HackitRx 2026 primary research
These files read three interviews against the team's own journey canvas, kept with the project research notes; the product it feeds is in the solution files.
Three interviews read against the team's seven-stage journey canvas. Patient 01 confirmed the journey's shape but moved where the uncertainty sits. Nurse 02 brought a stricter rule: site and trauma first, one dose for everything, report every bleed. Patient 03 answered the resulting prep sheet, then described the product he wants: a searchable site-first reference, a 24-hour check-in that doubles as the report, and a recovery track that ends at the physiotherapist. Interviewees are anonymised as Patient 01, Nurse 02 and Patient 03; hospitals as Hospital A, B and C.
A validation interview against the canvas's seven-stage journey, four personas and twelve pain cards, to find out how much a real patient recognised.
Adult with haemophilia A, extended half-life factor VIII, no inhibitor, reviewed twice a year at Hospital A. Treats himself at home from his own stock; keeps a written injury record.
Three team members walked him through the canvas stage by stage. Recorded and transcribed; quotes are his words, lightly trimmed.
One experienced adult, no inhibitor, no caregiver present. Recording split across two team members without full speaker separation for the second half; the clinician lane was not reached.
The canvas centres the decision at Stage 4: treat, monitor or go. Patient 01 has no decision there; once confident it is a bleed, he injects, even when A&E is next. His uncertainty lives in Stages 1 to 3, in "is this a bleed at all", and spikes on an unfamiliar symptom. Dose is not a calculation and over-treating is not a fear. What he wants is one trusted place consolidating bleed types and severity, personalisation later.
6 / 7
Stages hold
Stages 4 and 5 shift; Stage 7 holds thinly.
7 / 12
Pain cards validated
2 partial, 2 contradicted, 1 untested.
3 / 4
Personas supported
Clinician lane not reached.
0
Stage zero
Prophylaxis adherence stays as context, not a stage.
Each row is the canvas stage against what Patient 01 described doing there.
Stage 1
Canvas goal: determine if it is a bleed
HoldsSame question, with hope written in. Most triggers are a sensation on waking, then the rest of the morning to find out.
"Could I get away with it? Like, is this sensation not the bleed that I think of?"On the first question he asks himself
Stage 2
Canvas goal: determine if immediate escalation is required
HoldsTraces back what he did and which target joint. Probabilistic, mental, not information-seeking; he waits for external signs.
"You can't really verify until you really notice the external symptoms, the range of motion. For an ankle or knee you might see some swelling."On how a suspected bleed becomes a confirmed one
Stage 3
Canvas goal: determine bleeding severity
Holds, simplerTwo tiers: a bruise waits, anything internal or mobility-reducing does not. A short site list escalates on its own: neck, head, brain, organs, hip.
"If it's a bruise, then I just do a normal dosage. Anything that impacts your mobility, I will give the higher recommended dosage by the doctor."On the only severity split he uses
Stage 4
Canvas goal: self-treat, monitor, or head to A&E
ShiftsNo fork once confident: inject immediately. Monitoring happens earlier, in Stages 1 to 2. A&E does not replace the injection; he'd inject first since the ambulance takes time anyway, though this is his belief, not a confirmed protocol.
"The general idea is, if there's a bleed, then you should inject immediately."On what happens once he is sure
Stage 5
Canvas goal: determine treatment plan
ShiftsNo calculation. Two doctor-set numbers, prophylaxis and bleed dose; he picks one. Personalisation, for him, is already done.
"The doctor, during the consultation, they have already done the calculation for you."On why there is nothing to work out at the needle
Stage 6
Canvas goal: determine if the symptom is improving or requires escalation
HoldsHis real black box: severity is only known once the dose has worked or not, which is also how an inhibitor shows itself. Improvement is felt in one to two days; a verbal template has him redose at 24 hours regardless.
"The black box would be we don't know how severe the injury is. We can only perform an injection and then we have to wait and see."On the variable he has least control over
Stage 7
Canvas goal: determine the appropriate recovery plan
Holds, thinlyRecovery means physio and avoiding activities, but most people never go; it's discretionary, one session then memory. He spares the area about a week; prophylaxis does not change. No close-off protocol.
