One systematic literature search was undertaken to cover all review questions. The search was adapted from that used in the USPSTF review (21) and used terms relating to diabetes and intermediate hyperglycaemia, interventions for these conditions and population screening. The search was limited to RCTs and systematic reviews of RCTs. A data limit of 2019 was applied to the search for RCTs to identify studies published since the USPSTF review (21). No date limit was applied to the search for systematic reviews. The search was conducted in the MEDLINE and Cochrane (reviews, trials and protocol) databases.
Search terms
English
Search terms and combinations in English were used for a search carried out on 21 June 2022 in MEDLINE (Ovid MEDLINE ALL <1946 to 20 June 2022» (Table A1.1) and in the Cochrane Library (Table A1.2). The search strategy also included scrutiny of the references in the included RCTs and relevant systematic reviews.
Table A1.1
English search terms and combinations for the search in MEDLINE.
Table A1.2.
English search terms and combinations for the search in Cochrane Library.
Russian
The following search terms and combinations in Russian were used for a search carried out from May to July 2023: диспансеризация, ежегодные профилактические осмотры, скрининг, популяционный скрининг, сахарный диабет, глюкоза сыворотки, глюкоза плазмы, промежуточная гипергликемия, ранние нарушения углеводного обмена, преддиабет, гипергликемия натощак, нарушенная толерантность к глюкозе, гликированный гемоглобин, гликозилированный гемоглобин, школа пациентов.
Study selection
Inclusion criteria
Studies that satisfied the criteria outlined under the patient/population, intervention, comparison and outcomes model with a duration of at least 6 months of followup were included (Table A1.3).
Table A1.3.
Study inclusion criteria using the patient/population, intervention, comparison and outcomes model.
Exclusion criteria
The following exclusion criteria were used:
studies on people:
- –
younger than 18 years;
- –
with T2DM;
- –
who are pregnant;
- –
with a recent hospitalization or myocardial infarction;
- –
with known cardiovascular diseases or severe chronic kidney disease;
- –
taking antipsychotic drugs or glucocorticoids;
- –
living in an institution; and
- –
with medical conditions limiting their applicability to primary-care-based populations (e.g. those with acute illness);
studies where intermediate hyperglycaemia is self-reported or diagnosed according to methods other than HbA1c, fasting plasma glucose or oral glucose tolerance test;
qualitative studies, modelling studies, case series, case reports, uncontrolled observational studies, retrospective cohort studies, any other non-RCT design;
comparative effectiveness (head-to-head) trials of medications or behavioural counselling without another eligible control group;
studies not generalizable to primary care (e.g. inpatient hospital, emergency departments, nursing homes, other institutional settings, school-based programmes or occupational settings);
studies in which more than 10% of the sample do not meet the inclusion criteria; and
letters, editorials, communications, conference abstracts and publications that contain no numerical outcomes data.
Selection process
An enhanced rapid evidence assessment approach was taken. Titles and abstracts of records identified by the searches were screened by one reviewer. A second reviewer then independently assessed a random 20% sample of the titles/abstracts. The search retrieved 2879 records in English after removal of duplicates and 185 records in Russian after removal of inappropriate records. The full publications of those papers considered potentially relevant by either reviewer (819 English and nine Russian) were then sourced, although outcomes were not considered at this stage. The full text of these articles was then assessed against the inclusion/exclusion criteria by one reviewer, with a random 20% sample assessed independently by a second reviewer. Disagreements were resolved by consensus or through discussion with a third reviewer. In total, 18 RCTs (reported in 19 papers) fulfilled the inclusion criteria. Data were also included from 34 trials (reported in 50 papers) identified in the USPSTF review (Fig. A1.1) (21) and three trials (reported in three papers) identified through the scrutiny of other systematic reviews. In total, data were available from 51 RCTs. Annex 2 lists the studies that were excluded following full-text screening with the reasons for exclusion.
Fig. A1.1.
Study selection process.
Quality appraisal
Risk of bias was assessed using the Cochrane Risk of Bias tool 2, which considers the risk of bias across five domains: (i) the randomization process, (ii) deviation from the intended intervention, (iii) missing outcome data, (iv) measurement of outcomes, and (v) selection of the reported results (105). Quality appraisal was conducted by one reviewer, with a random 20% assessed independently by a second reviewer. Disagreements were resolved by consensus or through discussion with a third reviewer. Annex 3 contains results from the risk of bias assessment.
Data extraction
The following data were extracted from all included studies: design; setting; country; duration; age and sex of participants; condition (type of intermediate hyperglycaemia: impaired fasting glucose or impaired glucose tolerance); diagnostic test used and criteria; number of participants screened, randomized and included in the analysis; and intervention and numbers in the intervention and control arm. The following data were extracted from the study results for both the intervention and control arm, recording the absolute numbers and hazard ratios and confidence intervals (CIs) for the data: number of people with T2DM and intermediate hyperglycaemia, all-cause mortality, cardiovascular mortality, non-fatal myocardial infarction, nonfatal stroke, non-traumatic amputation, revascularization, nephropathy (chronic kidney disease), neuropathy (general), neuropathy (specific subtypes: peripheral, autonomic, proximal, mononeuropathy), vision (moderate/severe non-proliferative diabetic retinopathy, proliferative diabetic retinopathy, macular oedema, reduced visual acuity), foot ulcers and harms (any).
Data were extracted by one reviewer, with a random 20% checked by a second reviewer. All data extracted were entered into a piloted electronic data collection form. Any disagreements were resolved by consensus or discussion with a third reviewer.
Methods of analysis and synthesis
Findings of studies were summarized in text and tables. As all outcomes were binary (did occur/did not occur), relative risks (with 95% CI values) were calculated using the numbers available in the papers, with the baseline number as denominator. Where multiple papers reported the same outcome, the one with the longest followup period was included. Where studies had multiple intervention arms, these were combined where it was reasonable to do so (e.g. they were more- or less-intensive versions of the same type of intervention). Meta-analyses were conducted using Revman (106) where at least three similar studies were available. Random-effects models using the inverse-variance weighted method (DerSimonian and Laird method) were used to estimate pooled effects (107). A narrative summary was provided for all studies that could not be meta-analysed. Subgroup analyses were conducted in line with those of the USPSTF review and the same definitions for categorizing participants into subgroups was employed.
