Completed PRISMA-P checklist, showing recommended items to include in a systematic review protocol, and where in the protocol each item is located, can be found in Appendix 1.
Eligibility criteria
Studies will be included that test the effectiveness of digital behavioural interventions in increasing the rate of flu vaccination amongst pregnant women. Comparators in included studies will be either usual care, a wait-list comparator, a historical control group where the digital intervention is not present, a digital intervention that is not about flu vaccination, or comparison to a non-digital intervention. The outcome being studied is the rate of flu vaccination amongst pregnant women. Only original research studies will be included, with any systematic reviews, protocols, commentaries and conference abstracts being excluded. Included studies will be either a Randomised Controlled Trial, a non-randomised controlled trial, quasi-randomised controlled trial or other type of quantitative study that reports the rate of flu vaccination (for example, before and after trials) following the implementation of a digital intervention, which also contains a comparator group. Quantitative studies such as case series, case studies and case reports will also be excluded. No date or country restrictions will be placed on the search, but studies will be required to be published in English.
Table 1: Inclusion and exclusion criteria
Characteristic
|
Inclusion criteria
|
Exclusion criteria
|
Population
|
Participants are pregnant women, over the age of 16 (or where there is a range of ages included, the majority of participants are over the age of 16).
|
Any participants other than pregnant women, or studies that include only adolescents (under the age of 16).
|
Intervention
|
Studies testing the effectiveness of a digital intervention (an intervention that attempts to change pregnant women’s vaccination behaviour that are delivered via a digital or mobile device directly to participants) to increase the rate of flu vaccination (if multiple types of intervention are tested, at least one of these needs to be a digital intervention, and results must allow for the rate of vaccination by a digital intervention to be extracted). Appropriate statistical information about the effectiveness must be provided
|
No intervention is tested, none of the tested interventions are digital
|
Comparator
|
Studies comparing the effectiveness of a digital intervention (for example, text message, website, mobile app) to usual care, to a wait-list comparator, to a non-digital intervention, to a digital intervention that is not about flu vaccination, or to a historical control group without digital intervention
|
No comparator, control or usual care condition is present.
|
Outcome
|
Outcome being studied is the rate of flu vaccination (either actual vaccination behaviour or intention to vaccinate)
|
The rate of flu vaccination is not the outcome measure
|
Publication type
|
Original research studies only
|
Systematic reviews, protocols, commentaries, conference abstracts
|
Study design
|
Studies will be RCTs, non-RCTs, quasi-RCTs or other quantitative study
|
Other study designs including quantitative studies that report audits, surveys and similar, and case series, case studies or case reports) or those that do not report the rate of flu vaccination after the implementation of a digital intervention.
|
Outcome measures
The outcome measure will be the rate of flu vaccination amongst pregnant women after receiving a specific digital intervention, compared to the comparator group. This will either be self-reported flu vaccination status or rate obtained from electronic patient records.
In studies where the main outcome is intention to vaccinate rather than actual vaccination behaviour (uptake of vaccination), this rate will be used for this review.
Information sources
Electronic bibliography databases will be searched during April and May 2020. MEDLINE, Embase, Web of Science, Scopus, Cochrane database, PsycINFO and Cochrane Central Register of Controlled Trials (CENTRAL) will be searched for all eligible studies. Clinical trial registers will be searched for in-progress trials (e.g. ClinicalTrials.Gov) Search terms will include all possible terms relating to ‘vaccination’, ‘influenza’, ‘pregnancy’ and variations of ‘digital interventions’ to include interventions that contain significant influence from text messages, video, internet or mobile phone applications (apps) (13,14). Boolean strategies of ‘AND’ and ‘OR’ will be employed.
Reference sections of included studies will also be screened by hand to identify any other eligible studies, and papers citing included studies will be searched.
Search strategy
Specialist advice on the search strategy has been sought in developing the search strategy for this review, to ensure searches are comprehensive and capture all relevant studies in the searches. An example of the full search strategy for one database can be found in appendix 2. Keywords will be searched in both titles and abstracts.
Data management and screening process
Results of searches will be managed using Endnote software. Results from all databases will be combined, alphabetised and duplicates will be removed.
The first stage of screening will consist of all titles and abstracts being screened. Any studies that appear to meet the inclusion criteria will be subject to stage two of screening, which will consist of full text screening, where full papers will be obtained and screened against the eligibility criteria. Screening will be conducted by two researchers independently, and any discrepancies will be resolved with discussion. If a consensus is not reached, advice will be sought by a third researcher, until a full and final set of studies that meet the inclusion criteria are obtained. In addition to this, reference sections of all included studies will be screened for any studies that were not captured by the searches.
