The value of big data and its ability to provide insights are clearly becoming hallmarks of the 21st century. As healthcare  managers seek to understand what is happening at their organizations from a real time perspective, they are turning to data (and the metrics that can be developed from them) to get a clear picture of where they are and where they want to be. Increasingly, providers and payors are focusing on understanding how data can be leveraged to educate, assist, and transform Clinical Documentation Improvement (CDI) programs within the healthcare environment. To monitor the success, performance, and sustainability of its CDI program, an organization must incorporate analysis of Key Performance Indicators (KPIs). This analysis should include: evaluation of MS and APR DRG Payor Groups and their corresponding core metrics, real time assessment of Value Based Purchasing criteria and measures, and optimization of ICD-10 coding accuracy.


The traditional Clinical Documentation Integrity (CDI) model focuses on MS DRGs in the Medicare system. An initial and ongoing focus here will assist the organization in understanding and improving its CDI program.

Having an established methodology on metrics such as Case Mix Index (CMI), Complications and Comorbidities (CC/MCC Capture Rates) and overall volume analyses in the medical and surgical arenas will lead to a focused approach in documentation and coding improvement. This focus will improve a hospital’s bottom line and assist with achievement of higher quality scores.

In particular, the use of benchmarks based on reliable and comparable data is essential in guiding a CDI program. Benchmark analysis can identify areas in which the organization is lagging behind its competitors with respect to CMI or CC/MCC capture rates, or where it may be exceeding certain benchmarks, leading to risk. Using data provided by the Center for Medicare and Medicaid Services (CMS), an organization can prepare benchmarks for facilities that are similar in areas such as volume, type of service and teaching status, among others. The example in Table 1 below shows CMI and CC/MCC capture rate benchmarks for the 20th, 50th and 80th percentiles, developed for like facilities based on teaching status, disproportionate share percentages and total volumes.

Click here to view Table 1

In the example in Table 1 above, the items underlined in green show the Medical CMI Excluding Ventilators for different time periods.  In Year 2, the example facility’s CMI of 1.2088 is well above the 80th percentile of 1.1365. There is also a noticeable improvement from 1.1720 in Year 1 to 1.2088 in Year 2. There is risk associated with being over the 80th percentile; it could indicate potential over coding. It is recommended to do a random chart review of these accounts to ensure that they are coded appropriately and compliantly.

The opposite is seen when looking at Surgical Complications and Comorbidities (MCC% and CC%) underlined in red. Both of the MCC and CC capture rates of 15.8% and 23.9% are below the 20th percentiles of 17.6% and 25.2% respectively. Further analysis into the Surgical DRGs is necessary to see where this organization is falling behind. A deeper dive into service lines such as Cardiovascular, General Surgery, Orthopedics, etc. will show where the CDI team should focus its attention.

For example, in an effort to stabilize the high fluctuations in Relative Weights on the surgical side, the organization should consider looking at MCC and CC Capture Rates. The range for Surgical MS DRG relative Weights is from 0.7070 to 25.3518. Adding only one of the highest or lowest of these DRGs to the calculation of CMI can prove to bring huge swings and dilute the hospital’s actual performance.

While looking at traditional Medicare is a good start, expanding this program into other payors is the best bet in getting the highest quality, reimbursement and resource utilization from a CDI program.


APR DRG trending can illuminate clinical and quality performance even more specifically than MS DRG trending.  Using APR DRG data, a hospital can track and trend detailed information that will provide insight into Severity of Illness (SOI), Risk of Mortality (ROM) by DRG, and by various groupings of similar DRGs. SOI acts as a modifier on the DRG, moving it between four different categories (1 – Minor, 2 – Moderate, 3 – Major, and 4 – Extreme) and will affect reimbursement. Table 2 below shows Intracranial Hemorrhage as the SOI moves from 1 to 4 and the impact to the corresponding relative weights. (The DRG’s relative weight multiplied by the payor provided Base Rate gives the total payment for that encounter.)

Click here to view Table 2

Tracking not only the Overall, Medical, and Surgical CMI, but also the SOI percentages within those categories creates a clearer picture of performance (Table 3 below). The higher the SOI placed on an APR DRG, the higher the payment will be.

