Quality Not Quantity

Big Data research needs to be guided by content experts. Physicians hypothesized that getting outpatient medications and instructions to patients prior to discharge would reduce readmissions. This hypothesis was validated based on information gleaned from Big Data. Many hospitals now provide patients with medications and guidance prior to discharge.  Integrating physicians' knowledge into Big Data findings made this possible.

High rates of hospital readmissions have long concerned healthcare providers. A 2010 U.S. government study found that 1 in 4 patients with heart failure returned to the hospital within 30 days. To combat the problem, Centers for Medicare Studies (CMS) enacted the Hospital Readmission Reduction Program (HRRP) in 2012. The regulation penalizes hospitals by withholding up to 3% of regular reimbursements if they have a higher-than-expected number of readmissions within 30 days for specific conditions.

These regulations stimulated development of predictive models to uncover patterns of readmission using Big Data. However, Big Data's ability to gauge hospital readmissions was poor. With hundreds of variables and thousands of records, it was possible to detect associations and anomalies but few that could predict readmission. It took researchers with specific hypotheses based on social risk factors to determine that hospitals with in-house pharmacies had significantly reduced readmissions.  If patients were discharged with medications in-hand they were more likely to take them.  Today, hospitals such as Johns Hopkins Medicine hospitals have special pharmacy programs that provide bedside delivery and instructions for discharge medications.