- Split View
-
Views
-
Cite
Cite
Kaj Sparle Christensen, Tomas Toft, Lisbeth Frostholm, Eva Ørnbøl, Per Fink, Frede Olesen, Screening for common mental disorders: who will benefit? Results from a randomised clinical trial, Family Practice, Volume 22, Issue 4, August 2005, Pages 428–434, https://doi.org/10.1093/fampra/cmi032
- Share Icon Share
Abstract
Background. Outcomes of studies on mental health screening in primary care are conflicting. A feasible and effective case-finding approach could benefit both GPs and their patients.
Objectives. (1) to examine the effect of using a composite screening questionnaire (SQ) on GPs' recognition and provision of care, and (2) to outline useful strategies for case-finding.
Methods. 38 GPs in Aarhus County, Denmark, volunteered to participate in this trial. 1785 consecutive patients aged 18–65 years consulting with new health problems were included. Patients were screened before consultation using an SQ including scales for somatisation, anxiety, depression and alcohol abuse. Patients were randomised into one of two groups: 900 questionnaires were disclosed to and scored by GPs, 885 were blinded. Number of diagnoses, subjects of conversation, and actions taken were analysed. Additional analyses aimed to identify GP and patient factors that could predict improved outcomes.
Results. Overall, disclosure of SQ results increased GPs' recognition of mental disorders by 3.8% [95% confidence interval (CI) −0.5% to 8.0%], and 6.6% (95% CI 1.2% to 12.0%) for patients screened positive. There was a marked variation in GPs' detection rates, and for GPs with moderate or low recognition rates increases were significant (P = 0.001). Conversation on psychological topics increased by 3.2% (95% CI −0.7% to 7.1%), and by 7.0% (95% CI 1.8% to 12.3%) for patients screened positive. Rates of planned follow-up consultations increased by 3.9% (95% CI 0.6% to 7.3%) and by 4.9% (95% CI 0.7% to 9.1%) for patients screened positive. GPs' self-reported benefit from screening was related to better outcomes. A range of patient and GP factors suggesting added value from using SQs were identified.
Conclusion. GPs' recognition and provision of mental health care can be influenced by the use of composite SQs. Perceived benefit from screening may serve as a useful predictor of better patient management. Pragmatic case-finding approaches need further evaluation.
Christensen KS, Toft T, Frostholm L, Ørnbøl E, Fink P and Olesen F. Screening for common mental disorders: who will benefit? Results from a randomised clinical trial. Family Practice 2005; 22: 428–434.
Introduction
Mental disorders such as somatisation, anxiety, depression and alcohol abuse are prevalent in general practice, but often go unrecognised.1,2 Psychiatric screening questionnaires (SQs) have been suggested as possible instruments for improving recognition and clinical decision making.3 So far, only specific screens for depression and alcohol misuse have documented some positive effect on patient outcome,4,5 and it remains unclear whether composite screening tools should be used to identify a wider range of mental disorders.6 Our previous study7 demonstrated limited effectiveness of routine screening for common psychiatric disorders when results were compared to a standardised psychiatric interview. However, findings did suggest that screening may be useful for case-finding among a subgroup of patients with high SQ scores.
Use of RCTs in assessing the effectiveness of different screening instruments may be hampered by a range of pitfalls, e.g. use of (too) broad patient selection criteria, (self) selection of skilful GPs, and problems with increases in baseline detection (sensitisation and cross-contamination), all leading to (false) negative outcomes.8 In order to limit the screening criteria, research has tended to focus on identifying patients at risk of having a mental disorder9–11 rather than on identifying patients who the GPs would consider candidates for screening.12 Such case-finding strategies may be useful from an epidemiological point of view, but it remains unknown whether they are feasible and effective in daily clinical practice.13 A possible solution to the above-mentioned problems may be to investigate the GPs' self-reported benefit from screening results. It is of particular interest to learn how such measures link with recognition and care, and whether they can help identify candidates for case-finding.
The aim of the present paper was to investigate if routine screening for mental disorders influenced GPs' recognition and provision of care. Additionally, we investigated if GPs' self-reported benefit from screening could be a simple predictor of the diagnosis and treatment they offered. Based on GP and patient factors we finally outlined some implications for use of the screening instrument.
