You are here

Article Text

Article menu
Neuropsychiatry
Research paper
Screening for poststroke major depression: a meta-analysis of diagnostic validity studies
  1. Nick Meader1,
  2. Thirimon Moe-Byrne1,
  3. Alexis Llewellyn1,
  4. Alex J Mitchell2
  1. 1Centre for Reviews and Dissemination, University of York, York, UK
  2. 2Department of Cancer Studies and Molecular Medicine, Leicester Royal Infirmary, Leicester, UK
  1. Correspondence to Dr Nick Meader, Centre of Reviews and Dissemination, University of York, York YO10 5HU, UK;  nick.meader{at}york.ac.uk;

Abstract

Background Major depression is common in stroke patients and associated with increased rates of disability and mortality. Identifying depression may improve mental and physical health. The aim of this review was to determine the most accurate tool for detecting poststroke depression.

Methods Seven databases were searched up to November 2012. Two authors selected studies using International Classification of Disease  or Diagnostic and Statistical Manual diagnosis of depression as the reference standard. Two authors extracted data and assessed methodological quality. Included studies were synthesised using meta-analyses.

Results A total of 24 included studies provided data on 2907 participants. The Center of Epidemiological Studies-Depression Scale (CESD) (sensitivity: 0.75; 95% CI 0.60 to 0.85; specificity: 0.88; 95% CI 0.71 to 0.95), the Hamilton Depression Rating Scale (HDRS) (sensitivity: 0.84; 95% CI 0.75 to 0.90; specificity:0.83; 95% CI 0.72 to 0.90) and the Patient Health Questionnaire (PHQ)-9 (sensitivity: 0.86; 95% CI 0.70 to 0.94; specificity: 0.79; 95% CI 0.60 to 0.90) appeared to be the optimal measures for screening measures. However, the clinical utility of all tools was modest for case-finding.

Interpretation There are a number of possible instruments that may help in screening for poststroke depression but none are satisfactory for case-finding. Preliminary data suggests the CESD, HDRS or the PHQ-9 as the most promising options. Although it should be noted such scales should not be used in isolation but followed up with a more detailed clinical assessment. While there is promising data for the PHQ-2 in other populations, it performed less well than other measures.

  • Stroke
  • Psychiatry
  • Depression
  • Scales

https://doi.org/10.1136/jnnp-2012-304194

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Stroke is common with an annual prevalence of 700 000 in the USA.1 Improvements in medical treatment have increased the number of people surviving stroke and living with the consequences.2 A recent survey of 28 090 participants over 50 years old in Germany found a community prevalence of 7.2%.3 Mental health complications are common in acute and post-acute stages of stroke. Depression is more common in people with chronic physical illness.4–6 A systematic review of 51 studies found the prevalence of any depression (ie, major and/or minor depression) was 33% in people who have had a stroke.7 Depression is often persistent after stroke, with high risk of relapse even after remission over a long period of time.8

    People with depression after stroke experience greater impairment including greater reductions in activities of daily living9 ,10 and increased mortality11 ,12 compared with non-depressed stroke patients. Untreated poststroke depression may impair physical rehabilitation and recovery.13–15 Despite the increased prevalence and morbidity associated with depression after stroke there is evidence of under-recognition in this population. For example, Williams et al16 in a study of stroke patients found that 60% of people scoring over the cut-off for depression using the Geriatric Depression Scale (GDS) were not diagnosed or treated for depression.

    Stroke patients often have a high burden of physical symptoms and symptoms such as fatigue, and psychomotor retardation may be caused by stroke or depression. As such the phenomenology of poststroke depression may be different from non-comorbid depression. However, most studies have found only minor differences in presentation.17–19 A wide variety of depression case identification measures have been investigated in stroke patients. Most scales were not originally developed for stroke populations or indeed physically ill patients therefore it is important to examine scales that include somatic items (eg, Beck Depression Inventory (BDI), Hamilton Depression Rating Scale (HDRS)) and others that seek to avoid such items (eg, Hospital Anxiety and Depression Scale (HADS), Montgomery Asberg Depression Rating Scale (MADRS)). Recent meta-analyses have not shown any particular advantage for depression scales that omit somatic items in cancer patients.20 Currently there are no systematic reviews assessing the psychometric properties of such tools in this population. Therefore a systematic review and meta-analysis was conducted to address the limitations in the literature and to seek to draw conclusions concerning what are the most accurate and efficient methods of identifying depression in people after a stroke.