Click here to view Table 3

Risk of Mortality (ROM) does not have an effect on the reimbursement of a DRG, but can assist in ascertaining quality performance. ROM as seen in Table 4 below does not necessarily correspond to SOI. Taking the ROM of a certain specialty and comparing that with actual death rates within that same specialty can show opportunities for improvement in coding and quality of the services provided. If there is a large portion of DRGs in a service with a ROM of 2 and a large percentage of those patients are expiring, an analysis of the documentation of those patients would be warranted to explain why the ROM is not higher.

Click here to view Table 4


Beginning October 1, 2013, payments for Medicare beneficiaries in acute care hospitals were affected by the Hospital Value Based Purchasing (HVBP) program put in place by the Patient Protection and Affordable Care Act of 2010. Evaluation of the quality of the Federal FY 2013 inpatient admissions criteria shaped the amount of 2014 Medicare reimbursements. DRG payments were withheld by 1% and placed in an incentive pool. Reimbursement of that 1% was either reduced or increased by what CMS called an adjustment factor. This factor for 2014 reimbursements ranged from 0.9886 to 1.00881. Some hospitals will not recoup their DRG withholding, some will receive a bonus, and others will break even. Ensuring that HVBP criteria are met from year to year will secure the reimbursement of withholdings from CMS.

Click here to view Chart 1

CMS adopted four domains to evaluate quality in the HVBP program: Clinical Process of Care, Patient Experience of Care, Outcomes and Efficiency. For FY 2013, CMS used only two domains to calculate hospital quality scores: Clinical Process of Care and Patient Experience of Care. The domain scores were weighted at 70% and 30%, respectively. Each year CMS will add a new domain and adjust the weighting as illustrated in Chart 1 above. CMS will continue to use these 4 domains in subsequent years, but the individual data points and metrics that comprise each domain may be reviewed and updated.

Effectively measuring and comparing hospital data throughout the year using CMS guidelines will assist hospitals in meeting and exceeding the HVBP quality goals and maximizing the reimbursements made by CMS the following year. See Table 5 for information on the metrics used for evaluation of Federal FY 2015.

Healthcare organizations should seek to utilize analytics to monitor all of the VBP metrics within each domain in real time. This monitoring, paired with the real-time education, can result in rapid improvements to the quality metrics that will be reported to CMS, ensuring that future reimbursements are maximized. Additionally, analytics can help organizations to validate the data open for public view on by comparing it to the data the hospital reported to CMS. Discrepancies between the public data and the reported data can form the basis for an appeal to CMS that would change those scores and reimbursements.

Click here to view Table 5


Under ICD-10, hospitals have the ability to analyze much more specific and detailed data than in the past. ICD-10 captures new technologies and services, as well as reflects current clinical knowledge and practices. Effectively utilizing and understanding trends in ICD-10 data ensures continued success and provides for higher standards in quality of care.

In procedure codes alone, the 3,882 ICD-9 codes translated to 72,769 ICD-10 codes for FY 2015. In diagnosis codes, the 14,567 ICD-9 codes became 91,737 ICD-10 codes. That represented a 1,775% increase in the volume of procedure codes and a 530% increase in the volume of diagnosis codes. The expanded code set provides for a much more specific and detailed picture of a patient and impacts the depth and breadth of analytics a hospital can evaluate. Additionally, the ICD-10 system makes it easier to assess quality and allows healthcare providers to compare their own metrics with those of other organizations at a much grander and more precise scale than was previously possible.

Organizations must ensure that the right processes and procedures are in place to ensure documentation specificity under ICD-10. Data analysis is an effective tool that should be leveraged to identify deficiencies in ICD-10 coding.

For example, unspecified codes are under much more scrutiny in the ICD-10 world than they were under ICD-9, because it is rarer not to be able find a more appropriate code. Unspecified codes are only used when there is not enough information in the medical record to select a more specific code. Education and querying of physicians can help to ensure that appropriate codes are chosen. Organizations can use data analysis and benchmarks to minimize situations such as the use of unspecified codes.