"I think most people actually don't go through that."On post-bleed physiotherapy
The twelve canvas pain cards, scored against what he said.
Gum bleeds that may be nothing or run for days into anaemia; internal bleeds that only show as lost range of motion hours later.
His second opinion is ChatGPT, which he expects to hallucinate. He wants the same thing from a trusted source.
Black stool, blood in urine, a head bleed with no sensation to feel: his own words for where his intuition stops.
Doesn't defer once confirmed, but delays recognising: same behaviour, one stage earlier.
No risk in over-treating for him; only delay worries him.
Not uncertain, under-served, treats it as optional. Gap is access and habit, not information.
Admitted as a child but did not describe the handover. Not asked.
Two prescribed numbers, no arithmetic. May still hold for inhibitor patients or caregivers.
His own method: "you have to have it happen, then you know what to do."
His own framing: the information exists, it's just everywhere, nobody has put it in one place.
No clinicians in the patient group chat; the doctor is reached by email.
Confident now, at the cost of a childhood of bleeds. Caregivers and newer patients don't get the same intuition.
His biggest remaining uncertainty, in his own words: blood in urine, black stool, something spontaneous he'd never felt. Right person; wrong stage for the worry.
He could not describe a bleed sensation to someone who has never had one. New patients depend on personality: some speak up, some stay quiet.
A caregiver, in his view, is a half-trained nurse reading the bleed from outside a body they cannot feel. Of the four personas, the one with least to work with.
Nothing on handover or what a clinician would need from him. The clinician interview guide remains the route.
What was not on the board and now should be.
He reads every sensation against a profile of his own target joints, built over a childhood of episodes; caregivers never build it, being outside 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.
His typical trigger is a sensation on waking at nine or ten; by afternoon the range of motion is gone.
"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 suits caregivers of children; his episode is a morning one, so the re-check timer should think in hours.
He does not weigh over-treatment; a dose to a non-bleed just raises factor levels for a while. All his decision effort goes into shortening time to "yes, it is a bleed."
"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.
He would inject before A&E, since the ambulance takes time anyway. A firm belief, 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.
Gum bleeds that drag on for days, black stool, blood in urine, a head where the brain has no sensation to feel: cases where his system stops, and most patients have none 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.
A bad elbow bleed leaves him one-handed, and reconstituting factor is a two-handed job; his high-severity threshold is not moving or not fully aware.
"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.
He sees his doctor twice a year and otherwise emails, batching questions into one message. The informal nurse relationship he described belongs to a friend at another hospital.
"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: no clinical channel with a same-day answer between reviews. The group chat's repeat questions are a ready-made content list.
For anything unfamiliar he types symptoms into ChatGPT, withholding his identity, expecting hallucination. What would earn trust instead is a curated source with an authority behind it.
"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.
A good outcome, for him, is the scattered information on bleed types and severity in one place. A triage engine did not come up; personalisation is a later stage.
"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 his two doses on top. A decision engine can wait.
A year on it 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: frequency questions must ask about the last 12 months.
Preparation and scheduling are a real mental and physical factor, taking half an hour to an hour, and he doesn't always keep to the schedule.
"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 home card should name that cost, not assume treating is free.
Asked whether a patient mid-bleed has the mind space for a short questionnaire, his answer was no.
"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; titles must carry the meaning, since he does not read past them.
From what Patient 01 said, the app should:
A second validation call, this time with the clinician side of the journey.
Nurse 02, Hospital B. Mostly severe adult patients on prophylaxis, who call her themselves; caregivers and paediatric patients are not her cohort.
A seven-section nurse guide, run over a roughly 50-minute video call with four team members, Myo Myo leading.
One nurse, one centre. Mild, moderate and paediatric practice were not covered; no haematologist interviewed yet.
Her rule set is site and trauma first: is it a joint, was there a fall. The dose is not a decision, it's the prophylaxis dose already at home; only the number of days varies. For her cohort, recognising a bleed is already solved.
1 dose
The prophylaxis dose
What varies is the number of days, not the amount.
24 h
The key check
Re-check next morning against pain score.
0
Written plans for patients
No checklist or pamphlet exists at Hospital B.
