Carb counting for a child with type 1 diabetes, using Indian food

The problem is the food table, not the app

Every carbohydrate app has a database behind it, and a database is a list of foods somebody sat down and measured. The well-known ones were measured in Europe, America and Australia. They are good lists. They contain sandwiches, cereal, pasta and pizza, and they do not contain a thepla, a medu vada, a katori of rajma, or the two-roti-and-sabzi that half the country’s children eat for dinner.

A parent counting carbohydrate for an Indian child therefore does the work twice: once to find the nearest wrong thing in the app, and once in their own head to correct it. That is not a small tax. It happens three or four times a day, every day, for years, and it is usually done at speed while a child waits.

The fix is not a cleverer app. It is a food table written for the food. Kulraise ships one — a little over a hundred dishes, by the portion Indian kitchens actually serve them in, with the carbohydrate given as a range and the source of each figure named on the screen beside it. It is readable without signing in.

Why it is a range, and why it will stay a range

Two kitchens do not make the same thepla. One is bigger, one has more besan in it, one is rolled thinner because that is how somebody’s mother did it. The honest carbohydrate figure for one thepla is therefore not a number, it is a band — and a band is what Kulraise shows, tied to a named portion, because “one thepla” means nothing without a size and “22 g” means nothing without a thepla.

This matters more than it sounds. A single number carries an implied precision that the underlying measurement does not have, and precision is exactly what gets trusted. A range says what is actually known: somewhere between these two figures, and you and your child’s team decide how to use that.

It is also why Kulraise will not quietly fill a carbohydrate figure into a blood-sugar reading for you. A number you chose and a number an app chose look identical three weeks later in a clinic, and only one of them is evidence.

Where the figures come from, and how to tell

Every dish in the table carries the name of whoever stands behind its figure. Some come from the Indian food composition tables, some are read off a packet, and some are the figure a family’s own dietitian gave them, entered for that family. On the screen, the credit sits next to the number rather than in a footnote, because the useful question about any carbohydrate figure is “says who?” and it deserves an answer where it is asked.

A dietitian’s figure for your kitchen beats ours for everybody’s, always. If your child’s team has given you a number for the roti your house makes, that is the number to use, and the app is built to take it and keep the credit on it.

Where nobody has measured something, the fields stay empty rather than getting a guess. An empty cell is honest and a plausible invented one is not, and in a record that may be read by a doctor the difference is the whole point.

The book you should have, which is free and is not ours

ISPAD and Life for a Child published Healthy Eating and Carbohydrate Counting for Children and Adults with Type 1 Diabetes: Indian Foods in 2021. It is free to download, it was written by dietitians and paediatric endocrinologists, and it covers Indian food properly — portions, exchanges, how to teach the counting rather than do it for somebody.

If you are newly into this, that book will do more for you in an evening than any app will do in a month. Download it, print it, put it in the kitchen. Nothing on this page is a substitute for it and nothing on this page is trying to be.

What the evidence actually says about counting

There is an Indian randomised trial worth knowing about, and its headline result is not the one a company selling a carbohydrate app would choose. A hundred and twenty-five children aged six to eighteen were randomised to carbohydrate counting or a fixed meal plan and followed for six months. HbA1c fell in both groups and the difference between them was not statistically significant — 8.9% to 8.6% with counting, 9.1% to 8.8% without.

What did move was distress. On a standard measure of the burden of living with diabetes, the counting group scored significantly lower than the fixed-plan group. Read plainly: in that trial, counting did not buy better numbers, it bought a family a less miserable day.

That is worth saying out loud on a page like this, because it sets what a log is for. A log is not a treatment. It is a way of making the same day take less out of you, and of arriving at a clinic appointment with something real to talk about instead of a memory.

What Kulraise does here, and what it never will

It keeps a log. Food eaten, with a carbohydrate range attached and the portion in the words your family uses; blood-sugar readings with the times they were taken; insulin doses that somebody has decided and given, recorded after the fact; and a report that puts the lot in front of a doctor in a readable shape.

It does not calculate or suggest an insulin dose, and it has no route by which it could. Carbohydrate grams, a blood-sugar reading and a ratio from a doctor are the three inputs to a dose calculation, and this app deliberately puts all three on one screen, because whoever does that arithmetic should have them in one place. Doing the arithmetic is a medical device. Kulraise is not one, and does not try to be one.

The health side is free on every plan, for everybody, and always has been. Not a trial, not a tier — there is no setting anywhere in the app that turns it off, which is the strongest form the promise can take.