frailty as deficit accumulation - hkag ppt/ps1-2_rockwood.pdf · frailty as deficit accumulation...

Frailty as deficit accumulation 

Kenneth Rockwood MD, FRCPC, FRCPDivision of Geriatric Medicine

Dalhousie University & Capital District Health AuthorityHalifax, Canada

Read it as: Rockwood K, Mitnitski A: Frailty defined by deficit accumulation and geriatric medicine defined by frailty. Clinics in Geriatric Medicine 2011 Feb;27(1):17‐26.

Disclosures

With colleagues, I am applying to various Canadian government university‐industry schemes for funding to commercialize a version of the Frailty Index, based on a Comprehensive Geriatric assessment.

My colleagues and I are always on the look out for clever young doctors who have undergraduate degrees in engineering, physics, mathematics …

3

Rockwood et al. Can Med Association 1994; 150:499-507Rockwood et al. J Am Geriatric Society 1996; 44:578-82

Health

Attitudes towardHealth and

health practices

Resources

Caregiver

Illness

Disability

Dependence on Others

Burden on the caregiver

Frailty is complex and dynamic

5

Operationalizing frailtyVariables are highly specified:

prototype is the frailty phenotype– Slow mobility– Weakness– Weight loss– Decreased activities– Exhaustion

• Fried et al.,. 2001;56 J Gerontol A Biol Sci Med Sci (3):M146-56.

Variables are hardly specified: prototype is the Frailty Index– Count health deficits (30-100)

• age associated but does not saturate;

• associated with adverse outcome

• <5% missing data– Divide by the number of deficits

considered.• Mitnitski et al., ScientificWorldJ

2001;1:323-326.• Searle et al., BMC Geriatr 2008;8:24.

6

The building blocks of life do not age

http://www.cerritos.edu/earth-science/images/radioa1.gif

Frailty as deficit accumulation: with age, most problems become more common

(Canadian National Population Health Survey, n= 66,580)

0 20 40 60 80 1000

0.1

0.2

0.3

0.4

0.5

0.6

0.7

arthritis

0 20 40 60 80 1000

0.02

0.04

0.06

0.08

0.1

0.12

0.14vision problems

0 20 40 60 80 1000

0.05

0.1

0.15

0.2

0.25

0.3

0.35Mobility disability

0 20 40 60 80 1000

0.02

0.04

0.06

0.08

0.1

0.12

0.14

thyroid problems

Age (years)

Rockwood & Mitnitski Rev Clin Gerontol 2007;18:1-12.

The frailty index is calculated as:

The number of deficits that an individual has

The total number of deficits considered

10 deficits present in an individual = Frailty index score of 10/40 =0.25

40 deficits considered in total

9

Table 1. List of deficits used in the frailty index.

DeficitsLevels

1 Eyesight 52 Hearing 53 Help to eat 34 Help to dress 35 Ability to take care of appearance 36 Help to walk 37 Help to get in and out of bed 38 Help to go to the bathroom 39 Help to take a bath or shower 310 Help to use the telephone 311 Help to travel beyond walking distance 312 Help with shopping 313 Help to prepare own meals 314 Help to do housework 315 Ability to take medications 316 Ability to handle own money 317 Self-rated health 518 Troubles prevent normal activities 319 Lives alone 220 Having a cough 221 Feeling tired 222 Nose stuffed up or sneezing 223 High blood pressure 224 Heart and circulation problems 225 Stroke or effects of stroke 226 Arthritis or rheumatism 227 Parkinson’s disease 228 Eye trouble 229 Ear trouble 230 Dental problems 231 Chest problems 232 Trouble with stomach 233 Kidney trouble 234 Losing control of bladder 235 Losing control of bowels 236 Diabetes 237 Trouble with feet or ankles 238 Skin problems 239 Fractures 2 40 Trouble with nerves 2

Frailty index distribution

02

46

810

Per

cent

0 .2 .4 .6 .8Frailty Index

• Range = 0 to 0.66, mean 0.16•The higher the Frailty Index, the more

frail the individual

Measuring frailty as anindex of (40) deficits

National Population Health Survey - Mean Frailty Index at each cycle in relation to age

Frai

lty In

dex

(or p

ropo

rtion

of h

ealth

def

icits

)

Age (years)10 20 30 40 50 60 70 80 90 100

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

A

Pro

porti

onal

dis

tribu

tion

Rockwood et al., CMAJ 2011; E-pub April 28

5. Why the deficit count matters: transitions from n deficits to death during 5 years; Canadian Study of Health & Aging, N=8,547

0 5 10 150.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Num ber of deficits, n

The

prob

abilit

y of

dea

th

0 5 10 150.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Num ber of deficits, n

The

prob

abilit

y of

dea

th

Of 8,547 people at baseline, only 18 had >17/31 possible deficits, and only 7 (of 5586) had >17/31 at follow-up

Survival limit close to the frailtyIndex of about 0.7

A limit to of the number of deficits suggests exhaustion of reserve capacity – is it operationalizable clinically?

