Crohn’s Disease Activity Index Calculator
Crohn's Disease Activity Index Calculator
Eight weighted terms, three of them seven-day diary sums. Bring the summed counts: the published multipliers apply to the seven-day total, and applying them to a single day under-reports by sevenfold.
CDAI
8 terms → CDAISeven-day stool total 28, pain sum 10, well-being sum 9, one complication, taking loperamide, no abdominal mass, male with a haematocrit of 38%, weight 62 against a standard weight of 68
The eight terms
under 150 remission · 150–220 mild · 220–450 moderate to severe · 450 and above severe
- the multipliers are on SEVEN-DAY SUMS
- 2, 5 and 7 multiply the seven-day totals, not a single day’s value. The Merck Manual publishes the per-day equivalents instead — 14 per stool, 35 per pain point, 49 per well-being point — exactly seven times the published multipliers, which is the cross-check this page is built against
- the haematocrit term is sex-specific
- 47 minus the haematocrit for men, 42 for women, times 6. The trial footnote’s worked fragment: a male with a haematocrit of 40% has a deviation of 7 at 6 points each, so 42 points. Using the male reference for a woman adds 30 points
- the weight term is signed
- a POSITIVE deviation means weight LOSS and adds points; being heavier than standard subtracts them. It is a ratio, so the two weights may be in any unit as long as it is the same one
- two clamps that are not in the index
- both CARMEN analysis plans add footnotes — “if hematocrit subtotal is under 0, enter 0” and “if body weight subtotal is less than −10, enter −10”. Protocol conventions on top of Best’s published form, not part of it; the headline here is unclamped and the clamped total sits beside it
- 150, 220 and 450 are trial conventions
- none is a measured boundary. 220 and 450 are the CARMEN screen-failure rule; 150 is the remission definition everywhere
Worked example
Seven-day stool total 28, pain sum 10, well-being sum 9, one complication, taking loperamide, no abdominal mass, male with a haematocrit of 38%, weight 62 against a standard weight of 68
Diary terms: 2 × 28 = 56, 5 × 10 = 50, 7 × 9 = 63 → 169
Complication 20 × 1 = 20; loperamide 30 × 1 = 30; no mass 0 → 50
Haematocrit: male reference 47, so 6 × (47 − 38) = 54. Had this been a woman, 6 × (42 − 38) = 24, and the index would be 252 rather than 282 — the same blood count, 30 points apart
Weight: (1 − 62 / 68) × 100 = 8.8 → 8.8 points of weight loss
169 + 50 + 54 + 8.8 = 282 points, inside the 220–450 trial entry window
Stopping the loperamide alone would take it to 252 without anything changing in the bowel, which is the single largest switch in the index
The eight terms and their published multipliers
| Term | What is entered | Multiplier | Range of the term |
|---|---|---|---|
| Liquid or soft stools | Total over 7 days | × 2 | 0 to about 280 |
| Abdominal pain | Sum of 7 daily ratings, 0–3 each | × 5 | 0 to 105 |
| General well-being | Sum of 7 daily ratings, 0–4 each | × 7 | 0 to 196 |
| Listed complications | Count of 6 categories present | × 20 | 0 to 120 |
| Antidiarrhoeal use | 0 no, 1 yes | × 30 | 0 or 30 |
| Abdominal mass | 0 none, 2 questionable, 5 definite | × 10 | 0, 20 or 50 |
| Haematocrit | 47 − HCT% (men), 42 − HCT% (women) | × 6 | signed; −108 to +222 at the input limits |
| Body weight | (1 − weight / standard weight) × 100 | × 1 | signed; negative when heavier than standard |
What the thresholds are, and where each comes from
| Figure | Where it is stated | What kind of number it is |
|---|---|---|
| Under 150 | Clinical remission in every trial protocol read | A protocol endpoint definition. One error analysis put the 95% confidence interval of an individual score of 150 at 105–195. |
| 220 to 450 | The CARMEN screen-failure rule: under 220 or over 450 fails screening | A trial ELIGIBILITY WINDOW, repurposed as a severity band by convention. |
| Fall of 70 points | Response in most protocols; 100 points in the stricter ones | A change threshold, so it needs a baseline score and says nothing about a single value. |
| 150 to 450 = mild to moderate | The rifaximin protocol (NCT02240108) | Contradicts the CARMEN reading of 220. Both are printed here rather than reconciled. |
| 451 to 1100 = severe to fulminant | Merck Manual Professional Edition | An upper band with no real ceiling; the index is unbounded above. |
A seven-day diary, eight weights, and the term that dominates it
The CDAI was built in 1976 for the National Cooperative Crohn’s Disease Study, by regressing a physician’s global assessment on a long list of candidate variables and keeping the eight that survived. That origin explains both its shape and its weaknesses: the weights are regression coefficients against a clinician’s opinion, not measurements of inflammation, and three of the eight terms are sums over a seven-day patient diary because that is what the study collected.
Those three diary terms are where implementations go wrong. The multipliers of 2, 5 and 7 apply to the seven-day TOTALS. A page that takes a daily stool count and multiplies it by two produces a stool term seven times too small, and because the error is in the dominant part of the index it moves the answer by hundreds of points. This calculator therefore asks for the sums explicitly and refuses to pretend it can collect a diary. If you only have a day’s data, the index you can compute is the Harvey–Bradshaw index, which was designed for exactly that situation and agrees with the CDAI closely enough to be worth using.
