A hospital treats an episode. A clinic follows somebody — antenatal care over nine months, a course of physiotherapy over twelve sessions, diabetes review for years. This is built around that record rather than around the appointment, because the appointment is not where the care lives.
Everything derived from one date, so correcting that date corrects everything.
A vaccine fridge out of range at two in the morning is the case this exists for: by the time somebody opens the cold chain screen, the clinic has a fridge full of vaccine it may not use and no way to prove when it went wrong. The clinic is told instead. There is no dismiss button — an item goes when the fridge comes back into range, not when somebody looks at it — and saying “I am on this” records a name, sinks it below what nobody has taken, and lapses after four hours. Nothing here decides what is wrong: every check calls the same function the cold chain screen calls, so a second opinion about a fridge is impossible. And when a check cannot run it says so, because a short list because a read failed looks exactly like a short list because nothing is wrong.
Rabies is a hundred per cent fatal and a hundred per cent preventable, and the step between those two facts is the one that gets skipped. For a bite through the skin the vaccine alone is not enough: rabies immunoglobulin has to be infiltrated into and around the wound on day nought, because it is the only thing that neutralises virus already in the tissue while the vaccine takes a week to work. Given as an intramuscular injection in the buttock it does almost nothing — and the person has been sent home believing they had it. This software refuses that entry by name.
A category III exposure cannot be recorded without immunoglobulin or a reason in words. The reasons are separated into clinical and not: being previously immunised is clinical; the clinic having none is not. A stock-out is a purchasing decision, and it goes on a board with dates on it rather than disappearing into a column that only ever said “vaccine given”.
A clinic decides whether the skin was broken. The category follows, on the server, from the worst description given — because a person mauled by a dog has usually been bitten and scratched and licked, and a busy afternoon turns a bleeding bite into a category two. Any contact with a bat is category III on contact alone: a bat bite can be too small to see and too painless to feel, and a register that filed that as a lick would be filing a death.
Soap, running water, and a clock. It removes and inactivates virus at the site and it costs nothing, and no rabies register in the country has a field for it. This one does, it says so at the dressing table when the number is short, and it reports the proportion of wounds that got the full fifteen minutes.
The two are often days apart, and a schedule counted from the wrong one puts every subsequent dose on the wrong day. The gap between them is kept as its own number — and it is not a fact about the patient, it is a measurement of how quickly the clinic saw them.
The day-nought and day-three slots are three days apart, so a single injection falls inside the grace window of both. Letting it close them both turns a person who came once into a person who came twice, and quietly removes them from the list of people who need chasing. Each dose is consumed by the slot it is nearest to and is not available to any other.
Not at six. An animal alive at six days may still be incubating, and stopping on that is how somebody dies of a bite that was treated. “Lost track of it” is one of the answers, and it means finish the course — an animal nobody could watch is an animal nobody can clear.
Five registers already know who is late: the rabies course that stopped at the third of five, the child who missed measles, the woman due a follow-up, the person screened positive at a camp, the patient who has not been back since a high score. Every one of them puts the answer on a screen and waits for somebody to open it. This is the half that goes after them — one list, in the order it should be worked, and one button that sends every message that may be sent. Nothing recalculates a schedule: each row comes from the register that owns the clinical rule, so the two can never disagree about who is late.
In this country a woman’s telephone is very often her husband’s telephone. “Your copper-T is due for replacement” arriving on that handset is not an indiscretion — for a woman using a method he does not know about, the message is the disclosure. So each programme declares what a message about it may name, and the discreet ones name the clinic and nothing else: please come and see us when you can. The screen marks which rows those are, because the care taken in the message is undone by the first sentence of the telephone call that follows. Vaccination reminders stay open on purpose — a mother who reads only “come to the clinic” does not know which child, and does not come.
A rabies dose three weeks late has broken a course whose whole value was its timing. “Please come for your dose” tells that person they are still protected when nobody has decided that they are, so it is not sent: the row goes to a doctor instead. A child a year late for a vaccine still needs the vaccine, so that one is chased for a year. Chasing also stops after four attempts — a family chased a fifth time has answered, and teaching them to ignore the clinic costs the message that matters later. Nobody in the mortality register is chased at all, on any programme, for any reason.