    Methods

    Data sources and searches

    A search for studies assessing the validity of case identification instruments for depression was made using seven electronic bibliographic databases (Cochrane Central Trial Register, Cumulative Index to Nursing and Allied Health Literature (CINAHL), Embase, Health Management Information Consortium (HMIC), Medline, PsycINFO, Web of Knowledge). Each database was searched from inception to November 2012. The search was kept as broad as possible with subject headings and text word searches of depression terms such as ‘major depressive disorder’, ‘depression’; and observational studies terms such as ‘cross-sectional study’, ‘observational’, ‘prospective’, etc. Therefore no limitations were set for stroke populations in order to maximise sensitivity for picking up relevant references.

    References of included articles, tables of contents of relevant journals (including Stroke, Cerebrovascular Diseases, Journal of Affective Disorders, Acta Psychiatric Scandinavica) previous systematic reviews and meta-analyses of case identification for depression were checked for further studies. Additionally experts in the field were contacted for further information on relevant studies.

    Study selection

    We included validation studies of mood questionnaires including the BDI,21 ,22 Patient Health Questionnaire-9 (PHQ-9)23 with linear scoring, Patient Health Questionnaire-2 (PHQ-2)24 with linear scoring, the two stem questions25 (either ‘low mood’ or ‘loss of interest’ found in the Diagnostic and Statistical Manual-Fourth Edition (DSM-IV) and the International Classification of Disease Tenth Edition (ICD-10)), General Health Questionnaire-12 (GHQ-12) and General Health Questionnaire-28 (GHQ-28),26 Center of Epidemiological Studies-Depression Scale (CES-D),27 GDS,28 GDS-15,29 Zung Depression Self Rating Depression Scale,30 HADS-D,31 HDRS32 (17 and 21 item versions were analysed together due to lack of separate data). The HDRS is usually applied by clinicians whereas the remaining scales are usually self-report. The reference standard was diagnosis according to DSM (eg, DSM-IV33) or ICD (eg, ICD-1034) criteria. Out of all studies, 10 used Major Depressive Disorder alone and these were entered in a subanalysis. The remainder used a broader definition of depression, for example major or minor depression, major or minor depression or dysthymia. Studies that did not clearly state the comparator to be DSM or ICD diagnosis of depression, or did not provide sufficient data to be extracted in the meta-analysis were excluded.

    Data extraction and critical appraisal

    All published studies that met our eligibility criteria were assessed for methodological quality independently by at least two reviewers using the Quality Assessment Tool for Diagnostic Accuracy Studies-2.35 Data were extracted independently by at least two authors and 2×2 tables were constructed from which the primary outcomes were calculated, that is, sensitivity and specificity. To maximise the available data we extracted the most consistently reported and recommended cut-off points for each of the scales.

    Data-analysis

    Bivariate and Hierarchical Summary Receiver Operating Curve (HSROC) meta-analyses were conducted where possible using R 2.15.1 using the meta-analysis of diagnostic accuracy (MADA) and HSROC packages in order to obtain pooled estimates of sensitivity and specificity (≥0.8 were considered acceptable, ≥0.9 were considered good) and HSROC curves. We required a minimum of three studies to ensure stable estimates were computed. Review Manager 5 (Cochrane Collaboration) was also used for displaying data from the meta-analyses estimates generated above. Heterogeneity was assessed by visual examination of forest plots and prediction regions of HSROC plots. In addition, meta-regression and subgroup analyses were conducted to explore heterogeneity. The impact on accuracy for patients in acute (eg, hospital setting and within 6 months of stroke) versus post-acute (receiving outpatient or inpatient rehabilitation treatment) settings; and criteria for depression (major depressive disorder or any depression) was examined. Sensitivity, specificity, positive predictive values (PPVs) and negative predictive values (NPVs) were calculated. Real world clinical utility was also calculated using the clinical utility index (CUI) which takes into account frequency and discrimination of a test (see http://www.clinicalutility.co.uk). A positive CUI (CUI+) gives the ability of a test to confirm a diagnosis (case-finding) with minimal false positives and a negative CUI (CUI−) the ability of a test to rule-out a diagnosis, that is, to screen with minimal false negatives.36

    Results

    A total of 24 studies on 2,907 participants were included in the review (see figure 1).16 ,37–59 The median prevalence for any depression in these studies was 18% (range 8–46%). Mean age ranged from 55 to 80 years with some studies specifically targeting either younger or older participants which may have introduced heterogeneity into the analyses. Most studies had an approximately equal distribution of men and women. Included studies assessed a total of 18 different scales (see table 1 for study characteristics).

