AdverseEvents Inc. is one of a handful of companies that are commercializing the FDA's Adverse Event Reporting (AER) data. The data has become increasingly useful to researchers, physicians, pharmaceutical companies and, of course, individuals who are using it to make business and personal decisions about drug safety.
The FDA database contains a compilation of adverse event reports from manufacturers – which they are mandated to provide – and voluntarily submitted reports from clinicians and the general public.
What sort of background is needed to aggregate the data and re-format it into something more user-friendly?
In 1999 I started a company called Sagient Research which is in the business of aggregating, standardizing, presenting and selling enterprise-level data to financial services and pharmaceutical companies. I ran that company as the president and CEO for 10 years. We sold it on May 31, 2012, to Informa PLC (LSE: INF).
Having database expertise and web development skills is great – and we have that on our team -- but actually understanding the data, its shortcomings, how to correct for those, and how to create a marketable product at a high level is something that few people know how to do.
One of our key competitive advantages is having 13-plus years of experience in not only how to aggregate and clean data, but how to present that data for sale to multiple constituencies.
How many drugs are currently in your searchable database?
We have taken the 200,000-plus drugs that are presented in the FDA’s raw data feed and aggregated those down to approximately 4,500 brand name drugs.
What kinds of augmentation have you added to the AERS reports? Do you track journal articles concerning drug adverse events? What other sorts of information do you include in your reports?
We actively produce proprietary daily news on drug safety through our quarterly Drug Monitor Report. Unlike the automated news aggregation that is presented on some of the other sites, we actually have analysts who read the base material, such as news reports and journal articles, and then summarize the salient facts and layer in some analysis based on our view of the related adverse events data. We also publish regular in-depth articles for publication in scientific journals.
How is AEI different from other companies that aggregate the AERS data?
The other companies that have launched recently are publishing the raw AERS data with a web front end to make searching and viewing results easier. The problem with this approach is that the raw AERS data is fundamentally flawed. There is no control on the input of data into the FDA’s database, so there is no control on the output. We spent a year and a half developing an algorithm that reads through the data and standardizes it for proper use and presentation. Some of the main attributes of this system include:
- Aggregating generic, non-U.S. brand, misspellings, and data errors in the drug names into a single unified report. For example, there are over 700 different Lipitor identifiers in the FDA database. We aggregate all of the cases for those 700-plus identifiers into a single report so users can get the proper overall picture of the side effects.
- Removing duplicate cases. There are a huge number of duplicate cases in the raw data. Failing to remove those makes the data seem worse than it actually is and can skew research and comparison results.
- Identifying and separating out primary suspect cases from other case types. The other sites treat all cases within the database the same, but that’s a fallacy. If a drug has been identified as the primary suspect for causing a specific side effect, that’s very different than if a drug is identified as a secondary drug in the case. Treating those case types (and others) as equals for presentation purposes is troublesome.
- Identifying and separating cases out by reporter type, nationality, and other unique attributes.
- Timeliness. The FDA releases data every quarter. We strive to run new data through our algorithm and have it loaded onto our site within 3 business days of the release. The other sites are not regularly updating their data, often six months or more delayed.
- Proprietary searching and alerting tools. We’ve developed sophisticated custom searching and email alerting modules that allow users to access the data they need and eliminate the noise created by the sheer volume of data that is of no concern to them.
There are many other aspects to our process not listed above. While I’m pleased to see other sites reporting on this data (as we obviously feel it is very important), I’m concerned by the lack of proper treatment of the data before presentation on those sites.
How do you discourage news media from sensationalizing the data?
In every case, we speak to the reporter and assist them with procuring the data and making sure they understand what they are looking at and how to properly present it. We’ve been in many media reports since launch and I’m comfortable with the data that have been reported in those outlets.
I noticed you have no pharmaceutical or other ads. What is your business model?
Our initial model is the sale of our data and the related proprietary tools to enterprise and professional users (pharmaceutical companies, insurance companies, healthcare practices, research labs, universities, etc.).
The second phase of our business plan is the roll out of new services based on the data for the individual patient and healthcare provider.
Who are the biggest users of AEI so far? Which percentage of users are individuals, researchers, physicians, market research, pharmaceutical companies?
At the moment, it is the enterprise and professional users. Our data is being used in multiple research studies being prepared for publication as well as within the pharmacovigilance and drug safety departments of a number of pharmaceutical and biotech companies. As we roll out the consumer/physician facing modules, we expect that the sheer number of users and traffic will shift to be heavily dominated by those users.
I noticed that basic data for the public is free but that you do custom reports for companies and organizations. What is the price range for custom reports? What kind of reports do clients seek?
Enterprise and custom report pricing is based on a number of factors, including the number of users, the extent of the licensing agreement, etc. We don’t publish specific prices publicly for those levels of service.
For more information about other companies working with the AER data, see “Firms Make FDA Adverse Event Report Data Accessible.”