Cochrane Reviews are systematic reviews of research in health care and policy. They are internationally recognized as a high-quality source of evidence for decision-making. They collate and summarize all the best available research evidence on the effects of healthcare interventions or the accuracy of diagnostic tests into a systematic review.
Diagnostic Test Accuracy Reviews (DTA) provide essential information regarding the accuracy of the available diagnostic tools for key decision makers, including patients, clinicians, guideline developers and researchers.
The purpose of a Cochrane DTA review is very like the purpose of a conventional Cochrane systematic review. Using explicit, transparent, and systematic methods to pool together all the available evidence relevant to a DTA research question, we can obtain more precise estimates of the accuracy of diagnostic tests than single DTA studies alone can provide. Estimates of sensitivity and specificity may also be more generalizable than estimates from single studies.
Cochrane GNOC has recently published its DTA review: Cytology versus HPV testing for cervical cancer screening in the general population.
Among its many international collaborators, Cochrane GNOC Group works with key contributors who are interested in disseminating evidence about the accuracy tests for diagnosing cancer. We spoke with three contributors, Jo Morrison, Co-ordinating Editor of Cochrane GNOC, Kate Sanger, Head of Communications & Public Affairs, Jo's Cervical Cancer Trust and Nuala Livingstone, Editor of the Cochrane Editorial Unit, to discuss the importance of this Review:
How important was this Cochrane Review to produce?
Jo Morrison: The review compares the accuracy of two different methods of testing cervical smears in picking up pre-cancerous changes (known as high grade cervical intra-epithelial neoplasia – HG CIN). Both tests start with a smear test, taken by directly visualising the cervix and using a ‘broom’ or spatula to sweep cells off the cervix. The difference is in what is done with the test once it is in the pot. One test puts cells onto a slide so they can be examined under a microscope and is called cytology, often called a Pap test. The other test takes a sample of cells and checks for the presence of viruses that can lead to pre-cancerous changes in the cervix (HG CIN). These viruses are called high risk Human Papilloma Virus or HR HPV. The test for the virus looks for genetic material (either DNA or RNA depending on the test) using a process called the polymerase chain reaction (PCR) and does not rely on someone looking at cells down a microscope.
Kate Sanger: This review is timely, as in the UK it has been announced that testing for HPV as the primary cervical screening test is to be rolled out over the coming years. We are in a position where the implications of this change on the workforce, systems and indeed patients are at the top of the agenda. The aim of the screening programme is to reduce incidence and mortality from cervical cancer through identifying women who are more at risk of the disease and providing treatment where appropriate. The confirmation through this review that screening for HPV is more sensitive and therefore more effective at identifying cases of CIN 2 or greater is extremely welcome and positive.
What did they find?
Jo Morrison: This DTA review found 40 studies comparing the two tests in over 140,000 women who had a smear test. Testing a smear test for HR HPV is less likely to miss HG CIN, as it is a more sensitive test than cytology. However, if the test is positive there is a lower chance that there really is HG CIN, as it is a less specific test. For every 1000 women who had a smear test, 20 actually had HG CIN. Cytology (Pap test) would have correctly identified 12 of these women (but missed 8 women with HG CIN); HPV testing would have correctly identified 16 women (and missed 4). The test is therefore more sensitive and has fewer false negative tests than cytology. However, this increased sensitivity comes at a cost. The HPV test would have suggested that 101 women had HG CIN when in fact they did not, whereas only 29 women would have had an over-call (false positive) with cytology (Pap) testing. This would result in an increase in women needing to be referred to a clinic for colposcopy, were the HPV test used as a stand-alone test, causing increased anxiety for these women. This is because the test is less specific and has more false positive test results than cytology.
What does this mean for clinical practice?
Jo Morrison: For the UK, this will mean that we can change the cervical screening programme. Currently all smears are screened for cytology first and those with features of low grade dyskaryosis are ‘triaged’ using HPV testing – those with HR HPV detected are referred to colposcopy. And those with no HR HPV are returned to routine re-call. However, mathematically, it makes more sense to use the most sensitive test (HPV test) up-front and then use cytology on those found to have HR HPV. By introducing a second level test after HPV testing, fewer false positives will be generated and fewer women will require colposcopy than with HPV testing alone. However, this change will have major implications for the screening programme since: 1) more women may be referred to colposcopy; 2) overall many fewer smears will require cytological evaluation. This has repercussions for the current laboratory services, since many fewer cytopathology screeners will be required and the service could be streamlined to only a few laboratories providing this service across the UK (since there are minimum numbers of smears screened per screener to maintain expertise).
World-wide this has major benefits since HPV testing is much more easily deliverable in the developing world that cytological screening, which is where the heaviest burden of disease from cervical cancer lies.
Kate Sanger: For patients, there is a clear need for increased information provision regarding why the changes to the screening programme are beneficial and the different pathways women invited for screening may take depending on the outcome of their initial HPV test. For women testing negative for HPV it is essential that they fully understand the need to continue to attend screening when next invited; greater symptom awareness for all women must also be increased. It is especially important that women who test positive for high risk HPV, but do not have cell changes that require treatment, fully understand what it means, for example with women starting new relationships or who may have multiple sexual partners. This is essential to reduce uncertainty, distress and to ensure being ‘HPV positive’ does not get a negative stigma attached to it. Information and support need to be available to women at every stage with clinicians able to signpost women to external sources.