3x
"Report every bleed"
Said three times, unprompted, closing the call.
Each row is what the nurse guide asked, against what Nurse 02 said.
Section 1
Guide asked: who calls her, and how a novice differs from an experienced self-manager
Partly answeredMostly severe adults on prophylaxis; caregivers rarely call. The one group who genuinely cannot tell: older patients with knee arthritis, whose pain doesn't go after factor.
"Most of them know how to manage the bleed. Most of them."On her patients' own competence
Section 2
Guide asked: what guidelines and written resources exist for patients
AnsweredFollows the World Federation of Hemophilia table by bleed site, three to eight day ranges. No written patient checklist exists; patients work from RICE. She is, in practice, the whole loop.
"We did not pull out a set of guideline or checklist for the patient."On why there is no patient-facing plan
Section 3
Guide asked: when and how patients reach her, and what they ask
AnsweredPatients call day two or three, office hours only. Repeated questions: another dose, more days of tranexamic acid, or is it a bleed or arthritis. She does not reply after hours; urgent goes to A&E.
"If it's really urgent, I cannot do anything also. I cannot see your bleed. Even you text me, I cannot help you."On why she does not answer after hours
Section 4
Guide asked: where bleeds happen and what changes the urgency
AnsweredMost bleeds are target joints. Her first fork is spontaneous versus trauma: spontaneous means dose, rest, report next office hour; a fall means dose then A&E even off hours, for an X-ray. Intracranial bleeds come in immediately.
"The important is how soon the factor is administered."On what decides the outcome
Section 5
Guide asked: how she triages by phone, and what a tool must never decide alone
Partly answeredGive one dose the moment a bleed is felt, then A&E if unsure; any trauma or non-joint site goes straight there. Days vary, two to three for a small joint, five to seven for a big one, by pain score.
"In the benefit of doubt, always give yourself one dose first. Then you head down to the A&E if you are not sure."On the rule she opens with
Section 6
Guide asked: how recovery is tracked and when prophylaxis resumes
AnsweredThe key check is next morning by pain score; no relief usually means it isn't a bleed, or rarely an inhibitor, and the patient comes back. Prophylaxis never stops. Physio after every bleed is the ideal that never happens.
"The crucial is the 24 hours."On the one number that decides the next step
Section 7
Guide asked: where a tool can and cannot help
AnsweredFactor first, then trauma or spontaneous, then site; non-joint means dose then A&E. Closing line, said three times: all bleeds must be reported, since a spontaneous bleed on prophylaxis signals dosing needs to change.
"All bleeds must be reported to the healthcare providers."Said three times, closing the call
The same nine points, through a patient without an inhibitor and a nurse whose cohort is severe adults on prophylaxis.
| Patient 01 said | The nurse |
|---|---|
| Inject first, then A&E (his belief, unconfirmed) | Confirmed as the rule she teaches, for trauma and for every non-joint site |
| Two prescribed doses, no calculation | Hospital B patients have one dose, the prophylaxis dose; what varies is how many days. Both are no calculation at the needle |
| Over-treating is not a fear | "In the benefit of doubt, always give yourself one dose first" |
| Uncertainty lives in "is it a bleed" | For her cohort it lives in how many more doses, and how many days of tranexamic acid |
| Unfamiliar sites are the black box | Her answer: not a joint, dose, then A&E to find the source. No middle rung |
| Two-tier severity plus a site list | She does not grade severity remotely at all; site and trauma decide |
| Physio rarely happens, a week sparing the joint, prophylaxis unchanged | Physio ideal and never happens, rest one to two weeks, prophylaxis never stops. The two accounts agree |
| Care team is an email address | Nurse is the whole loop, office hours, and does not reply after hours by design |
| He logs injuries with pride | She wants every bleed reported to her, because a spontaneous bleed is a dosing signal |
Gerald's question on the call: "According to Nurse 02, patients generally know it is a bleed, so does our solution still need to help identify bleeds?"
For her cohort recognition is mostly solved; questions that reach her are downstream, how many more doses, how many days of tranexamic acid. Recognition is still the product for three groups: elderly arthritis patients, an unfamiliar site, and newer self-managers the Hospital B cohort does not contain. Her "one dose first if in doubt" rule is the recognition rule for everyone else, and belongs on the first screen.
