Tuesday, December 21, 2010

Un dialogue de sourds

In October, the Medical Device Manufacturers Association (MDMA) placed another shot across the bow of Group Purchasing Organizations (GPOs), by publishing a study that purported to show "... that GPOs fail to deliver on their primary justification for allowing them to receive kickbacks from suppliers, and that they do not in fact reduce costs relative to a world in which GPOs are funded by hospitals ...", and that the "... elimination of GPO protections could reduce health care costs up to $37.5 billion each year ..."

First, some general background on GPOs:
Derwood Dunbar on GPOs
Hospitals, group purchasing organizations, and the antitrust laws - Business
Group Purchasing Organizations – Seriously Reviewed By GAO and Others
Group Purchasing Organizations: An Evaluation of Their Effectiveness in Providing Services to Hospitals and Their Patients
GPOs save money web site - Understanding How Group Purchasing Organizations Work & Understanding GPOs and the Safe Harbor Provision

... and related to this particular spat:
War of words heats up in GPO controversy
Hospitals slam anti-GPO report
Nation’s Largest Hospitals Respond to Recent Medical Device-Sponsored Study

The MDMA's charges brought a swift riposte... A number of VHA members wrote a letter to the MDMA stating that its conclusion was "completely unbelievable." Further, they charged that "... contrary to your report’s assertion, changing the current GPO funding mechanism would mean additional costs for hospitals and other health care organizations, thus driving healthcare expenditures even higher. Neither the healthcare community nor the country as a whole can afford the type of disruption and increased cost that your organization is advocating..."

Other GPOs characterized the MDMA report as "propaganda" and as a "slap in the face," and accused the MDMA of self-serving, for example: "The only parties that stand to benefit from disrupting a working, competitive GPO market are the manufacturers who comprise the $200 billion medical device industry. They continue to attack the GPOs because what they do works. GPOs effectively drive prices down for our hospital members and no amount of MDMA propaganda will change that."

OK, so looking past the allegations and counter-allegations, past charges of "kickbacks" and "propaganda", etc., where does the truth lie? First let us look at the MDMA report.

The report analyzed over 8,100 historical records of capital purchases (of biomedical, dietary, imaging, information technology, laboratory, laundry, monitoring, oncology, and a number of other equipment categories) from a company called MEMdata, looking at medical devices purchased by multiple hospitals; and comparing the prices actually paid to GPO contract prices for the equipment. Crunching the numbers they found that the hospitals paid the GPO price 21.7% of the time; paid less than the GPO price 76.9% of the time; and (somewhat inexplicably) paid higher than the GPO price 1.4% of the time. Overall, hospitals were able to save between 10% and 14% off GPO prices, and in fact incumbent (GPO contracted) device vendors lowered their prices by an average of 7%. This situation was attributed to the financing mechanism - " ... economic theory explains that this failure (of GPOs to secure the best price) emerges from the way in which GPOs are compensated. The current system - which allows GPOs to share in the sales revenue of those with whom they are to negotiate for the best prices - creates a perverse incentive that limits the GPOs' ability to procure the lowest prices for its member hospitals..."

The MDMA then extrapolated from those numbers, concluding "... Based on the results from our empirical analysis, we conservatively estimate that changing the incentive structure by reapplying the anti-kickback statutes would reduce private U.S. health care expenditures by roughly $25 billion annually, and would reduce federal health care spending by roughly $11.5 billion annually." (i.e. a total savings of $36.5 billion.)

OK, so here is where the apples and oranges come in... The sample data that was analyzed was for capital equipment (device) purchases, and the savings factor calculated was applied to GPO medical/surgical supply volumes! So, what is the difference? Well, in addition to lowering overall transaction costs, in theory GPOs get better pricing by leveraging the larger, combined volumes of their members to achieve better pricing than an individual hospital should/would be able to achieve on its own. Now this works well for the members' medical/surgical supply purchases (whether they be 'commodities' such as needles, dressings, etc.; or physician preference items such as implants, etc.), as their purchase volumes can be reasonably well forecast. However, it is very different for capital equipment purchases - while a GPO can forecast and provide vendors with estimated volumes of supply purchases (which might be characterized as 'continuous') to achieve lower pricing levels, it is a different matter for equipment purchases (which are more "episodic" or "intermittent" in nature). Volumes here would depend on a number of factors, including depreciation and equipment life cycles, technology life cycles, differing fiscal years for various hospitals, etc. Once you buy, say, physiological monitors or syringe pumps, they last for perhaps five years, and in fact could be maintained for six, seven, or more... Thus while negotiating with an equipment vendor, the GPO is generally unable to forecast and leverage specific volumes (not really knowing which members will be in the market for particular models/modalities of equipment.... ). As a result, the GPOs' processes and resulting contract prices for capital equipment are more tentative, and most hospitals use the GPO contract price as a ceiling below which they negotiate individually with the equipment vendor.

