More than a quarter of us will experience an anxiety disorder at some point in our lives (Kessler et al 2005). These disorders disrupt lives and are costly for society (Fineberg et al 2003). Some people who experience anxiety disorders recover naturally, but many experience significant difficulties with anxiety on an ongoing or reoccurring basis. Those with persistent anxiety disorders may be particularly likely to benefit from treatment or intervention.
Understanding who is most likely to develop persistent anxiety disorders would help inform targeted prevention and early intervention. However, relatively little research has focused specifically on risk factors for persistent anxiety disorders (rather than risks for initial onset). Individual studies also tend to only focus on a small number of risks in a particular population or setting, which limits our ability to draw firm conclusions.
A new systematic review by Hovenkamp-Hermelink et al (2021) therefore set out to systematically review and evaluate evidence related to predictors of persistent anxiety disorders across the lifespan.
The authors systematically searched three databases (PubMed, PsycINFO, and Web of Science) to identify potentially relevant articles. Two authors then independently screened titles and abstracts, and where relevant reviewed full-texts. Key inclusion criteria were: 1) participants had an anxiety disorder at baseline; 2) the study reported a follow-up of at least 2 years, 3) the paper was in English, German or Dutch. Persistence was defined as “having an anxiety disorder diagnosis at both baseline and follow-up”.
Authors used the Quality In Prognosis Studies (QUIPS) tool (Hayden et al 2013) and an additional assessment of sample size to assess the risk of bias for each included study. These risk of bias assessments were then used to produce an overall judgement on study quality (high quality; moderate quality; low quality). Study characteristics, data on persistence, and predictors of a persistent anxiety disorder were extracted, and predictors of persistence were organised into the following groups: ‘clinical’ ‘psychological’, ‘biological’, ‘sociodemographic’ and ‘other’ characteristics.
For each predictor of persistence, the authors assessed the strength of evidence (strong evidence; moderate evidence; limited evidence; inconclusive evidence) using criteria related to the number and quality of studies, and consistency in findings. Consistent findings were defined as “similar findings (positive, negative, or no association) in 75% or more of the studies analysing that predictor”.
The search identified 5,389 unique abstracts and 48 studies were included in the review. About half of these studies included participants from one of three longitudinal adult cohort studies: the ‘NESDA’ cohort in the Netherlands, and the ‘NESARC’ and the ‘HARP’ cohorts in the USA. Most studies (34 out of 48) recruited and followed up adults, and a smaller number included children up to age 18 (7 studies), only adolescents (3 studies), or only older adults (4 studies). Follow-up periods ranged from 2 to 16 years, and studies either focused on a specific anxiety disorder (e.g. panic disorder), a group of anxiety disorders, or all anxiety disorders. The vast majority of studies were rated as ‘high quality’ (27 out of 48) or ‘moderate quality’ (18 out of 48). Where reported, percentage of participants with persistent anxiety disorders varied, but most studies (>70%) reported at least 30% of the sample had persistent anxiety disorders.
- The review found strong evidence that the following clinical characteristics predicted persistent anxiety disorders:
- more panic attacks,
- recent treatment seeking,
- and (in adults) comorbid ‘personality disorders’.
- There was also moderate evidence to support ‘poor clinical status’ after treatment as a predictor of persistence.
- Among adults, there was strong evidence that the following psychological characteristics predicted persistent anxiety disorders:
- higher avoidance,
- lower extraversion,
- higher anxiety sensitivity.
- Among both adults and children, the review identified moderate evidence that higher behavioural inhibition predicted persistent anxiety disorders.
- Notably, the review found consistent evidence that sociodemographic characteristics, (e.g. socioeconomic status) did not predict persistent anxiety disorders (i.e. no association).
- Where there was strong or moderate evidence related to biological characteristics (e.g. physical inactivity), these characteristics also did not predict persistence.
Evidence related to other predictors of persistence was largely limited or inconclusive because of a lack of studies and/or inconsistent findings.
The authors conclude:
this review of predictors of persistent anxiety disorders showed clinical and psychological characteristics, such as having panic attacks, comorbid personality disorders, seeking and receiving treatment, poor clinical status after treatment, higher severity and longer duration of avoidance behaviour, lower extraversion, higher anxiety sensitivity, and higher behavioural inhibition, to be the strongest predictors of a persistent course.
Strengths and limitations
This study used robust methods to provide a clear and detailed overview of the evidence related to predictors of persistent anxiety disorders. Including and clearly reporting on methods used to assess study quality and strength of evidence are particular strengths, and provide a helpful model for future, similar systematic reviews. Systematic reviews are often limited to papers published in English and so including papers in German and Dutch helps to address this common limitation.
Key limitations with this review are associated with existing evidence on the topic. The small number of studies focused on children and adolescents (given that anxiety disorders typically first emerge early in life) and the absence of studies in non-western countries are particularly striking, and limits our ability to draw conclusions for these populations. As the authors note, the varied follow-up duration (from 2-16 years), and failure to consider the intervening period between baseline and follow-up may have influenced results. This review also did not consider the potential overlapping or cumulative effects of predictors on persistence, and there may be important predictors neglected in existing studies and therefore not captured in this review’s results.
It is also important to note that this review does not distinguish between persistent specific anxiety disorders (e.g. presence of social anxiety disorder at baseline and follow-up) versus persistence of anxiety disorders generally (e.g. presence of at least one anxiety disorder at baseline and follow-up). This means that an individual with, for example, social anxiety disorder at baseline and generalised anxiety disorder at follow-up, would be classed as ‘not persistent’ in a study focused on social anxiety disorder, but ‘persistent’ in a study focused on all anxiety disorders. We do therefore need to be cautious comparing and collating findings related to predictors of persistence where study outcomes vary in this way.
Implications for practice
Findings from this review identify key clinical and psychological characteristics that place individuals at particular risk for developing persistent anxiety disorders. It will be important for those developing targeted approaches to prevention and intervention for anxiety disorders to investigate the potential benefit of assessing these risks to identify who best to target. However, critically, this review highlights key gaps in the existing evidence-base and areas for future research.
My research focuses on identifying and supporting children with anxiety disorders and so it is particularly clear to me that prospective international studies that track outcomes across the lifespan among children with anxiety disorders are needed. While research focuses on adult populations, we cannot be confident about which children are most at risk for persistent anxiety disorders and therefore particularly likely to benefit from support early in life.
Statement of interests
Hovenkamp-Hermelink, J. H., Jeronimus, B. F., Myroniuk, S., Riese, H., & Schoevers, R. A. (2021). Predictors of persistence of anxiety disorders across the lifespan: a systematic review. Lancet Psychiatry. 8, 428-443. https://doi.org/10.1016/S2215-0366(20)30433-8
Fineberg, N. A., Haddad, P. M., Carpenter, L., Gannon, B., Sharpe, R., Young, A. H., … & Sahakian, B. J. (2013). The size, burden and cost of disorders of the brain in the UK. Journal of Psychopharmacology, 27(9), 761-770. https://journals.sagepub.com/doi/full/10.1177/0269881113495118
Hayden, J. A., van der Windt, D. A., Cartwright, J. L., Côté, P., & Bombardier, C. (2013). Assessing bias in studies of prognostic factors. Annals of Internal Medicine, 158(4), 280-286.
Kessler, R. C., Berglund, P., Demler, O., Jin, R., Merikangas, K. R., & Walters, E. E. (2005). Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Archives of General Psychiatry, 62(6), 593-602. https://jamanetwork.com/journals/jamapsychiatry/article-abstract/208678