In his epic novel Ulysses, author James Joyce wrote that mistakes “are the portals of discovery”. He’s wrong. Okay, it’s slightly contentious (and perhaps slightly ambitious) to call out a literary giant, but hear me out…
In the pharmaceutical industry, creating a successful new drug takes 10 years on average and costs $2.6 billion. So, it’s vitally important that all information is created, captured and recorded correctly. But what about when it’s not?
We all know mistakes happen, and many of the mistakes that happen in the lab are preventable. The most common areas are also the easiest to eliminate: data errors. Data errors can be costly and time consuming to put right, with poor quality data, redundant data and lost data costing companies anywhere from 15%-20% of their operating budget. Sorry Mr Joyce, but data mistakes don’t sound like a portal of discovery to me, they sound like a very expensive recipe for disaster!
Here are our top tips for minimizing mistakes in scientific R&D:
Don’t put pen to paper
Use an electronic solution to reduce mistakes. These days, most people can type faster than they write, and it’s easier to read typed notes than handwriting too! Up to 52% of handwritten words are illegible, so it’s easy to see how this leads to problems with clean record keeping (the reproducibility crisis, anyone? More on that below). It’s also easier to add to typed notes and you can track additions over time.
Capture and store all data results
Capture all measurable results from research for reference. We’ve previously written about why scientists should share failures, and we maintain that recording what has and hasn’t worked is a simple but effective way to increase R&D productivity in laboratories. Failed experiments are nothing be ashamed of. In fact, whether relevant to the original hypothesis or not, colleagues might be able to use by-product data for other purposes. An electronic lab notebook (ELN) facilitates complete data capture, a process that isn’t as practical when completed using traditional note-taking methods. And, by recording all data electronically, you have a source of all information should anything ever go wrong, making it far easier to identify and remedy errors.
Simplify the data entry process
Restrict the number of options within a field to limit potential errors and help standardize information. With digital data entry, it is as simple as providing a drop-down list of fields, where only one is available for selection. This also has the advantage of standardizing formatting in your organization.
Audit data regularly
Proactively perform data audits on a regular basis to see what information needs to be fixed and which areas of your data entry processes need improvement. Integrating audits into your data management will also help to mitigate security and compliance issues.
Automate where possible
With advances in technology and software, automation has become more accessible, and more organizations are investing in automation strategies to improve their operations and services. By automating processes and streamlining data capture, organizations reduce the number one cause of error in a process – humans. A common cause of deviation stems from transcription errors – humans transcribing data from one system to another, typically via a paper-based system.
Use the cloud for collaborative projects
Collaboration projects can cause all kinds of logistical problems – with various parties needing to co-ordinate activities and transfer data and reports. There’s also likely to be someone within your organization tasked with processing these documents and ensuring they are distributed to each relevant corporate repository. This is a time-consuming process and can introduce more opportunities for data loss and corruption. By using a cloud-based solution, you can ensure everyone is using the same file types and remove any manual intervention (like copy and pasting information out of spreadsheets).
 PhRMA (2015) http://www.phrma.org/sites/default/files/pdf/rd_brochure_022307.pdf
 Dilworth IP (2015) https://www.dilworthip.com/2-6-billion-per-drug-patent-portfolio-strong-enough-protect/
 USGS https://www2.usgs.gov/datamanagement/why-dm/value.php