How do we make a diagnosis? Screening tools for anxiety disorders

shutterstock_179060606

Anxiety disorders are common in the UK, with the proportion of people in England with generalised anxiety disorder estimated to be 4.4% in the most recent Adult Psychiatric Morbidity in England survey (McManus et al, 2009).

Anxiety or stress, at appropriate levels, can make us perform better and keep our bodies ready for the evolved ‘fight or flight’ response. However, when this anxiety becomes too much or continues for too long, it can cause significant functional impairments.

As with most mental health problems, there is no ‘test’ to confirm anxiety disorders and it can be difficult to detect, therefore likely to be underdiagnosed. Knowing this, are there any screening tools suitable for anxiety disorders?

A group of North American researchers (Herr et al, 2014) decided to systematically review the evidence for screening tools for two anxiety disorders:

  1. Generalised anxiety disorder (GAD)
  2. Panic disorder
It's estimated that 4.4% of people in England have generalised anxiety disorder.

It’s estimated that 4.4% of people in England have generalised anxiety disorder.

Methods

The researchers included studies looking at people aged 18 or over, with GAD or panic disorder, who had been treated in general medical settings, focussing specifically on self-assessment tools. The studies compared the results of the tools with “an acceptable criterion standard”, meaning a diagnostic interview using criteria from DSM-III, ICD-9 or more recent revisions.

The authors looked for papers in key databases (MEDLINE, PsycINFO, and the Cochrane Library) published between 1980 and 2014. They only included English language studies from Western Europe, North America, New Zealand and Australia, arguing that this was to identify studies with the highest applicability to US populations. I would personally question this approach, given the diversity of the US population, but this was their method and there is no way of knowing what they might have missed.

Thirteen articles were selected, representing 10 studies. It was not specified within the main article what type of studies they were, although the authors did scrutinise the quality of the included studies. Most were based in primary care settings and the authors assessed applicability (setting, sample characteristics, anxiety disorder prevalence), test performance and quality (recruitment method, blinding, reference standard, sample size) and completed a Quality Assessment of Diagnostic Accuracy Studies and gave each a quality rating. Where 3 studies had examined the same tool, the results were combined and this seemed reasonable to do so.

This study aimed to systematically review the accuracy of self-report screening instruments in diagnosing GAD and panic disorder in adults.

This study aimed to systematically review the accuracy of self-report screening instruments in diagnosing generalised anxiety disorder and panic disorder in adults.

Results

Two screening instruments were identified as being appropriate for use in primary care and having good results.

For each instrument, specificity, sensitivity and likelihood ratios were calculated. A positive likelihood ratio (LR+) is the ratio of the likelihood of a positive test result in an individual with the condition to the likelihood of a positive test result in an individual without it. Similarly, a negative likelihood ratio (LR-) is the ratio of the likelihood of a negative test result in an individual with the condition to the likelihood of a negative test result in an individual without it.

For generalised anxiety disorder, the GAD-7 instrument, a 7-item rating scale:

  • Had good sensitivity (0.89, Confidence Interval (CI) 0.82 to 0.96)
  • Had good specificity (0.83, CI 0.80 to 0.85)
  • Had the best LR+ of all the instruments for GAD (5.1, CI 4.3 to 6.0)
  • Had the best LR- of all the instruments for GAD (0.13, CI 0.07 to 0.26)

For panic disorder, the PHQ, a 5-item yes or no screening tool:

  • Had the best sensitivity (0.81, CI 0.68 to 0.93) and specificity (0.99, CI 0.98 to 1.00) of all the instruments for panic disorder
  • Had the best LR+ of all the instruments for panic disorder (78, CI 29 to 210)
  • Had the best LR- of all the instruments for panic disorder (0.20, CI 0.11 to 0.37)

The Beck Anxiety Inventory-Primary Care showed promising results for detection of both GAD and panic disorder, although quality assessment showed it to be high risk of bias and small sample size (56 patients).

GAD-7 for generalised anxiety disorder and PHQ for panic disorder were singled out as the most appropriate for use in primary care. 

GAD-7 for generalised anxiety disorder and PHQ for panic disorder were singled out as the most appropriate screening tools for use in primary care.

Limitations

So, back to assessing GAD and panic disorder as separate entities: both of the studies looking at the GAD-7 and PHQ instruments were rated at the highest quality level. Great results, I hear you say! Have we found screening tools that can accurately detect and have very low chances of missing both generalised anxiety disorder and panic disorder? Well… not quite.

The bottom line is that neither instrument’s results have been replicated in more than one primary care population. As we know, research needs to be reproducible in order to be useful – a tool that produces one result in one group could easily produce a different result in another population if it has not be thoroughly tested.

Conclusions

What does this mean for us working in the woodland? The fact remains that you are your own best tool for diagnosis. There is not (and I believe, never will be) a satisfactory replacement for clinical assessment, judgment and a bit of elfish intuition. However, that doesn’t mean that you couldn’t do with a bit of help. If you suspect an underlying anxiety disorder and someone scores highly on these tools, it could point you in the right direction, even if it won’t tell you the definitive answer.

The authors acknowledge the major setback of not having found a study which replicated the above promising results, but their review is a good starting point. Both the GAD-7 and the PHQ are widely available and easily administered, so try them out yourself. They have potential for being useful additions to your clinical consultation.

We'd love to hear your thoughts on this new review.

We’d love to hear your thoughts on this new review.

What do you think?

I’d really like to hear your thoughts on this subject:

  • Should mental health diagnoses be further standardised? Should we be screening everyone in primary care?
  • If you are a primary care practitioner, do you use these tools, or would you, having read this article?
  • For those based in secondary care, how would you feel about referrals based on these instruments?
  • Last but not least, to those people affected by these conditions: how would you feel about filling in a questionnaire about your symptoms? Do you think it could help professionals you come into contact with?

Please share your thoughts with us using the comments form below.

Links

Herr NR, Williams JW Jr, Benjamin S, McDuffie J. Does this patient have generalized anxiety or panic disorder?: The Rational Clinical Examination systematic review. JAMA. 2014; 312(1): 78-84. [Abstract]

The GAD-7 and PHQ instruments are both freely available on the Pfizer www.phqscreeners.com website.

Generalized Anxiety Disorder 7-item (GAD-7) scale (PDF). Source: Spitzer RL, Kroenke K, Williams JBW, Lowe B. A brief measure for assessing generalized anxiety disorder. Arch Inern Med. 2006;166:1092-1097.

Spitzer RL, Kroenke K, Williams JB. Patient Health Questionnaire Primary Care Study Group. Validation and utility of a self-report version of PRIME-MD. JAMA. 1999;282(18):1737-1744.

McManus S et al (eds.). Adult psychiatric morbidity in England, 2007: results of a household survey. The NHS Information Centre for Health and Social Care. 2009.

Likelihood ratios from the Bandolier archives. (Examples of Likelihood ratios for readers who want to hone their skills in this area).

Share on Facebook Tweet this on Twitter Share on LinkedIn Share on Google+