Signal Detection Principles

"Volume 9A, Part I, Section 8 “Overall Pharmacovigilance Evaluation and Safety-Related Regulatory Action” describes the principles of signal detection and highlights the sources of information where a signal may be detected.

Signal detection procedures can range from individual case review, trend analysis of case reports to the use of complex statistical methodologies (e.g. Bayesian methods).

The methods used (including periodicity of review) should be determined, for example, by the product portfolio and the number of reports of suspected adverse reactions received. 
The MAH should, based on risk assessment, decide what signal detection methods are to be used for which products or types of products, and should document this risk assessment. 
All MAHs are expected to have in place systems and procedures for systematic signal detection that are adequately documented in formalised procedures (e.g. SOPs). 

Within this documentation it is useful to provide a definition of what constitutes a potential signal, in order to educate all personnel involved in the process and to identify what requires further evaluation. 
These documents should also provide detail about the roles and responsibilities of all personnel involved, the sources of information included in the analysis and the methods used for signal detection. 

In addition, formal written procedural documents should describe what actions MAHs intend to take based on the outcomes generated from signal detection activities. 
This could include escalation procedures following identification of a potential signal, such as further analysis that will be undertaken or referral to a company committee. 

Whilst there are no defined minimum requirements for signal detection, whichever method is employed by MAHs for signal detection, certain criteria should apply, namely: 

• The method used is appropriate for the MAH’s data set. For example, the use of complex statistical tools may not be appropriate for MAH’s with a small data set. 

• That data from all appropriate sources are considered. 
  
• MAHs have systems in place to assure the quality of their signal detection processes. 
  
• That any outputs from cumulative data review are assessed by an appropriately qualified person in a timely manner (and that the QPPV is informed of new information relevant to the evaluation of risk/benefit). 
  
• That the MAH takes timely and appropriate actions and decisions based on the outputs from cumulative data review. 
  
• That the MAH adequately documents signal detection and evaluation. 
  

Signal detection performed only at the time of PSUR production is unlikely to be adequate in most situations. 
However, for a generic substance with low numbers of ADRs (and other safety information) generated which is currently on a six-monthly PSUR cycle, this may be appropriate. 

Flexibility and Improvements

Every organization reaching a certain capability in processing reports will need to set up measures.

If you don't know what you are doing now, how will know what you'll be doing next?

It's very easy to account the reports ratio given by the number of staff you have and the number of reports you process. But this measure is very uneasy to analyze.

So now how do we handle that complexity? Pondering is the best way to go.

Every report has a certain number of data organized by categories: product, reporter, patient, etc.
Every category has structured data: Product's indication, product's dosage, patient's medical history.
Every structured data has more or less information (start/stop date partial or complete).

As you have defined the mandatory information, those are not accounted. This is key to make the measure really informative. What you're interested in are the information that will make the report more accurate, more difficult to process but more valuable for your analysis and organization. This will be the subject of our next post.

accidental exposure (with or without an AE) to study drug (non-marketed) by a non-subject

I would like to share with you this interesting topic.

Susanna Flores, Manager of Drug Safety at SuperGen asked the question and among the answers Bruce Graves Owner, Pharma GCP Support Services, Inc, Dr. Yeshpal Mathangi Clinical Data Reviewer at Clinnovo and Anuj Kapoor Medical Writer at Novartis provided great answers.

In a nutshell, this is a very rare situation, hence it won't require a SOP.

Without an AE, you should get as much as possible from the patient and supporting the investigator to make sure you collect the sequence of the event and save it into your database.

With an AE, you should report it as you'll do for events occuring during the study, making sure you clearly separate this single report from the rest of the study.

In the US, Poison Prevention Packaging Act (PPPA) administered by US Consumer Product Safety Commission is providing guidance on the situation.

CIOMS is working on the question too and Classify accidental exposure as an reportable event: Inadvertent or accidental exposure to study product with or without an ADR http://www.cioms.ch/march2008_pv_in_rpc_final_14dec2005.pdf.

