GLOBALRPH ADVANCED VANCOMYCIN CALCULATOR CLINICIAN COMPETENCY CHECKLIST Version: 1.8 Reviewed: 2026-10-17 Use deidentified or constructed cases. Complete within the institution's training and review process. This checklist does not certify a clinician or establish external clinical validation of the calculator. Reviewer: _______________________ Date: _______________________ Participant: ____________________ Calculator version: _________ For each task, record: demonstrated / needs review / not applicable. 1. SCOPE AND WORKFLOW [ ] Identify the intended adult intermittent-infusion use and unsupported pregnancy, pediatric, unsupported dialysis or temporary mechanical circulatory support, and changing-physiology scenarios. [ ] Choose correctly among empiric, current-no-levels, single-level, two-level steady-state, and first-dose workflows. [ ] Explain why selecting a steady-state workflow does not prove steady state. [ ] Identify whether the final result used Bayesian or conventional estimation. [ ] Explain when automatic conventional fallback is permitted and why it cannot resolve invalid data, unstable renal function, or unsupported CRRT changes. 2. INPUTS AND TIMING [ ] Convert height, weight, and creatinine units accurately; verify normalized values after entry. [ ] Translate an actual two-hour infusion followed by a sample one hour later into a sample time of 3 hours from infusion start. [ ] Explain the regular-dose reconstruction: three doses q12h start at elapsed hours 0, 12, and 24; a sample three hours after the latest start is elapsed treatment hour 27. [ ] Identify a history that cannot be represented by identical doses and one interval, including unequal loading, held, partial, delayed, or extra doses. [ ] Distinguish collection time from laboratory result-release time. [ ] Investigate infusion-line sampling, samples during infusion, duplicate concentrations, and inadequate time separation. [ ] Explain why the 15-minute spacing screen is only a software safeguard. [ ] Recognize that first-dose estimation excludes unrecorded residual drug. 3. MODEL AND RENAL ESTIMATES [ ] Explain automatic model-selection order and manually select an eligible model for a constructed overlapping-population case. [ ] Describe why Belabbas is a hematologic-malignancy option rather than an automatic prior for every cancer diagnosis. [ ] Identify Roberts' indexed Cockcroft-Gault proxy for measured urinary CrCl and its continuous-infusion source population. [ ] Distinguish Zhang's indexed CKD-EPI units from unindexed CrCl. [ ] Explain that Tsai does not directly correct clearance for effluent intensity, urine output, downtime, or changing settings. [ ] Read uncapped and used renal values, their weight basis, and their limits. [ ] Demonstrate that the manual CrCl override and amputation adjustment concern the conventional pathway rather than all Bayesian models. 4. OUTPUT INTERPRETATION [ ] Identify a prior-only estimate and distinguish prior spread from posterior uncertainty informed by measured levels. [ ] Explain that one Bayesian level fits CL while central volume stays fixed; two levels can fit both except Tsai keeps volume fixed. [ ] Interpret a parameter boundary, convergence warning, and unavailable range; explain why a range crossing parameter limits is withheld rather than clipped. [ ] Explain why the approximate range excludes model disagreement, inaccurate history, future renal changes, and uncertainty in fixed parameters. [ ] State why computational QA does not establish clinical calibration. [ ] Separate steady-state AUCss24, first-dose AUC0-24, and AUC0-infinity. 5. REGIMEN AND MIC EXERCISES [ ] Calculate the arithmetic example: daily dose 2,000 mg with CL 4 L/h gives AUCss24 500 mg*h/L. This example is not a prescribing recommendation. [ ] Transform CL limits 3-5 L/h for that same fixed daily dose into AUC limits of 400 to approximately 667 mg*h/L; explain the reversed endpoints. [ ] Distinguish current-regimen, exact-target, practical candidate, and custom regimen output; explain why an unrounded dose is not a medication order. [ ] Change infusion duration alone and explain why peak/trough may change while steady-state AUC is unchanged if dose, interval, and CL remain fixed. [ ] Enter MIC 0.5 and MIC 2 in a constructed case and confirm that MIC changes the ratio rather than automatically changing the exposure target or dose. [ ] Review above-target exposure and explain why the warning is not a binary prediction of toxicity in an individual patient. [ ] Compare eligible priors using the same regimen, identify exclusions, and explain why choosing the most