The well-being term is the one to be sceptical about. At seven points a day it can reach 196 on its own, and a formal error analysis of the index found that self-reported wellness contributed about 40% of the total variance of the CDAI. The same analysis estimated the combined standard deviation of the index at up to 23 points and put the 95% confidence interval of an individual score of 150 at 105 to 195. So the remission threshold cannot separate a score of 140 from one of 190 in one patient, and a 70-point fall — the conventional definition of response — is only about three standard deviations of measurement noise. That is not an argument against the index; it is an argument for using it as a difference between two visits rather than as a level.
Two single items swamp the arithmetic and are worth watching. Antidiarrhoeal use is a 30-point switch, so a patient who stops loperamide loses 30 points with no change in their bowel. The haematocrit term is sex-specific — 47 minus the haematocrit for men, 42 for women, times six — so using the male reference for a woman adds 30 points, and an iron-deficient haematocrit of 30 contributes 102 points in a man whether or not the iron deficiency has anything to do with active disease. This index scores symptoms, not inflammation. It can be high with a normal mucosa — bile-acid diarrhoea, a stricture, bacterial overgrowth, coexisting irritable bowel syndrome — and low with active endoscopic disease. That mismatch is why treat-to-target moved off symptom indices and onto objective markers. Check the faecal calprotectin interpreter and the CRP unit converter before reading a high CDAI as inflammation, and the Rutgeerts score if the question is recurrence after a resection. A score is not a diagnosis and a figure from a cohort is not a probability for one patient. This states what the number meant in a named study; the clinician in front of the patient decides what follows. Thresholds here are the published ones; reference intervals and assay units are method- and laboratory-dependent and your own laboratory’s interval takes precedence.
Frequently asked questions
Why does this page ask for seven-day sums instead of daily values?
Because the published multipliers apply to the seven-day totals. The stool term is twice the number of liquid stools over seven days, the pain term five times the sum of seven daily 0–3 ratings, and the well-being term seven times the sum of seven daily 0–4 ratings. A web page cannot collect a diary, so the sums have to be brought to it. The Merck Manual’s calculator takes daily values instead and multiplies by seven internally — 14 points per stool per day, 35 per pain point, 49 per well-being point — which is the same arithmetic stated the other way round.
What counts as a complication for the 20-point item?
Six published categories: arthritis or arthralgia; iritis or uveitis; erythema nodosum, pyoderma gangrenosum or aphthous stomatitis; anal fissure, fistula or abscess; other fistula; and fever over 37.8 °C in the past week. One point per CATEGORY, so two skin manifestations in the third group still score one point and 20 index points. Some secondary summaries collapse these into four groups, which gives a different maximum; the two trial statistical analysis plans this page follows both list six.
Where does the standard weight come from?
The trial footnote says the Metropolitan Life tables, programmed into the study’s diary device. That is a height-sex-frame lookup table, and this calculator cannot hold one — so the number has to come from you, and a figure from any other source should be named as such. A Devine ideal body weight is not the same construct and will give a different standard weight and therefore a different index.
Can the CDAI be negative?
Yes, in the published form. A patient with no symptoms, a haematocrit above the sex reference and a weight above standard has negative haematocrit and weight terms and no positive terms at all. Both CARMEN protocols floor the haematocrit subtotal at zero and the weight subtotal at −10, which prevents most of it; Best’s published form does not. Both totals are shown here so the difference is visible rather than hidden.
Does a CDAI under 150 mean the inflammation has settled?
No. The index scores symptoms, antidiarrhoeal use, a haematocrit and a body weight, none of which is a measure of mucosal inflammation. Patients reach remission by this index with active endoscopic disease, and patients with healed mucosa score high because of bile-acid diarrhoea, a stricture or coexisting irritable bowel syndrome. That gap is why faecal calprotectin, CRP and endoscopy are used alongside it. Note too that a patient needing steroids to stay asymptomatic is not counted as in remission however the arithmetic falls.
Related calculators
References
- Shire/Takeda. Statistical Analysis Plan, SHP647-306 (CARMEN CD 306), Table 11: Calculation of the Crohn’s Disease Activity Index. NCT03566823, clinicaltrials.gov.
- Shire/Takeda. Statistical Analysis Plan, SHP647-307 (CARMEN CD 307), Table 11 and footnotes: Crohn’s Disease Activity Index. NCT03627091, clinicaltrials.gov.
- Merck Manual Professional Edition. Crohn’s Disease Activity Index (CDAI) medical calculator. Rahway, NJ: Merck & Co.
- Jørgensen LGM, Fredholm L, Hyltoft Petersen P, et al. How accurate are clinical activity indices for scoring of disease activity in inflammatory bowel disease (IBD)? Clin Chem Lab Med. 2005;43(4):403–11.
- Sturm A, Maaser C, Calabrese E, et al. ECCO-ESGAR guideline for diagnostic assessment in IBD part 2: IBD scores and general principles and technical aspects. J Crohns Colitis. 2019;13(3):273–84. doi:10.1093/ecco-jcc/jjy114
Not medical advice. For healthcare professionals and education. Reference intervals vary by laboratory and assay — always use your own laboratory's. Never base a dose or a treatment decision on this page alone. Full disclaimer at calcengines.com/disclaimer/