In This Issue: Claims Turnaround and Decision Support

Issue summary and other notes...

Article 1 Claims Turnaround. Measuring, tracking, and improving claims turnaround.

Article 2 Decision Support in Claims. Best practices for setting up and managing a reporting environment in claims.

CH Tip Making Better Data Graphs. Graphs can be very powerful, especially when properly made.

Notes

The Arrow Keeps Going Up. According to Health Affairs, $7,498 will be spent on every woman, man and child in the United States in 2007, 20 percent of all spending. Costs are projected to increase to $12,782 by 2016. -- "Average 2016 health-care bill: $12,782" by Ricardo Alonso-Zalvidar Los Angeles Times Feb 21, 2007

Improving Claims Turnaround isn’t what you think

The pitfalls and best practices associated with improving claims turnaround.

Many payers struggle to improve claims turnaround time (TAT). As with other claims process improvement efforts such as reducing claims adjustments or increasing auto-adjudication, improving TAT require two things: Data and Approach.

Data

Claims systems do not typically measure claims turnaround in a way that facilitates understanding and improving TAT. Most claims operations have claims inventory reporting which categorizes claims by into aging buckets. Understanding which claims are over 40 days is not sufficient information to understand and improve TAT. Instead, the data needs to be configured to view all of the statuses of an individual claim, and how much time each claim spent at each status. By analyzing the turnaround data at the status level, appropriate opportunities to improve TAT can be identified.

Approach

Lets begin with a definition of TAT. For a given set of finalized claims (i.e. one week), TAT is defined as follows:

Effective TAT improvement efforts focus on total days turnaround by pend (status) instead of by claim. It is usually surprising to see which pends account for the largest volume of TAT days. For example, lets look at a Utilization Management pend. Odds are a significant portion of the claims over 40 days have a UM pend. However, the total UM days required to process all the claims for a given period probably accounts for 5% (or less) of the aggregate total days to process. Sixty to eighty percent of total turnaround time is associated with claims that take fewer than 20 days to process. As mentioned above, most claims operations support their staff with inventory reporting that is organized by aging buckets (i.e. claims over 40 days). Consequently, most TAT improvement efforts never leave this paradigm. Aged claims however, are a different claims process issue. Attack them separately. It might quiet a few noisy providers to pay those claims over 40 days faster, but it probably won’t substantially reduce average turnaround.

If UM pends (or aged claims) are not the right place to focus, where then? Focus on the pends with the most days. The most common pend (by claim volume) is probably the one with the most days behind it. Eliminating one of those top volume pends will likely have a significant TAT impact. Additionally, look for other system or batch steps (which all claims hit) which may represent significant portions of aggregate TAT.

Two final notes:

Cash flow is an issue. Be careful how you approach TAT improvements. A lower TAT means faster claims payments. Involve the Finance in decisions regarding appropriate TAT levels.

Management and incentives can significantly impact TAT. Turnaround issues are sometimes a function of other internal problems. Without going into how to set and manage performance goals for a claims operation, turnaround time improvement efforts should begin by making sure appropriate production goals are established and being managed towards.

Decision Support for the Claims Operation

Want to fix your claims reports. Here are a few tips...

Claims operations are a data rich environment. Examples of databases and decision support tools commonly found in claims operations include:

1. Claims Inventory Data and Reports. Most claims operations are managed by daily claims inventory/backlog. Managers adjust staff based on current claim backlog as it relates to a target. Typically, claims operations have significant inventory reporting tools including work queue, total backlog and aging reports. These reports help managers allocate resources to queues with the largest backlog of claims or a significant number of aged claims.
2. Other Production Reports. Claims operations run in weekly cycles. In addition to daily inventory reporting, it is quite common for there to be a series of weekly production reports. These typically outline summary volumes and rates such as total received claims, total paid claims, the AA rate, and the adjustment rate. These standard production reports tend to have three problems. First, they’re hard to change. Because they tend to be static reports created by IS, they require SQL programming to change. Second, they don’t facilitate drill down analysis. Understanding why your AA rate is 25% is not something you can figure out with a standard report. Finally, many of these static reports contain hidden business rules. For example, how are received claims counted? This often results in confusion and additional reconciliation.
3. Claims data warehouse. Most claims operations are supported by a large claims data warehouse. Such warehouses typically contain a year or two of claims data. They tend to be huge. Running queries or reports against them can take days or even weeks. Because the data loaded into them typically comes from multiple sources, there are often data integrity problems. Finally, the process to load a claims warehouse is typically significant and many are refreshed monthly.
4. HEDIS and other quality data. In any data rich environment with lots of decision support needs, multiple small databases (data marts) will emerge to meet specific reporting needs. Examples of such needs include HEDIS and customer reporting.

