Banking on data analytics

By Paula DeJohn

The days when arm twisting a vendor rep could yield one more discount off the list price, and when joining a group purchasing organization meant automatic double-digit savings from volume-aggregating contracts, are long gone. Yet the pressure to eke out more margin from dwindling reimbursements and rising supply costs goes on.

Hospital materials managers have a new tool in that effort, generically known as “data analytics.” Like electronic commerce, benchmarking systems, taxonomy standardization and in-house materials management information systems (MMIS), data analytics relies on computer technology to inject efficiency, accuracy and credibility into the never-ending cycle of product review, contracting, ordering, delivery and payment.

And, like the earlier technologies, it has early adopters and promoters who offer it as a holy grail in an effort to reduce supply costs. Others say it is overdue, noting other industries are far ahead in knowing how to use financial data, and still others say this is just the first step in a complex process that will change the face of materials management in health care.

Rapid returns

Summa Health System, Akron, Ohio, is well into its first year of using Spend Advisor, the data analysis system developed by Summa’s GPO, Premier, Charlotte, N.C.

Summa’s director of supply chain management Jay Mitzel reports, “First, we put together a contract management team. We used this as an initial tool.”

The move paid off almost immediately, as the system turned up $1 million in savings out of an annual purchasing volume of $65 million. “SpendAdvisor data accounts for half of it,” Mitzel says. “Real implementation only began in June and July, so it’s been a pretty rapid return on the investment.”

SpendAdvisor is Premier’s name for a system that compares ordering and payment data with contract prices, and identifies savings opportunities from possible higher tier eligibility based on volume, product and vendor standardization, and other statistical data that in the past would have required intensive manual labor and head-scratching.

Similar systems are Global Spend from VHA, Irving, Texas; Spend Compass from Advisory Board Co., Washington, D.C.; Contract Center from GHX, Louisville, Colo.; and Supply Chain Analytics from MedAssets, Alpharetta, Ga. All offer variations on the theme of spending analysis to identify savings opportunities.

In Summa’s case, Mitzel says he looked at three or four alternatives before selecting a data analytics product. His criteria were usability and effectiveness. “A cornerstone to effective supply chain management is the availability and accuracy of data, benchmarks and contract information,” he says. “It gives us a tool we can legitimately use.”

In fact, he expects the system to come in handy as Summa works to integrate the three hospitals it acquired in 2007. The lessons it learned through SpendAdvisor will give Summa benchmarks to analyze the new hospitals.

Among these lessons were the importance of visibility and reliability of  data—Summa was able to implement savings because it was able to demonstrate the effect of changing behavior on the bottom line.

“If it wasn’t working, I wouldn’t keep it, but I think it’s working,” Mitzel says.

Bill Marquardt, vice president of product development and adoption at Premier, says before SpendAdvisor was launched two years ago, there was a long beta process in which the Greater New York Hospital Association, a Premier affiliate, tested it in member hospitals.

The cleaner, the better

For hospitals embarking on the journey to data analytics, cleansing is an issue. Data analytics software companies want to see clean data, i.e, no mismatched product identifiers, and consistent product descriptions and hospital or department names.

 Not all providers of data products take this responsibility, while others include it in the package. For example, there is no cost to a Premier member to use the SpendAdvisor technology, but there is a charge based on the size of the hospital or IDN, for data collection, cleansing and uploading. VHA does not combine data cleansing with the analytics function, but it does provide a “data scorecard” that points out discrepancies it finds.

Spend Compass does the cleansing through an automatic function that avoids letting mismatched data creep in. 

For its input data, VHA prefers the receipts for all invoices paid. This is more comprehensive than purchase orders (PO), says Dan DeLay, senior vice president of supply chain analytics, because not all transactions are by PO. “It gives a better picture than a PO file.”

To use the Spend Compass system, a hospital exports data from its materials management information system. Spend Compass accepts raw data from a hospital’s item file, vendor file, 12-month rolling purchase order history, rolling 12-month accounts payable records and the hospital’s contracts. 

For input, the Premier system uses line-item, transaction-level data. It’s the materials manager’s choice whether to use purchase order records, invoice files or accounts payable receipts.

“Some hospitals are more confident in the purchase order price and others are more confident in receipt price,” Marquardt says.

What’s to come

At first, the bulk of potential savings comes from comparison of hospital spending with group or local contract pricing. Too often, hospitals find they are paying higher prices than necessary because vendors invoiced, and accounts payable paid, prices that did not reflect negotiated contracts or appropriate contract tiers. The initial “dashboard” reports show these discrepancies.

VHA takes the hospital’s receipt file and runs the data through its global spend analysis program. Then the member goes to www.VHA.com and logs in to the secure spend analysis site to view the reports.

The basic report is the executive summary, a graph of overall spending by product and segment, such as capital equipment, orthopedics, etc. For IDNs, a user can look at the whole system, individual hospitals or selected groups of hospitals.