The programme counts cataract operations. Districts are judged on cataract operations. Surgeons are paid for cataract operations. What nobody counts is whether the eye can see afterwards — and the surveys that have looked found poor results in twenty to thirty per cent, against a World Health Organization standard that at least eighty per cent should reach 6/18 or better. Here an operation is not closed by being done. It is closed by a best-corrected acuity recorded four to six weeks later, and until then it sits on a list of results nobody knows — a list whose size is the first number on the page.
The commonest cause of poor vision in India is not disease. It is not owning spectacles. A person who reads 6/60 across the room and 6/9 through a trial lens does not have a vision problem; they have an unmet prescription, and counting them as visually impaired both overstates the disease and hides the cheapest intervention in medicine. Blindness here is computed on the best-corrected vision in the better eye, and the software says at the desk not to put that person on a surgical waiting list.
A card with a hole in it. Ten seconds, no equipment. If vision improves through it the problem is refractive; if it does not, it is not. Cataract cannot be recorded from an unaided acuity alone — the refusal offers all three ways out: do the pinhole, record a corrected acuity, or write the cause as not established. The honest answer is always available, and the guess that fills these registers is not.
India’s programme has counted vision worse than 6/60 in the better eye; the World Health Organization counts worse than 3/60. The same person is blind in one register and not in the other, and every comparison between an Indian figure and an international one runs into it. Both numbers are computed, both are shown, and the count of people between the two definitions is shown as well.
The whole acuity scale is an ordering, not a set of numbers — 6/60 is the smaller fraction and the worse result. Nothing in this module compares acuities inline; every comparison goes through one function that knows the ranking, because a comparison written by hand comes out backwards in one of the two directions every single time.
The first review is the one the standard is defined on. A second reading taken three months later after a re-refraction is a different measurement, and overwriting the first with it is how an outcome figure improves without anybody operating better. A result taken too early is refused — the figure is about swelling. A result taken late is kept and marked late, because refusing it would throw away the only result that operation will ever have.
A cataract is removed once. A second entry for the same eye is a duplicate, and a duplicate inflates the count the whole programme is judged on while halving the apparent follow-up rate at the same time. The software refuses it and names the date of the first one.
India ran family planning on targets until 1996, and what targets produced was camps, quotas, and sterilisations performed on people who had not chosen them. The approach has been target-free ever since — and every register that still carries a column headed “achievement against target” reproduces the pressure whatever the policy above it says, because a health worker looks at the column and not at the policy. So the column does not exist here. A request carrying a target, a quota, an achievement or a camp is refused by name, at the top level of the request as well as inside it, and there is no role and no flag that reaches around it.
Three cycles of pills handed over in March is three cycles handed over. It is not three months of contraception, and nobody has ever checked. Couple-years of protection computed from supply are arithmetic on what left the shelf wearing the vocabulary of protection — which is why the national figure has never matched a household survey. Here the two figures sit side by side, named apart: couple-years from supply, and couple-years from methods somebody has actually looked at. The gap between them is the only honest measure of how much the register knows.
It is unknown. It may have been expelled in the first month, which is when most expulsions happen, and the register has been counting a protected couple ever since. Continuation is computed from the dates: in use, due to be repeated, lapsed, past its life, removed, or not checked since — and that last one is a state this software is willing to print, usually as the biggest number on the page.
“Not examined” is one of the findings a follow-up can record, because a visit where nobody looked is a visit where nobody looked. It is stored honestly and it does not move the last-seen date — so it cannot turn an unknown into an in-use, which is precisely what a tick-box register does every day.
About three months, until the semen analysis is clear. A couple sent home believing otherwise conceives, and the register said they were protected from the morning of the operation. The software says the date, at the desk, in the sentence that has to be spoken to the couple — not on a supervisor’s dashboard next month. A tubectomy is effective at once, and the two are not treated alike.
Twenty-two to forty-nine, and a nineteen-year-old is refused even when the request says thirty. The consent reference is required and the refusal says why: only the person being sterilised consents, and a spouse’s signature is not required and is not a substitute. A second sterilisation for the same person is refused as the duplicate it almost always is — a duplicate doubles the couple-years and hides a follow-up that never happened.
