A validation interview. The team already had a seven-stage journey, four personas and twelve pain cards on the canvas from the first HSS conversation. The job on Sunday was to find out how much of it a real patient recognised.

Each row carries the stage as drawn on the canvas, with its goal, against what Patient 01 described doing at that point. The chip is the team's call on whether the stage survives contact.

The twelve pain cards on the canvas, each scored against what one patient said. Validated means he described it unprompted or agreed when asked. Contradicted means his account runs the other way, at least for a patient like him.

1. Experienced adult facing unfamiliar symptoms

   He is this persona

   Asked for his biggest remaining uncertainty after an hour of describing a confident routine, he named the unfamiliar symptom: blood in urine, black stool, something spontaneous he has never felt. That is the persona card, in his own words. The canvas has the right person; it has him worrying in the wrong stage.

2. Less experienced or adult-acquired patient

   Supported, second hand

   He could not describe a bleed sensation to someone who has never had one, and said so. New patients depend on personality: some speak up, some keep quiet to avoid trouble. The repeat questions in the group chat come from this end of the population.

3. Family caregiver under time pressure

   Strongly supported

   His description of a caregiver was a half-trained nurse: able to prepare and inject, do first aid and post-bleed care, but reading the bleed from outside a body they cannot feel, from advice held in memory with the nuance summarised away. Of the four personas, this is the one he thought had the least to work with.

4. Triage nurse or ED clinician

   Not reached

   Nothing on handover, triage calls or what a clinician would need from him. The clinician interview guide on the canvas remains the route.

The canvas asked about stages. He answered about a life. These are the things that were not on the board and now should be.

1. 1

   Recognition is learned, and it is learned by bleeding.

   He carries a profile of his own target joints, ankle, elbow, knee, and reads every sensation against it. He called the process a Markov chain: gather evidence, raise the probability of the likeliest cause, act on it. It took a childhood of episodes to build. Caregivers never build it because they are on the outside of the body.

   "It's accumulated knowledge. It's like you have to have it happen, then you know what to do."

   So what: the app's job for an adult like him is small; for a caregiver it is the whole intuition he has and they do not.

2. 2

   The clock runs in hours, and it starts on waking.

   His typical trigger is a sensation when he wakes at nine or ten. If it is a joint, by the afternoon the range of motion is gone and the question has answered itself. The longer it is left, the worse it gets, so the useful window is half a day, not a night.

   "Usually by afternoon, if you're waking up at maybe 9 or 10am, you will know it's a bleed already."

   So what: the 2am framing is right for caregivers of children; for this patient the episode is a morning one, and the re-check timer should think in hours.

3. 3

   The needle is cheap, so the only cost is delay.

   He does not weigh over-treatment. A dose given to a non-bleed just raises factor levels for a while; surgery uses far more. So the whole of his decision effort goes into shortening the time to "yes, it is a bleed", never into avoiding an unnecessary injection.

   "I wouldn't say it's a risk, because it's just increasing the levels to a degree."

   So what: copy that warns about over-treating is solving a problem he does not have. Copy that shortens recognition is the product.

4. 4

   Inject first, then call the ambulance.

   Asked whether he would inject before going to A&E, he said yes, because the ambulance takes time anyway. Asked whether that is a universal protocol, he said he did not have the answer but would think so. It is a belief, held firmly, unconfirmed by a clinician.

   "It would be best to inject first, because you have to wait for A&E anyway."

   So what: this needs a haematologist's confirmation before it appears on any A&E-lane card. If confirmed, it belongs on the first one.

5. 5

   The unfamiliar symptom is the real black box.

   Gum bleeds that drag on for days until he felt lightheaded. Black stool from swallowed blood. Blood in urine. A head where the brain has no sensation, so he would only know from vision or consciousness. These are the cases where his system stops and, by his own account, most patients and caregivers have no system at all.

   "Suddenly you got black stool, or whatever. Where can it come from? That's the worst part."

   So what: the non-joint cards, GI, renal, head, oral, are the highest-value content in the app, and the ones no generic leaflet gives a patient in the moment.

6. 6

   Help means a second pair of hands before it means a hospital.

   His low-severity call for help is physical: a bad elbow bleed leaves him with one working hand, and reconstituting factor is a two-handed job. His high-severity threshold is when he cannot move or is not fully aware. Between those, he expects to manage.

