In June, the Danish Health and Medicines Authority (DHMA), issued its Guide to Individual Case Safety Reporting.
This is a 60 plus page “best practice” (their words) document describing the process for submitting individual case safety reports (ICSRs) to them using the E2B standard to “complement current EU legislation and guidance and provide practical advice to industry.”
The goal of this volume is to help industry and reporters increase the quality of safety submissions which then allows the health authorities to better analyze them and avoid missing signals or creating false signals.
Although this is a useful and laudable document for PV, it introduces changes to the more or less agreed upon ICSR preparations which complicate matters without necessarily making PV more efficient or better for public health. Many of the changes are bureaucratic, time-consuming and add little. This is an example of the fragmenting of a workable, acceptable system which adds time, cost and energy without necessarily improving the “product.” I’ll highlight some examples below.
Even if you are not involved in submissions for Denmark, this volume is worth reading since it does summarize the rules in place for EU for safety submissions. It is similar to the guide published by the British MHRA about 4 years and this Danish guide updates the changes since then. This guide discusses some differences from the EU procedure. These then, are additional requirements in Denmark and similar add-ons exist in other EU countries and vary from country to country unfortunately.
I would suggest your reading through the document. There are some points which are worth a comment and which I do below.
Multiple Reports
Page 9: Interestingly, the agency in DK wants multiple cases submitted if there were multiple suspect drugs and multiple suspect ADRs. They give a couple of examples:
“…a patient who received drug A and drug B. The patient experienced bleeding following treatment with drug A and hypokalaemia following treatment with drug B. The patient also experienced dizziness as an ADR to both drug A and drug B. In this case, three individual cases should be reported; one with suspect drug A and bleeding, one with suspect drug B and hypokalaemia and one with suspect drugs A and B and dizziness.”
In my experience, not all sponsors or MAHs handle cases this way. Many will put all the serious ADRs and the drugs into a single report. The example above creates three cases from one patient experiencing several SAEs/AEs at the same time. It is not clear that the dizziness in the example is a serious ADR though one presumes it is and thus warrants an expedited report.
If this refers to spontaneous reports, the HA is, in effect, requiring that a causality be done with each individual SAE and each individual suspect drug. This is not always possible. Causality, as we all know, is frequently not precise and multiple factors can combine to produce an SAE or several SAEs at the same time.
If this refers to SUSARs in clinical trials, where causality is done by the investigator and the sponsor, then multiple possibilities can occur making this very messy. For example, with the case above in a clinical trial, here are some possible causalities:
Drug |
Bleeding |
Dizziness |
Hypokalaemia |
A |
Sponsor: Yes
Investigator: No |
Sponsor: No
Investigator: Yes |
Sponsor: Yes
Investigator: Yes |
B |
Sponsor: No
Investigator: Yes |
Sponsor: Yes
Investigator: No |
Sponsor: Yes
Investigator: ? |
So do we have 6 cases to be submitted? This is, to me, unclear. It does not seem practical to have this SAE which might have produced a cascade of SAEs (e.g. bleeding producing low blood pressure producing dizziness with an unclear relationship to the hypokalaemia) recorded as 3 or 6 ICSRs. It would be very difficult to understand this case if it were fragmented into 3 or 6 cases. Would the same narrative be used in each or would each have to be tailored to the SAE being reported?
Multiple reports would certainly make frequency analyses of SAEs very difficult to do.
The agency also seems to think that causality assessments are more precise and relatable to a single drug that occurs in reality. If the agency is referring to spontaneous reports, they are, in effect, asking the reporter/MAH to make causality judgments. This would lead to three different ICSRs.
The agency says “judgment should be applied” which means different submitters will apply different judgments and which will produce inconsistency.
This is, at best, confusing and fragmenting. Not sure how it adds to public health particularly since all of this data is in searchable, structured databases. In which case this information can be found by customized searches rather than complex data entry.
Lack of Efficacy
Page 10: The agency notes that medical judgment should be used on a case-by-case basis to decide if individual reports raise a concern about lack of efficacy. This is not unreasonable but it is very hard to make causality judgments and come to conclusions based on a single spontaneous case. Rather a case series should be collected and the cases analyzed together to come to this judgment. Again, a lack of consistency in reporting from MAH to MAH may lead to difficulties and under or over-reporting.
ICSRs from Other Companies
Page 15: This section describes the situation where one company sends cases to another company. It presumes there is a business relationship between the two (or more companies) which is by no means always the case in the post-marketing situation. The agency wants the case submitted only once by the MAH.