Day one, day seven, day thirty after a sterilisation. Those are the ones where a missed visit is a court case rather than a data-quality problem, so they are flagged, counted separately, and sorted to the top with the telephone number in the row. The eligible-couple denominator is computed from the patient list rather than typed — a typed denominator is a denominator somebody chose.
The ninth question of the PHQ-9 asks whether the person has thought they would be better off dead. It scores nought to three like the other eight, and a register that adds it up with them has thrown away the only answer on the form that has to be acted on this afternoon. A person can answer “several days” to item nine and total one out of twenty-seven — minimal, on every scoring table ever printed — and that person is at risk today. Here item nine has its own box, outside the band, and it puts them on the risk list whatever the total says.
The Mental Healthcare Act 2017 presumes that a person who attempts suicide has severe stress, and provides that they shall not be tried or punished. The clinical record must not read like a charge sheet. This software records it as “survived an act of self-harm”, carries the section onto the record itself so it is in front of whoever opens it in a year, and states plainly that nothing in the record is a statement to the police. And that person is on the risk list even if they answered every question “not at all” — because somebody who has just survived that is not somebody who is well.
A person found at risk last week who scores nought this week is still owed the visit nobody made. Most registers replace the old row with the new one: the number improved, the row went green, and nobody went back. Here the open risk is carried forward onto today’s screen, the old row closes as carried rather than as resolved, and the software says so at the moment somebody records the better score.
Being assessed, making a safety plan, restricting means at home and referring are all things the clinic did. What closes a follow-up is being seen by the psychiatrist, being admitted, declining help, or genuinely not being reachable. The list is sorted by how long people have been waiting and turns red after a week — because at that point the clinic found somebody at risk of suicide and then did not go back.
Both instruments ask about the previous two weeks. A score taken in March is a fact about a fortnight that ended in March, and after a month this register stops showing it as current and puts the person on a list to be asked again. A stale number in a box looks exactly like a fresh one.
Fourteen is moderate on the PHQ-9 and moderate on the GAD-7 — but fifteen is moderately severe on one and severe on the other, and using a single table for both is the commonest error in a paper register. Each instrument is scored and banded on its own scale, and an anxiety score never lands in the depression prevalence.
People stop. People are called away. People decline a question. That is fine — what must not happen is a total from six of nine questions being read as if all nine were asked, because it is a low score for a reason that has nothing to do with the person. The software asks why, in words, and keeps the answer beside the total for ever.
Every growth register in the country records a weight and compares it with a line on a chart. A child sitting comfortably on that line who has stopped gaining is in more trouble than a small child tracking steadily below it — and gets a tick and a smile, because faltering is a fact about a sequence and the box on the form only holds today. Here every weighing is read against the weighings before it, and the child in the green who has stopped moving gets a list of their own.
They find overlapping but different children: roughly a third of severely malnourished children are picked up by only one of the two. So the rule is either, never both — a red tape alone is severe, a weight-for-height below minus three alone is severe. A register that waits for the two to agree before it refers misses that third while looking as though it is working, and this one says out loud when the measures disagree so nobody goes looking for a fault in the tape.
Press the top of both feet for ten seconds. A pit that stays is severe acute malnutrition whatever the scale says — and the scale usually says something reassuring, because the child is full of fluid. It is the one finding in this module that ignores everything else and cannot be overridden by a good number.
The arm tape is read from six months and not before: below that the cut-offs were never derived, and reading them anyway gives a confident wrong answer about a baby. So the age comes from the date of birth on the record and the day of the weighing, never from the form — and a child with no date of birth on file is refused outright, because a weight alone concludes nothing.
It is a visit that did not happen. If the scale was broken and the tape was in someone else’s bag, this register says “not enough measured” rather than putting a green row on the board. An empty finding that reads as reassurance is the most flattering wrong answer a growth register can give.
Not pedantry. A second row for the same day makes the interval between the two a gain of nothing, and turns a thriving child into a faltering one on the next read. If the first entry was wrong, correct it — the software says exactly that instead of silently accepting both.
A stunted child is the record of months of undernutrition. A wasted child is an emergency this week. Folding them together sends a stunted but currently well-fed child to a feeding centre that cannot help them, and buries the child who needed to go today. Two columns, two meanings, and the referral text is different for each.
