Lee 2019 Immune Effector Cell-Associated Encephalopathy (ICE) points: orientation 0-4 (year, month, city, hospital), naming 0-3 (three objects), following commands 1, writing 1, attention 1 (count backward from 100 by tens); total 0-10. Caller-assigned; magent does not perform the exam. Not ASTCT ICANS grading. Deterministic published arithmetic with citation; not a medical device.

Hemphill 2001 ICH Score from caller-assigned GCS, age, infratentorial origin, ICH volume (mL), and IVH; max 6. GCS 3-4 scores 2, 5-12 scores 1, 13-15 scores 0; age ≥80, volume ≥30 mL, infratentorial origin, and IVH score 1 each. Deterministic published score with citation; not a diagnosis or withdrawal-of-care protocol. magent does not read CT. Not FUNC and not ICH-GS.

Hassanein 2021 IDF-DAR fasting risk table: type 1 scores 1; duration ≥10 scores 1; hypoglycemia 0/1/3.5/5.5/6.5; HbA1c band 0/1/2; treatment 0/0.5/1/1.5/2/2.5/3; SMBG 0/1/2; DKA/HONC 0/1/2/3; MVD 0/2/6.5; eGFR band 0/2/4/6.5; pregnancy 0/3.5/6.5; frailty 0/3.5/6.5; labour 0/2/4; prior negative Ramadan 1; fasting ≥16 h 1. Low 0–3, moderate 3.5–6, high >6. Caller-assigned; not a fasting order or insulin dose.

Park 2016 ICE Table 5 simplified 6-month IE mortality score: constant 4, age bands (<=45: 0; 46-60: 2; 61-70: 3; >70: 4), and caller-assigned flags (dialysis +3, nosocomial +2, prosthetic +1, symptoms >1 month -1, S. aureus +1, viridans -2, aortic/mitral vegetation +1 each, NYHA 3-4 +3, stroke +2, paravalvular +2, persistent bacteremia +2, surgery -2). Returns score, quadratic mortality percent, and score >= 8. Not a diagnosis. magent does not examine the patient or decide surgery.

Barbour 2019 race-free full IgAN model: 5-year risk of 50% eGFR decline or ESKD from age, eGFR, MAP, proteinuria g/day, Oxford MEST, RAS blockade, and immunosuppression at biopsy. Risk = 1 - S0(60)^exp(LP) with eTable 3 coefficients. Caller supplies eGFR; magent does not compute CKD-EPI or score the biopsy. Not the race/ethnicity model. Not KFRE. Not a diagnosis or treatment order. Not a medical device.

Heng 2009 IMDC: KPS <80, diagnosis-to-treatment <1 year, hemoglobin <LLN, corrected calcium >ULN, neutrophils >ULN, platelets >ULN; 1 point each, max 6. Favorable 0, intermediate 1–2, poor ≥3. Deterministic published score with citation; not an RCC diagnosis or treatment recommendation. magent does not assign KPS or compare Hb, Ca, ANC, or platelets to institutional LLN/ULN.

Steyerberg 2008 IMPACT prognostic score for 6-month mortality and unfavorable outcome after traumatic brain injury. model=core (age, motor, pupils), extended (plus hypoxia, hypotension, Marshall CT, tSAH, EDH), or lab (plus glucose mmol/L and Hb g/dL). Points and printed logistic LPs from that paper. Not a diagnosis, not MRC CRASH (Perel 2008), and not magent chemotherapy CRASH.

Sanfilippo 2019 IMPEDE VTE for multiple myeloma: IMiD 4, BMI>=25 1, pelvic/hip/femur fracture 4, ESA 1, doxorubicin 3, dexamethasone none/low/high 0/2/4 (high if monthly >160 mg), Asian/Pacific Islander -3 (caller-assigned), VTE history 5, tunneled line 2, thromboprophylaxis none/prophylactic/therapeutic 0/-3/-4. Unclamped. <=3 low 3.3%, 4-7 intermediate 8.3%, >=8 high 15.2%. Not a prophylaxis order. Not SAVED or Khorana.

