The HeartQol questionnaire. Reliability, validity and responsiveness? Stefan Höfer, PhD, MSc Associate Professor Medical University Innsbruck
The HeartQoL Story Data collection: 2000 2010 Presentations: 2006 (EuroPrevent Athens) : : : : 22 Minutes 2012 (EuroPrevent Dublin)
Global HRQL Physical HRQL Emotional HRQL Range: 0 (low) 3 (high)
Reliable, valid, responsive Reliable: yes/don t know Valid: yes Responsive: yes
HeartQoL Project: International 21 countries 15 languages n= 6,384 Australia; Canada; USA SE: Portugal, Spain UK, Ireland NE: Denmark, Norway, Sweden WE: Austria, Belgium, France, Netherlands, Germany, Switzerland SE: Italy EE: Hungary, Poland, Russia, Ukraine
Sample mi primary: 37.3% angina primary: 33.1% heart failure primary: 29.6% Age: 62.5±11.3 mi primary: 59.7±11.7*** angina primary: 63.1±10.2*** heart failure primary: 65.1±11.5 ***
HeartQoL patient profiles: International cohort summary HeartQoL Age / % male 62.3 / 75% BMI 27.4 Smoking 15.1% Diabetes 20.6% Treated for 90.1% Hypertension 55.4% Treated for 95.9% Hypercholesterolemia 59.6% Treated for 92.1% EuroAspire III 61.9 / 73% 28.0 18% 28% 35% 61% 89% 46% 88%
Criteria of HRQoL Instruments Conceptual and measurement model Reliability Validity Responsiveness Interpretability Alternative forms Respondent and administrative burden Objectivity Cultural and language adaptations Scientific Advisory Board, Quality of life Resaerch, 2002
Conceptual & measurement model MacNew MI Q 27 items 3 sub-scales: Physical Social Emotional Sub-scale scores Total Score Seattle Angina Q 19 items 5 sub-scales Physical limitation Angina stability Angina frequency Treatment satisfaction Disease perception Sub-scale scores Minnesota Living with Heart Failure Q 21 items 2 sub-scales Physical Emotional Sub-scale scores Total score
Conceptual & measurement model MacNew MI Q 27 items 3 sub-scales: Physical Social Emotional Sub-scale scores Total Score Seattle Angina Q 19 items 5 sub-scales Physical limitation Angina stability Angina frequency Treatment satisfaction Disease perception Sub-scale scores Minnesota Living with Heart Failure Q 21 items 2 sub-scales Physical Emotional Sub-scale scores Total score 49 items
What we did Item reduction from 49 item to as few as possible: Clinical impact method Mokken scaling Item response theory Pre-testing the prototype instrument Decision about scoring method Determination of reliability Validation Finally: 14 items
Reliability Is an instrument free from random error? Internal consistency the precision of a scale: i.e.: Cronbach s alpha Reproducibility stability over time: i.e.: Test-retest reliability
Reliability Cronbach s a Total group Angina MI Heart Failure Global score 0.91 0.90 0.90 0.91 Physical score 0.90 0.89 0.90 0.90 Emotional score 0.81 0.80 0.81 0.82 Cronbach s a: Group comparison: >0.70 Individual comparison: >0.90
Reliability compared (i.e.) Cronbach s a Total group Angina MI Heart Failure Global score 0.91 0.90 0.90 0.91 Physical score 0.90 0.89 0.90 0.90 Emotional score 0.81 0.80 0.81 0.82 Angina HeartQoL - physical.89 SF-36 - PCS SAQ physical limitations MacNew physical HRQL.93.94.95
Validity Does an instrument measure what it purports to measure? Content-related: Evidence that the domain of an instrument is appropriate relative to its intended use Construct-related: Evidence that supports a proposed interpretation of scores based on theoretical implications associated with the constructs being measured Criterion-related: Evidence that shows the extent to which scores of the instrument are related to a criterion measure
Convergent Validity SF-36 PCS SF-36 MCS p-value H-physical ++ - <.001 H-emotional - ++ <.001 p-value <.001 <.001
Convergent Validity SF-36 PCS SF-36 MCS p-value H-physical 0.68** 0.36** <.001 H-emotional 0.28** 0.60** <.001 p-value <.001 <.001 Same pattern for each diagnosis Same pattern in each country
Discriminative Validity 3 2,5 2,4 2,3 2,5 2 1,8 1,9 1,9 1,5 1 non-depressed depressed 0,5 0 Angina MI Heart Failure Same pattern for each diagnosis Same pattern in each country
Discriminative Validity 3 2,5 2 2,3 2,1 2,2 1,9 1,5 1 CCS/NYHA II CCS/NYHA III/IV 0,5 0 Angina Same pattern for each diagnosis Same pattern in each country Heart Failure
Responsiveness Is an instrument able to detect change over time? Evidence on the changes in scores of the instrument Longitudinal data that compare a group that is expected to change with a group that is expected to remain stable
PCI: N=398 CR: N=383 3 Responsiveness 2,5 ** 2,4 2 2 ** 2 1,5 1 1,7 PC Rehab 0,5 0 Baseline Follow-up
Responsiveness 0,7 Effektsize statistics SRM 0,6 0,5 0,4 0,3 0,2 Medium SF-36 HeartQoL Small 0,1 0 PCI Rehab PCI Rehab Physical Scales Emotional Scales
Criteria of HRQoL Instruments Conceptual and measurement model Reliability Validity Responsiveness Interpretability Alternative forms Respondent and administrative burden Objectivity Cultural and language adaptations Scientific Advisory Board, Quality of life Resaerch, 2002
Interpretability Are the numbers produced easily understood? Meaningful benchmarks to facilitate interpretation of the scores ( norms )... Pre post change... Clinical important difference What does a change of 0.3 on the HeartQoL actually mean?
Summary: HeartQoL Reliable: yes/don t know Valid: yes Responsive: yes
HeartQoL - Future To check for: test-stability interpretability Comparing with current gold standard(s) i.e. MacNew Health-related quality of life instrument Application