SAFETY RISK MANAGEMENT PLAN (RMP) HAND BOOK

The 12 “Must-Know” Concepts before Beginning RMP Authoring

RMP authoring is beyond collating content; it is a strategy in itself that is legally binding.

It has been a couple of years now, the industry has evidenced major and minor overhaul to the new EMA’s Risk Management Plan (RMP) template since late 2012. This document has taken a proper form and shape in terms of logical flow, consistency, and transparency through EMA’s initiative to bring in a submission friendly modular format. The authority’s emphasis on the risk-benefit balance and the outcome of risk minimization (RM) measures are compelling the Marketing Authorization Holders (MAHs) to be more vigilant in their approach to present and submit an RMP.

Over a period of time, RMP authors have noticed that this comprehensive, elaborate, robust, and conservative Good Vigilance Practices (GVP) template and guidance have flaws lingering around some of the concepts and requirements which are either ambiguous or unexplained. These could only be understood or rather interpreted based on rapporteurs’ or Pharmacovigilance Risk Assessment Committee’s (PRAC) assessment reports. The aim of this article is to flash a spotlight over 12 must-know concepts or requirements of an RMP; nonetheless, this list is not a limited one. This could possibly clarify some of the common inadvertent slip-ups while RMP authoring and aid in avoiding health authority questions (HAQs) that are uncalled for.

Before beginning to read through the list, one must be clear on a simple yet significant concept that an RMP authoring is not a mere collation of inputs from different contributors; therefore, not the sole responsibility of a ‘Medical Writer’ to develop an RMP. RMP authoring is beyond collating content; it is a strategy in itself that is legally binding, as it deals with defining MAH’s commitments for ensuring patient’s safety by optimizing benefit risk balance. In essence, an RMP requires a “team of authors” from various departments, not confined to safety, clinical, non-clinical, epidemiology, regulatory, biostatistics, technical research and development, non-healthcare professionals. Here is the list of “must-know” concepts of an RMP in the chronology of their appearance within the EMA template which also conveys the above said point.

1. Product overview: Overview of versions under administrative information of this section should state the procedure number of the agreed RMP version and the RMP version under evaluation (if applicable). The irony with this requirement is to see how the “procedure number” is being misinterpreted by some of the MAHs where they actually stated the “internal SOP number(s)” relevant to RMP authoring.

To clarify the true elucidation, let us put it this way. Would EMA be really interested in knowing which internal SOP has the MAH followed for their RMP development? Thought provoking? Well, having said that this section intends to state the “Marketing Authorization procedure number” given in the EU for every submission that falls either under centralized, mutual recognition, or decentralized procedure. This indicates that the level of review being performed by the ‘Regulatory Lead’ in an RMP should be more stringent and diligent.

2. Epidemiology of the target indication: Misunderstanding around presenting concomitant medications in the target population versus medications commonly used for co-morbid conditions seems to be generating repetitive data in both the sub-sections within the same module.

• Example of co-morbid condition from GVP V: If a medicinal product intended for treating prostate cancer, the target population is likely to be men over the age of 50 years who are usually at a risk of myocardial infarction. The intent here is to identify if the drug might be increasing the risk of myocardial infarction as compared to target population of the same age group who are not taking the drug.

• On the contrary, a concomitant medication is one that is being commonly used for the disease itself or the complications of it. Example of concomitant medication from the template: Anti-hypertensives frequently used alongside hypoglycaemic medication in the treatment of diabetes. Therefore, the list of concomitant medications stated in the target population should not be the medications that are being used to treat co-morbid conditions as the requirements of these two sub-sections are discrete.

3. Conclusions of non-clinical data: This is a tabular list of safety concerns that are evidenced in the non-clinical setting, clearly characterized into one of the risk categories based on the evidence backed by clinical data. As the template says “list of safety concerns”, it should be clear enough that there is no requirement to provide further textual explanation before or within this table since the table above this list on “Key Findings” serves the purpose.

Coming to the key clarification point, this list of conclusive safety concerns should be appearing in the ‘SVIII Summary of safety concerns’ and ‘Module SVII identified and potential risks’ under respective category. It should be noted, this does not imply that the list of safety concerns in SVII and SVIII should be appearing in the non-clinical conclusions, if the findings were not observed in non-clinical setting. The simple rationale behind this is, not every safety concern will come out of non-clinical setting, but rather directly evidenced in a clinical setting. This approach holds true for conclusions of Modules SIV (Populations not studied in clinical trials) and SVI (Additional EU specific requirements).

4. Effect of exclusion criteria in the clinical trial development plan: Two things are being frequently seen in most of the RMPs despite straight forward, clear template instructions. One, including the comprehensive list of all exclusion criteria; two, listing them trial-wise. If one has to know “what not to do in an RMP”, this would be for sure in the list.

The principle behind this section is to portray only those exclusion criteria from the whole clinical trial program which are being medically assessed as having a potential effect on the target population. Meaning to say, only if those excluded populations are commonly found in the target population based on epidemiological evidence. For instance, diabetic population will have a sub-set of population with renal impairment. If a study involving anti-diabetic medication with renal impairment as one of the exclusion criteria, the impact of excluding this population in the target population should be discussed and justified.

Another point here to clarify is that not all safety concerns concluded from the list of exclusion criteria being discussed in this section will necessarily be categorized as missing information in Module SVIII. This is a direct interpretation from the template instructions: “If the missing information has been adequately investigated outside of the clinical programme this should be noted (with cross reference to the appropriate RMP section) in the comment section”. Implying, it could be an identified or a potential risk based on the potential magnitude of impact that is anticipated or evidenced of that exclusion criterion. Most often, this happens for approved products when evidence gets generated from the post-marketing setting.

5. Action taken by regulatory authorities and/or marketing authorization holders for safety reasons: This Module SV has a major discrepancy that is probably leading to lesser transparency in reporting actions taken for safety reasons. If one has noticed the template instructions that are very specific which states “List any significant regulatory action”, and “by regulatory authorities and/or marketing authorization holders”. The former has the word ‘significant’ which is not specified in the heading of this sub-section and the latter directly implies that actions taken by third parties (conducting cooperative group studies or Investigator-initiated trials) need not be listed though might be taken for safety reasons. Both scenarios have an ambiguity that lead to underreporting of actions taken for safety reasons.

However, since the modular format has an obligation to align the data and messages with other safety documents - PBRER and DSUR, most often it is expected to directly copy over the content from them. If one has carefully read through the guidance instructions, the definition of “actions taken” is a bit different for an RMP as compared to PBRER/DSUR. The definition in Section 3.3 of PBRER and DSUR is more elaborative and logical in covering all the actions taken for safety reasons beyond what is defined in the RMP. This could probably be due to the fact that RMP template is developed entirely by EMA and PBRER/DSUR is developed from ICH templates, bringing in this difference between these safety documents.

Nevertheless, the bottom line recommendation would be to keep it consistent and conservative in stating all actions taken to avoid any questions in the assessment report as long as they are due to a safety reason associated with the product.

6. Pediatric Investigation Plans (PIP): This is a philosophy driving the pediatric regulations which mandates that every product be tested in pediatric patients when pediatric use is anticipated. In the RMP context, reference to PIP is required if the target indication is one that is seen in pediatric population apart from adult population. Based on the feasibility and epidemiological data, the product may have a waiver or deferral in the EU.

The logic in this section should be consistent with the key message stated in the section ‘potential for pediatric off-label use’. When potential for pediatric off-label use of a product is positive, then it cannot have a waiver. If the product has a deferral or has no waiver, the key issues need to be addressed in Module SVI.

7. Identified and potential risks: This Module SVII which details different aspects of each safety concern in a tabular format does not provide clear instructions for some of the elements that include - nature of risk and preventability.

Unlike frequency (95% CI), severity, and seriousness which are self explanatory in their terminology, ‘nature of risk’ seems to lack clarity as to what should and should not be described under this element. Most often redundancy has been observed in the amount of data being presented under “nature of risk” which is repetitive of the data pertaining to the three elements stated above. Instead, nature of risk should entail discussion of the risk beyond frequency, severity, and seriousness. For example, time-to-onset, time-to-resolution, exposure adjusted rates, any complications, events requiring intervention and/or dosage adjustments, etc.

On the contrary, ‘preventability’ is self-explanatory and yet does not seem to be interpreted in its exact terminology by the authors. This needs to have a description of the measures that are being proposed by the MAH to prevent a specific risk and not measures that are to be taken to mitigate or reduce the impact of a risk after it has occurred. Authoring this section is no big deal as it is covered entirely in the label (SmPC). The basic tip here is to only understand the obvious difference between prevention and mitigation.

Another frequent observation is the team’s requirement to align the language in the risk table for each safety concern with the label. Before arguing further around this point, the preliminary question to ask is - If there is a compelling reason to do this, why do we have a separate RMP ‘part: Risk Minimization Measures?’ . To put it in a straight-forward way, the RMP GVP guidance has a specific table “Mapping between RMP modules and CTD” which states “CTD modules 2.5 and 2.7” and “SPC”. While it should be kept in mind that the majority of the data presented is to characterize each risk which basically comes from the CTD modules and the scope of SPC remains very narrow limited to “Preventability”.

Therefore, authors needs to reduce the amount of time ensuring verbatim alignment with label wording as majority of it speaks of mitigation or prevention measures which entirely has a separate place in the RMP. Instead, the focus should be on aligning the ‘key safety messages’ between the sources and the RMP.

8. Pharmacological class effects: Through template instructions, one can certainly understand that if a pharmacological class effect is already listed as a risk associated with that product, overall clinical data in comparison to that of the reference products of the pharmacological same class should be summarized. However, the last column of the table “Comment” lacks clarity as to what areas of the risk should be commented upon.

If one can catch up on the pulse of this section, the prime rationale is to highlight if these risks of the product that are considered to be pharmacological class risks are higher or lower compared to the reference products. If so, to detail what could be the potential reasons for these differences between the product and reference. It does not end here, but to evaluate if there are any confounding factors that may have led to these differences, for example, different frequencies due to differences in target populations.

Vice versa, when a class effect is not assessed to be a risk associated with the product, respective data similar to that of a risk table needs to be presented as an evidence to show that there is no association with the product although a class effect. The glitch here is that often the Medical Writer and the Statistician are not provided with this information until the late stages of an RMP development and so, this requirement remains unfulfilled. Obviously, MAH is bound to expect HAQ on this template non-compliance.

So, the recommendation would be to generate analysis for all pharmacological class effects as required by the template regardless of their association with the product. Again, it is the Safety Lead who should be diligent in providing this list to the team upfront for creating respective data outputs.

9. Pharmacovigilance (PV) plan: This part of the RMP is quite critical for the MAH as it details all the commitments made with the health authorities’ consensus which brings in the element of legality in this document. Three points in relevance to additional or targeted PV activities/studies in this section are probably overlooked by the authors, which leads to inconsistencies within the document and non-compliance with the template.

First, misinterpretation is often observed in this section where in authors believe that if a PV study is being proposed for a particular safety concern, then that study protocol should list in its objectives the assessment of that specific safety concern. This is definitely not true. The rationale being that, if MAH already has plans to conduct any study which collects and evaluates safety by default then they do not have to propose another study to evaluate a specific safety concern. In the end, why would the MAH want to spend additional resources (time, money, subjects, etc.) in conducting another study when the one that is already in place serves the purpose? Nevertheless, this may not be true by the virtue of a specific safety concern and not be feasible to evaluate in an already existing/proposed study. For example: Off-label use, use in special populations etc.

Second, usually there is an inconsistency in presenting the additional PV activities between PV table of each safety concern and summary table of PV activities. If HA assessment reports need to be void of questions on listing PV studies, the utmost prerequisite is that the medical writer should have an eye for detail in checking consistency between these two sub-sections. As the template directs, each PV study listed in the table of ‘summary of PV activities’ should be reflecting under the column of “Safety concerns addressed” all the safety concerns that have this particular study proposed in respective PV tables.

Last but not the least, a simple inconsistency is noted in stating milestones of PV studies; authors at times provide dates of finalization for one study and dates of submission for another. The clarification in line with the template heading “Submission date” is that for a completed study it is not the “date of finalization” of the published CSR (interim/final) that should be stated. The template clearly asks for the submission date of the CSR to the HA. A connotation to this is that if the study is ongoing/proposed, obviously the HA expects to see an anticipated date and of course not necessarily an exact date but at least an approximate month-year.

10. Post-authorization safety study (PASS): This is one of the critical concepts of an RMP that an author should be aware of. While it is a PV activity, it should be clearly and consistently stated as to which PV category it belongs to. These are the studies with a specific safety concern as a primary objective unlike other trials which may have efficacy or safety objectives, in general. These are often a part of marketing authorization (MA) obligation or mandated by HA and therefore, should be classified under category 1 or 2, as applicable.

Parallel to this concept, if a study has an objective to evaluate the effectiveness of risk minimization measure it should be considered as a PASS, by definition. In such a case, it would be a category 3 PV activity. Yet, it should be noted that such studies at times could be an obligation or a mandate and therefore, it is the medical writer’s responsibility to confirm with the ‘Regulatory Lead’ on a case-by-case basis before assuming under either category. In essence, the point here is to understand the thin red line while PV categorization for a PASS based on their regulatory triggers.

