Table 1. Coding scheme for protocol analysis.
| Code | Definition | Examples |
|---|---|---|
| Supportive | The cue was perceived to support the diagnosis | 1. Cue is considered a feature of the diagnosis:“ • [cue] is a presenting factor for [diagnosis A]” • “you can get [cue] with [diagnosis A]” • “[cue] is consistent with [diagnosis A]” 2. Cue is perceived to increase the likelihood of the diagnosis: • “[cue] makes [diagnosis A] more likely” • “[cue] would add to the diagnosis of [diagnosis A]” • “[cue] suggests/supports/favors [diagnosis A]” 3. Cue leads participant towards the diagnosis: • “given [cue], I’m now thinking [diagnosis A]” • “[cue] sends me towards [diagnosis A]” • “[cue] makes me want to explore [diagnosis A]” • “[cue] makes me consider [diagnosis A] as a possibility” |
| Non-supportive | The cue was not perceived to support the diagnosis | 1. Cue is not considered a feature of the diagnosis: • “[cue] is not a presenting factor for [diagnosis A]” • “You don’t usually see [cue] with [diagnosis A]” • “[cue] is not consistent with [diagnosis A]” 2. Cue is not perceived to increase the likelihood of the diagnosis:* • “[cue] doesn’t help me to know whether this is [diagnosis A]” • “[cue] tells me nothing about [diagnosis A]” • “[cue] doesn’t add anything to [diagnosis A]” • “[cue] is irrelevant to/has no bearing on [diagnosis A]” • “[cue] doesn’t suggest/support/favor [diagnosis A] 3. ”Cue leads participants away from the diagnosis: • “[cue] doesn’t make me think of [diagnosis A]” • “[cue] moves me a bit away from [diagnosis A]” • “I wouldn’t be looking for [cue] with [diagnosis A]” |
| Unclear | The cue was evaluated in relation to the diagnosis, but perceived support for the diagnosis was unclear | 1. Cue’s support for the diagnosis is ambiguous: • “[cue] might be because of [diagnosis A] and it might not be because of [diagnosis A]” • “[cue] doesn’t differentiate between [diagnosis A] and [diagnosis B]” 2. Cue’s support for the diagnosis is questioned rather than stated: • “could [cue] suggest [diagnosis A]?” |
| Nothing | The cue was not evaluated in relation to the diagnosis | N/A |
Each participant was assigned two codes per cue: one in relation to the diagnosis that was leading at the time (“Verb_Lead”) and one in relation to the diagnosis that was trailing at the time (“Verb_Trail”).
* Note: many of the utterances in this subcategory suggest that a cue is irrelevant to a diagnosis: the cue is not perceived to support the diagnosis but it is not perceived to negate it either. Such utterances could be coded as “unclear”. We categorized them as such in a second coding of the data and our findings did not change. Full details are available in the Supporting Information (S3 Text: point 8).