   "If it's a bad elbow bleed, then I will be stuck with just one healthy hand. That will be hard for me to prepare something."

   So what: one-handed use is a literal requirement, and "who can help you prepare it" is a real question for the setup flow.

7. 7

   The care team is an email address and a friend's nurse.

   He sees his doctor twice a year and otherwise writes email, saving every question for one message because he assumes the doctor is under pressure. The informal, back-and-forth nurse relationship he described belongs to a friend at a different hospital, not to him. The patient group chat has patients and caregivers only, and its questions repeat, mostly from parents of young children.

   "When I reach out to the doctor, I have to cover everything. You will just unload all the questions you want to ask at one call."

   So what: between reviews there is no clinical channel with a same-day answer. The repeat questions in the group chat are a ready-made content list; he offered to bring the team's questions there.

8. 8

   He already uses an AI, and already distrusts it.

   For something unfamiliar he types the symptoms into ChatGPT, withholding his identity and giving as little as he can, and treats the answer as a second opinion. He expects hallucination. He would still follow a "give a dose, wait, then go" answer as a precaution. What would make him trust a tool instead is a curated source with an authority behind it: his first answer was the government, then a named clinical institution.

   "Currently we are not recommended to take it at face value from these chatbots. They may hallucinate. Curated information, from a trusted source: these are the two values."

   So what: the source stamp on every card is the reason he would open this instead of a chatbot. Never ask for more identity than the card needs.

9. 9

   He logs injuries. He does not log knocks.

   He keeps a record of every injury, and said so with some pride. Asked whether he would log a head bump that felt fine, in case it turned into dizziness two hours later, the answer was that he is not that hardworking. The record is of confirmed events, never of suspicions.

   "I do record what injury I have. Not that hardworking, though."

   So what: the bleed log will be filled after the fact, from confirmed episodes. A "note a knock" feature will not be used unless it costs one tap and asks nothing.

10. 10

    Consolidate first. Personalise later.

    Asked what a good outcome would look like, his answer was the scattered information on bleed types and severity in one place, from which he can make his own interpretation. A triage engine did not come up. He agreed that a version tuned to his own profile would be better, and placed it as a later stage; a compact, curated reference would already be a big step.

    "A lot of information is everywhere. It's there, just that it's everywhere. Is there something that can help the patient consolidate all this important information at one place?"

    So what: version zero is a trusted, well-organised reference with the patient's two doses on top. A decision engine can wait, and a reference is also a far easier thing to get a clinician to sign.

11. 11

    Cost shaped the past, and independence shapes the present.

    As a child he was treated on demand, after a bleed rather than to prevent one, because of cost, which he knows is not the recommendation. Regular prophylaxis and less cost pressure came later. He linked adult self-management to independence as much as to experience: an adult takes charge of his own condition, and a child's parents admit him to hospital when the injection does not seem to work.

    "That is actually not the recommendation, but that is the circumstances due to the cost."

    So what: the adult and the parent are different products wearing one app. The parent's threshold for hospital is lower and their tolerance for waiting is shorter.

12. 12

    The extended half-life product cut his bleeds by more than half.

    A year on an extended half-life factor VIII product took him from 12 to 18 bleeds a year down to under six, provided he keeps to the prophylaxis schedule.

    "Provided we adhere to the prophylaxis."

    So what: the extended half-life product cut his bleeds, so frequency questions must ask about the last 12 months.

13. 13

    Adherence has a cost, and it is not oral medication.

    He does not always keep to the prophylaxis schedule, and said why. Preparation, timetable and scheduling are a real factor, both mental and physical, and sometimes he does not want to do it. Preparation and the injection together take half an hour to an hour.

    "It's not like oral medication, you don't just pop the pills... preparation, timetable, scheduling, all that is a big factor."

    So what: the needle costs half an hour to an hour, so the home card should name that cost rather than assume treating is free.

14. 14

    He named his own risk appetite, and called it generational.

    Being young and on prophylaxis makes him "somewhat of a regular person," against an older, on-demand cohort with permanent damage to their joints. He was blunt about where that confidence can lead.

    "It's like a gambling addiction. One more, one more try. But it's ultimately very detrimental to the body."

    So what: risk appetite is generational, so design for the conservative patient; the gambler still lands inside it.

15. 15

    The survey traces back to his own suggestion.

    Asked for an input sheet on the variables in his decision-making, he said not to take his own account as representative, and suggested the team circulate a ranking survey in the group chat instead. The survey the team built traces to this exchange.

