Indian Society for Probability and Statistics
Indian Society for Probability and Statistics
Letter from the President to members and friends of ISPS
Clinical Trials
(April 2007)
To: Members and friends of Indian Society for Probability and Statistics
From: Anil Gore, President, ISPS.
Dear Colleagues,
As you know, I keep writing mails to you about topics of common interest with the hope that it will initiate new thinking and action. Feedback is rare. But I persist. Today I wish to discuss a topic that can have a significant impact on all aspects of statistics - education, research, professional opportunities etc. in India.
Of late India seems to have broken out of the ‘Hindu rate of growth’, which hovered around 4% for decades, and now there is talk of even double digit growth rates. This is mainly due to performance in the manufacturing and service sectors of the economy. Our blue-eyed boys are the software people. But a quiet revolution is taking place in the health sector also that may have largely gone unnoticed by many including us statisticians.
One feature of the new paradigm is medical tourism. We have some fine hospitals and expert Doctors and Surgeons. Treatment is first class. And all this comes at a cost that is remarkably low. So, getting treated in India makes a lot of economic sense to people in many countries. Health insurance companies in the west are sure to take advantage of this high-quality-low-cost service. It will enhance their profits and competitiveness. Of course the service provider has to have excellent reputation. Apollo Hospitals is one company often mentioned in this context. In my own city of Pune, Jehangir Nursing Home (an affiliate of Apollo Hospitals) and Ruby Hall are two hospitals already in the game. In addition to treatment, there will be a plethora of remote services like diagnostics and expert advice.
There are two other major areas in the field of health that may see great expansion. They are drug manufacturing and drug discovery/testing. Our pharmaceutical companies such as Ranbaxi, Dr. Reddy’s Lab., CIPLA and others have already become global players. Earlier their important business was the so-called generics. These are medicines that are out of the patent restrictions. In other words, the medicines have been around long enough that proprietary rights of discoverers are over and the items can be manufactured by anyone. Cost of production of our companies is low and hence their competitiveness is high. In exceptional circumstances, patent rights may be relaxed for humanitarian considerations. You may have heard that some drugs for HIV-AIDS are in this category. Indian companies may manufacture these medicines to make them available to victims in Africa, Asia etc at an affordable price.
For all these years, activity in the area of drug testing was minimal in India. One possible reason was our patent regime. It was different from that in the west. We allowed production of a formulation by people other than the innovator as long as the process of production was different. Hence huge expense for trials did not seem worthwhile. Another effect of the regime was that the MNCs were very cautious in introducing a new drug in India for fear that it may be copied and hijacked. Now, the law has changed and is more in line with the west. So, there is much greater interest in shifting the work of developing and testing new drugs to India. Under the new law, MNCs seem to feel safe enough to outsource drug-testing work to India. This has opened up new opportunities for statistics.
If there exists a group of molecules/formulations potentially useful in a certain disease, it is necessary to screen it and test a potential drug before it is manufactured and sold to people. Over the last several decades, the process of testing has been refined greatly in the international arena. In most countries, a government department with a name such as ‘Drug Regulatory Authority’ lays down the law of the land as to how testing of a new drug should be done. Three markets for medicines, in the world, namely North America, Western Europe and Japan account for major part of the sales of modern medicines. Hence drug-testing rules used there are crucial for any commercial manufacturer of drugs. What matters most to us is that these rules provide a critical role for statistics. In brief, the tests have to be planned following statistical principles of experimental design and analyzed using good statistical methods and good software packages. You can check all this on the Internet. As an example you may want to visit the website www.fda.gov of the U.S. Food and Drug Administration.
How big is this activity? A few numbers may give some idea of the enormity of work going on in this field of clinical trials. All trials have to be registered before starting (so that we do not have the problem of failed trials being suppressed). Time gap between starting a trial and its successful completion and beginning of sale of the drug may be of the order of a decade and the cost in US may be several hundred million dollars. According to one estimate, total annual outlay on clinical trials in the world may be about 40 billion dollars (about 2 lakh crore rupees, or market capitalization of Infosys and TCS together). As you can imagine, this involves many hospitals, doctors and countless patients. Data collected are huge and the work of analysis is also very considerable.
About a decade ago, Pfizer, a giant MNC in pharmaceuticals, opened a center in Mumbai to handle part of the work of data management and statistical analysis of clinical trials done by (or commissioned by) them. The center has flourished. Novartis, another MNC in the same field has also followed suit. Others may come too. It must have been relatively easy to outsource statistical analysis. The big step is conducting the trial itself in India. This is also happening. Companies farm out part of the work to other companies called CRO (contract research organizations). Many such CROs (newly started as well as old and established ones like ‘Quintiles’ www.quintiles.com ) are now operating in India. According to one estimate total turnover of CROs in India will rise by two orders of magnitude in five years to reach 10,000 crore rupees per year. (Also visit www.informa-ls.com/ct-ceeasia for details of a conference on issues concerning clinical trials to be conducted in India and other countries)
What are the implications for the field of statistics education and research in India?