    View this table:
    Table 1

    Summary characteristics of included studies

    Figure 1
    Figure 1

    Study flow diagram.

    Quality assessment of the included studies is summarised in online supplementary appendix 1. The most common risk of bias issues identified for these studies were not reporting blinding of assessments, not reporting predefined cut-offs for diagnostic thresholds and a lack of information concerning drop out.

    Diagnostic validity meta-analyses

    Any depression

    Sensitivity and specificity data for scales with at least two studies are summarised in figure 2A,B). Meta-analysis was possible for nine scales (see table 2 for summary sensitivity and specificity estimates and online supplementary appendix 2 for summary HSROC plot). Although there were three studies assessing the GDS the analysis had difficulty converging, therefore pooled estimates for this scale were not reported. It was not possible to conduct meta-analyses of direct comparisons (ie, including only data where each study compared two scales). Comparisons of the bivariate estimates suggest the scales with the highest sensitivity were the PHQ-9 and the BDI. Indirect comparisons of scales on the summary HSROC plot (see online supplementary appendix 2) suggests the three best performing scales in terms of sensitivity and specificity were the HDRS, PHQ-9 and MADRS. The HSROCs for each of these scales overlapped at most thresholds. While GDS-15 could potentially be the best measure at some thresholds it was also the poorest performing at other thresholds (requiring higher specificity). This appears to reflect the very high heterogeneity associated with this scale and uncertainty concerning its diagnostic accuracy. Substantial heterogeneity was observed for sensitivity and specificity estimates of most other scales (see figure 2A,B).

    View this table:
    Table 2

    Evidence summary on meta-analysis of depression scales for any depression

    Figure 2
    Figure 2

    (A) Forest plots for scales with a minimum of two studies. (B) Forest plots for scales with a minimum of two studies.

    Figure 2
    Figure 2

    continued.

    PPV (converted from the bivariate estimates of sensitivity and specificity) was highest for CES-D and also had the highest utility for screening (ruling out a diagnosis with minimal false negatives) in clinical practice (CUI−=0.77) followed by the HDRS (CUI−=0.76). The test with the lowest utility was likely to be BDI (CUI−=0.576). NPV was quite consistent across studies. Using the CUI, the test with the highest utility for case-finding (ruling in a diagnosis with minimal false positives) in clinical practice was likely to be either the HDRS (CUI+=0.60) or PHQ-9 (CUI+0.58) and the test with the lowest utility for case-finding was the BDI (CUI+=0.46).

    Subgroup analysis of major depression

    There were six scales with sufficient data to meta-analyse data for major depression (see table 3). HDRS appeared to have the best balance between sensitivity and specificity. BDI and HDRS had the highest sensitivity, and HDRS, HADS-T and PHQ-9 showed the highest specificity.

    View this table:
    Table 3

    Evidence summary on meta-analysis of depression scales for major depression

    PPV was highest for HDRS and also had the highest utility for screening (ruling out a diagnosis with minimal false negatives) in clinical practice (CUI−=0.86), which can be considered ‘excellent’ qualitatively. As above NPVs were very similar across studies.Using CUI, no test had adequate ability to rule in a diagnosis with minimal false positives although the optimal choice appeared to be the HDRS (CUI+=0.49) followed by the PHQ-9 (CUI+=0.35).

    Subgroup analysis of acute and post-acute stroke

    There were six scales with sufficient data for meta-analysis for acute stroke patients (see table 4). Sensitivity was highest for HDRS and HADS-T had the highest specificity. PPV was highest for HDRS. The test with the highest utility for screening (ruling out a diagnosis with minimal false negatives) in clinical practice was the HDRS (CUI−=0.80). Using CUI, only the HDRS had good ability to rule in a diagnosis with minimal false positives.

