Biomedical Science Learning Guide
Evidence-Based Practice in Biomedical Science
A practical introduction to evidence-based laboratory medicine: how to turn research into safer testing, better clinical decisions and measurable benefits for patients.
Key message: a technically accurate result is not automatically a useful result. Evidence-based practice asks whether the right test is being used for the right patient, whether the result changes a decision and whether that decision improves care.
What is evidence-based practice?
Evidence-based practice is a structured way of making professional decisions. It brings together three things: the best available research, the knowledge and judgement of experienced professionals, and the needs and circumstances of the patient and healthcare service.
Best available evidence
Relevant, critically appraised research rather than the first paper found or the newest publication.
Professional expertise
Scientific knowledge, laboratory experience, clinical judgement and an understanding of local processes.
Patient and service context
Clinical need, patient preferences, equality, safety, access, resources and the likely effect on the care pathway.
The classic description of evidence-based medicine emphasises careful and transparent use of the strongest relevant evidence when caring for an individual patient.2 Evidence-based laboratory medicine applies the same idea to the selection, evaluation, interpretation and use of laboratory investigations.1
Evidence-based practice is not “following research blindly”. Evidence may be incomplete, uncertain or not fully applicable to the local patient population. Professional judgement is needed to decide whether the evidence is trustworthy, relevant and practical.
Why evidence-based practice matters in biomedical science
Laboratory results influence diagnosis, treatment, monitoring, discharge and referral decisions. An inappropriate test can create harm through false reassurance, unnecessary investigation, delayed treatment, anxiety or avoidable cost. A well-chosen test, used in a clear pathway, can support faster and safer care.
Evidence-based practice is also a professional expectation in the United Kingdom. The HCPC Standards of Proficiency for Biomedical Scientists include engaging in evidence-based practice, evaluating research, monitoring quality and using outcomes to improve care.3 Current IBMS professional guidance similarly links critical review of scientific advances with service development, audit and patient benefit.4
What evidence-based practice looks like at the bench
- Checking whether a new assay is fit for its intended clinical purpose.
- Reviewing evidence before changing a reference interval or decision threshold.
- Investigating whether an analyser bias could alter clinical classification.
- Assessing whether urgent turnaround time changes patient management, not only whether it meets a laboratory target.
- Using an audit to identify unnecessary repeat testing and then measuring improvement.
- Explaining the limitations of a result so that it is interpreted in the correct clinical context.
What questions can a laboratory test answer?
A request should begin with a clinical question. In the framework described by Price, Bossuyt and Bruns, laboratory investigations commonly support four broad types of decision.1
1. Is the condition present?
The result contributes to a diagnosis or increases the probability that a condition is present. This is sometimes described as a rule-in purpose.
2. Can the condition be excluded?
The result reduces the probability sufficiently to support a safe rule-out decision when used in an appropriate pathway.
3. What is likely to happen?
A prognostic test estimates future risk, such as deterioration, recurrence or response.
4. How is the patient progressing?
Monitoring tests assess disease activity, treatment response, toxicity or physiological change over time.
Screening is another important use of laboratory testing. It involves testing people who do not have recognised symptoms and therefore requires a carefully evaluated programme, a defined target population and a useful next step for people with positive results.
Diagram 1. From clinical question to patient outcome
The test has value only as part of a complete decision pathway.
A useful request can often be written as a sentence: “For this patient group, will this test provide information that changes a decision and leads to a beneficial action?”