Methods
Participants
Aarhus County has 600 000 inhabitants living in rural and urban areas. The county is served by 431 GPs working in 271 practices. All the GPs were invited to participate in an RCT on recognition and treatment of functional illness in primary care (the FIP study). Thirty-eight GPs (8.8%) working in 28 practices accepted to participate. Half of the participating GPs were randomised to an educational programme on somatisation.14 The present study on psychiatric screening is part of the FIP study and includes all the participating GPs.
Included in the study were consecutive patients aged 18–65 years presenting with a new health problem within a 3-week period (study period: 3 March to 1 May 2000). Excluded were patients of non-Scandinavian descent, patients unable to speak or read Danish, and patients who were too ill or demented to read and fill in questionnaires. Only patients enrolled in the National Health Care Programme, which covers 98% of the Danish population, were included. All participating patients are listed with one specific GP and must primarily consult this doctor/practice. Nearly all specialised treatment, including hospital admission, requires GP referral, except in emergency cases. The Danish health care system is almost entirely tax financed and most medical care is free.
Interventions
All the included patients were screened in the waiting room before consultation. A one page SQ including various rating scales was used: the SCL-90R somatisation subscale (SCL-SOM),15 the Whiteley-7 scale,16 the anxiety and depression (SCL-8) subscale17,18 and an alcohol abuse scale (CAGE).19 Additionally, patients were asked about their reasons for consulting the GP (whether for physical, psychological, or both sorts of complaints) and to complete the SF-36.20
The patients were then randomised to either blinding or disclosure of their SQs to their GP. All GPs were instructed how to rate the questionnaire at the beginning of the consultation. Information was provided in terms of positive predictive values (PPVs) on somatisation (symptom and illness worry score), mental disorder in general, depression and alcohol dependence/abuse.21 Immediately after the consultation, the GPs completed a questionnaire on their own assessment, subjects of conversation, actions taken, and self-reported benefit from disclosed screening results, if any.
Outcomes
Main outcome measures were recorded immediately after the consultations by all GPs: Additional analyses were made to: Supplementary analyses were made to clarify whether outcomes could be explained by GP randomisation to the educational programme on somatisation.
Any mental disorder diagnosed, i.e. somatisation, anxiety, depression or alcohol abuse. The GPs were asked to classify the main problem presented into one of five categories: physical disease, probable physical disease, medically unexplained symptoms, mental illness, or no physical problem. Later, this categorisation was dichotomised into ‘physical disease’ (two first categories) or ‘somatisation’. The GPs were specifically asked whether the patient suffered from a significant psychiatric disorder, i.e. anxiety, depression or alcohol abuse.
Subjects of conversation discussed during consultation, i.e. physical, psychological, social, or existential issues.
Actions taken, i.e. prescriptions/tests/referrals, recommended sick leave and planned follow up.
Study whether patients' screening scores (dichotomised positive score defined as SCL-SOM >3, or Whiteley-7 >0, or SCL-8 >0, or CAGE >0) had an impact on GPs' diagnosis and patient management.
Clarify whether differences in outcome measures were explained by GP differences in recognition rates.
Test if GPs' self-reported benefit from the actual SQ information could predict diagnoses, subjects of conversation, and actions taken.
Identify variables predicting increased benefit from SQ-disclosure and an unrecognised high SQ-score (dichotomised score on SCL-SOM >3, or Whiteley-7 >1, or SCL-8 >1, or CAGE >1).7
Sample size
No sample size calculations were made for the screening intervention. Calculations targeted the educational programme on somatisation, which was thought to be the intervention least susceptible to change outcomes. A series of simulation studies were performed to assess how many GPs and patients would be required. A significant difference (P < 0.05) in health between intervention and control patients was ascertained with a power of 68% on inclusion of 300 patients with a positive somatoform diagnosis (15 GPs per group, 10 patients per GP), or with a power of 80% on inclusion of 400 patients with a positive somatoform diagnosis (20 GPs per group, patient-to-GP ratio = 10). Prevalence results from other studies22 suggested that 2000–2500 patients were needed to reach a sample size of 400 patients with a positive somatoform diagnosis.
