Many healthcare companies are aligning their long term goals to collect data from various streams into logical data warehouse to get competent and increase its operating margins (James Manyika, 2011). A data-lake in HDFS (Hadoop distributed file system) is used to dump a variety of data including but not limited to EDI, structured proprietary data, unstructured/quasi-structured data as Facebook feeds, weblog feeds, website rating, customer sentiments, logs from fitness trackers, ECG blood pressure monitors and blood glucose monitors). Pig, Hive and custom MapReduce are used to streamline the dumped data into Logical data warehouses (usually in Hadoop or other MPP database) to be used effectively in detecting Fraud, reducing gaps in care management and increase ratings by studying customer sentiments. In the current article, I would like to discuss care gaps on a high level. Proactively covering care gaps will not only save on administrative costs but predominantly increases HEDIS score and STAR ratings which help in rate setting activity- a direct capital benefits for MCOs (Managed Care organisation). Also MCOs can exploit double savings from reducing higher hospitalization costs for its members and also spend significantly less on outreach programs by reaching out through emails and letters than spending on a professional counseling service for the same where applicable. On the flip side, members will also gain advantage by being on top of their health schedule and significantly reduce risk of hospitalization. While predicting for care gaps we should consider a lot of determinants such as member’s psychological condition (reluctant to some intrusive detection techniques) and their willingness to undergo prescribed clinical procedures on time. In addition , behavioral, social, ethnic and educational variables often play a key role (Medicine, 2014) in determining care gap scores. In recent trends, usage and advancements in personal wearables which monitors physical activity and heartrate have sky rocketed and can add significantly to the decision matrix. (kosla)
Machine learning models can be used to predict member population health and tackle care gaps and pro-actively follow up with members. In the research observations it is found that the population in medium risk are often neglected and rise to higher risk. Within the determined groups, we can also further classify higher risk individuals who require counseling intervention for regular checkups thus reducing huge operational costs. As described earlier it is a win-win situation for both care organization and members alike. Augmenting AI with case managers gives us distinct advantage.
Care businesses can now define a strategy through proposed methods to benefit cost savings and avail distinct advantage over competition.
follow me on twitter @selineni