Table 2.

Findings from systematic reviews including quantitative studies

Systematic review	Effectiveness of (breastfeeding, diet, physical activity, pharmacological) interventions, and screening on reducing diabetes	Stakeholders involved	Organisation
Buelo 2019 UK (34)	Physical activity (PA) 4/28 statistically significantly (SS) increased PA 14 had either mixed effectiveness or no changes in PA Reported intervention components and study quality varied greatly Interventions that incorporated childcare issues, social support and cultural sensitivities were associated with effectiveness	Healthcare professionals Doctors Practitioners Researchers	‐
Chasan‐Taber 2015 USA (27)	Breastfeeding, diet, PA 2/9 reported type 2 diabetes (T2DM) incidence Annual incidence rate 6.1% vs. 7.3% Incidence rate ratio (IRR) = 0.83, 95% confidence interval (CI) 0.47‐1.48 Breastfeeding vs. usual care (UC) 1/1 non‐significant (NS) difference Diet and exercise vs. placebo SS 53% risk reduction of T2DM incidence, p = 0.002 4/9 Diet vs. control, Low Glycaemic Index (GI) diet vs. UC SS improvements to on one or more dietary components 3 SS impact on weight change 4 NS impact on weight change 2 SS impact on Body Mass Index (BMI) change 1 NS impact on BMI Exercise vs. UC 3/9 SS impact on one or more measures of PA 4/9 Positive impact on biomarkers of insulin resistance (glucose measures) 2/4 NS	‐	‐
Feng 2018 China (26)	Breastfeeding 13 cohort studies included in the meta‐analysis (MA) 9/13 reported SS association with a lower T2DM risk Risk ratio (RR) 0.66, 95% CI 0.48‐0.90, I2 = 72.8%, p < 0.001) 3/13 Long‐term (>1–3 months (m) postpartum (pp) NS association with T2DM risk 1 USA study (RR 0.66, 95% CI 0.43‐0.99), SS regardless study design: prospective (RR 0.56, 95% CI 0.41–0.76); retrospective (RR 0.63, 95% CI 0.40–0.99), smaller sample size (RR 0.52, 95% CI 0.30–0.92, p = 0.024) Follow‐up (FU) >1 y (RR 0.75, 95% CI 0.56–1.00) (Adjusted RR 0.69, 95% CI 0.50–0.94)	‐	‐
Gilinsky 2015 UK (28)	Breastfeeding, diet, PA 3/13 reported on progression to T2DM (Ratner; Shek; Wein) Equally effective at reducing the rate of T2DM progression in women with previous gestational diabetes mellitus (pGDM) and without pGDM Numbers needed to treat higher among women with vs. women without previous GDM (pGDM) NS rate reduction in T2DM at 3 years (y) (Shek) and 51 m (Wein) Breastfeeding and sleep may offset T2DM risk after GDM MA found a SS 34% lower T2DM risk for any breastfeeding vs. no breastfeeding (Feng) Diet 6/11 favourable intervention effects PA 6/11 favourable intervention effects MA found SS weight loss was attributable to one Chinese population study (WMD = −1.06 kg (95% CI = −1.68−0.44) Lifestyle interventions NS change Fasting Blood Glucose (FBG) or T2DM risk Recruitment rates were poor but study retention good	Trained counsellor Exercise physiologist Dieticians Lifestyle behaviour case manager Research nutritionist Lactation consultant Peer educators (training and support from a multidisciplinary health professional team) Diabetes educators Research nurse	Hospital clinic and community health centre Hospital clinics
Goveia 2018 Brazil (29)	Breastfeeding, diet, PA MA found homogeneous (I2 = 10%), NS reduction of 25% T2DM incidence No beneficial changes in glycaemic levels (mean change from baseline of FBG, oral glucose tolerance test (OGTT) or haemoglobin A1c (HbA1c) Moderate reductions in weight (MD = −1.07 kg; −1.43−0.72 kg); BMI (MD = −0.94 kg/m2; −1.79 −0.09 kg/m2); and waist circumference (MD= −0.98 cm; −1.75 −0.21 cm) Only interventions soon after delivery (<6 months pp) were effective (RR =0.61; 95%CI: 0.40–0.94; p for subgroup comparison = 0.11) Effects were larger in studies with longer duration and FU Importance of maintaining support for lifestyle changes for a longer period, particularly given the women's frequently overwhelming tasks of motherhood	Lifestyle coach Nutrition coaching	Clinics Hospitals