Five areas of evidence-based laboratory medicine
The Tietz chapter groups the information needed for evidence-based laboratory medicine into five connected areas.1 Together, they move the focus from “Can the analyser measure it?” to “Does using it improve care?”
| Area | Main question | Laboratory example |
|---|---|---|
| Diagnostic accuracy | How well does the index test identify or exclude the target condition compared with an appropriate reference standard? | Sensitivity and specificity of a biomarker in the intended clinical population. |
| Outcomes | Does using the test improve a patient, clinical or service outcome? | Whether faster testing shortens time to a safe clinical decision. |
| Systematic review | What does the total body of relevant evidence show? | Combining diagnostic accuracy studies after checking their quality and comparability. |
| Economic evaluation | Are the extra benefits worth the extra resources, and from whose perspective? | Considering staff time, repeat visits, downstream imaging and length of stay rather than reagent cost alone. |
| Audit | Is the test being used correctly and producing the expected benefits in routine practice? | Measuring adherence to a testing pathway, intervening and then re-auditing. |
Analytical performance is essential, but it is only the beginning
Before a result can help a patient, the method must produce reliable measurements. Precision, bias, measuring range, interference, carryover, detection capability, sample stability and traceability may all matter. However, excellent analytical performance does not prove diagnostic usefulness or patient benefit.
Analytical validity
Does the method measure the analyte reliably under defined conditions?
Clinical validity
Does the result distinguish or predict the target condition in the intended population?
Clinical utility
Does using the result change management and improve an outcome that matters?
The five-step evidence cycle: Ask, Acquire, Appraise, Apply and Assess
A practical way to use evidence is the five-step evidence cycle. The final step feeds back into the next question, so evidence-based practice becomes continuous improvement rather than a one-off literature search.
Diagram 2. The evidence-based laboratory practice cycle
Each step should be documented clearly enough for another person to understand the decision.
1. Ask a focused question
Vague questions produce vague searches. Start by defining the population, setting, test, comparator and outcome. For diagnostic accuracy, the PICTO structure used in current NICE methods guidance is useful:5
P – Population
Who will be tested, and in which setting? Include symptoms, disease spectrum and relevant exclusions.
I – Index test
Which new or existing laboratory test is being evaluated, including the method and threshold?
C – Comparator
Which reference standard or alternative test will be used for comparison?
T – Target condition
Which disease, disease stage or clinical state should the test identify or exclude?
O – Outcome
Which measure of diagnostic performance will answer the question, such as sensitivity, specificity or a likelihood ratio?
For questions about impact rather than accuracy, a PICO structure may be more suitable: population, intervention, comparator and outcome.
2. Acquire the evidence
Search for the highest-quality evidence that answers the specific question. Useful sources may include NICE guidance, systematic reviews, professional guidance, peer-reviewed diagnostic studies, manufacturer documentation and local laboratory data. A manufacturer’s instructions for use are essential for understanding claims and operation, but independent evidence is still needed when judging clinical value.
| Question | Useful evidence | Important caution |
|---|---|---|
| Does the assay perform analytically? | Well-designed precision, bias, comparison, interference, stability and detection-capability studies. | Performance must be assessed at clinically relevant concentrations and under intended conditions. |
| Does the test detect the target condition? | Diagnostic accuracy study in a representative patient group, using an appropriate reference standard. | Studies using obvious cases and healthy controls can exaggerate accuracy. |
| Does using the test improve outcomes? | Randomised or carefully controlled impact studies, supported by implementation evidence. | A change in a biomarker is not automatically a patient-centred outcome. |
| Is the pathway good value? | Economic evaluation using relevant costs and consequences. | A cheaper reagent may create greater downstream cost or poorer outcomes. |
| Does it work locally? | Verification or validation data, service evaluation, incidents, feedback and audit. | Local data cannot repair a fundamentally weak evidence base, but they can show whether implementation is safe and effective. |
3. Appraise the evidence
Critical appraisal asks three simple questions:
- Can I trust the study? Look for bias, missing data, inappropriate comparisons and weak methods.
- What are the results? Examine effect size, diagnostic accuracy, confidence intervals and uncertainty, not only the p value.
- Does it apply here? Compare the study population, prevalence, setting, sample handling, analyser, pathway and thresholds with local practice.