Next is a table showing potential savings. These are calculated in three ways:

• By comparing hospital payments with what is available in the Novation contract and with the best tier for which a hospital is eligible
• By showing where standardization is possible by product and vendor, even for noncontract items
• By price leveling. In other words, finding the lowest price any hospital or department is paying for an item and standardizing it.

Then there are custom reports. For example, a user can download all data related to cardiology or other service line. The next time they upload data, the reports are automatically refreshed.

“Then we can begin to do trending,” DeLay says. That means watching how spending is changing for those products, if additional vendors are creeping into the mix. A materials manager can ask the department why it needs the additional vendors, and if doctors insist they are necessary, negotiate contracts with the new suppliers.

“Like any other application,” DeLay says, “the more you use it, the more ways you find to use it.”

 “We identify simple problems first,” Marquardt says. First is the simple contract discrepancy where a hospital has paid more than the contract price for a product.

Another is what Premier calls the “Price Activation Opportunities Report.” This is a list of contracted products that the hospital is buying without having officially signed up for the contract. The report shows how, by correcting that oversight, the hospital may qualify for a higher discount tier and become eligible for manufacturer and GPO rebates.

Child Health Corporation of America (CHCA) uses the pharmacy component of a data analytics system and estimates its hospitals saved $4.5 million in the first 18 months, partly by improving contract utilization.

Mississippi Baptist Health System, Jackson, Miss., reports finding more than $5 million in supply savings during 2006 by using data analytics software: $3.3 million from contract activation and better utilization, and $2 million by identifying products for which vendors were not offering their lowest prices.

While savings alone are good news, Marquardt notes, they also give materials managers leverage as they work with other departments to improve expense management strategies such as opening discussions about high-expense physician preference products.

Crozer-Keystone Health System, Springfield, Pa., began using VHA’s Global Spend Analysis system in 2007 and found $140,000 in potential savings in the first quarter.

“VHA continues to help me understand where we stand as a health system relative to our supply chain efficiency and whether or not we’re maximizing VHA/Novation contracts to our fullest advantage,” Anita Keenan, director of purchasing, says in a statement.

Like Premier, VHA developed its analytics tool internally, with member input.  VHA rolled out Global Spend at the AHRMM annual conference in 2006, and demonstrated it again at the 2007 meeting.

According to DeLay, 240 members, representing hundreds of hospitals, had signed up by late 2007, and “demand is increasing.” There is no charge to participate, he says: “It’s part of our core service to our members.”

Hospitals should be able to depend on analytics providers to help them find savings based on the spending data they upload.

If data analytics is to help the health care industry manage its supply chain, it cannot impose additional burdens on materials managers. All current providers insist they are the ones to provide that expertise.

 “Data analysis has been used in other industries for a long time,” VHA’s DeLay says, and health care suppliers are familiar with it, but it is new to providers. “It’s not because hospitals are dumb,” he explains. “It’s because they didn’t have the data.” Or rather, they lacked the technology to manipulate the data.

For data analytics to be viable, he says, three things must be present: “technology, data and talent, or people who know how to combine the data and technology.”

The real deal?

With only about two years of use in health care, the technology is in its infancy, and there are skeptics. One is Jack Schwartz, director of materials management at Antelope Valley Health and Hospital System, Lancaster, Calif., a VHA member. Schwartz has used Global Spend Analysis to run a couple of reports.

“It’s a kind of fool’s gold,” Schwartz says. The problem, he notes, is that a hospital may not be able to realize the potential savings the reports identify.

For example, the program is such that it is assumed a hospital can adjust its volume to reach the highest contract tier. In his case, with such products as pacemakers, “I’m never going to qualify for the highest tier,” Schwartz says.

From an industrywide perspective, Patrick Michael Plummer agrees. A former vice president for capital equipment at VHA and Amerinet, St. Louis, Plummer also has developed an equipment-pricing database and now consults for suppliers. “Over the last five years there’s been enormous progress made in these data systems,” he says, “but we’re still in the early stages.”

The current systems address the fundamentals: What a hospital is buying, how much of that product, from whom and the right price. “That is only where we are today as an industry,” he says, and much of that information is available just from cleaning up data.

“We are only scratching the surface,” Plummer says. “The next step, where the real value comes from, is to suggest comparable products that other institutions of comparable size are buying because they can get them at lower cost.”

That, he notes, means going beyond the contract list of a particular GPO and surveying the entire industry. “An individual GPO is never going to get there,” he adds. Beyond comparability, according to Plummer, is the issue of product choice, and that should be the next level of data analytics.

A hospital should be able to identify which products and vendors will work best for it based on its caseload in cardiology, orthopedics, oncology and other departments.

