AstraZeneca is a British pharmaceutical firm, operating in more than 100 countries, with primary foci on cardiovascular and metabolic diseases, oncology (cancer), and respiratory, inflammation and autoimmunity issues.

AstraZeneca employ around 51,500 people worldwide, make drugs in 16 countries, and had 2012 sales over $25.7 billion (Astrazeneca.com). In the past few years, the firm has made repeated news headlines as a possibly desirable target for takeovers, something that by the end of 2014 its shareholders had resisted.

The AstraZeneca stable of products includes several blockbuster drugs. Included in these is Crestor, which as of the middle of 2014 was the bestselling cholesterol drug worldwide and the 9th highest selling drug overall of 2013. Crestor alone enjoyed 2013 sales of some $6 billion for AstraZeneca.

Statistical testing in clinical trials lies at the heart of the development process of pharmaceutical drugs such as Crestor. The basis for a drug’s fundamental usefulness, marketability, and approval by oversight bodies such as the U.S. Food and Drug Administration (FDA) is sufficient evidence that a drug has a significantly positive effect in whatever health issue it targets, and evidence that it does not cause unduly deleterious side effects. In balance with the huge costs of drug development (see the introductory case of Chapter 18 for more on this), the statistical tests that help discover drugs and get them to market are at the center of the industry.

Typically, many rounds of clinical trials will seek to demonstrate that a drug has substantial benefits to patients in contrast to not taking the drug, taking other drugs, taking less of the drug, and the like. These tests often involve measuring patient symptoms or the like over time, and comparing changes in groups that take the drug with groups that do not get the drug but instead get a placebo. In other tests, instead of looking at symptoms, researchers will measure the incidence of undesirable side-effects from the drug.

Generally, the success of such tests is a key gateway for whether a drug gets accepted for marketing by regulators or not. The tests also indicate the extent to which the drug (seemingly) works well with relatively few side-effects, and therefore how popular it will be with consumers and doctors, and therefore whether it will sell well.

When implementing these clinical trials, one fundamental finding in a given study might be something like “CRESTOR given as a single daily dose (5 to 40 mg) over 6 weeks significantly reduced serum TG levels” (rxlist.com). I have no idea what a serum TG level is, but the key thing here is that is was ”significantly” reduced. The statistical question will revolve around the size of the reduction and whether this effect of the drug is big enough to be both substantial and financially feasible.