Ending the Paper Chase

July 27, 2005
Pioneers are taking e-notebooks beyond the R&D lab.
Electronic notebooks (ELNs) have become a fixture of drug discovery laboratories, improving search capabilities, simplifying report generation, enhancing intellectual property protection and facilitating collaboration. Until recently, fundamental workflow and regulatory differences between R&D and manufacturing prevented many drug makers from using ELNs in post-discovery operations.Responding to these different needs, ELN technology is evolving along two parallel paths: one optimized for discovery, the other for good laboratory and manufacturing practices (GLP and GMP) environments. Will these paths ever converge? Probably not, or at least, not for a while, experts agree (see "One Size Fits All?" below).

Compliance, operating cost reductions and the need for “right the first time” documentation will drive the use of electronic notebooks in manufacturing — in addition to keeping your staff from being suffocated beneath mountains of paper.

Nevertheless, pioneers have taken paperless compliance into more quality assurance and control laboratories, and a few are using it on the manufacturing plant floor. For example, Novartis has equipped operators and quality staff at its Suffern, N.Y. facility with ELNs for the past year and a half.What exactly is an ELN? “The term means different things to different people,” says consultant Guy Talbot, based in Cedar Grove, N.J., who had spearheaded Purdue Pharma’s implementation of ELN technology in R&D and quality applications (see "Purdue Pharma Goes Paperless," Pharmaceutical Manufacturing, June 2004). Talbot defines ELN as a software application that accepts raw data in much the same way as a LIMS (laboratory information management system), but which also allows contextual annotation and data mining, while facilitating collaboration.ELNS are integrating with laboratory information systems (LIMS) and other key IT applications. “The heart of ELNs is integration,” says Michael Elliott, president of Atrium Research, Inc. (Wilton, Conn.). “Any vendor that doesn’t offer that is not going to get very far.”Most ELNs operate on open XML platforms and interface easily with document management systems (DMS), corrective and preventive action (CAPA), manufacturing execution systems (MES), enterprise resource planning (ERP) platforms and other software.After a lab function has been completed and signed off on, for example, an ELN can automatically transmit data to an MES and become a means of controlling operations. A document management system can be linked with a notebook to ensure that lab analysts are using the proper procedures, then, once the lab has results, data can be sent back to the DMS, Elliott says.The newer ELNs, on both the R&D and manufacturing sides, offer improved capabilities. At this point, Elliott says, ELNs are forcing more manufacturers to rethink their entire IT infrastructure. “Large pharmaceutical companies tell us that, as ELNs mature, they’re covering requirements that LIMS used to handle,” he says.In research, the notebook serves as the nexus for laboratory work and everything associated with it: files, structures, spectra, printouts, raw data, structure-activity relationships, as well as notes. Where research notebooks needn’t tie in instrumentation as intimately, GLP and GMP notebooks must interface with instrumentation, sensors and machinery, and fit within the validation environment for those devices. In an analytical laboratory, the method includes instrument calibration and validation data, which may be viewed along with the result from any test.As soon as you go into manufacturing, “everything changes with respect to documentation,” notes Ken Rapp, CEO of VelQuest (Hopkinton, Mass.). VelQuest’s SmartLab product is primarily used in QA/QC labs, which must meet both good laboratory and good manufacturing practices. “In GMP, the integration platform, by law, is the method,” he says.“The method and instrument data are inseparable,” says Rapp. “The instrument has to be integrated directly into the method.” VelQuest’s ELN, for example, contains a library of hundreds of commonly used analytical instruments, and utilizes these as a Microsoft Windows computer would use printer drivers.Variability is simply not an option. GMP users are looking not just for a data repository, but “metadata” — all the information about a data point that becomes relevant in a highly regulated environment. And more companies are approaching R&D ELN projects this way, with a view to making data accessible throughout their organizations.Right First Time and ROICompliance, operating cost reductions and the need for “right the first time” documentation will drive ELN use in manufacturing. These goals have already been met or exceeded by ELNs deployed in QC analytical laboratories.At Forest Labs in Dublin, Ireland, for example, quality manager Dennis McDaid says that in the year since Forest’s quality department has implemented ELNs, the technology has improved quality and compliance, as well as staff morale, and allowed management to monitor and trend data continuously. Forest expects the technology to pay for itself in another 1.5 to 2 years, based on a 30% reduction in product testing time, and 50% reduction in resources.Those same benefits could make ELNs extremely