Structured Data: Big Data vs. Actionable Data

Structured Data in action

Data has always been collected in healthcare. Previously most of the information was in handwritten or typed forms that necessitated extractors to go through, in a time and financially expensive process to find the data points they wanted to measure and compare. With the push for EHRs, and other electronic solutions (including mTuitive's own electronic reporting software), this data is now structured and one of the advantages is that they are accessible and able to be manipulated for whatever reports that need to be run for studies, demographic reviews, or what have you. This rush to capture everything electronically has led to the boom of "Big Data," a buzzterm that has infiltrated almost all levels of healthcare. It refers to the fact that now, in theory, all of a patient's health information is now available as individual data points that can be used for running reports, or sending between health providers, or standardizing the way people read the various forms, it's all available and it's a lot of information per patient now blown up to include the majority of the population. However, right now it's still a novelty. It's still a matter of amassing large amounts of data. But is that actually helping healthcare? Does all of this data serve a purpose or is it simply a glut of information?

There's an excellent article by Ted Quinn for MedCityNews called "The healthcare data debate: Forget big data and think actionable data." In it, Quinn asks: what's the point of collecting as much data as we are? Is it simply a preventable thing because we don't know what we'll need the data for or, worst case, is it simply collecting all of this data because we can? The conclusion that Quinn reaches is that now that we've been amassing huge amounts of data for a few years, it's time to refine the process. This is similar to the debate between interoperability and data liquidity that we visited in this blog a few months ago. In both cases there's an ostensible goal that needs to be achieved - collecting vital information, connecting different health systems to work together - but we have outgrown the baseline manner in which we have been proceeding.

Now the question is how can you turn all this data collecting into actionable data. Physicians resent having to be turned into recorders, so they are already an eager audience for this transformation. The answer is that developers need to work closer with health providers to understand what information needs to be captured and what information is already being captured elsewhere. This is why data liquidity and not just interoperability is of vital importance to making the data more actionable. If systems allow these points of data to come across, not just as PDFs or block of text, then structured data will reduce the amount of redundancy in data entry and can in fact kick off other processes. For example, if a surgeon fills out an electronic operative report, then that should kick off the discharge summary form complete with fields already filled out based on what's in the operative report. The data then becomes part of the process that is transferred and now becomes a helpful element in the physician's workflow.

Furthermore, developers need to work closer with physicians to determine what information is actually worth capturing. Sure, in some instances there are protocols & various pieces of information mandated to be captured in order to conform to institutional, professional, or national standards. But outside of those pieces of required info, developers need to not make assumptions of the type of information that physicians need to capture and what is important for other physicians to see. In our own experience, we work closely with domain experts - that is to say physicians well versed in their fields so we can tailor the experience and the content to their needs. The more companies that take this approach, that actually consult with healthcare providers to find out what they need, the better the data will be and the more impactful technologies will become.

Workflows need to be improved, and technology needs to play a part in lessening the amount of retraining and recording with which health workers are currently being tasked. But the first step is the hardest for most companies: they have to actively listen and understand the needs of the end users. Not just assume what should happen or how personnel should respond to various prompts, but actually get to know the processes and seek for ways to remove redundancies and make the entire experience one that provides actionable data. Data has always been around in healthcare, now it's time for healthcare providers and technology companies to collaborate to figure out how to make it all work for them.

photo credit: Algorave patches via photopin (license)