In Papua New Guinea, 1500+ women die every year from childbirth-related causes – 80 times higher than in Australia. And these deaths are, mostly, preventable.
Background: Syndromic management of sexually transmissible infections is commonly used in resource-poor settings for the management of common STIs; abnormal vaginal discharge (AVD) flowcharts are used to identify and treat cervical infection including Neisseria gonorrhoea and Chlamydia trachomatis. A systematic review and meta-analysis was undertaken to measure the diagnostic test performance of AVD flowcharts, including both World Health Organization (WHO)- and locally-adapted AVD flowcharts. Methods: A systematic search of multiple electronic databases was conducted to locate eligible studies published between 1991 and 2014. Flowcharts were categorised into one of 14 types based on: 1) use of WHO guidelines or locally-adapted versions; 2) use of risk assessment, clinical examination or both; and 3) symptomatic entry. Summary diagnostic performance measures calculated included summary sensitivity, summary specificity and diagnostic odds ratio. Results: Thirty-six studies, including data on 99 flowcharts, were included in the review. Summary sensitivity estimates for WHO flowcharts ranged from 41.2 to 43.6%, and for locally adapted flowcharts from 39.5 to 74.8%. Locally adapted flowcharts performed slightly better than the WHO flowcharts. A difference in performance was not observed between use of risk assessment or clinical examination. The AVD flowchart performed slightly better when it was not restricted to symptomatic women only. Conclusions: There was considerable variation in the performance of the AVD flowchart but overall it was a poor diagnostic tool regardless of whether risk assessment or clinical examination was included, or whether the flowchart was WHO or locally developed. Many women were treated unnecessarily and many women with cervical infection were not detected. We caution against their continued use for management of cervical infection.
Part of this work was conducted while Caroline van Gemert was a Master of Public Health student at the University of Melbourne. Caroline van Gemert receives support through an Australian Government Research Training Program Scholarship. Freya Fowkes is supported by an Australian Research Council Future Fellowship. Margaret Hellard is supported by a National Health and Medical Research Council Principal Research Fellowship. The Burnet Institute gratefully acknowledge the contribution to this work of the Victorian Operational Infrastructure Support Program.