Randomisation and blinding
The medical secretaries were instructed in registration and inclusion procedures. Having obtained the patient's informed consent, the secretary broke a sealed non-transparent envelope with an SQ and a GP questionnaire. Having completed the SQ, the patient was randomised to either blinding or disclosure of the SQ to the GP. A colour code on the GP questionnaire clearly stated whether the secretary should give the SQ to the GP (disclosure), or put it back in the envelope (blinding).
Statistical analysis
All randomised patients were included. Unanswered questions on the SQ were automatically scored zero, based on the assumption that patients only responded to questions relevant to them. Our analysis was based on random allocation, using complete data sets only. All statistical analyses were performed with STATA 7.0 and SPSS 10.0 for Windows. We investigated differences in baseline characteristics using unpaired t-tests and Chi-squared tests. Differences in main outcome measures and subgroup analyses were calculated using general linear modelling (family: Bernoulli and link: Identity) and logistic regression models adjusting for clustering at the GP level.23
Ethics
This study was approved by The Scientific Ethics Committee in the County of Aarhus, the Danish Data Protection Agency and the Scientific Research Evaluation Committee of the Danish College of General Practitioners.
Results
Participant flow
Among 2424 patients assessed for eligibility, 227 met the exclusion criteria, 274 patients refused to participate, and 138 did not participate for other reasons (time pressure in clinic, patients not bringing their reading glasses, etc.). Each patient was included only once; thus 1785 patients (81.2%) joined the study (Fig. 1).
Baseline data
The 38 participating GPs had spent fewer years in practice [mean 10.3 versus 14.1 years, likelihood ratio test (LR-test) < 0.005] than non-participating GPs, and participating GPs more often reported having participated in longer courses (duration more than 3 days) in communication skills or psychological therapy (52.8% versus 39.5%, LR-test < 0.05). No significant differences were found between participating and non-participating GPs regarding length of postgraduate psychiatric training, type of practice, GP age or gender.
Patients who refused to participate had a mean age of 42.2 years compared with 38.8 years among included patients (P < 0.001, t-test) whereas we found no statistically significant gender differences. No significant differences were found neither in age nor gender between included patients and patients not participating for other reasons. Equally, no statistically significant differences were found between the two randomisation groups with respect to baseline variables.7
Main outcome measures
Overall, disclosure resulted in an absolute increase of 3.8% [95% confidence interval (CI) −0.5 to 8.0] in GP registration rates of any mental disorders evaluated, and 6.6% (95% CI 1.2 to 12.0) for patients screened positive (Table 1). GPs with low or moderate detection rates at non-disclosure showed statistically significant increases (9.4%, 95% CI 0.02 to 0.16 and 6.0%, 95% CI 0.00 to 0.12) by disclosure whereas GPs with high detection rates at non-disclosure had no significant benefit (−2.2%, 95% CI −0.07 to 0.02) from screening disclosure. Among patients screened positive, the increase in rates of recognition of depression and alcohol abuse was significant.