Fortunately, the data challenge is not insurmountable. By following two simple rules of thumb, a decision support solution designed specifically to understand and track AA and TAT can be deployed:

1. Focus narrowly on the data required to understand the key claims process issues: auto-adjudication, turnaround, and claims adjustments. Instead of using or loading whatever data is available, start by specifically defining the needed data. The data requirements for appropriate AA and TAT analysis are as follows. Get two types of data. Claims data comes in two types – data about the claim (provider, member, procedure, billed amount, etc…) and data about the process (examiner/analyst ID, date received, pend statuses, date paid, etc…). This second type, data about the process, is fundamental to understanding turnaround, auto-adjudication and adjustments. For example, analyzing how particular provider contracts or benefit plans affect auto-adjudication requires a picture of entire lifecycle of a claim, from receipt date, through its various statuses or pends, to paid date. Limit the data to a four-week sample. Too many decision support tools in claims operations fall prey to the temptation to get more data than is needed (there’s lots of data, why not get it all?). Four rolling weeks of data will support robust analysis of claims turnaround and auto-adjudication. Set it up so that each week the previous week is added and the oldest week is archieved.
2. Implement a decision support tool using off-the-shelf technologies developed for business users. There are hundreds of decision support technologies and tools and a thicket of jargon (OLAP, ROLAP, Portals…). Don’t be confused. Two things are needed A leading relational database. Most payers have established one of the leading packages as an internal standard. Its much easier to use an established standard than to go with a new package. Pick a leading query/OLAP tool. There are hundreds of analysis tools. Within the top ten vendors they are more or less equal. Make sure a particular tool supports ad hoc analysis and the easy creation of new standard reports or views. Also, make sure most, if not all of these functions can be managed by a minimally trained business user. The best trick I’ve used to screen amongst the tools is to visit other companies using a particular tool. If ‘normal’ business users in those companies are struggling with (or not using) the tool in question – try another tool vendor.

How were the case study payers using these tools to improve AA or reduce TAT? One of them was literally printing huge stacks of green bi-fold output and manually tabulating in order to perform root cause analysis. Another created a few new custom reports, which largely highlighted the need for more information. Rather than endlessly create new reports, management elected to focus on claims process improvement projects other than AA and TAT. The final case study payer had determined that a new database was needed but was mired in the process of developing an agreed upon specification for the database (no one could agree how much data was needed). It is also illuminating to review how each of the three had approached changing referral rules as a means to increase AA. Two of the three had abandoned hope of understanding the nature of their referral process. One had actually gone as far as to create a new report listing the claim numbers of all claims pended for the lack of a required referral for a particular week. This report was then run and printed. An analyst then went online and looked up each claim to determine how many were later paid.

CH Tip: Making Better Data Graphs

Graphs are a way to make your point. Edward Tufte shows you how...

I’m not a numbers person. I’m a graphs person. Graphs are easier to process. I’ve found that claims managers overwhelmingly prefer paper and online reports to be formatted as graphs (with supporting data available if needed). A well-executed graph makes it much easier to understand a problem or a complex relationship. Columns of numbers don’t give up their secrets so easily. Though graphs may be widely appreciated, surprisingly few how-to or best practice books are available. Head and shoulders above what is available stands Edward Tufte. He has published three books on the subject, all three of which are exceptional.