    View this table:
    Table 4

    Evidence summary on meta-analysis of depression scales for acute and non-acute stroke

    There were four scales where meta-analysis was possible in post-acute stroke patients. CES-D had the highest PPV and the highest utility for screening (CUI−=0.71) followed by the HDRS. None of the scales were rated as adequate for rule-in utility.

    Meta-regression findings

    Given the small number of studies for each scale there was very limited statistical power to identify potential covariates. No statistically significant effects were found on acute/post-acute stroke or type of depression assessed for any scale. It was not possible to conduct meta-regression for the HDRS or HADS-T as the models were unable to converge.

    Discussion

    Twenty-four studies were identified on screening for depression in stroke patients. Meta-analyses were conducted on the HDRS, HADS-D, MADRS, PHQ-9, PHQ-2, BDI, HADS-T, GDS-15 and CESD. Ideally we looked for independent verification of scale performance from multiple centres but data was often sparse, typically with three or four studies per scale included in analyses. The three scales that appeared optimal in this population were CES-D, HDRS and the PHQ-9. However, comparisons are limited by the lack of studies that directly compared one scale with another. No tool was adequately able to rule-in a diagnosis with few false positives. Conversely, the ability to exclude (rule-out) depression and confirm non-depressed status was high, with most scales making only one false negative error out of every 20 negative results. In clinical practice this means that no current scale is an adequate proxy of interview defined depression and all positive screening results should be followed by a thorough clinical assessment, diagnostic interview or repeat screen.

    There have been no meta-analyses or systematic reviews of case identification instruments in stroke populations. However, the results of the present systematic review are largely consistent with others on case identification for depression in people with chronic physical health problems,20 ,62 showing inadequate case-finding ability. In high risk, high prevalence settings case-finding ability would improve as a function of discrimination (in this case PPV) and as a function of frequency in depressed patients (sensitivity). Given the interest in ultra-short screening techniques, a surprising result from our meta-analysis was that the PHQ-2 did not appear to perform as seen in other populations. The sensitivity and specificity of the PHQ-2 in stroke were 79% and 76% but we found 96%, 89% in cancer settings63 and 92% and 69%i in primary care, respectively. However in all settings where the prevalence is low rule-out performance of the PHQ2 may be adequate including patients with chronic physical health problems.

    One of the limitations of this systematic review concerns the lack of studies for most included scales. It is therefore possible that the conclusions of this review may change once further research has been conducted. In particular, data on two item measures such as the PHQ-2 and the two stem questions, found by systematic reviews in relatively similar populations to have good psychometric properties, were limited. A further limitation of the review was the substantial heterogeneity identified for most scales in the meta-analysis. However, this is a problem found in most meta-analyses of case identification measures.61 Further primary studies are required before a more detailed exploration of heterogeneity is possible. A further potential limitation is that there are no available diagnostic criteria specifically for poststroke depression. Although as argued above there is limited evidence that poststroke depression differs from major depressive disorder.

    Further research is needed to confirm the optimal tool for identifying depression after a stroke. Studies should focus on measures with easy interpretation and high acceptability for healthcare professionals and stroke patients. For example, although the HDRS performed well, it usually is considered a scale requiring significant time from the clinician. Similarly the BDI and GHQ-28 are probably too long for regular administration in clinical practice. Ultra-short measures such as the PHQ-2 and the two stem questions are often preferred by clinicians. New studies are beginning to investigate the use of visual-analogue scales such as the Distress Thermometer.60 It is important to note that diagnostic accuracy is no guarantee of real world benefit. Scales should be embedded into screening programmes assessing unmet needs, and desire for specialist help and then subjected to randomised controlled trials of screening versus diagnosis-as-usual.

    In conclusion, there appear to be a number of potentially useful scales but none with wholly adequate rule-in and rule-out performance.

    Acknowledgments

    We would like to thank the following for responses to requests for further data: Dr Ann Charlotte Laska, Danderyd Hosptial, Sweden; Prof Nadina Lincoln, Univerisity of Nottingham, UK; Dr Ulrike Sagen, Telemark Hospital, Norway.