“Can someone tell us what sources we should be using based on our consumption? What’s the lowest cost I can get to based on consumption across the board, for the whole institution?” The answers, Plummer adds, will influence the choice of GPO as well.

“This is where it is going,” he says. “It’s going to evolve pretty fast once it starts.”

The GPO-neutral model

Hackensack (N.J.) Medical Center selected the GPO-neutral Spend Compass because it covered the entire spending history, rather than concentrating on GPO compliance, and because of the Advisory Board Co.’s focus on best practices, says Paul Onufer, chief supply chain officer, in a statement.

Hackensack signed up for Spend Compass in April 2006, began running reports in June, and by the end of the year had made giant strides toward its goal of knocking 2 percent off its $80 million annual purchasing volume. For 2007, estimated savings are $3 million.

The hospital derived the savings from a combination of improving contract compliance, renegotiating pricing, and standardization. That would have been impossible before, Onufer says, because Hackensack did not have the resources to compile the spending information. Now, the system does it for the hospital.

He and his staff review each report to identify where prices have changed and drill down to the transaction level to track the root cause.

They have found areas where prices paid varied wildly, and used this information to remind buyers to stick to contracts, to renegotiate with vendors or simply to convert to other products. In the process, supplier accountability has been easier to enforce, Onufer says.

“Suppliers who change prices frequently are quickly identified and the situation is addressed,” he says. 

Troy Wasilefsky,  senior director at The Advisory Board Co., says the system is free to hospitals that are members of the company, and other hospitals may participate for a fee based on size.

In addition to the spend and price reports is a feature Wasilefsky calls “guided analytics,”  which he defines as “pre-determined algorithms designed to pull the needles out of the haystack for the organization.”

One algorithm example is vendor performance. Usually, materials managers review their vendors only when a contract is expiring. Spend Compass flags performance criteria every month, highlighting vendors who have fallen short.

“It’s a much more proactive way to manage vendor relationships,” Wasilefsky says.

In addition to the electronic reports, Advisory Board Co. assigns an advisor who is a supply chain expert to each customer. Other hospitals have found data analytics help through GHX’s Contract Center, which is also available for a fee.

St. Joseph Health System, Orange, Calif., is a member of MedAssets but since 2006 has been using the GHX system. With 14 hospitals and annual purchasing volume of $495 million, St. Joseph saves $500,000 annually through data analysis.

According to Susan Wilson-Bromley, corporate director of data operations, Contract Center is worth the fee. “The knowledge and visibility of each item, each order, each price, minutes after the order is placed, has enabled us to correct potential invoice discrepancies before they happen,” she says.

Wilson-Bromley says GHX reps have worked with St. Joseph to customize reports to make them more meaningful at the IDN level, and has continued to enhance the system based on suggestions from hospital users.

Future prospects

If, as Plummer suggests, data analytics is only the first step in true optimization of  data to manage the supply chain, the next step will not be possible until data analytics systems are fully up to speed.

As Schwartz of Antelope Valley says, “Global Spend Analysis is a work in progress.”

Mitzel, who is generally happy with Summa’s data analytics system, is still looking forward to enhancements that will make it more useful.

He would like to see more attention to other variables in the data comparisons, such as product utilization and even clinical outcomes.

Mitzel believes that wider adoption of data analytics will make current systems more useful as well: “As the database becomes bigger, data becomes more relevant.” 

For now, taking that first step is important, he says. “There are a lot of tools available, but I think this is a critical one. Anyone who’s not doing contract management is probably struggling a little.”

Premier’s Marquardt says improvements are already on the way. One, just launched in November 2007, is a tier calculator that will tell users which tier they should be on. Premier also is working on customizing spend reports for each department such as cath labs.

Another enhancement due out soon is related to product conversion. “We need to assign responsibility for implementation of conversion. Materials management has to sell the department on changing products. Reports will monitor changes and track savings. It closes the gap between potential savings and implemented savings.”

At VHA, DeLay says he is aiming for “more functionality in the reports.”

The next upgrade, he says, will most likely be automated data transfer from the user’s MMIS. For example, a hospital could arrange to upload the file on the 30th of every month automatically.

Plans also call for the ability of reports to flag products a materials manager wants to track.

DeLay says data analytics has an even broader reach. “The difference between health care and other industries,” he explains, “is that in health care, we manage by initiative.

For example, we want to reduce anesthesia costs, so we do a project. Then we forget about it and move onto another project, say cardiology. Later, the previous savings erode. We need ongoing data management. Until now, we couldn’t do this because we didn’t have the data. Now it’s available.”

The ongoing reports make it possible to review spending issues continuously, rather than periodically.

When data analytics is fully accepted, DeLay says, “the world is going to look different in terms of materials management.”  

Paula DeJohn is a freelance writer based in Denver.

This article 1st appeared in the January 2008 issue of Materials Management Magazine.

To respond to this article, please click here.