attractive for manufacturing, especially considering the amount of time and resources spent on drug batch documentation today. ELNs would minimize the time that operators spend functioning as “scribes” and allow them to focus on production.For manufacturing, the move from paper to electronic compliance may be as inevitable as banking’s move from the old paper “passbooks” of yesteryear, to ATM software. Rapp believes that ATM software, which allows updating bank records securely from anywhere in the world, represents the kind of function and functionality that ELNs will ultimately provide pharmaceutical manufacturing. VelQuest is now at work on a new manufacturing ELN technology, Smart Batch, which is being evaluated by two drug manufacturers.Who’ll be first?Ultimately, adoption of ELNs deeper into manufacturing requires that a few companies lead the way.AstraZeneca’s Westborough, Mass. manufacturing site has assumed the role of test facility for the company’s electronic data systems evaluations in analytical labs. Recently-installed products include a LIMS, a chromatography data system (CDS), and an ELN from VelQuest. The key drivers for deploying the VelQuest system were greater efficiency and error reduction. “In a quality laboratory you want to make sure you do everything perfectly,” says Mike Stroz, director of quality assurance laboratory automation.The company’s adoption of an ELN was part of a much larger IT deployment which was occurring more or less simultaneously. Very early on, AstraZeneca recognized that getting an ELN, CDS and LIMS to work together, and validating operability and interoperability, would not be easy.After identifying his group’s data management needs, Stroz and his team set out to find vendors who could provide applications that worked “out of the box,” and which could integrate with the other pieces of AstraZeneca’s analytical IT program. One of the applications was VelQuest’s SmartLab, which is targeted to analytical laboratories.“We knew what data we wanted to manage, how we wanted it to flow, and through which applications,” says Stroz. Avoiding customization was key to making everything work together. “We didn’t try to make our LIMS system do what it wasn’t meant to do. We let it manage our stability program but didn’t use it in metrology or training systems. The VelQuest product handled our lab bench data right out of the box, eliminating paperwork and worksheets previously done by hand.”To get the ball rolling, Stroz picked the three largest products, which make up 65% of Westborough’s production volume, as the test processes. The largest product served as the pilot.Pre-ELN, in 2002, Stroz’s group logged about 20 errors per month in its analytical labs. Most were of the type that reviewers would immediately notice, like mistakes in calculation, transcription or data entry. Still, correcting each error took time. By simply managing workflow better, Stroz reduced errors to 10 or12 per month, or about one per analyst. The ELN reduced errors to zero.“All those little errors that take five to twenty minutes to fix have been eliminated,” says Stroz. “Plus we don’t have to worry about calculations, signature or pagination.”When you’re done, you’re doneAnother benefit of the ELN was to reduce analysis turnaround (cycle) times of 10% to15%. Stroz identified two components of cycle time: analyst time and review time. Routine analyses, like pH measurements, take as long with the ELN as with paper, but complex methods are faster due to automation. The greatest benefits come during data review, which pre-ELN involved reviewers poring over pages of chromatography packets, worksheets and notebooks, checking and recalculating along the way.Now, when an analysis is completed and signed, everything is perfect. Reviewers can spend more time analyzing top-level issues, like whether how chromatography is integrated and data compiled. “It frees reviewers to look at our data as a whole more critically, in terms of trending and analyzing what we’re doing,” Stroz comments. “And, as a result of being able to trend data, we can address possible changes before they become problem.”AstraZeneca’s greatest ELN success came with an assay content uniformity method, which previously required analysts to spend about two hours per assay on calculations and notebook preparation. Now calculations are performed instantaneously, as data is entered, saving 10-12 hours per week. “Notebook time for this assay has been reduced to zero,” says Stroz. “That’s our absolute best case.”Are AstraZeneca’s production suites ready to enjoy similar efficiencies? Not quite, but the company is working on it. “Our next step is to see how ELNs might cross over from the lab into a manufacturing environment,” Stroz comments.Stroz recounts a comment made by one of his analytical chemists shortly after the ELN was installed: “When you’re done, you’re done.” When Stroz asked what that meant, the analyst explained that for the first time, as soon as the method was completed and the data entered, nothing related to record-keeping needed to be done. “That’s when it struck home that we really had something here,” he says.Lab to plant?Opinions differ as to whether one notebook product can handle both research and manufacturing, or intermediate activities such as scaleup and process development. Although a crossover product