Study design:
A systematic review will be performed using guidelines from the Cochrane Handbook on Systematic Reviews for Intervention, and described according to the Preferred Reporting Items for Systematic Reviews and Meta-analysis Protocol (PRISMA-P) guidelines (19, 20). If adequate homogeneity is observed, a Meta Analysis of the data will be performed.
Study registration:
In accordance with PRISMA-P guidelines, our systematic review was registered with the International Prospective Register of Systematic reviews (PROSPERO) on September the 19th 2018 with registration number: CRD42018105973.
Data source and search methods:
The search strategy will be developed and performed by an experienced information specialist and peer-reviewed by a second research librarian (21). The strategy will combine subject headings (e.g., MeSH), title and abstract terms for concepts related to ultrasound, point-of-care testing and training programs. We will search the following electronic databases: Ovid MEDLINE (1946-), Ovid Embase (1974-), Cochrane Library (inception-), CINAHL via EBSCOhost (1937-), Ovid ERIC (1965-) and Science Citation Index via Web of Science (1900-). We will also search the ClinicalTrials.gov register and Conference Proceedings Citation Index (1900-). All reference lists from relevant systematic reviews identified by the search will also be screened. Finally, we will hand search conference proceedings from the past two years from CHEST, Canadian Critical Care Forum, European Society of Intensive Care Medicine, and Society of Critical Care Medicine. No language or date limits will be applied. Search results will be exported in RIS format to Mendeley reference manager (by Mendeley Ltd, the Relx™ group, Elsevier) for primary screening. Duplicates were removed prior to screening. See appendix for the MEDLINE strategy.
Inclusion criteria:
We will include full text publications describing POCUS curriculum implementation, assessment, and evaluation with the following characteristics: retrospective, prospective observational and interventional studies, systematic reviews, and meta-analysis examining POCUS to detect acute medical and surgical conditions including cardiac, pulmonary, abdominal, and venous pathology. For the purposes of generalization, we are seeking to examine common POCUS applications across acute care medical specialties: cardiac, pulmonary, abdominal, and venous pathology and the ultrasound-guided procedures of thoracentesis and paracentesis. While there are myriad applications of POCUS, we believe examining these “core” modalities will yield critical details and be comprehensive, but ensure this review remains technically feasible. For the same reason, we will limit procedural POCUS to thoracentesis and paracentesis.
Healthcare professionals will include undergraduate medical students, postgraduate trainees (residents, registrars) and attending physicians in the acute, inpatient setting: Critical care, emergency medicine, anesthesiology, pediatrics, surgical and internal medicine.
Exclusion criteria:
We will exclude studies examining POCUS curriculum implementation, assessment, and evaluation with the following characteristics: in the outpatient setting, those which do not report the outcomes of participant assessment, if designed to exclusively validate a training simulator, studies involving a training program whereby the emphasis is already on diagnostic imaging modalities like cardiology or radiology, and those exploring non-POCUS modalities.
We will also exclude those studies evaluating ultrasound-guided procedures other than thoracentesis and paracentesis, and where the training population includes physician assistants, nurse practitioners, nurses, paramedics, or other allied healthcare providers. Primary literature published as abstracts only, will be excluded.
Screening and study selection:
Screening and study selection will occur in two phases. The first phase will be conducted by two independent authors, through screening titles and abstracts identified in the primary search. Articles screened for possible inclusion, will be tracked using Mendeley reference manager (by Mendeley Ltd, the Relx™ group, Elsevier). Disagreements will be resolved through discussion and if unresolved, a third party will be involved.
The second phase will involve full text screening. Articles identified as eligible during the initial screening, will be retrieved, and independently assessed by the same two authors. Full text screening will be done using the Covidence online platform (Covidence.org, Melbourne, Australia – a Cochrane technology platform). Disagreement will be resolved through discussion and if unresolved, a third party will be involved. To keep the literature review relevant and up to date, the search strategy and screening could be repeated should enough time lapse prior to the conclusion of the Systematic Review.
Methodological quality assessment and data extraction:
At this stage, we anticipate that a 3rd reviewer will be added to manage the expected volume of included full text articles as we transition to methodological quality assessment and data extraction.
Data extraction tools will be designed, using standardized forms, which will be divided into 2 major extraction categories: Quality assessment and general extraction with assessment validity. Data extraction will be done by 3 independent reviewers after calibration. Calibration will be accomplished by individually extracting data on the first 10% of studies in each category, followed by reviewing the same 10% as a group. Differences will be discussed and if unresolved, a 4th reviewer will be consulted. Once adequate correlation is achieved, data extraction will continue independently.
Quality assessment will be undertaken by applying the Medical Education Research Study Quality Instrument (MERSQI) score, specifically designed for quality assessment of educational studies (22). Studies will be scored using the MERSQI score categories: Study design, sampling, data type, validity of evaluation instrument, data analysis and finally outcomes. Scores range from 5–18, with 18 being the highest possible (22). Each study will be scored independently between 5 and 18.
Further data extraction, with focus on the learner population, number of participants, setting, POCUS application, methods of instruction, duration of intervention, methods of assessment, and program evaluation will be extracted using standardized forms. The evaluation of learning will be categorized according to the four levels of the Kirkpatrick model (22): Reaction, learning, behavioural change, and organization performance. Additionally, assessment validity data, using the key elements in the validity argument of Kane as our framework (23) will be extracted. Kane’s framework is founded on arguments for validity regarding: Scoring, generalization, extrapolation, and implication of assessment tools employed.
If additional educationally significant data elements are identified during our review, they will be added on a post hoc fashion, given the expected heterogeneity of studies. Extraction of quantitative and qualitative data will be done by a single reviewer with verification by a second. Disagreement will be resolved through discussion. In the case of unresolved matters, a third party will be involved.
Data synthesis and analysis:
The primary analysis of the qualitative data will be a narrative synthesis with an emphasis on the learner population, setting, POCUS application, methods of instruction, duration of intervention, methods of assessment and program evaluation. The curriculum evaluation and outcomes will be categorized according to the four levels of the Kirkpatrick model (22): Reaction, learning, behavioural change, and organization performance. Where available, we will report on assessment validity employing the key elements from Kane’s framework for validity arguments on scoring, generalization, extrapolation, and implication of assessment. Graphs and tables will be used as necessary in the qualitative analysis.
Studies will, where possible, be pooled in analysis by using 2x2 tables constructed from MERSQI quality score, population, duration of intervention, methods of instruction, methods of assessment, outcomes, Kirkpatrick level and validity arguments. We will use random effects model to pool effect sizes for exposures amenable to analysis, as it associated with the primary outcomes. Dichotomous outcomes will be reported as pooled odds ratios and 95% CI’s based on the DerSimonian-Laird random-effects model. Continuous outcomes will be standardized using calculated weighed standardized mean differences with the 95% Cis. Using the I2 statistics and chi squared test, we will assess the heterogeneity among studies. The thresholds for interpretation of the I2 will be in accordance with the definitions presented in the Cochrane Handbook for Systematic Reviews of Interventions. The presence of publication bias will be explored using funnel plots.