It is a raised reading on one occasion, taken on a plastic chair by somebody who walked two kilometres in the sun to get there. Hypertension is raised readings on separate days. Every screening register in the country collapses the two: the person is entered as hypertensive, put on the line list, counted in the prevalence, and — because they were never actually diagnosed — never treated. Here the word “diagnosed” is refused until the register holds two days, and the refusal says which day it already has.
The loophole every register falls through. A nurse who takes it twice at one sitting has honestly recorded two raised readings and has not established anything. The software counts separate days, not readings — and the blood pressures come from the register rather than from the request, so a screen that would rather supply its own two occasions cannot.
Worse, not merely useless: the person has been told something is wrong with them and then left with it, the programme has spent the money and bought a number, and the clinic believes it has done the work. So the first thing on this screen is everybody who screened positive and has not been seen, longest wait first, with the telephone number in the row. Past a month the row turns red.
It does not take anybody off the waiting list. Neither does “attended”. What closes a follow-up is being diagnosed, being ruled out, refusing to go, dying, or moving away — and the software says so out loud at the moment somebody records a referral and expects to be finished.
The Community Based Assessment Checklist is arithmetic on six answers, and the waist thresholds are 90 cm for men and 80 cm for women. Using the male numbers for everybody under-scores exactly the women the programme is worst at reaching. The total is worked out on the server, shown with its working so a person told they are high risk can see which four points it was, and a half-filled checklist is flagged — because it scores low for the same reason a blank one does, and a low score is the answer that sends somebody home.
A random blood sugar of 140 is a reason to look again. Diabetes is a fasting sugar of 126 or a random one of 200, confirmed at a facility. The screening cut-off is deliberately generous because a second visit is cheap and a missed diabetic is not — and reading a screening threshold as a diagnostic one is how a programme manufactures patients.
Nothing about a normal reading last March says anything about this March. The rescreen list is computed from the date of the last screen rather than from an appointment, so somebody screened once at a camp and never seen again appears on it — which is the entire population a camp-based programme loses. Coverage is withheld until you tell the software how many people are eligible, because coverage of the people who happened to turn up is 100% for ever.
Every store in the country watches the top of the range: the power cut, the hot afternoon, the door left open. The damage happens at the bottom of it. Pentavalent, hepatitis B, DPT, Td, IPV and PCV are built on an aluminium adjuvant that aggregates when it freezes and does not come back when the temperature does. The vial afterwards is clear liquid, intact seal, valid expiry, monitor still at stage one — and it will be drawn up, injected into a child, recorded as a dose, counted in the coverage figure, and do nothing.
It is the single commonest entry in an Indian temperature register, and there is no path through this software by which it clears anything. A freeze event is recomputed from the register every time it is asked, and the freeze-sensitive stock that was in that appliance stays named until somebody shake-tests it. Note also that a refrigerator at one degree has frozen nothing — the range starts at two so somebody notices, and a store that condemns its stock every time the thermostat drifts stops believing the alarm.
The square on the label darkens with every hour of heat the vial has ever seen, and it never lightens. That is its whole value, so a batch recorded at stage three cannot be recorded at stage two next week — the software refuses, names both readings, and points out that this is either a different batch or an earlier mistake. Both are worth finding out before the session rather than after it. And the previous stage is read from the store, not sent with the request.
Suspected freezing is settled by shaking the vial beside one of the same batch that was frozen solid on purpose, and watching how the sediment falls. A control from another batch or another maker has a different formulation and a different sedimentation, and comparing against it gives a confident answer that means nothing — which is worse than not testing. Refused here, by name, with the reason.
BCG and measles are powders until somebody adds the diluent; from that minute they are four-hour vaccines and they are not eligible for the open-vial policy, whatever the multi-dose label says. A multi-dose liquid gets twenty-eight days from opening — or its expiry date, whichever comes first, and the expiry usually comes first. That is the half of the policy everybody forgets.
Ice packs straight from the deep freezer are at about minus twenty. Pack them against pentavalent and Td and the vaccine freezes on the way to the village — and the refrigerator register shows nothing wrong afterwards, because nothing was wrong with the refrigerator. The software refuses the carrier until the packs have been conditioned, and tells you what conditioning is: leave them out until the ice has begun to melt and the pack sloshes when you shake it.
