This interactive forest plot shows estimated associations and 95% confidence intervals (CI) related to the ARIC Resilience Relevance Framework. Here the estimates provide information on the Adversity of the specific combinations of stressors, physical function measures, and downstream outcomes.
More information about the definitions and data used may be found at the following locations:
More information about the definitions and data used may be found at the following locations:
Myocardial infarctions, heart failures, and strokes were ascertained as adjudicated diagnoses from physicians.
The New Cancer Diagnosis stressor was derived
by looking at the afucomp7k_g variable
(Year of Cancer Diagnosis) in the AFU-SAFU
dataset.
The Hospitalization stressor was derived by
looking at variables of the form
afucomp_hosp_mmyyyy* (Hospitalization Dates) in the AFU-SAFU dataset.
The Self Reported Poor Overall Health stressor
was derived by looking at the
afucomp6_a variable (Self-Described Health Compared to Others at Similar
Age) in the AFU-SAFU dataset.
The Surgery, Fracture, Infection,
and Pneumonia stressors were derived by looking
at variables of the form
afucomp_hosp_reason* (Hospitalization Reasons) in the AFU-SAFU dataset.
The Widowed stressor was derived by looking at
the afucomp40a_d variable (marital status) in the AFU-SAFU dataset.
The Fall stressor was derived by looking at the
gnb19 variable (Experienced a fall within the past 6 months of form
completion date) in the GNB dataset.
Self-reported falls were classified as present if participants responded “yes” to the following question: “In the past 12 months, did you fall?”
The Dr. Fall stressor was derived by looking at
the gnb22 variable (Saw a Doctor for the fall experienced in the past 6
months) in the GNB dataset.
The Tired (Fatigued) stressor was derived by
looking at variables the gnb14-gnb18 in
the GNB dataset.
The Physical Ability stressor was derived by
looking at variables the gnb2-gnb13 in
the GNB dataset.
The Depression stressor was derived by looking
at variables gne23 and
gne24 in the GNE dataset.
Depression was defined as self-reported depressive symptoms by answering “yes” to “During the past month, have you been bothered by feeling down, depressed or hopeless?” or “During the past month, have you been bothered by little interest or pleasure in doing things?”
The Poor Memory stressor was derived by looking
at variables gne13-gne19 in
the GNE dataset.
Subjective memory complaints were self-reported on a questionnaire by answering yes to “Do you have any complaints concerning your memory?” “Do other people find you forgetful?” “Do you ever forget names of family members or friends?” “Do you often forget where things are left?” “Do you often use notes to avoid forgetting things?” “Do you ever have difficulties in finding particular words?” “Did you ever lose your way in your neighborhood?” “Do you think more slowly than you used to?” “Do your thoughts ever become confused?” or “Do you have concentration problems?”
The Social Support stressor was derived by
looking at variables gne26 and
gne27 in the GNE dataset.
Poor social support was defined by participants that were widowed or had a low support system, assessed by answering “no” to “Can you count on anyone to help you when you need to make difficult decisions or talk over problems?” or “Can you count on anyone to help you with daily tasks like grocery shopping, housecleaning, cooking, telephoning, or giving you a ride?”
More information about the definitions and data used may be found at the following locations:
Good Gait Speed is defined as a gait speed greater than or equal to one meter per second.
Good Balance is defined as having an SPPB Balance measurement of 4.
Good Chair Stands is defined as having an SPPB Chair Stand measurement of 4.
Good Overall SPPB is defined as having an overall SPPB measurement of 12.
Non-Exhaustion (a good Exhaustion measurement)
is based on exhaustion from depression. For Visit 5,
it is defined as having EXHAUSTCOMP==0 in
the
DERIVE52 dataset. For Visit 6, it is defined as having
EXHAUST61==0 in the in
DERIVE61 dataset. The Frailty Exhaustion criteria was met (i.e.,
EXHAUSTCOMP==1 or
EXHAUST61==1) if the participant
responded 2 ("Much or most of the time")
to CES3 (“I felt everything I did was an
effort”) or CES11 (“I could not get
going”) in the
CES dataset.
