Risk Factors and Precipitants of Severe Disability Among Community-Living Older Persons

Key Points Question What are the relative contributions of traditional risk factors and intervening illnesses or injuries (ie, precipitants) associated with severe disability that develops progressively vs catastrophically? Findings In this cohort study of 754 nondisabled community-living persons aged 70 years or older who were followed longitudinally for nearly 19 years, the associations of intervening illnesses or injuries with severe disability were much more pronounced than those of the risk factors, with hazard ratios up to 20-fold greater for progressive disability and 177-fold greater for catastrophic disability. Meaning The findings of this study suggest that to reduce the burden of severe disability, more aggressive efforts will be needed to prevent and manage intervening illnesses or injuries and to facilitate recovery after these debilitating events.

Abbreviations: SD, standard deviation a The number of hospitalizations are provided for each of the three disability groups.
b Grouped into distinct diagnostic categories, as described in the "Methods" section, and presented in order of highest to lowest rates in the not severe disability group.
c For each calculation, the numerator includes the number of admissions during the follow-up period for the specific reason, while the denominator includes the number of admissions for all reasons combined.
d Includes psychiatric admissions and responses that could not otherwise be categorized.
© b Grouped into distinct diagnostic categories, as described in the "Methods" section, and presented in order of highest to lowest rates in the not severe disability group.
c For each calculation, the numerator includes the number of emergency department visits during the follow-up period for the specific reason, while the denominator includes the number of emergency department visits for all reasons combined.
d Includes unclassifiable complaints such as feeling "weak," "tired," or generally unwell, with no other specifying features.
© 2020 Gill TM et al.JAMA Network Open.Abbreviations: SD, standard deviation a The number of episodes with restricted activity are provided for each of the three disability groups.
b Presented in order of highest to lowest rates in the not severe disability group.
c For each calculation, the numerator includes the number of restricted activity episodes during the follow-up period for the specific reason, while the denominator includes the number of restricted activity episodes for all reasons combined.Because participants could provide more than 1 reason for their restricted activity, the percentages for each type of severe disability do not sum to 100.
© 2020 Gill TM et al.JAMA Network Open.e Assumes that all of the 113 decedents experienced catastrophic severe disability only in the last month of the personinterval that was truncated by death.

eTable 1 .
Reasons for Hospitalization According to Type of Severe Disability eTable 2. Reasons for Emergency Department Visits According to Type of Severe Disability eTable 3. Reasons for Restricted Activity According to Type of Severe Disability eTable 4. Sensitivity of Associations Between Precipitants and Progressive Severe Disability to the Competing Risk of Death eTable 5. Sensitivity of Associations Between Precipitants and Catastrophic Severe Disability to the Competing Risk of Death This supplementary material has been provided by the authors to give readers additional information about their work.© 2020 Gill TM et al.JAMA Network Open.

eTable 1 .
Reasons for Hospitalization According to Type of Severe Disability a

eTable 3 .Reason
Reasons for Restricted Activity According to Type of Severe Disability a

eTable 4 .
Abbreviations: CI, confidence interval a Follow-up was truncated by death in 113 (3.2%) of the 3550 person-intervals.bCalculated using pooled logistic regression with a complementary log-log link and generalized estimating equations with robust standard errors based on an autoregressive correlation structure, as described in the Methods.The models included the same set of factors as in Figure1.cThe occurrence of progressive severe disability was imputed for each month from the time of death through the end of the person-interval, with censoring at the first occurrence of the outcome for all 113 decedents.d Assumes that none of the 113 decedents experienced progressive severe disability from the time of death through the end of the person-interval.eAssumes that all of the 113 decedents experienced progressive severe disability only in the last month of the personinterval that was truncated by death.
Abbreviations: CI, confidence interval a Follow-up was truncated by death in 113 (3.2%) of the 3550 person-intervals.bCalculated using pooled logistic regression with a complementary log-log link and generalized estimating equations with robust standard errors based on an autoregressive correlation structure, as described in the Methods.The models included the same set of factors as in Figure1.c The occurrence of progressive severe disability was imputed for each month from the time of death through the end of the person-interval, with censoring at the first occurrence of the outcome for all 113 decedents.d Assumes that none of the 113 decedents experienced progressive severe disability from the time of death through the end of the person-interval.
2020 Gill TM et al.JAMA Network Open.
Abbreviations: SD, standard deviation a The number of emergency department visits that did not lead to hospitalization are provided for each of the three disability groups.These visits were identified from 371 (not severe disability), 43 (progressive severe disability), and 72 (catastrophic severe disability) participants.Reasons are missing for 34 (3.4%), 1 (1.4%), and 5 (5.6%) emergency department visits in the not severe, progressive severe, and catastrophic severe disability groups, respectively.
eTable 5. Sensitivity of Associations between Precipitants and Catastrophic Severe Disability to the Competing Risk of