📋 Safety & Governance

Acid-Base Clinical Calculator
Safety & Governance

Tool: AcidBase Calculator
Engine: TheraIQ v2.1
Last updated: June 2026
This document describes the clinical positioning, intended use, safety design decisions, and known limitations of the Theracalc Acid-Base Clinical Calculator. It is intended for licensed clinicians, clinical pharmacists, and institutional reviewers evaluating the tool for clinical or educational use.
1 — Intended Use & Clinical Positioning
  • The Theracalc Acid-Base Calculator is a clinical decision support tool intended for use by licensed healthcare professionals — clinical pharmacists, physicians, nurse practitioners, and advanced practice clinicians
  • It is designed to assist with pattern recognition, formula verification, and stepwise acid-base interpretation at the bedside or during clinical review
  • It is not intended to replace clinical judgment, patient-specific assessment, physical examination, or current institutional protocols
  • It is not a diagnostic tool and does not constitute medical advice
Important — clinician responsibility All outputs — including diagnoses, fluid recommendations, infusion rates, and drug dosing — are estimates based on population-derived formulas. Individual patient responses may vary significantly. Always verify calculations independently and consult current evidence-based guidelines.
2 — Safety Language Design
  • Non-prescriptive language: all intervention outputs use guidance-level verbs ("consider," "may," "reassess") rather than prescriptive directives ("give," "order," "administer")
  • Explicit disclaimers: every output section is accompanied by a clinical decision support notice reminding clinicians that the tool is advisory only
  • Provisional flagging: when ABG is absent, all downstream interpretations are explicitly flagged as provisional pending confirmation
  • Context-sensitive suppression: clinical indices of low decision value in a given pattern (e.g., UAG in clear lactic acidosis) are demoted in the output hierarchy to reduce cognitive noise
  • No third-party AI at runtime: all logic executes locally via the TheraIQ engine — no external AI model generates or modifies recommendations at runtime
3 — Known Limitations
  • Population-based formulas: all compensation equations are derived from population studies and may not reflect individual physiology, particularly in chronic compensated states
  • No outcome validation: the tool has not undergone prospective clinical validation against patient outcomes
  • BMP-only limitation: without pH and PaCOâ‚‚, primary disorder classification is estimated from HCO₃ pattern alone — mixed disorders may be missed
  • Static reference ranges: normal ranges (AG cutoff 12, HCO₃ 22–26) are bedside standards and may differ from individual lab reference ranges
  • No renal-hepatic interaction modeling: the tool does not account for the effect of combined organ dysfunction on acid-base status
  • Chronic compensation: the tool distinguishes acute from chronic respiratory compensation but relies on user context for correct selection
Mixed disorder detection Triple disorder detection and complex mixed patterns are flagged but require careful clinical correlation. The engine identifies these patterns algorithmically; final interpretation requires clinician judgment and full clinical context.
4 — Bicarbonate Therapy Safety Guidance
  • The tool explicitly discourages routine bicarbonate infusion in lactic acidosis per current evidence (BICAR-ICU trial)
  • Bicarbonate is flagged as not recommended unless pH <7.10 with cardiovascular collapse or select specific indications
  • The tool explains the mechanism — bicarb generates COâ‚‚ acutely, worsening intracellular acidosis — to support clinician understanding rather than blind compliance
Reference Jaber S, et al. Sodium bicarbonate therapy for patients with severe metabolic acidaemia in the intensive care unit (BICAR-ICU). Lancet. 2018;392(10141):31–40.
5 — Fluid Recommendation Safety
  • Balanced crystalloid (LR) is recommended over normal saline for resuscitation in HAGMA — consistent with SMART and SALT-ED trial evidence
  • The tool explicitly addresses the "LR has lactate" misconception — LR does not meaningfully worsen serum lactate in lactic acidosis
  • Volume recommendations (e.g., 30 mL/kg) include explicit caveats that final fluid choice must follow hemodynamic status and local protocol
  • High-risk populations (CHF, CKD, advanced age) are flagged as warranting modified fluid approaches with earlier vasopressor consideration
6 — Lactate Trajectory Guidance
  • Serial lactate monitoring is explicitly recommended with defined trajectory targets
  • Target: ≥10% lactate clearance at 1 hour, ≥20% at 2 hours (Jansen et al, JAMA 2010)
  • Failure to clear is flagged with a differential: occult source, inadequate source control, impaired hepatic clearance, mitochondrial toxin, or persistent hypoperfusion
  • Escalation triggers (ICU, vasopressors) are explicitly stated when MAP and urine output targets are not met
Reference Jansen TC, et al. Early lactate-guided therapy in ICU patients. Am J Respir Crit Care Med. 2010;182(6):752–761.
7 — Regulatory & Institutional Considerations
  • Theracalc operates as a browser-based reference tool — all computation runs locally in the clinician's browser; no patient data is transmitted to Theracalc servers
  • The tool does not store, process, or transmit protected health information (PHI) and is not intended to function as a HIPAA business associate
  • Healthcare organizations should evaluate Theracalc under their own policies for browser-based clinical reference tools
  • The tool is classified by major security and threat-intelligence providers in non-malicious categories (Medicine, Information Technology)
For institutional IT and security teams Additional documentation including security questionnaire responses and network whitelisting support is available upon request at [email protected]