    "Don't take my one as representative."

    So what: the survey came from him, so he is the first distribution channel.

16. 16

    He doubts anyone answers a questionnaire mid-bleed.

    Asked whether a patient in the middle of an acute episode would have the mind space for a short questionnaire, his answer was no, with a hint that there is a more intuitive way to ask than a list of questions. He was blunt about how content gets read at all.

    "Would you have that mind space and patience to go through a short questionnaire? ... I look at title, I never read."

    So what: no questionnaire at 2am; the first screen is a tap on a body, and titles must carry the meaning because he does not read past them.

A second validation call, this time with the clinician side of the journey: a haemophilia nurse who sees the same kind of patient Patient 01 is, from the other end of the phone.

Each row carries what the team's seven-section nurse guide asked, against what Nurse 02 said in response. The chip marks whether the section's questions were fully put to her.

The same nine points, read once through a patient without an inhibitor and once through a nurse whose cohort is severe adults on prophylaxis.

1. 1

   Inject first is now confirmed. The nurse teaches injecting before A&E for trauma and every non-joint site. Keep the haematologist question open for edge cases, but the cards can stop calling it unconfirmed.

2. 2

   Add a trauma question to the night spine. NGT-06 asks when, NGT-07 asks whether it is familiar; neither asks whether there was a fall or a knock, and a trauma fall routes to A&E after the dose even at a target joint.

3. 3

   Report every bleed. The receipt back to the nurse (OUT-09) moves from a nice loop to the nurse's own rule: make telling the haemophilia nurse on the next working day the last step on every home card.

4. 4

   Carry a days-of-treatment question, not a severity one. The day two and three check-ins (OUT-13 series) should ask how many more days of factor and tranexamic acid, answered by the nurse in office hours, using the pain score before and after the first dose.

5. 5

   No second-dose path. If the first dose brings no relief, the nurse's reading is that it is probably not a bleed, or rarely an inhibitor, and the patient should come back; OUT-13b and JNT-17 stay as written.

6. 6

   Add infection to the urine lane. URN-03 should add a GP visit to rule out infection as the next-day step; URN-04, no tranexamic acid, stands unchanged.

7. 7

   Let the bleed dose default to "same as usual". Hospital B patients carry one prophylaxis dose, so profile doses can offer a single number rather than a calculation. Correction from interview 03: Hospital B is not uniformly one number; Patient 03, an Hospital B patient, carries two. Soften to one or two numbers, set by the centre.

8. 8

   Point recovery at a physio appointment, not exercises. OUT-12 should reflect that prophylaxis resumes on its fixed days and never stops, that the gym is off for one to two weeks, and that physio intensity is a physiotherapist's call the card should prompt for, not describe.

9. 9

   Add three library entries. Gym sessions on trough days cause muscle bleeds, gum bleeds after dental scaling respond to tranexamic acid, and kidney stones are a common cause of blood in the urine at Hospital B; the nurse gives all three verbally today with nothing written down.

10. 10

    Update the persona ledger. The adult patient is now validated twice, by Patient 01 and by the nurse's cohort; caregivers remain unvalidated, since her adult patients rarely have one call for them; acquired haemophilia stays out of scope, by disease rather than by choice.

Persona ledger

1. Adult patient: validated twice, by Patient 01 and by the nurse's cohort.
2. Caregivers: still unvalidated; her adult cohort gave no caregiver signal.
3. Acquired haemophilia: stays out of scope, by disease rather than by choice.

The team debriefed for about thirty minutes straight after the call with Nurse 02, on 7 September 2026. Rachel led the analysis, Ariel facilitated and owns the problem statement and the next interview, and Myo Myo and Gerald were both on the call. The debrief assumed the Patient 01 call was not recorded; both Gerald's and Rachel's recordings exist, and Rachel's covers the lunch segment, which matters for the Roundtable video.

Thirteen calls the team made in the thirty minutes after the nurse hung up, working the cohort she described against the Patient 01 page and the canvas.

1. 1

   Stage 1 to 3 move earlier than the Patient 01 page has them. Rachel: "the moment something feels off, whether or not it is familiar or unfamiliar, it is a jab first. That stage ends quite fast. It is not a decision point." Myo Myo asked the nurse about severity and got the same answer: "the first thing you notice, you just jab." Recognition and severity collapse into one move for her cohort.

