• 
• 

Three Part Question

In [adult stroke patients] is [the EQ-5D] a [valid measure of quality of life]

Clinical Scenario

We routinely use EQ-5D as a patient experience based outcome for our stroke and neuro service but is it a valid measure for our service?
Neurology was considered to be too wide an area and therefore we narrowed the question to stroke. We did not restrict the question to community dwelling patients as it was felt this would restrict the search.

Search Strategy

Medline, CINAL and AMED limited to English and humans. The relevant papers reference lists were also checked.
Search terms used; Stroke*, cerebrovascular accident, CVA, Cerebral infarction, Subarachnoid hemorrhag*, subarachnoid haemorrhage*, EQ-5D, quality of life, QoL, Euroqol-5D

Search Outcome

Altogether 173 papers were found of which 8 were relevant to the study question.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Golicki, D et al 2015 Poland	The study compared the validity of the EQ5D-5L and EQ5D-3L in 408 acute stroke patients during hospitalization	Single center cross sectional study (level 2)	Modified Rankin Score (mRS) Bartel Index (BI) VAS (Visual Analogue Scale)	Moderate to strong correlations were found between 5L and mRS, BI and EQVAS The study supports the EQ-5D-5L as a valid generic health outcome measure in patients with acute stroke. Shows some psycho metric advantages of the EQ-5D-5L in comparison with EQ-5D-3L	Quality of life (QoL) questionnaires were administered in a fixed order rather than randomly. No questionnaires were presented between 5L & 3L. The risk is that memory effects may affect the comparison of the two versions.
Whynes, D.K. et al 2013 UK	1462 patients were recruited approximately 90 days post stroke to assess the presence of differential item functioning within the EQ5D	Tested presence of differential item functioning in EQ5D using the data from a large multi-national trial in acute stroke	The study mapped the following clinical outcome measures onto the EQ5D: Modified Rankin Score Barthel Index Zung Depression scale	It shows it is a sensitive scale in stroke patients but that only UK EQ-5D scores should be compared due to regional variances. The differential item functioning with respect to patient versus proxy responses to the EQ-5D was detectable but the impact on mean index scores insignificant.	Inter regional variances were reported and greater proxy variance but this is not isolated to the EQ5D and has also been noted in other QoL measures i.e. SF36 and ascribed to sociocultural differences.
Dorman, P.J. et al 19 1997 Scotland	The validity of the EuroQol was assessed in 152 patients with acute stroke. 92 completed questionnaires independently (with mild to moderate stroke) and 60 by proxy (patients with significant motor deficits)	Single group prospective assessment via questionnaire	Euro Qol Frenchay activities index HADS (hospital anxiety & depression scale) VAS pain scale Modified single questions Bartel Index GPCS disability index	The EuroQol appears to have acceptable concurrent and discriminant validity for the measurement of health related QoL after stroke. -Concurrent validity good- patients who reported problems on EuroQol also reported dysfunction with standard test for that domain -HRQol worst in domains for patients with partial anterior circulation strokes and best in patients with posterior circulation strokes showing discriminant validity.	Only used on those >1yr post stroke. Many of our client group are earlier than this. 6 ruled out due to reduced communication
Sang-Kyu Kim et al 2015 South Korea	1462 patients were recruited approximately 90 days post stroke to assess the presence of differential item functioning within the EQ5D The study looked at the minimally important differences in EQ5D and SF-6D indices to explore the responsiveness of EQ-5D and SF 6D in stroke. 487 community subjects who had suffered their first ever or recurrent stroke and were aged 50 years or older.	Anchor based approach using observational longitudinal survey data over 10 month period	Modified Rankin Score & Bartel Index (BI) were used as anchors SF36 EQ-5D	Minimally important difference (MID) estimates for the EQ-5D ranged from 0.08-0.12 and for SF36 ranged from 0.04-0.014 The MID changes depended on which anchor was used. The study primarily focuses on the MID but also considers the responsiveness of the EQ-5D and SF36. It concluded that EQ-5D can show changes in health state and therefore may be useful in interpretation of intervention effects on Health related QoL	The data used was from 2009 but the article published in 2015 Different interviewers did the assessments at different time points but all were trained to reduce variability. This should not impact on the EQ-5D as this is self or by proxy administered. The article looks at EQ-5D-3L but we use EQ-5D 5L
Hunger, M. Sabariego, C. Stollenwerk, B. Cieza, A. Leidl, R. 2012 Germany	Analysis of the validity, reliability and responsiveness of the EQ-5D in German stroke patients. 210 stroke survivors (5.7 weeks post stroke) before and 183 after participating in an education programme while undergoing inpatient rehabilitation	Single blind RCT anchor based approach Patients divided into sub groups BI 35-65 BI>65	EQ-5D compared with other measures: Stroke impact scale Mobility ADL Physical domain, social HADs	The EQ5D showed reasonable validity, reliability and more limited responsiveness in stroke patients. Ceiling effect in higher functioning group (BI>65) Good correlation SIS on mobility (rs=-0.74), self care (rs= -0.65) and usual activities (rs= -0.39) Unable to correlate pain domain as no comparable domain in SIS. Test- retest not responsive to changes in health status in serial assessment of individual patients, people value different domains depending on recovery.	Part of another published RCT therefore methodology not included in this paper but referred to original study.
King et al 2009 USA	Validity and reliability of the EQ-5D in 178 Neurosurgical patients with cerebral aneurysms. 55% had survived a SAH	Rank order methods. Multivariate linear regression.	Glasgow Outcome Scale Rankin Scale Barthel Index Physical performance test scores SF12 HADS EQ5D-3L	Construct validity was confirmed by statistically significant association between EQ5D and outcomes Reliablity was demonstrated by Cronbach’s ? of 0.70	No specific pain measure was used to map onto the EQ5D The study sample was drawn from 1 academic institution. Not all eligible patients enrolled and some did not complete all of the instruments.
Pickard, A.S. et al 2004 USA	Comparison of the responsiveness of 5 health related quality of life measures in 124 stroke patients within the first 6 months post stroke	Observational longitudinal cohort study	EQ-5D VAS & index score SF-6D Health Utilities Index (HUI) Mark 2 (HUI2) & Mark 3 (HUI3) BI mRS	EQ-Index is more responsive and required smaller sample size to detect MCID than EQ-VAS EQ-5D (index) has a ceiling effect in higher functioning patients	Not randomized Exclusion criteria included aphasia and cognitive impairment and therefore not reflective of our stroke population.
Pinto et al 2011 Brasil 2011 Brasil	Validation of the EuroQol on a convenience sample of 67 stroke patients.	Convenience sample from stroke clinic	EQ5D Modified Barthel Index NIHSS	Demonstrates EQ-5D is reproducible and valid EQ5D showed good correlation with both stroke severity (NIHSS, r= -0.404, P<0.001) and degree of impairment on ADL (mBI, r= 0.512 P,0.001) Good interobserver agreement (k>0.60) in all dimensions evaluated P<0.01 Demonstrates a significant correlation exists between QoL and level of function	Uses the 3L rather than the 5L Convenience sample therefore mainly patients who could access outpatients and therefore presenting usually with less severe strokes. Low quality evidence