Decousus 2011 IMPROVE admission bleeding-risk score: ulcer 4.5, prior bleeding 4, platelets <50 4, age >=85 3.5, hepatic failure 2.5, severe renal failure 2.5, ICU/CCU 2.5, CVC 2, rheumatic disease 2, current cancer 2, age 40-84 1.5, male 1, moderate renal failure 1. Max 30.5. Increased risk at >=7. Not a VTE-prophylaxis order. magent does not examine the patient.

Spyropoulos 2011 7-factor associative IMPROVE VTE score during hospitalization: previous VTE 3, known thrombophilia 2, lower-limb paralysis 2, current cancer 2, immobilized >=7 days 1, ICU/CCU stay 1, age >60 1. Max 12. Bands: 0-1; 2-3 (observed VTE 1.5%); >=4 (observed VTE 5.7%). Not a VTE diagnosis or prophylaxis order. Not the 4-factor admission predictive model. Not IMPROVEDD. magent does not examine the patient.

Gibson 2017 IMPROVEDD VTE score: Spyropoulos 2011 7-factor IMPROVE items (previous VTE 3, known thrombophilia 2, lower-limb paralysis 2, current cancer 2, immobilized >=7 days 1, ICU/CCU 1, age >60 1) plus 2 if caller-assigned D-dimer >=2x ULN. Max 14. low 0-1 vs high >=2 (heightened VTE through 77 days when >=2). Not a VTE diagnosis or prophylaxis order. Not the 4-factor admission model. Not Padua or Caprini. magent does not assay D-dimer.

IMWG 2014 myeloma requiring therapy (Rajkumar 2014). Met when clonal BMPC ≥10% or biopsy-proven plasmacytoma plus ≥1 MDE: hypercalcemia, renal insufficiency, anemia, osteolytic lesion (CRAB), or BMPC ≥60%, involved:uninvolved FLC ratio ≥100 with involved FLC ≥100 mg/L, or >1 focal MRI lesion ≥5 mm (SLiM). Not a diagnosis. Not smouldering MM, MGUS, POEMS, or amyloid. Not Greipp ISS or R-ISS. magent does not assay calcium, creatinine, hemoglobin, or FLC, and does not read MRI.

Kumar 2016 IMWG measurable serum+urine M-protein response. First-match PD, sCR, CR, VGPR, PR, MR, or SD from caller-assigned flags. sCR is CR plus normal FLC ratio and no clonal BM cells by IHC. CR is negative serum and urine IFE, no plasmacytomas, and BMPC <5%. VGPR is electrophoresis-negative or ≥90% serum M-protein reduction with urine M-protein <100 mg/24 h. Not a diagnosis. Not MRD. Not FLC-only. magent does not assay labs.

Greenes 2001 infant scalp score: age points (0-2 months → 3; 3-5 → 2; 6-11 → 1; 12-24 months → 0) plus caller-assigned hematoma size (0 none, 1 small, 2 medium, 3 large) and location (0 frontal, 1 occipital, 2 temporal/parietal). Total 0-8. magent does not examine the scalp, measure the hematoma, or decide CT. Does not apply later ≥4 clinically important TBI cuts.

Aerts 2017 INTERCHEST: six equally weighted predictors for CAD in primary-care chest pain (age ≥55 male or ≥65 female +1; physician suspected serious +1; history of CAD +1; pain on exertion +1; pressure quality +1; pain reproducible on palpation −1). Range −1 to 5. CAD considered absent iff score <2. Published CAD probability 2.1% (95% CI 1.1–3.9) when <2 and 43.0% (35.8–50.4) when ≥2 in the 13.2% prevalence setting. Not HEART, not EDACS. magent does not examine the chest.

Mandelbrot 1988 intrauterine RBC transfusion volume: fetoplacental_volume_ml = 1.046 + estimated_fetal_weight_g * 0.14; volume_ml uses the unrounded fetoplacental volume times (hct_final_percent - hct_initial_percent) / hct_donor_percent. Hematocrit arm only; the paper also allowed platelet count — magent does not implement platelets. Caller-assigned estimated fetal weight; magent does not estimate weight from biometry. Not a medical device or transfusion order.