Inject first is now confirmed. The nurse teaches it before A&E for trauma and every non-joint site; cards can drop "unconfirmed."
Add a trauma question to the night spine. NGT-06/NGT-07 don't ask about a fall; a trauma fall routes to A&E after the dose even at a target joint.
Report every bleed. Make telling the nurse next working day the last step on every home card (OUT-09).
Carry a days-of-treatment question, not a severity one. The day two/three check-ins (OUT-13) should ask how many more days, using pain score before and after the first dose.
No second-dose path. No relief means probably not a bleed, or rarely an inhibitor; OUT-13b and JNT-17 stay as written.
Add infection to the urine lane. URN-03 adds a GP visit to rule out infection; URN-04 unchanged.
Let the bleed dose default to "same as usual". Correction from interview 03: Hospital B is not uniformly one number. Soften to one or two numbers, set by the centre.
Point recovery at a physio appointment, not exercises. OUT-12: prophylaxis never stops, gym off one to two weeks, physio intensity is a physiotherapist's call.
Add three library entries. Trough-day gym muscle bleeds, dental-scaling gum bleeds respond to tranexamic acid, kidney stones as a cause of blood in urine, all given verbally, nothing written.
Update the persona ledger. Adult patient validated twice; caregivers unvalidated; acquired haemophilia stays out of scope.
Persona ledger
About thirty minutes straight after the Nurse 02 call, on 7 September 2026. Rachel led the analysis, Ariel facilitated. The debrief assumed the Patient 01 call was not recorded; both Gerald's and Rachel's recordings in fact exist.
Twelve calls the team made in the thirty minutes after the nurse hung up.
Stage 1 to 3 move earlier than the Patient 01 page has them. Rachel: "it is a jab first ... not a decision point."
Patient 01's wait-and-see reads as the anomaly, not the norm. Ariel's counter: some observe first for the pain and mental load of infusing, not uncertainty.
Decision: design for the conservative patient. Rachel: "those that are more wait and see will come under it anyway."
The real question moves from "is this a bleed" to "how severe" and "trauma or spontaneous." That decides the order of screens.
Risk assessment comes first, before symptoms. "When was your last prophylaxis dose" surfaces early, since a missed dose raises bleed probability.
Logging becomes an input, not only an output. Pull from an existing log, Patient 03's runs in Excel, rather than ask a frightened person to remember.
Bleed reporting is a must-have. Rachel's doubt: "how many actually report?" A reminder plus the day-after check-in is the feature.
The problem statement needs a post-mortem. Rachel questioned what "self-service" means and how big the off-hours pain really is, since most calls happen in office hours.
The "digital nurse" framing does not fit what the nurse described. Rachel: "they are not asking her what to do. They have done it."
Ariel's hypothesis: a hidden question. "Is there a chance I do not have to infuse first ... I can afford to wait?" Unconfirmed.
Interview 03 changes focus, to pain and needs. Rachel: "we need more flavour and need around the experience of it."
The Patient 01 page places the uncertainty in "is this a bleed"; the debrief finds that stage barely exists for the conservative patient. Both can be true: the nurse's cohort has infuse-on-suspicion trained in, Patient 01 has not. The survey's q15 and q16 are the test.
"How big is the pain outside clinic hours" challenges the whole PS2 premise. The nurse confirmed there is no after-hours channel by design, and Patient 01's episodes start on waking. The honest reframe locates the off-hours pain in the 24 hours after the dose, how many more days, whether to report it, more than a single night-time panic, which is where the app's day-2 check-ins already sit.
The team's "risk assessment first" and the build's "red flags first, then body map" are compatible if risk facts, last dose, trauma or not, are asked in setup rather than mid-night. Rachel's point about memory under stress argues for setup.
A third interview, video call, with an Hospital B patient on a different factor product and dosing rule than Patient 01. In the last twenty minutes, at his own request, he stopped answering the prep sheet and started describing a product.
Patient 03, adult with haemophilia B, factor IX, Hospital B. Two fixed doses: nine for prophylaxis, twelve for a bleed. Target joints right elbow and ankle; kept an Excel log for years.