Data will be extracted from all included studies. This will be conducted by two researchers independently, using a pre-defined extraction form. Eligibility for inclusion in the meta-analysis will also be determined. The following information will be extracted from each included study: Name of author, year of publication, study design, study setting, participants, details about intervention (including the mode of digital intervention such as text message, video, mobile phone app), comparison/ control condition, rate of flu vaccination, size of effect of intervention (if reported). Any discrepancies between data extraction carried out by the two researchers will be discussed, and a third author will be consulted if a consensus is not reached.
Quality assessment
Risk of bias in Randomised Controlled Trials will be assessed using the Cochrane Risk of Bias tool (15). Studies will be assessed on the domains below and classified as either low risk, medium risk or high risk of bias:
- Sequence generation
- Allocation concealment
- Blinding of participants, personnel and outcome assessors
- Incomplete outcome data
- Selective outcome reporting
- Other sources of bias.
Risk of bias in non-randomised controlled trials (including quasi-randomised controlled trials and quantitative studies) will be assessed using the Cochrane Risk of Bias in Non-randomised Studies of Interventions (16). Studies will be assessed on the domains below and classified as either low risk, moderate risk, serious risk, critical risk of bias or no information to make a judgement:
Pre-intervention
- Bias due to confounding
- Bias in selection of participants into the study
At intervention
- Bias in classification of intervention
Post intervention
- Bias due to deviation from intended intervention
- Bias due to missing data
- Bias in measurement of outcomes
- Bias in selection of the reported result
Data synthesis
Key information that is extracted from each included study will be synthesised. This will consist of descriptive information about the type and content of intervention in each condition as presented in the paper. The rate of flu vaccination will also be extracted from included studies and will be synthesised and discussed to determine if digital interventions are effective at increasing flu vaccination amongst pregnant women. Discussion about the risk of bias of included papers will also be included.
Data analysis
Statistical information about the rate of flu vaccination in pregnant women will be presented, to establish whether digital interventions are more effective at increasing flu vaccination amongst pregnant women, than other types of interventions or comparison groups.
If studies report the rate of intention to vaccinate instead of actual vaccination behaviour (vaccination uptake), this data will be reported instead. Separate analysis for vaccination and intention to vaccinate will be conducted. If studies report both vaccination rate and intervention to vaccinate rate, then the actual vaccination rate will be used.
If sufficient statistical information is provided by included studies, a meta-analysis will be conducted to establish the pooled and weighted size of the effect of digital interventions in increasing flu vaccination amongst pregnant women. Standardised study effect sizes, using standardised mean differences (with 95% confidence intervals) will be calculated for the increase of flu vaccination rates in digital intervention conditions, compared to other intervention types or control/comparison conditions, using a random-effects meta-analysis model, where the statistical information provided allows. It applicable, separate meta-analyses will be conducted for vaccination intention and actual vaccination uptake. In studies where insufficient statistical information is include, study authors will be contacted to ask for further information. If no further information is provided then the study will be excluded from the meta-analysis.
Heterogeneity
As it is anticipated that interventions within included studies will differ, random effects model will be used for the meta-analysis. Heterogeneity between studies will be assessed using the chi-squared statistic and I2, to examine what percentage of variability between studies is due to heterogeneity rather than chance. Where I2 is over 40% moderate heterogeneity between studies is indicated (17).
Subgroup analysis
If sufficient studies are included, and there is sufficient variation between the types of digital intervention used to try to increase flu vaccination amongst pregnant women, subgroup analysis will be performed to determine which type of digital intervention (for example, comparing text messages, mobile apps, websites) has the greater effect on increasing flu vaccination amongst pregnant women.
If there are sufficient studies, and sufficient variation between studies, additional sub-group analysis will be performed to determine the rate of self-reported flu vaccination verses validated flu vaccination (from medical records). This will be compared with the combined rate of flu vaccination, to determine if there is any difference between the rate depending on the way it is measured.
If there is sufficient variation in the risk of bias ratings of included studies, sensitivity analysis will be performed to determine whether results of the meta-analysis would change if high-risk studies were removed from the analysis.
Publication Bias
Funnel plots will be examined to identify the presence of publication bias. If missing studies as a result of publication bias are detected, a Trim and Fill analysis will be completed to account for the missing studies in the effect size (18).