This protocol is developed in accordance with PRISMA-P [Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocols] guidelines (15) (see Additional file 1). In accordance with the guidelines, the umbrella review is registered in PROSPERO (International Prospective Register of Systematic Reviews; registration number: ) in which any important protocol amendments will be noted.
Search strategy
We plan to search in four electronic databases: Embase (1974 to present), PsycINFO (1967 to present), PubMed (1966 to present) and Cochrane Database of Systematic Reviews (1992 to present). Furthermore, reference lists from included papers will be manually screened for further eligible papers.
Four domains will be used in the following search: I. psychiatric disorder, II. prevalence, III. diabetes and IV. systematic review. The search terms that capture the different psychiatric disorders will be taken from the DSM-III, DSM-IV-TR (axis I and axis II) (16, 17), DSM5 (section II) (18) and the ICD-10 Classification of mental and behavioral disorders (F00-F99) (19). The four domains will be included in a four-step search combining relevant search words and MeSH terms with Boolean Logic operators (OR and AND). Search terms within each domain are combined with the operator “OR” and the different domains are combined with the operator “AND”. The search strategy has initially been targeted PubMed and will subsequently be adapted for the other databases. Additional file 2 shows the complete search string that will be used.
Eligibility criteria
In the umbrella review, we will include systematic reviews focusing on cross-sectional and longitudinal observational studies in humans. The following eligibility criteria will be used:
- A systematic review (with or without a meta-analysis), describing a systematic search string and eligibility criteria. Posters and abstracts describing a systematic review are not included.
- Focus on an adult population (≥18 years) with one or more psychiatric disorders (measured by e.g. diagnostic interviews, hospital records, medical prescription or self-reported measures). Psychiatric disorders include psychiatric diagnoses mentioned in the ICD- or DSM classifications or elevated levels of clusters of psychiatric symptoms (DSM-III or DSM-IV, axis I or axis II; DSM-5 section II; ICD-10, F00-F99). Studies focusing on psychotropic medication that describe the prevalence of type 2 diabetes are also included. Studies that focus on a single psychiatric symptom and/or distress not described in the ICD- or DSM classifications will not be included in the umbrella review (e.g. work-related stress and short sleep duration).
- Measure of prevalence of type 2 diabetes (measured by e.g. diagnosis, medical reports, medical prescription or self-reported measurement). If no information is available regarding subtypes of diabetes, we expect that the majority will have type 2 diabetes and include these studies.
Only publications in English, a Scandinavian language, German or Dutch will be included. No exclusion criteria related to publication date will be set.
After removing duplicates, two independent reviewers will screen titles and abstracts to identity potential eligible papers. If at least one of the reviewers qualifies the systematic review as potentially eligible after reading the abstract, the full text will be retrieved and assessed according to the eligibility criteria. If there are any disagreements after the full-text screening, the two reviewers will discuss until consensus is reached. If necessary, a third reviewer will be involved in the discussions. Reference lists of the included publications will be manually screened with the aim to identify additional eligible publications. A PRISMA flowchart will describe the selection process. The screening of studies will be handled in EndNote and Covidence (20).
Data extraction and data-synthesis
The data from the included systematic reviews will be extracted using a data extraction form that has been developed a priori. From each of the eligible systematic reviews the following information will be extracted: first author, year of publication, country, type of study, type of assessment of psychiatric disorder(s), type of assessment of type 2 diabetes, the number of studies included in the systematic review, total number of participants, and primary findings, including information on the prevalence of type 2 diabetes, odds ratio´s and relative risk ratio’s, when possible. Furthermore, the number of cases and controls in each of the primary studies included in systematic reviews will be extracted. If the number of participants in the primary studies is not reported in the systematic reviews, the information will be retrieved from the papers describing these primary studies.
All systematic reviews will be summarized, and a narrative data-synthesis will be performed and presented for each of the psychiatric disorders identified in the umbrella review. Additionally, to summarize the reported prevalence of type 2 diabetes, we plan to conduct a meta-analysis of prevalence (a random effects model and 95 % confidence interval) for each of the psychiatric disorders (21). Our meta-analyses will be based on information from the primary studies included in the systematic reviews to make sure that the primary studies will only be included once (22). We plan to assess between-study heterogeneity (quantified as I2 metric) (23, 24) and publication bias (small study effects, Egger´s test) (25).
Risk of bias assessment
The strength of an umbrella review depends on the quality of the included systematic reviews. Therefore, it is important to assess the risk of bias in the included systematic reviews. Risk of bias assessment will be performed with the Risk of Bias in Systematic Reviews (ROBIS) tool (26).