Continued from Part 1.
We continue with the analysis of the Institute of Safe Medication Practices’ recent report on FDA and FAERs of late January 2015 entitled “A Critique of Key Drug Safety Reporting System.”
Gaps in FDA FAERS Coverage
The report also comments on some FAERs issues.
AEs in Children
Too few cases. Children (defined to be under 18 years old) are 24% of the population and get 7% of dispensed prescriptions but account for only 3% of the AEs.
Comment: One wishes the cutoff had been lower, say 12 years or age or even less. Or perhaps even better several tranches such as 0-3 years, 3-6 etc. No question though there are too few reports.
This is not really a criticism, I think, of FDA or FAERs but an observation that the current medical system does not handle medications for children very well. Manufacturers are hesitant to study new drugs in kids for multiple reasons even though FDA and the EMA and other agencies provide incentives (e.g. patent extension) for studies in children. HCPs may be hesitant to report AEs, particularly if they are prescribing a drug which is not FDA approved for children – which is a very common practice amongst pediatricians.
The report notes that the CDC estimates that 4% of the 4 million live births in the US have some sort of anomaly or birth defect. This would suggest some 160,000 defects. The report also notes that there is “no credible estimates of drug-related birth defects, but could not rule out that they number in many thousands.” Of the reports sent to FDA there were few (about 3,200) and only 25% were “reasonably complete.”
Comment: Although possible, there is no clear basis to say there would be several thousand such cases. It is not clear pregnant women take as many drugs as the general population and extrapolations are always problematic where there is no solid data. Certainly though, an area sorely in need of data and study.
An excellent point is made about generic drugs. First 86% of outpatient prescriptions in 2013 were for generics. The original brand name manufacturers now sell very little (1%) of the original product (about 2.1 million brand name prescriptions and 206 million generic prescriptions cited in the ISMP report) but supplied 68% of all SAE reports on these products. This touched big pharma primarily including Pfizer, AstraZeneca, Merck and Janssen for five products (Zoloft, Prilosec, Cozaar, Lipitor and Risperdal). The implication is that generic houses do not get or report many AEs.
Comment: There were, amongst the cases received, very few SAEs were received and most of the SAEs were for the branded products. This is somewhat surprising; there is no reason to think that the branded products are “more toxic” than the generics. For this set of products, about 2300 SAE cases were reported of which about 1600 were from the branded products. The report discuss some of the reasons for this.
Clearly there is under-reporting of SAE cases as a product ages during its lifecycle and as generics come on the market. What is not discussed is whether this is a critical issue. Although a drug’s safety profile is never fully and finally known, clearly the safety profiles of these old products are better known than drugs newly approved and marketed. The question is then: how much effort (and money, time etc.) should be put into covering the safety profile of old drugs and new generics? Should the bulk of the resources be put into new products? An on-going debate.
Duplicate Literature Reports
This is a known issue. The current laws in the US require that all manufacturers of approved products with reporting requirements scan the published medical literature for serious AEs. When found, unless the manufacturer or brand name is clearly indicated, each company must database and report the case to FDA unless they can clearly show it is not their drug. This means that every generic manufacturer will end up reporting most of the literature cases producing many duplicates for FDA to read. For some drugs there can be scores of manufacturers and many duplicate reports. FDA is trying to eliminate this but in practice it is hard to put into place operationally.
The report notes that the FDA, now in the digital age, is behind in updating and modernizing the laws and regulations which were made in a paper-based era years ago. The ISMP report does not mention social media, but there are some hints now that Twitter, Facebook and other social media sites and apps may have a major role to play in the field of drug safety in the new, digitally-connected world.
Patient Death Reporting
A very interesting point is made by the report. The pharma world is changing now with an increase in new drugs for life-threatening disease (that is, very ill patients), orphan drugs and new cancer treatments. This is increasing the number and proportion of death reports, many or most of which cannot determine whether the drug played a role. They cite a particular example where several thousand reports came in from patient support programs and for which follow up with the health care provider produced very low response rates. The company was cited in a warning letter by FDA. The report notes that FDA is applying “an outdated regulation to a situation not contemplated in current regulations and guidance.”
