Table 3.

Summary of results

Authors	Year of publication	No of studies	Intervention	Comparison	Formance (AUC, C	Validation	Outcomes: incidence, prevalence, mortality	Quality assessment score
Abbasi et al18	2012	16	25 type 2 diabetes risk prediction models.	Prospective cohort study, with a case cohort study random subcohort.	C-statistics: 0.74–0.92	External validation cohort	Type 2 diabetes morbidity (incidence)	6
Barber et al19	2014	12	18 type 2 diabetes risk prediction models.	Seven risk tools were validated using an external dataset. 14 of the tools were validated internally using resampling techniques such as bootstrapping and cross-validation. 4 of the tools also split their dataset into a ‘training’ set, which was used to create the tool. One risk tool was validated using a partially independent dataset; all cases were used in both the ‘training’ and ‘test’ cohorts, however, the non-cases were split into these two cohorts.	AUROC: 0.69	Internal and external validation cohort.	Pre-diabetes morbidity (incidence).	6
Beswick et al29	2008	110	110 risk scoring methods with potential for use in primary prevention (among them, 70 methods specifically designed for use in primary prevention).	internal?	AUROC min 0.57, max 0.88	Internal and convergent validation	CHD, MI/sudden ischaemic death CHD death, CVD, stroke, CHD, CVD death MI or CHD death CHD death CHD Death all causes Stroke (ABI) MI Heart attack CVD MI or stroke (ABI) Sudden death Hard CHD All causes of death CHD (CAD) fatal MI non-fatal MI sudden coronary death chronic CHD IHD fatal/ non-fatal MI CHD death or non fatal MI hard CHD hard CVD hard cerebrovascular macrovascular disease fatal MI/ sudden death MI (fatal or non-fatal).	11
Collins et al35	2011	39	Use of 47 different risk predictors.	10 studies randomly split the cohort into development and validation cohorts. 8 of these studies split the original cohort equally into development and validation cohorts. 21 studies conducted and published an external validation of their risk prediction models within the same article, and 8 of these studies used two or more data sets in an attempt to demonstrate the external validity of the risk prediction models.	NA	Internal and external validation cohort	Diabetes morbidity (incidence and prevalence).	6
Cortes-Bergoderi et al24	2012	5	Framingham Novel RPM based on local cohort WHO/ISH cardiovascular risk score system Chagas RPM.	Framingham not for Chagas.	C-stat MAX 0.81 (95% CI, 0.72 to 0.90); MIN 0.69	N\A (cChagas only)	General mortality, CVD mortality and morbidity, (incidence), and CHD mortality and morbidity (incidence).	6
Damen et al28	2016	212	363 different models.	Internal?	For 143 (39%) models, discrimination was reported as a Cstatistic or area under the receiver operating characteristic curve; D statistic 5 (1%) other (sensitivity, specificity, etc) 24 (7%).	80 of the 363 developed models (22%) were internally validated; 132 (36%) models were externally validated	Fatal or non-fatal CHD Fatal or non-fatal: CHD, CVD, myocardial infarction and stroke fatal: CHD, CVD and non-CHD Fatal CVD Fatal or non-fatal stroke fatal or non-fatal: CHD, myocardial infarction and stroke; claudication; coronary artery bypass grafting; percutaneous transluminal coronary angioplasty; transient ischaemic attack fatal CHD.	7
Damen et al31	2019	38	Framingham Adult Treatment Panel III, the Framingham Wilson model and the pooled cohort equations.	Internal	OE from 0.58 to 0.79; C-STAT from 0,58 to 0,82.	External	Fatal or non-fatal MI, Fatal or non-fatal stroke, fatal, hard or non-fatal CHD, angina pectoris, TIA, ASCVD.	9
Echouffo-Tcheugui et al30	2015	13	22 RISK SCORES: FHS risk score Health ABC Score Kaiser Permanente FHS Risk score 1 FHS Risk score 2 FHS Risk score 3 CHS score 1 CHS risk score 2 MDCS score 1 MDCS score 2 MDCS score 3 MDCS score 4 MDCS score 5 MDCS score 6 MDCS score 7 ABC score 2 Proactive trial score CHS? FHS ARIC NAVIGATOR study British Regional Heart Study.	Internal	AUC from 0.71 to 0.87	HL χ2 from 2.98 to 9.45	Heart failure incidence.	6
