Cellulitis or a noninfectious mimic in a hospitalized adult. Answer the findings below and the score updates as you go.
Runs entirely in your browser. Nothing you enter is sent anywhere, stored, or logged; there is no server and no analytics. Research use only, not yet validated for clinical decisions.
Answer every finding to read the score.
Rule-out below 30%. Rule-in at 65% or above. Thresholds were fixed before evaluation.
Each finding adds its coefficient to the log-odds when present. The bars are those additions; the probability is the sum passed through the logistic curve.
24 852 adults admitted to Beth Israel Deaconess Medical Center between 2008 and 2022 (MIMIC-IV v3.1), one index admission each, with a discharge diagnosis of cellulitis or erysipelas (47.4%) or 1 of 18 pseudocellulitis conditions such as deep vein thrombosis, venous stasis, contact dermatitis, and gout. Lesion laterality was read from the discharge summary by a language model run on local hardware; 17 024 patients had a processed note. Every tier is logistic regression; the coefficients on this page are the model.
A held-out fifth of those patients (3446 with a processed note, 45.4% cellulitis). The Bedside tier reached an area under the curve of 0.72 (95% CI, 0.70 to 0.74) against 0.60 (0.59 to 0.62) for the complete ALT-70 score in the same patients, a paired difference of +0.12 (0.10 to 0.14), P < .001. Rapid 0.73, Standard 0.75, Comprehensive 0.76. In external hospitals (eICU-CRD, 2536 patients) a structured version reached 0.69 against 0.54 for the ALT-70; in an emergency department cohort without any examination finding available (3314 patients) 0.56 against 0.52. Calibration slope 1.00, intercept -0.01.
Clinicians assessing a hospitalized adult with a red, warm, swollen limb who want a transparent estimate to set beside their examination. The reference standard was the coded discharge diagnosis, not expert adjudication. Discrimination was lowest in patients 70 years or older (0.70), the group most likely to have a mimic, so a rule-in result at older age still deserves a second look.
Children, outpatients, and patients with an obvious alternative diagnosis, necrotizing infection, or sepsis. It does not replace examination, ultrasound for suspected clot, or a dermatology consult, and it should not be used to withhold antibiotics from an unstable patient. No tier uses a culture result, so nothing here depends on a test that has not come back.
This calculator accompanies a manuscript under review and has not been validated prospectively. It is provided for research and education. The authors accept no responsibility for clinical decisions made with it.
The Bedside tier is logit = -1.076 + 1.530·unilateral + 0.187·leukocytosis - 0.654·age ≥ 70 - 0.562·malignancy + 0.399·diabetes, and probability = 1 / (1 + e-logit). Worked example: unilateral involvement, diabetes, age 74, no leukocytosis, no malignancy. logit = -1.076 + 1.530 + 0.399 - 0.654 = 0.199; e-0.199 = 0.820; probability = 1 / 1.820 = 0.549, shown as 55%, indeterminate. The hand score for the same patient is +3 + 1 - 1 = +3 (unilateral +3, diabetes +1, leukocytosis 0, age 70 or older -1, active malignancy -1), and in the test set 60% of patients with a hand score of +3 had cellulitis. The Rapid tier uses the same form with 10 coefficients, printed in the source.
Wu E, Lee A, Youssef A, English J, Mason E, Xing EP, Pugliano-Mauro M. CELTIC, a tiered interpretable model for distinguishing cellulitis from its mimics in hospitalized adults. JAMA Dermatol. 2026; submitted. cellulitiscalc.com