    References

    1. ↵
      1. Whyte EM,
      2. Mulsant BH
      . Post-stroke depression: epidemiology, pathophysiology, and biological treatment. Biol Psychiatry 2002;52:253–64.
      OpenUrlCrossRefPubMedWeb of Science
    2. ↵
      1. Lee S,
      2. Shafe AC,
      3. Cowie MR
      . UK stroke incidence, mortality and cardiovascular risk management 1999–2008: time-trend analysis from the General Practice Research Database. BMJ Open 2011;1:e000269.
      OpenUrlAbstract/FREE Full Text
    3. ↵
      1. Jungehulsing GJ,
      2. Muller-Nordhorn J,
      3. Nolte CH,
      4. et al
      . The prevalence of stroke and stroke symptoms: a population-based survey of 28,090 participants. Neuroepidemiology 2008;30:51–7.
      OpenUrlCrossRefPubMed
    4. ↵
      1. Egede LE
      . Major depression in individuals with chronic medical disorders. Prevalence, and correlates and associates of health resource utilisation, lost productivity and functional disability. Gen Hosp Psychiatry 2007;29:409–16.
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      National Collaborating Centre for Mental Health. Depression in people with a chronic physical health problem. London/Leicester: Royal College of Psychiatrists’/British Psychological Society, 2009.
    6. ↵
      1. Kessler RC,
      2. Berglund R,
      3. Demler O
      . et al. The epidemiology of major depressive disorder: results from the National Comorbidity Survey Replication (NCS-R). JAMA 2003;289:3095–105.
      OpenUrlCrossRefPubMedWeb of Science
    7. ↵
      1. Hackett ML,
      2. Chaturangi Y,
      3. Parag P,
      4. et al
      . Frequency of depression after stroke: a systematic review of observational studies. Stroke 2005;36:1330–40.
      OpenUrlAbstract/FREE Full Text
    8. ↵
      1. Ayerbe L,
      2. Ayis S,
      3. Rudd AG,
      4. et al
      . Natural history, predictors, and associations of depression 5 years after stroke: the South London Stroke Register. Stroke 2011;42:1907–11.
      OpenUrlAbstract/FREE Full Text
    9. ↵
      1. Chemerinski E,
      2. Robinson RG,
      3. Kosier JT
      . Improved recovery in activities of daily living associated with remission of post-stroke depression. Stroke 2001;32:113–17.
      OpenUrlAbstract/FREE Full Text
    10. ↵
      1. Schmid AA,
      2. Kroenke K,
      3. Hendrie HC,
      4. et al
      . Poststroke depression and treatment effects on functional outcomes. Neurology 2011;76:1000–5.
      OpenUrlCrossRef
    11. ↵
      1. Gump BB,
      2. Matthews KA,
      3. Eberly LE,
      4. et al
      . For the MRG Depressive symptoms and mortality in men: results from the Multiple Risk Factor Intervention Trial. Stroke 2005;36:98–102.
      OpenUrlAbstract/FREE Full Text
    12. ↵
      1. House A,
      2. Knapp P,
      3. Bamford J,
      4. et al
      . Mortality at 12 and 24 months after stroke maybe associated with depressive symptoms at 1 month. Stroke 2001;32:696–701.
      OpenUrlAbstract/FREE Full Text
    13. ↵
      1. van de Weg FB,
      2. Kuik DJ,
      3. Lankhorst GJ
      . Post-stroke depression and functional outcome: a cohort study investigating the influence of depression on functional recovery from stroke. Clin Rehabil 1999;13:268–72.
      OpenUrlAbstract/FREE Full Text
    14. ↵
      1. Nannetti L,
      2. Paci M,
      3. Pasquini J,
      4. et al
      . Motor and functional recovery in patients with post-stroke depression. Disabil Rehabil 2005;27:170–5.
      OpenUrlCrossRefPubMedWeb of Science
    15. ↵
      1. Fruehwald S,
      2. Gatterbauer E,
      3. Rehak P,
      4. et al
      . Early fluoxetine treatment of post-stroke depression—a three-month double-blind placebo-controlled study with an open-label long-term follow up. J Neurol 2003;250:347–51.
      OpenUrlCrossRefPubMedWeb of Science
    16. ↵
      1. Williams CL,
      2. Rittman MR,
      3. Boylstein C,
      4. et al
