Back

 Industry News Details

 
Five Challenges in Using Predictive Analytics to Improve Patient Outcomes Posted on : Oct 08 - 2015

In the increasingly patient-centric world of healthcare, predictive analytics has taken firm hold as a means for healthcare organizations to improve patient outcomes. Predictive models for determining patient responses to medication, lowering hospital readmission rates, assessing risk of disease breakout, and other uses are being implemented for superior disease management and care delivery. Similarly, in the area of med-tech, predictive analytics is about determining the likelihood of device failures, responding proactively to improve machine downtime, assessing consumable sales forecasts, and other medical device and customer business issues. Innovations in predictive analytics are enabling med-tech companies to differentiate themselves from the competition and improve business results like never before.

Currently, med tech companies are strengthening their data collection and analytical capabilities to indirectly determine patient conditions, through devices such as heart monitors, biosensors and other wearable devices. They are also using these new capabilities to directly predict device performance and usage, such as machine failure, which is critical to maximize clinical availability and minimize costs due to downtime. The costs of downtime can be significant based on the Complexity of the treatment process, the scale of operation, the number of failure occurrences, and the opportunity cost of providing better care.

Implementing a successful healthcare analytics program requires tackling the following challenges:

1) The need to select predictor variables:

Predictions are obtained by developing a predictive model which is essentially a statistical model predicting a value or happening of an event, based on information from explanatory variables. The implicit assumption here is that a lead time exists between when data of explanatory variables are obtained, and when the predicted event could happen. This lead time helps companies take corrective action or mitigate risk. View More