Research report
Validity of the Hospital Anxiety and Depression Scale to assess depression and anxiety following traumatic brain injury as compared with the Structured Clinical Interview for DSM-IV

https://doi.org/10.1016/j.jad.2008.06.007Get rights and content

Abstract

Background

Rating scales are often used in the assessment of depression and anxiety in traumatic brain injury (TBI), but few have been validated for use with this population. Overlap of symptoms between such disorders and TBI may lead to under- or over-diagnosis of depression or anxiety.

Methods

100 participants with mild to severe TBI, and 87 informants, were interviewed using the SCID-IV (Axis I). The HADS was administered at the same time.

Results

According to the SCID-IV, 34 participants were diagnosed with major depression and 36 with an anxiety disorder. Higher HADS scores were associated with a greater likelihood of depression and anxiety. However, the “clinical” categories of the HADS did not strongly correspond with the clinical diagnoses of depression and anxiety. Compared with SCID diagnoses, the depression subscale of the HADS had a sensitivity of 62% and a specificity of 92%. The anxiety subscale had a sensitivity of 75% and a specificity of 69%. Positive predictive and negative predictive values were calculated.

Limitations

This study included mostly moderate to severe TBI individuals, recruited from a rehabilitation hospital. Therefore, they may not necessarily be representative of the entire TBI population.

Conclusions

The HADS was a reliable measure of emotional distress in this TBI sample; however the cut-off scores and categories were not useful in predicting caseness of depression and anxiety. Clinicians should be mindful of the sequelae of TBI that may confound the scores yielded in rating scales and should follow up with a psychiatric interview when diagnosis is unclear.

Introduction

Psychiatric disorders occur with increased frequency following traumatic brain injury (TBI), but the reported rates of disorders are highly variable (Hibbard et al., 1998, Deb et al., 1999). This variability may reflect methodological differences between studies in terms of inclusion criteria, injury severity, time post-injury or the assessment tools used. The latter represents a major concern as most, if not all, psychiatric rating scales have not been validated for use in the TBI population.

Methods of assessment of depression and anxiety include rating scales and structured clinical interviews, based on DSM, ICD-10 or other criteria. Rating scales represent a time-efficient and easy method of measurement of current emotional state, requiring no specific training. Psychiatric interviews allow for more detailed assessment of the development and fluctuation of symptoms over an extended period. In the case of medical illness or injury, the assessor has to determine the aetiology of somatic symptoms. This is especially important in the case of TBI, which can cause symptoms also associated with depression or anxiety, including fatigue (Ziino and Ponsford, 2006), sleep disturbance (Vela-Bueno and Bixler, 2006), slowed thinking (Timmerman and Brouwer, 1999), attentional problems (Veltman et al., 1996, Brewer, 2003), planning and problem solving difficulties (Mamelak, 2000), increased aggression, mood swings or flattened affect (Eames, 1990). Socially withdrawal may result from impaired interpersonal skills (Ponsford et al., 1995). The extent to which symptoms endorsed on these rating scales are indicative of longer term psychiatric disturbance as opposed to the TBI itself remains unclear.

The Hospital Anxiety and Depression Scale (HADS) was designed as a brief self-report measure of the severity of depressive or anxious states (Snaith and Zigmond, 1994), not as a clinically diagnostic tool. The HADS has been employed in numerous studies of anxiety and depression following TBI (Medd and Tate, 2000, Hoofien et al., 2001, Powell et al., 2002). The reliability and validity of the HADS have been established in a number of non-TBI studies (Herrmann, 1997, Bjelland et al., 2002). It has satisfactorily classified symptom severity and presence or absence of anxiety and depression in patients with somatic, psychiatric and medical conditions and the general population (Bjelland et al., 2002), although it failed to adequately predict caseness in another study of depression associated with Hepatitis C (Golden et al., 2007). The HADS has been shown to discriminate between psychiatric and non-psychiatric groups of adolescents (White et al., 1999), and to be an acceptable screening tool in several languages (Herrero et al., 2003, Aro et al., 2004), and in various medical populations, including those with cancer (Osborne et al., 2004), non-cardiac chest pain (Kuijpers et al., 2003), Parkinson's Disease (Marinus et al., 2002) and stroke (Aben et al., 2002).

