The American health care delivery system is reprehensible for the degree to which it tolerates the under-treatment of those in need and supports the over-treatment of those who are entitled. It invests vast wealth in its own entropy. I don’t want to belabor all this shamefulness. The best we can do is to superimpose rationality on the current system—iron clad, science supported, and patient driven rationality with the goal of assuring health and providing recourse when that assurance falls short. We are advantaged by a cadre of physicians who are culled from the ranks of the best and the brightest and who would like nothing better than to do what is right by their patients. The moral charge to our society is to design a system that exists for no reason other than to provide for the wellbeing of both the sick people and the sick peoples amongst us1. To begin to do so demands confronting 3 of the current system’s most intransigent and least recognized moral lapses: licensing overtreatment, institutionalizing conflictive relationships, and promoting perverse incentives.
Failing #1: Putting the Health Care Cart in Front of the Health Care Horse
In the debates regarding health care reform, one goal promoted by all is “quality.” If only we could perform up to standards, the institution of American medicine would be supreme and vindicated. That being the case, why is it that eliciting “quality” in the treatment of heart attacks with carrots and sticks does not improve outcome?2 I argue that the American approach to coronary artery disease is unsurpassed in terms of Type II medical malpractice, my term for doing the unnecessary even if it is done very well.
Quality of performance is a compelling goal only if the performing benefits patients! The majority of the “health care dollar” expended in the United States does not benefit patients. So many high ticket items that are trumpeted as the triumphs of U.S. medicine are little more than a scam. The first priority for reform is to revoke the licensure of the profitably useless. We have the science to do so, as I detailed in an essay posted on The Health Care Blog last month. Furthermore, just as we have the scientific methodology and the moral imperative to demand clinically meaningful efficacy before pharmaceuticals are licensed, we have the scientific methodology and moral imperative to demand as much of devices and elective procedures before they are foisted on our fellow human beings.
We also have a moral imperative to assist patients in making informed medical decisions. Some aspects of enabling to this end are more quantifiable than “bedside manor.” A recent Randomized Controlled Trial (RCT) of the different ways to express evidence of benefit is an example.3 A large number of people were asked to place a value on a drug taken for five years to prevent heart attacks when the effectiveness was described in three different ways: postponement of a heart attack by two months for all patients, postponement by eight months for one in four patients, or sparing one in every thirteen patients from a heart attack. When the data on effectiveness was presented in the first format, 69 percent of the subjects said they’d take the pill; in the second format, 82 percent consented; and in the last format, 93 percent consented. Of course, all of these estimates of risk-reduction are comparable. But the last format, the “number needed to treat” (NNT) proved most persuasive.
Cognitive psychology suggests there are two influences on the way we comprehend risk4. We bring an analytic system to bear that takes advantage of whatever mathematical skills we can muster. We also bring a good deal of emotion to the table. NNT is the metric that resonates with most people’s risk construct. I personally prefer knowing the actual data and deriving the absolute risk reduction. But a 2 percent absolute risk reduction implies an NNT of about fifty and a 5 percent reduction is an NNT of about twenty. So the subjects in the RCT and I play in the same ballpark.
Failing #2: Contract Research Organizations (CROs)
A New Drug Application (NDA) to the FDA is a weighty document. At its core is the analysis of at least one RCT, which purports statistically significant evidence that the new drug is not a placebo and, increasingly, that it as good as if not somehow better than an alternative drug. Today, pharmaceutical firms outsource much of the application, including the conduct of the RCT, to CROs, spawning a multi-billion dollar service industry. My qualms about the marriage between pharmaceutical firms and CROs relate to conflicts of interest. Science tests null hypotheses: a negative result is assumed and a positive result is a surprise begging further analysis. The pharmaceutical business model wants a positive result and is disappointed when none is forthcoming, often a very costly disappointment. One might well imagine that the corporate responses of the manufacturer and the CRO to a positive result would be more conducive to negotiating another contract than a negative result. Given the judgment involved in analyzing the enormous data sets generated by large RCTs, one need not postulate malfeasance to imagine some prejudice at work. There is no mystery that whenever one compares RCTs supported by industry with RCTs supported by public moneys for the same drug, same class of drug, even for meta-analyses on the same class of drug, the industry supported studies are more likely to be positive. We are institutionalizing conflicts of interest, which percolate down to the academic health centers5 and on through the “thought leaders” to the practicing doctors6. Pharmaceutical firms are forever entrenched in the private sector. But just as the licensing of their products is public sector function so, too, should RCTs be a public sector responsibility. We should establish Clinometric Centers at medical schools and national RCT steering committees. All these entities should be staffed by academicians committed to the relevant issues in epidemiology. There is no shortage of willing and competent academics. Funding for staffing should be from federal coffers; the cost will be more than compensated by the savings that derive from keeping trivially effective and me-too drugs from the nation’s medicine chests. The steering committees would prioritize RCTs by the promise of and need for a new agent, and offer the Clinometric Centers the opportunity to bid to participate. The Clinometric Centers would function as public CROs and as the only agent to prepare and submit an NDA.
