For Researchers
Project BigLife algorithms are published in peer-review science literature. We embrace open science with additional web appendices, visualization tools and other resources to describe how the algorithms. These resources support further development, validation and use the algorithms. The resources support further development, validation and use the algorithms. The resources are on are on our Github repository and in the algorithm publications.
Our Calculators
Life Expectancy / Mortality Population Risk Tool (MPoRT)
"Unhealthy behaviours place a major burden on Canadian life expectancies," said lead author Dr. Doug Manuel, senior scientist at The Ottawa Hospital and professor at The University of Ottawa, and a senior core scientist at the Institute for Clinical Evaluative Sciences (ICES). "This study identified which behaviours pose the biggest threat."
The study found:
• 26 % of all deaths are attributable to smoking
• 24 % of all deaths are attributable to physical inactivity
• 12 % of all deaths are attributable to poor diet
• 0.4 % of all deaths are attributable to unhealthy alcohol consumption
The life expectancy calculator was developed using a statistical model to estimate the risk of death associated with smoking, unhealthy alcohol consumption, poor diet, and physical inactivity in Ontario. "Our approach is a new way of measuring the impact of health problems on life expectancy," said Dr. Manuel. "In an era of big data, we should be moving beyond the old methods that have remained largely unchanged for the past 60 years."
References
Manuel DG, Perez R, Sanmartin C, Taljaard M, Hennessy D, Wilson K, . . . Rosella L. (2018). Development Measuring Burden of Unhealthy Behaviours Using a Multivariable Predictive Approach: Life Expectancy Lost in Canada Attributable to Smoking, Alcohol, Physical Inactivity, and Diet. PLoS Medicine, 13(8), e1002082. doi:
Manuel DG, Abdulaziz KE, Perez R, Beach S, Bennett C. Personalized risk communication for personalized risk assessment: Real world assessment of knowledge and motivation for six mortality risk measures from an online life expectancy calculator. Inform Health Soc Care. 2017:1-14. https://doi.org/10.1080/17538157.2016.1255632
Manuel DG, Perez R, Bennett C, Rosella L, Taljaard M, Roberts M, . . . Manson H. Seven more years: The impact of smoking, alcohol, diet, physical activity and stress on health and life expectancy in Ontario. An ICES/PHO Report. Toronto: Institute for Clinical Evaluative Sciences and Public Health Ontario; 2012. ISBN: 978-1-926850-33-7
Cardiovascular Disease Population Risk Tool (CVDPoRT)
A predictive algorithm for 5-year risk of incident cardiovascular disease. Developed and validated using the 2001 to 2008 Canadian Community Health Surveys (CCHS). Cardiovascular incidence is a first hospitalization or death for major cardiovascular event. The main predictors are health behaviours (smoking, diet, physical activity and alcohol consumption). The model is currently calibrated for Canada 2013 with provisions to calibrate to other countries.
There were 104 219 respondents aged 20 to 105 years, 3 709 cardiovascular events, and 818 478 person-years follow-up in the combined derivation and validation cohorts.
5-year cumulative incidence - males = 0.026, 95% confidence interval [CI] 0.025–0.028; females = 0.018, 95% 0.017–0.019.
Discrimination - c-statistic: male model = 0.82, 95% CI 0.81–0.83; female model = 0.86, 95% CI 0.85–0.87.
Calibration - overall population 5-year observed cumulative incidence versus predicted risk: males = 0.28%; females = 0.38%. Calibration slope females: 0.9734, SE 0.0698; for males: 0.9295, SE 0.0731. Observed versus predicted < 20% difference in predefined policy-relevant subgroups (206 of 208 groups) at P
Trial registration: ClinicalTrials.gov, no. NCT02267447
Reference data
Additional reference data includes:
• An interactive algorithm visualization tool that displays the contribution of the main predictors and the role of age and exposure interactions.
• CVDPoRT model in XML format (Predictive Modelling Markup Language);
• 500 beta coefficients to calculate statistical uncertainty; and,
• Fictitious examples (n = 20 000) of exposure input and output, including intermediate calculation steps.
See the API/developer page for additional information.
References
Manuel, D., Tuna, M., Bennett, C., Hennessy, D., Rosella, L., Sanmartin, C., . . . Taljaard, M. (2018). Development and validation of a cardiovascular disease risk-prediction model using population health surveys: the Cardiovascular Disease Population Risk Tool (CVDPoRT).