4. Apply the evidence
Application requires professional judgement. Consider whether the method is compatible with the local population, sample pathway, staffing, information systems, turnaround requirements and clinical protocol. Discuss changes with the multidisciplinary team and include the patient or service-user perspective where relevant.
5. Assess the result and improve
Define success before introducing the change. Measures might include analytical quality, sample rejection, time to result, time to decision, repeat requests, downstream testing, incidents, patient experience and cost across the whole pathway. Review the data, act on problems and re-audit.
Understanding diagnostic accuracy
A diagnostic accuracy study compares an index test with a reference standard used to establish the best available classification of the target condition. Results can be arranged in a two-by-two table.
| Index test result | Target condition present | Target condition absent |
|---|---|---|
| Positive | True positive (TP) The test is positive and the condition is present. |
False positive (FP) The test is positive but the condition is absent. |
| Negative | False negative (FN) The test is negative but the condition is present. |
True negative (TN) The test is negative and the condition is absent. |
Sensitivity
Of the people who truly have the condition, how many test positive?
Sensitivity = TP / (TP + FN)Specificity
Of the people who do not have the condition, how many test negative?
Specificity = TN / (TN + FP)Positive predictive value
When the test is positive, how often is the condition actually present?
PPV = TP / (TP + FP)Negative predictive value
When the test is negative, how often is the condition actually absent?
NPV = TN / (TN + FN)Positive likelihood ratio
How much more likely is a positive result in someone with the condition?
LR+ = sensitivity / (1 – specificity)Negative likelihood ratio
How much less likely is a negative result in someone with the condition?
LR- = (1 – sensitivity) / specificityDo not interpret these measures in isolation. Predictive values change with disease prevalence and pre-test probability. Sensitivity and specificity can also change with the patient spectrum, setting and selected threshold. A result that is helpful in specialist care may perform differently in screening or primary care.
Where does an ROC curve fit?
A receiver operating characteristic (ROC) curve shows the trade-off between sensitivity and specificity across different thresholds. The area under the curve summarises discrimination, but it does not identify the best clinical threshold on its own. The consequences of false positives and false negatives, and the intended use of the test, must also be considered.
How to appraise a diagnostic accuracy study
Ask the following questions before accepting the reported sensitivity, specificity or predictive values:
- Patient selection: were participants representative of people who would receive the test in practice, ideally recruited consecutively or randomly?
- Avoiding an artificial comparison: did the non-disease group include realistic alternative diagnoses rather than only healthy volunteers?
- Index test: was the method and threshold defined before the reference result was known?
- Reference standard: was it appropriate, applied consistently and interpreted without knowledge that could create bias?
- Flow and timing: did all participants receive both assessments within a suitable interval, and were withdrawals explained?
- Uncertain results: were indeterminate, invalid and missing results reported rather than quietly removed?
- Precision of estimates: were confidence intervals reported?
- Applicability: are the population, setting, specimen, analyser, threshold and pathway similar to the intended local use?
The current STARD checklist contains 30 essential reporting items intended to make diagnostic accuracy studies more complete and transparent.6 STARD helps readers see what was done, but it is not itself proof that a study has low risk of bias. QUADAS-2 is an established tool for judging risk of bias and applicability across patient selection, the index test, the reference standard, and flow and timing.7
Important distinction: good reporting makes appraisal possible; good study design makes the result more trustworthy. A clearly reported weak study is still a weak study.
Common mistakes when using laboratory evidence
Relying on one paper
A single study may be small, biased or inconsistent with the wider evidence.
Confusing statistical with clinical significance
A small difference can be statistically significant without changing a clinical decision.
Ignoring pre-analytical factors
Patient preparation, timing, transport, storage and sample quality may limit real-world performance.
Treating manufacturer claims as the whole evidence base
Claims must be understood, independently appraised and confirmed locally where required.
Assuming accuracy proves benefit
A test may classify disease accurately but fail to change management or improve outcomes.
Forgetting the denominator
Percentages are difficult to interpret without sample size, patient selection and confidence intervals.
