New techniques for drug safety development: predicting adverse effects and diagnostic testing

The technique seems simple enough. A group of researchers used a mathematical model to create a network with links between known adverse drug effects (ADEs) and a variety of drug safety parameters and pharmacological information. The data included over 800 different drugs and adverse effects considered to be drug-related from a 2005 survey. They compared their results with adverse effects reported in 2010 from the same drug safety database, and found their predictions to be right on target.

The computational analysis involved a linear regression model to determine if the computer could actually predict adverse side effects purely based on the network content. The relationship between their predictions in 2005 and the actual effects that were presented in 2010 was strong enough to consider this to be a viable method for predicting unknown side effects of new drugs.

Another technique sure to grace the news pages of 2012 is companion diagnostic testing. The push for this is coming from the investors of drug development companies. Basically, the investors want objective proof that the drug is effective as well as safe.

Regulatory agencies such as the Food and Drug Administration (FDA) are responsible for ensuring the safety of new drugs before they are approved for human use, but efficacy testing may be conducted separately and include bias or unverifiable results. The changes investors are pushing for aim to prove or disprove claims made by drug developers. It’s also a way of gauging the demand for a drug, and may ultimately limit the target market for drug manufacturers. This could hurt the income stream for drug manufacturers, big pharma companies, and sales reps.

How can new techniques save money and boost profits? If the technique increases costs, that increase will ultimately be passed along to end users. If the technique can save money, reduce waste, or add value in some other way, there is going to be a cost for that value somewhere along the line.

With the predictive network method used for adverse drug effects, the innovation is clearly intended to increase safety measures which will resonate loudly with regulators, investors, and manufacturers alike. On the other hand, companion diagnostic testing is a great way to develop the market for a drug, but there are extra costs and time involved which ultimately add to the cost of drug development.

The regulatory process for new drugs is long and costly. If evidence suggests additional testing for safety pharmacology may be valuable, then the agency will withhold approval until the research is complete. Additional testing can slow the approval process and ultimately cut into profits for investors and drug manufacturers. The longer it takes for a drug to hit the market, the more expensive it will be for patients who need the meds. If the delay is for safety reasons, then by all means, the testing must continue. But if the delay is for marketing or profitability reasons, then it is important for regulators and investors to balance the needs of the industry with the patients’ need for treatment.

Citation: A. Cami, A. Arnold, S. Manzi, B. Reis, Predicting Adverse Drug Events Using Pharmacological Network Models. Sci. Transl. Med. 3, 114ra127 (2011).