Signal Detection

WHO defines a signal as, “Reported information on a possible causal relationship between an adverse event and a drug, the relationship being unknown or incompletely documented previously. Usually more than a single report is required to detect a potential signal, depending upon the seriousness of the event and the quality of the information”

CIOMS VI defines it as, “A report or reports of an event with an unknown causal relationship to treatment, that is recognised as worthy of further exploration and continued surveillance”

Sources of signals: Signals can be generated through various sources. Given below is a list of possible sources of signals

• Clinical Studies- Any clinical study with the product, whether company sponsored or otherwise, both pre and post marketing, is a rich pool of information for astute observers. Discussion with investigators often leads to identification of suspect situations which need to be explored further
• Post marketing information from prescribers, consumers, other regulatory bodies, ECs, IRBs
• Single cases, or case series in aggregate review, PSURs
• Medical Literature, internet, newspapers other media
• WHO database or other regulatory databases

Confirming the signal: Having gathered data from various sources it is almost impossible to manually screen all the data. Complex statistical modeling, apart from routine statistical methods are required to confirm that the signal exists. e.g. Latest techniques like Empirical Bayesian Neural network, Proportional Reporting Ratio(PRR) and MGPS (Multi-Item Gamma Poison Shrinker ), using exclusive software, have been developed

This is called data mining where spontaneous reports are systematically screened for interesting associations. Another method is disproportionality analysis again towards the same goal of detecting “higher than expected” drug-event frequencies without having actual exposure data

Signal evaluation: Various associations and possible signals are prioritised based on frequency, seriousness, impact on or risk to patient. In today’s litiginous society, companies also have to guard their reputation and protect against liabilities. Having prioritised the signals they need to be further evaluated to ascertain their cetainity, frequency, seriousness

Further evaluation could include

• Sub group analysis of existing data
• Advanced data-mining
• Pharmacoepidemiologic studies to corroborate findings
• The signal could be evaluated as part of a new safety study
• Flag the adverse event and monitor it in all ongoing studies
• Design an exclusive preclinical study in an animal model to study the signal
• Use of latest pharmacogenetic techniques including biomarker research to obtain quicker and specific answers

Possible outcomes after signal evaluation : Depending upon the strength of the signal, possible outcomes could range from no action at all to withdrawal of the drug from the market, with many intermediate actions in between. These are generally decided after a discussion with the regulatory authorities. Since every authority refers the matter to its own set of experts and also due to conditions typical to that population, the action may not be the same in all countries. The classic example is that of the antiamoebic Iodochlorohydroxyquin which is banned in major countries but is allowed to be marketed in India. Summary of possible actions is,

• No action if signal is of no consequence.
• If there is a level of uncertainity, there could be increased monitoring for that adverse event
• Change product information
  • Addition of new event
  • Modification of current wording
  • Addition of a frequency descriptor
• Restriction of use
• Withdrawal from the market or aborting development plans if not yet marketed
• Passing on Information of the change in prescribing information to all stakeholders like ECs, IRBs, doctors, regulatory authorities, licencee partners, consumers

References

• http://ec.europa.eu/health/human-use/pharmacovigilance/index_en.htm
• Reporting ADRs, CIOMS publication 1999
• Post-approval safety data management: definitions and standards for expedited reporting .E2d . ICH publication 2003
• Karen Hedenmalm. Hazards of Drug Therapy. Upsala Dissertations 2005;downloaded 2012jan
• Council for International Organizations of Medical Sciences,(CIOMS VI, 2005)

__________________________________________________________________

Comments

Detailed information to begin with. Need more articles mainly casestudies on some scenarios that challenge and improve signal detection system

Dr Bhavana Bhagat

___________________________________________________________________