Preliminary Step: Initial Data Analysis
Prior to implementing PROMETHEUS, an initial data analysis of the health plan’s claims data must be performed as it is essential for determining the areas in which quality of care can be improved and costs can be reduced. The initial data run can help implementation partners determine which Evidence-Informed Case Rates (ECRs) are of the highest priority, from both a quality improvement and cost savings perspective. Potential Pilot partners review the initial outputs in order to determine the percentage of dollars spent on Potentially Avoidable Complications (PACs) across all 21 conditions and procedures. The PROMETHEUS developmental database, as well as our regional variation analysis offers pilot site partners the opportunity to compare their results to our benchmark data, via the ECR Tool.
LESSON LEARNED: Completeness and Accuracy of Claims and Eligibility Data
Before a pilot implementation can begin a sample of data from each potential data source should be reviewed, to ensure the necessary data specifications are in place. This is an essential component to implementing PROMETHEUS, and it is actually one of the first things that should be looked at when an organization is interested in implementing PROMETHEUS. Payers need to be sure they have claims data that are clean and accurate. A certain level of data quality is required for data that is run through the SAS-based programs and for data that is input into the ECR Engine. If member identifiers or other information is missing from the data, the SAS-based output and the outputs from the Engine will be compromised, the data outputs and analyses will be inaccurate and will produce erroneous conclusions. Potential pilot sites should refer to the Data Specifications, which detail each data requirement that is essential for the success of pilot. This checklist should be used in the review of each data source before the implementation can move forward. Until it is verified that the initial data analysis comes from clean and accurate data, an organization should not take any further steps until they work through the issues. In our experience, health plans that did not have the data specifications in place faced problems when it came to analyzing their data, and had to spend time correcting the errors and running the analysis all over again. This delayed some pilot sites a year or more.
PROMETHEUS has developed a SAS-based ECR analytic package that a potential implementation site can run in-house if they have SAS capabilities. Interested organizations must download the freeware, run it internally on their claims data according to the PROMETHEUS data specifications, and then upload the results to the ECR Tool. The ECR Tool will then validate the data, compare it to our benchmarks, and produce an opportunity report. This ECR Tool is located on the HCI3 website. See below for the link to a sample opportunity report. For more in depth analysis, a potential pilot site can also decide to have a PROMETHEUS data partner (MassPro or MedAssets) perform the ECR analysis. A Data Use Agreement (DUA) and Business Associates Agreement (BAA) will need to be in place prior to any data transfer. Prometheus can provide template DUAs and BAAs (see the link to template DUAs and BAAs below) but if the site usually employs a standard agreement and would prefer to use their own agreement as a template, this can help to expedite the agreement process. In past experience, the timeframe for execution of legal agreements varies widely and so PROMETHUES recommends beginning the process as soon as possible.
Another option is for a plan to run the data in house. This requires a licensing agreement with HCI3 to license the full analytic package. Please see below for the link to the licensing agreement. Once all the necessary agreements are in place the site submits claims data in SAS input format to be run through the ECR programs. The advantage of running the data by this method is that the ECR analytic outputs are more in depth. In addition to providing the user with a breakdown of the dollars spent on the chosen ECRs, an organization will also receive provider-level outputs that can help an organization understand where the sources of variation are, in terms of cost and quality. See below for a link to a template report that can be created with these outputs.
These initial outputs and results, whether generated by the website ECR Tool, MassPro, MedAssets, or internally can be reviewed within the potential pilot site organization, or can be sent to HCI3 to facilitate interpretation and create reports. The results can then be reviewed with the organization, with or without the help of HCI3. From here, it can be decided whether to implement PROMETHEUS, and the organization can reach out to HCI3 if they have not already done so.
Useful Tools:
ECR Tool This is the HCI3 ECR Tool where a potential pilot site can download the freeware, run their claims data, and receive an opportunity report.
Sample automated report from ECR tool This sample report is generated from the HCI3 ECR tool.
New ECR analytics report This is the template report that includes a more in-depth analysis. This type of report is created when a PROMETHEUS Data Partner runs the claims data, or when a potential pilot site licenses the full ECR analytic package and runs their claims data in-house.
Sample Data Use Agreement and Business Associates Agreement These are drafts of the legal agreements required between a potential pilot site, HCI3 and a PROMETEHUS Data Partner when a site decides to have a PROMETHEUS Data Partner run their claims data.
Sample License Agreement This is the sample Licensing Agreement required for a potential pilot site to run their claims data in house on the full ECR analytic package.
Data Specifications for Payers This document contains a "checklist" of the data specifications according to which payers need to format their claims data in order to run the analysis for the SAS ECR Analytics v3.5.2.
ECR Instruction Book This document contains the instructions on how to format claims data so it meets the data specifications for the SAS ECR Analytics v3.5.2 package.
Suggested Data Warehouse Fields for Robust Analysis This document contains the suggested claims data warehouse specifications that HCI3 came up with through working with health plans to analyze data. These specifications are not PROMETHEUS-specific, but are designed for any health plan wishing to create a data warehouse for robust data analysis. The focus is on creating a data warehouse that allows data to be easily extracted and formatted for analysis.
