umu.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • harvard1
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Predicting 10-year cardiovascular disease morbidity and mortality in northern Sweden from a multietiological perspective -: the application of boosted regression tree in public health. Ka Chun Tsang Supervisors: Joacim
Umeå University, Faculty of Medicine, Department of Public Health and Clinical Medicine, Epidemiology and Global Health.
2016 (English)Independent thesis Advanced level (degree of Master (Two Years)), 10 credits / 15 HE creditsStudent thesis
Abstract [en]

Background: Statistical models derived from machine learning techniques perform better than traditional statistical models in prediction, but with results that are difficult to interpret for clinical application. Among them, boosted regression trees is shown to have good prediction and comparatively easier to interpret with visualisation aids.

Introduction: Cardiovascular morbidity and mortality prediction is important for health promotion. The variables selected in most existing predictions models however stay at the physical aspect, despite proof of moderation effects brought by contextual factors, like socio-economic status, occupational health, and quality of life.

Objective: To apply the boosted regression trees to predict cardiovascular morbidity and mortality with predicting variables from a wider range than the physical aspect.

Design: Using cross-sectional data for 40 to 60 year old adults from the Vasterbotten Intervention Programme database built since 1989 in northern Sweden, and, the hospitalised and death registry. At the beginning, all variables, except the detailed dietary section, are included for boosted regression tree analysis to examine the relative influence on predicting CVD mortality.

Result: On top of existing variables used by Framingham score, mean arterial blood pressure, year of birth, and detailed smoking habit and quantity should be added for CVD prediction. Meanwhile, none of the psychosocial variables are found to be strong predictors.

Conclusion: Traditional predictors are useful in identifying one’s CVD risk. The result suggests that the use of modern statistical tools, like boosted regression tree, may help improve the predictability.

Place, publisher, year, edition, pages
2016. , 44 p.
Series
Centre for Public Health Report Series, ISSN 1651-341x ; 2016:41
Keyword [en]
Predicting 10-year cardiovascular, disease morbidity, mortality, northern Sweden, boosted regression
National Category
Public Health, Global Health, Social Medicine and Epidemiology
Identifiers
URN: urn:nbn:se:umu:diva-131699OAI: oai:DiVA.org:umu-131699DiVA: diva2:1075334
External cooperation
Västerbottens läns landsting - Margareta Norberg
Educational program
Master's Programme in Public Health
Presentation
2016-05-23, Umeå Universitet, Umeå, 08:46 (English)
Supervisors
Examiners
Available from: 2017-02-17 Created: 2017-02-17 Last updated: 2017-02-17Bibliographically approved

Open Access in DiVA

No full text

Search in DiVA

By author/editor
Tsang, Ka Chun
By organisation
Epidemiology and Global Health
Public Health, Global Health, Social Medicine and Epidemiology

Search outside of DiVA

GoogleGoogle Scholar

Total: 1 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • harvard1
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf