1 can self-report measures reliably predict audiometric measures in hong kong older adults? 9 oct...
TRANSCRIPT
1
Can self-report measures Can self-report measures reliably predict reliably predict audiometric measures audiometric measures in Hong Kong older adults?in Hong Kong older adults?
9 Oct 20059 Oct 2005
Kevin YuenKevin Yuen1212, Michael Tong, Michael Tong1212, Alex Lee, Alex Lee1212, , Peter TangPeter Tang11, Andrew van Hasselt, Andrew van Hasselt1212
11Institute of Human Communicative Research,Institute of Human Communicative Research,Division of Otorhinolaryngology, Dept of Surgery,Division of Otorhinolaryngology, Dept of Surgery,The Chinese University of Hong KongThe Chinese University of Hong Kong&&22Hear Talk FoundationHear Talk Foundation
2
ObjectivesObjectives
• To compare the prevalence of hearinTo compare the prevalence of hearing problem estimated from self-reportg problem estimated from self-reports and pure-tone audiometrys and pure-tone audiometry
• To investigate the performance of To investigate the performance of selself-reports: (1) a single question, (2) Hf-reports: (1) a single question, (2) HHIE-SHIE-S in identifying individuals with h in identifying individuals with hearing loss, against the standards froearing loss, against the standards from pure-tone audiometrym pure-tone audiometry
3
Hearing Handicap Inventory for the Elderly – Screening Hearing Handicap Inventory for the Elderly – Screening (HHIE-S)(HHIE-S)(Ventry & Weinstein, 1983)(Ventry & Weinstein, 1983)
• Self-administered 10-item questionnaireSelf-administered 10-item questionnaire• Aim at detecting emotional and social problems Aim at detecting emotional and social problems
associated with impaired hearingassociated with impaired hearing• Subjects respond to questions about circumstances Subjects respond to questions about circumstances
related to hearing by stating whether the situations related to hearing by stating whether the situations presents a problempresents a problem– ““no” (score 0)no” (score 0)– ““sometimes” (score 2)sometimes” (score 2)– ““yes” (score 4)yes” (score 4)
• Total HHIE-S score range from 0 to 40Total HHIE-S score range from 0 to 40
4
HHIE-S questions (English version)HHIE-S questions (English version)
• Does a hearing problem cause you to feel embarrassed when you Does a hearing problem cause you to feel embarrassed when you meet new people?meet new people?
• Does a hearing problem cause you to feel frustrated when talking to Does a hearing problem cause you to feel frustrated when talking to members of your family?members of your family?
• Do you have difficulty hearing when someone speaks in a whisper?Do you have difficulty hearing when someone speaks in a whisper?• Do you feel handicapped by a hearing problem?Do you feel handicapped by a hearing problem?• Does a hearing problem cause you difficulty when visiting friends, Does a hearing problem cause you difficulty when visiting friends,
relatives, or neighbors?relatives, or neighbors?• Does a hearing problem cause you to attend religious services less Does a hearing problem cause you to attend religious services less
often than you would like?often than you would like?• Does a hearing problem cause you to have arguments with family Does a hearing problem cause you to have arguments with family
members?members?• Does a hearing problem cause you to have difficulty when listening to Does a hearing problem cause you to have difficulty when listening to
TV or radio?TV or radio?• Do you feel that any difficulty with your hearing limits/hampers your Do you feel that any difficulty with your hearing limits/hampers your
personal or social life?personal or social life?• Does a hearing problem cause you difficulty when in a restaurant with Does a hearing problem cause you difficulty when in a restaurant with
relatives or friends?relatives or friends?
5
HHIE-S questions HHIE-S questions (Chinese translation version)(Chinese translation version)• 在遇見新相識的人時,聽力問題有否讓你感到尷尬?在遇見新相識的人時,聽力問題有否讓你感到尷尬?• 在和家人交談時,聽力問題有否讓你感到受挫折?在和家人交談時,聽力問題有否讓你感到受挫折?• 當別人喁喁細語時,你有否感到聆聽困難?當別人喁喁細語時,你有否感到聆聽困難?• 聽力問題有否令你感到殘缺?聽力問題有否令你感到殘缺?• 聽力問題有否令你在探望朋友聽力問題有否令你在探望朋友 ,, 家人或鄰居時感到困難?家人或鄰居時感到困難?• 聽力問題有否令你參加宗教聽力問題有否令你參加宗教或其他活動或其他活動較你希望能參加的較你希望能參加的
為少?為少?• 聽力問題有否引至你和家人或朋友吵架?聽力問題有否引至你和家人或朋友吵架?• 聽力問題有否令你聆聽電視或收音機時感到困難?聽力問題有否令你聆聽電視或收音機時感到困難?• 你認為任何的聽力問題有否影響你的個人或社交生活?你認為任何的聽力問題有否影響你的個人或社交生活?• 和家人或朋友在餐廳時,聽力問題有否令你感到困難?和家人或朋友在餐廳時,聽力問題有否令你感到困難?
6
Question 8Question 8
Does a hearing problem cause you to have difficultDoes a hearing problem cause you to have difficulty when listening to TV or radio?y when listening to TV or radio?聽力問題有否令你聆聽電視或收音機時感到困難?聽力問題有否令你聆聽電視或收音機時感到困難?
AnswerAnswer ScoreScore
““yes” yes” ”” 有”有” 44
““sometimes”sometimes” ““ 間中有”間中有” 22
““no”no” ““ 沒有”沒有” 00
7
Single QuestionSingle Question
Do you think you have a Do you think you have a problem with your heariproblem with your hearing?ng?
你覺得你的聽力有問題你覺得你的聽力有問題嗎?嗎?
8
SubjectsSubjects
• 1016 subjects participated in the “Ear and 1016 subjects participated in the “Ear and Hearing Assessment Project for the Elderly” Hearing Assessment Project for the Elderly” (June to Oct 2004)(June to Oct 2004)
• Data from 911 subjects were analyzed in this Data from 911 subjects were analyzed in this studystudy– 99 subjects excluded (reported signs of dementia)99 subjects excluded (reported signs of dementia)– 6 subjects excluded (incomplete data)6 subjects excluded (incomplete data)
• Gender distributionGender distribution– Male (n=369; 40.5%); female (n=542; 59.5%)Male (n=369; 40.5%); female (n=542; 59.5%)
9
Age and gender distribution Age and gender distribution
Age category
>8076-8071-7566-7060-65
Per
cent
age
(with
in g
ende
r)40
30
20
10
0
Gender
Male
Female
10
Definition of Definition of hearing losshearing loss vs vs hearing handicaphearing handicap
• Hearing loss by Hearing loss by Pure Tone AudiometryPure Tone Audiometry– pure tone average (PTA) of 500,1k, 2k & 4kHz of the pure tone average (PTA) of 500,1k, 2k & 4kHz of the bettbett
er earer ear– Hearing loss defined at 4 cut-off pointsHearing loss defined at 4 cut-off points
• PTA PTA ≥ 25dBHL ≥ 25dBHL • PTA PTA ≥ 40dBHL≥ 40dBHL• PTA PTA ≥ 55dBHL≥ 55dBHL• PTA PTA ≥ 70dBHL≥ 70dBHL
• Hearing handicap by Hearing handicap by HHIE-SHHIE-S and and Single QuestionSingle Question– HHIE-S > 8, >10, >12 & >16HHIE-S > 8, >10, >12 & >16– Single Question -> Answered “yes” Single Question -> Answered “yes”
11
Comparison of prevalence ratesComparison of prevalence rates
• Definition of hearing lossDefinition of hearing loss– PTA >25dBHL (0.5 – 4kHz)PTA >25dBHL (0.5 – 4kHz) of the better ear of the better ear
Age rangeAge range NN Prevalence ratePrevalence rate
Framingham Cohort Framingham Cohort
(1978-79)(1978-79)57-8957-89 22932293 47%47%
Framingham Cohort Framingham Cohort
(19(1983-8583-85))63-9563-95 16621662 42%42%
Wisconsin Cohort Wisconsin Cohort
(1993-95)(1993-95)48-9248-92 37533753 32%32%
Blue Mountains Cohort Blue Mountains Cohort (1997-99)(1997-99)
55-10055-100 20152015 30%30%
