Charlson Comorbidity Index Calculator
Charlson Comorbidity Index Calculator
The original 1987 weights, with the hierarchy the index intends and the one-year mortality each stratum carried in the 559-patient derivation cohort. Quan’s 2011 weights are printed alongside, because they give this patient a different number.
Charlson Comorbidity Index (1987 weights)
19 conditions, 0 to 33A 68-year-old with a previous myocardial infarction (1), peripheral vascular disease (1), a transient ischaemic attack with no residual deficit (1), dementia (1) and biopsy-proven non-cirrhotic liver disease (1)
Scoring
Published strata: 0 · 1–2 · 3–4 · 5 or more
- the hierarchy
- three pairs are mutually exclusive and the severe member replaces the mild one rather than adding to it: diabetes with and without end-organ damage, mild and moderate-or-severe liver disease, and localised and metastatic tumour. That is why each pair is one select on this page
- 0 to 33, not 0 to 37
- the nineteen weights sum to 37 and that is the range quoted almost everywhere, but 37 requires scoring both halves of all three exclusive pairs at once, which no patient can be. Applied properly the maximum is 33. The 37 figure is not wrong about the arithmetic, only about patients — and 33 plus the 4 maximum age points of the age-adjusted index is also 37, which is a coincidence that keeps the confusion alive
- which weight set
- Two weight sets are in current use and they give different totals for the same patient. State which one a number came from, every time. This page implements the ORIGINAL 1987 weights. Quan’s 2011 re-derivation changed eight of the ten conditions that keep the same name between the two formulations, sent five to zero, and caps at 24
- chart or code
- the conditions were defined for chart review and are now most often ascertained from ICD codes. The two give different scores for the same patient: coding cannot see the five-year window on the tumour item, cannot distinguish a compensated from a decompensated cirrhosis without looking, and depends on what the discharge summary happened to list
- derived to predict one-year mortality
- in 559 patients on a general medical service in 1984, and validated against ten-year survival in 685 patients with breast cancer. It is now used as a covariate in almost every outcome study there is, which is a use its derivation does not license and its performance in each of those populations is a separate question
- the ten-year survival formula
- a construct of the form 0.983 raised to the power of e to the 0.9 times the score is quoted everywhere alongside this index. Two independent reproductions were read for this page and NEITHER attributes it to any paper, including the two Charlson papers they cite. An exponential in the score is an assumption, not a measurement, and this page therefore prints what the 1987 paper actually published — the four stratum rates above — and computes no survival percentage
Worked example
A 68-year-old with a previous myocardial infarction (1), peripheral vascular disease (1), a transient ischaemic attack with no residual deficit (1), dementia (1) and biopsy-proven non-cirrhotic liver disease (1)
1 + 1 + 1 + 1 + 1 = 5 points on the 1987 weights
5 falls in the published 5-or-more stratum, which carried a one-year mortality of 85 per cent among the 82 patients in it in 1984
Now score the same patient on Quan's 2011 weights: myocardial infarction 0, peripheral vascular disease 0, cerebrovascular disease 0, dementia 2, mild liver disease 2 — a total of 4
Same patient, same five diagnoses, and the total moves from 5 to 4 and out of the top stratum. Three of his conditions are worth nothing on the newer weights and two are worth double
That is the single most important thing to know about this index: a Charlson score is meaningless without the weight set it was computed on. Two weight sets are in current use and they give different totals for the same patient. State which one a number came from, every time.
Change the transient ischaemic attack to a stroke with residual hemiplegia and the 1987 total becomes 6, scoring the hemiplegia item instead of the cerebrovascular one and not both
The nineteen original conditions, grouped by weight
| Weight | Conditions | Count |
|---|---|---|
| 1 | Myocardial infarction · congestive heart failure · peripheral vascular disease · cerebrovascular disease · dementia · chronic pulmonary disease · connective tissue disease · peptic ulcer disease · mild liver disease · diabetes without end-organ damage | 10 |
| 2 | Diabetes with end-organ damage · hemiplegia · moderate or severe renal disease · any tumour without metastasis · leukaemia · lymphoma | 6 |
| 3 | Moderate or severe liver disease | 1 |
| 6 | Metastatic solid tumour · AIDS | 2 |
1987 weights against Quan’s 2011 re-derivation
| Condition | 1987 weight | 2011 weight | 2011 hazard ratio |
|---|---|---|---|
| Myocardial infarction | 1 | 0 | 0.99 |
| Congestive heart failure | 1 | 2 | 1.91 |
| Peripheral vascular disease | 1 | 0 | 1.10 |
| Cerebrovascular disease | 1 | 0 | 1.10 |
| Dementia | 1 | 2 | 2.39 |
| Chronic pulmonary disease | 1 | 1 | 1.28 |
| Connective tissue disease | 1 | 1 | 1.30 |
| Peptic ulcer disease | 1 | 0 | 1.08 |
| Mild liver disease | 1 | 2 | 1.94 |
| Moderate or severe liver disease | 3 | 4 | 3.83 |
| Diabetes without end-organ damage | 1 | 0 | 1.12 |
| Diabetes with end-organ damage | 2 | 1 | 1.22 |
| Hemiplegia | 2 | 2 | 2.26 |
| Moderate or severe renal disease | 2 | 1 | 1.43 |
| Any tumour without metastasis | 2 | 2 | 2.28 |
| Metastatic solid tumour | 6 | 6 | 6.01 |
| AIDS | 6 | 4 | 3.69 |
Two weight sets, one name, and different numbers
The Charlson index was built to do one thing: adjust for the burden of chronic disease when comparing outcomes between groups of patients. Nineteen conditions were selected from the charts of 559 patients admitted to a general medical service at a New York hospital during a single month of 1984, each was given a weight from the relative risk of death it carried, and the weighted sum predicted one-year mortality in four steps: 12 per cent at a score of 0, 26 per cent at 1 to 2, 52 per cent at 3 to 4, 85 per cent at 5 or more. A second cohort of 685 patients followed for ten years reproduced the gradient for death from comorbid disease. Those eight numbers are what the index was validated on.
