Brief report
Determinants of early identification of suicidal ideation in patients treated with antidepressants or anxiolytics in general practice: A multilevel analysis

https://doi.org/10.1016/j.jad.2006.09.012Get rights and content

Abstract

Background

General practitioners (GPs) play a key role in identifying and managing patients with suicidal tendencies. Few studies, however, examine both GP and patient characteristics and GP practices associated with suicide assessment. This article aims to evaluate 1) GPs' success in early identification of suicidal ideation (SI) in patients starting antidepressant or anxiolytic treatment, and 2) patient- and GP-related factors associated with this success.

Methods

Survey of 144 GPs practising in southeastern France and of consecutive adults consulting them during June–October 2004 and prescribed antidepressant or anxiolytic treatment. Data were collected from GPs (consultation-questionnaires focusing on their prescription, diagnosis and detection of SI) and patients (self-administered questionnaires including the Hospital Anxiety and Depression scale and social and demographic characteristics). We used multilevel logistic regression to analyse factors associated with SI detection.

Results

GPs completed consultation-questionnaires for 713 patients, 405 of whom completed self-administered questionnaires. Eighty-nine patients (22%) reported SI; in 43 cases (48%) SI had not been detected by the GP. GPs detected SI more frequently when they had completed continuing medical education about depression, when patients had higher depressive symptom scores, and when consultations were relatively long.

Limitations

Study limited to patients receiving initial prescriptions for antidepressants or anxiolytics.

Conclusions

The percentage of undetected SI in this study population was high. Additional training of GPs increases the chances of detecting SI. Medical training and continuing medical education should include better instruction about SI risk factors and diagnosis, including non-major depressions, and stress that screening requires sufficient consultation time.

Introduction

General practitioners (GPs) have a key role in identifying and managing patients with suicidal tendency (Gaynes et al., 2004): significant proportions of people in psychological distress consult their GPs rather than mental health professionals (Paykel et al., 2005); half to two thirds of successful suicide completers visited their GP in the weeks or months before their death (Schulberg et al., 2004). Appropriate GP training programs on the diagnosis and treatment of major depressive disorders (MDD), as in Gotland, can reduce population rates of suicide and suicidal behaviour, although these findings are not always replicated (Rutz et al., 1992, Mann et al., 2005).

While there is some controversy about routine primary care screening for suicide risks in the general population (McNamee and Offord, 1994, Gaynes et al., 2004), several authors strongly recommend that physicians should be alert to suicidal ideation (SI: death wishes, consideration or planning of suicide) among patients with MDD or other suicide risk factors (Schulberg et al., 2004). SI, one of the strongest predictors of suicide, is a key factor in identifying suicide risk (Schneider et al., 2001, Gaynes et al., 2004).

Few studies, however, simultaneously examine both GP and patient characteristics and GP practices associated with suicide assessment (Schulberg et al., 2004, Matthews et al., 1994, Isometsa et al., 1995, Williams et al., 1999). Better understanding of the factors associated with GPs' suicide assessment practices may help in designing training programs. This article intends to evaluate 1) GPs' success in early identification of SI in patients for whom they are starting an antidepressant or anxiolytic treatment; and 2) patient- and GP-related characteristics associated with this success.

Section snippets

Population

The 550 members of a representative panel of GPs (Verger et al., 2005) practising in southeastern France were asked in May 2004 to participate in this study, which surveyed consecutive patients aged 18 years or older who consulted them in June–October 2004, whom they started on an antidepressant or anxiolytic, and who agreed in writing to participate. Patients were considered to “start” these drugs if they had had no such treatment during the preceding 6 months.

