Data Collection Nightmares in Clinical and Nutritional Research: The Researcher's Guide

Nightmares in Data Collection

‍

Without even trying, I'm sure you can think of lots of reasons to collect data from people. If you’re a researcher, you might have a specific research question that needs addressing in a particular population or cohort. If you’re a marketer you might want consumer insights into your product's suitability in the marketplace. If you’re a student, you might be collecting some survey or questionnaire data to perform some qualitative or quantitative analysis. Whatever our reasons, we must be confident that the data we collect is accurate, timely, trustworthy and stored safely at all times.

Speaking to researchers in academia and industry from 43 different countries at a recent research conference, the trials and tribulations associated with data collection - past and present - were communicated very passionately. In fact, many researchers indicated that they were either not able to conduct their research to the best standards, or that their setup was too improper to allow them to do this important work in the first place.

A Brief History of Data Collection

‍

Scroll and Quill

‍

Old reliable - the paper and pen - has certainly had its glory moments and continues to add value to research everywhere. Mass-producing paper printouts and distributing these to research participants ensures top deliverability. The high literacy of most participants also assures that people from all walks of life can participate in human research, facilitating unbiased data collection.

Filling in data is one thing, but problems begin to arise with compiling the data into usable formats. Retrieving and digitising paper forms come with many challenges.

‍

Are records fully legible?

• Is it a tick? a cross? a number "1" or a letter "I"? Is that a coffee stain smear on the paper records?

‍

‍

Are the data entries validated?

• I recently heard of a story where a participant had filled out a personal address in their "Sex" field. Why is there open text in a number-only field?

‍

‍

Did any crucial paper work get lost?

• Many researchers have reported loss of documents in the postal system, or, failures of physical documents to be retrieved and delivered to the correct hands.

‍

The $50,000 - $100,000/year Leak

‍

We've covered just some of the difficulties when compiling physical data records for analysis and, for these reasons, it makes sense to have a team dedicated to data auditors. These persons are charged with the laborious task of checking all data and responses from all participants in the study. Auditors conducting this important work particularly throughout trials are very common, but last checks on study close are also useful. These checks on their own can reveal many human errors that have already occurred during data collection from participants or study team members collecting on their behalf, in addition to many new errors from the data auditors themselves as they digitise paper records to the best of their abilities.

‍

While this is often accepted as an invisible cost, some researchers suggested that this data-checking and auditing paper > digital logs might cost them, in human hours, between $50,000 - $100,000 per year - truly staggering figures for work that can be automated.

‍

Digital and Hybridised Approaches

‍

Moving forward to digital data collection at source (eSource) and many of these problems have faded. Enter electronic data capture (EDC). With digital forms or data entry on laptops and smartphones, characters are characters - there is almost always clarity on what data was entered, even with a different font family. If the user is competent in using the device or platform, there shouldn't be any problems.

Adding to this the ability to set required fields, field limitations and field validation (e.g. dropdown option limitations, integer-only fields, ranges of acceptable values and more), EDC at source can be incredibly powerful. The true drawbacks, however, lie in the platform or technology of choice. When the platform is limited in some way, does not incite enough confidence or isn't up to regulatory compliance standards, users sometimes turn to hybridized approaches which can muddy the data collection waters.

A good example of this is the still common practice for many sites to undertake paper and pen consent yet collect other data electronically. Other common cases include side-by-side data capture on paper and digitally 'just in case' technology fails.

‍

All-In-One Solutions

‍

Now, there are increasingly popular all-in-one solutions offering a plethora of services to end users - whether they are participants, research team members, sponsors and more.

These products or services might offer digital eConsent, electronic patient-reported outcomes (ePRO), EDC and even randomisation, recruitment and participant retention programmes. But with more functionalities can come more complex workflows. A good platform of choice should:

‍

• Be user-friendly and enjoyable for all users
• Secure and compliant with any necessary data collection regulations (patient data, personally identifiable data, adhering to marketing practices)
• Save time and effort
• Reduce redundancy
• Affordable enough at-scale

Creating a solution for your data collection can be difficult. There are myriad of methods one can use. Whether it's a bespoke software solution, a hybrid of paper, PDFs, web portals, emails and Microsoft software services, it's hard to know where to begin. We recently spoke to researchers from academia and industry from 43 countries and quizzed them about their data collection. Key themes that were communicated regularly about in conversations were:

• What they used/were using - and why
• Why they didn't/couldn't use certain methods or tools
• What features and tools they liked
• What they thought hindered them in the industry as a whole
• What is needed for greater efficiency
• What is needed for best practices and regulatory compliance

From our many conversations, these recurring topics emerged consistently and researchers were stuck on some pretty major pain points. From PhD candidates and PIs to sponsors and sites, here are the top 10 most communicated data collection topics from academia and industry in the Americas, Europe, Asia, Africa and Australia.

‍

Top 10 pain points in data collection

‍

1) Not another "Portal"...(!)

2) Nothing talks to anything else

3) Platforms we have tried are too difficult to use

4) Platforms we use have no support or training and no "humans" to talk to

5) IT teams, self-hosting and support are required for our platforms of choice

6) Data encryption and security concerns

7) Regulatory compliance - what is up to scratch?

8) Nothing does everything that we want

9) Software can be expensive and there is no budget for this

10) What we use is unreliable or untrustworthy

Conclusions?

‍

It's clear that data collection is a complex field and many different research programmes or preferred styles of working exist. These might be specific to country, culture or industry type, however, one thing was abundantly clear from researchers - regardless of their background: Things are ever-changing and can always be improved....

‍

In a future blog post, we will discuss these points in depth and work through some potential remedies to these pain points.