Improving Patient Care with Real-time Analytics
Identifying patterns in medical data is crucial to establish a solid care plan with the patient and continuously improving care plan options as new patterns emerge.
An important part of improving patient care is a matter of getting involved in the patient’s medical history (personal and family), and having access to technologies that will assist in recognizing patterns of behavior and evidence-based population health driven decision support. Identifying patterns in medical data is crucial to establish a solid care plan with the patient and continuously improving care plan options as new patterns emerge.
Another option to improving patient care is to establish a competitive advantage for the providers by communicating to them how they compare to other care providers in their organization in matters of keeping patients healthy (e.g. better treatment and outcome metrics, preventing readmission).
The last option is leveraging “what if” calculators to brainstorm behavior-driven outcome options with the patient, and applying tactics such as CBT (Cognitive Behavioral Therapy) to influence (steer) the patient’s decision making process. For example: Sharing a picture [on the patient portal] of a sunny beach with the patient, as a reward for taking his or her medication as prescribed, is a visual indicator stating, “You will be healthy enough to go on vacation,” versus a big red “X” over the sunny beach picture, indicating, “No vacation in sight with the way you are following your doctor’s care plan.”
Healthcare institutions have been collecting patient data since the early 60’s in some form of electronic format. Today, centuries later, our data warehouses are filled with a fast amount of longitudinal patient data across numerous generations of care. Data storage is no longer calculated in mega bytes or tera bytes. They are being calculated in peta bytes (1 PB = 1000000000000000B = 10005 B = 1015 B = 1 million gigabytes = 1 thousand terabytes).
There are many factors contributing to this growth rate in patient care related data stored in our data warehouses. Primarily, this growth rate can be contributed to the number of patients (e.g. chronic deceases have increased in recent years) and electronic medical record systems that are collecting more data points. Aside from that, we have research related health data and integrated data points through Health Information Exchange HIE type data transactions.
So how will we put all that data to good use?
As humans, we are born pattern recognizers. We analyze patterns all the time. A simple crossing of an intersection is a good example of real-time analytics in action: in the seconds before we step into the road, we are collecting a vast amount of data points, comparing it with our history data (experience storage) to determine if we should even be here (remembering if there is a positive or negative experience associated with this particular intersection), and determining a go or no-go. Once we have a “go,” we precede with follow-up steps walking us closer into the intersection. While we are doing this, we are constantly collecting new information, comparing it to our history data, looking to discover patterns that could be important for our survival (making it through the intersection alive). Patterns such as “heavy traffic pushing into the intersection,” “everybody walks with me,” “no one walks, I am the only one walking,” or “all cars stopped, expect the red car approaching fast form the right.” Our brain is looking for patterns all the time at no additional [operating] cost to us.
Basically, our given ability to recognize patterns is always at an “ON” state and thus for free in terms of operational cost. So, how can we use our ability to analyze in our daily practice of care?
Traditionally, we have been analyzing data through complex ETL (Extract, Transform, Load) processes in our data warehouses. Although valuable for research and trend type analytics, these types of analytics do not lend itself practically, and cost sustainably, to real-time patient care (at the point of care). We have to produce new innovative and cost effective ways of moving the discovery of patterns and communication of data realization closer to the patient in order to improve patient care. There are various ways of accomplishing this.
One way is to provide the care team with an integrated mechanism for discovering patterns in the patient’s history, during the point of care, by combining analytics dashboards (e.g. pie-charts, lists, views) with longitudinal patient medical trends inside the EMR’s user interface required to capture new data points. Traditional EMR systems keep analytics as a separate concern, removing the natural capability of real-time pattern discovery.
Considering the shift in healthcare from the episode-based care model to the outcome (value) -based care management, I believe it is crucial to update our Electronic Medical Record systems with the ability of integrated analytics. We have to fuse the EMR and Analytics.
It is the same from the patient’s point of view. Analytics should be incorporated in the traditional patient portal, giving the patient a unique ability to look at his or her health record in a more analytical way (natural way). For example by discovering how they compare to other patients in their county (defined population), their occupation category and their particular age group. Or, given their medication history, why did the dosage increase over time, and compared to other more successful outcomes (in similar patients), what could they do to lower the prescribed medication dosage?
In my view, analytics of today is no longer “just about the data”, it is largely about the ability to discover patterns in the data and being able to make that knowledge available in a usable manner, so it can become actionable by the receiver.
A smart man once said, “Data is the secret sauce.” Although I can agree with this statement, I believe that, in today’s market, data is a given; we have lots of it (and more and more every day). Data is what flour is to bread. The secret is in the method of kneading and baking the flour.
Chief Technology Officer, Marshfield Clinic, Oliver Degnan, will be presenting at the PDS 2012 Technology Conference this Sept 20, 2012 at 2:15 PM. Learn more about Degnan’s session and register today by visiting www.PDS2012.com.