Are Legacy COA Instruments Still Fit for Purpose?

The following insights are synthesized from TransPerfect LifeSci Talks, hosted by Mark P. Wade, Practice Leader COA SME at TransPerfect Life Sciences, with guest Kathrin Fischer, PhD, Patient-Centered Outcome Strategy Lead at Boehringer Ingelheim and board member of ISOQOL.

Clinical outcome assessments (COAs) are fast becoming the backbone of patient-centered drug development. Many of the instruments we rely on today were built decades ago; however, much has changed since then. In a recent episode of TransPerfect LifeSci Talks, host Mark P. Wade sat down with Kathrin Fischer, PhD, to ask whether those established measures still do the job they were originally intended to do.

Familiar Doesn't Mean Fit for Purpose

Fischer's starting point was that teams often use legacy instruments simply because everyone knows them, and more often than not, sponsors assume suitability without checking. Her short answer to whether they are fit for purpose today is "not always."

The older instruments came from a different mindset. They were built on classical test theory, with the aim of capturing the whole quality-of-life experience in a single tool. That produced long, generic instruments meant to work for everyone, an approach that doesn't match patients' experiences in many populations. Moreover, legacy instruments overlooked some key cultural differences. For example, when measuring dexterity, some measures refer to using a knife and fork with both hands, a Western concept. However, in some cultures, utensils such as chopsticks are used with one hand, representing a very different concept to measure. Fischer was clear that these measures still have value in some contexts. The key is a clear rationale for why you’re using a given instrument and what you plan to do with the results, whether in a clinical trial or in clinical practice.

Concepts That Have Aged

Both speakers noted that some legacy items no longer resonate. Fischer said she can answer certain physical functioning items, but they don't reflect her own experience, which raises a content validity question. Wade added that daily life has changed with increasing time spent in front of screens. Many instruments were created before screens dominated daily life, and this lived reality should be reflected.

There’s also a cultural dimension. Many measures were developed in Western populations, yet Wade pointed out that roughly 70% of studies now run outside the United States and beyond Western Europe. Items that translate well linguistically may not work conceptually in other countries and cultures.

Computer Adaptive Testing: Opportunity or "Frankenstein"?

Wade asked whether assembling instruments from different items, which some critics call "Frankenstein" instruments, is cause for concern. Rather than concern, however, Fischer sees these tools as a big opportunity and prefers to call them "dynamic" rather than static. What does this mean exactly?

A traditional instrument asks a series of items, sometimes as many as 40, and a respondent generally has to answer all of them for a score to be calculated. In computer adaptive testing (CAT), an algorithm selects the items most informative for that particular patient. Fischer's example illustrates that asking someone visibly wearing glasses whether they wear glasses adds nothing. Because items are drawn from a large bank representing the full construct and scored on a common underlying scale, results remain comparable even though patients see different questions. Items can also shift over time as a patient's condition changes, for example, when someone responds well to treatment and becomes more active.

So why isn't CAT more widespread? Fischer identified several hurdles:

• Conceptual worry. Some fear the subconcepts being measured shift from patient to patient, which risks comparing "green apples with red apples."
• Sponsor caution. Although the FDA's PFDD Guidance 3 now discusses CAT, Fischer hasn't yet seen a sponsor readily adopt it.
• Practical barriers. Setup is complex, and the eCOA technology has to support it.
• Communication. It is unclear how other stakeholders will respond, and teams worry about explaining the data to clinicians who are used to static instruments.

Wade acknowledged the sponsor’s perspective: a familiar instrument that has supported approvals before feels safe, and a novel approach adds perceived regulatory risk. Fischer countered that sticking with old measures carries risk, too. If patients' lived experiences have shifted, an older instrument may eventually stop capturing meaningful change.

Patient Burden and the Battery Problem

The conversation then turned to what happens when trial protocols pile on instruments. Wade, drawing on his eCOA background, listed the consequences:

• Modality matters. From self-reporting handheld devices to ClinROs on tablets and ProxyROs for patients who rely on outside assistance, all these modalities present their own considerations and must be planned well in advance.
• Patients can disengage, producing poor-quality data (good data in, good data out; bad data in, bad data out).
• Overlapping instruments ask near-identical questions, which frustrates participants. Using batch design and staggering completion dates can alleviate some of the burden.
• Recall problems and sheer volume add burden. See Parking Lot Syndrome.

Wade went as far as saying this may undercut some of the benefits that originally motivated the move to electronic collection.

Fischer agreed the problem is real and described what she sees in practice. Static instruments force you to administer a whole tool even when only one domain is relevant to patients. Stakeholders often have favorite instruments, which leads to a "potpourri" of patient-reported outcomes (PROs) in one trial. Data quality checks raise hard questions, such as how long a patient must view an item before the response can be trusted. Timestamps sometimes suggest an item could not possibly have been read.

She also added a counterpoint. The number of instruments is not the same as the number of items. A domain-based approach, which selects targeted measures for exactly the concepts of interest, can yield five instruments with fewer total items than one broad, multidimensional measure. And questions that feel relevant to patients are far less burdensome than those that feel disconnected from their experience. There are also legitimate reasons to include overlapping measures, such as testing which instrument performs best, as long as patients are aware of the purpose.

Working with Patients as Partners

On patient involvement, both speakers welcomed the growing role of advocacy groups. Fischer reframed the idea of "educating" patients, noting that patients educate the industry through their experience, and she would like to see more genuine collaboration. She also urged mindfulness, since many patients already carry a heavy load. Explaining practical constraints, such as why copyrighted instruments can't simply have words changed, can help patients understand the process.

The Magic Wand

As Wade normally does, he ended the podcast by asking, “If you had a magic wand what’s the one thing you would change while using a measure in a clinical trial?” Fischer chose advancing computer adaptive testing, with the important caveat that the concerns raised by skeptics are valid and need to be addressed. That means bringing patients and all other stakeholders to the same table and building greater collaboration.

She also pointed to ISOQOL as a forum for that kind of work. Its 33rd Annual Conference takes place in Prague, October 18–21, 2026, and brings researchers, patients, industry, and regulators together around the science.

Key Takeaways

To summarize some of the main points outlined above:

1. Don't assume a familiar instrument is appropriate. Check that it fits the concept, context of use, and population.
2. Content validity and cultural relevance deserve ongoing attention, especially in global trials.
3. CAT offers precision and lower burden, but adoption depends on building stakeholder understanding and trust.
4. Think about total patient burden by considering items, overlap, and relevance rather than just counting instruments.
5. Involve patients early and treat them as partners.

Want to hear the full conversation? Watch the complete episode of TransPerfect LifeSci Talks with Kathrin Fischer, PhD, and Mark P. Wade here. Fischer and Wade go deeper on computer adaptive testing, patient burden, and what it takes to bring patients in as true partners.