Video call, about 81 minutes. Ariel led the experience half, a second teammate the problem-statement half, Gerald closed on what he wants built, against the prep sheet above.
One experienced factor IX patient. Recording starts a few minutes late; the last twenty minutes are a stakeholder brief, not observation.
He confirmed factor first without hesitation, even before A&E, for a second product and centre. Unresolved is everything after: what to do once the first dose is in, and how a joint gets back to baseline. He described a searchable site-first reference with the don'ts, a 24-hour check-in that doubles as the report to his nurse, and a recovery track ending at the physiotherapist.
2 doses
Fixed numbers, no calculation
Nine for prophylaxis, twelve for a bleed, set by his care team.
1
Off-hours bleed in the past year
Handled alone, in a car in Malaysia.
2 months
Bleed to physio
The grey zone where range of motion was lost.
0
Target-joint bleeds reported
He tells the nurse about unusual sites only.
The twelve prep questions, put to Patient 03 row by row.
Debrief 1
Asked: what columns, when you fill it in, what you have ever looked back at
AnsweredOne all-in-one sheet, kept for years: inventory, dose per day, and bleed site and amount. His own template, not the hospital's; its job is flagging low stock, since weekly prophylaxis needs no reminder.
"Already in my head, every Sunday."On why the sheet isn't for remembering the schedule
Debrief 2
Asked: the last time something felt off, what did you do first and why
AnsweredTarget joint: knew it was coming, infused as soon as he reached a place to do it. Blood in urine: factor first, same dose, hydrate, monitor. No emotion at the moment.
"It's more like, do I, is this most likely a bleed or not ... it's a bit autopilot for me."On what he feels in the first ten minutes
Debrief 3
Asked: have you ever waited rather than infused, and what were you weighing
AnsweredYes, at night, and on ambiguous muscle sensations, cold compress first. He named the cost himself: asleep and immobile, he cannot react for six or seven hours instead of two.
"If it's late in the middle of the night, I wait until the morning ... I'm not in the right state of mind to give an IV infusion."On why he waits till morning
Debrief 4
Asked: after a bleed, do you tell the nurse, every time, what decides whether you call
Answered"No. I don't." Target joints managed alone; unusual sites only. Hospital B's monthly log is submitted when he remembers, and nobody chases it.
"If we report a bleed, then we just manage on our own, then there's nothing that they can do."On why reporting doesn't change what happens next
Debrief 5
Asked: is it the decision, the needle, being alone, or the morning after
Answered, reframedNot the decision or the needle. No one to ask "then what next" after the first dose, and the night infusion itself, done groggy.
"There was no pain, no symptoms ... I just treated it like a bleed."On the one frightening night, and why it wasn't so frightening
Debrief 6
Asked: what would you want to do without a clinician
AnsweredLook up the site, get the rule and the don'ts, manage the night on that plus his own experience, then contact the nurse next working day. Doesn't want the tool to decide for him.
"I would just search for it. Then it comes up something like, do not use tranexamic acid, right there."On what he wants instead of a clinician
Debrief 7
Asked: what did your doctor say that you can no longer remember
Partly answered"Sometimes I forget," and messages Nurse 02, which happens often. Couldn't recall who taught him the three A&E triggers.
"Just to double-check that we're doing things correctly."On why he messages Nurse 02
Debrief 8
Asked: any symptom you would want permission to wait on
Answered, indirectlyMuscle sensations after exertion. For the urine he gave himself the lower dose and skipped A&E because there was no pain; the nurse later confirmed it.
"I'm more inclined to think it might be muscle ache."On what tips a muscle ache toward a bleed
Carried 1
Asked: prophylaxis, physio, and the week after a bleed
AnsweredCautious, shifts weight-bearing, resumes exercise late. Prophylaxis untouched. His ankle physio session was two months away; nothing for the elbow.
"We often don't think about the optimal recovery."On the week after a bleed
Carried 2
Asked: how you judge it worked, and what you did the one time it did not
Partly answeredNo pain score, no fixed clock. Never had a bleed he couldn't settle at home apart from the open wound.