Canadian Medical Association Journal,190(29), E871-882. doi:doi: 10.1503/cmaj.170914
Taljaard, M., Tuna, M., Bennett, C., Perez, R., Rosella, L., Tu, J. V., . . . Manuel, D. G. (2014). Cardiovascular Disease
Population Risk Tool (CVDPoRT): predictive algorithm for assessing CVD risk in the community setting. A study protocol.
BMJ Open, 4(10), e006701.doi:10.1136/bmjopen-2014-006701
Risk Evaluation for Support: Predictions for Elder-life in the Community Tool (RESPECT)
The algorithm used in this calculator is called RESPECT. RESPECT is short for Risk Evaluation for Support: Predictions for Elder life in the Community Tool.
RESPECT was developed on data from 491,277 older people in Ontario, Canada, who used home care between 2007 and 2013. The dataset contains detailed health information from the standardized Resident Assessment Instrument for Home Care (RAI-HC), which case managers and nurses use to assess the needs of home care users. The calculator includes a wide range of questions that reflect different trajectories in physical health and cognition, including age, sex, cognitive impairment (memory decline), diseases (e.g. diabetes, cardiovascular disease, dementia, cancer), sociodemographic factors (marital status, level of education), health status (e.g. difficulties with activities of daily living), symptoms of reduced physiologic reserve (e.g. weight loss), use life-sustaining therapies (e.g., dialysis), and health care use (e.g., number of hospitalizations and emergency department visits).
The algorithm is calibrated to 1.3 million assessments and 112,823 deaths. RESPECT accurately predicts a wide range in life expectancy; we ranked and classified our cohort based on their six-month risk of death, which spanned from 1.5% to 98%. This translates to an average survival of 4 weeks (interquartile range [IQR] of 11 to 84 days) for the frailest people in our cohort to 8.1 years among people with the few physical limitations and chronic health conditions (IQR of 5.9 years to 9.4 years).
RESPECT is incorporated into the Project Big Life scoring engine. The scoring engine generates calculations on any combination of responses to the algorithm questions. The main output of the engine is the actual health experience of Ontarians who completed the 1.3 million assessments. These health experiences can be translated into a wide range of patient-oriented measures. For more information on how we developed the RESPECT–End of Life algorithm, click here.
References
Hsu AT, Manuel DG, Taljaard M, Chalifoux M, Bennett C, Costa AP, . . . Tanuseputro P. Algorithm for predicting death among older adults in the home care setting: study protocol for the Risk Evaluation for Support: Predictions for Elder-life in the Community Tool (RESPECT). BMJ Open 2016;6:e013666. doi:10.1136/bmjopen-2016- 013666
Amy T. Hsu, Douglas G. Manuel, Sarah Spruin, Carol Bennett, . . . Peter Tanuseputro: Predicting death in home care users: derivation and validation of the Risk Evaluation for Support: Predictions for Elder-Life in the Community Tool (RESPECT). CMAJ Jul 2021, 193 (26) E997-E1005; DOI: 10.1503/cmaj.200022
Sodium Calculator (Salt)
The Sodium Calculator consists of 26 questions to capture how often sodium-containing foods are consumed; it provides users with rapid, detailed, personalized information on dietary sodium.
The Sodium Calculator questions were developed by examining the sources of sodium in the diet using the Canadian Community Health Survey (CCHS)-Nutrition 2004/2015. The sodium content (mg/100g) of foods and beverages that made the greatest contribution to sodium intakes were collected and combined using the University of Toronto Food Label Information Program (FLIP) 2013/2017 and the Canadian Nutrient File 2010/2015. FLIP is a branded food composition database of 10,487 and 16,761 packaged foods for 2013 and 2017, respectively, whereas the Canadian Nutrient File is the national food composition database of 5,807 (2010) and 5,690 (2015) foods maintained by Health Canada. The University of Toronto Restaurant-FLIP was used to derive the sodium content of fast-food and restaurants foods (n=4,836 and 12,271 for years 2013 and 2016, respectively). Foods were classified into sodium-focused categories that included food group (e.g. bakery), and major (e.g. bread) and minor (e.g. bagels) sub-categories.
To estimate the amount of sodium consumed from the different food categories, the Sodium Calculator uses algorithms based on average portion sizes consumed by 13 age and sex groups, as reported in the CCHS: all 4-8 years, and males and females 9-13 years, 14-18 years, 19-30 years, 30-51 years, 51-70 years and 71 years and older. Median sodium levels were derived from the FLIP and Canadian Nutrient File databases and were weighted according to the number of products in the food sub-category. Correction factors of 20% for packaged foods and 10% for restaurant foods were applied to the derived sodium estimates to approximate habitual sodium consumption.