Mitnitski, Bao, Rockwood. Mech Ageing Dev 2006;127:490-3. Rockwood & Mitnitski Mech Ageing Dev 2006;127:494-6.

For men & women, deficit accumulation is highly    related with mortality (r>0.95); men have a higher  

death rate than women 

0

0.2

0.4

0.6

0.8

18-

year

Dea

th R

ate

0.1 0.2 0.3 0.4 0.5 0.6 0.70

Frailty Index

women

men

Shi et al., BMC Geriatr. 2011 Apr 20;11:17

Mitnitski, et al., J Am Geriatr Soc, 2005;53:2184-9

Mea

n ac

cum

ulat

ion

of d

efic

its

Legend

65 70 75 80 85 90 95

ALSACSHA-screenCSHA-examNHANESNPHSSOPSBreast cancerCSHA-instMyoc InfarctUS-LTHSH70-75

0.1

0.2

0.3

0.5

1.0

0.05

Age (years)

Clinical and institutional samples, n=2,573

The slope is ~0.03Community samplesn=33,559

Log scale

Deficits accumulate characteristically, both between groups (community vs. institution/ clinical) and  within groups*

Slope <0.01

15

Failure kinetics of systems with different levels of redundancy

From Gavrilov & Gavrilova Sci Aging Knowledge Env, 2003; 28:1-10

The rate of deficit accumulation slows as the value of the Frailty Index (here based on Comprehensive 

Geriatric Assessment) increasesLo

g of

the

Frai

lty In

dex

Age, years

70 75 80 85 90 95 100-3

-2.5

-2

-1.5

-1

-0.5

0

Rockwood, Rockwood, Mitnitski., J Am Geriatrics Soc, 2010;58:318-323

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.80

100

200

300

400

500

600

700

800

900

Frailty Index

Cou

nt

Distribution of Frailty Indexes at Each Wave

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.80

0.02

0.04

0.06

0.08

0.1

0.12

0.14

Frailty Index

Den

sity

Proportional Distribution of Frailty Indexes at Each Wave

Distribution of the Frailty Indexin 4 successive waves of the Chinese Longitudinal Health and Longevity Study;Subjects aged 80-99 years; n= 6664

Bennett et al., submitted

18

5‐year transitions between different states of health (empty circles), replicated 5 years later 

(solid circles)*

Legend:Empty circles: CSHA-1 CSHA-2Solid circles: CSHA-2 CSHA-3

Goodness of fitr = 0.99

*Mitnitski, Bao, Rockwood., Mech Ageing Dev 2006, 127;490-493

Number of deficits

0 5 100

0.1

0.2

0 5 100

0.1

0.2

0 5 100

0.1

0.2

0 5 100

0.1

0.2

0 5 100

0.1

0.2

0 5 100

0.1

0.2

0 5 100

0.1

0.2

0 5 100

0.1

0.2

0 5 100

0.1

0.2

0.3

0 5 100

0.1

0.2

0.3

0 5 100

0.1

0.2

0.3

0.3

0.3

The

trans

ition

pro

babi

litie

s

0 5 100

0.1

0.2

0 5 100

0.1

0.2

0 5 100

0.1

0.2

0 5 100

0.1

0.2

0 5 100

0.1

0.2

0 5 100

0.1

0.2

0 5 100

0.1

0.2

0 5 100

0.1

0.2

0 5 100

0.1

0.2

0.3

0 5 100

0.1

0.2

0.3

0 5 100

0.1

0.2

0.3

0.3

0.3

The

trans

ition

pro

babi

litie

s

n=0 n=1 n=2 n=3

n=7n=6n=5n=4

n=8 n=9 n=10

)1(!)( )(

ndn

k

nk PeknP

The model

Four parameters of the model and their Interpretation

0k0 5 10

0

0.2 Average number of deficits given zero deficits at baseline

nn kk 110 10

0

0.2

0.4 The difference between the average number of deficits at the two incremental deficit numbers at baseline

0 10 n

nPP dnd 20lnln The intercept and the slope in the probability of deathas a function of the number of deficits at baseline

How can we assess frailty in older adults who are ill?

Defining frailty by counting deficits: data from a medical history & examination 

What is added by a Comprehensive Geriatric Assessment 

Learning from other complex systems applications

24

Which patient is the more frail?Which patient is the more acutely ill?