      . Qualitative and quantitative measurement of depression in veterans recovering from stroke. J Rehabil Res Dev 2005;42:277–90.
      OpenUrlCrossRefPubMed
    17. ↵
      1. Lipsey JR,
      2. Spencer WC,
      3. Rabins PV,
      4. et al
      . Phenomenological comparison of poststroke depression and functional depression. Am J Psychiatry 1986;143:527–9.
      OpenUrlPubMedWeb of Science
    18. ↵
      1. Anderson G,
      2. Vestegaard K,
      3. Riis J,
      4. et al
      . Incidence of post-stroke depression during the first year in a large unselected stroke population determined using a valid standardized rating scale. Acta Psychiatr Scand 1994;90:190–5.
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      1. Cumming TB,
      2. Churilov L,
      3. Skoog I,
      4. et al
      . Little evidence for different phenomenology in poststroke depression. Acta Psychiatr Scand 2010;121:424–30.
      OpenUrlCrossRefPubMed
    20. ↵
      1. Mitchell A,
      2. Meader N,
      3. Davies E,
      4. et al
      . Meta-analysis of screening and case finding tools for depression in cancer: evidence based recommendations for clinical practice on behalf of the Depression in Cancer Care consensus group. J Affect Disord 2012;140:149–60.
      OpenUrlCrossRefPubMed
    21. ↵
      1. Beck AT,
      2. Steer RA,
      3. Brown GK
      . Beck depression inventory. 2nd edn. Manual, San Antonio: The Psychological Corporation, 1996.
    22. ↵
      1. Beck AT,
      2. Guth D,
      3. Steer RA,
      4. et al
      . Screening for major depression disorders in medical inpatients with the Beck Depression Inventory for Primary Care. Behav Res Ther 1997;35:785–91.
      OpenUrlCrossRefPubMedWeb of Science
    23. ↵
      1. Spitzer RL,
      2. Kroenke K,
      3. Williams JB,
      4. et al
      . Validation and utility of a self-report version of the PRIME-MD: the PHQ primary care study. JAMA 1999;282:1737–44.
      OpenUrlCrossRefPubMedWeb of Science
    24. ↵
      1. Kroenke K,
      2. Spitzer RL,
      3. Williams JBW
      . The Patient Health Questionnaire-2: validity of a two-item screener. Med Care 2003;41:1284–92.
      OpenUrlCrossRefPubMedWeb of Science
    25. ↵
      1. Whooley MA,
      2. Avins AL,
      3. Miranda J,
      4. et al
      . Case-finding instruments for depression. Two questions are as good as many. J Gen Intern Med 1997;12:439–45.
      OpenUrlCrossRefPubMedWeb of Science
    26. ↵
      1. Goldberg DP,
      2. Williams P
      . A user's guide to the general health questionnaire. Windsor: NFER-Nelson, 1988.
    27. ↵
      1. Radloff LS
      . The CESD scale: a self report scale for research in the general population. Appl Psychol Meas 1977;1:385–401.
      OpenUrlCrossRef
    28. ↵
      1. Yesavage JA,
      2. Brink TL,
      3. Rose TL,
      4. et al
      . Development and validation of a geriatric depression screening scale: a preliminary report. J Psychiatr Res 1982;17:37–49.
      OpenUrlCrossRefPubMedWeb of Science
    29. ↵
      1. Sheikh JI,
      2. Yesavage JA
      . Geriatric Depression Scale (GDS): Recent evidence and development of a shorter version. In: Brink TL.ed. Clinical gerontology: a guide to assessment and intervention in NY: New York, The Haworth Press 1986: 165–73.
    30. ↵
      1. Zung WWK
      A self-rating depression scale. Arch Gen Psychiatry 1965;12:63–70.
      OpenUrlCrossRefPubMedWeb of Science
    31. ↵
      1. Zigmond AS,
      2. Snaith RP
      . The hospital anxiety and depression scale. Acta Psychiatr Scand 1983;67:361–70.
      OpenUrlCrossRefPubMedWeb of Science
    32. ↵
      1. Hamilton M
      . The hamilton depression rating scale. J Neurol Neurosurg Psychiatry 1960;23:56–62.
      OpenUrlFREE Full Text
    33. ↵
      American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4th edn. (DSM-IV). Washington, DC: APA, 1994.
    34. ↵
      World Health Organization. The ICD-10 classification of mental and behavioural disorders. Geneva: World Health Organization, 1993.