"I just monitor. If it's not getting better, and if it's actually getting worse, then I'll give another dose."On how he judges the 24 hours
Carried 3
Asked: one number or two, and who set them
AnsweredTwo, set by nurse and doctor: nine and twelve. For the urine, nine was enough. Factor VIII dosing differs by half-life; he can only speak for factor 9.
"We kind of work around certain fixed doses."On the two numbers
Carried 4
Asked: what makes you open a card instead of the group chat
Answered, in his own framingWFH guidance made findable, locally validated by Nurse 02 or the haematologist. He offered to bring the nurse in to trial the workflows.
"A bit subjective ... the advice that we give may not be suitable for this patient's context."On why the group chat isn't enough
The same canvas, run against a second patient with a different product and a different centre.
| Stage | What he does | Verdict |
|---|---|---|
| 1. Trigger | Target joint: he knows before it shows. Unfamiliar site: "you wake up and the toilet bowl is red." | Holds |
| 2. Interpretation | Joint bleeds are easy to diagnose. Muscle sensations are correlated with activity, cold-compressed, and weighed until the probability tips, the nurse's trauma-or-spontaneous fork built in. | Holds |
| 3. Severity | Copy-paste from the target joint unless a red line is crossed: open wound, head injury or concussion, trauma to the torso. Urine colour is his own severity gauge. | Holds, thinner than Patient 01's two-tier list |
| 4. Management decision | Factor first, always, even before A&E, because the nurse and haematologist are at Hospital B; an ambulance goes to the nearest hospital, and only Hospital A, Hospital B and Hospital C have haemophilia care. | Holds; the centre-choice rule is new |
| 5. Treatment and dosage | Two fixed numbers, no calculation. Twelve for joints; the nurse dropped him to nine for the urine. Night delay is the failure mode, not the arithmetic. | Holds |
| 6. Monitoring | "Factor first, but then what next": another dose, go down for evaluation, or which medicines to avoid. He monitors by feel, redoses if worse, calls the nurse next working day for unusual sites only. | Shifts, the uncertainty sits after the first dose |
| 7. Post-bleed recovery | His strongest material, "the grey zone": a two-month gap to physio, no idea how to scale exercise down and back up, cautious enough to lose range of motion in the elbow. | Holds, thinly, and is where he'd move the product |
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.
| Point | Patient 01 | Nurse 02 | Patient 03 |
|---|---|---|---|
| First move | Could I get away with it; wait and watch through the day | One dose at the first sensation, benefit of the doubt | Factor as soon as he can, for a target joint; probability call for muscles; waits till morning at night |
| Dose | Two numbers, no calculation | One number, days vary | Two numbers, no calculation, nurse can drop it to the lower one |
| Unfamiliar site | The black box | Dose, then A&E to find the source | Copy-paste the target-joint protocol, call the nurse next day, A&E only for the three red lines |
| Red lines to A&E | Neck, head, brain, organs, hip | Any trauma, any non-joint site | Open wound, head injury, torso trauma |
| Which A&E | Not discussed | Not discussed | Hospital B, never the nearest; only three hospitals have haemophilia care |
| Reporting | Logs injuries with pride, tells no one | Every bleed must be reported | Own Excel log; Hospital B monthly log submitted when he remembers; nurse told about unusual sites only |
| After hours | Group chat, ChatGPT | Nothing by design; A&E if urgent | Nothing; manage on experience, nurse next working day |
| Recovery | A week sparing the joint, physio rarely | Physio ideal and never happens | Two months to physio, lost range of motion, "the grey zone" |
| Care team | An email address | The nurse is the whole loop | The nurse is the single point of contact and the bottleneck; wants physio and doctors on the platform |
| Caregivers | A half-trained nurse | Almost never call her | Parents of young children are a different population with their own set of needs |
| Style | Wait and see | Conservative | "Our styles are very different ... each of us manage our conditions very differently" |
Asked directly, in the last twenty minutes and at his own request, he laid out a product.
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." Hospital A has no nurse at all.
The gap is the middle ground. "There is not really a middle ground that we can refer to before we go A&E or wait for the next working day."
The shape is a searchable site-first reference, with the don'ts. "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." The WFH guidelines, in his words, are "buried in a 100-plus-page PDF."
