Challenges Facing Comparative Effectiveness Research

 Healthcare reform has become a leading political issue and the use of competitive effectiveness research (CER) to reduce costs is embroiled in the debate. Comparative effectiveness is simply an evaluation of different treatment options available for a given medical condition. For example, a study may compare competing drugs, or perform an analysis to see if the benefits of a surgical procedure outweigh its risks. On the surface finding what medical treatments work and which ones don’t, would not appear to be controversial, but as noted in my last commentary, there’s plenty of disagreement when it comes to CER.

CER critics say it will lead to healthcare rationing, with government boards deciding what treatments will be paid for or denied and in effect dictating to doctors how to practice medicine, The other side argues that CER will avoid keeping the medical community in the dark about the relative value of different medical treatments, inform patients and providers so they make good decisions and substantially reduce government funding. They also point out that there was nothing in the enabling legislation about government controls interjected into the debate by the opposition.

Perhaps even more important is the fact that CER is not something new and it is not confined to clinical trials. Other research methods such as cohort or case-control trials can also be used and innovative analytical techniques that can maximize the value of information that has been gathered are also legitimate CER approaches.

There have been numerous CER trials and neither the disasters nor the promises made by proponents and opponents have occurred. But it is important to keep in mind the monumental task CER faces in terms of finding differences for available treatment options for drugs, devices, and procedures such as surgery. The cost and time to do the trials will be large. Most drugs are on the market because they were found to be more effective than placebo, but those differences were relative large and it will be harder to locate the smaller differences that exist among competing active drugs. Furthermore doctors want the testing done in more real world environments rather than the artificial settings used to gain a drug’s approval for marketing. They want a variety of patients tested not the homogeneous set used in the early clinical trials, the use of measurements that are meaningful to patients not those that are precise but impractical, an environment that corresponds to a clinical practice not one that is overly controlled and unrealistic, trials that treat patients for a long period of time not ones that end when a treatment difference first emerges. They want more information on sub-groups of patient broken down by age, gender, disease severity, length, ethnicity, concomitant disease, complications etc. But adding these contrasts will also mean there must be a huge increases in the required sample sizes. Under this tangle of factors that need to be considered, clear cut differences will be a luxury and proponents of treatments that come out second best will have ample evidence to challenge a negative finding. Add in the fact that there are many treatment options available and it should be clear that the task for CER is daunting but not overwhelming. The point to keep in mind is that the goal of CER is valid and making some progress is better than remaining ignorant regarding what is the best medical treatment option.

In addition, the results from CER must also affect clinical practice. The best way to accomplish that goal is also laden with perplexing problems, but that is a subject for a later commentary.

The Stimulus Bill and Medical Research

The medical field did well when money from the stimulus bill was dolled out. The bulk of the money was set aside to get medical information computerized, but who what have believed that medical research would be an important recipient as well. The economic recovery plan allocated $1.1 billion for a critical research area that has been woefully neglected – comparing the relative effectiveness of medical treatments. When the FDA approves a drug or medical devise, key questions remain unanswered – just how different is the new product compared to the current available therapies in respect to effectiveness and tolerance? Sadly, the answer is – we don’t know because the necessary studies were not carried out. Knowing what treatments work best and sharing that information with patients, their doctors and government healthcare officials woulkd appear to be a worthwhile goal that everyone would support. However, look at it from the viewpoint of the pharmaceutical company. Drug companies are not eager to have their drugs compared to a competitor’s – they’ve already spent millions getting the product approved by the FDA relying primarily upon placebo controlled trials so why now bare the cost of comparative trials that would be even more costly and run the risk of showing their newest product was second rate.

This issue clearly has political overtones as well. Conservative republicans don’t like the idea of comparative research. They have trouble with the government’s role because it would be one more intrusion into private industry by a federal agency plus there is the enormous cost for such trials that would have to be paid for by the government. They worry that comparative effectiveness may become cost effectiveness. They fear it will be used by insurers, private as well as governmental, to deny coverage for treatments that do not fare well in a comparative trial. Nonetheless, with the cost of healthcare relentlessly increasing and prescription drug costs representing a major factor in the increase, a way to rein in these costs will not go away.

SMOKING AND LUNG CANCER

Commentary – No. 1

Smoking and Lung Cancer

Can you imagine a headline that reads: “Smoking Does Not Cause Lung Cancer?” All of us would be stunned – are they nuts? That’s a ridiculous statement!

