What does this career entail?
Statisticians (and statistical programmers) in the pharmaceutical industry are certainly not just "number crunchers"; instead, they are key players in all areas of drug development, from initial research and development right through to manufacturing of pharmaceutical products. They spend a lot of time working with people from different disciplines, including doctors, scientists, production managers and marketing teams. They also have to work with government agencies. They work in many different geographical locations, all round the world.
Pharmaceutical statisticians carry out a wide range of activities. These often begin with the design of scientifically sound experiments. There will then be the analysis of the collected data; the statisticians have primary responsibility for this throughout all stages of a drug's development. And at the end come the vital tasks of arriving at correct interpretations of the data analyses, writing summaries for formal documentation, and presenting the results to senior managers and regulatory authorities as necessary.
Development of a new or improved drug typically passes through many stages, often called clinical trials. Pharmaceutical statisticians are closely involved with them all. Here is an outline of the various stages, showing how the statisticians are an integral part of the processes
Pre-clinical. Potential drugs are often adapted from naturally occurring proteins or molecules, or they might be found using computer models. By whatever route they are discovered, there are many different areas of research that must be gone through before they can be given to humans. The main objective of this pre-clinical stage is to make sure that a drug is as safe as can possibly be known, thus allowing it to go forward to the clinical stage and be given to humans. Many of the necessary experiments are governed by regulatory requirements. While this implies that experimental designs and data analyses are sometimes fairly routine, it also means that the statisticians have more time to concentrate on the many non-routine tasks that always arise. These might for example include analysing data from a non-standard molecular toxicology experiment, or training research scientists to use statistical computing programs to analyse data from routine experiments, or developing new systems for dealing with the outcomes of the experiments.
Clinical. This stage consists of investigating a drug in humans. The experiments are generally categorised into four sequential phases:
Phase I experiments are undertaken on healthy volunteers typically 8 to 24 of them. The main objective here is to assess the safety of the drug in humans and understand what the drug does to the body as well as what the body does to the drug. Statisticians are involved in designing the experiments, identifying what data should be collected and how they should be analysed, and writing formal reports of the experiments with appropriate statistical and scientific interpretations.
Phase II experiments are performed if the drug is considered to be safe after analysis of the phase I stage. These experiments involve people who have the disease that it is thought the drug alleviates. The aim here is to find the most effective and safe dose. Usually around 100 to 200 people are involved. Statisticians need to spend a lot of time designing these experiments, which will typically involve several different doses of the drug as well as any alternative drugs that are already available. Furthermore, the general efficacy (how well the drug works) and overall safety of the drug are still being worked out; this means that a great deal of information is required from the experiments, and this brings up many important issues that have to be addressed in the experimental design as well as in the data analysis. As with phase I experiments, the statisticians are also involved in writing the formal reports of the experiments.
Phase III experiments investigate the drug in large numbers of humans who have the disease, comprising anything from hundreds to thousands of people. Statisticians use the data collected in phase II to estimate how many people should be included. They also search the published literature for any relevant information about the disease. This helps to gain an understanding of how variable the results of the investigations are likely to be, and this gives further information on how many people should be included so as to ensure that estimates of responses to the drug are as accurate as possible.
The statisticians then prepare formal documents to guide the experiments. These will for example include guidance on aspects such as stopping the experiment early. This is very important as the complete experiment might take say 2 or 3 years, but if it appears at an early stage that the drug is not performing well or if it is performing so well that an early release to the market might be justified then an early termination could be appropriate. The statisticians also prepare documentation about the data analyses that will be undertaken, perhaps indicating alternative methods of analysis.
Once the data have been collected, statisticians perform the analyses and write formal reports.
In addition to these responsibilities, the statisticians summarise all the data collected throughout the whole of the drug's development, so that the information can be shared with the regulatory authorities. This enables the statisticians to work with regulatory assessors some of whom will themselves, of course, be statisticians to give answers to many important questions about the safety and efficacy of the drug, before it can be deemed to be safe and effective and allowed to be sold in the market.
Phase IV experiments are primarily commercially oriented. Inferences now need to be made about use of the drug in a wide population. The data collected may be for different formulations of the drug or different devices for taking the drug. The aim of these experiments may be to show not only that the drug is safe but also that it has superior effectiveness or could be part of a cheaper healthcare package for a particular type of person. Again the statisticians are centrally involved in designing the experiments and analysing and interpreting the data. Long-term data are also collected, so as to see whether there are any possible safety issues with taking the drug over several years or whether there might be any very rare after-effects of concern which might not have been picked up in the earlier studies even with many hundreds or thousands of people.
Manufacturing. First the drug needs to be made on a fairly small scale for use in the clinical trials. If it passes the trials successfully, it will need to be produced on a larger scale perhaps a very much larger scale for sale in the market. Statisticians help improve manufacturing efficiencies during the development stages, making sure that the raw materials are being combined within acceptable quality control limits and identifying how the process can be scaled up from a laboratory bench level to large-quantity manufacturing. When the drug is being sold in the market, statisticians will help make sure that the production quality of the drug is maintained within acceptable levels. This involves the design of experiments for investigating and improving industrial processes; statistical process control and general statistical modelling arise naturally from this work.
The pharmaceutical industry has come to realise how important statisticians are. As a result, other opportunities are arising all the time. For example, statisticians are supporting areas such as pharmacology and, more recently, they have been instrumental in cost-effectiveness modelling.
Statisticians are also involved in the development and improvement of devices and diagnostics. An example is involvement in assessing different devices for patients to take their own ECG readings and then transmit them to a specialist centre. Another example is assessment of a device for measuring tremor in arms and legs so that this can be done when the patient is asleep; these measurements also are then transmitted to a specialist centre for analysis. Statisticians look at factors such as ease of use of such devices, and hence the likelihood of error (perhaps by patients somehow inadvertently entering incorrect results, or even erasing the results by mistake). The general reliability of the devices is obviously also important, as well as an understanding of the natural variability to be expected in such readings.
All in all, there is plenty of scope within the pharmaceutical industry for statisticians to make a real difference and expand their roles!