Skip to main content
Last updated: 27 July, 2023

Evidence-based medicine

Australian pharmacists obtain much of their clinical information from a range of trusted publications, such as Therapeutic Guidelines and the Australian Medicines Handbook, and bodies such as the Pharmaceutical Society of Australia, NPS MedicineWise, the National Health and Medical Research Council, and the Australian Government Department of Veterans’ Affairs. Information from these sources and from clinical guidelines issued by peak bodies has been critically appraised using the principles of evidence-based medicine (EBM). The information is presented in a manner that facilitates the application of EBM to individual patients.

EBM involves combining the best available scientific evidence with medical professional and patient preferences to make clinical decisions. EBM should be considered the integration of three elements:

  • the best available evidence
  • professional judgement to interpret the relationship between clinical trial participants and individuals in the pharmacist's care
  • the individual patient’s characteristics and relevant clinical history.

The following definitions and principles of EBM are outlined to assist pharmacists in the interpretation of clinical information.

Levels of evidence

The ‘level of evidence’ reflects the risk of bias in a particular study design that is used to answer a clinical question. Studies are graded according to a level-of-evidence hierarchy, from robust to progressively less robust designs. The National Health and Medical Research Council1 has ascribed the following hierarchy of levels of evidence to different study designs:

I a systematic review of two or more randomised controlled trials

II a randomised controlled trial

III-1 a pseudo-randomised controlled trial—alternate allocation or some other method

III-2 a comparative study with concurrent controls—for example, a cohort or case–control study

III-3 a comparative study without concurrent controls—for example, a comparison of two or more single-arm studies

IV a case series with either post-test or pre-test–post-test outcomes.

The evidence base is assessed in terms of the level and quality of the evidence. The quality of evidence depends on the study methodology and reflects how well a study was conducted to eliminate bias. This is influenced by how participants were selected, allocated to groups, managed and followed up, and how the study outcomes were measured.

Study design

Incorporating EBM into clinical care requires an understanding of the main research designs underlying the published evidence (see Study types and designs). Each study design has inherent strengths and limitations. Not all studies will be able to demonstrate the same types of findings. For example, randomised controlled trials (RCTs) can demonstrate causality between treatments and outcomes, and prove or disprove hypotheses, whereas observational studies cannot.

Although systematic reviews or RCTs, if well conducted, may provide high-level evidence, practitioners often need to draw on research designs lower in the evidence hierarchy when better-quality evidence is not available.

Systematic review and meta-analysis

A systematic review involves the systematic location and critical appraisal of multiple research studies, allowing evidence to be combined. Such reviews are distinct from traditional (descriptive) literature reviews in that they are based on a strict scientific design in an effort to minimise bias and ensure reliability of the findings. When studies deal with a similar study question and use similar methods, the results can be combined to calculate an overall estimate of effect (called 'meta-analysis'). The results of meta-analyses provide high-level evidence to inform practice.

Systematic reviews and meta-analyses are particularly useful when no large RCTs have been conducted but a number of small ones have. Systematic reviews and meta-analyses that use high-quality RCTs are considered to produce high-level evidence. However, not all published systematic reviews have used a reliable methodology, so findings can sometimes be misleading.2

Randomised controlled trial

An RCT is considered the gold-standard study design for measuring the effect of an intervention. Participants are randomly allocated to receive either an intervention or a control (e.g. placebo or standard treatment). Randomisation minimises the likelihood that the effects observed are the result of factors other than the intervention.

Clinical trials are designed to have one or more study end points. An end point is a measure that determines whether the intervention under study has had an effect (e.g. whether a tumour shrinks after treatment with chemotherapy).2,3,4 The study end point can be very important. Whereas RCTs are designed to investigate the primary end point, the study may investigate multiple secondary end points.

Using information from RCTs to inform clinical decision making about treatment options for an individual patient requires careful consideration of the people who were included in the RCT and how they compare with the patient. It is also important to critically appraise the quality of the study, including the study design (including sample size), end points and statistical analysis.

Observational study

Observational data from cohort or case–control studies can be necessary if it is not possible to compare participants in a treated group with a control group. Caution should be used when considering the results because of the potential for bias or confounding. Observational studies often reflect real-world practice and are relevant to the clinical setting.

Cohort study

A cohort study involves two groups (cohorts) of patients: one that is exposed to the variable of interest (e.g. smoking) and one that is not. The cohorts are followed to observe for the outcome of interest (e.g. lung cancer). Cohort studies are useful when trying to determine whether exposure to a particular factor is likely to cause specified events. Prospective cohort studies, which track participants forward in time, are more reliable than retrospective cohort studies.2,3,4

Case–control study

Case–control studies identify people with a disease or condition (cases) and people from the same population who do not have that disease or condition (controls), and look at how exposure to a suspect agent differed between the two groups. The exposure could be to a drug or other therapeutic intervention, an environmental factor or a behavioural factor. Case–control studies are retrospective.

A case–control study can be used to identify risks and trends, and to suggest possible causality for a disease or a particular outcome (e.g. an adverse drug effect). Such studies can only generate odds ratios, not relative risk ratios.2,3,4

Study types and designsStudy types and designs