We all know smoking cigarettes causes lung cancer. It took a while to find this out, but even tobacco companies have given up fighting the claim. Early evidence that smoking had harmful effects was based on reports of lip cancer among pipe smokers as far back as the late 1700s. However, the connection wasn’t taken seriously. It was written off as a result of the heat produced by the pipe. In the early 1900s lung cancer, which had been a rare disease, was seen much more often. Cigarette smoking was identified as a possible cause. But since nonsmokers also developed lung cancer this possibility received little support. And look at all those people who smoked and didn’t get lung cancer.

It would have been nice to research this issue using clinical trials, but that was impossible. Can you guess why? In a clinical trial you recruit volunteers to serve as subjects and create two groups – a control group and an experimental group. Subjects would be randomized to one or the other group. All the subjects in the control would not be allowed to smoke and all the subjects in the experimental group would be required to smoke for the duration of the trial which could last years. This kind of study can’t be done. For ethical reasons you could not require people to smoke and on practical grounds you’d have a hard time keeping other people from smoking if they wanted to. Other approaches were required.

In the 1940s smoking was fine from an environmental, social and medical perspective. Cigarettes were advertised using sport stars and movie idols (including future President Ronald Reagan). An ad claimed over 10,000 physicians smoked Lucky Strikes and another proclaimed that while smoking Camels, throat specialists never have throat irritation. The message was clear: smoke all you want – it’s good for you.

However, this tranquil picture began to change during the 1950s with the publication of a number of studies that examined the medical records of patients who already had lung cancer to those of a group similar to the lung cancer cases, except they did not have the disease. Comparisons between the two groups revealed that smoking was present in a larger proportion of the cancer patients. In another pivotal study, researchers followed British doctors without lung cancer to see how they fared. Almost 70% of the physicians in the country participated and were observed for 2 ½ years. Dividing the doctors into smoker and non-smoker groups, the rate of lung cancer was highest among the smokers. A similar but even larger study in the U.S. came up with a comparable result.

The evidence against smoking became even stronger in the late 50s and into the 60s. The research condemning smoking involved different settings and countries - Denmark, France and Japan. The period of observation was also extended reaching upwards of 10 years. All studies came to the same conclusion – smoking tobacco increased the risk of getting lung cancer. Furthermore, studies consistently showed that the more a person smoked, the greater the risk of lung cancer. They also demonstrated that the longer a person smoked, the greater the chance that they’d get lung cancer. Major research organizations in the U.S., U.K. and Canada concluded that there was a casual relationship between smoking and lung cancer. The weight of the evidence became overwhelming placing pressure on governments to act. In 1962 the U.S. government took action , the Surgeon General officially announced that smoking was a major cause of lung cancer and required warning labels on all cigarette packages.

End of story? No, not really. Could there be some other factor that was the villain. If so, what could it be? How about a genetic trait that causes both a desire to smoke and a predisposition to lung cancer? Those with this genetic factor would likely take up smoking and they also had a good chance of developing lung cancer. However, the cause of the cancer would be the genetic factor, not smoking. This possibility was put forward by a renowned scientist in the U.K., Ronald Fisher. Knighted for his research work and considered the “father” of modern day statistics, Fisher was held in high esteem all over the world. How in the world could Fisher’s concern be tested – any ideas?

The genetic issue was laid to rest when the results from studies of identical twins became available. The twins had the same genetic make-up, but for each twin pair, one was a smoker and the other wasn’t. The twin study found that the twin who smoked was much more likely to develop lung cancer than the twin who didn’t. However, there still remains the smallest, slightest, infinitesimal possibility that the cigarette is not a cause of lung cancer even today. Re-examine the concern expressed by Fisher and replace the genetic connection with some event in a person’s life that triggers a physiological or anatomical change (e.g. exposure to a chemical byproduct or an environmental pollutant). The change has two effects. One it makes smoking a desirable habit and two it causes lung cancer. Yes, this idea is far-fetched, highly unlikely and it’s even fair to say I’m grasping at straws. However, even if the likelihood of such an event is one chance in a million or a billion or a trillion, it is not absolutely impossible either. But it is fair to say that smoking as a cause of lung cancer has been proven beyond a reasonable doubt with emphasis on the qualifier, “beyond a reasonable doubt”.