Timmerman 2008 simple ultrasound rules. Caller assigns five M-rules (irregular solid tumor; ascites; ≥4 papillary structures; irregular multilocular-solid ≥100 mm; very high color Doppler) and five B-rules (unilocular cyst; solid component <7 mm; acoustic shadows; smooth multilocular <100 mm; no Doppler flow). Malignant if any M and no B; benign if any B and no M; otherwise inconclusive. magent does not read ultrasound. Not an ovarian-cancer diagnosis. Not RMI. Not Swede. Not SRRisk.

Shipp 1993 IPI for aggressive non-Hodgkin lymphoma: age >60, Ann Arbor III or IV, extranodal sites >1, ECOG ≥2, LDH > ULN; 1 point each, max 5. Low 0–1, low_intermediate 2, high_intermediate 3, high 4–5. Deterministic published score with citation; not a lymphoma diagnosis or treatment choice; not age-adjusted IPI, not R-IPI, not NCCN-IPI. magent does not assay LDH.

Condoluci 2020 IPS-E for asymptomatic early-stage chronic lymphocytic leukemia: unmutated IGHV 1, ALC >15 ×10^9/L 1, palpable lymph nodes 1; max 3. Low 0, intermediate 1, high 2–3. Time-to-first-treatment score, not a treatment start order; not CLL-IPI, Rai, or Binet. magent does not assay IGHV, count lymphocytes, or palpate nodes.

Greenberg 2012 IPSS-R for MDS: cytogenetics (very_good 0, good 1, intermediate 2, poor 3, very_poor 4; caller assigns the Schanz subgroup), BM blasts, hemoglobin g/dL, platelets, and ANC; max 10. very_low ≤1.5 through very_high >6. Deterministic published score with citation; not a diagnosis or transplant listing; not IPSS 1997 or IPSS-M. magent does not karyotype.

Brahmer 2018 ASCO Table 2.1 irAE colitis grade from stools_over_baseline (0-40) and four caller-assigned flags. First-match: grade 4 if life_threatening; else 3 if stools >= 7 or incontinence or hospitalization_indicated or limiting_self_care; else 2 if stools >= 4; else 1. Returns icpi_action continue, hold, hold_consider_discontinue_ctla4, or permanently_discontinue. Ostomy-output bands not implemented. magent does not count stools. Not a medical device, not a diagnosis, not a steroid dose.

Brahmer 2018 ASCO Table 4.4 irAE diabetes from glucose_mg_dl (50-1000, mg/dL only) plus caller ketosis, t1dm, severe_unable_adl, and life_threatening. First-match: G4 if life_threatening or glucose >500; else G3 if severe_unable_adl or glucose >250; else G2 if ketosis or t1dm or glucose >160; else G1 if glucose >99; else G0. ICI continue (0-1), hold_until_control (2), hold_admit (3-4). Not DKA diagnostic criteria. magent does not assay glucose. Not a diagnosis. Not an insulin dose.

ASCO 2018 (Brahmer) immune-related hepatitis grade from caller AST, ALT, and total bilirubin as folds of laboratory ULN (mg/dL only). Lab grade is the max of transaminase and bilirubin Table 2.2 bands; decompensated liver upgrades to grade 4; symptomatic compensated liver dysfunction upgrades to grade 3. ICI action: continue (0-1), hold (2), or permanently discontinue (3-4). Not a medical device, not a diagnosis, not DILI/RUCAM, not R factor, not a steroid dose; magent does not assay enzymes.

Brahmer 2018 ASCO Table 4.1.1 irAE primary hypothyroidism grade from tsh_miu_l (0.01-150 mIU/L) and three caller-assigned flags. First-match: grade 3 if severe_or_life_threatening or unable_to_perform_adl (paper G3-4 as grade 3); else 2 if moderate_symptoms or TSH >= 10; else 1 if TSH < 10. Returns icpi_action continue (grade 1) or hold_until_baseline (grade 2 or 3). Not hyperthyroidism (Table 4.1.2). magent does not assay TSH/FT4. Not a medical device, not a diagnosis, not a levothyroxine dose.