. | Blinding (%) . | Disclosure (%) . | Δ%** . | 95% CI*** . | P value*** . | |||||
---|---|---|---|---|---|---|---|---|---|---|
All patients | n = 885 | n = 900 | – | – | 0.72 | |||||
Female % | 61 | 59 | – | – | 0.72 | |||||
Mean age, SD | 38.2, 12.9 | 39.3, 12.9 | – | – | 0.09 | |||||
Diagnoses | ||||||||||
Somatisationa | 176 (20.0) | 187 (21.1) | 1.0 | −2.7 to 4.8 | 0.59 | |||||
Anxietyb | 86 (9.7) | 95 (10.6) | 0.9 | −1.9 to 3.7 | 0.54 | |||||
Depressionc | 87 (9.8) | 108 (12.0) | 2.2 | −0.7 to 5.1 | 0.14 | |||||
Alcohol abused | 21 (2.4) | 36 (4.0) | 1.6 | 0.0 to 3.3 | <0.05 | |||||
Any diagnosis of a,b,c,d | 242 (27.5) | 278 (31.3) | 3.8 | −0.5 to 8.0 | 0.08 | |||||
Any diagnosis of b,c,d | 152 (17.2) | 186 (20.8) | 3.6 | −0.1 to 7.2 | 0.06 | |||||
Screened positive | n = 514 | n = 530 | ||||||||
Somatisationa | 145 (28.4) | 155 (29.7) | 1.3 | −3.5 to 5.9 | 0.59 | |||||
Anxietyb | 72 (14.0) | 87 (16.5) | 2.5 | −1.9 to 6.9 | 0.27 | |||||
Depressionc | 72 (14.0) | 103 (19.5) | 5.5 | 1.8 to 9.2 | <0.01 | |||||
Alcohol abused | 19 (3.7) | 36 (6.8) | 3.1 | 0.4 to 5.9 | 0.02 | |||||
Any diagnosis of a,b,c,d | 197 (38.6) | 236 (45.2) | 6.6 | 1.2 to 12.0 | 0.02 | |||||
Any diagnosis of b,c,d | 127 (24.7) | 173 (32.8) | 8.1 | 3.1 to 13.2 | <0.01 | |||||
Screened negative | n = 371 | n = 370 | ||||||||
Somatisationa | 31 (8.4) | 32 (8.7) | 0.3 | −3.8 to 4.4 | 0.87 | |||||
Anxietyb | 14 (3.8) | 8 (2.2) | −1.6 | −4.3 to 1.1 | 0.24 | |||||
Depressionc | 15 (4.1) | 5 (1.4) | −2.7 | −5.2 to −0.2 | 0.04 | |||||
Alcohol abused | 2 (0.5) | 0 (0) | −0.5 | −2.0 to 1.3 | 0.15 | |||||
Any diagnosis of a,b,c,d | 42 (12.2) | 45 (11.5) | −0.7 | −5.4 to 4.0 | 0.77 | |||||
Any diagnosis of b,c,d | 25 (6.8) | 13 (3.5) | −3.2 | −6.3 to −1.2 | 0.04 |
. | Blinding (%) . | Disclosure (%) . | Δ%** . | 95% CI*** . | P value*** . | |||||
---|---|---|---|---|---|---|---|---|---|---|
All patients | n = 885 | n = 900 | – | – | 0.72 | |||||
Female % | 61 | 59 | – | – | 0.72 | |||||
Mean age, SD | 38.2, 12.9 | 39.3, 12.9 | – | – | 0.09 | |||||
Diagnoses | ||||||||||
Somatisationa | 176 (20.0) | 187 (21.1) | 1.0 | −2.7 to 4.8 | 0.59 | |||||
Anxietyb | 86 (9.7) | 95 (10.6) | 0.9 | −1.9 to 3.7 | 0.54 | |||||
Depressionc | 87 (9.8) | 108 (12.0) | 2.2 | −0.7 to 5.1 | 0.14 | |||||
Alcohol abused | 21 (2.4) | 36 (4.0) | 1.6 | 0.0 to 3.3 | <0.05 | |||||
Any diagnosis of a,b,c,d | 242 (27.5) | 278 (31.3) | 3.8 | −0.5 to 8.0 | 0.08 | |||||
Any diagnosis of b,c,d | 152 (17.2) | 186 (20.8) | 3.6 | −0.1 to 7.2 | 0.06 | |||||
Screened positive | n = 514 | n = 530 | ||||||||
Somatisationa | 145 (28.4) | 155 (29.7) | 1.3 | −3.5 to 5.9 | 0.59 | |||||
Anxietyb | 72 (14.0) | 87 (16.5) | 2.5 | −1.9 to 6.9 | 0.27 | |||||
Depressionc | 72 (14.0) | 103 (19.5) | 5.5 | 1.8 to 9.2 | <0.01 | |||||
Alcohol abused | 19 (3.7) | 36 (6.8) | 3.1 | 0.4 to 5.9 | 0.02 | |||||
Any diagnosis of a,b,c,d | 197 (38.6) | 236 (45.2) | 6.6 | 1.2 to 12.0 | 0.02 | |||||
Any diagnosis of b,c,d | 127 (24.7) | 173 (32.8) | 8.1 | 3.1 to 13.2 | <0.01 | |||||
Screened negative | n = 371 | n = 370 | ||||||||
Somatisationa | 31 (8.4) | 32 (8.7) | 0.3 | −3.8 to 4.4 | 0.87 | |||||
Anxietyb | 14 (3.8) | 8 (2.2) | −1.6 | −4.3 to 1.1 | 0.24 | |||||
Depressionc | 15 (4.1) | 5 (1.4) | −2.7 | −5.2 to −0.2 | 0.04 | |||||
Alcohol abused | 2 (0.5) | 0 (0) | −0.5 | −2.0 to 1.3 | 0.15 | |||||
Any diagnosis of a,b,c,d | 42 (12.2) | 45 (11.5) | −0.7 | −5.4 to 4.0 | 0.77 | |||||
Any diagnosis of b,c,d | 25 (6.8) | 13 (3.5) | −3.2 | −6.3 to −1.2 | 0.04 |
Missing data on diagnosis a: 0.7% (n = 6) and 1.3% (n = 12) respectively, missing data on diagnoses b,c,d: all 0.1% (n = 1) and 0.4% (n = 4) respectively.
