|
Conflicts of interest
Clinical research funding supported by pharmaceutical companies in the United States grew from $1.5 billion in 1980 to $22 billion in 2001.
There are significant benefits of pharmaceutical funding of research:
- money for research into improving treatment
- some support for basic research
- emphasis on applied research
- competition to produce the best product
- support for academic medicine
- support for individual researchers
As funding from the National Institutes of Health has declined, pharmaceutical companies have filled the gap left by declining federal funding for science. In this situation, journalists are understandably interested in whether researchers are motivated by money or by higher desire to serve mankind.
Conflict of interest exists when a primary ethical or professional interest clashes with (financial) self-interest. Doctors’ professional duty is to the patient or to the best, unbiased research.
There is potential for conflict of interest in two forms of support for academic research:
- research funding
- “soft money”—honoraria, speaking fees, consulting contracts. All universities have rules about this area, such as a maximum amount that can be accepted, but none have outright bans [on accepting soft money]. The more prestigious the doctor or researcher, the more likely it is that the pharmaceutical company [pursues him or her].
Both kinds of support can ultimately affect clinical practice.
Industry uses dollars to
- bias studies to sell products
- exert control over published research, which guides practice
- influence FDA decisions and clinical guidelines setting, and
- influence continuing education.
In a survey asking doctors whether their prescribing behavior was affected by gifts they received from pharmaceutical companies, only 9% of doctors said gifts influenced their own behavior but more than 70% believed such gifts had impact on their colleagues’ choices.
Even former Surgeon-General C. Everett Koop [accepted money from industry]. He [supported a study claiming that] latex gloves don’t cause allergies. It turns out he got $1 million from the latex industry. But he denied that the funding influenced him, saying, “Maybe some people would be affected by a million dollars, But not me. I have a reputation to uphold.”
Industry knows what it is doing. “If you want to prove a point, fund a study,” said a pharmaceutical representative.
Gifts are effective because reciprocity is a powerful human desire. Even a pen with a Paxil logo can have an influence, in that it’s seen as a gift and something is owed to the giver. Sometimes gifts are much larger, too.
Another issue is that continuing medical education courses (which doctors need to maintain licensure or certification) are often little more than public relations events for drug manufacturers. In fact, the distinction “is becoming obsolete,” according to Neil Kendle of Lowe Fusion Healthcare [a leading communications consultant to the pharmaceutical industry].
The pharmaceutical industry funds 80% of clinical research in this county, and most of these studies are done because a company wants to expand the marketing of a drug [or device] to doctors and the general public.
How do conflicts of interest and funding sources affect research outcomes?
A 1986 study found that “published results of research supported by pharmaceutical companies tend to favor new drug treatments over traditional therapy and patented products over generic.” (Davidson, RA. Source of funding and outcomes of clinical trials. J Gen Intern Med 1986;1:155-158.)
A later meta-analysis came to the same conclusion, noting that “strong and conclusive evidence shows that industry-sponsored research tends to draw pro-industry conclusions” and that “industry preferentially supports trial designs that favor positive results.” (Berkelman, JE et al. JAMA 2003;289:454-465.
An article about how financial conflicts influenced outcomes of clinical trials involving calcium channel antagonists found that authors are more likely to be positive about a sponsor’s drug if they get money from them. (Stelfox HT et al. New England Journal of Medicine, 1998;338:101-106.)
Results of sponsored studies influence drug-buying choices in two main ways
- Doctors are influenced by continuing education courses, visits from pharmaceutical reps, journal articles, and display ads in journals.
- Advertisements on TV and in print urge consumers to request specific drugs.
What can be done [to change this pattern]?
One option is to increase Federal funding for medical research; another is to cut way back on the amount of academic research – which would not necessarily be bad because much of it is not really good science, anyway.
How does sponsorship shape results?