The bottom line? The fact that most hospitals pay below the "GPO price" for capital equipment is in majority a reflection of this market weakness, and is not due to the GPO funding mechanisms, as posited by the MDMA study. Further, the assumption that the 'discount' achieved off GPO price for capital equipment can be similarly achieved on all supply purchases is speculative in the extreme. Reading the study it is clear that the authors make no distinction between supplies and equipment. For example, one of the authors has stated "We looked at the price of medical supplies (sic!) after GPOs have ostensibly secured best price. If original GPO auctions are done efficiently, there's no price room." Well, they manifestly did not look at the price of medical supplies, they looked at the prices of equipment purchases! And since the GPO negotiations for equipment purchases are not done as efficiently as those for supplies, there is "price room."

Having looked at the (flawed) MDMA study, next let us look at the response of the GPOs and hospitals. Rather than calmly rebutting the MDMA study by pointing out its flaws and/or weaknesses (e.g. as in the above), the response has been rather shrill... And some of the counter-claims that have been made seem a tad unlikely. For example, the hospital letter cited earlier claimed that " changing the current GPO funding mechanism would mean additional costs for hospitals and other health care organizations..." It is not at all clear why this is valid, and this seems to be an example of the GPOs engaging in "apples to oranges" comparisons of their own!

The graphic above (courtesy of 'GPOs Save Money') shows the funding flows. First, a GPO member hospital uses a GPO contract to purchase supplies from a vendor. The vendor periodically remits a percentage (up to 3%) of the purchase to the GPO in the form of administrative fees. The GPO funds its operations from the fees collected, then at the end of the year passes the "excess" back to the hospital, usually in the form of a cash distribution called a "patronage dividend" (or some similar term.) Note: the GPO will likely use portions of the administrative fees for items other than purely costs related to contracting activities; examples might include quality collaboratives, clinical and operational benchmarking, patient safety related activities, IT infrastructure build, and so on.

The percentage of the administrative fees that are 'surplus' and that are that are returned to the various member hospitals differs between the different GPOs, but an approximation of 50 cents returned on every dollar of administrative fees generated is probably in the ball park. As such, it is not at all clear why switching from the current 'administrative fee' model to a service subscription model would result in any additional costs for a member hospital. Initially the members' fees could be set at 60% of the current level of the administrative fees they are generating, to provide the GPO operational cost plus a small cushion (and they are generating these fees by paying them!). It is unclear to this blogger why this should result in the apocalypse that the GPOs claim would occur, viz. "... Limiting GPO administrative fees would require hospitals to pay more so that GPOs could continue to operate and negotiate on their behalf. This would require billions of additional dollars..." (see here). In fact it would make much more explicit the costs related to belonging to the GPO - the fee would be a clearly visible line item in somebody's budget vs. being embedded "invisibly" within, and spread across, every purchase transaction.

Now it is true that if GPOs did not exist and if every member had to have their own staff, etc. to negotiate their own contracts for supplies and equipment, then there would be a rather large incremental cost increase and overall transaction costs would rise due to a duplication of effort... However, no one (including the MDMA) is arguing for this, so it is rather a GPO straw man!

So, bottom line, while there is a solid argument to be made in changing the current GPO funding structure, the MDMA study is not it. And, the GPO special pleading is somewhat self-serving as well!

Quick note: The MDMA study does go into some other points... and also has some other statements that can be contested. However, for the purposes of this blog entry I am only addressing/discussing the main point made by the study...

No comments:

Post a Comment