Hong Kong Cohort Hong Kong Cohort (2004)(2004)
60-9260-92 911911 82%82% (for all (for all, N=911, N=911))
71%71% ( (the group answerethe group answeredd “no” from “no” from the the Single Single Question, n=336)Question, n=336)
12
Odds ratios for the presence of hearing loss at Odds ratios for the presence of hearing loss at different PTA cut-off levelsdifferent PTA cut-off levels
Odds ratioOdds ratio
PTA ≥ PTA ≥
N=911N=911 FactorFactor 25dBHL25dBHL 40dBHL40dBHL 55dBHL55dBHL 70dBHL70dBHL
Age Age 60-6560-65 1.01.0 1.01.0 1.01.0 1.01.0
groupgroup 66-7066-70 1.41.4 1.51.5 2.52.5 2.32.3
(years(years))
71-7571-75 2.7***2.7*** 2.5***2.5*** 3.8**3.8** 5.95.9
75-8075-80 8.3***8.3*** 6.5***6.5*** 8.5***8.5*** 7.07.0
> 80> 80 17.6***17.6*** 9.8***9.8*** 13.5***13.5*** 18.3***18.3***
SexSex FF 1.01.0 1.01.0 1.01.0 1.01.0
MM 1.5*1.5* 1.5**1.5** 1.7*1.7* 2.4*2.4*
* p < 0.05, ** p < 0.01, *** p < 0.0005.
13
Prevalence of hearing loss (measured) vs Prevalence of hearing loss (measured) vs hearing handicap (estimation)hearing handicap (estimation)
Criterion for presence of hearing problem
Pre
vale
nce
(%
)
100
80
60
40
20
0
Estimated prevalence-> Single question > HHIE >8
Prevalence from HHIE >8 (estimation) similar to PTA ≥ 40 (measured)
Hearing handicap
Hearing loss82
3834
13
4
63
14PTA cut-off level (dBHL)
>=70>=55>=40>=25
Mea
sure
d (a
udio
met
ry)
- es
timat
ed p
reva
lenc
e (%
) 80
60
40
20
0
-20
-40
-60
-80
Estimated prevalence
HHIE >8
Single Question
Comparison of Comparison of measured (audiometry) vs estimated prevalencemeasured (audiometry) vs estimated prevalence
McNemar Test
p < .001 for all comparisons
* except for PTA >=40 vs HHIE >8 p = .04*
15
Screening performance characteristicsScreening performance characteristics
YesYes NoNo
YesYesTTrue rue
PPososFFalsealse N Nee
gg
NoNoFFalsealse
PPososTTrue rue NNee
gg
Hearing problem from HHIE-S / Single Question ?
Hea
ring
prob
lem
fro
m a
ud
iom
etry
?
Sensitivity = Sensitivity = TP TP / ( / ( TPTP + + FNFN ))Specificity = Specificity = TNTN / (/ ( TNTN + + FPFP ))
Positive predictive value (PPV)Positive predictive value (PPV)= = TP TP / (/ (TPTP + + FPFP ))Negative predictive value (NPV)Negative predictive value (NPV)= = TNTN / ( / ( TNTN + + FNFN ))
Positive likelihood ratio (PLR)Positive likelihood ratio (PLR)= sensitivity/ (1-specificity)= sensitivity/ (1-specificity)
Negative likelihood ratio (NLR)Negative likelihood ratio (NLR)= (1-sensitivity)/ specificity= (1-sensitivity)/ specificity
16
HHIE-SHHIE-S vs vs Single QuestionSingle QuestionDefinition of Definition of
hearing handicaphearing handicap
Definition of hearing lDefinition of hearing lossoss
PTA 0.5-4kHz, better PTA 0.5-4kHz, better earear
SensitivitySensitivity(%)(%)
SpecificitySpecificity(%) (%)
AccuracyAccuracy(%)(%)
PPVPPV(%)(%)
NPVNPV(%)(%)
PLRPLR(%)(%)
NLRNLR(%)(%)
HHIE-S > 8HHIE-S > 8 ≥≥40dBHL40dBHL 5252 7777 6767 5858 7272 2.22.2 0.60.6
HHIE-S > 16HHIE-S > 16 ≥≥40dBHL40dBHL 3434 9191 6969 6969 6969 3.73.7 0.70.7
Single Single QuestionQuestion ≥≥40dBHL40dBHL 8383 4949 6262 5050 8282 1.61.6 0.40.4HHIE-S > 8HHIE-S > 8 ≥≥55dBHL55dBHL 7272 7272 7272 2828 9494 2.52.5 0.40.4