seems difficult to conceptualize, one gets the impression that acceptance of all-purpose products will be more a matter of will than of technology (see "Planning for Paperless," below).The hard-core segregators say such products are impossible: “The only thing research and manufacturing ELNs share is security and the ELN name,” says Michael Elliott of Atrium. “In terms of functionality, they’re completely different.”Atrium has published two reports in the last six months on the ELN market. The organization estimates, through its survey of 170 pharm and biotech decisionmakers, that most ELN deployments take place in drug discovery, followed by ADME/Tox labs. Approximately 25% of new installations occur in manufacturing, mostly in QA/QC.The fence-sitters say convergence between lab and plant is feasible, but difficult. “It’s possible to customize anything to do anything on an IT platform,” says Ken Rapp of VelQuest. “If you push it, sure you can use a GMP ELN for research, and perhaps the other way around, but it will take time and money to do it right.” VelQuest is interested in pursuing an all-purpose ELN, but for now prefers to keep the two functions separate. “We will not oversell a product designed for one activity to the other,” says Rapp.Bristol-Myers Squibb’s New Brunswick, N.J. development facility has been evaluating ELNs in support of Phase I to Phase IV development work, especially for analytical and process optimization. The facility’s interest in ELNs arises in part from the proliferation of automated equipment and instrumentation, and the difficulty in recording their outputs into paper notebooks.“Many of our tests and experiments involve electronic, automatic systems controlled by PCs and data systems, but the paper notebook hasn’t changed in 100 years,” says Steve Conder, Ph.D., director of pharmaceutical development informatics. “However, most of the data in analytical laboratory notebooks is not handwritten, but computer printouts.”To Conder, a major deficiency of today’s ELN marketplace is the lack of a single application that is everything to everyone. “No vendor yet has produced the application that everyone wants, or would use,” he says. So companies like BMS are likely to source data capture and management systems piecemeal, through targeted applications. BMS uses VelQuest’s SmartLab for analytical applications and a Symyx IntelliChem package for process chemistry and recipe development.“We still have a way to go to get to the equivalent of a blank notebook page that allows a scientist, in a very extemporaneous manner, to design an experiment and capture data,” Conder notes. He does not believe that ideal is achievable today because existing technology cannot support so many diverse features under one application. “It’s much better to market a product that satisfies at least part of the market, and does it well, and which users can integrate with other systems.”Right now the frontier of ELN work is the analytical lab, including QA/QC activities, where day-to-day operations are well defined, and control over data and operations most critical. Beyond that lies the vast wilderness of the “blank page” — development territory — where few vendors and even fewer users dare tread. Conder flatly states that this is where the future of ELNs lies. If vendors can develop a product that accommodates both strict adherence to methods and blank-page creativity, ELNs for manufacturing cannot be far behind.“Productivity improvements have been expected all along from ELNs, based on the enormous quantity of mechanical work that needs to be done to satisfy paper notebooks,” Conder states. “The freedom people will enjoy, with the extra time, to think about what they’re doing and the science behind it will spur real changes in the fundamental way we work.”Not just replacing paper“Replacing paper notebooks with a computer screen doesn’t solve all the problems of managing information,” notes Jeffrey Spitzner, Ph.D., chief science officer at Rescentris (Columbus, Ohio), a provider of research ELNs. Software usually trumps paper by virtue of reduced record-keeping and related efficiencies, but an application people live with every day must be more than that. “Software should also facilitate the day-to-day things people do in the lab, and help you accomplish those activities both individually and as groups” by decreasing the burden of recordkeeping, he adds.Something like Rescentris’s flagship product, CERF (Collaborative Electronic Research Framework), might be one approach to a cross-discipline, cross-functional ELN. CERF is based on ontologies, which are defined as a set of ideas, objects, interrelations and constraints that define some domain of interest. According to Spitzner, ontologies help CERF understand data, workflows, experiments and protocols in context, all the while accommodating tasks that are common and individualized without requiring new coding.Although he believes that CERF could work quite well in a manufacturing environment, the idea is probably a hard sell. Spitzner recalls a recent meeting with a manufacturer of chemical, cosmetic and pharmaceutical products. “They have 37,000 employees worldwide, and perhaps ten of them are researchers. Even though there was very high interest, the manufacturing people nixed the purchase because they are so risk- and change-averse.”


