Study of quality-of-life trajectories at late life: impact of statistical models and study design
| dc.contributor.author | Liu, Yixiu | |
| dc.contributor.examiningcommittee | Tate, Robert (Community Health Sciences) St. John, Philip (Internal Medicine) | en_US |
| dc.contributor.supervisor | Jiang, Depeng (Community Health Sciences) | en_US |
| dc.date.accessioned | 2019-09-04T12:49:04Z | |
| dc.date.available | 2019-09-04T12:49:04Z | |
| dc.date.issued | 2019 | en_US |
| dc.date.submitted | 2019-08-27T03:42:17Z | en |
| dc.date.submitted | 2019-09-03T23:37:30Z | en |
| dc.degree.discipline | Community Health Sciences | en_US |
| dc.degree.level | Master of Science (M.Sc.) | en_US |
| dc.description.abstract | The importance of understanding the developmental pathway of physical functioning for old population is well-recognized. To accurately estimate the true trajectory of physical functioning, a well-designed longitudinal study is needed. The relationship between physical functioning trajectories and survival are not well-understood, which requires more advanced statistical approaches. The purpose of this study is to investigate the impact of statistical approaches and study design in a longitudinal study on health and aging. Growth curve models were first conducted only using a sample of 200 living males from the Manitoba Follow Up Study (MFUS) to identify the shape of physical functioning trajectory. Then using all available sample over 12 years from MFUS, we compared three statistical approaches: 1) the growth curve model; 2) the extended Cox model; and 3) the joint model for longitudinal and survival data. We investigated the impact of study design on parameter estimations of physical functioning trajectory and survival function. The overall trend of physical functioning was decreasing at an accelerating rate over year. According to the best joint model, those with worse physical functioning or higher declining rate would have higher risk of death. Joint model approach provides better performance in aging studies with an interest on the association between longitudinal markers and the risk of death. The impact of study design on parameter estimations for describing the longitudinal trajectory is minimal as soon as we have enough data points to estimate the individual shape of trajectory (e.g., three points for linear growth and four points for quadratic growth) and large sample size. The influence of data collection cycle on the association of current longitudinal marker and risk of death is relatively small. The influence of data collection cycle on the association of derivative of longitudinal markers and the risk of death in the survival submodel is large. | en_US |
| dc.description.note | October 2019 | en_US |
| dc.identifier.uri | http://hdl.handle.net/1993/34141 | |
| dc.language.iso | eng | en_US |
| dc.rights | open access | en_US |
| dc.subject | Aging longitudinal studies | en_US |
| dc.subject | Joint models | en_US |
| dc.subject | Growth curve models | en_US |
| dc.subject | The extended Cox model | en_US |
| dc.subject | Frequency of data collection | en_US |
| dc.subject | Dynamic prediction | en_US |
| dc.title | Study of quality-of-life trajectories at late life: impact of statistical models and study design | en_US |
| dc.type | master thesis | en_US |