They give two examples. Here is one:
“Company A receives a case on a suspect drug where company B is MAH. Company A should not report the case to DHMA as company A does not have reporting obligations. Company A should send the case to company B and ask them to report the case to DHMA.”
This presumes that A and B have an agreement on handling cases. However, if the drug is manufactured by more than one MAH or if there are generics and if it is not clear which company manufactured the suspect drug, it is not clear who did or should submit.
When one company sends a CIOMS I or MedWatch form to another company, it is rare for the sender to tell the other company whether they sent it to which HAs. More than likely it is simply the form with no additional comments. How, then, should the receiving company handle the case? Unless they are sure that the sender sent the case to the HA(s), the receiver should most likely submit the case to be sure they remain in compliance. There is often ambiguity in who the MAH is and whether the sender really did send the case to the HAs. If electronic (E2B) reports are exchanged, are all fields (including HA submission fields) included?
If indeed the two companies are business partners, then this is not an issue in most cases as a phone call or email will resolve the submission question. But if the companies are competitors, it may not be possible to get full information.
Patient Characteristics
The Danish agency alters the ICH Guideline E11 – patient age groups:
ICH Guideline E11:
- Preterm newborn infants
- Term newborn infants (0 to 27 days incl)
- Infants and toddlers (28 days to 23 months incl.)
- Children (2 to 11 years incl.)
- Adolescents (12 to 16-18 years (dependent on region))
Danish Agency:
- Adolescent (12-17 years incl.)
- Adult (18 to 65 years incl.)
- Elderly (>65 years)
It is not at all clear that these additions add anything. What is probably the most controversial and, arguably, potentially misleading is the “elderly” definition. This presumably lumps everyone from 65 to ~100 years of age together. Not clear if this helps.
Timing
Page 24. “It states DHMA has a validation so that reactions reported to have started before the suspect drug was administered are rejected.”
Although logical, it is not always that simple. Sometimes the start date of the reaction and even of the drug are not precisely known or able to be pinpointed (“My chest pain started a few days ago.”). Also, the well-described anticipatory nausea and vomiting can occur even before the administration of certain chemotherapeutic agents in cancer therapy.
Thus there will be instances where this validation mechanism may reject a valid case.
MedDRA
Page 26: The agency writes: “In line with the guidance from the EMA and MSSO, DHMA will implement new MedDRA versions on the first Monday of the second month after release. Any cases submitted with the new version of MedDRA prior this date will be rejected.”
This obviously can cause problems for companies that switch more quickly to the new version. Some companies, for clinical trial cases, do not update MedDRA until the end of the trial to avoid using multiple versions of MedDRA in a study running a year or more.
ADRs vs AEs
Page 31: The agency states: “All reaction terms classified in a case report must be suspected to have a causal relationship to the drug.”
Presumably this refers to clinical trial cases as all spontaneous cases are presumed to have some level of causality suspected by definition.
Rule-outs and Provisional Diagnoses
Page 33: The agency notes it codes provisional/rule-out diagnoses in its case handling but does not expect companies to necessarily do this. Each company can do as they wish and document this in working practices.
Again, the lack of consistency will produce different ways of handling this from company to company making coding and signaling harder, particularly if it is not clearly indicated that a code is provisional or if that information is buried deep in the narrative.
The Narrative
Page 35: This is a fine summary of the ideal narrative indicating that “The narrative should serve as a comprehensive, stand-alone ‘medical report.’ The information should be presented in a logical time sequence; ideally with the source of the report and presented in the chronology of the patient’s experience.”
This is fine for simple cases but less clear for complex cases with multiple follow-ups. Should the new data be appended? When should the whole narrative be rewritten to combine the 5 multiple follow up reports into a single, cogent “stand alone medical report” similar to a hospital discharge summary for example?
This is the tricky part and is not addressed, alas companies handle narratives in multiple ways.
Literature Reports
Page 45: This section appears to have been prepared before the change being put in place whereby the literature searching is done by the EMA.
Comments:
These are some of the key issues I see in this otherwise fine document. There are others, some of which are technical and relate to the structure of the fields in electronic reports. If you submit in Denmark you must familiarize yourself with this document and implement the required and possibly the suggested changes.
This is a useful document worth reviewing, particularly in regard to the E2B handling of various fields.
As noted above, the intent of these requirements is good. The issue is that each individual agency in the EU and elsewhere in the world, may make some minor (or major) changes to what are more or less standardized CIOMS/ICH/EMA/FDA requirements which add cost, time and burden to those doing PV. Such is life.