Clinical Governance — Known Limitations
This page is a frank and complete disclosure of what OpenBook Clinical can and cannot do. Read this before using OpenBook Clinical in any clinical context.
Professional Obligations — Read This First
Using this tool does not discharge your professional obligations under AHPRA registration standards. You remain solely responsible for all clinical decisions made in your practice. OpenBook Clinical is a reference and decision support tool — it cannot assess your patient, cannot apply professional judgement, and cannot be held accountable for clinical outcomes.
A clinical decision support tool for evidence retrieval and synthesis
OpenBook Clinical searches academic databases in real time and synthesises peer-reviewed literature to help you find relevant research faster.
A tool to help AHPRA-registered clinicians find relevant research faster
Designed specifically for Australian allied health professionals registered with AHPRA. Evidence is graded against the NHMRC evidence hierarchy with preference for Australian guidelines and context.
A supplement to clinical reasoning — not a replacement
Evidence synthesis is one input into a clinical decision. Your professional training, direct assessment of your patient, and application of clinical judgement are not substitutable by any AI tool.
Not a diagnostic tool
OpenBook Clinical cannot diagnose conditions. It does not have access to your patient's clinical information, examination findings, imaging, or test results. Any synthesis output must not be interpreted as a diagnosis.
Not a treatment prescriber
OpenBook Clinical does not prescribe treatments or recommend specific interventions for specific patients. Evidence summaries describe what the research literature says — applying that evidence to an individual patient requires professional clinical judgement.
Not a replacement for clinical supervision
Evidence retrieval does not substitute for the supervision, mentoring, and professional oversight that are required components of safe clinical practice, particularly for early-career clinicians.
Not a clinically validated tool against patient outcomes
OpenBook Clinical has not been clinically validated against patient outcomes. We have not published or completed a formal study demonstrating that its use improves clinical outcomes, reduces adverse events, or changes practitioner behaviour in measurably beneficial ways.
Not endorsed by AHPRA, any professional college, or university
OpenBook Clinical is not endorsed by the Australian Health Practitioner Regulation Agency (AHPRA), by any AHPRA-regulated professional college (including the APA, SPA, ESSA, DAA, or AASW), or by any Australian university. Any reference to these organisations in our content is for context only.
AI hallucination — citations may be incorrect
The AI may state incorrect facts confidently. It may produce plausible-sounding citations that reference papers that do not exist, have incorrect metadata (wrong author, year, or title), or misattribute findings to the wrong paper. Always verify citations independently before including them in clinical documentation.
Database gaps — not all journals are indexed
Our sources cover major academic databases, but many specialist, niche, or paywalled journals are not indexed. Grey literature, conference proceedings, theses, and non-English publications are systematically underrepresented. A negative result (no papers found) does not mean no evidence exists.
Evidence currency — indexing lag
Papers published in the last 1–6 weeks may not be indexed. Very recent publications (especially those published in the last 4 weeks) are likely absent from search results. If the most current literature is critical, supplement with a direct PubMed or database search.
Synthesis quality — AI may oversimplify or miss nuance
AI synthesis of complex clinical topics may oversimplify findings, miss important clinical nuance, conflate different patient populations, or fail to adequately communicate uncertainty or heterogeneity in the evidence base. Treat synthesis outputs as a starting point, not a final summary.
Study design classification — automated NHMRC grading
Automated NHMRC evidence grading is performed using heuristic analysis of abstracts and metadata. This can misclassify non-standard study designs, adaptive trials, N-of-1 studies, pragmatic trials, and papers with atypical terminology. An incorrect evidence grade is possible and should be verified for critical citations.
Your Professional Obligations Remain
Using this tool does not discharge your professional obligations under AHPRA registration standards. You remain responsible for clinical decisions made in your practice.
Specifically, you are obligated to apply your professional training, maintain competence in your scope of practice, seek supervision where appropriate, and make clinical decisions that account for the individual circumstances of your patient. No AI tool, however well-designed, can substitute for these obligations.