Worked example: introducing a new high-sensitivity cardiac troponin assay
This example shows how a Biomedical Scientist could structure an evidence-based evaluation. It is an educational framework, not a clinical protocol; the actual pathway must follow current national guidance, manufacturer instructions and local governance.
1. Ask
In adults assessed for possible acute coronary syndrome in the emergency setting, will the proposed assay and associated pathway support safe, earlier decisions compared with the current service?
2. Acquire
Gather current clinical guidance, systematic reviews, diagnostic and outcomes studies, manufacturer documentation, relevant analytical standards, and local data on workload, timing and patient pathways.
3. Appraise
- Analytical performance near the concentrations used for clinical decisions.
- Imprecision, bias, interference, measuring range and lot-to-lot consistency.
- Diagnostic performance in a population similar to local patients.
- Sample timing and the effect of serial measurements.
- Effects on time to decision, admissions, repeat testing and safety outcomes.
- Uncertainty, conflicts of interest and limitations of the available studies.
4. Apply
Agree acceptance criteria before local verification. Confirm manufacturer claims under local conditions; perform additional validation when the intended use or process is modified. Update the clinical pathway, requesting rules, LIMS, units, interpretive comments, SOPs, training and risk assessment. ISO 15189:2022 provides the overarching quality and competence framework for medical laboratories.8 Current UK validation guidance stresses that local performance must match the intended patient population, workflow and clinical use.9
5. Assess
After implementation, monitor quality-control performance, sample issues, turnaround time, repeat measurements, time to clinical decision, incidents, complaints and unexpected effects. Compare the findings with the pre-agreed standard and re-audit after corrective action.
Why this example matters: installing a more sensitive assay without changing the decision pathway, communication and training may fail to deliver the expected benefit. The intervention is the whole test-and-action pathway, not merely the analyser reagent.
Clinical audit closes the evidence-to-practice gap
Research asks what should work. Audit asks whether agreed good practice is happening here. A basic audit cycle is:
- Choose the question. Focus on a clear problem, risk or improvement opportunity.
- Define the standard. Use an authoritative guideline, policy, specification or justified local benchmark.
- Measure current practice. Agree inclusion criteria, data fields and sample size before collecting data.
- Compare and explain. Identify the size of the gap and likely causes.
- Implement change. This may involve training, pathway redesign, electronic requesting controls or clearer reporting.
- Re-audit. Check whether the change was sustained and whether it produced the intended outcome.
Student-friendly audit example
A department suspects that a test is being repeated sooner than the locally agreed minimum interval. A student could help define the standard, collect anonymised request data under supervision, identify the clinical areas with the highest non-compliance, support an educational intervention and then compare the re-audit results. The value lies not only in calculating a percentage but in linking the finding to workload, sample collection, patient experience and the risk of unnecessary follow-up.
Reflective questions for CPD or training discussion
- Which laboratory test in your section has the greatest effect on an urgent clinical decision?
- What evidence supports its intended use and decision threshold?
- Which pre-analytical or analytical factor could weaken that evidence in local practice?
- What patient or service outcome would show that the test is genuinely useful?
- How could you audit that outcome without including identifiable patient information in your learning record?
For further learning, visit the LabPathPro Learning Centre, explore the Clinical Chemistry section, or review the guide to CPD for Biomedical Scientists.
Summary: the practical principles to remember
- Begin with a specific clinical or laboratory question.
- Choose evidence that matches the type of question being asked.
- Separate analytical performance, diagnostic accuracy and clinical utility.
- Check internal validity, external validity and risk of bias.
- Interpret sensitivity, specificity and predictive values in the correct population and setting.
- Consider the complete pathway: request, sample, result, decision, action and outcome.
- Combine research with professional expertise, patient needs and local constraints.
- Measure what happens after implementation and use audit to improve practice.
