Systematic Review of Bipolar Disorder: Data Extraction and Quality Appraisal, and Synthesis Strategy
Conducting a systematic review needs to be thorough and systematic in every sense as the name implies.
The author lists underneath the techniques he used to conduct a systematic review of pscyhosocial interventions for bipolar affective disorder to help other researchers design their strategies for future systematic reviews in this field.
Data Extraction strategy A detailed standard data extraction sheet was developed, to assist with systematic data collection on the following: 1) The study population 2) Criteria for study eligibility 2) Interventions carried out on treatment and control arms 3) Randomisation, blinding procedures and allocation concealment 4) Sample size 5) Outcome measures chosen and outcome data 6) Important clinical, and design features and follow up methods 7) Statistical analysis 8) Results 9) Drop outs and intention to treat analysis issues The extracted data were directly inserted in to data extraction tables in respective sections.
Data were categorised under population, intervention, treatment of controls and outcomes.
Data on quality of the studies were collated in a separate set of tables that will be discussed under quality assessment strategy.
Quality appraisal strategy Once the primary studies were selected, a more detailed quality assessment was carried out, to allow the reviewer to assess differences in the execution of studies within design categories.
This information was considered crucial for subsequent synthesis and interpretation of results.
Detailed quality assessments was based on a quality assessment instrument that was developed using Cochrane reviewer's handbook and CRD report 4 to match all assessment requirements of studies taking in to account the context.
The developed quality assessment instrument was a tabulated checklist of factors that needed to be assessed for each study.
The quality assessment instrument was derived taking in to consideration the factors that could bias review results.
The quality items within the instrument were assigned numerical scales up to 5, so numerical assessments of quality could be obtained.
Synthesis strategy Data synthesis involved collating and summarising results of the included primary studies.
The data extraction sheet was used to extract all necessary information to be inserted in the data extraction tables.
The tables were only numbered and the names of authors were not initially included in the table to help with minimising potential bias introduced by reviewer's knowledge of authors of the study, during quality appraisal process.
After completion of the quality appraisal process, the names of the authors were added to the tables for completeness of information.
Details of each step of the review process was reported explicitly with names of searched databases, number of citations, and the inclusion /exclusion process was reported with every necessary detail.
Details of excluded studies were reported in a separate list with explicit reasons for exclusion of each study.
Country of origin and context of each study were explained clearly.
The range of mean age of study participants, and range of gender of participants of the included studies were reported.
Extracted information from the studies (i.
e.
intervention, population, and context, sample sizes, outcomes, study quality) was tabulated in a manner consistent with the focused question of the review.
Tables were structured to highlight similarities and differences between study outcomes.
Data from studies were analysed in subgroups relevant to interventions that were carried out by the primary studies.
Information regarding interventions, participant populations, outcome measures that the trialists considered, was reported in subgroups of studies that applied similar interventions.
The reviewer endeavoured to provide a comparative account of the included studies rather than reporting critical appraisals of 10 included studies separately.
Quality of studies was reported after application of detailed quality checklist with descriptive information (ie, adequacy, inadequacy) and scores.
Total scores were reported and quality scores were cross tabulated to demonstrate the overall picture, and make comparison possible.
Findings of the included studies regarding effectiveness of the intervention under question was discussed collectively to provide an overall picture of impact of "psychosocial management" on course of bipolar disorder.
The reviewer put thorough effort to carry out a meta-analysis, with the view of reporting a quantitative account of effectiveness of psychosocial management of bipolar disorder.
Synthesis was descriptive - non-quantitative - for reasons of "feasibility of meta-analysis and examination of heterogeneity".
Explicit reasons were given why a meta-analysis was not carried out.
However, every effort was made whenever possible to complement the descriptive synthesis with a quantitative summary.
After summarising the results of the studies regarding effectiveness, implications of this review for future practice, policy and research was discussed.
Potential weaknesses of this review were clearly mentioned to warn the reader of the potential caveats.
The author lists underneath the techniques he used to conduct a systematic review of pscyhosocial interventions for bipolar affective disorder to help other researchers design their strategies for future systematic reviews in this field.
Data Extraction strategy A detailed standard data extraction sheet was developed, to assist with systematic data collection on the following: 1) The study population 2) Criteria for study eligibility 2) Interventions carried out on treatment and control arms 3) Randomisation, blinding procedures and allocation concealment 4) Sample size 5) Outcome measures chosen and outcome data 6) Important clinical, and design features and follow up methods 7) Statistical analysis 8) Results 9) Drop outs and intention to treat analysis issues The extracted data were directly inserted in to data extraction tables in respective sections.
Data were categorised under population, intervention, treatment of controls and outcomes.
Data on quality of the studies were collated in a separate set of tables that will be discussed under quality assessment strategy.
Quality appraisal strategy Once the primary studies were selected, a more detailed quality assessment was carried out, to allow the reviewer to assess differences in the execution of studies within design categories.
This information was considered crucial for subsequent synthesis and interpretation of results.
Detailed quality assessments was based on a quality assessment instrument that was developed using Cochrane reviewer's handbook and CRD report 4 to match all assessment requirements of studies taking in to account the context.
The developed quality assessment instrument was a tabulated checklist of factors that needed to be assessed for each study.
The quality assessment instrument was derived taking in to consideration the factors that could bias review results.
The quality items within the instrument were assigned numerical scales up to 5, so numerical assessments of quality could be obtained.
Synthesis strategy Data synthesis involved collating and summarising results of the included primary studies.
The data extraction sheet was used to extract all necessary information to be inserted in the data extraction tables.
The tables were only numbered and the names of authors were not initially included in the table to help with minimising potential bias introduced by reviewer's knowledge of authors of the study, during quality appraisal process.
After completion of the quality appraisal process, the names of the authors were added to the tables for completeness of information.
Details of each step of the review process was reported explicitly with names of searched databases, number of citations, and the inclusion /exclusion process was reported with every necessary detail.
Details of excluded studies were reported in a separate list with explicit reasons for exclusion of each study.
Country of origin and context of each study were explained clearly.
The range of mean age of study participants, and range of gender of participants of the included studies were reported.
Extracted information from the studies (i.
e.
intervention, population, and context, sample sizes, outcomes, study quality) was tabulated in a manner consistent with the focused question of the review.
Tables were structured to highlight similarities and differences between study outcomes.
Data from studies were analysed in subgroups relevant to interventions that were carried out by the primary studies.
Information regarding interventions, participant populations, outcome measures that the trialists considered, was reported in subgroups of studies that applied similar interventions.
The reviewer endeavoured to provide a comparative account of the included studies rather than reporting critical appraisals of 10 included studies separately.
Quality of studies was reported after application of detailed quality checklist with descriptive information (ie, adequacy, inadequacy) and scores.
Total scores were reported and quality scores were cross tabulated to demonstrate the overall picture, and make comparison possible.
Findings of the included studies regarding effectiveness of the intervention under question was discussed collectively to provide an overall picture of impact of "psychosocial management" on course of bipolar disorder.
The reviewer put thorough effort to carry out a meta-analysis, with the view of reporting a quantitative account of effectiveness of psychosocial management of bipolar disorder.
Synthesis was descriptive - non-quantitative - for reasons of "feasibility of meta-analysis and examination of heterogeneity".
Explicit reasons were given why a meta-analysis was not carried out.
However, every effort was made whenever possible to complement the descriptive synthesis with a quantitative summary.
After summarising the results of the studies regarding effectiveness, implications of this review for future practice, policy and research was discussed.
Potential weaknesses of this review were clearly mentioned to warn the reader of the potential caveats.