To our knowledge, only one study has attempted to validate the HADS in a population of individuals with TBI (Al-Adawi et al., 2007). Al-Adawi et al. compared the sensitivity and specificity of subscales from an Arabic translation of the HADS in predicting diagnoses on the CIDI in an Omani sample of 68 individuals with mild-severe TBI. Using scores > 7 to indicate caseness, they found poor tradeoffs between sensitivity and specificity, as reflected in low ROC curve areas (.53 for the depression subscale; not reported for anxiety). Their HADS-D subscale had a sensitivity of 18% and specificity of 97%, and the HADS-A subscale had a sensitivity of 33% and specificity of 100%. Al-Adawi et al. suggested their findings may have been affected by cultural differences in expression of emotions. Translation into the local dialect or the reading aloud of questions to participants may also have influenced responses. In light of these findings it would seem important to investigate the reliability and validity of this tool in the English-speaking TBI population.

The aims of the current study were:

  • 1.

    To examine the relationship between TBI participants' scores on the depression and anxiety subscales of the HADS and the presence or absence of clinical diagnoses of depression and anxiety according to DSM-IV criteria, as measured by the Structured Clinical Interview for DSM-IV (SCID-IV). It was hypothesised that higher scores on the HADS-D (depression) would be associated with greater likelihood of clinical diagnosis of depression, and higher scores on the HADS-A (anxiety) scale with presence of clinical diagnosis of anxiety. It was also hypothesised that those with SCID-diagnosed depression or anxiety would have scores in the “clinical” range on the HADS-D and HADS-A respectively.

  • 2.

    To examine and compare the sensitivity and specificity of different cut-off scores on HADS subscales in the prediction of anxiety and depression diagnoses according to the SCID-IV.

  • 3.

    For each possible cut-off score on the HADS subscales, to establish the positive predictive values (PPVs) and negative predictive values (NPVs) associated with that score.

Section snippets

Participants

Participants were enrolled in a study of the nature and frequency of psychiatric disorders following TBI (Whelan-Goodinson et al., in submission). They were 100 community-based participants with mild to severe TBI sustained 0.5 to 5.5 years previously, aged between 19 and 74. Participants were included if they had a Glasgow Coma Scale (GCS) score of 3–14, were sufficiently proficient in English to complete the interview and rating scale and had no history of previous TBI or neurological

Participants

Of the 100 participants, 71% were male. Mean age at assessment was 37.18 years (SD = 14.19, range 19–74) and average years of education was 11.70 (SD = 2.65, range 6–18). Average time post-injury of assessment was 2.98 years (SD = 1.47, range 0.5–5.5). Participants had a mean lowest GCS score of 9.10 (SD = 4.12, range 3–14), with 35% scoring 13–14, 20% 9–12 and 45% 3–8. Mean duration of PTA was 20.77 days (SD = 17.85, range 1–77), with 9% having PTA < 24 h, 20% PTA 1–7 days, 42% 8–28 days and 29% > 

Discussion

This study aimed to examine the sensitivity and specificity of the HADS against the “gold standard” diagnoses of the SCID-IV. As hypothesised, higher HADS subscale scores were associated with a greater likelihood of a clinical diagnosis, as participants with depression and anxiety had higher mean HADS-D and HADS-A scores respectively as compared with non-depressed and non-anxious participants. More severe depression was also associated with significantly higher HADS-D ratings. These findings

Role of funding source

The funding source for this study had no role in study design, data collection, data analysis or had any contribution to the manuscript or decision to submit the manuscript for publication.

Conflict of interest

No conflict declared.

References (38)

  • APA

    Diagnostic and Statistical Manual of Mental Disorders, 4th Edition Text Revision

    (2000)
  • AroP. et al.

    Validation of the translation and cross-cultural adaptation into Finnish of the Abdominal Symptom Questionnaire, the Hospital Anxiety and Depression Scale and the Complaint Score Questionnaire

    Scand. J. Gastroentero.

    (2004)
  • AshmanT.A. et al.

    Psychiatric challenges in the first 6 years after traumatic brain injury: cross-sequential analyses of Axis I disorders

    Arch. Phys. Med. Rehab.

    (2004)
  • BrewerT.

    Are you paying attention?

    Am. J. Nurs.

    (2003)
  • CollinsonP.

    Of bombers, radiologists and cardiologists: time to ROC

    Heart

    (1998)
  • DebS. et al.

    Neuropsychiatric consequences of traumatic brain injury: a comparison between two age groups

    Brain. Inj.

    (2007)
  • DebS. et al.

    Rate of psychiatric illness 1 year after traumatic brain injury

    Am. J. Psychiatry

    (1999)
  • EamesP.

    Organic bases of behaviour disorders after traumatic brain injury

  • FirstM.B. et al.

    Computer-assisted SCID Clinician Version (CAS-CV): Software Manual

    (2004)
  • Cited by (0)

    View full text