ASCO 2018 (Brahmer) irAE nephritis grade from creatinine vs baseline (mg/dL). First-match: 4 if dialysis or life-threatening; 3 if fold >3, creatinine >=4.0, or hospitalization; 2 if fold >2; 1 if fold >=1.5 or delta >=0.3; else 0. ICI action continue (0), consider_hold (1), hold (2), permanently_discontinue (3-4). Not a medical device, not a diagnosis, not KDIGO/AKIN, not a steroid dose; magent does not assay creatinine.

Brahmer 2018 ASCO Table 3.1 irAE pneumonitis. First-match: G4 life-threatening respiratory compromise; else G3 hospitalization, oxygen, limiting self-care, or all lobes/>50%; else G2 symptomatic, >1 lobe, 25-50% parenchyma, or limiting instrumental ADL; else G1. ICPi: hold_if_radiographic_progression (G1), hold (G2), permanently_discontinue (G3-4). Caller-assigned flags. magent does not read CT or pulse oximetry. Not SITC 2021, not a CTCAE instrument, not a diagnosis, not a steroid dose.

Injury Severity Score (Baker 1974): ISS is the sum of squares of the three highest caller-assigned AIS body-region grades (head, face, chest, abdomen, extremity, external; each 0-6). Any AIS 6 (unsurvivable) is ISS 75. Deterministic published arithmetic with citation; not a triage order and not NISS (which squares the three worst injuries regardless of region).

Rodeghiero 2010 ISTH/SSC bleeding assessment tool (BAT). Fourteen caller-assigned Table S1 symptom integers (0-4; cns_bleeding 0, 3, or 4 only). magent sums only (max 56) and does not take the bleeding history. This 2010 letter did not identify an abnormal bleeding score; magent does not apply later Elbatarny 2014 / WFH ≥4 male / ≥6 female / ≥3 child cutoffs. Not a VWD diagnosis and not a medical device.

Taylor 2001 ISTH SSC overt DIC (four lab items; max 8). Platelets >=100=0, 50-<100=1, <50=2. fibrin_related_marker none/moderate/strong = 0/2/3 (caller-assigned; magent does not grade D-dimer). PT sec above ULN <3=0, >=3 and <6=1, >=6=2. Fibrinogen >=1.0 g/L=0, <1.0=1 (XOR g/L or mg/dL÷100). Band <5 not compatible with overt DIC; >=5 compatible. Not a diagnosis. Not the non-overt serial score.

Misra 2013 ITAS2010: weighted sum of 44 caller-assigned Takayasu clinical-activity items that are new or worsening within 3 months and ascribed to active vasculitis. Seven items score 2 (diastolic hypertension, stroke, bruits, pulse inequality, pulse loss, claudication, carotidodynia); the other 37 score 1. Max 51. active_disease iff score ≥ 2. magent does not examine the patient. Not ITAS-A, not a diagnosis, and not a medical device.

INS gravity drip rate: gtt/min = (volume_ml × drop_factor_gtt_ml) / time_min, rounded half away from zero to an integer. Drop factor 10, 15, or 20 gtt/mL macrodrip or 60 gtt/mL microdrip. Also reports mL/hour. Not a pump program, drug order, or device. Deterministic published arithmetic with citation.

Powers 2018 Table 6 Class III Harm IVT. contraindicated iff any of 15 caller flags: CT ICH; recent ischemic stroke, severe head trauma, or intracranial/spinal surgery; posttraumatic infarction; history of ICH; SAH presentation; GI malignancy or bleed 21d; coagulopathy; LMWH 24h; DOAC not cleared; GP IIb/IIIa; infective endocarditis; aortic arch dissection; intra-axial neoplasm. Not Class I eligibility, time window, or a dose. magent does not read CT or assay labs. Not a treatment order.

