- Physician Community Needs Assessments: "Lies, Damn Lies and Statistics?" Part II of II
- August 29, 2008
- Law Firm: Holland & Hart LLP - Denver Office
Recently, a national chain hospital located in a highly competitive market in Los Angeles, California used a home office prepared CNA developed in another state to attract an invasive cardiologist to move to L.A. and start up a practice. The corporate recruiter announced that he had given the physician a copy of the CNA to "hammer home" the need. Neither the CEO of the hospital or even the recruiter had much of an understanding of the methodology utilized in the development of the CNA. They knew only that they could look through the two-inch thick document to two particular pages where corporate had concluded that in 2007, based on 2003 data there would be an objective need for 7 to 8 additional cardiologists in the hospital’s GSA. This conclusion was based solely on statistical data and benchmarks from a number of different sources and virtually no anecdotal or qualitative survey confirmation in the market whatsoever.
The weight of the evidence in the CNA including many multi-colored pie charts and graphs was impressive, but misleading. There were many pages in the CNA devoted to HPSA’s and MUA maps in or near Los Angeles, but there were no HPSA’s or MUA’s in the hospital GSA. There was page after page of demographic data in bar charts and pie charts, but none of this data linked in any way to the need conclusion.
The actual methodology involved the computer "slicing and dicing" of U.S. Census data for the service area available from commercial sources by zip code and the application of various physician to population ratios to determine manpower by specialty needs in the district based on the population size, then applying an anticipated growth rate factor to determine the level of demand in 2007 for cardiologists.
This approach while ostensibly "objective" on the surface contained a number of significant flaws in logic and application. Starting with the identification of the existing cardiologists in the market, the names of cardiologists with office address located in the GSA, the data did not identify cardiologists whose office addresses were outside the GSA, but who were serving people within the GSA. The data did not account for the potential compensating "oversupply" of cardiologists in the next zip code because of the exixtence of a larger, more prominent cardiology center, with a statewide, even regional draw.
Accuracy in the physician count compilation in the GSA from all available sources is key to the integrity of the study.
No projection of current and future physician need can be accurate without this core data.
* * *
The only way to conduct an accurate physician count in some medical markets is to profile each physician’s practice, usually through telephone canvassing of the physicians or their office administrators.
Moody, "Demonstrating Community Need for Physicians". Health Lawyer’s Weekly, October 10, 2003, Vol. 1, Issue 27.
The "benchmark" data regarding community demand for physicians was dated and widely divergent in conclusions. The CNA compiler calculated the median conclusions of need from these sources in order to be "conservative" and to deflate the "outliers." The basic question remains as to whether this median had any meaning. The data identified the current existence of 18 cardiologists in the market. The six sources of benchmark data bases concluded that based on their ratios, there was a current range from an oversupply of 1.5 cardiologists to a current "need" for 20.1 (well over twice the total identified supply of cardiologists identified as being in the GSA from the public sources used). This wide band of contradictory conclusions begs the question as to fundamental validity of the data. The limited number of samples (6) also gives concern. (If you flip a coin only five times and it comes up heads on 4 occasions, is it reasonable to conclude that there is an 80 percent chance of it landing on heads in the future or is the change likely to approach 50 percent at 100 tosses? In other words, the larger the sample, the greater the validity).
One might suspect some problem in the data when it suggests that there was a current need for well over twice the number of cardiologists in the GSA, without a public outcry and a major boom in cardiologist compensation in the area. (Neither circumstance surfaced). In 2005 the compiler arrived at a 2003 median need in the GSA of 3.3 physicians based on 2003 data and projected an oversupply of 7.7 in 2007. The data did not differentiate between cardiologists, interventional cardiologists or cardiovascular surgeons. It just lumped them together. This was because of the lack of differentiation in the underlying benchmarks themselves. There is a considerable difference if the need, if any, relates to cardiovascular surgeons rather than say an interventionalist.
One of the benchmarks utilized was the Graduate Medical Educational Adversary Committee Study (GEMENAC). This is a survey adopted in the early 1980’s to project the future needs for medical services based upon theoretical models of per capita healthcare consumption. It was specifically identified as one of a number of criteria for determining community need for physicians in the famous Herman Hospital closing agreement with the federal government. Another study by Hicks and Glenn, of the University of Missouri, used Federal data regarding patient visits per specialty in a population of annual patient visits. They then divided the result by the number of annual patient visits seen per specialists based upon a Medical Group Management Association’s count to arrive at a number representing the appropriate physician to population ratio.
The reliance upon these statistics to determine Community Need cannot reasonably be used alone without further context to confirm the existence of an existing demand.
The fact that GEMENAC ratios pre-date the proliferation of managed care and that the population has aged considerably since 1980 only partly explains why these and other ratios are problematic. The real reason that they are not definitive is that they suggest standardization is achievable in the patchworks of healthcare markets that make up a most heterogeneous whole.
* * *
Physician-to-population ratios, then, are only a signpost or general indicator of physician need, rather than a definitive benchmark. A reliable projection of community need requires a thorough analysis of local conditions… ."
Moody, "Demonstrating Community Need for Physicians," supra.As a case in point, the CNA made no adjustments for a 17% Kaiser Permanente market share in the hospital’s GSA. Kaiser uses a physician population ratio for cardiologists that is only 43 percent of the national ratios identified by the Governmental Accounting Office. See Weiner, Jonathon P., "Prepaid Group Practice Staffing And U.S. Physician Supply: Lessons For Workplace Policy, Health Affairs, 4 February, 2004, at 43-49, Ex. 42 at 46. The Kaiser physician practice data is compelling.
Even though PGP’s are a relatively small segment of U.S. practice, managers and policymakers have looked to them since their inception as an example of how best to "staff up" to provide care for defined populations. Within the disaggregated U.S. health care system, there are few other instances where a "numerator" of providers and a "denominator" of consumer patients are so clearly demarcated.
Weiner, supra, at 43.
The failure to research and validate statistical findings with the existence of actual indicia of demand in the GSA may render a CNA as practically useful for regulatory and recruiting purposes as the Los Angeles phone book. Failure to access the input of local physicians is inexcusable if the CNA is going to be relied upon for any substantial purpose in regulatory compliance, not to mention as an inducement for prospective physician recruits.
Local physicians usually have the best insight into local demand for medical services. Physicians know if there are long wait times to schedule their patients with specialists. They also know which specialty services their patients need that currently are not available in the community. A survey of established physicians can reveal how busy they are relative to national averages, whether their practices are open to certain populations such as Medicare patients, and what recruitment needs they see in the community.
It is perhaps emblematic of CNAs based solely upon statistical data that the real purpose and design of the drill is to provide statistical cover for the recruitment of physicians to enhance the competitive standing of the recruiting hospital without a real concern for the existence of a real need for that physician’s services. A recruitment in those circumstances substantially increases the risk of the already dicey investment of a physician in a start up practice in the GSA and the potential of a fraud in the inducement claim being brought against the hospital when the physician practice fails and the hospital is confronted with the need to sue the physician to recover the initial subsidy when she leaves the market in order to survive. The practice also leaves the hospital with the potential of explaining to the Office of Inspector General how and why its CNA, in the parlance of Benjamin Disreaeli, is more than "lies, damned lies and statistics."