1. Availability of jobs for holders of M.Sc.(Statistics) degrees. Many pharmaceutical companies and CROs recruit such students. Campus interviews have become a regular feature in Pune and some other universities. Neither teachers nor students have experience in this respect. Teachers should get to know how students of MBA prepare for interviews. Aspects such as group discussion, essay writing, case studies are quite neglected in traditional teaching of MSc statistics. Some correction is overdue.
2. Holders of M.Phil. and Ph.D. degrees in statistics are also needed. But supply is short. These individuals will become supervisors and group leaders and hence will get better emoluments. We should make our students aware of this. Any perception that time spent in research is ‘wasted’ has to be corrected.
3. Training in commercial software packages for statistical calculations needs to be emphasized. In particular SAS programming is a widely used skill in analysis of clinical trials. Fresh recruits can be trained but those who already know it will have a competitive edge. Hence launching of short programs in statistical computing (SAS and SPSS along with some other packages) would be worthwhile. (Do think about organizing an evening program of two hours per day for four to six weeks. You may find takers in many disciplines including economics and management. The program can easily become self-supporting.)
4. It would be useful to ensure that topics in statistics that are directly relevant to clinical trials are duly emphasized in our teaching; for example crossover designs, (group) sequential tests, bio-equivalence testing, multiple testing and issues arising from there. Further, specific attention should be paid to increase skills in report writing. Our students are often weak in written and oral communication, which becomes a handicap.
5. An elective course on statistical methods in clinical trials may be launched. This will require extra effort by faculty. Teacher training programs can be thought of as aid in initial stages. Experienced academicians and scientists from companies can be invited to speak. If you need help, let us discuss the matter.
6. Small research projects of relevance to clinical trials can be taken up. I will give below an example of the kind of research that may be directly relevant and can be taken up by any statistics group in a college or university.
7. Students of M.Phil. can be asked to review selected recent material on clinical trials in journals like Biometrics.
8. Statisticians employed in such work can be encouraged to work towards a research degree such as M.Phil. and Ph.D. They will bring to the table domain knowledge and clearer view of problems of interest to the industry. Faculty can complement them with expertise in theory.
9. Companies are keenly aware of the important role that universities will play in development of this sector. After all we are talking about knowledge-based activities. Companies are therefore favorably disposed towards any ideas of collaboration. At Pune we have a relationship with Pfizer group in Mumbai. Possibilities of collaboration with Glaxo are being discussed. The message is that teachers of statistics should be bold in approaching pharmaceutical companies in the vicinity. Any interaction is bound to be of mutual interest.
10. In allopathic field, while MNCs are clear about need for statistics, my experience is that many government laboratories lack awareness and skill in statistics. Several laboratories of the CSIR are seriously involved in drug related research. But awareness of statistics is inadequate (to say the least). Expertise in statistics is almost non-existent. Hence extensive statistics training for chemists, biologists and other scientists in these drug research laboratories is essential. We at Pune had prepared an ambitious proposal to carry out such a training program with provision for it to be turned into a collaborative activity. This was done with encouragement from Dr. Kumar from IISc. But it all came to nothing due to a sudden and inexplicable loss of interest from the CSIR. Perhaps things may click with you. Hence I strongly recommend that everyone should think about exploring the possibility of collaboration with institutions like the Central Drug Research Institute, Lucknow.
[Project: As stated earlier, clinical trials are extremely expensive. Hence even a small reduction in cost is of interest to industry. There are three prongs of a clinical trial: time needed, cost involved and power to detect efficacy of a drug. So, any research that can reduce time and/or cost or increase power is of interest. {Time is very precious in clinical trials because duration of the trial directly cuts into time available to market the drug under patent protection.} One reason for delay in a trial is non-availability of enough patients. Planners have some idea of how many patients they need and in what period. Very often, hospitals have no idea about the number of patients that can be recruited and they give crude (and generally inflated) estimates. If patient recruitment falls short, trial just has to wait. So, we take up a hospital, select a disease (asthma, heart attack, hypertension, cancer of different types, diabetes etc.), study records of patient arrivals and develop a suitable model for patient arrival rates (stochastic process, queue, time series etc). In addition to giving a better estimate of patient arrival, such a model may also throw new light on the epidemiology of the disease. ]
My last point is the breadth of the field. If drug testing in allopathic research needs statistics, the need is equal or more in naturopathy, ayurved, homeopathy and other alternative medicine systems. Awareness of the need for scientific trials is increasing among followers of such alternative systems. [Pune is home of the ‘National Institute of Naturopathy’ and practitioners there appear to be looking for ways to assess efficacies of their methods. Some of us at Pune are currently participating in many trials of this type. Some examples are: a trial of karela (bitter gourd) on diabetes, herbal tea on cholesterol and ‘abhyangam’ on stress] But unfortunately, awareness of statistics may be much less. I would really urge my colleagues to adopt a missionary spirit and to proactively connect with such people. This is a big opportunity if we decide to go out and grab it. It is an ideal situation because we can expect material gains while doing something for social good. I would like to draw your attention to a very relevant guest editorial in Current Science (January 10, 2006) by Prof. M S Valiathan.
I invite members to write to me their personal experiences and opinions regarding this whole field of clinical trials.