"Is this something temporary... is it specific to right now? I don't know."
| Pattern | Seen in | P1 | P2 | P3 | P4 | P5 |
|---|---|---|---|---|---|---|
| Own-account scheduling: no online option, told to call | 4 of 5 + 1 partial | Yes | Yes | Yes | Partial | Yes |
| Messaging: could not find the entry point, or could not send | 5 of 5 | Yes | Yes | Yes | Yes | Yes |
| Specialty booking not completed on the child's account | 4 of 5 + 1 partial | Yes | Partial | Yes | Yes | Yes |
| Confusion over which account or organization they were in | 4 of 5 | Not seen | Yes | Yes | Yes | Yes |
| Offered an explanation for the inconsistency | 5 different | Vague | Org. | Patient age | Visit type | Temporary |
Participants P1 to P5. "Partial" means the pattern appeared in part, or the task could be started but not completed.
What the research revealed
The starting hypothesis was that provider configuration explains what patients see. It is part of the story, but the larger problem is that the app never says why.
- No reason given. In 4 of 5 sessions, scheduling on the participant’s own account ended at a phone number. People could not tell whether the block was temporary, personal or by design.
- Different even within one provider. One participant with both accounts at the same provider found scheduling closed on their own and open on their child’s.
- Five people, five explanations. Each participant offered a different theory for the same behavior.
- Switching organizations is a task of its own. Four participants stumbled when moving between organizations.
How I ran it
- Moderated, in-person testing. Each participant did 4 tasks on their own account, then switched once to their child’s, stopping just before anything real was submitted.
- Debrief interviews. I asked for explanations before suggesting any of my own, so I could tell what participants came up with unprompted.
- Analysis. I logged moments of hesitation and confusion, then checked every quote against the full transcript.
What could be different
These are directions drawn from observed behavior, not tested solutions.
- Explain unavailability at the point of failure. Say why, say whether it is temporary, and put the next step beside it.
- Clarify messaging entry points. One clear path to the care team, and an explanation when sending is limited.
- Make referral and eligibility rules visible. Say so on the department list instead of listing it and failing.
- Show who and where you are. Display the active profile and organization, and make switching reversible.
What went well
- Taking tasks to real depth meant participants met real branches with no real-world side effects.
- One control case challenged my own hypothesis.
- An open question followed by silence produced five different theories.
What I’d do differently
- Most participants had rarely booked care for themselves. I’d also recruit people who actively use their own account.
- A blocked screen capture hid the app from the recording for about five minutes. I’d test capture before every session.
- Most sessions ran 35 to 46 minutes. I’d budget 45 from the start.