0 10 20 30 40 50 60 700

0.2

0.4

0.6

0.8

1

A Frailty Index based on a Comprehensive Geriatric Assessment identifies a group at the highest risk of dying (some of whom live 18 months).

FI-CGA

0.1

0.2

0.3

0.40.5

Survival time (months)

Sur

viva

l pro

babi

lity

Rockwood, Rockwood, Mitnitski, J Am Geriatric Soc 2010;58:318-323

Parks et al., J Gerontol Biol Sci, in press, 2011.

27

Studies of frailty as deficit accumulation.

There is remarkable consistency in: • how deficits accumulate with age. • the limit to how many things can be wrong.• how deficit counts change over time.

Some clinical lessons:•How can we count what people have wrong with them?• Does our clinical intuition about the “stability” of deficit accumulation mislead?

28

AcknowledgmentsFunding sources:• Fountain Innovation Fund of

the QEII Health Sciences Foundation

• Canadian Institutes of Health Research

• Mathematics of Information Technology and Computer Science program, National Research Council

• Alzheimer Society of Canada• National Natural Science

Foundation of China• Dalhousie Medical Research

Foundation

Colleagues & students:• Arnold Mitnitski• Nadar Fallah• Xiaowei Song• Ruth Hubbard• Melissa Andrew• Michael Rockwood• Samuel Searle• Paige Moorhouse, Laurie

Mallery

29

Fig. 1. Cumulative distributions of frailty index scores for people defined as ‘robust’, ‘pre-frail’,

and ‘frail’*

0 0.2 0.4 0.6 0.80

0.2

0.4

0.6

0.8

1

Frailty index value

Cum

ulat

ive

Pro

porti

on

Robust

Pre-frail

Frail

0.30

0.12

0.45

*Rockwood, Andrew, Mitnitski. J Gerontol Med Sc, 2007;62:738-743.

30

Fig. 1. Cumulative distributions of frailty index scores for people defined as ‘robust’, ‘pre-frail’,

and ‘frail’*

0 0.2 0.4 0.6 0.80

0.2

0.4

0.6

0.8

1

Frailty index value

Cum

ulat

ive

Pro

porti

on

Robust

Pre-frail

Frail

0.30

0.12

0.45

*Rockwood, Andrew, Mitnitski. J Gerontol Med Sc, 2007;62:738-743.

31

Fig. 2. Cumulative distributions of frailty index scores by number of phenotypic items present.

0 0.2 0.4 0.6 0.80

0.2

0.4

0.6

0.8

1

Frailty Index value

Cum

ulat

ive

prop

ortio

n

Rockwood , et al., J Gerontol Med Sc, 2007;62:738-743.

0.12

0.22

0.34

0.45

0.54

32

Survival curves by CHS definition (Panel A) and for each CHS level (Panels B-D) by FI value cut-point*)

10 20 30 40 50 60 700

0.2

0.4

0.6

0.8

1

A.

0 10 20 30 40 50 60 700.3

0.4

0.5

0.6

0.7

0.8

0.9

1B. ‘Robust’

0 10 20 30 40 50 60 700.3

0.4

0.5

0.6

0.7

0.8

0.9

1C. ‘Pre-frail’

0 10 20 30 40 50 60 700.3

0.4

0.5

0.6

0.7

0.8

0.9

1

D. ‘Frail’

Rockwood , et al., J Gerontol Med Sc, 2007;62:738-743

FI<0.25

FI>=0.25

FI<0.25

FI>=0.25

FI<0.25

FI>=0.25

FrailPre-frail

Robust

33

Additional comparisons of the CHS definition and the FI: FI stratified by CHS (Panel E); Institutionalization of the

Robust stratified by FI (Panel F)

Time (months)

Pro

babi

lity

of a

void

ance

of in

stitu

tiona

l car

e

(n=700)

(n=52)

0 10 20 30 40 50 60 700.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Rockwood, et al., J Gerontol Med Sc, 2007;62(7):738-742.

Frailty Index

Cum

ulat

ive

dist

ribut

ion

0 0.2 0.4 0.60

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.60

0.2

0.4

0.6

0.8

1

Empirical CDFA B

How crucial are the exact components of the CHS definition of frailty?

Rockwood, et al., J Gerontol Med Sc, 2007;62(7):738-742.

None8.6%

Disability & Comorbidity

48.1%

Only Disability

18.5%

Only Comorbidity

24.8%

None5.2%

Disability & Comorbidity

59.7%

Only Disability

17.5%

Only Comorbidity

17.5%

Prevalence of Disability and Comorbidity in frail older

adults Frailty Index (Frail >0.25FI) Frailty Phenotype (Frail ≥ 3 Phenotypic Frailty

Criteria)

Theou et al. (in preparation)