Why flexibility is now trendy in Pharmaco-Vigilance?

When one hears the word flexibility, the image coming to mind is the one of those acrobats bending and rolling over themselves.

In an organization, flexibility should be more like gearing up and down in a car depending on the road and the traffic's conditions.

Of course, every organization would like to be at full throttle.
This is not possible.
Why? As we all know organizations are based on individuals. Full throttle for an individual can not last long and the most important when dealing with PV is consistency.

Regulation, do not forget Pharmaco-Vigilance is a regulated activity, so regulation does not expect you to be good at 80% or 95%. Auditors, sent by management, or authorities, look for every single defect in your organization and they expect you to be at 100% compliance (reports on time). So if you want to be there, you need every one available and focused. You do not want exhausted individuals, painful meetings, passive-agressive approaches.

Question: How do you make your organization flexible enough to bend itself to outer constraints (regulation, compliance) inner constraints (data quality, team management) and achieving sustainability in the proceeding?

First Step: Closer is Better
Everyone agrees when it comes to information you need to be close to the source of it. Not only when people are using their own language it helps understanding the subtilitites )localization/translation) but it also insures you can provide protection under local laws to privacy.

Second Step: Doing good once better than doing several
Processing information must b an industrialized activity. Yes you may end up having queries (questions to the source of the information). Definitely if you control the quality of the data entered compared to the data given (called QC) you should do it once.

Third Step: Beat it when it's hot
When you have to correct/fix something you should do it as close to the step when this information/action was entered/done. More you wait, more expensive the fix will be.

Fourth Step: Empower thy users
You will have to build throughout your orgnization a structure/tool to enhance the knowledge about the defects/failures it achieved previously. If you know why somethign went wrong and if you describe it precisely, chances are you won't repeat it.

In a nutshell, you'll start being flexible when you'll be empowering sooner, closer and better. Giving power to users to make decisions, helping them to understand the failures, collecting the information where it should, doing things in order and once, that the first phase to gain flexibility.

The next one (in another post) will be adaptation.

PV Trends

When it comes to the future, Pharmaco-Vigilance is structured via the guidances and drafts provided by the agencies.

Now every department of every company gets its own dynamic.

The main reasons are varying. It can be a great team work, it can be a strong leadership. It might result from acquiring another company. Whatever, you name it.

• So what are the trends in PV now?
• Where are we?
• Where do we go?
• How fast will we be there?

In the next, posts I will detail what I feel are the most important results we will see in the near future.

What do you learn from PV?

When you are prescribed drugs, even when you take them over the counter without a prescription, the benefits and the risks associated with this drug are monitored.

Not only by the doctors in universities or in centers doing medical research, but by yourself. Getting a rash or a painful throat after taking a drug will lead yo to call the 800 number printed on the box. From there people will process the information and eventually this call will end up being declared to authorities like FDA in the US.

Whomever is calling , whether a nurse, a physician or yourself, will help the manufacturer of the drug to better understand how this active product is behaving in the field.

Even after years in service, someone taking a drug may describe something never seen before.

Don't be mislead, this will not bring any reward to you or the manufacturer. It will cost a lot of time and money to every one involved. Nevertheless, it will guarantee for future patients that we know how things are away from the labs.

Why I like it so far?

I've been working in PV for 30 months now.
I enjoyed it and I'm still learning.

Main aspect of this field:
The bar is getting higher.

As PV is mostly regrouping Mds and Phds from pharmacy to medecine, you have to deal with people used to facing more or less any kind of evolution (chemistry and biology are definitely challenging fields).
Second aspect of the job: people are open to new approaches.

As companies are distributing and sometime manufacturing, products all over the world, you are dealing with a mix of culture and positions that create wonderful teams.
Third aspect of the job: every team has its own dynamic.

Last but not least, you are dealing with a regulated (agencies/laws/auditors) environment where stakes are high.
Fourth aspect of the job: your decisions do matter.

So I do consider to be very lucky to have had the opportunity to get into this domain at a point in my career where I had definitely learned already from other fields and bringing something to the new one.