Optimizing Healthcare Through Clinical Documentation
The Science of Medicine moves forward almost entirely on peer-reviewed research, based upon controlled clinical studies…with a little (dubious) help from anecdotal evidence. These studies are expensive, narrowly focused, and dependent upon the perspectives or whims of individual researchers.
There is a wealth of valuable data hidden in the electronic health records (EHRs) of healthcare providers across the world. These data are not being analyzed. Therefore, we are not learning from ongoing experiences. This is tragic, and it is arguably the primary reason healthcare is such a mess. Cost is out of control, and the quality is in no way commensurate with these excessive costs.
The only way to reverse the apocalyptic trend, so that we can bring about the quality and cost effectiveness that we would all like to see, is to institute Continuous Quality Improvement (CQI). This starts off by analyzing real-life clinical encounters. Figure out what are the early signs of serious diagnoses. Determine what treatments work best for various conditions and under what circumstances. Then we can add this knowledge to what medical science currently knows from controlled studies, to develop a Clinical Guidance System (CGS). The CGS would be functional at the point of care, providing diagnostic and treatment advice to caregivers – based largely upon statistical analysis of how things turned out for similar situations in the past.
The CGS is more than a clinical decision support system. It actually provides proactive advice at each decision point, with supporting data. The clinician still makes the decisions, overriding CGS guidance if the “art of medicine” so dictates. But the results of all decisions, whether in concert or contrast with the CGS, are tracked so that the decision algorithms can be refined on an ongoing basis. CQI for healthcare. It’s a pretty picture. But there are problems that must be confronted.
The first problem is that most clinical data are in the form of free text, not codified. Therefore, they are not natively ready to be analyzed or mined. But there are now Natural Language Processing (NLP) engines, which can extract clinical facts from a narrative and convert them to standardized codes. They are maybe only 90% accurate. But since the errors are mostly random, or, at least, haphazard, these data can still be effectively exploited for research purposes. Any significant findings are actually likely to be somewhat understated, as a result of the significance-dilution from unbiased errors. In any case, intriguing but surprising results can always be confirmed by controlled studies.
The second problem is confounding variables. Retrospective, uncontrolled studies like these are potentially impacted by tertiary factors that could be correlated with the data being analyzed, perhaps both the cause and effect. Statistical techniques, however, enable the control of these variables. Nonetheless, there is always the chance that some unknown variable will confound the results. Although this needs to be acknowledged, it does not substantially diminish the value of this research. We just need to do be careful about invoking the appropriate controls.
Another problem, arguably the most severe, is that outcomes are poorly documented. This absolutely needs to change. How can we improve what we do if we don’t measure the outcome of our activities? In the meantime, there are some usable documented outcomes out there; time to mortality being the cleanest. But lab results, imaging, biometric tests, pain indices, etc. also provide decent outcomes to evaluate. Once the importance of outcomes is understood, that should lead providers into more prevalent and rigorous documentation of them, even including the subjective assessments of well-being by the patients themselves.
The data and technology are out there. Let’s put them together. The time has come to make healthcare all that it can be.
About The Author
Joe Weber is Exec VP of Clinitech. He has been CEO or VP/Marketing for both health systems and biotech companies, starting his career as Associate Director of Ambulatory Care at Cook County Hospital. He has a BA in Biology from Brandeis, an MS in Biostatistics from Columbia, and an MS in Management from M.I.T.