She holds an M.S. in Statistics from Rutgers University and a B.A. from Cornell University. She is a member of the American Statistical Association and the New York Area SAS Users Group. Her contributed article and contact information follow.

———————————————————–

For many in businesses that rely on clinical research, the ‘sizzle’ of a headline or claim seems to be enough. But imagine the following scenario:

A prospectus sits on your desk. There is a sentence “The formulation shows a p-value of .023, significance at 95%, at a dose of 10 mg. Sounds good…its statistically significant…move on. But then you get a call six months later. A regulatory reviewer questions the randomization technique. But you thought everything was okay at 95% confidence. They did the statistics.

Statistics are the language for reporting results of efficacy and safety trials. We report significance, confidence and p-values, the probability of an event under certain conditions. These statements make their way into press releases and business documents.

In truth, statisticians play a role before the study is started. They advise on the study design, data collection and implementation. While the study is in progress, they bridge communication between the computer professionals such as programmers and data managers and the medical professionals such as clinical investigators and research associates.

In the jargon of the industry, the research on human subjects is referred to as Phase 1 through Phase 4 clinical while other activities such as drug discovery, manufacturing and toxicology get lumped into “non-clinical”. Clinical statisticians do work in a structure of tables, listings and graphs that adheres to the requirements of the FDA and various other national and international organizations. The roles vary for non-clinical statisticians. There is more interest in applications of statistics in early stage research and manufacturing processes.

One can find various definitions of the word ‘statistics,’ including the vernacular usage as any display of numbers. The American Statistical Association in its career section website (www.amstat.org/careers) answers the question “What is statistics?” as follows–Statistics is the scientific application of mathematical principles to the collection, analysis, and presentation of numerical data. Statisticians talk in the language of science-observation, estimation, hypotheses, generalization and modeling. Unlike other approaches to data analysis that aim only for a mathematical optimization, statistical analysis concerns itself with the scientific context of the data collection such as the measurement ranges and precision of lab equipment. It is not black-box computation.

In the early 20th century a mathematical framework for statistics emerged built on the backbone of probability theory, an older branch of mathematics. The expansion of this mathematical framework continues in universities around the world. While clinical research sometimes uses new techniques, it relies on basic ideas like distributions of averages and randomization. Randomization is a plan for treatment assignment that is independent of human choice. We know the probability that a subject is placed in a treatment group. This assignment lets the researcher use probability-based statistics to assess the effect of a treatment and whether it is just a result of ‘chance.’

As a professional statistician, most people find what I do mysterious. They remember a dreaded required course. If older they think of grayish books with grids and curves. If younger they think of the horrible software they were forced to use. Fortunately these classes, the books and the software are now a lot more applicable, even fun.

But you don’t have to take a new-fangled level 200 class to bring make clinical statistics more relevant. A good start is thinking about the numbers that express the result. This will make your questions more effective. A desired clinical change can be expressed as a change in a decimal number, a count, or a percent. A ‘20% drop’ can explain going from ‘200 to 160′ or ‘15 to 12′. This is where the ‘scientific context’ becomes important. The expression of a clinical result is the end-product of the clinical practices, the regulatory guidelines, previous company studies, and the judgment of the current statistician, and perhaps a previous statistician.

My ending comment is stay aware. A short phrase in a business or legal document is a small blip in the thousands of pages in a clinical trial submission. The minds of regulatory reviewers are at once specialized and focused but tuned into current trends. Clinical statistics can become pivotal in the approval process.

Georgette Asherman
Applied Statistician
Direct Effects, LLC.
www.directeffects.net
201 673-4301

Installing More Flashing Red Lights

In “FDA Turns Attention to Detection,” Jill Wechsler concentrates on upcoming changes to the FDA’s Adverse Event Reporting System (AERS). AERS is the FDA’s current surveillance method for drugs currently available to the public, tracking some 4 million case reports and growing by about 300,000 submissions per year. While the information is extensive, its quality varies and has been identified as a major shortcoming in the current iteration of the system. (1)

Regardless of the quality of data, however, the FDA defends AERS’ use because it covers all FDA-regulated products and subsequently reaches a large portion of patients, including any that might be using a drug for an off-label purpose. It is considered especially effective at highlighting rare, unexpected drug safety problems, a sentiment echoed by a majority of healthcare professionals and cited as a reason that the system should be strengthened and augmented by other safety systems, not replaced.