    35. ↵
      1. Whiting PF,
      2. Rutjes AWS,
      3. Westwood AE,
      4. et al
      . QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med 2011;155:529–36.
      OpenUrlCrossRefPubMedWeb of Science
    36. ↵
      1. Mitchell AJ
      . Sensitivity×PPV is a recognized test called the clinical utility index (CUI+). Eur J Epidemiol 2011;26:251–2.
      OpenUrlCrossRefPubMed
    37. ↵
      1. Aben I,
      2. Verhey F,
      3. Lousberg R,
      4. et al
      . Validity of the beck depression inventory, hospital anxiety and depression scale, SCL-90 andhamilton depression rating scale as screening instruments for depression in stroke patients. Psychosomatics 2002;43:386–93.
      OpenUrlCrossRefPubMedWeb of Science
    38. ↵
      1. Agrell B,
      2. Dehlin O
      . Comparison of six depression rating scales in geriatric stroke patients. Stroke 1989;20:1190–4.
      OpenUrlAbstract/FREE Full Text
    39. ↵
      1. Johnson G,
      2. Burvill PW,
      3. Anderson CS,
      4. et al
      . Screening instruments for depression and anxiety following stroke. Acta Psychiatr Scand 1995;91:252–7.
      OpenUrlCrossRefPubMedWeb of Science
    40. ↵
      1. Laska AC,
      2. Martensson B,
      3. Kahan T
      . Recognition of depression in aphasic stroke patients. Cerebrovasc Dis 2007;24:74–9.
      OpenUrlCrossRefPubMed
    41. ↵
      1. Lightbody CE,
      2. Baldwin R,
      3. Connolly M,
      4. et al
      . Can nurses help identify patients with depression following stroke? A pilot study using two methods of detection. J Adv Nurs 2007;7:505–12.
      OpenUrl
    42. ↵
      1. Parikh RM,
      2. Eden DT,
      3. Price TR,
      4. et al
      . The sensitivity and specificity of the Center for Epidemiologic Studies Depression Scale in screening for post-stroke depression. Int J Psychiatry Med 1988;18:169–81.
      OpenUrlCrossRefPubMedWeb of Science
    43. ↵
      1. Quaranta D,
      2. Marra C,
      3. Gainotti G
      . Mood disorders after stroke: diagnostic validation of the poststroke depression rating scale. Cerebrovasc Dis 2008;26:237–43.
      OpenUrlCrossRefPubMed
    44. ↵
      1. Roger PR,
      2. Johnson-Greene D
      . Comparison of assessment measures for post-stroke depression. Clin Neuropsychol 2009;23:780–93.
      OpenUrlCrossRefPubMedWeb of Science
    45. ↵
      1. Sagen U,
      2. Vik TG,
      3. Moum T,
      4. et al
      . Screening for depression and anxiety after stroke: comparison of the Hospital Anxiety and Depression Scale and the Montgomery Asberg Depression Rating Scale. J Psychosom Res 2009;67:325–32.
      OpenUrlCrossRefPubMed
    46. ↵
      1. Shinar D,
      2. Gross CR,
      3. Price TR,
      4. et al
      . Screening for depression in stroke patients: the reliability and validity of the Center for Epidemiologic Studies Depression scale. Stroke 1986;17:241–5.
      OpenUrlAbstract/FREE Full Text
    47. ↵
      1. Tang WK,
      2. Ungvari GS,
      3. Chiu HFK,
      4. et al
      . Screening post-stroke depression in Chinese older adults using the Hospital Anxiety and Depression Scale. Aging Ment Health 2004;8:397–9.
      OpenUrlCrossRefPubMed
    48. ↵
      1. Tang WK,
      2. Chan S,
      3. Chiu H,
      4. et al
      . Can the Geriatric Depression Scale detect poststroke depression in Chinese elderly? J Affect Disord 2004;81:153–6.
      OpenUrlCrossRefPubMedWeb of Science
    49. ↵
      1. Tang WK,
      2. Ungvari GS,
      3. Chiu HFK,
      4. et al
      . Detecting depression in Chinese stroke patients: a pilot study comparing four screening instruments. Int J Psychiatry Med 2004;34:155–63.
      OpenUrlCrossRefPubMed
    50. ↵
      1. Berg A,
      2. Lonnqvist J,
      3. Palomaki H,
      4. et al
      . Assessment of depression after stroke: a comparison of different screening instruments. Stroke 2009;40:523–9.
      OpenUrlAbstract/FREE Full Text
    51. ↵
      1. Healey AK,