2. 2

   Patient 01's wait-and-see reads as the anomaly, not the norm. Rachel's read is higher risk tolerance, resignation, "Ah, not again," and a childhood of on-demand treatment. Ariel's counter, from the HSS report: some patients do observe first and jab only if it worsens, and the reason may be the preparation, the pain and the mental load of infusing, rather than certainty about the bleed.

3. 3

   Decision: design for the conservative patient. Rachel: "if you cover a more conservative approach, those that are more wait and see will come under it anyway." Safe to be safe.

4. 4

   The real question moves from "is this a bleed" to "how severe" and "trauma or spontaneous." Both agreed severity and the trauma fork are what decide the order of screens.

5. 5

   Risk assessment comes first, before symptoms. A starting question such as "when was your last prophylaxis dose, which day" surfaces early: schedules differ by patient, and a missed dose raises the probability of a bleed. Rachel wants these background facts up front so the flow can scale risk.

6. 6

   Logging becomes an input, not only an output. If a patient already keeps a log, Patient 03's runs in Excel, pull from it rather than ask a frightened person to remember. Inertia around moving an existing system is expected; minimal fields, dose dates, means minimal personal data.

7. 7

   Bleed reporting is a must-have. Rachel's doubt about the nurse's rule: "how many actually report?" If the patient does not call and the nurse's own check-in misses it, the bleed goes undocumented on both sides, the pattern of repeated spontaneous bleeds is never seen, and the prophylaxis dose stays wrong. A reminder to report, plus the day-after check-in already in Gerald's design, is the feature.

8. 8

   The problem statement needs a post-mortem. Ariel read the current one aloud: patients and caregivers lack an accessible 24/7 self-service reference tool for symptom triage, site-specific acute bleed management and structured post-bleed recovery outside clinic hours; home care relies on memory rather than a centralised trusted resource; families face high anxiety, deferred bleed reporting and recovery management, and delayed intervention that leads to long-term joint damage. Rachel raised two open questions on it. What does "self-service" mean, and what exactly does the patient want to do for themselves? And what is the real pain outside clinic hours: "It sounds like a lot of the calls happen during work hours, and it is fine. It is managed just fine. I am starting to wonder how big really is the pain outside of clinic hours."

9. 9

   The "digital nurse" framing does not fit what the nurse described. Patients call her after they have infused, inside the 24-hour recovery window, asking about a second dose or whether to come in, well past the acute moment. Rachel: "They are not asking her what to do. They have done it." The journey map has the nurse contact at the wrong stage.

10. 10

    Ariel's hypothesis: the problem statement may carry a hidden question. "Is there a chance I do not have to infuse first, because for this specific symptom I can afford to wait?" Nobody on the call could confirm it, and HSS has said they will not spell out expectations.

11. 11

    Interview 03 changes focus, to pain and needs. Rachel: "We have a very good overview of process already. Now we need more flavour and need around the experience of it." The team will ask about what patients are not saying they need; Ariel adds her usual product probes, the actual problem and what is currently done that falls short.

12. 12

    Survey refinement waits for interview 03. Ariel restated the survey's purpose: to validate whether "information is everywhere" is a real problem, how often bleeds happen outside clinic hours, and what people actually do.

13. 13

    Presentation is Sunday 13 September, at the Hackers Roundtable. Per the organisers' brief: one page summarising who, what, when, where and how; a walkthrough of each chosen feature linked to a specific challenge or friction; video of the interviews as a learning journey. Patient 03's call will be recorded and screenshotted; Patient 01's was not recorded, so the photo stands in.

1. Design for the conservative patient: safe to be safe, and the more wait-and-see patients come under it anyway.
2. Bleed reporting is a must-have: on the strength of the nurse's own rule, said three times on her own call.
3. Survey refinement waits: interview 03 clarifies the problem first, the survey follows after.
4. Record interview 03 and screenshot it: the Patient 01 photo stood in once because his call was not recorded; it should not have to again.

Interview 03: Patient 03, Wednesday 9 September. Focus is pain and needs over process; he is the meticulous self-logger, his log runs in Excel, so he is also the test case for logging as an input. Answered on 9 September; the answers are read row by row in interview 03 below.

From the debrief

1. 1

   Walk us through your Excel: what columns, when you fill it in, what you have ever looked back at.

2. 2

   The last time something felt off, what did you do in the first ten minutes, and why that?