In one sentence: evidence-based laboratory medicine means using the strongest relevant evidence, together with professional judgement and patient context, to ensure that laboratory testing leads to better decisions and better outcomes.
Frequently asked questions
What is evidence-based practice in biomedical science?
It is the use of critically appraised research, professional expertise and patient or service context to make safe, transparent laboratory decisions. It applies to test selection, method evaluation, interpretation, service development and audit.
What is evidence-based laboratory medicine?
Evidence-based laboratory medicine is the application of evidence-based practice to laboratory investigations. Its purpose is to ensure that tests are reliable, appropriate and used in pathways that improve clinical or service outcomes.
Is a systematic review always the highest level of evidence?
Not automatically. The best design depends on the question, and a review is only as trustworthy as its methods and included studies. For example, analytical performance, diagnostic accuracy, patient impact and economic value require different forms of evidence.
What is the difference between diagnostic accuracy and clinical utility?
Diagnostic accuracy describes how well a test classifies the target condition. Clinical utility asks whether using the test changes management and produces a worthwhile outcome. A test can be accurate without being useful in a particular pathway.
What is the difference between validation and verification?
Verification confirms that an established method can meet specified performance claims under local conditions. Validation provides broader evidence that a method or modified use is fit for its intended purpose. The required approach depends on the device, intended use, modification and applicable governance.
How can a Biomedical Science student demonstrate evidence-based practice?
A student can formulate a focused question, search reputable sources, compare study methods, explain limitations, relate the evidence to a local SOP or pathway, discuss patient impact and propose an audit. Any portfolio or assessment submission must remain the student’s own work and follow university, IBMS and workplace rules.
References
- Price CP, Bossuyt PMM, Bruns DE. Introduction to clinical chemistry and evidence-based laboratory medicine. In: Burtis CA, Ashwood ER, Bruns DE, editors. Tietz fundamentals of clinical chemistry. 6th ed. St Louis (MO): Saunders Elsevier; 2008. p. 1-18.
- Sackett DL, Rosenberg WMC, Gray JAM, Haynes RB, Richardson WS. Evidence based medicine: what it is and what it isn’t. BMJ. 1996;312(7023):71-2. doi: 10.1136/bmj.312.7023.71.
- Health and Care Professions Council. Standards of proficiency: biomedical scientists [Internet]. London: HCPC; 2023 [cited 2026 Jul 22]. Available from: https://www.hcpc-uk.org/standards/standards-of-proficiency/biomedical-scientists/
- Institute of Biomedical Science. Good professional practice and conduct in biomedical science. Version 8 [Internet]. London: IBMS; 2025 [cited 2026 Jul 22]. Available from: https://www.ibms.org/asset/D65F08EF-B6A2-4D79-B699CEA706C55F62/
- National Institute for Health and Care Excellence. Developing NICE guidelines: the manual [Internet]. London: NICE; 2014 [updated 2025 Oct 23; cited 2026 Jul 22]. Available from: https://www.nice.org.uk/process/pmg20
- Bossuyt PM, Reitsma JB, Bruns DE, Gatsonis CA, Glasziou PP, Irwig L, et al. STARD 2015: an updated list of essential items for reporting diagnostic accuracy studies. BMJ. 2015;351:h5527. doi: 10.1136/bmj.h5527.
- Whiting PF, Rutjes AWS, Westwood ME, Mallett S, Deeks JJ, Reitsma JB, et al. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med. 2011;155(8):529-36. doi: 10.7326/0003-4819-155-8-201110180-00009.
- International Organization for Standardization. ISO 15189:2022 Medical laboratories – requirements for quality and competence. 4th ed. Geneva: ISO; 2022.
- Institute of Biomedical Science, Association for Laboratory Medicine, Medicines and Healthcare products Regulatory Agency. Validation guidance for medical laboratories [Internet]. London: IBMS; 2026 [cited 2026 Jul 22]. Available from: IBMS validation guidance for medical laboratories