Article type
Year
Abstract
Background: although various groups such as Cochrane Effective Practice and Organisation of Care (EPOC) provide guidance on selecting the best observational studies for inclusion in a systematic review, this guidance is typically focused on study design alone and does not address the potential of the study to address important gaps in randomized controlled trial (RCT) evidence.
Objectives: to describe the development of the seven-item Portland Observational Study Screening Tool of Interventions (POST-I) to prioritize the best observational studies and to describe our experiences using POST-I in a Cochrane Systematic Review on fludrocortisone for orthostatic hypotension.
Methods: our review team used the Cochrane Handbook of Systematic Reviews of Interventions (Version 5.1), 2013 CARE case report guidelines, and principles of rigorous speculation to inform the development of the POST-I. One review author used the POST-I to screen nine included observational studies and then discussed these assessments with other review authors.
Results: the POST-I (Figure 1) first assesses whether a study fills a gap in RCT evidence. If not, the study is low priority. If yes, the POST-I asks whether the study has a control group. If yes, the POST-I assesses whether study authors applied inclusion/exclusion criteria uniformly across groups; if groups were balanced at baseline, and if not, whether appropriate analyses were used to control for these imbalances. If yes, the study is high priority and if not, it is low priority. If the study does not have a control group, the POST-I asks whether the study was particularly careful at measuring events/outcomes, or the effect was unlikely to be due to confounding alone. If yes, the study is high priority and if not, it is low priority. In our review, using the POST-I helped to prioritize two out of nine observational studies. Both studies examined long-term harms of fludrocortisone therapy, one discussed use of fludrocortisone in a rare population, and one measured long-term effects of fludrocortisone on blood pressure. We highlighted findings from these two studies in our review and only briefly described the other seven observational studies.
Conclusions: the POST-I can be used as a starting point for review authors seeking a transparent and efficient method of prioritizing the best observational studies for a systematic review. We plan to further refine the tool through conversations with methodological experts outside our review team.
Patient or healthcare consumer involvement: the POST-I can help review authors identify the best evidence of the effects of interventions on additional populations and outcomes not previously included in systematic reviews, which has the potential to improve evidence-based practice and patient outcomes.
Objectives: to describe the development of the seven-item Portland Observational Study Screening Tool of Interventions (POST-I) to prioritize the best observational studies and to describe our experiences using POST-I in a Cochrane Systematic Review on fludrocortisone for orthostatic hypotension.
Methods: our review team used the Cochrane Handbook of Systematic Reviews of Interventions (Version 5.1), 2013 CARE case report guidelines, and principles of rigorous speculation to inform the development of the POST-I. One review author used the POST-I to screen nine included observational studies and then discussed these assessments with other review authors.
Results: the POST-I (Figure 1) first assesses whether a study fills a gap in RCT evidence. If not, the study is low priority. If yes, the POST-I asks whether the study has a control group. If yes, the POST-I assesses whether study authors applied inclusion/exclusion criteria uniformly across groups; if groups were balanced at baseline, and if not, whether appropriate analyses were used to control for these imbalances. If yes, the study is high priority and if not, it is low priority. If the study does not have a control group, the POST-I asks whether the study was particularly careful at measuring events/outcomes, or the effect was unlikely to be due to confounding alone. If yes, the study is high priority and if not, it is low priority. In our review, using the POST-I helped to prioritize two out of nine observational studies. Both studies examined long-term harms of fludrocortisone therapy, one discussed use of fludrocortisone in a rare population, and one measured long-term effects of fludrocortisone on blood pressure. We highlighted findings from these two studies in our review and only briefly described the other seven observational studies.
Conclusions: the POST-I can be used as a starting point for review authors seeking a transparent and efficient method of prioritizing the best observational studies for a systematic review. We plan to further refine the tool through conversations with methodological experts outside our review team.
Patient or healthcare consumer involvement: the POST-I can help review authors identify the best evidence of the effects of interventions on additional populations and outcomes not previously included in systematic reviews, which has the potential to improve evidence-based practice and patient outcomes.
PDF