      2. Kneebone II,
      3. Carroll M,
      4. et al
      . A preliminary investigation of the reliability and validity of the Brief Assessment Schedule Depression Cards and the Beck Depression Inventory-Fast Screen to screen for depression in older stroke survivors. Int J Geriatr Psychiatry 2008;23:531–6.
      OpenUrlCrossRefPubMed
    52. ↵
      1. Lee AC,
      2. Tang SW,
      3. Yu GK,
      4. et al
      . The smiley as a simple screening tool for depression after stroke: a preliminary study. Int J Nurs Stud 2008;45:1081–9.
      OpenUrlCrossRefPubMedWeb of Science
    53. ↵
      1. Lincoln NB,
      2. Nicholl CR,
      3. Flannaghan T,
      4. et al
      . The validity of questionnaire measures for assessing depression after stroke. Clin Rehabil 2003;17:840–6.
      OpenUrlAbstract/FREE Full Text
    54. ↵
      1. Narding P,
      2. Leentjens AFG,
      3. van Kooten F,
      4. et al
      . Disease-specific properties of the Hamilton Rating Scale for Depression in patients with Stroke, Alzheimer's Dementia and Parkinson's Disease. J Neuropsychiatry Clin Neurosci 2002;14:329–34.
      OpenUrlCrossRefPubMedWeb of Science
    55. ↵
      1. O'Rourke S,
      2. MacHale S,
      3. Signorini D,
      4. et al
      . Detecting psychiatric morbidity after Stroke: comparison of the GHQ and the HAD scale. Stroke 1998; 29:980–5.
      OpenUrlAbstract/FREE Full Text
    56. ↵
      1. Sivrioglu EY,
      2. Sivrioglu K,
      3. Ertan T
      . Reliability and validity of the Geriatric Depression Scale in detection of post-stroke minor depression. J Clin Exp Neuropsychol 2009;31:999–1006.
      OpenUrlCrossRefPubMed
    57. ↵
      1. van Ginkel J,
      2. Hafsteinsdóttir T,
      3. Lindeman E
      ,et al. An efficient way to detect poststroke depression by subsequent administration of a 9-Item and a 2-item patient health questionnaire. Stroke 2012;43:854–6.
      OpenUrlAbstract/FREE Full Text
    58. ↵
      Williams LS, Brizendine EJ, Plue L, et al. Performance of the PHQ-9 as a Screening Tool for Depression After Stroke. Stroke 2005;36:635–8.
    59. ↵
      1. Kang HJ,
      2. Stewart R,
      3. Kim JM,
      4. et al
      . Comparative validity of depression assessment scales for screening poststroke depression. J Affect Disord. Published Online First: Nov 17 2012. doi:10.1016/j.jad.2012.10.035.
    60. ↵
      1. Turner A,
      2. Hambridge J,
      3. White J,
      4. et al
      . Depression screening in stroke: a comparison of alternative measures with the structured diagnostic interview for the diagnostic and statistical manual of mental disorders, fourth edition (major depressive episode) as criterion standard. Stroke 2012;43:1000–5.
      OpenUrlAbstract/FREE Full Text
    61. ↵
      1. Macaskill P,
      2. Gatsonis C,
      3. Deeks JJ,
      4. et al
      . Chapter 10: analysing and presenting results. In: Deeks JJ, Bossuyt PM, Gatsonis C. eds.Cochrane handbook for systematic reviews of diagnostic test accuracy version 1.0. The Cochrane Collaboration, 2010. http://srdta.cochrane.org/sites/srdta.cochrane.org/files/uploads/Chapter%2010%20-%20Version%201.0.pdf (accessed 29 January 2013).
    62. ↵
      1. Meader N,
      2. Mitchell A,
      3. Chew-Graham C,
      4. et al
      . Case identification of depression in patients with chronic physical health problems: a diagnostic accuracy meta-analysis of 113 studies. Br J Gen Pract 2011;61:808–20.
      OpenUrlCrossRef
    63. ↵
      1. Mitchell A
      . Are one or two simple questions sufficient to detect depression in cancer and palliative care? A Bayesian meta-analysis. Br J Cancer 2008;98:1934–43.
      OpenUrlCrossRefPubMedWeb of Science

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

      Files in this Data Supplement:

    Footnotes

    • Funding None.

    • Competing interests None.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • ↵i Data revised for 2012 (currently unpublished).