Two readings a day, morning and evening. A day with one reading is not half monitored — it is a day whose other twelve hours nobody saw, and those are the hours the compressor runs unattended. A chart drawn through the readings a register happens to have draws a healthy line straight across the week the clinic was shut, which is the one week worth looking at. Here the gaps are gaps.
A patient is cured when they finished the course and a sputum smear at the end was negative and there was a positive one at the start for it to have converted from. A patient who finished and was never tested is treatment completed, which is a different word for a different thing. Type “cured” without the evidence here and the register refuses, and tells you which of the three is missing. Elsewhere it just saves.
The oldest rule in tuberculosis chemotherapy and the one most often broken with the best intentions. The patient is not improving, so a drug is added; the organism was already resistant to what it was getting, the new drug faces the whole bacillary load alone, and resistance to it follows within weeks. This is how a curable tuberculosis becomes an incurable one. A regimen changed for failure must bring at least two drugs the patient is not taking — and whether it is failing is read from the sputum results, not from the form.
Thirty consecutive days of interrupted treatment is the programme’s definition, and it is computed from the dose dates rather than accepted as a judgement. The same arithmetic produces the list of houses to visit this morning: one day in the intensive phase, a week in the continuation phase, longest gap first, telephone number in the row. It is the only part of a TB register that saves a life on the day it is read.
Dosing is by weight band. A patient who starts at 42 kg and gains nine is now under-dosed — while swallowing every tablet in front of a health worker. That is the quiet way to manufacture resistance in somebody doing everything right. The band is recomputed from the latest weighing on every read, never stored as a decision made in month one, and the screen says the tablets change today.
Tuberculosis has been notifiable by every healthcare provider in India since the Gazette notification of March 2018. An episode here carries either the Nikshay number or a sentence saying why it does not yet — and the register counts the days, which only go up. A field somebody has to fill in is the whole difference between a rule and a reminder.
Cure is defined only for microbiologically confirmed pulmonary cases, so extra-pulmonary and clinically diagnosed patients are excluded from both halves rather than counted as failures. Below twenty of them the percentage is not reported at all: a district that treated nine patients does not have a cure rate, it has nine patients. The denominator is shown so you can see why.
Under the Telemedicine Practice Guidelines 2020 the medicines a practitioner may prescribe are not one list but four, and which applies depends on the mode and on whether this is the first time. A List A medicine on a first teleconsultation needs video — not the telephone, not text — because the permission rests on the practitioner being able to see the patient. Every teleconsultation product in the country treats the mode as a preference.
Narcotics and psychotropics scheduled under the NDPS Act 1985. Not over video, not on a follow-up, not by the most senior physician in the state. It is the one rule in the Guidelines with no exception attached, so there is no permission, no role and no flag in this software that reaches around it — which is the only honest way to implement a rule written without one.
The Guidelines say so in terms: a patient who initiates the teleconsultation has consented by initiating it. A patient the clinic telephoned has not, and explicit consent must be taken and recorded. Every product treats these the same and they are opposite. A caregiver consulting on somebody’s behalf is lawful; a stranger doing it is not, and the recorded relationship is the only thing that tells them apart.
The follow-up rules are the wider ones, so a form that let a clinic declare its own consultation a follow-up would see every consultation declared one. It is checked against the register — earlier teleconsultations from that registration number for that patient, because two people called Dr Menon is a thing and a typed name is not an identity — and the reason it was treated one way or the other is stored in words, for the audit that asks why a List A drug went out over the telephone.
Choosing it clears the prescription, on the screen and on the server. A patient told to come in and handed medicines is a patient who decides they need not — and for an emergency, where the Guidelines permit first aid advice and require immediate direction to in-person care, a teleconsultation that ends in a prescription instead is the failure.
Required on every teleconsultation prescription under the Guidelines, and the prescription is assembled on the server rather than in a browser because it leaves the building. A pharmacist who cannot verify who wrote it is being asked to dispense on the strength of a photograph.
A measles dose at seven months does not produce immunity: the mother’s antibody is still circulating and neutralises it. The child is written down as vaccinated, is not vaccinated, is counted in the coverage figure, is never recalled, and gets measles. Every immunisation register in the country records the dose. Almost none records whether it worked.
