Background
Methods of scoping review
Results
Authors/ year | CVD calculators | Study design | Level of evidence |
---|---|---|---|
Mendis et al. / 2007 [19] | WHO/ISH risk charts | A Hypothetical cohort for different 14 sub-regions | Level II prognostic study |
Hippisley-Cox et al. / 2007 [10] | Framingham risk score QRISK1® ASSIGN | A Prospective open cohort study | Level I prognostic study |
Beswick et al. / 2008 [24] | A total of 70 risk scoring methods | Series of systematic reviews | Level I prognostic study |
Hippisley-Cox et al. / 2008 [18] | Framingham risk score QRISK1® QRISK2® | A Prospective open cohort study | Level I prognostic study |
Collins et al. / 2009 [25] | Framingham risk score QRISK1® | A Prospective open cohort study | Level I prognostic study |
Collins et al. / 2010 [17] | Framingham risk score QRISK1® & QRISK2® | A Prospective open cohort study | Level I prognostic study |
Mendis et al. / 2011 [26] | WHO/ISH risk charts | A Cross-sectional population-based survey | Level IV prognostic study |
Siontis et al. / 2012 [27] | Framingham risk score ASSIGN SCORE PROCAM Reynolds risk score QRISK1® & QRISK2® | A Systematic review of comparative predictive model studies | Level I prognostic study |
Kariuki et al. / 2013 [28] | Framingham non-laboratory Gaziano non-laboratory WHO/ISH non-laboratory Swedish Consultation-based UK General Practice model | A Systematic review of non-laboratory-based CVD algorithms | Level I prognostic study |
Otgontuya et al. / 2013 [29] | WHO/ISH risk charts | A Retrospective cohort study in 3 Asian countries | Level II prognostic study |
Bansal et al. / 2014 [30] | JBS3 CVD risk calculator ACC/AHA ASCVD) Risk Framingham risk score WHO/ISH risk charts | A Cross-sectional study | Level IV prognostic study |
Selvarajah et al. / 2014 [31] | Framingham risk score SCORE-high and -low WHO/ISH risk charts | A Retrospective cohort study | Level II prognostic study |
Ofori et al. / 2017 [32] | ACC/AHA ASCVD) Risk Framingham risk score WHO/ISH risk charts | A Cross-sectional study | Level IV prognostic study |
Bonner et al. / 2018 [33] | 73 unique CVD risk calculators | A Systematic review to measure the validity, understandability, and actionability of online CVD risk calculators | Level I prognostic study |
Kaptoge et al. / 2019 [34] | 2019 WHO risk charts | A Retrospective cohort study from 21 global regions | Level II prognostic study |
Hasabullah et al. / 2020 [21] | Framingham risk score SCORE ACC/AHA ASCVD) Risk QRISK® | A Cross-sectional study | Level IV prognostic study |
Hosein et al. / 2020 [35] | Framingham risk score ASSIGN QRISK2® | A Cross-sectional study | Level IV prognostic study |
Summary of existing CVD risk calculators
Characteristics | QRISK® | WHO/ISH risk charts |
---|---|---|
Derivation study | Prospective cohort database (version 19 of the QRESEARCH database) | Hypothetical cohorts for different 14 sub-regions derived from the WHO Comparative Risk Assessment study [37] |
Statistical method | Cox proportional hazards models | Non- specified Modelling approach |
Population of study | 2.29 million patients | Not applicable |
Sample size | 1,535, 583 patients (Simple random sampling) | Not applicable |
Duration of the study | 1 January 1993 to 31 March 2008 | 2007 |
Age of participants | 35—75 | 40—79 |
Country of validation | United Kingdom | Not applicable |
Ethnicity involved | White/not-recorded, Indian, Pakistani, Bangladeshi, other-Asian, black-African, black-Caribbean, Chinese, Mixed | Not involved |
Number of variables | 14 | 6 |
Type of variable | • Fixed Risk Factors • Modifiable Risk Factors • Ongoing Clinical Conditions | • Fixed Risk Factors (Age, sex) • Modifiable Risk Factors + Diabetes |
Applicability and ease of use | • Online available calculator • Clinical and lab variables are needed for more accurate personalized risk | Two paper-based versions of WHO/ISH risk charts for each sub-region |
Discrimination in AUROC | 0.792 | 0.613 (Discrimination in Asian population) |
10-year risk estimate | Percentage & Categories | Categories |
Risk categories | • < 5%, • 5- 10%, • 10–20%, • > 20% | • Green < 10% • Yellow 10% to < 20% • Orange 20% to < 30% • Red 30% to < 40% • Deep Red > 40% |
Costs involved | • Involve more costs for lab and clinical examination required for complete risk assessment | • Fewer costs are involved, especially for charts without blood cholesterol investigation |