Dubois 2014 IWG-2 typical AD research criteria. typical_ad iff early hippocampal-type episodic memory, no better explanation, and in-vivo AD pathology (CSF Abeta1-42 plus T-tau or P-tau, amyloid PET, or a PSEN1/PSEN2/APP mutation). Typical AD only; not atypical phenotypes. Not NINCDS-ADRDA, not NIA-AA ATN, not DSM. magent does not examine the patient, assay CSF, or read PET. Not a diagnosis or medical device.

Gewitz 2015 AHA Jones criteria (Table 7). jones_arf iff evidence_of_preceding_gas and the episode rule: initial 2 major or 1 major plus 2 minor; recurrent also allows 3 minor. Majors/minors follow population_risk for arthritis vs arthralgia, fever, and ESR/CRP. Prolonged PR does not count with carditis. Joints count in major or minor, not both. magent does not examine the patient or read an echocardiogram. Not Jones 1944. Not a diagnosis. Not a medical device.

Karnofsky 1949 Performance Status: caller-assigned integer 10, 20, 30, 40, 50, 60, 70, 80, 90, or 100 (normal; minor signs or symptoms; normal activity with effort; cares for self; occasional assistance; considerable assistance and frequent medical care; disabled; severely disabled; very sick, hospitalization necessary; moribund). Grade 0 (dead) is not scored. Returns score, max 100, and a stable label. Not a performance-status assignment and not ECOG; magent does not observe the patient.

Kasymjanova 2010 prognostic index for advanced NSCLC from caller-supplied CRP in mg/L and WBC in ×10^9/L. CRP is elevated if crp_mg_l > 10 (10 is not elevated). WBC is elevated if wbc_10e9_l > 11 (11 is not elevated). PI 0 if neither marker is elevated, PI 1 if exactly one is elevated, PI 2 if both are elevated. Not Glasgow Prognostic Score, not mGPS (McMillan), not a cancer diagnosis, and not a chemotherapy order. magent does not assay CRP or WBC.

Katz 1963 Index of Independence in Activities of Daily Living: six caller-assigned independence flags (bathing, dressing, toileting, transferring, continence, feeding). Each independent function scores 1; max 6. magent sums the published 0–6 count and does not observe the patient. Not the Barthel Index, not Katz letter grades A–G, and not a rehab order or placement decision. Not a medical device.

AHA 2017 classic/complete Kawasaki disease (McCrindle). Complete when fever_days >= 4 and at least 4 of 5 principal features are true (oral mucosal changes, bilateral conjunctival injection, rash, extremity changes, cervical lymphadenopathy). Does not implement the incomplete-KD lab/echo algorithm or the 3-day clinician exception. magent does not examine the patient or read an echocardiogram. Not a medical device and not a diagnosis.

KDIGO 2012 Table 2 AKI stage 0-3 from baseline and current creatinine in mg/dL and a caller-assigned urine-output category. Stage is the worse of the two limbs; RRT or a caller pediatric eGFR <35 flag is stage 3. magent does not assay creatinine, measure urine, collect weight, or compute eGFR (no MDRD). Not a medical device, not a diagnosis, not AKIN 2007, and not RIFLE.

Chen 2013 kinetic eGFR (KeGFR) from two plasma creatinines, hours between them, caller-assigned steady-state creatinine and corresponding eGFR/CrCl, and required max_delta_pcr_mg_dl_day (Chen usual default 1.5). KeGFR = (SSPCr * CrCl / MeanPCr) * (1 - 24*dPCr / (dTime_hours * MaxdPCr/Day)), clamped at 0 when that (1 - ratio) term is negative. Creatinine mg/dL only. Not a CKD-EPI result. Not a dialysis prescription. magent does not compute MDRD or CKD-EPI.

Tangri 2011 4-variable Kidney Failure Risk Equation: 2- and 5-year risk of dialysis or transplant from age, male, eGFR, and urine ACR (mg/g). Risk = 1 - S0^exp(LP) with that paper's 4-variable coefficients. Caller supplies eGFR; magent does not compute CKD-EPI or assay ACR. Not the 8-variable model, not Tangri 2016 country S0, no race. Not a dialysis or transplant order. Not a medical device.