It counts, or it does not and must be repeated — computed on the server from the date of birth and the dose before it, and the request cannot set it. A dose given too early is recorded rather than refused: refusing it would mean the injection happened and nothing was written down, which is the worst of the three outcomes. The card shows three states, not two, and the third is in red.
The error in the other direction, just as common. A child who missed the second pentavalent and comes back at eight months is started again from the first — two extra injections, two more months unprotected, and wrong: no series in the national schedule needs restarting however long the gap. The module says so where somebody would otherwise reach for the first dose.
Not from appointments. A child whose family never came back has no appointment to be late for, and a recall list made from the appointment book contains precisely the children who are already coming. Furthest behind first, with a telephone number — and the ones owed a repeat listed apart, because “you are late” and “that one has to be done again” are different conversations.
The proportion of recorded doses that did not count, printed beside the coverage. A clinic where one dose in twelve was given too early has a coverage figure inflated by exactly that much and those children on nobody’s list. It is deliberately not a target: a clinic told to reduce it will stop recording the verdict rather than stop giving early doses, and then the children are invisible instead of merely uncounted.
A pregnancy is not a property of a visit — it spans nine months and thirty encounters. It gets its own record, with the visits hanging off it, so “how far along is she” has one answer rather than living in whichever note somebody last typed it into.
The stage, the expected date and every due date on the schedule are worked out from the programme’s dating. A scan revises the dates and all of them move at once. Storing the stage on each visit means correcting thirty records and missing two.
It is its own action, it asks why, and every correction is kept with who made it. A due date that moved three weeks is a question somebody will ask, and the answer has to include what it moved from.
The screen opens on who is overdue and who is due now — with a phone number to ring. Steps still ahead are deliberately absent: a list that includes everything is a list nobody works through.
Nobody attends on exactly week 20. Each scheduled step carries a window, so a patient inside it reads as due and one past it reads as overdue, which are different problems.
The antenatal schedule most Indian clinics work to is offered as a starting point, not imposed. A clinic that books differently changes the schedule rather than asking us to change the software.
The same clinical core, not a cut-down version of it. These are the modules our hospital product runs, unchanged.
One number, one person, for life. Allergies drawn first and in red, struck through rather than deleted. Duplicates merged rather than removed, combining both allergy lists.
Booked slots checked on the server, walk-ins given a token, and a waiting room ordered by when people actually arrived rather than by their appointment time.
Notes, vitals and prescriptions. A prescription matching a live allergy cannot be signed unless it is explicitly acknowledged, checked against the record at the moment of signing. A signed note is final; corrections are addenda.
Batches with expiry dates. Nothing expired leaves the shelf, stock goes out earliest-expiry-first, and Schedule H1 dispensing is written into a register with the prescriber and the patient.
Reference ranges that know the patient — a haemoglobin of 12.5 is low for a man and normal for a woman — and a critical value that blocks the report until the callback is logged.
The morning board: the waiting room in arrival order, who is due for recall, the lab worklist, pharmacy alerts and money outstanding — the same figures as the screens behind them, because they are produced by the same code. A card that cannot read its number says so rather than showing zero.
A week as a grid, because “is she in on Thursday” is a question you answer by looking rather than by reading a list of dates. Three states a day, not two: on, away, and nothing recorded — a hospital that has never filled the roster in must not have every day read as “off”, or the board is a lie and everything it drives is noise.
Recorded once and read by the outpatient and theatre booking forms, from the same rule, so a consultant marked away cannot be quietly booked over. Nothing is refused: a hospital books over leave every week — leave is cancelled, a colleague covers, the patient is somebody’s own — and software that refused would be overruled by writing the appointment on paper, leaving no record at all.
Overlapping leave for the same person merges: two rows for one absence read as somebody away twice, and deleting one leaves the other quietly warning. Leave that begins the day after another ends stays separate, because a conference and annual leave are two facts and one of them may need accounting for.
Between the sum insured on the card and the amount a hospital receives sit four separate reductions, and every one of them produces the argument at the discharge counter. This works the number out on admission instead: non-payable items first, then the proportionate deduction, then the co-payment, then whatever is left of the sum insured after earlier claims this year — each with the sentence that explains it, on a sheet meant to be printed and handed over.