Echouffo-Tcheugui et al33	2013	11	Johns Hopkins Framingham score Women’s Health Study (WHS) Inclusive model WHS simplified model with lipids WHS- simplified model Whitehall II risk score Whitehall II Repeat measures risk score Whitehall II Average blood pressure measure risk score Whitehall II usual measure risk scores ARIC/CHS Score Iran BP risk score Taiwan BP clinical risk model Taiwan BP biochemical risk model Korean risk model Swedish risk model.	Internal?	MAX 0.803, MIN 0.707	External? C-statistic in validation studies (0.71 to 0.81)	Hypertension morbidity (incidence).	9
Fowkes et al26	2008	20	Baseline ABI measurements.	Framingham risk score	N\A		General mortality, CVD mortality and CHD morbidity (incidence).	6
Hu et al21	2016	12	12 type 2 diabetes risk prediction models.	To be seen	AUROC: 0.66–0.91	Internal and external validation cohort	Diabetes morbidity (incidence).	8
Noble et al22	2011	43	94 type 2 diabetes risk prediction models.		AUROC: 0.74–0.85	Internal and external validation cohort	Diabetes morbidity (incidence).	5
Siontis et al27	2012	20	Framingham ASSIGN score SCORE score PROCAM score Reynolds risk score QRISK1 QRISK2.	Internal?	AUC MIN 0.55 MAX 0.83	N\A	CVD mortality, CHD incidence, cerebrovascular disease incidence, CABG or PTCA.	9
Sun et al32	2017	26	Anthropometric indices risk prediction ARIC/CHC risk score biomarker-based risk-prediction model China risk prediction model 1 China risk prediction model 2 China risk prediction score conditional model Demographic indices risk prediction model one for men demographic indices risk prediction model two for men demographic indices risk prediction model three for women demographic indices risk prediction model 1 for women demographic indices risk prediction model 2 for women demographic indices risk prediction model 3 for women Framingham risk score genetic risk prediction model 1 genetic risk prediction model 2 genetic risk prediction model 3 IDH risk prediction model InterASIA risk prediction ISH risk prediction model Japanese risk prediction model Japanese risk score sheet Johns Hopkins KoGES risk score Korean genetic risk score marginal model northeastern Han Chinese genetic risk score Prediction for men Prediction for women rural India risk score SHIP risk model Swedish genetic risk model Swedish nongenetic risk model Swedish risk model 2 Taiwan BP clinical risk model the ‘usual’ blood pressure risk score the average blood pressure risk score the PFAA index TLGS risk multivariable models TLGS risk prediction for IDH TLGS risk prediction for ISH TLGS risk score Whitehall II repeat measures risk score Whitehall II risk score WHS inclusive risk prediction WHS Simplified Model WHS Simplified Model with Lipids.	Internal?	AUC=0.767, 95% CI(0.742, 0.792)	Variable	Hypertension morbidity (incidence)	9
Tzoulaki et al25	2009	79	Application of FRS plus a candidate additional predictor.	Framingham risk score	AUC 0.77	n.p.	General mortality, CVD mortality and morbidity (incidence) and CHD mortality and morbidity (incidence).	5
Yoshizawa et al23	2016	18	Non-blood-based risk model for type 2 diabetes	None of the included studies met all criteria on the QUADAS list. Among five studies that had both derivation and validation cohorts, three studies conducted the calibration in both cohorts; two studies did it only in derivation cohort.	AUC: 0.71–0.79	Internal and external validation cohort.	Diabetes morbidity (incidence)	8

ABC, Age, Biomarker, Clinical History score; ABI, Ankle-Brachial Index; ARIC, Atherosclerosis Risk in Communities study; ASCVD, Arteriosclerotic Cardiovascular Disease; ASSIGN, Assessing cardiovascular risk using SIGN (Scottish Intercollegiate Guidelines Network) guidelines score; AUC, AUROC, area under the curve/area under the receiver-operating characteristic curve; BP, blood pressure; CABG, coronary artery bypass grafting; CHD, coronary heart disease; CHS, Cardiovascular Health Study; C-STAT, C-statistic; CVD, cardiovascular disease; D-STAT, D-statistic; FHS, Framingham Hearth Study; FRS, Framingham Risk Score; ISH, Isolated systolic Hypertension; MI, Miocardial Infarction; N/A, Not Available; PROCAM, Cardiovascular Risk PROCAM Score; PTCA, Percutaneous transluminal coronary angioplasty; RPM, Risk Prediction Model; TLGS, Tehran Lipid and Glucose Study.