Non-Unintentional Weight Loss (a good
Unintentional Weight Loss measurement) is based on
weight loss and BMI. For Visit 5, it is defined as
having WEIGHTLOSSCOMP==0 in the
DERIVE52 dataset. For Visit 6, it is defined as having
WTLOSSCOMPA61==0 in the in
DERIVE61 dataset. The Frailty Weight Loss criterion was met (i.e.,
WEIGHTLOSSCOMP==1 or
WTLOSSCOMPA61==1) when a PPT lost more
than 10% of their weight as measured at the previous
visit or if their BMI was less than 18.5.
Non-Low Grip Strength (a good Low Grip Strength
measurement) is based on grip strength, conditional on
gender and BMI. For Visit 5, it is defined as having
GRIPCOMP==0 in the
DERIVE52 dataset. For Visit 6, it is defined as having
GRIPSTRENGTHCOMP61==0 in the in
DERIVE61 dataset. The Low Grip Strength criterion was met (i.e.,
GRIPCOMP==1 or
GRIPSTRENGTHCOMP61==1) if grip strength
was lower than the specified cutpoints adjusted for
gender and BMI.
Non-Slow Walking (a good Slow Walking
measurement) is based on slowness by walking. For
Visit 5, it is defined as having
WALKCOMP==0 in the
DERIVE52 dataset. For Visit 6, it is defined as having
WALKSPEEDCOMP61==0 in the in
DERIVE61 dataset.
Non-Low Grip Strength (a good Low Grip Strength
measurement) is based on grip strength, conditional on
gender and BMI. For Visit 5, it is defined as having
GRIPCOMP==0 in the
DERIVE52 dataset. For Visit 6, it is defined as having
GRIPSTRENGTHCOMP61==0 in the in
DERIVE61 dataset. The Low Grip Strength criterion was met (i.e.,
GRIPCOMP==1 or
GRIPSTRENGTHCOMP61==1) if grip strength
was lower than the specified cutpoints adjusted for
gender and BMI.
Non-Low Physical Activity (a good Low Physical
Activity measurement) is based on low physical
activity during leisure time. For Visit 5, it is
defined as having PACCOMP20==0 in the
DERIVE52 dataset. For Visit 6, it is defined as having
LOWENERGYCOMPCOMP61==0 in the in
DERIVE61 dataset.
A participant was considered to be Robust if
all ARIC frailty indicators were marked as being
absent. For Visit 5, it is defined as having
FRAILTY51==0 in the
DERIVE52 dataset. For Visit 6, it is defined as having
FRAILTY61a==0 in the in
DERIVE61 dataset.
Loosely speaking, we define a Relative Prevalence
Ratio (RPR) with respect to the model
logistic y i.x as follows:
RPR = [Pr(y==1|x==1)]/[Pr(y==1|x==0)].
The Good Function RPRs were derived from models
of the form
logistic good`measure'6 i.`stressor` $adjustors if
good`measure'5 == 1
with
adjustors = "i.male age5 i.black bmi5". Here, we have good`measure'6==0 if a
participant either had a poor measurement at visit 6
or if the participant was known to be dead at visit 6.
The Good Function RPRs (Condtional on Being Alive at Visit 6) were defined similarly to the Good Function RPRs, except for the fact that participants who were known to be dead at Visit 6 were excluded from analysis.
The
Downstream Outcome Odds Ratios and RPRs were
derived from models of the form
logistic event i.g_`measure`_only6 if
good`measure'5 == 1 & `stressor' == 1
with
event = !missing(dateofdeath). Here,
g_`measure`_only6
is an indicator for good performance at V6 and is
defined only for participants who were alive and
performed at V6.
The Downstream Outcome Hazard Ratios were
derived from models of the form
stcox i.g_`measure'_only6 if good`measure'5 == 1 &
`stressor' == 1
after running
stset time_to_event_fromv6, failure(event ==
1).