本指南目前僅提供英文版本。
The Diabetes Code → daily measurementThe Diabetes Code by Jason Fung: its causal model, and measuring glucose daily
The Diabetes Code argues that type 2 diabetes is a dietary disease, that lowering a glucose number with medication treats the symptom, and that the cause can be corrected. Those are Fung's claims. The practical measurement gap is narrower: HbA1c summarizes glucose exposure over months, while meals generate glucose responses every day. This guide explains Fung's model, what the evidence supports, and how Prime Factors helps you observe the glucose side of daily eating between lab measurements.
Fung's model of type 2 diabetes
Jason Fung is a nephrologist, a kidney specialist, who came to diabetes through the kidney failure it causes. In The Diabetes Code he argues that too much insulin, not too much glucose, is the root of the disease: years of sugar and refined carbohydrate keep insulin high, the liver and pancreas fill with fat, cells stop responding, and glucose rises last. On this model a drug that lowers glucose without lowering insulin is treating the reading, not the disease. Fung uses the words reversal and cure. This guide keeps his words where it describes his position, and uses the clinical term remission where it describes the evidence.
Insulin first
High insulin, driven by what and how often we eat, comes before insulin resistance and before high glucose.
Fat in the wrong places
Liver and pancreas fat are, in his account, the mechanism that turns high insulin into a disease that feeds itself.
Treat the cause
If the cause is dietary, the correction is dietary. Medication that only moves the number leaves the process running.
The two levers he prescribes
Fung's prescription follows from the model. Lower the insulin stimulus by changing what you eat, and give the body long stretches without food so stored energy gets used. He presents both as central, and he is explicit that they are meant to change the disease, not manage it.
What you eat
Cut sugar and refined carbohydrate, with fructose singled out as the worst offender for the liver. Protein in moderation; natural fats are not the enemy.
When you eat
Intermittent fasting, from skipped meals to multi-day fasts under supervision, to bring insulin down for long enough that the body draws on what it has stored.
Fasting is Fung's lever, not a Prime Factors one, and it carries real risk for anyone on insulin or a sulfonylurea. Nothing in this guide is a fasting protocol. If you are on glucose-lowering medication, changes to eating pattern belong with your clinician.
What the evidence supports
Fung's causal story is stronger than the consensus, but the practical claim underneath it, that diet can put type 2 diabetes into remission, has been tested. Here is what the trials say, and what they do not.
Low-carbohydrate diets and remission
A 2021 meta-analysis of 23 randomized trials found that at six months, 57% of people with type 2 diabetes on a low-carbohydrate diet met a remission definition of HbA1c below 6.5%, against 31% on control diets.
What it does not establish: The effect was smaller when remission required no medication, faded by twelve months, and was weaker in trials that included insulin users. Adherence past six months is the open question.
BMJ: low-carbohydrate diets and type 2 diabetes remission, meta-analysisIntermittent versus continuous restriction
In a 12-month trial of 137 adults with type 2 diabetes, two low-calorie days a week lowered HbA1c about as much as cutting calories every day.
What it does not establish: Equivalent, not superior, and medication was reduced at baseline under a protocol. It does not show fasting works through a special insulin mechanism, only that it is one workable way to eat less.
JAMA Network Open: intermittent versus continuous energy restriction in type 2 diabetesRemission through weight management
In the DiRECT trial, 46% of people with recent type 2 diabetes reached remission at one year through a structured weight-management programme in primary care, against 4% with usual care.
What it does not establish: Total meal replacement, not fasting, and remission tracked weight lost. It shows that the disease is not fixed; it does not show which dietary route is best.
The Lancet: DiRECT primary-care weight management and remissionTaken together: type 2 diabetes can go into remission with sustained dietary change, low-carbohydrate approaches do this at least as well as alternatives in the short term, and the hard part is staying with any of them. None of this makes remission a promise or removes the need for a clinician.
From a three-month average to meal-level feedback
Fung's model emphasizes that eating happens daily while HbA1c reflects roughly three months of glucose exposure. A continuous glucose monitor can add meal-level feedback for the glucose side of that model. It does not measure insulin, which is central to Fung's explanation, and a lower glucose peak does not by itself prove insulin fell. What it can show is how glucose responded around a meal and whether a change produced a repeatable difference.
HbA1c summarizes months. A meal-level glucose response gives you feedback today.
What a sensor can contribute
How specific meals and eating patterns move your glucose, and whether the changes Fung proposes actually change your response.
What a sensor cannot establish
Insulin levels, whether your disease is in remission, or whether any medication should change. Those are lab and clinician questions.
A daily loop on the lever you share with Fung
Of Fung's two levers, one maps directly onto what Prime Factors records for every meal: what carbohydrate is in it, in what amount and alongside what, which the app calls Carbohydrate Composition. Together with food sequence and after-meal movement, that is the daily behavior the app scores. You log the meal, the sensor records the response, and the app turns each response into a glucose-load score that can be compared across days. It uses those scores to estimate and project HbA1c, checks the projection against a later lab result, and continues estimating from the meal log after the sensor phase. This makes the glucose response to daily eating visible between lab measurements without claiming to measure Fung's proposed insulin mechanism.
Strategy
Prime Factors keeps lab HbA1c as the outcome anchor, the same number Fung says is not the whole story. A later lab result helps validate whether the meal-level pattern translated into the intended longer-term direction.
Tactics
Carbohydrate Composition, Food Sequence and After-Meal Movement. Fasting is not among them; it stays with Fung and, if you are on medication, with your clinician.
Feedback
Every meal scored, comparable across days, so a change in what you eat shows up as a repeatable difference or does not, within the week rather than at the next blood test.
Prime Factors does not diagnose insulin resistance, does not measure insulin, does not advise on medication and does not promise remission. It provides meal-level glucose feedback on part of the dietary territory Fung discusses, anchored to a later lab result.
Measure the response between lab results
The Starter Program opens with a four-week sensor phase around the meals you already eat. It is designed to build repeated evidence about how meal changes affect your glucose response, show an estimate of where your HbA1c may be heading and support continued meal logging after the sensor phase.
Explore the Starter ProgramArticle notes and medical boundary
This guide explains a public framework and a limited product role. Fung's causal model and his reversal and cure language are his; the evidence cited here supports remission through sustained dietary change, not a specific mechanism. Prime Factors does not endorse every Diabetes Code claim, and glucose data should be read alongside other clinical information.
Prime Factors is a behavioral tracking and control system, not medical advice, diagnosis or treatment, and it does not replace your clinician. Do not use this guide to start fasting, change medication, or manage an urgent or unsafe glucose reading. The Starter Program is not currently suitable if you have type 1 diabetes or use insulin.
Sources and attribution
Framework descriptions refer to The Diabetes Code: Prevent and Reverse Type 2 Diabetes Naturally by Jason Fung. The book's positions are attributed to its author; they are not automatically scientific consensus or Prime Factors recommendations. Research links are cited separately.