Automated systematic reviews

Systematic reviews are considered the highest-level evidence to underpin clinical decisions, guidelines, and policy decisions because they synthesise all of the existing evidence that answers a specific question, using methods that are transparent and reproducible.

Systematic reviews include the following steps:

  1. Preparation: formulate the question to be answered, write protocol
  2. Retrieval of evidence: design search strategies, run the searches, retrieve results
  3. Appraisal of evidence: screen the evidence for relevance to the question, include only the pertinent evidence
  4. Synthesis of the evidence: extract from relevant studies the information that answers the question, assess the quality (“risk of bias”) of the relevant studies, pool (“meta-analyse”) the evidence together
  5. Report: write up the results of the systematic review

As a team, we offer expertise in all areas of systematic reviews, including:

  • development of methodology for conducting systematic reviews, including quality assessment tools and reporting standards (e.g. GRADE, CONSORT, STARD, AMSTAR, etc.);
  • development of Systematic Review Automation tools to accelerate the process of conducting a systematic review (SRA helper, de-duplicator, word frequency analyser, Polyglot Search Syntax Translator, etc.); 
  • methodological advice on systematic reviews;
  • preparing protocols, and conducting systematic reviews for a wide range of bodies (e.g. NHMRC, MSAC, PBAC, GIN, WHO, Cochrane);
  • preparing and reviewing clinical practice guidelines for a variety of agencies both domestically and internationally (e.g. NHMRC, AHRQ, NICE).

Our team includes: Paul Glasziou, Tammy Hoffmann, Chris Del Mar, Elaine Beller, David Henry, Jenny Doust, Rae Thomas, Sharon Sanders, Anna Scott, Ray Moynihan, Justin Clark, Alex Bannach-Brown, Paulina Stehlik, Rebecca Sims, Loai Albarqouni, and Mina Bakhit.

Case studies

Systematic reviews are time- and resource-intensive to produce, taking on average 45 weeks to complete, and typically requiring 5 or more personnel. We recently developed and successfully piloted a process for significantly accelerating this timeline by integrating the use of Systematic Review Automation tools with the expertise of 4 experienced systematic reviewers (two researcher clinicians, an information specialist, an epidemiologist) to complete a full systematic review in less than 2 weeks.

In 2016, our team was commissioned by the WHO to conduct a systematic review of evidence of the impact of the use of antimicrobials in food animals, on both food animals and humans. This systematic review of evidence involved: scoping of the review, preparing the systematic review protocol, coordinating with the WHO to reach consensus on the proposed methodology, conducting literature searches, screening of the identified literature, extracting data, appraising evidence quality, evidence synthesis, preparing the final report documenting the processes and findings, and two presentations of findings at the WHO meeting in October 2016. This systematic review underpinned the resulting guidelines, “WHO guideline on use of medically important antimicrobials in food-producing animals,” which was published in November 2017. (WHO guidelines on use of medically important antimicrobials in food-producing animals. Geneva: World Health Organization; 2017

Cochrane reviews are internationally recognised as the gold standard for high quality information about healthcare. In the past 3 years, staff have collectively authored nearly 100 Cochrane protocols and subsequent systematic reviews and overviews of reviews. These have included reviews such as antibiotics for acute otitis media and antibiotics for acute pharyngitis which have both led to important changes to guidelines worldwide.    

In 2015-16 our team was contracted by the NHF to evaluate evidence relating to the diagnosis, monitoring and management of hypertension in adults. We conducted this evidence evaluation, which comprised the identification and appraisal of systematic reviews and primary studies relating to 8 clinical questions. This evaluation underpinned the development of the NHF Guideline for the diagnosis and management of hypertension in adults, published in 2016.