Failing #3: The Insurance Industry’s Many Moral Hazards
There is a general belief that health insurance companies have our interests at the forefront of their business models. For me, no such naiveté remains. The majority of the people in the United States who have health or Workers’ Compensation insurance are insured by their employer and serviced by their insurance company. The insurance company negotiates a fee, some percentage of the total cost, and takes that off the top of the money that flows from the employer to the providers of health care. The insurance company is a cost-plus provider at the service of the employer, not the insured. A minority of Americans purchases policies for themselves at great expense since their insurer has no “deep pocket” employer assuming the risk.
This creates perverse incentives for the insurer. The more money spent on “health care,” the more money flows into the coffers of the insurer to pay the enormous salaries and other aspects of “overhead” that are features of this industry. Why do you think the health insurance companies are so willing to promote screening programs and to underwrite the most expensive of me-too agents and unproved devices and procedures? Nearly all screening (from cholesterol to PSA to mammography to HBA1c to…) leads to little if any benefit for the screened, but to largesse for the screeners and more for the underwriters. If we stopped underwriting such examples of Type II Medical Malpractice as interventional cardiology for coronary artery disease or surgery for regional spine disorders, we’d discover that the system would fare poorly but the patients would be better off
Onward toward Assuring Health and Insuring Disease
It will require monumental public enlightenment to overcome the entrenched fiscal and political constraints to rational health care. In order to serve the health and welfare of the citizenry we must redirect enormous sums of the current “health care dollar” away from underwriting profitable futility. This will dramatically affect the incomes and lifestyles of almost everyone gainfully involved in the status quo. But there is no other way. I am proposing a national system that operates as a conglomerate of independent units. This has nothing to do with market considerations; I have no need to demonstrate yet again that free market principles do not apply to health care in the United States. Rather, the multiplication in the system is to serve the range of philosophies of “health care” that this reform countenances.
Each state will create a fund, an Enablement Fund. It may be more palatable given the precedents if this fund is a percentage of wages, 12 percent, but I prefer a comparable fund derived from income taxation. Tax revenues will be tapped to cover those who are not wage earners or the dependents of wage earners. The moneys will distribute to two accounts, a Disease Insurance Account and a Health Assurance Account. The Disease Insurance Account is a shared risk pool designed to supplant health insurance as we know it until one is eligible for federal Medicare insurance. The Health Assurance Account is not a shared risk pool. It is the repository for all moneys not expended from the Disease Insurance Account to be managed as are state pension funds. Ownership in the Health Assurance Account is prorated to contributions, perhaps progressively. The Health Assurance Account is available for licensed services in the state relating to health that do not qualify for underwriting by the Disease Insurance Account. This will become clear below.
Disease Insurance Account: The intellectual engine for managing the disease insurance account is the same Clinometric Centers discussed above as participating in drug and device licensure trials. As with the trials, the underwriting process must be as free of conflicts of interest as possible. There are two safeguards built into the system. One relates to the ability to compare administrative decisions across the states. The other follows from the Clinometric Center’s fiduciary role. The size of the Enablement Fund is fixed by the GDP of the state (or cluster of states). In administering the Disease Insurance Account the Clinometric Center cannot spend more than the fund has available. Furthermore, the administration cannot “save” money; it can save “your” money as money not expended in the Disease Insurance Account reverts to the Health Assurance Account. Nonetheless, there will be penalties for accepting gifts, consultative fees, and similarly morally bereft activities.
The guiding principle of the disease insurance account is to indemnify all that is demonstrated to be efficacious. The exercise is to calculate NNTs by taking advantage of evidence statements from the Cochrane Collaboration, the ACP Journal Club, the FDA’s trial data repository, and whatever else might be relevant without duplicating these available and ongoing exercises in evidence based medicine. Agreeing on prerequisite NNTs for effectiveness requires a good deal of consensus building in each Clinometric Center. Personal philosophies must confront statistical uncertainties:
- There is variability in the magnitude of effect between the trials that find evidence for an effect. Hence, there is often a range of NNTs.
- What NNT does one need in order to declare the effectiveness clinically meaningful? I would argue that an NNT of fifty is the upper limit of reliability and of meaningfulness when the outcome is “hard” (death, renal failure, myocardial infarction, etc.) I would personally be more comfortable with an NNT of twenty or less for my cutoff. Furthermore, the effect has to be clinically meaningful in the context of life course epidemiology. If the intervention spares one in twenty from death by colon cancer but spares none of these from death from something else at the same time (all-cause mortality), the clinically meaningful NNT is not twenty since there is no meaningful result. Death is death.
- NNTs for “soft” outcomes are more problematic. How much better do you have to make patients feel, and how many patients, before you would consider the effect likely to be reproducible let alone clinically meaningful? That depends in part on the reliability and validity of the measure of subjective improvement. If there is a reasonable measure, how many more patients must improve when treated with an active agent rather than with a placebo, or an old standby? Again, consensus must be reached as to the quality of these measures and their relevance to clinical effectiveness before analytic modeling is sensible8.