QRISK® | WHO/ISH risk charts | |
---|---|---|
Advantages | 1. Dynamically updated CVD risk score developed from annually updated anonymized e-health records to reflect changes in the population characteristics 2. Various ethnicity and deprivation groups are included in calculating the CVD risk score for population groups most likely disadvantaged by the other risk algorithms 3. Additional modifiable risk factors and ongoing clinical conditions are included to quantify the CVD risk score for every individual patient 4. CVD Risk scores could be saved through digital online platforms if embedded within mobile or computer applications 5. Upgradeable to a national comprehensive CVD risk factors profile and a rank-ordered recall list | 1. Charts can be used even in resource-constraint healthcare settings due to simple variables 2. Allow improvement and effectiveness of CVD risk assessment even in the countries that do not possess sophisticated technology for the use of online calculators through established health information systems 3. The WHO/ISH risk charts provide an optimal visual illustration when explaining the implications of elevated CVD risk to the patients through colour-coded CVD risk categories [24] |
Disadvantages | 1. Some data are usually missing due to the submission of several CVD risk factors; an estimated QRISK® score is usually calculated using the previously recorded data and predicted values based on the patient's ethnicity, age, and sex 2. High costs are needed for incorporating into a primary health setting with an appropriate computer system and costly screening investigations and examinations are needed for complete “actual” QRISK® CVD risk assessment 3. Patients presented with a QRISK® score acknowledged their CVD risk level, but it is usually unclear if they correctly understood the 10-year CVD percentage risk. It needs after-score recommendations and lifestyle modifications to prevent the poor patient recall of CVD risk, confusion or misunderstanding [39] | 1. Results will be only applicable to the country with the largest population in the region 2. Charts underestimate the risk in several groups of people, e.g. Hypertensive with blood pressure persistently ≥ 160/100 mm Hg, people with blood cholesterol ≥ 8 mmol/l, diabetics, people with a history of diagnosed ischemic heart disease or renal patients [40] 3. The WHO/ISH risk charts often incorrectly categorize most people into the low CVD risk group due to no prior validation, leading to a higher under-treatment rate and more complications and costs to incur [31] 4. The over-simplification and absence of information on validation or discrimination index have a real effect on the sensitivity, specificity and predictive accuracy of the WHO/ISH risk charts [31] |
Variables | QRISK®2 | WHO/ISH | |
---|---|---|---|
Fixed Risk Factors | Age | X | X |
Gender | X | X | |
Ethnicity | X | – | |
Relevant family history | X | – | |
Modifiable Risk Factors | Smoking status | X | X |
Systolic blood pressure | X | X | |
HDL Cholesterol level | X | – | |
Total Cholesterol Level | X | X | |
Body mass index | X | – | |
Deprivation score | X | – | |
Ongoing clinical Conditions | Ongoing Diabetes | X | X |
Ongoing hypertensive medication | X | – | |
Ongoing Rheumatoid arthritis | X | – | |
Ongoing chronic kidney disease | X | – | |
Ongoing Atrial Fibrillation | X | – |
AUC (95% CI) | |||
---|---|---|---|
Study | Models | Men | Women |
Collins et al. (2010) [17] | Framingham risk score | 0.75 (0.747 to 0.753) | 0.774 (0.771 to 0.777) |
QRISK1® | 0.771 (0.768 to 0.774) | 0.799 (0.796 to 0.802) | |
QRISK2® | 0.773 (0.770 to 0.776) | 0.801 (0.798 to 0.804) | |
Collins et al. (2009) [25] | Framingham risk score | 0.752 (0.749 to 0.755) | 0.770 (0.766 to 0.774) |
QRISK1® | 0.762 (0.759 to 0.765) | 0.789 (0.785 to 0.793) | |
Hippisley-Cox et al. (2008) [18] | Framingham risk score | 0.779 (0.776 to 0.782) | 0.800 (0.797 to 0.803) |
QRISK1® | 0.788 (0.786 to 0.791) | 0.814 (0.811 to 0.817) | |
QRISK2® | 0.792 (0.789 to 0.794) | 0.817 (0.814 to 0.820) | |
Hippisley-Cox et al. (2007) [10] | Framingham risk score | 0.759 (0.756 to 0.764) | 0.774 (0.771 to 0.778) |
QRISK1® | 0.767 (0.763 to 0.772) | 0.788 (0.785 to 0.791) | |
ASSIGN | 0.764 (0.760 to 0.769) | 0.784 (0.781 to 0.787) |