Take a ₹10,000 room on a policy that allows ₹5,000 and the associated medical expenses are payable at half. It is the reduction nobody expects and the one that produces the largest shortfalls. IRDAI settled what it does not touch — medicines and consumables, implants and devices, diagnostics — and getting that wrong overcharges the patient in one direction and underpays the hospital in the other, so the exemption is written down rather than remembered.
Requests carry the diagnosis and the estimate, because one without them comes back queried and costs three days. A query can still be approved; a denial is final and a fresh request is raised instead, because reopening a denial loses the reason — which is the document an appeal is built on. Emergencies are marked as emergencies, and do not queue behind next Tuesday’s elective.
PM-JAY and Aarogyasri do not pay a share of a bill — they pay a fixed rate for the procedure, and the hospital is paid that whatever its own bill came to. So none of the four reductions applies: no room-rent proportion, no co-payment, no sum insured running down. The beneficiary pays nothing: not the difference, not the non-payable items, not a co-payment. Balance billing a beneficiary is prohibited, and software that computed a patient share would be producing the exact demand the scheme forbids.
Where the bill exceeds the package, the difference is the hospital’s. It is shown as its own figure on the estimate rather than folded into anybody’s column, because it is what a hospital actually uses to decide whether to stay empanelled — and because the one place it must never appear is on the beneficiary’s side of the sheet.
When a TPA settles less than it approved, that gap is a dispute with the payer, not automatically the patient’s debt. It sits visibly as undecided until somebody with the authority writes it off, bills it, or appeals it — and the record says who decided and when. Claims needing an answer are listed by how long they have waited rather than by how much they are worth, because a small query about to lapse needs attention and a large claim submitted yesterday does not.
The counter opens on who owes money rather than on a search box. Consultations, dispensing and tests are billed from what was recorded, so nothing reaches a bill that is not already in the notes — and a test that was ordered and then stopped does not appear on one.
Clinical services are exempt and the software treats them that way. Medicine sold across the counter to somebody who is not admitted is an ordinary taxable sale at the drug’s own rate. Every taxed line prints the reason beside it, so the person at the window can answer why without ringing anybody.
A concession comes off the charges and never off the tax, and it cannot be given without a reason — it is the one line on a bill nobody can account for later and the one most worth accounting for. A bill settles itself when the balance reaches zero.
The group, for a practice with more than one clinic: each one and the sum of them. A clinic that cannot be read is named rather than dropped from the total, because a partial figure presented as whole is worse than no figure.
Three tiers, so a receptionist can register a patient without being able to change a clinical record.
Grant a doctor once at the group and they reach every clinic, including the ones opened next year. No re-inviting anybody when the third branch joins.
Our support administrators, your administrators, and your desk and clinical staff scoped to what their job needs. Roles are data we can change, not code we have to deploy.
Hosted, backed up, monitored and updated by us. Certificates, upgrades and capacity are not on anybody's list at your end.
Your programmes, your schedules, your recall intervals. The system matches the clinic; the clinic does not rearrange itself around the system.
Structure, roles, schedules and branding. Configuration rather than a rewrite, which is why this takes weeks and not quarters.
Your existing register imported and reconciled with you — including the duplicates, which are merged deliberately rather than silently.
On your own address, with your staff trained on it, and us reachable when something is not obvious.
All of the above is in production: the patient register, the appointment book and queue, consultations with the allergy check enforced at signing, the pharmacy with expiry and the Schedule H1 register, the laboratory with patient-specific reference ranges and the critical-value callback log, and care programmes with derived staging and a recall list.
What it does not do yet, said plainly rather than left to be discovered after you have bought it. There is no patient-facing portal or app — patients get printed prescriptions and reports, not a login — no online payment (the counter records what was taken, it does not collect it), and no electronic claim submission — the desk, the estimate and the claim ledger are here, but a request still goes to the payer the way you send it today and the reply is typed back in. If you need any of them now, say so and we will tell you when they land rather than selling you a date.
It keeps working when the connection does not. Screens stay readable from the last known state, and anything that would write to a record refuses out loud rather than appearing to succeed — because a note that silently fails is a note somebody believes they wrote.