The idea is to apply underlying principles of data science to improve processes or predict outcomes. Ideally, organizations can improve the quality of care while reducing cost, all by looking at what the data is telling them.
Why Analytics is Important in Healthcare
Many large teaching hospitals have been tapping into the power of big data for quite some time. However, as payment models continue to shift away from fee-for-service and toward value-based purchasing, the number of hospitals collecting and relying on data to drive business decisions only will continue to grow. Readmissions are a focus area for many institutions. For example, using data analytics at the point of care helps providers determine the risk of readmission, allowing them to tailor discharge planning accordingly. Analytics also can be used to identify the factors that affect patient safety indicators, core measures, case-mix index, hospital-acquired conditions, severity of illness, risk of mortality, and more.
Analytics and ICD-10
Predictive analytics will become even more prevalent as the industry transitions to ICD-10-CM/PCS. Greater specificity inherent in this new code set undoubtedly will lead to developments in chronic condition management, population health management, larger-scale accountable care organizations (ACOs), and even bioinformatics and personalized medicine. Although it may be too soon to predict just how powerful ICD-10 will be in terms of analytics, it’s not difficult to surmise where the most significant opportunities may lie. Any code sets that will be expanded in ICD-10 have the potential to enhance research and process improvements. Some examples include the following:
- Glasgow coma scale score — Capturing these scores within ICD-10 diagnosis codes will benefit trauma centers in particular in terms of research, tracking expenditures, and more. It also helps trauma registries capture richer data.
- Diabetes code expansions — These expansions will be extremely beneficial in terms of treatment and programs for this common condition.
- Combination codes for pressure ulcers — In ICD-10, one code captures both the location of the ulcer (including laterality) as well as the stage. ICD-9 requires two codes. Having one code that captures all of this information will make analyses much more efficient and comprehensive.
- Seventh character extension to capture more information about injuries — Thisincludes whether the encounter is an initial, subsequent, or sequela visit. This data will greatly assist with injury surveillance and the ability to target public health programs based on need.
- Changes in acute myocardial infarction (AMI) coding — In ICD-9, an AMI is coded as acute when it occurs within eight weeks of initial onset. However, ICD-10 has shortened the timeframe for AMI to four weeks but established a new category to identify a second acute MI that occurs within four weeks. This coding guideline change may enhance research and bring greater attention to the prevalence of acute MIs and what causes them, as well as provide data to potentially improve treatment options such as cardiac rehab.
Looking Ahead
It certainly will take time for organizations to develop strategies to capitalize on the data analytics possibilities with ICD-10. For the first few months or even years after implementation, the attention will be primarily on getting paid – that is, ensuring clean claim submission and reducing any denials. However, with the ensuing wealth of available information, other parties likely will use that data to improve healthcare and for other defined purposes, not just to track and bill for services.
Data analytics is only as accurate as the information on which it is based. This is where health information management (HIM) can make a big difference. Is the data accurate and complete? Can it be mapped between ICD-9 and ICD-10? Does the data provide an accurate picture of what’s going on?
HIM can work with clinical documentation improvement (CDI), physicians, and hospitals’ leadership to make the most of this data revolution and capitalize on all that ICD-10 has to offer.
