When Dante opened his Divine Comedy with a descent through the nine circles of hell, he was not only writing about sin and punishment. He was also describing layers of difficulty, each more constraining than the last, until the traveller reached the core.

In our work on Clinical Outcome Assessments (COAs), we have stumbled into a similar metaphor. Patients, when faced with COA items, do not simply “answer questions”. They descend through successive circles of burden, each one demanding more mental effort, until the act of reporting becomes a test of endurance as much as a record of experience.

We visualise it like this: At the centre sits the words: the building blocks of what makes the item, nouns, verbs, etc. and the concepts that are contained within these words. This is our starting point. We are not yet in any of the nine circles of burden. But this starting point is what we based our Cognitive Burden Simulator (CBS) on. It was just a thought experiment that we wanted to use for a course that we have on cognitive debriefing. We wanted to see how many concepts a patient could conceivably deal with, cognitively of course, without them feeling overwhelmed. We wanted to parse individual COA items by merging Chomskyan syntactic trees (really popular in the 80s and 90s) with mindmaps to simulate the mental gymnastics that people perform to make sense of COA items. The CBS was sterile, divorced from the real world, but was useful to illustrate the kind of parsing, recalling, attributing and mapping that a hypothetical patient may do when tackling a COA item.

Of course, real-world patients are rarely confined to this clean puzzle. We realised that the CBS was just the core of the problem – the words and concepts. Around the core radiated nine additional circles each adding a new dimension of difficulty: working memory limits, visual design, cultural processing, delivery mechanisms, contextual stressors, adaptation quality, entrenched formatting conventions, participant cognitive impairment and cultural-technological erosion. There will inevitably be more outer circles but we thought of nine. And we call them The Nine Circles of Burden.

The Centre: The Words and Concepts

Before we trace the circles, we must start at the core. The Cognitive Burden Simulator represents the fundamental cognitive acts involved in reading and answering a COA item. By mapping a sentence into a syntactic tree and then overlaying a mindmap of possible decision branches, we visualised the unseen work of:

• Parsing and comprehension: breaking the question into grammatical and conceptual units.

  
  

• Recollection: searching memory for relevant experiences within the timeframe of the item.

  
  

• Attribution: deciding whether those experiences were due to the health condition or something else.

  
  

• Quantification/classification: judging frequency, intensity or impact.

  
  

• Response mapping: aligning that judgement to the categories offered.

Even at this “sterile” core, answering is far from trivial. A binary Yes/No may require four chunks of working memory: timeframe, definition, recall, mapping. A frequency scale requires more. This is where our model began.

But in practice, patients must also traverse the outer circles of burden too.

Circle 1: Working Memory Capacity

The first circle is the natural bottleneck of human cognition: working memory. We have already encountered the generous span of Millers’s 7 ± 2, Cowan’s stricter biological limit of 4 ± 1 and the real-world conditions, such as stress, fatigue, pain and anxiety, that often shrink it further to just 2-3 chunks.

But COAs often demand more than this:

• Even simple binary items hover at the edge of Cowan’s window.

  
  

• Frequency scales often exceed it.

  
  

• Multi-symptom checklists, ranking tasks, hedged questions blow past even Miller’s generous upper bound.

Patients under pressure are expected to juggle more than their cognitive systems allow.

Circle 2: Visual Processing

Even before content is interpreted, visual design imposes load.

• Readability: Long stems packed with clauses (“Over the past 7 days, including today…”) stretch comprehension.

  
  

• Legibility: Small fonts, poor contrast, or cluttered pages slow reading and increase errors.

  
  

• Layout and alignment: Shared-stem grids require exact row-by-row tracking. A misaligned tick is not a lapse of memory but of visual burden. This is compounded in the ‘one-size-fits-all’ approach to eCOA layout once translated into other languages where different visual processing strategies are typically used.

These “design details” directly affect how much cognitive effort is required.

Circle 3: Cultural and Linguistic Processing

COAs are usually anchored in English, assuming a Western, linear reading style. Patients worldwide approach them with different information processing habits and cultural lenses.

• Directionality: Left-to-right (English, French) versus right-to-left (Arabic, Hebrew) versus top-to-bottom (Japanese, traditional Chinese). A Western grid feels alien to RTL or top-to-bottom readers.

  
  

• Processing style: East Asian readers often scan holistically, considering relationships; Western readers proceed linearly.

  
  

• Vocabulary and metaphor: Words like flashback or gnawing pain may lack equivalents. Cultural stigma may make loneliness or depression harder to endorse.

A “universal” COA item is not universal at all. It is always culturally situated.

Circle 4: Delivery Mechanisms

The mode of delivery changes the experience:

• Paper COAs allow scanning across the page but risk misalignment and skipped items.

  
  

• eCOAs typically present items one by one, reducing clutter but forcing patients to hold stems in memory. Small screen can hide options, scrolling fragments context and inconsistent orientation disrupts flow.

  
  

• Interaction friction: Paper needs a tick. eCOA needs taps, swipes, drag-and-drop. Each of these are a micro-task.

The same question can feel much harder, or easier, depending on its format.

Circle 5: Contextual Stressors

Patients rarely answer COAs in calm, idealised settings. They bring their contextual burdens with them. Where are they completing the COA? After medication? In a clinic wearing a hospital gown that doesn’t quite close and feeling exposed?

Contextual stressors like anxiety and stress narrow attention. Pain and fatigue sap working memory capacity. Time pressure in clinic pushes people to satisfice (“sometimes” as a default response).

And first-time exposure to a COA adds meta-load: figuring out the “rules” of the instrument.

These conditions shrink effective working memory to 2-3 chunks or even less. Even simple items may collapse under these pressures.