Comment: Excellent point. The way safety data is handled is markedly different now and the old rules don’t seem to be very useful. There are many new sources of safety reports today compared to twenty years ago when most came in as spontaneous reports. Everyone is aware of this issue as well as the tie-in to social media and the globalization of safety information and reporting. Expect to see lots of discussion and changes (but more slowly) in the way safety is regulated.
Top Priorities for Reform
The report says that FDA’s AE reporting system needs a “thorough overhaul, if not a comprehensive redesign.”
Comment: Hard to disagree. Probably true for the EU also even though they just redid their system.
The report correctly notes that quality improvement should be a priority. Manufacturers should be able to meet the 85% completion rate of FDA.
Comment: Hard to disagree that quality improvement should be a priority. However, here I think careful thought should be put into the definition of quality. One can also argue that quality (and effort) should differ depending upon the importance of the report. Serious reports should be higher priority than non-serious, unexpected higher than expected etc. This is actually very hard to put into a set of rules since sometimes a non-serious case is important and sometimes a serious case is less so. The focus on the event date is useful only from a compliance (15 day report) point of view. I’d prefer to see, perhaps, an increase in effort to collect co-medications, history and risk factors for the SAE and a good, concise narrative, particularly for the subset of important new drugs, drugs given to very ill patients, orphan drugs or drugs where the clinical trial safety dataset is small etc. Maybe big data and electronic hear records will move us in this direction.
In addition, it is hard to measure quality. Counting (number of boxes checked off) is one way to measure quality but perhaps more important (but very hard to do) is a case with a concise, cogent narrative rather than a long data dump, as well as a complete co-medication list and medical history.
The report here is focusing on patient support lines, restricted distribution pharmacies, on-call nursing support and other new methods of collecting safety data. It is suggested that a “serious of pre-approved protocols for manufacturer-initiated contacts” would collect better data.
Comment: Hard to disagree that new means to collect data that is not spontaneous are needed. This will be more than just “pre-approved protocols” (this is not defined in the report). One suspects that social media and other “big data” related methods will come into play here.
The report notes that generics account for 86% of dispensed prescriptions and that we are not getting many safety reports. Those that do come in are often duplicate literature reports. This should be a priority for improvement.
Comment: Fully agree again. No one knows how to do this yet; the report states that a solution is “not immediately evident”.
The ISMP summarizes some of the ongoing research and thoughts to improve drug safety including better and easier physician AE reporting with electronic health records. They note that new systems to “provide better intensive postmarket surveillance for high-risk drugs rather than the present situation, which produces large numbers of reports of dubious quality and value” are needed.
Comment: Again, hard to disagree. There are efforts underway in the US, EU and elsewhere along these lines. They are fairly thin, scattered, uncoordinated and do not have a very high priority. There is little academic effort. In fact, there is little work being done in what I might call “operational pharmacovigilance”. There is work being done on genomics, prediction of high risk patients etc. One needs to “kick start” (and better fund) these efforts.
There is an extensive review of sofosbuvir (Sovaldi) suggesting issues with the quality of manufacturer reporting and “an uncontrolled key pivotal clinical trial” which impair signal detection and make cases that are identified difficult to evaluate. The reader is referred to the report for the information on this.
ISMP & Funding
Finally there is a brief summary of the ISMP funding sources and capsule bios of the authors.
This is an excellent and important paper from a highly trusted, independent organization. They make points that most of us in the PV world are well aware of in one way or another. However, these points are minimally understood by non-PV personnel in the pharmaceutical world and not at all understood by the public and most of the health profession (alas).
The authors have done a fine review of FDA FAERS data to make their points. This study should be used as a kick-off for work on solutions. Perhaps this paper could form the basis of a major conference of all the major stakeholders to get feedback and, more importantly, brainstorm on the next steps and solutions.
As one can readily imagine, this paper has received much coverage and pushback from the companies and manufacturers. The FDA has also commented. Most of the comments are quite defensive along the lines of: we are doing the best we can with an outmoded system and, of course, changes need to be made, there isn’t a lot of funding, there are other health priorities etc.. See the NY Times article referenced at the beginning for some more discussion about this.
There are many stakeholders involved and the issue is global not just in the US with the EU and several other regions/countries being players. Given the major number of deaths, injuries, hospitalizations and other problems associated with medications, this is an area of public health that is crying out for more and better data collection and analysis.