Sponsorship can introduce bias into studies in many different ways. Some of these are
- Recruiting patients
- Randomization to treatment groups or treatment vs placebo
- Blinding of patients and researchers
- Analysis
- Publication
The industry-supported research manages to distort results in two main ways: (1) hiding in plain sight, and (2) suppressing data.
Ways of hiding in plain sight:
- randomization bias
- subgroup analysis, data dredging, post-hoc analysis
- straw man comparators
- combination endpoints
- shifting goal posts
- publication bias
- underpowering
- segregation of data
- surrogate endpoints
- failure to look at all-cause mortality
In an industry-sponsor clinical trial examining possible links between antidepressant treatment and suicidal ideas in adolescents, the study was designed to make the drug look as benign as possible. Youngsters who rapidly developed bad reactions to the drug during the initial “screening” phase of the study were eliminated from the experiment and set aside. The remaining participants were randomized to treatment or placebo groups, and later the teens who had been set-aside were integrated into the placebo group. Data from the “screening” phase were excluded from analysis
The published article reported that “most adverse effects were not serious,” although the reality was that seven children had to be hospitalized for side effects. A sales memo from the drug company [Pfizer] claimed “remarkable efficacy and safety in the treatment of adolescent depression.” Reviewers missed this design flaw, perhaps because they don’t give careful scrutiny to results that they expect to see – such as “antidepressants are safe.”
What I take away from this is that we should be skeptical about studies of antidepressants in kids, who are especially likely to be labeled and medicated for psychiatric and learning disorders.
Another way to hide data in plain sight is to make “straw man” comparisons:
- The sponsor’s drug will always look better compared to a placebo than to an existing treatment for the same ailment.
- If the goal of a head-to-head comparison is to demonstrate efficacy, use a low dose of the competing drug.
- If the goal is to demonstrate safety, use a high dose of the competing drug.
Risperidone and other “atypical” antipsychotic agents were hailed in the 1990s, at least by one researcher with a conflict of interest, as “a potential breakthrough of tremendous magnitude.”
In reality, these newer drugs are no more effective than the old ones, but they did have different side effects – including potentially lethal weight gain and diabetes.
The artful use of relative and absolute expressions of risk also played a role in the selling of risperidone. One 1997 journal article touted a 26% decrease in in-patient days, which sounded good until closer examination showed that the change was from 5 days to 3.7 days – which is clinically insignificant. Nor surprisingly, three of the 6 authors of this study worked for a “research foundation” set up by the drug manufacturer.
Major depression and similar conditions are serious diseases that need treatment, and people are desperate for appropriate care. Yet it is easy for doctors to mis-prescribe for these patients, because they are bombarded with biased information from companies, who send drug reps to their offices.
The public and professional right to know
Suppressing data is another way of distorting research findings.
Publication bias is one aspect of that, and it occurs when industry manages to:
- nix negative trials
- selectively report only positive results
- republish positive results
Data revealing negative aspects of the sponsor’s drug often go unpublished because the company is the legal owner of this information and can choose to keep it quiet.
Corporations have so much say in what gets published and what doesn’t that the former [CQ—or current?] editor of The Lancet, Richard Horton, said in 2004 that “Journals have devolved into information-laundering operations for the pharmaceutical industry.”
The growth of market-driven research has brought about
- more me-too drugs,
- more polypharmacy (concurrent treatment of one patient with multiple medications)
- longer term use of drugs
- more inappropriate use of drugs
- lower treatment adherence
- more risk of side effects
The overall impact is that private industry now leverages the direction of more than 80% of clinical research.
Everyone is accountable to someone else, and its easier to think critically about the dynamics of modern clinical research if you remember the types of fiduciary responsibility involved:
- Doctors are responsible to patients
- Journalists are responsible to the public
- Government is responsible to the public
- Lawyers are responsible to clients
- Pharmaceutical company CEOs are responsible to shareholders
Tips for reporters: here are good questions to ask whenever reporting on a clinical trial:
- Who funded the study?
- Who was in charge of the data?
- Who wrote the paper?
