Protocol: Systematic review of interventions initiated in response to disasters and national emergencies and their impact on short-and long-term diabetes outcomes

April 17, 2020

Anticipated start date 1/5/2020

Not yet started

Named contact: Jamie Hartmann-Boyce; jamie.hartmann-boyce@phc.ox.ac.uk


Review question

How do interventions enacted in response to disasters and national emergencies affect diabetes outcomes during and following these events?

Searches

Searches will be based on those run for: https://www.cebm.net/covid-19/managing-diabetes-during-the-covid-19-pandemic/ and https://www.cebm.net/covid-19/supporting-people-with-long-term-conditions-ltcs-during-national-emergencies/

Condition being studied

Diabetes (any type)

Participants/population

People with diabetes (no restriction on age or type of diabetes); (for qualitative studies only) carers for people with diabetes and healthcare providers

Interventions/exposures

Any intervention used to respond to national emergencies/disasters

Comparators/control

Studies will be included regardless of comparator group

Types of study to be included

There will be no restrictions on study design

Context

This review focusses on interventions implemented in response to disasters and national emergencies and on their impacts on in people with diabetes. The definition of disasters and national emergencies is broad but to be considered must impact systems for healthcare delivery, communication, physical activity, transportation, and/or food supply. This would include most natural disasters (e.g. large storms and floods, earthquakes) as well as new infectious disease epidemics/pandemics and civil and international conflict/unrest.

Main outcomes

Main outcomes include: HbA1c; diabetes-related hospital admissions; complications from diabetes; quality of life; and mortality.

Additional outcomes

Mental health indexes (including anxiety, depression, and diabetes-related distress); Diabetes self-efficacy; Severe hypoglycaemic events; Diabetic ketoacidosis; Cardiovascular outcomes (including AMI, stroke); Blood pressure; Hypertension; Barriers, facilitators, and patient experience as captured through qualitative data

Measures of effect

Risk ratios will be used for dichotomous outcomes. Mean differences, and where required (e.g. quality of life) standardized mean differences, will be used for continuous outcomes.

Data extraction, selection and coding

Two reviewers will independently screen studies at title/abstract and at full text stage using Covidence. Discrepancies will be resolved by discussion or through referral to a third reviewer. Data extraction will follow Cochrane methods (though this is not a Cochrane review) using a pre-specified and piloted data extraction form. Again, discrepancies will be resolved by discussion or referral to a third reviewer. Fields will include: study dates; study type; study funding source; authors conflicts of interest; country; setting; details of disaster/emergency; population characteristics (n; diabetes type; measures of diabetes control; age; gender; socioeconomic status; co-morbidities); intervention details; and outcomes as specified above

Risk of bias assessment

Risk of bias tool will depend on study type. The Cochrane risk of bias tool (ROB1) will be used for randomized controlled trials. The Cochrane tool for non-randomized studies of interventions will be used for studies reporting quantitative data that are not randomized ROBINS-I). The Joanna Briggs quality assessment tool will be used for qualitative studies.

Strategy for data synthesis

We anticipate that studies and data will be so heterogenous as to preclude statistical synthesis in most or possibly all cases. Where meta-analysis is not possible, quantitative data will be synthesised using recent Cochrane guidance on synthesis without meta-analysis. We anticipate use of effect direction plots as set out in the Cochrane Handbook though synthesis method will be to some extent guided by the nature and availability of the quantitative data. Where meta-analysis is deemed appropriate, we will use random effects models, calculating pooled risk ratios and 95% confidence intervals for dichotomous outcomes, and mean differences and standardised mean differences for continuous outcomes. Qualitative data will be synthesised using framework analysis.

Analysis of subgroups or subsets

Studies will be grouped by intervention type, by type of diabetes, and by socioeconomic status of participants

Dissemination plans

Results will be published on an ongoing basis and hosted on the CEBM website. Final results will be published in an academic journal. A short briefing document will be made for policy makers, including a GRADE summary of findings table. A lay summary will also be prepared and will be reviewed by people with diabetes.