Digital Connected Care: Translating Data into Knowledge

The Digital Connected Care article series elevates the conversation from tech talk to the practical application of remote patient monitoring in clinician designed workflows with evidence of improved outcomes without increasing staff burden.

A 65-year-old woman with uncontrolled hypertension conducts self-measured blood pressure (SMBP) and takes at least two measurements a day—or over 60 measurements total—between her monthly provider visits. In checking his blood sugar regularly before and after meals, an 80-year old man with Type 2 diabetes collects over 500 data points between clinic visits. A 15-year-old asthmatic uses their smart inhaler to take maintenance and rescue medication doses and enters the number of trigger exposures into a mobile health app. In a six-month period, there may be close to 1,000 values for review. After multiplying across a patient population, a thorough examination of these voluminous, yet valuable data sets can be unmanageable for a clinician. Moreover, the data deluge offered by patient-generated health data (PGHD), can be a burden to already overwhelmed providers and, perhaps, even ignored. 

A previous article in this series, Surge Care for the 21^st Century, states that access to accurate, real-time data anytime day or night is key to providing hospitals with surge capacity to expand venues for care. To best leverage PGHD and to support optimal clinical decision-making, the data must be processed, standardized and structured prior to presentation. A fundamental element of any practical remote patient monitoring (RPM) solution is to translate PGHD into organized information that facilitates the interpretation of that data which, in turn, can be used to build a knowledge base and a set of rules contextualized to the given clinical situation. With these principles in mind, novel healthcare technologies will be in better position to support clinicians in making the most of the data generated by their patients.

Step 1: Assessing information needs

In a similar fashion to other clinical information synthesis endeavors, the first step consists of understanding the information needs of the clinician(s) using the PGHD. A helpful framework is the three-phase Needs Assessment Model by Altschuld and Kumar. The pre-assessment phase begins with forming a group to understand what the concerns are and what is already known. The next phase calls for the full assessment and identification of solutions. This can be accomplished by surveying, interviewing or simply observing the clinicians during clinical-decision making activities. Some questions to guide the inquiry are:

• Are all types of PGHD needed to support the clinical decision?
• Does every single data point need to be surfaced?
• What PGHD should be prioritized?
• Do gaps in the collected data exist?

The final phase of the assessment consists of developing an action plan for the solution, communication plans and methods to evaluate if the information needs are eventually met.

Step 2: From data to information

After assessing the needs, the next step is to process the data deemed useful into information. This entails attributing meaning to the PGHD by way of connection to the clinical situation. The resulting information should provide answers to questions related to "who," "what," "where" and "when." In the case of blood pressure measurements, the data may be trended by time of day—e.g. those taken in the morning and those taken at night. Blood glucose values may be best organized by showing those that are out of range or displayed as part of the meal log. Asthma PGHD processing may be dependent on whether it is a controlled or uncontrolled asthma patient or may benefit from being viewed alongside environmental triggers, such as humidity or pollutants.

Step 3: Knowledge and understanding patterns

Knowledge is the interpretation of data and information. The final step is assembling the information in a way that provides decision-makers with the tools they need to have a full understanding. This includes the identification of actionable findings from the data and information and provide a level of predictability as to what will happen next. Using the 65-year old woman with hypertension as an example, suppose she has a known history of heart disease and is now experiencing chest pain. Any sudden drop in blood pressure may indicate a heart attack. This interpretation of the PGHD, along with other information, provides knowledge that a clinician would want to act on immediately. Ultimately, the interpretation of PGHD will benefit from integration with other knowledge sources, such historical population health analyses, medical literature and clinical guidelines. This is essential to designing automated clinical decision support systems that truly benefit the clinical practice. Going one step further, this collection of data, information and knowledge can be transformed into insight used as a foundation for applied artificial intelligence.

In summary, PGHD can provide insights as to what happens in between clinic visits. We can collect as much PGHD as available, but it doesn’t do any good if clinicians have information overload and don’t know what to do with that information. We must thoughtfully and systematically collect, store and analyze PGHD so it can be used to support clinical decision-making and improve patient outcomes.

HIMSS Accelerate Health is working with a community of healthcare providers and system integrators to develop and deploy the underlying tools and infrastructure that supports the effective application of PGHD to a broad range of workflows that allow for successful transitions to RPM models of care. You are invited to participate in this effort by joining the HIMSS Innovation Organization, Personal Connected Health Alliance.

Previous Blogs in This Series:

1. The Innovation Journey 
2. Workflows by the Clinician, for the Clinician 
3. SMBP to Manage Hypertension 
4. Surge Care for the 21^st Century