What has happened since is that the index became a covariate in tens of thousands of studies it was never derived for, and acquired a second set of weights. Quan and colleagues re-derived them in 2004 Calgary discharge data and published the result in 2011. Five conditions — myocardial infarction, peripheral vascular disease, cerebrovascular disease, peptic ulcer disease and uncomplicated diabetes — no longer predicted death and went to zero. Dementia and mild liver disease doubled. AIDS fell from 6 to 4. Both weight sets are in current use, often without being named, and they give different totals for the same patient: the worked example on this page scores 5 on the original and 4 on Quan’s, crossing a stratum boundary without anything about the patient changing.
The second thing that moves a Charlson score is how the conditions were found. The items were written for chart review. Most modern use reads ICD codes instead, and coding cannot see the original’s five-year window on the tumour item, cannot tell a compensated cirrhosis from a decompensated one without reading the notes, and records only what the discharge summary listed. A chart-based and a code-based score on one patient are different measurements with the same name.
Two things this page deliberately does not do. It does not print a maximum of 37, because reaching 37 means scoring both halves of three mutually exclusive pairs. And it does not compute a ten-year survival percentage: the exponential construct quoted everywhere alongside this index is attributed by nobody, and two independent reproductions were read that print it while citing papers which do not contain it. Quan’s weights have their own page, and the age adjustment is a third.
A score is not a diagnosis and a cohort figure is not this patient’s outcome: a stratum in which 52 per cent died within a year describes that stratum, not which 52 per cent. This page reports what a published instrument scored and what that score predicted in a named cohort. It recommends no action. Every weight, cut-off and outcome figure here comes from a named derivation cohort, and cohorts differ in case mix, era, coding and outcome definition; where your own institution’s protocol or analysis plan differs, it takes precedence.
Frequently asked questions
Which Charlson weights does this calculator use?
The original 1987 weights from Charlson, Pompei, Ales and MacKenzie, with the three mutually exclusive pairs handled as the index intends. Quan and colleagues’ 2011 re-derivation is a different weight set with a different maximum and it has its own page; the comparison table above prints both.
What is the maximum Charlson score?
33 as the index is meant to be applied. The nineteen weights sum to 37, which is the figure usually quoted, but 37 requires scoring mild and severe liver disease, complicated and uncomplicated diabetes, and localised and metastatic tumour all at once — three impossible combinations. Quan’s weights cap at 24.
Does a Charlson score of 5 mean an 85 per cent chance of dying within a year?
No. 85 per cent is what happened to the 82 patients who scored 5 or more among 559 admitted to one New York medical service during one month of 1984. A score is not a diagnosis and a cohort figure is not this patient’s outcome: a stratum in which 52 per cent died within a year describes that stratum, not which 52 per cent. Case mix, treatment and coding have all changed since, and a modern cohort at the same score does not have the same mortality.
Where does the Charlson ten-year survival formula come from?
As far as this page could establish, nowhere citable. The construct — 0.983 raised to the power of e to the 0.9 times the score — appears in reproduction after reproduction, and the two read in full for this page both print it without attributing it, while citing Charlson’s 1987 and 1994 papers. No page here computes it. The published figures are the four stratum rates in the bands.
Should the score be computed from the chart or from ICD codes?
Whichever your analysis plan states, and state it. The conditions were defined for chart review; ICD-based algorithms (Deyo, Romano, Quan’s own ICD-9 and ICD-10 mappings) are what most administrative work uses, and they systematically miss what coding does not capture. The two approaches give different scores for the same patient and the difference is not noise.
Related calculators
References
- Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373–83.
- Portal of Medical Data Models (medical-data-models.org), form 16088: Charlson Comorbidity Index, citing Charlson et al. 1987. medical-data-models.org/16088 (accessed 9 October 2026).
- Quan H, Li B, Couris CM, Fushimi K, Graham P, Hider P, Januel JM, Sundararajan V. Updating and validating the Charlson comorbidity index and score for risk adjustment in hospital discharge abstracts using data from 6 countries. Am J Epidemiol. 2011;173(6):676–82.
- CASRAI. Charlson Comorbidity Index (guide). casrai.org (accessed 9 October 2026).
- FPnotebook. Charlson Comorbidity Index. fpnotebook.com (accessed 9 October 2026).
- Omni Calculator. Charlson Comorbidity Index (CCI) Calculator. omnicalculator.com/health/cci (accessed 9 October 2026).
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/