ONE SIZE FITS ALL?

Is a one-size-fits-all ELN possible? Can one notebook provide the structure needed for manufacturing and the creativity demanded in R&D? Most everyone says no, that overcoming the structure vs. creativity barrier is just too difficult.

EKM (Solana Beach, Calif.) has a broader vision of ELNs than most vendors or end users. Its LABTrack ELN, says scientific division CEO Richard Stember, replaces paper records of every type, from research through QA/QC and manufacturing, wherever it is used. LABTrack was the result of EKM’s 2004 acquisition of Avatar Consulting, which introduced the application in 1998. EKM’s scientific division, which sells LIMS systems and ELNs to life science and chemical companies, was the result of the merger.

LABTrack overcomes the seemingly diverse form and format requirements of research and manufacturing by behaving like a word processor that interfaces to a relational database of the customer’s choice. Researchers who fire up LABTrack see a blank page, just as with paper. Creating a method template for an analytical lab entails inserting pre-written templates into LABTrack and filling in the form as usual. Most paper-based QA/QC labs already rely on word processor-generated templates, Stember says.

LABTrack template pages are stored and managed in HTML 4, which combines text and form entry elements such as drop down lists, check boxes and radio buttons. Rules can be introduced that require data entry in a specific order, within a certain range, with or without over-ride capability. Data of any type can be entered, including photographs, sounds and spectra. And it’s all 21 CFR Part 11 compliant.

“It’s about as close as software can get to a real paper notebook and still maintain very powerful data management functions,” says Stember. “The usability factor has been a tremendous asset in LABTrack’s success.”

What about the naysayers? “They’re probably not familiar with what we’re doing,” he says.

Most ELNs tend to limit their usefulness based on scientific discipline, or by where in the company they are used. Other products are little more than fleshed-out LIMS, with little or no flexibility, and are therefore not true notebooks.

Stember sees the commingling of LIMS and ELNs as a barrier to full functionality, and perhaps to regulatory compliance as well. “LIMS are suited to a production environment, but they are not a replacement for a paper lab notebook," he says. "The number of companies that stopped using paper as a result of purchasing LIMS is zero or close to it. The reason is that these products, by their very nature, have not addressed work flexibility and interface issues. Electronic lab notebooks are supposed to capture more data, a more complete data flow, and make it available throughout an organization. Anecdotal data never finds its way into a LIMS report. Only something that resembles a real notebook can do that.”

PLANNING FOR PAPERLESS

ELN adoption depends on the corporate environment and culture. In manufacturing, ELNs compete at some level with LIMS, quality applications, ERPs, process software and even process analytics. “Where any number of technologies can serve in place of electronic notebooks, confusion stymies adoption,” notes Michael Elliott of Atrium Research. Confusion over electronic signature requirements has also slowed adoption. “No one is sure if Part 11 requirements will be relaxed or tightened,” he says.

From his bird’s-eye view of ELNs in pharmaceutical manufacturing, Elliott offers the following recommendations for would-be deployers:
  1. Have a clear understanding of business requirements five years in the future. What processes are currently in place? What do you want them to look like down the road? “Only then should you be looking at specific technologies,” Elliott says.

  2. Don’t try to automate existing operations that may be too inefficient or broken to benefit from electronic documentation. “Just replacing paper won’t do it,” he says.

  3. Implement based on your organization’s culture. Are workers and management primed for the change?

  4. Institute performance metrics that will enable assessment of the project at every stage, as it unfolds.

  5. Approach electronic data systems through a stepwise approach. Don’t try to solve every problem right off the bat. Pick the system or process that is most inefficient. Fixing it will not only eliminate a major headache, but will encourage the organization to continue pursuing electronic data systems for remaining operations.

  6. Don’t expect to find a homogeneous computing environment. The equivalent of desktop application suites probably doesn’t exist in an ELN. Better to solve 90% of data capture needs perfectly, through three or four packages, than to attempt to solve all and find that nothing works quite right, or the different modules don’t communicate that well.

  7. Don’t expect convergence between research and manufacturing ELNs any time soon. Convergence will come slowly but surely.

Atrium Research recently conducted surveys of IT users about their interest in ELNs and related information systems. While ELNs are clearly in the early adoption stage, their potential is great. In many manufacturing operations, ELNs will eventually supplant LIMS systems, says Atrium's Michael Elliott. For more information, visit www.atriumresearch.com.

About the Author

Angelo De Palma | Ph.D.