Assessing the validity of the PHQ9, HADS, BDIII and QIDSSR16 in measuring severity of depression in

1

Assessing the validity of the PHQ-9, HADS,
BDI-II and QIDS-SR16 in measuring severity of
depression in a Scottish sample of primary care
patients with a diagnosis of depression Cameron
IM1, Crawford JR2, Lawton K3, Sharma, S4, DuToit
S4, Hay S4, Winning, S4, Mitchell K4,
Shivaprasad S4, Cardy A3, Reid IC1 1Applied
Health Sciences (Mental Health), 2School of
Psychology, 3Centre of Academic Primary Care,
University of Aberdeen, 4Royal Cornhill
Hospital, NHS Grampian
Background
Convergent and discriminant validity
? NICE Guidelines1 and the Quality and Outcomes
Framework (QOF)2 of the new General Medical
Services (nGMS) emphasise the importance of
measuring the severity of depression in primary
care in order to target the condition with an
appropriate intervention. ? There is an absence
of UK psychometric assessment of the QOF scales
(Patient Health Questionnaire (PHQ-9), Hospital
Anxiety and Depression Scale (HADS) and Beck
Depression Inventory (BDI-II)) in terms of their
suitability for measuring severity of depression.
? Preliminary investigations of HADS Depression
subscale (HAD-D) and PHQ-9 have demonstrated a
lack of concordance between the scales
categorisation of severity of depression, leading
to a lack of confidence in their validity3.
The scales demonstrated good convergent validity
in that they all correlated highly with the HRSD.
They also correlated highly with the HADS
Anxiety subscale (HAD-A). This is to be expected
between such closely related constructs as
anxiety and depression however (apart from
BDI-II) correlations were significantly higher
between the scales and HRSD, than between the
scales and HAD-A.
Study aim
To assess the psychometric properties of the
PHQ-9, HADS, BDI-II and Quick Inventory of
Depressive Symptomatology (QIDS-SR16), relative
to the clinician administered Hamilton Depression
Rating Scale (GRID-HAMD 17-item)4 in primary care
patients with a depression diagnosis.
Convergence of severity banding
Methods

• Adult patients with a GP diagnosis of depression
  from nine practices in Grampian completed a
  booklet consisting of the PHQ-9, HADS, BDI-II and
  QIDS-SR16 and were assessed by a psychiatrist
  with the HRSD (GRID-HAMD 17-item). Where more
  convenient to participants the assessment was
  made over the telephone.
• Participants were randomised to complete the
  scales up to one day before, or up one day after,
  the assessment.
• Additionally, participants completed the scales
  three months following.
• A sub-sample of participants had their
  assessment audio recorded to allow investigation
  of inter-rater reliability of the HRSD.
• Statistical analyses were carried out using SPSS
  Version 17 and Clinimetric Toolkit (CMT) Version
  1. A concurrent analysis was made of the four
  scales to assess internal consistency, factor
  structure, convergent and discriminant validity
  and responsiveness to change.

The chart shows the convergence in severity
banding between each scales and the HRSD. HAD-D
tended to categorise participants in a milder
category than HRSD whereas PHQ-9, QIDS-SR and
BDI-II tended to categorise participants in a
more severe category. Wilcoxon Signed Rank test
for related samples showed these differences to
be significant for each measure relative to the
HRSD (plt0.001) .
Results
Responsiveness to change
Participants
To date 175 participants have returned the 3
month follow up questionnaire. All but the PHQ-9
observed a difference over this time.
? From October 2007 and May 2009 273 patients
participated . ? Mean age 49.8 (s.d. 13.9), 187
(68.5) female. ? One consultant psychiatrist,
three specialist registrars and two staff grade
psychiatrists administered the HRSD assessments.
? 232 (85) assessments were made face-to-face
and 41 (15) over the telephone. ? The mean HRSD
score was 12.9 (s.d.7.6), representing
depression typical of primary care (i.e. mild to
moderate severity).
Discussion
Internal Consistency
All scales showed adequate internal consistency
(Cronbachs agt 0.8).
All scales had satisfactory internal consistency
and factor structure however not all showed
discriminant validity (BDI-II) or responsiveness
to change (PHQ-9). None of the scales converged
with the HRSD in terms of severity banding
indicating that current severity cut-offs for the
scales are untenable if the intention is to use
them to link to evidence based treatments with
regard to depression severity.
Acknowledgements
We thank NHS Quality Improvement Scotland for
funding this study, the patients and staff of the
participating practices, Kirsty Sykes for
secretarial support and the Scottish Primary Care
Research Network.
Factor Structure
? Principal Components Factor Analysis assessed
the homogeneity of each scale. Most items within
each scale had substantial loading on the first
factor.
References
1 National Institute for Clinical Excellence.
Depression management of depression in primary
and secondary care (NICE guideline). Clinical
guideline 23. London National Institute for
Clinical Excellence, 2004. 2. NHS Employers and
the General Practitioners' Committee. Quality and
Outcome Frameworks guidance for GMS contract
2009/10 2009. . 3. Cameron IM, Crawford JR,
Lawton K, Reid IC. Psychometric comparison of the
PHQ-9 and HADS for measuring depression severity
in primary care. Br J Gen Pract 20085832-6. 4.
GRID-HAMD-17 (http//www.iscdd.org/)