Instruments and procedure

When we contacted GPs, we

Results

Of the 550 GPs on our panel, 144 (26%) participated. They did not differ from non-participants in age (p = 0.16) or place of practice (rural/urban) (p = 0.98) but were more frequently women (31.9% vs 21.4% of non-participants, p = 0.01). They completed 713 consultation-questionnaires: 439 (61.5%) for patients prescribed antidepressants and 274 (38.5%) anxiolytics (alone). Of the 713 patients, 405 (56.8%) returned their questionnaires. They did not differ in age (p = 0.28), gender (p = 0.50), or

Discussion

Our results confirm previous findings that primary care physicians detect SI relatively infrequently (Matthews et al., 1994, Isometsa et al., 1995, Williams et al., 1999). The association observed between CME training in MDD and SI detection is consistent with previous findings (Pfaff et al., 2001). Although the transversal nature of our study prevents any causal inference, this suggests that the doctors with specific training may actively ask about SI, perhaps because of better confidence in

Acknowledgments

This study received support from the Southeastern France Regional Union of Private Practitioners, (URML-PACA), the Funds for quality insurance of private medical care (FAQSV-URCAM-PACA), the Foundation for Public Health of the MGEN (national teachers' union mutual insurance fund), the Southeastern Regional Direction for Sanitary and Social Affairs (DRASS-PACA), and the 2002 call for proposals of CNRS/MIRE/DREES/INSERM. We thank Jo Ann Cahn for reading the manuscript and improving our English.

References (26)

  • S.K. Goldsmith et al.

    Reducing Suicide: A National Imperative

    (2002)
  • H. Goldstein

    Multilevel Statistical Models

    (1995)
  • M. Hamilton

    Development of a rating scale for primary depressive illness

    Br. J. Soc. Clin. Psychol.

    (1967)
  • Cited by (23)

    • Work intensity in men and work-related emotional demands in women are associated with increased suicidality among persons attending primary care

      2018, Journal of Affective Disorders
      Citation Excerpt :

      About half of persons who complete suicide and two thirds of those who attempt suicide visited a GP in the preceding month (Houston et al., 2003; Raue et al., 2006). GP's recognition and management of suicidal patients can be improved (Fanello et al., 2002; Feldman et al., 2007; Houston et al., 2003; Marquet et al., 2005; Poma et al., 2011; Verger et al., 2007). Recently, social epidemiologists studied individual and area-level factors for suicidal behaviors, in line with Durkheim's historical theory (Durkheim, 1985), advising the use of multilevel methods (Rezaeian et al., 2005) and reporting heterogeneity among studies (Rehkopf and Buka, 2006).

    • Attempted and completed suicide in primary care: Not what we expected?

      2015, Journal of Affective Disorders
      Citation Excerpt :

      Suicidal ideas expressed at the last consultation are more frequent than reported previously (Houston et al., 2003; Marquet et al., 2005); but the underexpression is still high and concerning for men. The main factor is that GPs do not assess suicidal ideation (Vannoy and Robins, 2011), notably in case of lack of continuing medical education, shorter consultations and moderate depressive symptoms or anxiety symptoms (Verger et al., 2007). As expected, suicidal patients mainly consulted their GP before the event (Houston et al., 2003; Luoma et al., 2002; Stenager and Jensen, 1994).

    • An evidence-based approach to managing suicidal patients in the patient-centered medical home

      2014, Cognitive and Behavioral Practice
      Citation Excerpt :

      Part of managing any chronic disease includes understanding prevalence rates for the condition of interest (i.e., suicidal ideation or behaviors) within a population and measuring program outcomes within the clinic population. For example, prevalence rates for suicidal thoughts and behaviors among primary care patients typically range from 2% to 5% of patients (Cooper-Patrick, Crum, & Ford, 1994; Olfson,Weissman, Leon, Sheehan, & Farber, 1996; Pfaff & Almeida, 2005), although higher rates (3% to 7%) have been found in urban settings (Olfson et al., 2000; Zimmerman et al., 1995) and among higher risk subgroups such as patients with mental health diagnoses (12% to 30%; Bruce et al., 2004; Bryan, Corso, Rudd, & Cordero, 2008; Verger et al., 2007). Understanding the relative risk of one’s population is an important first step for considering how to best develop one’s approach to suicide risk management from a population health perspective.

    View all citing articles on Scopus
    View full text