HHIE-S > 16HHIE-S > 16 ≥≥55dBHL55dBHL 5555 8787 8383 3939 9393 4.24.2 0.50.5Single Single QuestionQuestion ≥≥55dBHL55dBHL 9292 4141 4848 1919 9797 1.61.6 0.20.2
HHIE-SHHIE-S better in better in ruling ruling ININ the presence of hearing loss the presence of hearing loss (FP) (FP)Single QuestionSingle Question better in better in ruling ruling OUTOUT the presence of hearing loss the presence of hearing loss (FN)(FN)
17
HHIE-SHHIE-S vs vs Single QuestionSingle Question
HHIE-SHHIE-S better in better in ruling in the presence of ruling in the presence of hearing losshearing loss, with better, with better
– PPV and PLRPPV and PLR
Single QuestionSingle Question better in better in ruling out the ruling out the presence of hearing losspresence of hearing loss, with better, with better
– NPV and NLRNPV and NLR
18
Receiving-operating characteristic curves Receiving-operating characteristic curves of different PTA cut-off levels with of different PTA cut-off levels with
HHIE-S scoreHHIE-S score
1-Specificity (false positive rate)
1.0.8.6.4.20.0
Se
nsi
tivity
(tr
ue
po
sitiv
e r
ate
)
1.0
.8
.6
.4
.2
0.0
PTA cut-off level
>=70 dBHL
>=55 dBHL
>=40 dBHL
>=25 dBHL
Area under curve (AUC) =
•Discriminating power of an HHIE-S score at each PTA cut-off levels.
•Probability that a random person with measured hearing loss (audiometry) has a higher HHIE-S score than a random person without the hearing loss
Area under curve
0.83
0.78
0.69
0.67
HHIE-S
cu
toff
incr
ease
HHIE >8HHIE >8
HHIE >16HHIE >16Perfect accuracyAUC = 1
ChanceChanceAUC = 0.5AUC = 0.5
19
HHIE-S > 8HHIE-S > 8 vs vs HHIE-S > 16HHIE-S > 16
Definition of Definition of hearing handicaphearing handicap
Definition of Definition of hearing losshearing loss
PTA 0.5-4kHz, PTA 0.5-4kHz, better earbetter ear
SensitivitySensitivity(%)(%)
SpecificitySpecificity(%) (%)
AccuracyAccuracy(%)(%)
PPVPPV(%)(%)
NPVNPV(%)(%)
PLRPLR(%)(%)
NLRNLR(%)(%)
HHIE > 8HHIE > 8 ≥≥40dBHL40dBHL 5252 7777 6767 5858 7272 2.22.2 0.60.6
HHIE> 16HHIE> 16 ≥≥40dBHL40dBHL 3434 9191 6969 6969 6969 3.73.7 0.70.7
HHIE > 8HHIE > 8 ≥≥55dBHL55dBHL 7272 7272 7272 2828 9494 2.52.5 0.40.4
HHIE> 16HHIE> 16 ≥≥55dBHL55dBHL 5555 8787 8383 3939 9393 4.24.2 0.50.5
20
Revised HHIE-S cut-off scoreRevised HHIE-S cut-off score
• HHIE-S > 16HHIE-S > 16 vs vs HHIE-S > 8HHIE-S > 8– Better specificity, accuracy, (PPV) and PLRBetter specificity, accuracy, (PPV) and PLR– Worse (sensitivity)Worse (sensitivity)– Similar NPV, NLRSimilar NPV, NLR
according to the comparisons between 95% confidencaccording to the comparisons between 95% confidence intervalse intervals
For PTA cut-off levels For PTA cut-off levels ≥ 40 & ≥ 55 dBHL≥ 40 & ≥ 55 dBHL
21
Combined “HHIE-S - Single Question” screening tool
HHIE-SHHIE-S FF PP FF PP
SQSQ PP PP FF FF
CombinedCombined PP FF
HHIE-S HHIE-S PP PP FF FF
SQSQ PP FF PP FF
CombinedCombined PP FF
P – PassP – PassF – FailF – Fail
SUPER POWERSUPER POWERSCREENER ?!SCREENER ?!
raise sensitivityraise sensitivitylower FN lower FN raise FPraise FP
22
HHIE-S – Single QuestionHHIE-S – Single Question
SUPER POWER SCREENING TOOL ?!SUPER POWER SCREENING TOOL ?!