- Do you have financial conflicts of interest with the manufacturer?
- Do you have any conflicts of interest with any drug or biotech company, and if so, which ones and what type?
- Who might have a different interpretation of your results (who is your harshest critic)?
How can we improve the conduct of clinical research, so that the public benefits more and the companies less? Here are two ideas: (1) provide more government funding and (2) channel money being spent on unnecessary or ineffective treatments (such as statins for people with no history of heart disease) into research.
Cancer Screening
Here’s a story that reporters in any community can write: Doctors are using more and more ancillary services to replace revenues lost because managed care plans, including Medicare, limit the amount physicians can be paid for office visits.
Doctors are buying or leasing MRI machines and CT scanners, and the number of mobile “whole body” CT scanners is increasing. These imaging tests are among the fastest growing costs in U.S. medicine. Payment systems are driving doctors forward to doing this; doctors profit from these services. An example is the rise in whole body imaging. It’s a bad idea: there’s no value from the imaging, it does much harm, and it raises costs.
In terms of cancer screening, it’s important to differentiate between what makes sense and what is known. [Theoretically, we] screen for cancer [on the assumption that] if we catch it early, we can save lives—that might or might not be true and might be true for some people [but not others].
Clinically obvious cancers are the tip of the iceberg. Screening allows us to look under water and see lots and lots of tiny cancers with no clinical manifestations as yet. The theory is that if only we see the cancer early, we can remove it and solve the problem. But the natural history of these two groups of tumors may be very different.
Let’s look at the prostate cancer “pseudoepidemic’ in the United States. The incidence of prostate cancers detected increased dramatically with PSA screening. Is it really more cancer or just more screening?
More than a million men have been diagnosed with prostate cancer who wouldn’t otherwise have been diagnosed without the PSA test. However, during the same period, deaths from prostate cancer remained stable. So, with the PSA test, we are catching more cancer, treating it early, but the death rates haven’t done down.
If we’re catching so many cancers so early, why is there no improvement in death rates? Three types of bias in cancer screening studies might explain this:
- Selection bias (just screening unhealthy people)
- Lead time bias (screening but not making no difference in outcome)
- Length bias
Of the three, lead-time bias is the most likely in this case.
With prostate cancer, whether you get early treatment or late treatment doesn’t seem to make any difference. So, if there’s nothing to do, why are we screening at all?
We are diagnosing many people, some of whom would have survived anyway, no matter whether they were getting treated. Many have survived prostate cancer and died of other causes.
Some cancers are rapid progressing and others are slow progressing. Slow-growing tumors are caught well via screening, but rapid ones are not. Consequently, we are treating many people with cancer who might not become sick with the cancer.
Comparing the rates of PSA screening in the United States and The United Kingdom between 1968 and 1995, we have more cases (because we do more screening) but the mortality is the same in both countries.
The problem is that we don’t know that PSA screening doesn’t work—we need randomized clinical trials to determine.
But we also need to be clear that we are not necessarily reducing deaths with treatment, but the treatment [has serious side effects like] a high rate of causing incontinence, impotence, and fecal problems. In our quest to reduce mortality, we may be causing more morbidity.
Is the screening worth doing? In the interim, how to deal with the question? Doctors and patients should talk honestly about decision-making. Patients should take advantage of resources that can help them choose the best individual course of treatment. [Peer support groups are available in many locales and some hospitals have interactive computer programs that help patients weigh pros and cons of various treatment.]
The issues are similar for ovarian and lung cancer screening and mammography—the evidence goes both ways. We don’t know the answers about colonoscopy—the evidence is not clear.
Advice to journalists: if you are doing screening stories, seek conflicting views. Ask hard questions re benefits and risks. What happens if the sources lie? Seek unbiased sources of information. Suggested sources for getting straight answers include:
- www.overtreated.com for a list of medical experts with no conflicts of interest
- www.cspinet.org for information on integrity in science, including researchers’ ties to industry
|
return to Knight Health home 