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. Ideally we should be contacting the nurses."
The second half of the product is recovery. "Every bleed inherently creates some damage ... if we bleed enough times, how do we get it back to baseline?"
Bring the care team onto the same platform. A prompt both ways: "24 hours past your infusion, report your progress," and "the team can prompt us, we think you should come down for a physiotherapy assessment." Office hours only.
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." He'll ask Nurse 02 who could take part; doctors are "very hard to get hold of".
Outcomes, in HSS's words. "The ultimate long-term aim is to reduce the long-term joint or muscle damage," and reduce caregiver anxiety. He warns off the obvious metric: "reducing bleeds is quite tricky ... I cannot find a direct correlation."
Where the line is. He put the question back to the team: "you cannot have a fully self-help tool because each person may be different." Team's answer: no diagnosis, validated content, decision stays with the human. He agreed and offered Nurse 02 for a trial run.
Two populations he wants remembered. Mild on-demand patients, who go straight to A&E if they cannot infuse; and parents of young children, "deciding, if the child reports I'm feeling weird, should I give factor or not." Teens can use the adult content.
Read as a spec: a site-first reference with the don'ts, locally signed; a 24-hour check-in that doubles as the report; a recovery track ending in a physio prompt; a care-team view that reads the log in office hours. The team's own answer on the call is narrower: it stops at item 4.
Where the line sits, by his account and the team's: no diagnosis, decision stays with the human, guidance calibrated to severity with built-in escalation, local validation layered on top of the WFH text.
The debrief expected the meticulous logger to be the test case for logging as an input. He is: his sheet is inventory first, bleeds second, and he wants someone to read it and call him in, not a reminder to fill it. Logging as receipt to the nurse is confirmed, on condition the care team acts on it.
The debrief's "design for the conservative patient" holds for his target joints and not at night; a nine-vial reconstitution at 2am is a different failure from Patient 01's "could I get away with it". For him the night card's one job is to make the infusion smaller.
The debrief asked how big the pain outside clinic hours really is. His answer: one off-hours bleed in the past year, plus a red-urine night some years ago, both handled alone. Real and rare, sitting in "then what next" after the first dose, not the night itself.
The debrief's hidden question, permission to wait, is answered: yes, for muscle aches with a trigger, and nobody gave it to him. He took it.
Search is a first-class entry, not only the body map. A search box on the home screen and a "do not" line at the top of every site card, URN-04's tranexamic acid rule the model.
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 nurse sees the answer, and say so on the card.
Recovery gets a track, not a card. OUT-12 runs past day three: the physio prompt, and how to scale exercise down and back up. Content needs a physiotherapist's signature; he's sourcing one through Nurse 02.
A care-team view is now asked for by a patient, not only a nurse. A read-only log page, office hours only, would meet his ask. Scope as the v3 spike.
The night card should shrink the infusion, not argue for it. A kit checklist (vials, saline, butterfly, sterilising, disposal) so reconstitution isn't worked out from memory at 2am. Confirm with the nurse before shipping.
Which hospital is a card. He goes to Hospital B A&E, not the nearest, because the nurse and haematologist are there. Print the named centre on the A&E card; check whether SCDF will honour it.
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.
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 usual/upper doses match his.
Cover-nurse variation is an argument for the signed card. Advice differs when the main nurse is on leave; a centre-signed card is the same answer every week.
Outcome claims for the pitch, in HSS's words. Reduce long-term joint and muscle damage, reduce caregiver anxiety. Do not claim fewer bleeds.
Corrections to earlier files
Persona ledger
The longer backlog: what a confident adult and one nurse's cohort could not settle between them, and who can.
A HackitRx 2026 entry by Team Bleedwise: Ariel Tan, Rachel Koh, Myo Myo Swe Oo and Gerald Ho. See the team.
Everything on this page comes from three accounts, two patients and one nurse, and the team's own debrief, and should be read as three data points and one internal discussion, not a finding about all patients or all nurses. Compiled 6 to 9 Sep 2026 for HackitRx 2026 research. Companion to the patient journey file, the case files and the solution files. Illustration AI-generated (Codex CLI), art-directed to the Patient Files design system.
