The Promise of Integrated Drug Pipelines From InfoDesk: Will Multisearch Work?
Posted On September 13, 2010
The benefit to being cohorts with a popular editor: insight into potential new products. In this case, a new tool under development by InfoDesk promises to integrate pipeline data from multiple drug information sources into a single set of results. Think of it: searching three or four databases and only dealing with a single output.
This could be a huge time saver to those researchers like me who download the drug profiles from hundred of drugs into several files, one per database, and then use BizInt’s patented Smart Charts software to combine them into a single set of results. Anyone using databases outside the Big 5 (Thomson Pharma, Prous Integrity, Pharmaprojects, ADIS R&D Insight, IMS R&D Focus), currently need to create their own customized solutions in house. Other databases—ask someone? I refer to such as Inteleos and MedTRACK as quick examples.
And, did you just read that right? A single umbrella search system with a single output product that combines all those pipeline databases into one simplistic program? Really?
This is what InfoDesk has decided to undertake. The company purchased a drug pipeline integration platform from Elsevier that it is currently redeveloping for release early next year. I’ve read a white paper discussing their plans, taking what they have learned from a sampling of information professionals and business development experts. The result: “InfoDesk is developing a tool that integrates multiple drug pipeline data sources with real-time intelligence to create a single set of integrated search results.”
The core of its solution cannot be a secret—the basic concept of umbrella mapping of taxonomies. The proposed framework has not yet been seen, so this researcher has a number of concerns.
First, will there be a limit to the number of databases included? Will the decision be made for the Big 5, leaving out the smaller companies that many times focus on the niche markets of oncology and diabetes? It would be highly beneficial to many if InfoDesk found ways to incorporate other resources into its proposed system.
Second, how does InfoDesk plan to connect access to these systems? Contracts and access points abound based on the platform. There are IP address verifications for some pipeline systems and username/passwords for others. There are organizations that maintain long-standing concurrent access for a select group of users, and provide individual access for others. The scope of contracts and access for my company alone can be a headache, and we’re not even international with language barrier issues.
Ah, the third question: will this be offered in multiple languages? Even English varies from country to country. Hematology and haematology are the U.S. and U.K. terms. Larger companies will undoubtedly have the need for German (Hämatologie), Spanish (hematologia), and other languages. Currently it’s a training issue for some users on select internationally-based platforms to think about alternate spellings.
My chief concern is echoed by a colleague of mine at a small, California-based genetics firm. Just how detailed of a search can this platform really handle?
Many researchers using the pipeline database programs are interested in the basic searches: how many drugs are in breast cancer in phase III trials? What is the background of the drug, what other indications are being pursued? How many entities are being developed by Companies A, B, and C?
General searches could be easily handled with a simplified platform. Several drop-down boxes, one for the country, another for indication, another for phase….press search for the results. Push a couple more buttons, create a single spreadsheet, and the user is happy.
Other companies have created such a simplified search system: Dialog1 was an early product. The back-end is a set of complex search strings set up by professionals but users only see a more Google-like interface with simple search boxes. For anyone looking for general information in the larger industry areas, such as biotechnology and engineering, this is a great one-site shop. Professional researchers often find the system cumbersome and restrictive. We know what we want and how to get it, so we want the options provided by the full Dialog Classic system.
It’s not that one is a better product than the other. It’s simply a matter of understanding the target audience—general researcher versus dedicated scientist. Will our dedicated users be willing to sacrifice the ability to search the biomarker field in one system in order to combine the results from several systems? At what point do my scientist patrons decide that sorting through a massive quantity of potentially non-relevant drugs resulting from a simplified search process will be easier or more useful than if they utilized the powerful search capabilities that start with focused results from multiple databases that they have to manually combine?
That brings up the last issue on my plate, one that will—for my company—be the make-it-or-break-it purchase of the proposed InfoDesk system: what exactly will the output resemble?
Routine users of the Big 5 databases will understand the inference here. No one system outputs the same data fields and the same number. For example, Pharmaprojects allows users to create customized reports on every field in the database, the limitation being lengthy description fields not fully exporting all the time (which is caused in part by a restriction in Excel 2003). The number of drug entities that can be downloaded is hampered only by users’ web connections and speed of their computers.
Prous, which is about to be rebranded Thomson Reuters Integrity, provides users a simple drop-down report option with check-boxes for less than 30 fields. Users can only export 200 entities at one time. In order to export more, users must run several exports, choosing first drugs 1-200, then 201-400, and so forth. Will this be an issue with a single output system?
Thomson Pharma, built off the IDdb3 database, is designed primarily for users to be on the system for a lengthy period of time, using click-throughs to stay updated on news and trials. The output feature is poor, allowing users to export just 6 data items at one time. Depending on the part of the database a user is in, the six items change. Users can, in theory, export up to 15 different items by running three different exports and combining them into one file. This is a major restriction that could greatly impact the plausibility of a single export system.
What happens with the different data fields? What if one system says the highest drug phase is Phase II and another system claims the drug is no longer under development? The BizInt system provides rows for individual databases, a great summary chart but tricky to sort. Will InfoDesk have a specific (user chosen) database act as the default data, with a macro adding in the additional data in adjacent columns or rows?
There are many unanswered questions, but this is also in part because the development of the InfoDesk solution is far from complete. This is a proposed system, the result of a survey of key people within organizations. I commend the InfoDesk developers for asking about the unmet needs, for analyzing the results, and for thinking about possible solutions. I’m sure I’ve raised a few questions that are about to cause them headaches.
I myself would love to have a single umbrella search system for my organization. It would mean training users on just one simplified platform, one manner of searching, and provide better desktop access in a standardized language where the output will be systematic. The time savings would easily argue paying extra for such a product.
But…going back to the user with the country search by indication—will the proposed platform work? I believe the answer is the infamous 80-20 ratio. If InfoDesk can overcome the major problems, I feel that about 80% of searches within the pipeline databases will be addressable. I feel many people want basic information and aren’t the hyper-searchers that so many professionals can be.
The database providers will argue that a basic search output will not provide users with a complete picture. They will point out, rightfully so, that one system provides better company profiles than another. One focuses better on clinical trials, another has detailed therapeutic indication profiles. And, of course, news! Users always need updated news with links to clinical and business resources to stay fully informed on a competitor’s important drug, something a basic output cannot do.
In the end, some users will always need direct access to the databases. Some companies may find this percentage to be about 10%. Others may find it higher, say 30%—averaging out at the rest of the 80-20 ratio. There are just unique search options within each system that cannot be combined, and too many benefits accessed directly.
There is a target audience for the InfoDesk proposed product. Smart developers will figure out who that audience might be, and aim for them first, rather than trying to please everyone. I look forward to hearing about future developments, a defined timeline, and—of course!—the opportunity to beta test the first system released.