3. 3

   Have you ever waited rather than infused? What were you weighing?

4. 4

   After a bleed, do you tell the nurse? Every time? What decides whether you call?

5. 5

   Outside clinic hours specifically: what is the worst part? Is it the decision, the needle, being alone with it, or the morning after?

6. 6

   When you say you want to self-serve, what would you want to do without a clinician?

7. 7

   What did your doctor tell you at the last review that you can no longer remember exactly?

8. 8

   Is there any symptom for which you would want permission to wait?

Carried over from these files

1. 1

   Stage 7 recovery: what you do the week after, whether prophylaxis resumed on schedule, whether anyone mentioned physio.

2. 2

   The 24-hour check: how you judge it worked, and what you did the one time it did not.

3. 3

   Dose: one number or two, and who set them.

4. 4

   Trust: what would make you open a card at night instead of the group chat or ChatGPT.

Rules for the room: record the call, take a screenshot, bridge new answers with "from Nurse 02 we heard X, does that match you", and do not ask what solution he wants.

A third interview, this time a video call, with an Hospital B patient on a different factor product and a different dosing rule than Patient 01, at the one centre of the three studied that has a dedicated haemophilia nurse. In the last twenty minutes, at his own request, he stopped answering the prep sheet and started describing a product.

The twelve prep questions built after the nurse call, put to Patient 03 row by row. The chip is the team's read on how fully each one landed.

The same canvas, run against a second patient with a different product and a different centre.

The same points, now read through three accounts: a patient on factor VIII at Hospital A, a nurse's cohort at Hospital B, and a second patient, on factor IX, at Hospital B itself.

Asked directly, in the last twenty minutes and at his own request, he laid out a product. Quotes below are his.

1. 1

   The problem is a single point of contact. "A lot of our queries go through one single point of contact, which is the haemophilia nurse ... only available on office hours on weekdays. Public holidays and weekends, we don't have access." Cover nurses when she is on leave "may not be as experienced". Hospital A has no nurse at all.

2. 2

   The gap is the middle ground. "Right now we are just using what works for our target joints and applying it to scenarios where we are not familiar with. The alternative is go straight to emergency department. So there is not really a middle ground that we can refer to before we go A&E or wait for the next working day."

3. 3

   The shape is a searchable site-first reference, with the don'ts. "If I know where the bleed is, I would just search up that particular site ... blood in the urine, I would just search for it. Then it comes up something like, do not use tranexamic acid, right there. That's a very clear don't." Not prompted, not diagnosed: searched. "The WFH guidelines are buried in a 100-plus-page PDF. Most of us will not be able to find what we need there."

4. 4

   It is a first gate, not a replacement. "If it's off-hours, we can just refer to it. Then, based on our own experience, we'll just manage it. And ideally the scenario should be that we should be contacting the nurses."

5. 5

   The second half of the product is recovery. "We are very focused on treating the bleed. But is the way we are managing optimal in terms of recovery? Every bleed inherently creates some damage ... if we bleed enough times, how do we get it back to baseline?" And on the joint: "when you have a bleed at a specific spot, it tends to bleed more frequently because that part is already weakened."

6. 6

   Bring the care team onto the same platform. "Whether we can get the haemophilia care team on board this solution such that they can also track the progress of the bleed and the recovery phase, instead of them needing to proactively contact us or we need to proactively tell them." Prompts: "24 hours past your infusion, report your progress, worsening, better, or whatever." And back the other way: "the team can prompt us, we think you should come down for a physiotherapy assessment. Otherwise we'll just manage it on our own." Only in office hours: "they should only be using this after they come back to work."

7. 7

   Bring the physiotherapist and the doctor in, not only the nurse. "The haemophilia care team is a multidisciplinary team ... the physiotherapist will be one that I can think of because there is a gap there. The doctors also, possibly." He will ask Nurse 02, back next week, which physiotherapist would take part. Doctors are "very hard to get hold of".

8. 8

   Outcomes, in HSS's words. "The ultimate long-term aim is to reduce the long-term joint or muscle damage." Every bleed reviewed by the care team one way or another, and recovery optimised so the same spot does not bleed again. And: "reduce the anxiety whenever you're met with a bleeding episode, especially for caregivers of the children or the youth." He warns off the obvious metric: "reducing bleeds is quite tricky ... if you are trying to use a digital tool to reduce bleeds, I cannot find a direct correlation."