- The analyses of NNTs must be repeated periodically as new data appear.
- All these considerations feed-back into the design of future drug and device trials. Today’s enormous trials seeking effects that are likely to be considered trivial will become history.
- A good deal of the practice of medicine is “common practice”. Much has never been subjected to scientific testing and may never be. Is such to be indemnified and to what extent?
Speaking of money and resources, there is no doubt that there will be enormous savings when the Enablement Fund replaces the current mess. The bizarrely expensive business of NDAs will be streamlined. The marketing and “educational” budgets of manufacturers and providers will become transparent and easy to regulate. Nearly all today’s elective high ticket items will no longer be underwritten. Furthermore, starting this account from scratch allows for innovative and cost-saving IT that will eliminate a good deal of the current administrative overhead. There is no reason to tolerate the repugnant incomes garnered by many employed by the “health care industry.” The disease insurance account can stipulate the overhead percentage it is willing to underwrite.
Notice I am not recommending the vaunted “co-pay.” If the intervention is effective, it’s provided. However, I predict a backlash from all the purveyors and consumers of the ineffective, a backlash that could prove litigious. My solution is to have two levels of co-pay: zero or 100 percent. The latter is the co-pay for anything deemed to offer inadequate effectiveness; the discretionary fall-back for the claimant is the funds available in the Health Assurance Account, as we’ll discuss. Notice, too, that I am not countenancing stratified health care. The Disease Insurance Account underwrites what is effective for patients whether employed, unemployed, injured, or disabled. What has been indemnified by Medicaid and Workers’ Compensation will come under the purview of the Disease Insurance Account. Ethical medical ministration knows no such distinctions.
Health Assurance Account: Health is difficult to define. It is not the enigma it once was thanks to life-course epidemiology.8 Health in the resource advantaged world requires the wherewithal to feel comfortable in your skin. You need to find a station in your community that satisfies to an important degree. You need to be able to maintain that station by virtue of the earnings from work that satisfies you to an important degree. You need to have the personal resources to cope with the predicaments that are bound to come up. High on this list of predicaments are the occasional encounter with painful episodes. Health does not require a low cholesterol, low blood pressure, low body mass index, normal kidney function, normal cardiac output, etc.
The Health Assurance Account exists to foster health. It accumulates based on individual contributions to the Enablement Fund minus the expenditures of the Disease Insurance Account. It is yours to do whatever you wish related to health that is provided by licensed professionals in your state. If you never touch it, it will revert to your pension at retirement. If you are not convinced that cholesterol screening or coronary artery stents are as ineffective as determined by the Clinometric Center, you can use your Health Assurance Account funds toward the purchase of such.
The Health Assurance Account will have an advisory component. If you are an entry-level employee, it might be very wise for you to expend Health Assurance funds for instruction in English as a second language or to acquire skills that afford job mobility. These moneys are available to aid you if you don’t otherwise have the wherewithal for attaining a station in life that promotes your healthfulness. There is also a role for the Health Assurance Account in educating the public as to the hazards of not attaining a satisfying station in life.
Rational Health Care is NOT Rationing
It’s a very long tunnel, but there is a light at the end.
Each of the following books harbors earlier iterations of the Health Assurance-Disease Insurance Plan. In each Dr. Hadler provides extensively referenced discussions of its scientific underpinnings:
The Last Well Person. How to stay well despite the health care system.
Worried Sick. A prescription for health in an overtreated America.
Stabbed in the Back.
Confronting back pain in an overtreated society.
1. Bayer R, Gostin LO, Jennings B, Steinbock B (Eds). Public Health Ethics. Oxford: Oxford University Press; 2007.
2. Glickman SW, Ou F-S, DeLong ER, Roe MT, Lytle BL, Mulgund J, Rumsfeld JS, Gibler WB, Ohman EM, Schulman KA, Peterson ED. Pay for performance, quality of care and outcomes in acute myocardial infarction. JAMA 2007; 297:2373-80.
3. Halvorsen PA, Selmer R, Kristiansen IS. Different ways to describe the benefits of risk-reducing treatments. Ann Intern Med 2007; 146:848-56.
4. Slovic P, Finucane M, Peters E, MacGregor DG. Risk as analysis and risk as feelings: some thoughts about affect, reason, risk and rationality. Risk Analysis 2004; 24(2):1-12.
5. Ehringhaus SH, Weissman JS, Sears JL, Goold SD, Feibelmann S, Campbell EG. Responses of medical schools to institutional conflicts of interest. JAMA 2008; 299: 665-71
6. Hadler NM. Ethics Forum. Would physician disclosure of all industry gifts solve the conflict-of-interest problem? Negative Constructive. American Medical (AMA) News, January 7, 2008.
7. Braithwaite RS, Roberts MS, Justice AC. Incorporating quality of evidence into decision analytic modeling. Ann Intern Med 2007; 146:133-41.
8. Marmot M. The Status Syndrome. How social standing affects our health and longevity. New York: Henry Holt; 2004.