Elixhauser Comorbidity Index Calculator

Elixhauser is a condition list, not a score. Collapsing it to one number means picking a published weighting, and the published weightings are different instruments: this page uses van Walraven’s and prints the alternatives beside them.

Last updated 8 Oct 2026

Elixhauser index (van Walraven weights)

21 weighted conditions, minus 19 to 89

Congestive heart failureThe heaviest positive weight after liver disease and the two cancer items: adjusted odds ratio 1.96 for in-hospital death, which rounds to 7 points. The Swiss re-derivation gives it 13.

Cardiac arrhythmiasWeight 5 here, from an adjusted odds ratio of 1.71 for in-hospital death; 6 on the Swiss weights.

Valvular diseaseA NEGATIVE weight. Valvular disease carried an adjusted odds ratio of 0.91 in van Walraven’s data, so it subtracts a point. Six of the twenty-one weighted conditions are negative, and an implementation that treats the index as a count of conditions gets those six backwards.

Pulmonary circulation disordersWeight 4 here, from an adjusted odds ratio of 1.48 for in-hospital death; 6 on the Swiss weights.

Peripheral vascular disordersWeight 2 here, from an adjusted odds ratio of 1.26 for in-hospital death; 3 on the Swiss weights.

ParalysisWeight 7 here, from an adjusted odds ratio of 1.93 for in-hospital death; 11 on the Swiss weights.

Other neurological disordersWeight 6 here, from an adjusted odds ratio of 1.83 for in-hospital death; 10 on the Swiss weights.

Chronic pulmonary diseaseWeight 3 here, from an adjusted odds ratio of 1.36 for in-hospital death; 3 on the Swiss weights.

Renal failurePlus 5. Elixhauser’s original has one renal failure item; the AHRQ refinement splits it into moderate and severe. If you are coding from the AHRQ software, check which version’s item list your weights belong to.

Liver diseasePlus 11, the largest single positive weight outside the cancer items. Note that Elixhauser’s liver item is ONE condition, where the Charlson index splits mild from moderate-or-severe. The two indices are not simply different weightings of the same list.

LymphomaWeight 9 here, from an adjusted odds ratio of 2.55 for in-hospital death; 9 on the Swiss weights.

Metastatic cancerPlus 12, the maximum single weight. With lymphoma at 9 and solid tumour at 4 as SEPARATE items, a patient with a metastatic solid tumour and a lymphoma scores 21 from malignancy alone — Elixhauser does not impose the Charlson index’s hierarchy, and the items are not mutually exclusive.

Solid tumour without metastasisWeight 4 here, from an adjusted odds ratio of 1.47 for in-hospital death; 10 on the Swiss weights.

CoagulopathyWeight 3 here, from an adjusted odds ratio of 1.30 for in-hospital death; 9 on the Swiss weights.

ObesityMinus 4 points, from an adjusted odds ratio of 0.64. This is the obesity paradox as a weight: coded obesity was associated with lower in-hospital mortality in these data. It is an association in administrative data, not a benefit, and the usual explanations are collider bias and the fact that obesity is coded more often in patients well enough to be weighed.

Weight lossWeight 6 here, from an adjusted odds ratio of 1.85 for in-hospital death; 6 on the Swiss weights.

Fluid and electrolyte disordersPlus 5. A reminder that this is an index of CODED conditions, not of chronic disease: fluid and electrolyte disorders are mostly acute, and a present-on-admission flag is the only thing separating a comorbidity from a complication here.

Blood loss anaemiaWeight -2 here, from an adjusted odds ratio of 0.81 for in-hospital death; -5 on the Swiss weights.

Deficiency anaemiaMinus 2. Deficiency anaemia and blood loss anaemia both carry negative weights under van Walraven and much larger negative weights (minus 7 and minus 5) under the Swiss re-derivation. Where two published weightings disagree in magnitude on the same condition, the number they produce is not comparable.

Drug abuseMinus 7 points, the largest negative weight in the set, from an adjusted odds ratio of 0.50. Drug abuse is coded in a younger population; the weight is about who carries the code, not about the drug.

DepressionMinus 3 under both weightings, one of the few conditions on which they agree exactly.

12pointsExample

Congestive heart failure (7), renal failure (5), fluid and electrolyte disorders (5), deficiency anaemia (−2) and depression (−3)

