributing guest articles to our website. Contact us at our contact page on our website.
Managing data to care for patients is a total farce. The doctor creates streams of data, usually festooned with errors carried forward from earlier entries that purpetuate errors so as to maximize payment.
One of my partners was a master of the short note, until you needed all the garbage attached for a level 5 E&M. Every word meant something about the patient’s care. Who cares if we regurgitate for the umpteenth time the appendix is out and we counseled them against smoking. None of that has a thing to do with care.
Our hospitalists uae a template based record for all H&P’s, daily progressnotes and discharge summaries. They are totally useless. In the ED we dictate a narrative. We say what we think is wrong, not what fits a template.
Why be enslaved to dictation machines?
See Lawrence and Lincoln Weed’s “Medicine in Denial.”
___
“Essential to health care reform are two elements: standards of care for managing clinical information (analogous to accounting standards for managing financial information), and electronic tools designed to implement those standards. Both elements are external to the physician’s mind. Although in large part already developed, these elements are virtually absent from health care. Without these elements, the physician continues to be relied upon as a repository of knowledge and a vehicle for information processing. The resulting disorder blocks health information technology from realizing its enormous potential, and deprives health care reform of an essential foundation. In contrast, standards and tools designed to integrate detailed patient data with comprehensive medical knowledge make it possible to define the data and knowledge taken into account for decision making. Similarly, standards for organizing patient data over time in medical records make it possible to trace connections among the data collected, the patient’s problems, the practitioner’s assessments, the actions taken, the patient’s progress, the patient’s behaviors and ultimate outcomes. Two basic standards of care, and corresponding tools, bring order and transparency to medical decision making:
• First, from the outset of care, relevant patient data must be chosen, and its implications determined, based on the best available medical knowledge, independent of the limited personal knowledge of the practitioners involved. Patient data must be systematically linked to medical knowledge in a combinatorial manner, before the exercise of clinical judgment, using information tools to elicit all possibilities relevant to the problem situation, while defining and documenting the information taken into account. Practitioners’ clinical judgments may add to, but must not subtract from, high standards of accuracy, completeness and objectivity for that information.
• Second, in complex cases, particularly in cases of chronic disease, the organization of data in medical records must be optimized for managing multiple problems over time. This means that each medical record must begin with a complete list of carefully defined patient problems, and that other clinical information in the record must be linked to the problem or problems to which it relates. Without that structure for the medical record, decisions are made out of context, follow-up and coordination of care are haphazard, and records are not usable for rigorous clinical research.
With these two basic standards of care, and the information tools needed to implement them, practitioners and patients can manage the flood of detailed information required for sound decision making over time. With this detailed information, made usable for research in structured electronic medical records, medical care can become increasingly refined and individualized. In contrast, so-called “evidence-based medicine” is derived from large population studies that fail to account for the medical uniqueness of each patient.
Enforcing the necessary standards and tools depends on changing medicine’s culture of professional autonomy for highly educated physicians. Indeed, the concept of a physician as we know it is not viable. All practitioners must submit to meticulous definition and control of their inputs to care (a principle recognized by the patient safety movement). The primary barrier to this cultural change is graduate medical education and credentialing. These social institutions (1) fail to define, disseminate
and enforce high standards of quality for provider inputs to care, (2) inhibit effective design and use of information technology to manage clinical information, and (3) suppress competition among providers who might otherwise exploit information technology to generate remarkable advances in patient care and medical knowledge.”
Unstructured data is a big part of the problem. That arises from the most relevant data being entered in narrative form. I did an EMR experiment on a medically complex patient and read all 640 physician notes in the file on my day off. It took me about 5 hours but lead to 5 additional relevant diagnoses and it clarified a critical path about how some of the medications were changed. The signal to noise is very poor and there are petabytes of data that are useless templates and scripted notes to meet coding and reimbursement requirements. That entire process needs to be revolutionized.
There also needs to be greater use of relevant graphics. Large amounts of information can be put into easily interpreted graphical form. Would you rather take a painstaking history of episodes of atrial fibrillation or be presented with this graphic before seeing the patient?
http://gdpsychtech.blogspot.com/2012/03/timelines.html
Lots of room for improvement. Any idea about how much of that 25 EB of data is actually useful information?