To Imply or Not To Imply

While the FDA works on safety, Congress is staging an investigation into the celebrity endorsements that are accompanying more and more direct-to-consumer (DTC) ads on TV. In “Attack Mounts on DTC Ads,” Jill Wechsler outlines a new effort by Michigan Democrats John Dingell and Bart Stupack to challenge Pfizer’s use of celebrity spokesman Dr. Robert Jarvik, a prominent contributor to the invention of the artificial heart, in commercials for Lipitor (atorvastatin), the blockbuster anti-cholesterol drug. I recall the commercial vividly, its impact perhaps resting in the subconscious part of my brain, and I remember assuming that Jarvik was giving his testimony based on the background of a heart specialist.(2)

Contrary to my belief, Jarvik is not a heart specialist and it’s probably inappropriate to imply as such; it is because of this misleading advertising to consumers that Dingell and Stupack have chosen to investigate. Also on their list for investigation is Vytorin commercials, a joint venture by Merck and Schering-Plough. The Congressmen have been provided documentation by the FDA, and the investigation is currently ongoing.

Rx Email: Redux

In Europe, the United Kingdom is attempting to halt false advertising from another source: illegitimate online pharmacies. According to “UK Tackles Faux Pharmacies Online” by Sarah Houlton, the Royal Pharmaceutical Society of Great Britain (RPSGB) has launched a new logo that will help certify an online pharmacy is a legitimate source of medication for consumers. You may recall I examined the ever-expanding problem of online pharmacies in “Rx Email,” and indicated that the increasing prevalence of this problem necessitates action by agencies around the world; with an estimated 2 million people now use the internet to buy their medications in the UK, it is becoming obvious that the issue likely needs to be dealt with now.

Registered participants in this program will receive a new logo and registration number to be used on their home page. The logo will be a link leading to a list of legitimate sources for online drugs as maintained by the RPSGB. Almost 50 online pharmacies now participate, and the hope is that the system will discourage the dishonest traders. (3)

The magazine, as usual, gave me plenty of material to think about. Does the FDA AERS need to be augmented? Or should the original system be re-examined and re-built to encourage a more efficient and effective process? Should drug companies be allowed to use celebrity endorsements and imply claims that they cannot make? Or is this type of marketing no different than the marketing of a car (buy our car and you will be cool) or some other product? Where do the differences lie? Could the method being developed by the RPSGB really discourage the sales of illegitimate drugs online, or is it simply another roadblock for these bogus sources to overcome?

Regardless of the answers to these and other questions that may arise in my mind as I continue to read, it appears there is increasing effort by governmental organizations to keep their consumers safe from spurious, possibly false and sometimes illegal dissemination of information. No agency is perfect, but it is encouraging to know there are people within them who are doing their best to make me a little safer as a consumer.

What do you think?

(1) See “FDA Turns Attention to Detection,” Jill Wechsler, Pharmaceutical Executive Magazine, March 2008, p. 14.

(2) See “Attack Mounts on DTC Ads,” Jill Wechsler, Pharmaceutical Executive Magazine, March 2008, p. 18.

(3) See “UK Tackles Faux Pharmacies Online,” Sarah Houlton, Pharmaceutical Executive Magazine, March 2008, p. 18.

Well, smokers at risk for Parkinson’s disease got an interesting reason not to kick their habit this week according to a report released by the University of California Los Angeles School of Public Health. According to the study, there is a protective effect generated by tobacco use against the degenerative nerve disease, which has been borne out by several studies on the topic over the past 47 years. Further, the study indicates a persistent protective effect, with smokers who had quit up to 25 years earlier also demonstrating better protection against the disease.

This kind of effect seems to be specific to the action of smoking tobacco. It is not believed that a nicotine patch or any other anti-smoking remedy could be linked to the benefit. This is mainly because in studies done with animals, the carbon monoxide or other agents in tobacco smoke seem to promote the survival of brain neurons that produce dopamine, a chemical lacking in Parkinson’s patients responsible for muscle movement. The other possibility is that the smoke ironically prevents the creation of toxic substances that normally would interfere with proper neurological functioning.

You can see more information about the report, including study statistics, here on the Reuter’s site.

Nick Naylor would jump at the opportunity to grasp this study and use it to his advantage. But upon reading about the results of this report (and really, it was only the summary I got from Reuters), I wondered whether this study was going to empower people who are truly addicted to have another reason not to quit.