Scoring

van Walraven score = sum of the weights of the conditions present
Range −19 to 89; each weight is 10 × ln(adjusted odds ratio), rounded

a list, not a score: Elixhauser is a condition list, not a score. The 1998 publication kept the comorbidities as a set of separate covariates on purpose, because each one affected outcomes differently across patient groups; turning the list into one number means choosing a published weighting, and the published weightings are different instruments giving different numbers.
the nine with no weight: van Walraven’s backward selection kept 21 of the 30 comorbidities at an alpha of 0.05. The others — complicated and uncomplicated hypertension, complicated and uncomplicated diabetes, hypothyroidism, peptic ulcer disease excluding bleeding, AIDS or HIV, rheumatoid arthritis and collagen vascular disease, alcohol abuse and psychoses — score nothing, so they are not items on this page. They are still Elixhauser comorbidities, and they carry non-zero weights under other weightings
six negative weights: valvular disease −1, obesity −4, blood loss anaemia −2, deficiency anaemia −2, drug abuse −7, depression −3. A patient can score below zero, and the minimum is −19. Any implementation that counts conditions rather than summing weights gets all six backwards
derived on: about thirteen years of admissions (1996–2008) at a single Canadian hospital, against in-hospital death, with the weights taken as the adjusted odds ratios on a logarithmic scale. One hospital, one country, one outcome
the alternatives, named: AHRQ publishes two further weightings of its own refined 38-measure version of the list — an in-hospital mortality index derived from about 23.8 million 2018 discharges across 45 states, and a 30-day readmission index. Their per-condition weights are NOT printed here, because the Appendix B tables could not be extracted from AHRQ’s own user guide across three attempts, and this project does not print a weight it has not read. Sharma et al.’s 2021 Swiss weights WERE read and are in the table above
discrimination: in 6,094,672 adult cases from 102 Swiss general hospitals, 2012–2017, with in-hospital mortality of 2.3 per cent: a c-statistic of 0.757 for age, sex and hospital type alone, 0.850 adding Charlson weights, 0.863 adding van Walraven’s Elixhauser weights and 0.867 adding newly derived Swiss weights. The weighted Elixhauser beats the Charlson index by about 0.013 of a c-statistic, and locally refitted weights beat van Walraven’s by a further 0.004

Worked example

Congestive heart failure (7), renal failure (5), fluid and electrolyte disorders (5), deficiency anaemia (−2) and depression (−3)
7 + 5 + 5 − 2 − 3 = 12 points on van Walraven's weights
The same five conditions on Sharma et al.'s Swiss weights: 13 + 8 + 5 − 7 − 3 = 16
The same five conditions counted rather than weighted: 5
One condition list, three numbers — 12, 16 and 5 — and none of them is the Elixhauser index, because there is no such single number. Elixhauser is a condition list, not a score. The 1998 publication kept the comorbidities as a set of separate covariates on purpose, because each one affected outcomes differently across patient groups; turning the list into one number means choosing a published weighting, and the published weightings are different instruments giving different numbers.
Note where the two weightings diverge most on this patient: deficiency anaemia is worth −2 under van Walraven and −7 in Swiss data, and heart failure 7 against 13. The ordering of the conditions is similar; the scale is not
Add uncomplicated hypertension, uncomplicated diabetes and hypothyroidism — three conditions that are not items on this page. The van Walraven total stays at 12, because all three carry a weight of zero; the Swiss total falls by 6 to 10, because they do not

The thirty Elixhauser comorbidities under two published weightings

Hypertension appears as two rows, which is how both sources render that single Elixhauser comorbidity. Every van Walraven weight is 10 times the natural logarithm of its own odds ratio, rounded — all twenty-one of them, which is how this table was checked rather than trusted. The two weightings disagree in magnitude on almost every row and in sign on five.
Comorbidity	van Walraven adjusted odds ratio	van Walraven weight	Swiss weight
Congestive heart failure	1.96	7	13
Cardiac arrhythmias	1.71	5	6
Valvular disease	0.91	-1	-1
Pulmonary circulation disorders	1.48	4	6
Peripheral vascular disorders	1.26	2	3
Hypertension, uncomplicated	not significant	0	-4
Hypertension, complicated	not significant	0	-3
Paralysis	1.93	7	11
Other neurological disorders	1.83	6	10
Chronic pulmonary disease	1.36	3	3
Diabetes, uncomplicated	not significant	0	1
Diabetes, complicated	not significant	0	-1
Hypothyroidism	not significant	0	-3
Renal failure	1.63	5	8
Liver disease	2.97	11	16
Peptic ulcer disease excluding bleeding	not significant	0	0
AIDS/HIV	not significant	0	0
Lymphoma	2.55	9	9
Metastatic cancer	3.30	12	17
Solid tumour without metastasis	1.47	4	10
Rheumatoid arthritis/collagen vascular disease	not significant	0	-1
Coagulopathy	1.30	3	9
Obesity	0.64	-4	-6
Weight loss	1.85	6	6
Fluid and electrolyte disorders	1.61	5	5
Blood loss anaemia	0.81	-2	-5
Deficiency anaemia	0.80	-2	-7
Alcohol abuse	not significant	0	-3
Drug abuse	0.50	-7	-5
Psychoses	not significant	0	-4
Depression	0.73	-3	-3

