January 27, 2016

Big Data at work in healthcare

Novel Predictive Models for Metabolic Syndrome Risk: A “Big Data” Analytic Approach

Reducing Metabolic Syndrome Risk Using a Personalized Wellness Program

Metabolic syndrome is somewhat fuzzily-defined, let's say it's a concept-frame that can lead to a condition like chronic heart disease, stroke and diabetes. The results of the estimates from a big data application provide clear messages about the need for personalized wellness programs. Both issues are covered in the quoted articles. There is no reason to delay its application.
Let's add the key statement:
The ability of the models to produce highly individualized risk profiles for overall risk of metabolic syndrome and by specific risk factors allows for more successful patient engagement in subsequent care management programs. Figure 2 shows 2 different individual risk profiles. Subject ID 423262 was a 46-year-old male with current out-of-range metabolic syndrome risk factors of high-density lipoprotein (HDL) and waist circumference. He had a 92% predicted probability of developing metabolic syndrome within 12 months, and a 73% probability of developing abnormal blood glucose as a third specific metabolic syndrome risk factor during the study period. Subject ID 107975 presented a contrasting profile. He was a 37-year-old male with 2 out-of-range metabolic syndrome risk factors—HDL and triglycerides—but had only a 40% predicted probability of developing metabolic syndrome within 12 months. For this subject, abnormal blood glucose was also the most likely abnormal factor to develop next, but carried only a 26% likelihood.