Guo 2016 China (30)	Diet, PA Incidence of T2DM (FBG, or HbA1c). 5 lifestyle intervention vs. UC Annual mean T2DM incidence ranged from mean = 6.0% vs. mean = 9.3% NS, Effect size ranged from 0.05 – 0.40 among these 5 studies 7/10 evaluated FBG between the two groups 1 revealed a SS decreased FBG in the intervention group 5 effect size ranged from 0.004 to 0.50 2/10 evaluated HbA1c between group 1 SS decrease of HbA1c 7/10 reported at least a small effect size (> 0.20) on T2DM development 1 woman with GDM enrolled in Diabetes Prevention Program (DPP) had 12‐year interval (mean) on T2DM development (Ratner) Majority (75%) of studies only immediate or interim efficacy Increasing PA / Decreasing sedentary activity Pp weight gain/ Improving dietary outcomes Risk perception of T2DM	Trained counsellor Dietician Research nurse Exercise physiologist Case manager Diabetes educators Nutritionist Physicians healthcare professionals Trained interventionists	‐
Jones 2017 USA (18)	Diet, PA Diet, weight 7/8 SS reduced weight and hip and waist circumference, NS decreased weight, decreased dietary fat (Ferrara), Decreased weight 1 y FU (Nicklas) SS reduced total fat intake, total carb. intake and GI load (Reinhardt) NS decline in weight and insulin resistance; no changes in glucose levels (Kim), NS change in weight, BMI or insulin resistance (McIntyre) NS clinical improvement in eating behaviours, NS changes in glucose metabolism or body composition (Peacock) PA 8/8 No differences (Ferrara, Smith, Nicklas, Kim) NS % of women achieved goals, targets were not attained (Cheung) NS increased PA, majority failed to reach recommended PA levels (McIntyre) NS clinical improvement in PA (Peacock) NS changes in total level (Reinhardt)	Researchers Clinicians Communities Dieticians Lifestyle coach/ interventionist exercise physiologist	Home‐based settings
Middleton 2014 New Zealand (31)	Screening pp Postal reminders sent to, respectively: GDM women, GDM women and physicians, or physicians only Proportion of women having their first OGTT pp RR 3.87 (1.68–8.93) RR 4.23 (1.85–9.71) RR 3.61 (1.50–8.71) Proportion of women diagnosed with T2DM or showing impaired glucose tolerance or impaired FBG pp RR 1.57 (1.01–2.44) RR 1.78 (1.16–2.73) RR 1.69 (1.06–2.72) Low‐quality evidence for a marked increase in uptake of testing for T2DM Important to determine whether increased test uptake rates increase women's use of preventive strategies such as lifestyle modifications Other forms (email and telephone) reminders need to be assessed; more understanding of why some women fail to be screened pp is needed	Clinicians Health professional Physicians	Clinics University‐affiliated tertiary centre
Morton 2014 UK (25)	Breastfeeding, diet, PA, pharmacological Breastfeeding 15‐y risk of T2DM in women who breastfeed for >3 m vs. <3 m: 42% (95% CI 28.9–84.7) vs. 72% (60.5–84.7%) Protective effect on T2DM development remained SS after multivariate analysis (Hazard ratio (HR) 0.55, 95% CI 0.35–0.85, p < 0.001) SS decreased T2DM incidence after intensive lifestyle intervention with regular, individualized FU (3y): RR 0.50; p = 0.006 (Ratner) Diet and exercise RR 0.63 (95% CI 0.35–1.14) p = 0.12 (Wein) Diet and exercise RR 0.77 (95% CI 0.51–1.16) (Shek) Blood glucose 2 hour (h) post−75 g, load from baseline Low‐GI diet: Change in blood glucose, p = 0.025 (Shyam) Diet Effects of 3 dietary patterns (Tobias): 1‐unit interquartile range associated with 15% reduction, HR 0.84 (95% CI 0.73–0.96), The alternate Mediterranean diet, HR 0.86 (95% CI 0.73–1.03), p = 0.01; Dietary Approaches to Stop Hypertension), NS 17% reduction HR 0.77 (95% CI 0.64–0.93) alternate Healthy Eating Index Adjusted for BMI PA Comparing highest vs lowest quartiles of total PA over 16 y FU: SS 28% reduction in progression to T2DM (RR 0.72, 95% CI 0.55–0.96, p = 0.01) Women >7.5 metabolic equivalent hours/w vs. <7.5/w: SS 29% reduction in risk (RR 0.71, 95% CI 0.59–0.86, p < 0.001) (Bao) Pharmacological interventions Metformin SS 50% reduction in T2DM incidence >3y FU compared to UC (p = 0.006) HR 0.45 (95% CI 0.25–0.83) p = 0.009 (Buchanan), RR 0.47; p = 0.002 (Ratner) Troglitazone Prevention of Diabetes (TRIPOD) study (n = 266) on Hispanic 400 mg. SS reduction incidence, FU 30 m (HR 0.45, 95% CI 0.25–0.83, p = 0.009) Troglitazone (200 or 400 mg) in 42 Latino women, SS improvement in insulin sensitivity, FU 12 w 88 ± 22 (200 mg) 40 ± 22 (400 mg)/ 4 ± 14%, p = 0.03 (Berkowitz) SS decreased levels of fasting insulin concentrations, 20% ± 9% (400 mg) vs. +/−7% (200 mg) and 10% ± 10% (placebo), p = 0.03	Dieticians	‐