The diagnostic difference is calculated by subtracting percentage of blinded from percentage in disclosed.
Confidence intervals and p values obtained by general linear modelling adjusted for clustering at GP level.
SD = standard deviation.
. | Blinding (%) . | Disclosure (%) . | Δ%** . | 95% CI*** . | P value*** . | |||||
---|---|---|---|---|---|---|---|---|---|---|
All patients | n = 885 | n = 900 | – | – | 0.72 | |||||
Female % | 61 | 59 | – | – | 0.72 | |||||
Mean age, SD | 38.2, 12.9 | 39.3, 12.9 | – | – | 0.09 | |||||
Diagnoses | ||||||||||
Somatisationa | 176 (20.0) | 187 (21.1) | 1.0 | −2.7 to 4.8 | 0.59 | |||||
Anxietyb | 86 (9.7) | 95 (10.6) | 0.9 | −1.9 to 3.7 | 0.54 | |||||
Depressionc | 87 (9.8) | 108 (12.0) | 2.2 | −0.7 to 5.1 | 0.14 | |||||
Alcohol abused | 21 (2.4) | 36 (4.0) | 1.6 | 0.0 to 3.3 | <0.05 | |||||
Any diagnosis of a,b,c,d | 242 (27.5) | 278 (31.3) | 3.8 | −0.5 to 8.0 | 0.08 | |||||
Any diagnosis of b,c,d | 152 (17.2) | 186 (20.8) | 3.6 | −0.1 to 7.2 | 0.06 | |||||
Screened positive | n = 514 | n = 530 | ||||||||
Somatisationa | 145 (28.4) | 155 (29.7) | 1.3 | −3.5 to 5.9 | 0.59 | |||||
Anxietyb | 72 (14.0) | 87 (16.5) | 2.5 | −1.9 to 6.9 | 0.27 | |||||
Depressionc | 72 (14.0) | 103 (19.5) | 5.5 | 1.8 to 9.2 | <0.01 | |||||
Alcohol abused | 19 (3.7) | 36 (6.8) | 3.1 | 0.4 to 5.9 | 0.02 | |||||
Any diagnosis of a,b,c,d | 197 (38.6) | 236 (45.2) | 6.6 | 1.2 to 12.0 | 0.02 | |||||
Any diagnosis of b,c,d | 127 (24.7) | 173 (32.8) | 8.1 | 3.1 to 13.2 | <0.01 | |||||
Screened negative | n = 371 | n = 370 | ||||||||
Somatisationa | 31 (8.4) | 32 (8.7) | 0.3 | −3.8 to 4.4 | 0.87 | |||||
Anxietyb | 14 (3.8) | 8 (2.2) | −1.6 | −4.3 to 1.1 | 0.24 | |||||
Depressionc | 15 (4.1) | 5 (1.4) | −2.7 | −5.2 to −0.2 | 0.04 | |||||
Alcohol abused | 2 (0.5) | 0 (0) | −0.5 | −2.0 to 1.3 | 0.15 | |||||
Any diagnosis of a,b,c,d | 42 (12.2) | 45 (11.5) | −0.7 | −5.4 to 4.0 | 0.77 | |||||
Any diagnosis of b,c,d | 25 (6.8) | 13 (3.5) | −3.2 | −6.3 to −1.2 | 0.04 |
. | Blinding (%) . | Disclosure (%) . | Δ%** . | 95% CI*** . | P value*** . | |||||
---|---|---|---|---|---|---|---|---|---|---|
All patients | n = 885 | n = 900 | – | – | 0.72 | |||||
Female % | 61 | 59 | – | – | 0.72 | |||||
Mean age, SD | 38.2, 12.9 | 39.3, 12.9 | – | – | 0.09 | |||||
Diagnoses | ||||||||||