11. Post-authorization efficacy studies (PAES): During the initial phase of new RMP template, this concept was perceived by the industry that it might not be critically evaluated by the Agency. The inclusion of this requirement clearly signifies that RMP does not only focus on safety rather, risk-benefit ratio. Until recently, most of the RMPs were observed to be dumping numerical data of efficacy which is not the right approach. The essence of this requirement is to not only evaluate efficacy in the target population but also discuss any differences between the sub-populations.

Authors should ensure that they adhere to the guidance requirements unless they may not have data to discuss further. Nonetheless, it should not be excluded for the same, but to make negative statements while discussing appropriate rationale. This would showcase to the HA reviewer that there might be opportunities to evaluate efficacy in the postmarketing setting. Authors might be speculating that it would elicit potential obligations of MA leading to an increase in the cost of RMP implementation. Not that the MAH should be inclined to develop an RMP strategy with a view of cost-cutting or be conservative, but the recommendation is to scientifically and medically justify the decision either way in view of patient safety.

As the HA started to review this requirement more critically especially for the newer products, MAHs seem to be justifying the risk-benefit balance with relevant discussion around the applicability of efficacy in the target population for various sub-groups.

12. Risk Minimization measures: Though occasional, it is quite astonishing to evidence authors still figuring out the difference between a ‘routine’ and an ‘additional risk minimization (RM)’ activity. Specifically, to see how the “Dear Doctor Letters” or “Dear Healthcare Professional Letters” are not listed as additional RM activity by the authors. This activity is something which definitely is not done for all safety concerns or all products. Implying, if anything is not proposed and implemented routinely, it should be inferred that it is an additional RM activity. It does not end here to classify it but to go beyond it and confirm if it is Global activity or a Local activity. One can blindly include within the RMP a global activity regardless of the region. The key is to have clarity on a local activity depending on the region since if a local activity is not applicable within the EU, it does not find a place within in the EU RMP.

Being such an extensive guidance and template, the list cannot be limited to just 12 elements discussed so far. With every new RMP, each author would realize that it is unique in itself. It always seems new every time a new RMP is authored with a new learning at every step.

Some authors might perceive the RMP template to be too bureaucratic and redundant if they have not understood the rationale behind this repetition with a different approach across modules. One needs to be receptive to the fact that regulatory authority did not develop the RMP to have redundancy within the same module, part or section; rather to showcase optimal and logical correlation of the data and its flow from one module to another. Stakeholders involved in developing regulatory guidance(s) are aware that these are living documents, dependant on the dynamic regulatory landscape with the advent of wide range of new therapeutics. Therefore, RMP authors should invest some time in understanding the RMP GVP guidance, religiously follow them, and learn to tackle the flaws on a case-by-case basis.

Based on the conventional organizational framework in the industry, the contributors are always aligned per product but the medical writer is not. Therefore, the quantum of exposure on RMP authoring that the medical writer has, would always be more than any other RMP contributor as they work on more than one product. So, from the contributor’s perspective, the key driver is the ‘Medical Writer’. To live up to this expectation, the writer is responsible to train the “team of authors” on RMP requirements based on various triggers.

Likewise, it is the team who would have a broader knowledge of a particular product and its regulatory lifecycle than the medical writer has. This leads to medical writer’s basic expectation that the relevant contributions on the scientific or technical aspects of the product in an RMP would be provided the contributor(s) without having to follow-up frequently. Therefore, it is the team’s responsibility as a whole to collaborate and integrate one another while learning and sharing, which would aid in the development of a ‘HAQ free RMP’ at least in terms of template compliance.

To conclude, authors should not overlook the fact that RMP is legally binding and that they have to be watchful of every word and strategy. Furthermore, it not only is a proof of regulatory commitments of MAH that have monetary implications but is also critically interdependent on the label which is directly proportional to the cost of RMP implementation. Excelling ‘the art of RMP authoring’ cannot be as complex as it seems provided a ‘logical and scientific approach’ is followed while applying the guidance in any scenario of an RMP development.


Author’s Note: This article as it is evident, is only a limited guidance and displays intent to initiate thought provoking collaborative discussions across the industry. If one may have alternate opinions and/or further recommendations in their experience on this topic are encouraged for an open dialogue.