What the choice of weighting does to discrimination

From 6,094,672 adult cases in 102 Swiss general hospitals, 2012 to 2017, in-hospital mortality 2.3 per cent. Weights refitted locally beat van Walraven’s by 0.004, which is the authors’ own argument that weights do not transport between countries.
Model	c-statistic for in-hospital mortality
Age, sex and hospital type alone	0.757
Plus Charlson weights	0.850 (95% CI 0.849–0.851)
Plus van Walraven Elixhauser weights	0.863 (0.862–0.864)
Plus locally derived Swiss Elixhauser weights	0.867 (0.865–0.868)

Thirty conditions, and the problem with adding them up

Elixhauser and colleagues went looking in 1998 for the comorbidities that actually predicted length of stay, charges and in-hospital death in administrative data, using a statewide California inpatient database of 1,779,167 stays. They found thirty, defined them by ICD-9-CM code, and deliberately did not combine them into an index — because each one affected outcomes differently in different patient groups, and collapsing them threw that away. The thirty were meant to enter a model as thirty separate covariates.

They are now routinely collapsed anyway, because thirty indicator variables are awkward and a single number can be tabulated. Doing it requires choosing a weighting, and there are several. van Walraven and colleagues fitted one in 2009 against in-hospital death across about thirteen years of admissions at one Canadian hospital, keeping twenty-one of the thirty conditions and taking each weight as ten times the logarithm of its adjusted odds ratio; the score runs from minus 19 to 89 and is what this page computes. AHRQ publishes two more over its refined thirty-eight measure list, one for mortality and one for readmission, and Sharma and colleagues fitted another in Swiss national data. These are not alternative renderings of one instrument: they take the same input and return different numbers, and the worked example above returns 12, 16 and 5 depending on which is applied.

Two features of the weighted version surprise people. Six weights are negative — valvular disease, obesity, both anaemias, drug abuse and depression — so a patient can score below zero, and any implementation that counts conditions instead of summing weights inverts all six. And nine of the thirty conditions carry no weight at all under van Walraven, including hypertension and uncomplicated diabetes; they are not items on this page, and they do carry weights under other weightings, which is why a total computed with one weighting cannot be repaired into another.

What the choice buys is real but modest. In 6,094,672 Swiss admissions the c-statistic for in-hospital mortality was 0.757 with age, sex and hospital type alone, 0.850 adding Charlson weights, 0.863 adding van Walraven’s and 0.867 adding weights refitted in that population. The weighted Elixhauser beats the Charlson index; local refitting beats van Walraven. Which weighting to use is a question about your data source and your outcome, not about which index is best.

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

Is there a single Elixhauser comorbidity score?

No, and that is the most important thing about it. Elixhauser is a condition list, not a score. The 1998 publication kept the comorbidities as a set of separate covariates on purpose, because each one affected outcomes differently across patient groups; turning the list into one number means choosing a published weighting, and the published weightings are different instruments giving different numbers. This page uses van Walraven’s 2009 weights; AHRQ’s mortality and readmission indices and Sharma et al.’s 2021 Swiss weights are three further published choices and they give different numbers.

What is the range of the van Walraven Elixhauser score?

Minus 19 to 89. Twenty-one of the thirty comorbidities carry a weight; fifteen are positive and six are negative, so the score can be below zero. The positive weights sum to 89 and the negative ones to minus 19.

Why do some conditions have a negative weight?

Because they were associated with LOWER in-hospital mortality in the derivation data. Obesity carries minus 4 from an adjusted odds ratio of 0.64 and drug abuse minus 7 from 0.50. These are associations in coded administrative data, not benefits: coded obesity marks a patient well enough to be weighed, and coded drug abuse marks a younger one.

Elixhauser or Charlson?

In 6,094,672 Swiss admissions the weighted Elixhauser discriminated in-hospital mortality slightly better than Charlson weights — a c-statistic of 0.863 against 0.850 — on top of 0.757 for age, sex and hospital type alone. But they were derived for different outcomes from different data, Charlson imposes a severity hierarchy where Elixhauser does not, and the answer depends on the question.

Why are the AHRQ weights not printed on this page?

Because they could not be read. AHRQ’s own user guide for the refined software holds them in Appendix B, and three attempts at different offsets all returned text that ended before the tables. The 38 measure names and the derivation description did come through and are used. A weight nobody has read is not printed here.

Related calculators

References

Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36(1):8–27.
van Walraven C, Austin PC, Jennings A, Quan H, Forster AJ. A modification of the Elixhauser comorbidity measures into a point system for hospital death using administrative data. Med Care. 2009;47(6):626–33.
Sharma N, Schwendimann R, Endrich O, Ausserhofer D, Simon M. Comparing Charlson and Elixhauser comorbidity indices with different weightings to predict in-hospital mortality: an analysis of national inpatient data. BMC Health Serv Res. 2021;21:13.
Agency for Healthcare Research and Quality, Healthcare Cost and Utilization Project. Elixhauser Comorbidity Software Refined for ICD-10-CM: user guide v2022.1. hcup-us.ahrq.gov (accessed 9 October 2026).
Wikipedia. Elixhauser Comorbidity Index. en.wikipedia.org (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/