Scott Weingarten, MD, MPH is president and CEO of Zynx Health.
What was your reaction when you saw the proposed Meaningful Use criteria?
When you look at what was published on December 30, I don’t think there were any huge surprises based on what we saw over the summer that was published in June and July. There was a continued focus on clinical decision support, which we anticipated based on the earlier information. We believe that that is a good thing.
We think clinical decision support can lead to improved quality and safety of care, as well as less costly care. We believe that’s where the greatest benefit will occur as a result of the investment in healthcare information technology. We were pleased to see that clinical decision support remain prominently featured in the December 30 publications.
Do you think the requirement to create five clinical decision support rules is a good step toward using systems to guide physicians?
I do. Initially, I believe it was one. Now it’s five. I think the bar is still very low for clinical support rules. I think in order to get the clinical return on investment and the cost benefits that the federal government would like to achieve, I think they’re going to need more than five, but I think it’s a good start.
To many doctors, when they hear clinical decision support, their experience or their perception is that it’s just intrusive, unhelpful, impersonal warnings that stop them from doing what they want to do. What has to be done to turn that into that vision of having it impact quality and cost?
They have to look carefully at both the sensitivity and the specificity of the clinical decision support, meaning, look at very carefully that there are as few false positives as possible. I think at least historically, largely with drug/drug interactions, there are many false positive alerts that in many cases can color our view about the value of clinical decision support.
When one goes beyond drug/drug alerts and really looks at the broader potential for clinical decision support, and if one really focuses on those aspects of care that provide great clinical benefit… For example, evidence-based clinical processes that have been shown to reduce mortality, reduce morbidity, and improve quality of life or safely reduce costs when one pays careful attention to the specificity, or ensuring that there are as few false positives as possible. I think the annoyance factor will go down and I think the benefits will increase.
Do you think it’s the content provider or the application vendor who needs to refine that sensitivity/specificity and the ultimate presentation of whatever the result was?
I think that it’s really teamwork. When I say teamwork or collaboration, I think the content vendor needs to really give a lot of thought to optimizing the sensitivity and specificity of the clinical decision support. I think the healthcare information technology supplier needs to have the functionality to optimize the specificity and sensitivity of clinical decisions.
Also the client, in some cases, can pick and choose which components of clinical decisions support that they would like to utilize. Thinking about the benefits, or them really analyzing the potential benefits of turning on clinical decision support, should occur before they select what form of clinical decision support.
A good example, as you mentioned earlier, would be the five rules. Making sure that the five rules are those that really will favorably impact care at their organization, have the greatest clinical benefit, and yet the rules will be as specific as possible.
One of the things that it seemed was fairly clear in the initial proposed criteria was that the rules needed to be user-maintainable rather than just a black box that you take as they come. Was that a surprise?
No, I think you want them to be maintained by the user because I think that different organizations, depending on local practice, there’ll be some rules that provide greater benefit than at other organizations.
Let me give you an example of what I mean. Let’s say an organization has already achieved the ceiling effect and eligible patients with chronic heart failure are being treated with either an ACE inhibitor or an ARB. Well, having a rule will provide very little benefit, just because the care is consistent with best practice or evidence-based practice. Another organization where they have not achieved those benefits, where far fewer appropriate patients are treated with ACE inhibitors or ARBs for chronic heart failure, may have an opportunity to save many lives by providing that rule.
Really, the point being that different organizations will achieve different benefits with different types of clinical decision support, depending on the size of the gap between optimal or evidence-based practice in their current practice. Practice varies, as I think has been very well described, organization to organization. Therefore, I think having the user select which clinical decision support rules have the potential to provide the greatest benefit for their organization, and potentially maintaining the information, to me, makes sense.
When hospital-based vs. practice-based doctors create their initial five rules, how will their priorities differ?
I think that the types of rules that are likely to be created in the ambulatory setting or by physicians in their offices will reflect the patterns of outpatient care. My guess is we will see a number of alerts and rules for chronic illness in the form of disease management rules or preventive care; where I think the rules will be quite different in the hospital, which will reflect acute illness requiring hospitalization.
My guess is in the ambulatory setting, we might see more rules related to chronic illnesses such as asthma or diabetes. In the hospital, we might see more rules and alerts that relate to the more common reasons why patients hospitalize, such as chronic heart failure or community acquired pneumonia.
What do you think about rules for nurses when charting or documenting?
There’s a fair amount of nursing practice that can be evidence-based. There are many good studies showing that certain nursing practices, when faithfully adhered to, will lead to better patient outcomes than other nursing practices.
Decades of nursing research support what processes are best for patients, and I think it makes a lot of sense to have rules and alerts to inform nurses, when appropriate. That will lead to the best possible nursing care. I would agree with that. I think alerts and rules are just as important for nurses as they are for physicians and other healthcare providers.
What kind of work has been done, or what kind of interest is there in background alerts based on collecting electronic data that indicate problems, the imminent harm type of rule?
My guess is that the initial rules may not be quite that sophisticated, but I think there’s tremendous benefit that can occur when these background rules — for example, will identify patients based on physiologic parameters, hemodynamic variables where they’re at risk of rapid clinical deterioration — to alert the physician to observe the patient closely and possibly prescribe new treatments.
I think that when we get there — when the field matures, when we’re consistently applying those rules — I think there’s great potential to quickly identify problems that may not have been identified by treating clinicians and to reduce morbidity and mortality. I think in the not-so-distant future there will be many more of those rules, and that will advance patient care significantly.
What advice would you give to hospitals, in general, about creation and maintenance of order sets?
I think it’s hard. One is I would advise them to create, update, and maintain order sets. There are good data in the scientific, peer-reviewed literature that shows that evidence-based order sets reduce mortality, reduce morbidity, and can safely reduce costs. I would advise hospitals to do it.
Second, is there are data showing that physician productivity can be increased when physicians use order sets for common diagnosis rather than write each order one by one. There’s the benefit of improving care and potentially improving productivity, but I think you need the order sets to be viewed as credible by the medical and nursing staff.
They have to be updated frequently. I think they have to be evidence-based. I think if they are not maintained and the information is highly perishable, that when the information goes out of date clinicians are smart and figure it out. They say, “Hey, how come this order set doesn’t reflect the findings in this article published in the New England Journal of Medicine a month ago? What’s going on here? Don’t we want to provide the best possible care to our patients?”
Clinicians, appropriately, can be critical when information in order sets is out of date. The organization needs to come up with a very methodical, disciplined approach to update and maintain the order sets and retain credibility with the clinicians.
There was a statement recently, by Eric Schmidt of Google, that seemed to imply that in his mind as a technologist, practice of medicine is simple as looking up reference information, correlating it to patient information, and out pops a diagnosis and a treatment plan. How can the art of medicine be reconciled with the support that software can provide to those who actually practice it?
I’m an internist. I’ve taken care of a number of patients in my life. Medical care is complicated. I think that it’s complicated for a number of reasons.
In some cases, there may not be evidence to support a particular treatment decision. In some cases, the evidence might be conflicting. In other cases, the patient may have many co-morbid illnesses.
It’s often not as simple as a patient that has one illness and therefore, this recommendation will always work for this patient. Many patients who are hospitalized have many different co-morbid illnesses which increase the complexity of clinical decision support.
Then finally, patient preferences are very important. I may suggest to a patient that there’s a particular drug I would like the patient to take, but the patient may have had a bad experience with the drug in the past, or may know someone with a very bad experience from that drug in the past. Therefore, for that patient, that drug may not be the most appropriate.
I think medicine is highly complex. Evidence-based information is critically important to informed care; but at the end of the day, what evidence-based medicine does is inform the best possible care. Each doctor and nurse has to understand the preferences and beliefs of his or her patients to make sure that the care is individualized to lead to the best possible care for any individual patient.
Has the industry moved enough toward guiding the caregiver, rather than warning them of conditions? In other words, helping them make a decision, rather than telling that they’ve made the wrong one?
I think so. I think order sets will help guide clinicians to making the right decision. They really do not tell clinicians they’ve made the wrong decision, so I think as an industry we’re heading in the right direction. I think that’s exactly what you want to do. You want to guide them to make the right decisions, rather than you made the wrong decision.