Comment(s)

EQ-5D is not a stroke specific measure but it is the most widely used generic health status questionnaire used in stroke. It’s key limitation is that it does not include questions around speech or cognition which are often issues for stroke patients. Two studies excluded patients with aphasia. The literature published has established its psychometric properties in stroke patients, reasonable construct, concurrent and discriminant validity, accuracy for predicting outcomes and responsiveness in longitudinal studies. Most of the articles compared it to health state linking this as a reflection of the patient’s quality of life. The studies reviewed often analysed retrospective data collected where the questionnaire had been used as part of a large study. EQ-5D is designed for self or proxy completion. We did not specify in our question whether the patients were to be able to complete the questionnaire themselves. Many of the studies use proxies and reported it to be reliable when compared to the patients’ health state. We have not used proxies to complete EQ-5D questionnaires but we now could consider using them in our practise. Our service uses the EQ-5D-5L version as we felt it would be more sensitive to change than the 3L. This was confirmed by the Golicki, D et al 2015 study.

Clinical Bottom Line

Based on the current best evidence there is strong evidence to suggest that the EQ 5D is a valid measure of quality of life.

References

1. Golicki, D. Niewada, M. Buczek, J. Karlinska, A. Kobayashi, A. Janssen, M.F. Pickard, S.A. Validity of EQ-5D-5 L in stroke. Quality of life research (2015) 24:845-850
2. Whynes, D.K. Sprigg, N.Selby, J. Berge, E. Bath, P.M. Testing for differential item functioning within the EQ-5D. Medical decision making. 2013; Vol 33 no 2 p252-260
3. Dorman, P.J. Waddell, F. Slattery, J. Dennis, M. Sandercock, P. Is the EuroQol a valid measure of health-related quality of life after stroke? Stroke 28:1876-1882
4. Kim, S-K, Kim, S-H. Jo, M-W. Lee, S-I Estimation of minimally important differences in the Eq-5D and SF-6D indices and their utility in stroke Health and Quality of life outcomes 2015;13:32
5. Hunger, M. Sabariego, C. Stollenwerk, B. Cieza, A. Leidl, R. Validity, reliability and responsiveness of the EQ-5D in German stroke patients undergoing rehabilitation. . Quality of life research 21:1205-1216
6. King, J.T. Tsevat, J. Robert, M.S. Measuring preference-based quality of life using the EuroQol EQ-5D in patients with cerebral aneurysms. Neurosurgery vol65, No3, p565-573
7. Pickard, S.A. Johnson, J.A. Feeny, D.H. Responsiveness of generic health-related quality of life measures in stroke Quality of life research 4:207-219
8. Pinto, E.B. Maso, L. Vilela, R.N.R. Santos, L.C. Oliveira-Filho, J. Validation of the EuroQol quality of life questionnaire on stroke victims Arquivos de neuro-psiquiatria vol 69, No 2B p320-323