Somatisationa | 176 (20.0) | 187 (21.1) | 1.0 | −2.7 to 4.8 | 0.59 | |||||
Anxietyb | 86 (9.7) | 95 (10.6) | 0.9 | −1.9 to 3.7 | 0.54 | |||||
Depressionc | 87 (9.8) | 108 (12.0) | 2.2 | −0.7 to 5.1 | 0.14 | |||||
Alcohol abused | 21 (2.4) | 36 (4.0) | 1.6 | 0.0 to 3.3 | <0.05 | |||||
Any diagnosis of a,b,c,d | 242 (27.5) | 278 (31.3) | 3.8 | −0.5 to 8.0 | 0.08 | |||||
Any diagnosis of b,c,d | 152 (17.2) | 186 (20.8) | 3.6 | −0.1 to 7.2 | 0.06 | |||||
Screened positive | n = 514 | n = 530 | ||||||||
Somatisationa | 145 (28.4) | 155 (29.7) | 1.3 | −3.5 to 5.9 | 0.59 | |||||
Anxietyb | 72 (14.0) | 87 (16.5) | 2.5 | −1.9 to 6.9 | 0.27 | |||||
Depressionc | 72 (14.0) | 103 (19.5) | 5.5 | 1.8 to 9.2 | <0.01 | |||||
Alcohol abused | 19 (3.7) | 36 (6.8) | 3.1 | 0.4 to 5.9 | 0.02 | |||||
Any diagnosis of a,b,c,d | 197 (38.6) | 236 (45.2) | 6.6 | 1.2 to 12.0 | 0.02 | |||||
Any diagnosis of b,c,d | 127 (24.7) | 173 (32.8) | 8.1 | 3.1 to 13.2 | <0.01 | |||||
Screened negative | n = 371 | n = 370 | ||||||||
Somatisationa | 31 (8.4) | 32 (8.7) | 0.3 | −3.8 to 4.4 | 0.87 | |||||
Anxietyb | 14 (3.8) | 8 (2.2) | −1.6 | −4.3 to 1.1 | 0.24 | |||||
Depressionc | 15 (4.1) | 5 (1.4) | −2.7 | −5.2 to −0.2 | 0.04 | |||||
Alcohol abused | 2 (0.5) | 0 (0) | −0.5 | −2.0 to 1.3 | 0.15 | |||||
Any diagnosis of a,b,c,d | 42 (12.2) | 45 (11.5) | −0.7 | −5.4 to 4.0 | 0.77 | |||||
Any diagnosis of b,c,d | 25 (6.8) | 13 (3.5) | −3.2 | −6.3 to −1.2 | 0.04 |
Missing data on diagnosis a: 0.7% (n = 6) and 1.3% (n = 12) respectively, missing data on diagnoses b,c,d: all 0.1% (n = 1) and 0.4% (n = 4) respectively.
The diagnostic difference is calculated by subtracting percentage of blinded from percentage in disclosed.
Confidence intervals and p values obtained by general linear modelling adjusted for clustering at GP level.
SD = standard deviation.
Conversation on psychological issues increased by 3.2% (95% CI −0.7 to 7.1) when SQ information was available and by 7.0% (95% CI 1.8 to 12.3) for patients screened positive.
Providing GPs with screening information increased the rates of planned follow-ups by 3.9% (95% CI 0.6 to 7.3), and by 4.9% (95% CI 0.7 to 9.1) for patients screened positive (Table 2). No significant differences were found in the rates of prescriptions/tests/referrals, or recommended sick leave.