How robust is this evidence?
Nuala Livingstone: Overall, I would consider this evidence to be robust. The review authors have provided detailed, and carefully considered information in this review to assist the user in assessing the ‘robustness’ of the evidence.
Firstly, they assessed the methodological quality of the included studies using the QUADAS tool, as recommended by the Cochrane Diagnostic Test Accuracy Working Group. This assessment revealed that most the included studies in this this review were ‘low risk of bias’. Review authors also assessed the quality of the body of evidence for each outcome in the Cochrane review using the GRADE considerations. This assessment considers the methodological quality of the included studies (as assessed by QUADAS) along with the precision, consistency, completeness and applicability of the evidence available. Based on these factors, review authors judged the overall quality of the evidence in this review to range from moderate to high. Evidence was downgraded from ‘high’ to ‘moderate’ quality because the sensitivity values among included studies were slightly inconsistent, with values ranging from 52%-94% for Cytology, and 61% to 100% for HPV testing.
Kate Sanger: It would be interesting to see a breakdown of the research by age group, as women under 25 and over 65 were part of the study and in the UK these women are not invited for screening. Data on women over 65 could contribute to evidence as to whether there should be a consideration of increasing the screening age in the UK. With women under 25 further research into the impact of the HPV vaccination on incidence will be particularly important, especially following the potential introduction of Gardasil 9.
How clinically relevant is it, and how can the evidence be used?
Jo Morrison: This review is highly clinically relevant and affects all women in a screened population – ideally it should be relevant to all women world-wide, since universal cervical screening would be a major contributor to global health and the burden of disease is largely in less developed countries.
What would your message be to clinicians and policy-makers in this area or cervical screening?
Jo Morrison: HPV-testing has increased sensitivity, but less specific than cytology. Used as the up-front test in a two-stage screening test, it has the potential to reduce the number of false-negative tests, thereby reducing the number of ‘interval’ cancers, which are missed by cytologyical screening. There are also potentials for cost savings, although this review did not perform an economic evaluation.
What would your message be to patients?
Nuala Livingstone: The high sensitivity of this test should provide reassurance that a negative result on this test is highly likely to be accurate. This finding is supported by moderate to high quality evidence, which suggests we can be fairly certain in the result. However, a positive result will not necessarily prove to be true.
Jo Morrison: HPV testing is much less likely to miss changes than conventional cytology, although this may come at the price of more women needing referral to colposcopy for someone to look at their cervix and possibly perform biopsies. It is hugely reassuring to have a negative HPV test, since this shows that the risk of having CIN is very low. There is the possibility that the interval between smear tests could be extended, reducing the frequency that women need smears, which nobody enjoys! However, this would need to be tested in a different sort of study than a DTA study.
What further research would you like to see, and what would this tell us?
Jo Morrison: The results of studies comparing screening programmes using conventional cytological screening up-front versus HPV screening up-front are in progress and initial results demonstrate that this is likely to reduce the number of cancers missed by the screening test due to false negative smear results.
Kate Sanger: This review provides evidence which will hopefully increase activity to ensure the roll out of HPV primary in the UK is not delayed. It also provides a range of considerations that policy makers and clinicians need to address to ensure the physical and emotional health of the women in their population is protected. We must however not lose sight that across the UK attendance of screening is in worrying decline and addressing this, regardless of the screening test, must remain a priority.
Nuala Livingstone: The results of this review have provided detailed an important information regarding the statistical evidence available for this research question. However, I would like to see further research that provides information regarding the clinical importance of the evidence, and the long-term implications for women.
Jo Morrison, Nuala Livingstone, Kate Sanger.
Jo, Nuala and Kate’s words stated and expressed in this blog are entirely personal, and do not represent any official views or opinions of Cochrane.
CGNOC are committed to providing reliable evidence required to make important decisions on issues concerning Gynaecological cancers, Neuro-oncology and related topics.
CGNOC's main task is to facilitate the identification of reports of controlled clinical trials and to carry out systematic reviews which are then published on the Cochrane Library. The reviews are updated as new evidence is identified. This ensures that Cochrane reviews are relevant, useful, usable and used.
Evidence relating to the care of people with cancer comes from a number of sources and so a multi-faceted approach is taken, including searching electronic databases of clinical trial reports, handsearching printed journals, and scanning of the 'grey' literature i.e. conference proceedings and abstracts. Every effort is made to find unpublished trials to avoid bias.
f you would like more information about proposing a review title or getting involved with the Cochrane Gynaecological, Neuro-oncology and Orphan Cancer Group, please contact us:
For more information, please contact:
Jo’s Cervical Cancer Trust is the only UK charity dedicated to women affected by cervical cancer and cervical abnormalities. Our vision is a future where cervical cancer is a thing of the past. We want to see cervical cancer prevented and the impact for everyone affected by cervical abnormalities and cervical cancer reduced. We do this through providing the highest quality information and support, and campaigning for excellence in cervical cancer treatment and prevention. www.jostrust.org.uk
August 2017: The Cochrane Official Blog is curated and maintained by the Communications & External Affairs Department. To submit items for publication to the blog or to add comments to a blog, please email email@example.com.