Peacock 2014 Australia (36)	Diet, PA, pharma logical Summary of identified studies Diabetes incidence rate SS decreased in the intervention group (5.4%) vs. placebo group (12.1%), p < 0.001 Diet NS returning to pre‐pregnant weight Intervention SS more effective in women without excessive gestational weight gain, p = 0.04 SS Weight reduction (95% CI: −7.6 to −0.5) and changes in dietary intake Reduction in weight in participants, p = 0.03 Eating patterns were changed during the index GDM pregnancy (protein p = 0.01, fibre p = 0.002) but not sustained pp PA SS leisure time PA increased in first year in women post GDM (p = 0.002) NS differences in PA and weight loss NS average time of PA (mean 60 (0–540) min/week) increased NS 10,000 steps on 5 or more days not reached Pharmacological Lifestyle changes (58% {48–66, 95%CI}) and Metformin (31% {17–43, 95%CI}) reduced the incidence of diabetes Lifestyle intervention (p = 0.002) and Metformin (p = 0.006) reduced the risk of T2DM compared to placebo and control Results supported a class effect of Thiazolidinedione drugs to enhance insulin sensitivity, reduce insulin secretory demands and preserve pancreatic b‐cell function in intervention group, p = 0.01 Group sessions demonstrated a potential to improve perceptions of healthiness in women but NS
Pedersen 2017 Denmark (20)	Diet, PA No specific intervention or components were found superior NS reduction of T2DM incidence (tendency only) SS pooled estimate of absolute risk reduction (−5.02 per 100 (95% CI: −9.24;−0.80) SS effect in the subgroup of participants >40 y (T2DM incidence 8% in intervention group vs. 20% in control group, n = 175, p = 0.018 Tendency of poorer effect starting during pregnancy or very early pp (≤6 w) vs. interventions started >6 w pp SS changes were found for PA but not for diet Biomarkers of insulin resistance Generally, results were consistent within trials 2 showed NS effect on fasting glucose in spite of a SS intervention effect on other measures of insulin resistance	Trained dieticians Exercise physiologist Trained research nurse	Medical centres Fitness centres
Tanase‐Nakao 2018 Japan (33)	Breastfeeding 6/9 reported results in favour of breastfeeding regards to T2DM incidence, 3/9 reported null results 2–4 w pp breastfeeding tends to lower the risk of T2DM compared with women with shorter period. SS effect with FU>2 y FU<2 y = OR 0.77, (95% CI 0.01–55.86) 2–5 y = OR 0.56, (95% CI 0.35–0.89) >5 y = OR 0.22, (95% CI 0.13–0.36) Exclusively breastfeeding for 6–9 weeks pp lower the risk compared with women giving formula feeding (OR 0.42, 95% CI 0.22–0.81)	‐	‐
Van der Heuvel 2018 Netherlands (38)	Screening pp eHealth in GDM care has evolved most notably of all perinatal appliances of eHealth the last 3 years (smartphone‐facilitated remote blood glucose monitoring, management of medication schedules through Web‐based or SMS‐facilitated feedback systems, and telephone review service to support and supervise glycaemic control) Decrease in planned and unplanned visits by 50% to 66%, whereas no unfavourable differences in glycaemic control, maternal, and neonatal outcomes occurred Advantages of eHealth implementation in perinatal care: Patient satisfaction and engagement, fewer clinic visits, clinician satisfaction, remote monitoring, access to care in low‐ and middle‐income countries Disadvantages and indistinct impacts: reimbursement, legal issues, technical issues, limited A‐level evidence, health outcome and costs pp screening after GDM with telephone FU (RCT) (Roozbahani) SS reduced FBG levels in mothers with GDM and increased the rate of pp screening test	Obstetricians	Outpatients clinics Hospitals Tertiary hospital

Abbreviations: BMI, body mass index; CI, confidence interval; DPP, Diabetes Prevention Program; FBG, Fasting Blood Glucose; FU, follow‐up; m, month; GI, Glycaemic Index; h, hour; HbA1c, haemoglobin A1c; HR, Hazard ratio; IRR, incidence rate ratio; MA, meta‐analysis; MD, mean difference; NS, not significant; OGTT, oral glucose tolerance test; PA, physical activity; pGDM, previous gestational diabetes mellitus; pp, postpartum; p, p‐values*; RR, risk ratio; SS, statistically significant; T2DM, type 2 diabetes; UC, usual care; w, week; y, year.

^* p‐values and authors of primary studies only if reported in the systematic review.