24
Conclusion 1Conclusion 1
The prevalence of The prevalence of hearing handicap hearing handicap (HHIE-S >8)(HHIE-S >8)
is comparable with is comparable with
the prevalence of the prevalence of hearing loss hearing loss (PTA (PTA ≥ 40dBHL)≥ 40dBHL)
25
Conclusion 2Conclusion 2
The prevalence of hearing handicap The prevalence of hearing handicap increases with age.increases with age.
The prevalence of hearing handicap is The prevalence of hearing handicap is higher in male than in female.higher in male than in female.
26
Conclusion 3Conclusion 3
• HHIE-S better in ruling HHIE-S better in ruling ININ the presence of he the presence of hearing lossaring loss
• Single Question better in ruling Single Question better in ruling OUTOUT the pres the presence of hearing lossence of hearing loss
• HHIE-S >16HHIE-S >16 has a better screening characteri has a better screening characteristics profile than stics profile than HHIE-S >8HHIE-S >8 for screening PT for screening PTA level A level ≥ 40 and ≥ 55 dBHL≥ 40 and ≥ 55 dBHL
27
Conclusion 4Conclusion 4
For older adults reported For older adults reported
• withwith hearing handicap (HHIE > 16), the probability with PT hearing handicap (HHIE > 16), the probability with PTA A ≥ ≥ 40dBHL is 69% 40dBHL is 69% (Positive predicative value).(Positive predicative value).
• withoutwithout hearing handicap (HHIE hearing handicap (HHIE ≤≤ 16), the probability with 16), the probability with PTA PTA << 40dBHL is 69% 40dBHL is 69% (Negative predicative value).(Negative predicative value).
Overall accuracy = 69Overall accuracy = 69 %%
29
Kevin YuenKevin Yuen
Division of Otorhinolaryngology, Division of Otorhinolaryngology,
Department of SurgeryDepartment of Surgery
The Chinese University of Hong KongThe Chinese University of Hong Kong
30
Screening performance characteristicsScreening performance characteristics
YesYes NoNo
YesYesTrue True
PosPosFalse NeFalse Ne
gg
NoNoFalseFalse
PosPosTrue NeTrue Ne
gg
Hearing problem from HHIE-S / Single Question ?
Hea
ring
prob
lem
fro
m a
ud
iom
etry
?
Sensitivity = Sensitivity = TP TP / ( / ( TPTP + + FNFN ))Specificity = Specificity = TNTN / (/ ( TNTN + + FPFP ))
Positive predictive value (PPV)Positive predictive value (PPV)= = TP TP / (/ (TPTP + + FPFP ))Negative predictive value (NPV)Negative predictive value (NPV)= = TNTN / ( / ( TNTN + + FNFN ))
Positive likelihood ratio (PLR)Positive likelihood ratio (PLR)= sensitivity/ (1-specificity)= sensitivity/ (1-specificity)
Negative likelihood ratio (NLR)Negative likelihood ratio (NLR)= (1-sensitivity)/ specificity= (1-sensitivity)/ specificity
31
Receiving-operating characteristic curves Receiving-operating characteristic curves of different PTA cut-off levels with of different PTA cut-off levels with
HHIE-S scoreHHIE-S score
1-Specificity (false positive rate)
1.0.8.6.4.20.0
Se
nsi
tivity
(tr
ue
po
sitiv
e r
ate
)
1.0
.8
.6
.4
.2
0.0
PTA cut-off level
>=70 dBHL
>=55 dBHL
>=40 dBHL
>=25 dBHL
Area under curve (AUC) =
•Discriminating power of an HHIE-S score at each PTA cut-off levels.