9. 9

   Where the line is. He put the nurse's question back to the team: "where does the line need to be drawn ... you cannot have a fully self-help tool because each person may be different ... if you have a severe bleed, you may not be able to fully use that self-help advice." The team's answer on the call: no diagnosis, a support tool with validated content, the decision stays with the human. He agreed and offered Nurse 02 for a trial run, "because the World Federation guidelines are meant to be broad."

10. 10

    Two populations he wants remembered. Mild patients on demand, not on prophylaxis, whose recommendation differs and who, if they cannot infuse, go straight to A&E. And parents of young children: "they are the ones managing everything ... deciding, if the child reports I'm feeling weird, should I give factor or not ... from another person's perspective, they may interpret it differently." Phase-by-phase is fine by him; teens starting self-infusion can use the adult content; "most kids will grow up to become adults anyway."

1. 1

   Search is a first-class entry, not only the body map. He described typing the site and reading the don't. The library screens need a search box on the home screen and a "do not" line at the top of every site card, with URN-04's tranexamic acid rule the model. His words for the format: "simple to understand", "a very clear don't".

2. 2

   The 24-hour check-in becomes the report. OUT-13's better / same / worse is exactly the prompt he asked for. Wire it so the answer is what the nurse sees, and say so on the card: this goes to your care team when they are next in. That is the receipt Gerald wants and the platform he wants, in one screen.

3. 3

   Recovery gets a track, not a card. OUT-12 should run past day three: the physio prompt, a note that the joint that bled is now the one most likely to bleed again, and the question he could not answer alone, when and how to scale exercise down and back up. Content for that needs a physiotherapist's signature; he is sourcing one through Nurse 02.

4. 4

   A care-team view is now asked for by a patient, not only a nurse. Even a read-only log page for the nurse, office hours only, would meet his ask. Scope it as the v3 spike, not v2.

5. 5

   The night card should shrink the infusion, not argue for it. Add to NGT or the profile: a checklist for his own kit (vials, saline, butterfly, sterilising, disposal) so the reconstitution is not worked out from memory at 2am, and the honest line, in his words, that asleep and immobile he cannot react for six or seven hours. Confirm with the nurse before shipping any of that.

6. 6

   Which hospital is a card. He goes to Hospital B A&E because the nurse and the haematologist are there; an ambulance takes you to the nearest. The profile's named centre should print on the A&E card with the line "tell the ambulance crew your centre"; check with the nurse whether SCDF will honour it.

7. 7

   Mild, on-demand patients need a lane. Profile: on prophylaxis or on demand. On demand with no factor at home routes to A&E for any bleed. One sentence on SET-11's dose screen covers it.

8. 8

   The lower dose is a real option for non-joint sites. The nurse dropped him from twelve to nine for the urine. The profile's two doses, usual and upper, match his; the urine card can say "your usual dose is enough unless your nurse says otherwise", which is what she told him.

9. 9

   Cover-nurse variation is an argument for the signed card. He raised it unprompted: when the main nurse is on leave the advice can differ. A card signed by the centre is the same answer every week. Use the line in the pitch.

10. 10

    Outcome claims for the pitch, in HSS's words. Reduce long-term joint and muscle damage by getting every bleed reviewed and recovery done properly; reduce anxiety for caregivers. Do not claim fewer bleeds.

Corrections to earlier files

1. Patient 01 block, the group chat insight: Patient 03 confirms the group chat carries "my kid's having a bleed, what should I do" at night and that the answers are subjective. Kept as written.
2. Nurse block, build item 7, "same as usual" dose: Hospital B is not uniformly one number. Patient 03, an Hospital B patient, carries two. Softened to one or two numbers, set by the centre.
3. UX files, failure mode 6, the profile as the verbal consultation written down: his consultation includes three A&E triggers he cannot source. The profile can carry them as "what your team told you", to be confirmed by the nurse.
4. The Wispr manual note: "Hospital B has no haemophilia nurse" is wrong. Hospital B and Hospital C each have a dedicated haemophilia nurse; Hospital A, where Patient 01 is followed, does not.

Persona ledger

1. Adult patient: validated three times, by Patient 01, by the nurse's cohort, and by Patient 03, and now in two styles, wait and see and conservative.
2. Caregivers: still unvalidated, and Patient 03 says parents of young children are a different population with their own set of needs.
3. Mild on-demand patients: new, raised by Patient 03, unbuilt.