Circle 6: Adaptation and Wording Quality

The sixth circle is about the instrument itself.

• Ambiguity: Terms like impact or disruption have multiple interpretations and often do not translate well.

  
  

• Overlap: Descriptor lists with near-synonyms, such as stabling versus cutting pain. These confuse and, again, may not translate well at all.

  
  

• Translation gaps: some concepts, e.g. quality of life, do not map directly.

  
  

• Poor eCOA adaptation: Stems split across screens, hidden options, inconsistent scales.

Here, the burden is not inherent to cognition or context. It is created by poor writing and poor localisation.

Circle 7: Presentation Conventions and Formatting Habits

This circle is one of the most pervasive and the most avoidable: bad-but-common presentation practices. This is similar but different to the circle of burden involving readability. It is similar in the sense of it dealing with (usually flouted) principles of Information Design and different because it is about how the information is presented. Here are some examples:

• ALL CAPITALS: The intention of this is for emphasis: “Please rate your pain over the past 7 days when it was AT ITS WORST”. It is pervasive in other patient-facing or health professional-facing documents and this is backed up by legislation in some cases. For example, the Slovak wording -  by legislation - for ‘For clinical trial use only’ on IMP labelling is “VZORKA NA KLINICKÉ SKÚŠANIE”, the Brazilian equivalent is “APENAS PARA USO EM UM ESTUDO CLÍNICO”. The rationale is to create contrast, emphasis, to draw the eye to this important notice. But, in reality, block capitals only slow reading and cannot be adapted across scripts.

  
  

• Underlining: Similar to all caps, taught in classrooms as emphasis. In digital contexts, it can be mistaken for hyperlinks and, on paper, it slows down reading.

  
  

• Bold and italics: these are excellent tools for emphasis but only when used sparingly. But overused, it loses its impact. Bold can often be misread as headings and italics are absent in many scripts.

  
  

• Worse than any of these is when all of them are brought together, which is exceedingly common in COAs: “Please rate your pain over the past 7 days when it was AT ITS WORST” We are sure that anyone even casually involved in COA development or testing would have seen this before.

  
  

• Dense blocks of text: multi-clause stems overwhelm memory. On small screen, scrolling fragments them further.

  
  

• Jargon and abstraction: Words like impairment, intermittent or even impact alienate patients.

  
  

• Overloaded scales: seven-point anchors blue distinctions: “sometimes” versus “occasionally” collapses in translation in many languages,

These conventions feel “normal” because we learn them very early on, at primary school, as the standard way of presenting, organising and highlighting text. They are reinforced by institutions and COA developers. But they are counter-intuitive: they are meant to clarify but they actually obscure.

Circle 8: Participant Cognitive Impairment

Many patients completing COAs are already living with cognitive impairment.

• Neurological conditions: Alzheimer’s, Parkinson’s, multiple sclerosis, stroke, traumatic brain injury.

  
  

• Psychiatric conditions: Depression, schizophrenia, ADHD.

  
  

• Treatment-related effects: “Chemo brain”, post-anaesthesia fog, long-COVID.

These reduce working memory, attention and executive control. A binary Yes/No item that feels trivial for a healthy person may be overwhelming for someone with mild cognitive impairment.

The paradox is that COAs are often used precisely in these populations. Yes, the very tools meant to capture their experience are least accessible to them.

Circle 9: Digital and Technological Offloading

Finally, the outermost circle reflects cultural-technological erosion of cognition.

• The Google effect: People remember where to find information, not the information itself.

  
  

• Digital amnesia: We no longer memorise phone numbers or details because devices store them.

  
  

• AI cognitive debt: Reliance on generative AI outsources not only memory but reasoning, reducing natural rehearsal and synthesis.

Patients accustomed to externalising memory may find recall-based COAs especially alien. What once felt natural (holding four chunks in mind) now feels burdensome because everyday life no longer exercises that capacity.

The Nine Circles of Burden

Bringing these together, we now have a full picture:

• Centre: the words and the concepts of the item; CBS baseline processing: parsing, recall, attribution and mapping.

  
  

• Circle 1: Working memory limits.

  
  

• Circle 2: Visual processing demands.

  
  

• Circle 3: Cultural and linguistic processing.

  
  

• Circle 4: Delivery mechanisms.

  
  

• Circle 5: Contextual stressors.

  
  

• Circle 6: Adaptation and wording quality.

  
  

• Circle 7: Presentation conventions and formatting habits.

  
  

• Circle 8: Participant cognitive impairment.

  
  

• Circle 9: Digital and technological offloading.

Each circle adds difficulty. Some burdens are natural: working memory, context, impairment. Others are design-related: adaptation, delivery mechanisms, presentation. Still others reflect cultural shift, such as digital offloading. But together, they accumulate until answering becomes a descent through cognitive and cultural obstacles.

Conclusion

The Cognitive Burden Simulator showed us the sterile puzzle at the centre: the mental gymnastics of parsing, recall and mapping. But real-world patients do not live at the centre. They traverse the Nine Circles of Burden, each layer adding weight.

The tragedy is that many of these burdens are avoidable. All caps, underlining, jargon, dense stems, overloaded scales, poor translations and thoughtless eCOA adaptation persist not because they help but because they are habits, codified by regulators and caried forward uncritically.

Like Dante’s descent, each circle reveals not only new obstacles but also the consequences of unexamined tradition. If COAs are to capture the patient  voice authentically, they must be designed not according to habit but according to cognitive accessibility, visual clarity, cultural sensitivity and respect for impairment.

Only then can patients report their experiences without descending through burden and only the can the data reflect their reality rather than their struggle with the instrument itself.

Thank you for reading,

Mark Gibson, Madrid, Spain

Nur Ferrante Morales, Ávila, Spain

September 2025