Statistics play a critical role in the decisions we make in our day-to-day lives. An opinion often seems less credible without some statistic or personal experience behind it. But some statistics, such as these that were gathered by this report, seem almost counterintuitive to me. I mean, will doctors potentially recommend smoking as a treatment for Parkinson’s in the future? Dr. Gregory House, of the television show “House MD”, once prescribed “smoking” as a treatment for inflammatory bowel disease, a scene I found amusing in the context of the TV program, but I would be horrified to see in an actual clinic. Yet, there are doctors that would prescribe drugs that could have a far more dangerous (and immediate) effect.

I am always weary of statistics presented to me by doctors, sales people, and anyone else who could have an agenda. There seems to be a statistic for everything now, and the onus is on the consumer to actually examine the facts and understand the nuances of a decision before following a certain path. In that way, I agree with the libertarian view as mentioned on the Wikipedia Web Site of “Thank You for Not Smoking”, and it is becoming increasingly important for people to make decisions based on all the information that’s available.

How do you feel about smoking as a treatment rather than as a social habit and the statistics that have been presented for its use in treatment of Parkinson’s patients? Would you suggest smoking to someone who might be suffering from the early stages of Parkinson’s? What are the moral implications of such a suggestion?

———————————————————–

A major activity for statisticians in pharmaceuticals is making recommendations about sample size. Why does one control/active study have 28 subjects in each treatment arm while another has 58? This is usually attributed to a method called ‘power analysis.’ The statistician will say something like ’58 subjects are needed to get a significant effect at critical value of .05 at 90% power.’ This sounds authoritative, but what does it mean in terms of clinical activity? The statistician will say that there are enough subjects to find the effect 90% of the time-or in other words-it will only be overlooked 10% of the time. The clinicians and scientists accept this, but most likely are still perplexed.

In these discussions, the main idea of performing the study, finding a desired effect, can be overlooked. In the straightforward example of a control/active study, the effect size, also called the delta, is the difference between the means of the endpoint of interest. This delta is an assumption about an unknown, not an observed value. In designing a study, the investigator has to think what could be a desired impact to make this new active product worthwhile. Usually a big impact would be desirable and easy to see. But a well-designed study should find a small delta if it is of clinical interest, such as drop of a few points in average cholesterol or blood pressure.

The starting assumption (or null hypothesis) is a difference of 0 between the two means of the groups. (There can be other null hypotheses, but this is the most common and straightforward.) With real people and real lab results, we shouldn’t see the same average, for two groups, even when the impact of both treatments is the same. The difference of the active mean and the control mean will be a continuous real number such as 6.91 or 1.24 or -3.21. Since there are different types of averages, the word ‘mean’ is used for the classic technique of the sum divided by the number of observations. The statistical test will show if this observed difference implies a real difference that isn’t 0.

We can do a study with a number of subjects that is convenient and affordable like 10 or 20. This happens in academic research sometimes. We will observe differences of means like 6.91 or 1.24 or -3.21. This data can go into a statistical package and show a significant effect. This ‘significance’ merely means that we can comfortably reject that the true difference is 0. But if there is no significant effect, how do we know that we aren’t missing out on finding the true effect? Should we recruit more people and redo the study? Since there was no power calculation it is hard to judge.

The ‘power’ of power analysis is the framework it provides. Since we can’t test the whole population, we will never true the effect size. We observe a non-zero difference which can be larger or smaller than the true effect size. We can only say that is not zero, not that this difference equals the true effect size from the active product. But suggesting a possible effect size makes the results grounded in the clinical context. The power number, either pre-set or calculated, shows the strength of the design. The sample size is obtained from these inputs.

These calculations are now done with several popular software packages. The desired effect size and power will derive the recommended sample size. As the power goes up, the required number of subjects increases. As the effect size goes up, the required number of subjects decreases. This sounds odd to some people. But if the true effect is bigger, fewer subjects can show an average distance just as far away from 0. Detecting smaller effects, closer to 0, requires more subjects.

Besides effect size and power, the critical value and a suggested standard deviation, a calculation of the spread of the data, is required for a sample size calculation. The critical value is typically .05 or .01, either one-sided or two-sided. The power analysis can be done for two-sided or one-sided tests. Two-sided tests identify a non-zero difference in either direction while one-sided tests look for either a positive or negative change. How do we know a standard deviation if we haven’t collected any data? Statisticians rely on previous studies or published results.

This is a very brief summary of a very large topic. While I compare two means as an example, the idea is applicable to treatment means within a subject, more than two treatment groups, and non-clinical studies in-vivo or in-vitro. Besides testing means, power analysis is used for counts such as safety results. Most of all don’t be afraid to ask questions.

Georgette Asherman
Applied Statistician
Direct Effects, LLC.
www.directeffects.net
201 673-4301