•Probability that a random person with measured hearing loss (audiometry) has a higher HHIE-S score than a random person without the hearing loss
Area under curve
0.83
0.78
0.69
0.67
HHIE-S
cu
toff
incr
ease
HHIE >8HHIE >8
HHIE >16HHIE >16
32
Prevalence of HHIE-S >8 Prevalence of HHIE-S >8 with age and genderwith age and gender
Prevalence of hearing handicap
increases with agePearson Chi square = 14.2, p <.01Linear-by-linear association = 6.3, p <.001
higher in Male than in FemalePearson Chi square = 5.2, p <.05
Male
Female
Gender
60-65 66-70 71-75 76-80 >80
Age category
10
20
30
40
50
Per
cen
tag
e w
ith
HH
IE >
8
33
Prevalence of HHIE >8 Prevalence of HHIE >8 with degree of hearing losswith degree of hearing loss
Prevalence of hearing handicap Prevalence of hearing handicap
increases with hearing lossincreases with hearing loss
Pearson Chi square = 116.8, Pearson Chi square = 116.8, pp <.001 <.001
Linear-by-linear association = 111.1, Linear-by-linear association = 111.1, pp <.001 <.001
normalmild
moderatemoderately severe
severe/ profound
Hearing loss severity from PTA
0
25
50
75
100
Per
cen
tag
e w
ith
HH
IE >
8
34
HHIE >8
Single question
Estimated prevalence
-50
-25
0
25
50
Me
asur
ed
(aud
iom
etry
)
- es
tim
ate
d p
reva
lenc
e (%
)male female
total
>=25 >=40 >=55 >=70
PTA cut-off level (dBHL)
-50
-25
0
25
50
Me
asur
ed
(aud
iom
etry
)
- es
tim
ate
d p
reva
lenc
e (%
)
Measured (audiometry) vs estimated prevalence –by gender
35
Screening performance characteristics Screening performance characteristics of of HHIE-SHHIE-S and and Single QuestionSingle Question
Measure Subgroup Group Meaning
Sensitivity •reported with problem•fail audiometry
All fail
audiometry
Prob with true hearing loss identified by the report
Specificity •reported without problem•pass audiometry
All pass audiometry Prob without true hearing loss correctly identified by the question
Positive predictive value
•reported with problem•fail audiometry
All reported with problem
Prob of report to identify disorder
Negative predictive value
•reported without problem•pass audiometry
All reported without problem
Prob of report to identify normality
36
Sensitivity and specificitySensitivity and specificity
HHIE-S>8HHIE-S>10
HHIE-S>12HHIE-S>16
Single Q0
25
50
75
100
Se
nsi
tivity
(%
)
HHIE-S>8HHIE-S>10
HHIE-S>12HHIE-S>16
Single Q0
25
50
75
100
Sp
eci
ficity
(%
)
• Sensitivity – increase with PTA cut-off– Single question better than HHIE-S
• Specificity– decrease with PTA cut-off– HHIE-S better than Single question
37
Positive and negative Positive and negative predictive values (PPV & NPV) predictive values (PPV & NPV)
HHIE-S>8HHIE-S>10
HHIE-S>12HHIE-S>16
Single Q0
25
50
75
100
+ve
Pre
dic
tive
Va
lue
(%
)
HHIE-S>8HHIE-S>10
HHIE-S>12HHIE-S>16
Single Q0
25
50
75
100
-ve
Pre
dic
tive
Va
lue
(%
)
• PPV – decrease with PTA cut-off
• NPV– increase with PTA cut-off
38
Positive and negative Positive and negative likelihood ratios (PLR & NLR) likelihood ratios (PLR & NLR)
HHIE-S>8HHIE-S>10
HHIE-S>12HHIE-S>16
Single Q
1
2
3
4
+ve
Lik
elih
oo
d R
atio
HHIE-S>8HHIE-S>10
HHIE-S>12HHIE-S>16
Single Q0.0
0.2
0.4
0.6
0.8
-ve
Lik
elih
oo
d R
atio
• PLR – Increase slightly with PTA cut-off– Increase with HHIE-S cut-off
• NLR– decrease with PTA cut-off– Increase slightly with HHIE-S cut-off