Abstract
To determine the extent to which excess mortality after fractures attributable to particular causes at specific skeletal sites can be predicted using data about all medical diagnoses, we conducted a historical cohort study among 1991 Olmsted County, Minnesota, residents aged ≥50 years who experienced any fracture in 1989 to 1991 and who were followed passively for up to 22 years for death from any cause. We used a machine learning approach, gradient boosting machine (GBM) modeling, to determine whether the comorbid conditions present at the time of fracture and those that arose subsequently could, in aggregate, identify patients at the greatest increased risk of death. During 21,867 person-years of follow-up, 1245 deaths were observed when 1061 were expected (standardized mortality ratio, 1.2; 95% confidence interval [CI] 1.1-1.2). Patients presented with a median history of 26 comorbid conditions each as assessed by the Clinical Classification Software system and 57 each over the total duration of follow-up. Using all available information, the excess deaths could be predicted with good accuracy (c-index ≥0.80) in 89% of the GBM models built for patients with different types of fracture; in one-third of the models, the c-index was ≥0.90. The conditions most prominent in the GBM prediction models were also reflected in the specific causes of death that were elevated, suggesting the influence of confounding on the relationship. However, the predominant comorbid conditions were mainly those responsible for mortality in the general population, rather than the specific diseases most closely associated with secondary osteoporosis. To reduce long-term deaths in the fracture population as a whole, a more general approach to the fracture patient is indicated.
| Original language | English (US) |
|---|---|
| Pages (from-to) | 1681-1690 |
| Number of pages | 10 |
| Journal | Journal of Bone and Mineral Research |
| Volume | 29 |
| Issue number | 7 |
| DOIs | |
| State | Published - 2014 |
Fingerprint
Keywords
- AGING
- FRACTURES
- GENERAL POPULATION STUDIES
- MORTALITY
- OSTEOPOROSIS
- STATISTICAL METHODS
ASJC Scopus subject areas
- Orthopedics and Sports Medicine
- Endocrinology, Diabetes and Metabolism
- Medicine(all)
Cite this
Predictors of excess mortality after fracture : A population-based cohort study. / Melton, L. Joseph; Atkinson, Elizabeth J.; St. Sauver, Jennifer; Achenbach, Sara J.; Therneau, Terry M; Rocca, Walter A; Amin, Shreyasee.
In: Journal of Bone and Mineral Research, Vol. 29, No. 7, 2014, p. 1681-1690.Research output: Contribution to journal › Article
}
TY - JOUR
T1 - Predictors of excess mortality after fracture
T2 - A population-based cohort study
AU - Melton, L. Joseph
AU - Atkinson, Elizabeth J.
AU - St. Sauver, Jennifer
AU - Achenbach, Sara J.
AU - Therneau, Terry M
AU - Rocca, Walter A
AU - Amin, Shreyasee
PY - 2014
Y1 - 2014
N2 - To determine the extent to which excess mortality after fractures attributable to particular causes at specific skeletal sites can be predicted using data about all medical diagnoses, we conducted a historical cohort study among 1991 Olmsted County, Minnesota, residents aged ≥50 years who experienced any fracture in 1989 to 1991 and who were followed passively for up to 22 years for death from any cause. We used a machine learning approach, gradient boosting machine (GBM) modeling, to determine whether the comorbid conditions present at the time of fracture and those that arose subsequently could, in aggregate, identify patients at the greatest increased risk of death. During 21,867 person-years of follow-up, 1245 deaths were observed when 1061 were expected (standardized mortality ratio, 1.2; 95% confidence interval [CI] 1.1-1.2). Patients presented with a median history of 26 comorbid conditions each as assessed by the Clinical Classification Software system and 57 each over the total duration of follow-up. Using all available information, the excess deaths could be predicted with good accuracy (c-index ≥0.80) in 89% of the GBM models built for patients with different types of fracture; in one-third of the models, the c-index was ≥0.90. The conditions most prominent in the GBM prediction models were also reflected in the specific causes of death that were elevated, suggesting the influence of confounding on the relationship. However, the predominant comorbid conditions were mainly those responsible for mortality in the general population, rather than the specific diseases most closely associated with secondary osteoporosis. To reduce long-term deaths in the fracture population as a whole, a more general approach to the fracture patient is indicated.
AB - To determine the extent to which excess mortality after fractures attributable to particular causes at specific skeletal sites can be predicted using data about all medical diagnoses, we conducted a historical cohort study among 1991 Olmsted County, Minnesota, residents aged ≥50 years who experienced any fracture in 1989 to 1991 and who were followed passively for up to 22 years for death from any cause. We used a machine learning approach, gradient boosting machine (GBM) modeling, to determine whether the comorbid conditions present at the time of fracture and those that arose subsequently could, in aggregate, identify patients at the greatest increased risk of death. During 21,867 person-years of follow-up, 1245 deaths were observed when 1061 were expected (standardized mortality ratio, 1.2; 95% confidence interval [CI] 1.1-1.2). Patients presented with a median history of 26 comorbid conditions each as assessed by the Clinical Classification Software system and 57 each over the total duration of follow-up. Using all available information, the excess deaths could be predicted with good accuracy (c-index ≥0.80) in 89% of the GBM models built for patients with different types of fracture; in one-third of the models, the c-index was ≥0.90. The conditions most prominent in the GBM prediction models were also reflected in the specific causes of death that were elevated, suggesting the influence of confounding on the relationship. However, the predominant comorbid conditions were mainly those responsible for mortality in the general population, rather than the specific diseases most closely associated with secondary osteoporosis. To reduce long-term deaths in the fracture population as a whole, a more general approach to the fracture patient is indicated.
KW - AGING
KW - FRACTURES
KW - GENERAL POPULATION STUDIES
KW - MORTALITY
KW - OSTEOPOROSIS
KW - STATISTICAL METHODS
UR - http://www.scopus.com/inward/record.url?scp=84903275594&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84903275594&partnerID=8YFLogxK
U2 - 10.1002/jbmr.2193
DO - 10.1002/jbmr.2193
M3 - Article
C2 - 24677169
AN - SCOPUS:84903275594
VL - 29
SP - 1681
EP - 1690
JO - Journal of Bone and Mineral Research
JF - Journal of Bone and Mineral Research
SN - 0884-0431
IS - 7
ER -