American physicians are puzzled by why the push towards electronic medical records (EMR) has failed. Millions of dollars were spent by multiple administrations in Washington, yet the quality of care hasn't improved. 250,000 patients continue to die each year due to patient error. Some hospitals spend as much as $10 million per year just in software licensing. The cost is always justified because the promise is increased revenues with a lower cost (more efficiency and lower re-admissions).
Predominantly, the systems remain "billing systems" and not decision support systems because the goal of the American healthcare system is overly focused on profit and reducing costs. (i.e. the accountants are running the hospitals). Contrast that perspective with Canada where the politicians are essentially responsible for the health of the population. So, it is fair to say that the reason that the systems have not improved is largely because we the people haven't demanded it.
The problem with healthcare can be blamed on many reasons including: incompatibility between systems, lack of standardized quality control metrics and lack of good data. In IT terms, we know that systems fail for one of two reasons: the process is poorly designed or not followed, or the IT system is dated and inflexible to respond to the customer's demands. Jenna's Law: Inflexible technical architectures cost more than flexible ones. Also, in IT, we have another rule Garbage In = Garbage Out. When the data coming into the system is essentially garbage, one can expect there to be misdiagnoses and medical errors.
Jenna's Law: Inflexible technical architectures cost more than flexible ones
Having had discussions with physicians all around the country, we found that largely all of the health data collected to date is not terribly useful with the exception of tangible data such as lab results and machine outputs. More shocking is that the patient's testimony is not part of the record. Rather, the data is interpreted by a different nurse and physician every time the patient visits the clinic or hospital. Instead of having accurate data to base a decision support system on, we have very little accurate data. To base a decision support system on this data would be a very bad idea and most likely riddled with errors and failure.
With new technologies emerging such as voice capability, we believe there are better ways to directly capture data from the patient and to make it part of their electronic record. Using this data, we think opens up a number of possibilities for decision support systems. Quality control managers may be able to assess doctors and increase the quality of their patient outcomes. In the US, physicians are largely rated on re-admission rates which doesn't paint a complete picture especially when the patient returns, but to another facility. As our Chief Data Scientist, Charu Kapil PhD, points out, there is no flag in the schema to indicate misdiagnoses. So, an EMR could include 4 clinical visits making the patient's condition seem chronic when they were simply misdiagnosed three of those four times. We conclude that lack of re-admission is not proof of quality care or a positive patient outcome.
In our company, we are actively building the data schema of the future that we believe will become the standard. From that system, we believe that we can finally achieve the type of decision support system tools that physicians and patients have been asking for.
Jenna Bourgeois is CEO of Dynamics Intelligence which specializes in the healthcare and public sector verticals. For further reading, check out these articles: DI Announces Data Science As A Service (DSAAS) Offering For Healthcare and Improving Patient Outcomes With Voice/Dynamics 365