Khorana 2008 Blood predictive model for chemotherapy-associated thrombosis: very_high site (stomach or pancreas) 2, high (lung, lymphoma, gynecologic, bladder, testicular) 1, other 0; platelets >=350 1; hemoglobin <10 g/dL or ESA 1 (once); WBC >11 x10^9/L 1; BMI>=35 caller flag 1. Max 6. 0 low; 1-2 intermediate; >=3 high. Not a diagnosis. Not Padua or Caprini. Caller assigns cancer site; magent does not compute BMI.

Killip and Kimball 1967 myocardial-infarction heart-failure class 1-4 as assigned by the caller. Returns the published class, description (no heart failure; S3 and basilar rales; pulmonary edema; cardiogenic shock), and 1967 CCU hospital mortality (6%, 17%, 38%, 81%). Not a diagnosis; magent does not examine the patient. Not GRACE or contemporary registry mortality.

Kim 2008 NEJM MELD-Na from creatinine, bilirubin, INR, sodium, and dialysis. Then-current UNOS MELD: dialysis sets creatinine 4.0 else clamp 1.0-4.0; bilirubin and INR floor 1.0. Sodium bound 125-140. MELDNa = MELD - Na - 0.025*MELD*(140-Na) + 140. Not OPTN 2016 MELD-Na (Na 125-137 only if MELD>11). Not Kim 2021 MELD 3.0. Not a transplant listing or allocation order.

King's College criteria (O'Grady 1989) for poor prognosis in fulminant hepatic failure. Acetaminophen: arterial pH <7.30, or PT >100 s plus creatinine >300 umol/L plus grade III/IV encephalopathy. Non-acetaminophen: PT >100 s, or any 3 of age <11 or >40, unfavorable etiology (non-A non-B or drug reaction), jaundice-to-encephalopathy >7 days, bilirubin >300 umol/L, PT >50 s. PT seconds, not INR. Not a transplant listing order.

Kocher 1999 four predictors for children with irritable hip: history of fever (oral temperature >38.5 C), non-weight-bearing, ESR >=40 mm/h, and serum WBC >12000/uL; 1 point each, max 4. Returns published predicted probability of septic arthritis vs transient synovitis. Deterministic published score with citation; not a diagnosis, not Caird CRP, not adults, and not an aspiration order.

Kruis 1984 weighted IBS versus organic score. History: pain/flatulence/irregularity triad +34, duration >2 years +16, characteristic pain descriptors +23, alternating constipation/diarrhea +14. Organic flags: abnormal physical or pathognomonic history -47, ESR >10 mm/h -13, leukocytosis -50, anemia -98, blood in stool -98. IBS iff score >= 44. Caller assigns flags. magent does not examine the patient or run labs. Not a diagnosis, not Manning, not Rome.

Daugirdas 1993 second-generation single-pool variable-volume Kt/V. Kt/V = -ln(R - 0.008 * t) + (4 - 3.5 * R) * UF / W, with R = post_bun_mg_dl / pre_bun_mg_dl. Numeric Kt/V to 2 decimals. Not a dialysis prescription, not equilibrated eKt/V, not stdKt/V, and not a urea-reduction-ratio tool. magent does not assign KDOQI or other adequacy targets.

Lundell 1999 Table 1 Los Angeles classification. Caller-assigned flags. none if no mucosal break (not minimal changes); D if break(s) involve ≥75% of circumference; C if continuous between fold tops but <75%; B if longest break >5 mm not between fold tops; A if ≤5 mm not between fold tops. magent does not interpret images. Not Savary-Miller, GERD diagnosis, Barrett/stricture, or 1994 LA (D was fully circumferential). Not a medical device.

LACE (van Walraven 2010 CMAJ Table 3) from length of stay in days (<1 0; 1 1; 2 2; 3 3; 4-6 4; 7-13 5; >=14 7), acute (emergent) admission (true 3), caller-supplied Charlson comorbidity index (0-3 identity; >=4 5), and ED visits in the previous 6 months (0-3 identity; >=4 4). Score 0-19 only; no unpublished high-risk class. Not a discharge order. magent does not compute Charlson.