convenient AUC is not a model-selection rule. 6. CLINICAL RESPONSE AND DOCUMENTATION [ ] Work through a poor-fit case: actual administrations, collection times, contamination, renal/fluid change, CRRT change, sample information, and eligibility; explain when an additional concentration is useful. [ ] Choose a reassessment plan using clinical changes and local monitoring policy rather than an automatic every-three-or-five-day schedule. [ ] Generate a report and reconcile its method, inputs, elapsed times, warnings, uncertainty, and clinician-selected regimen. [ ] Explain that elapsed event times are not patient calendar dates and that report-generation time is not a dose or collection timestamp. [ ] Describe PHP server processing versus browser calculation and that the application does not save entered inputs; handle downloaded reports according to the institution's chosen destination and retention process; recognize that independent hosting and browser retention are not guaranteed. [ ] Complete a handoff that identifies the selected model, unresolved inputs, active regimen, next assessment, and responsible clinician. RECOMMENDED LOCAL CASE SET Include each retained prior; zero, one, and two levels; a general adult and elderly adult; BMI around 30 and 40; obesity plus critical illness; hematologic malignancy with AML/neutropenia; CrCl near 120 and 150; conventional manual CrCl; first-dose and steady-state pairs; clinician-selected regimen; MIC values other than 1; supported uninterrupted CRRT; and invalid timing/history/CRRT cases that must be rejected. Agree on expected results and tolerances before scoring. TRAINING OUTCOME Demonstrated tasks: ______ Tasks requiring review: ______ Unresolved questions / remediation: ________________________________________________________________________ ________________________________________________________________________ Follow-up reviewer/date: ________________________________________________ Further reading: globalrph-vanco-advanced-help.html and CLINICAL_REVIEW.txt. Their reference lists identify the guideline, five priors, first-dose study, method comparison, model-switch study, and workflow webinar used in this guide. V1.6 ADDITIONAL EXERCISES [ ] Match the calculated-case snapshot to the intended clinical record. [ ] Choose Edit clinical inputs, confirm that the earlier result is removed, change one input, and recalculate before using the revised result. [ ] Change a custom dose or fractional interval, then download a report; confirm that the report evaluates the currently displayed draft. [ ] Leave an active CRRT urine or downtime value blank and confirm rejection; explain why explicitly entered zero is different from missing information. [ ] Identify a candidate outside AUC 400-600 mg*h/L and explain why candidates inside that range rank before candidates outside it. [ ] Review infusion administration separately from exposure. Confirm the generated 1,000 mg reference duration is 1.75 h under the reviewed conservative default; preserve an actual historical 1 h infusion in the dose history. [ ] Explain why obesity-specific loading reference takes precedence in an obese ICU case, while the result remains a reference requiring review. [ ] Recognize proposed regimens withheld for unreliable fitting or computational operating-limit violations, and use diagnostic reports appropriately. [ ] Explain that Goti/Belabbas primary-table verification is incomplete and that repeated synthetic calculations cannot establish patient prediction accuracy. V1.8 ADDITIONAL EXERCISES [ ] Read 65.00, 80.000, 1.0000, and 2.50 as 65, 80, 1, and 2.5 in the calculated snapshot while confirming the editable entries and calculation values have not been rounded or silently changed. [ ] Select each infusion policy before calculation. Confirm 1000 mg uses 1.75 h under the conservative default and 1 h under the supplied bands. [ ] Test both 2.5 h and 3 h choices for a generated dose above 2000 mg. [ ] Preserve an actual historical infusion while changing the generated policy; confirm the recorded history and fitted PK parameters are unchanged. [ ] Explain why a dose-band duration can trigger an administration review even though it follows the chosen schedule, and check applicable local policy. [ ] Calculate average steady-state concentration as AUC24 / 24 and distinguish it from a measured concentration, peak, or trough. [ ] Open the equation reference from the app and identify the method actually used, its unit definitions, and a difference from the legacy Word/PDF sheet. [ ] Leave a required policy choice invalid and confirm a field-specific error.