References
Arcand JA, Abdulaziz K, Bennett C, L’Abbé MR, Manuel DG. (2014) Developing a Web-based sodium screening tool for personalized assessment and feedback. Applied Physiology, Nutrition and Metabolism 39(3) doi:10.1139/apnm-2013-0322
Chronic Kidney Disease
CKD risk equations have been previously adopted into clinical practice, but there are no equivalent methods for members of the general public to engage with and identify their risk self reported and readily available risk factors.
The CKD calculator is a web based too designed to spread awareness and help people improve their health behaviours as they relate to CKD. The calculator asks general questions about a person’s health, and behaviours to determine score that describes a person’s risk of developing CKD in the next 8 years. Detailed clinical information such as blood pressure or laboratory measurements are not necessary for an accurate risk assessment, to help ensure that the calculator is appropriate for all audiences.
To do this We identified a preliminary list of candidate predictors using information from a systematic literature search and availability in the Canadian Community Health Survey (CCHS). Data from 22,200 individuals in Ontario, Canada was used to develop a model, which was then validated using 15, 522 individuals from the UK Biobank to ensure predictive performance. As many individuals may not know their baseline eGFR level, we developed separate models for individuals with and without a baseline eGFR.
The time-averaged c-stat for models with and without baseline eGFR were 83.5 (95%CI 82.2 to 84.9, without 81.0 95%CI79.8 to 82.4) The c-stat ranged from 86.3 to 77.6 from years 1 to 8 with a baseline eGFR and 82.8 to 71.2 from years 1 to 8 without a baseline eGFR.
The model calibration slope was 1.002(95%CI 0.944-1.061) and 0.991 (95%CI 0.923-1.06) with and without baseline eGFR, respectively and over prediction among those with a high estimated risk.The predicted 5-year risk was 0.0967(with baseline eGFR) and 0.0972 (no baseline eGFR) compared to an observed risk of 0.0977. Across years 1 to 8, the Brier score ranged from 1.4 to 10.2 for the model with baseline eGFR and 1.4 to12.0 without.
References
Noel AJ, Eddeen AB, Manuel DG, Rhodes E, Tangri N, Hundemer GL, Tanuseputro P, Knoll GA, Mallick R, Sood MM. A health survey-based prediction equation for incident CKD (Development and External Validation of The Chronic Kidney Disease Population Risk Tool). Submitted to CJASN.
Dietary Pattern Calculator
In development of the DietaryPattern Calculator (DiPaC), a scoping review was conducted to identify currently available short diet quality assessment tools. The identified tools mainly focused on individual nutrients or food groups or were developed for a specific population, but few ascertained overall dietary patterns.
Subsequently, 24-hour dietary recalls from the nationally representative Canadian Community Health Survey(CCHS)-Nutrition 2015 (n = 13,958) were used to derive and validate a dietary pattern informed by the scoping review using weighted partial least squares.The dominant dietary pattern in CCHS-Nutrition 2015 was characterized by high consumption of fast foods, carbonated drinks, and salty snacks and low consumption of whole fruits, orange vegetables, other vegetables and juices, whole grains, dark green vegetables, legumes, and soy.
DiPaC, which demonstrated high validity and intermediate reliability (internal consistency = 0.47–0.51) can be used by the public, clinicians, and researchers for a quick and robust assessment of diet quality.
References
Jessri M, Jacobs A, Ng Alena (Praneet), Bennett C, Quinlan A, Nutt C, Brown J, Hennessy D, Manuel DG. (2023) Development and Evaluation of the Dietary Pattern Calculator (DiPaC) forPersonalized Assessment and Feedback. Canadian Journal of Dietetic Practice and Research https://doi.org/10.3148/cjdpr-2023-013
Tools
Project Big Life Planning Tool
The Project Big Life Planning Tool is a web-based application that has been developed for health planners to use our algorithms for population analytics.
https://www.projectbiglife.ca/planning-tool
CCHS Flow
cchsflow supports the use of the Canadian Community Health Survey (CCHS) by transforming variables from each cycle into harmonized, consistent versions that span survey cycles (currently, 2001 to 2014).
https://big-life-lab.github.io/cchsflow/
Algorithm Viewer
An interactive algorithm visualization tool that displays the contribution of the main predictors and exposure interactions.
http://algorithm-viewer.projectbiglife.ca/#/
Github
Check out PBL's Github repositories - Big Life Lab Flow (BLLFlow) - a workflow for open, reproducible research.
https://github.com/Big-Life-Lab