. | Blinding (%) . | Disclosure (%) . | Δ%b . | 95% CIc . | P valuec . |
---|---|---|---|---|---|
All patients | n = 885 | n = 900 | |||
Prescription/test/referral | 563 (63.7) | 582 (65.0) | 1.3 | −1.9 to 4.4 | 0.43 |
Recommended sick leave | 39 (4.4) | 41 (4.6) | 0.2 | −1.7 to 2.0 | 0.86 |
Planned follow up | 198 (22.4) | 236 (26.3) | 3.9 | 0.6 to 7.3 | 0.02 |
Screened positive | n = 514 | n = 530 | |||
Prescription/test/referral | 318 (61.9) | 339 (64.3) | 2.5 | −2.2 to 7.1 | 0.30 |
Recommended sick leave | 27 (5.3) | 28 (5.3) | 0.1 | −2.8 to 2.9 | 0.97 |
Planned follow up | 134 (26.1) | 163 (30.9) | 4.9 | 0.7 to 9.1 | 0.02 |
Screened negative | n = 371 | n = 370 | |||
Prescription/test/referral | 245 (66.2) | 243 (65.9) | −0.4 | −8.0 to 7.3 | 0.93 |
Recommended sick leave | 12 (3.2) | 13 (3.5) | 0.3 | −2.2 to 2.8 | 0.83 |
Planned follow up | 64 (17.3) | 73 (19.8) | 2.5 | −3.8 to 8.8 | 0.44 |
. | Blinding (%) . | Disclosure (%) . | Δ%b . | 95% CIc . | P valuec . |
---|---|---|---|---|---|
All patients | n = 885 | n = 900 | |||
Prescription/test/referral | 563 (63.7) | 582 (65.0) | 1.3 | −1.9 to 4.4 | 0.43 |
Recommended sick leave | 39 (4.4) | 41 (4.6) | 0.2 | −1.7 to 2.0 | 0.86 |
Planned follow up | 198 (22.4) | 236 (26.3) | 3.9 | 0.6 to 7.3 | 0.02 |
Screened positive | n = 514 | n = 530 | |||
Prescription/test/referral | 318 (61.9) | 339 (64.3) | 2.5 | −2.2 to 7.1 | 0.30 |
Recommended sick leave | 27 (5.3) | 28 (5.3) | 0.1 | −2.8 to 2.9 | 0.97 |
Planned follow up | 134 (26.1) | 163 (30.9) | 4.9 | 0.7 to 9.1 | 0.02 |
Screened negative | n = 371 | n = 370 | |||
Prescription/test/referral | 245 (66.2) | 243 (65.9) | −0.4 | −8.0 to 7.3 | 0.93 |
Recommended sick leave | 12 (3.2) | 13 (3.5) | 0.3 | −2.2 to 2.8 | 0.83 |
Planned follow up | 64 (17.3) | 73 (19.8) | 2.5 | −3.8 to 8.8 | 0.44 |
Missing data on actions taken 0.1% (n = 1) and 0.4% (n = 4) respectively.
The difference in actions taken is calculated by subtracting percentage of blinded from percentage in disclosed.
95% confidence intervals and P values obtained by general linear modelling adjusted for clustering at GP level.
. | Blinding (%) . | Disclosure (%) . | Δ%b . | 95% CIc . | P valuec . |
---|---|---|---|---|---|
All patients | n = 885 | n = 900 | |||
Prescription/test/referral | 563 (63.7) | 582 (65.0) | 1.3 | −1.9 to 4.4 | 0.43 |
Recommended sick leave | 39 (4.4) | 41 (4.6) | 0.2 | −1.7 to 2.0 | 0.86 |
Planned follow up | 198 (22.4) | 236 (26.3) | 3.9 | 0.6 to 7.3 | 0.02 |
Screened positive | n = 514 | n = 530 | |||
Prescription/test/referral | 318 (61.9) | 339 (64.3) | 2.5 | −2.2 to 7.1 | 0.30 |
Recommended sick leave | 27 (5.3) | 28 (5.3) | 0.1 | −2.8 to 2.9 | 0.97 |
Planned follow up | 134 (26.1) | 163 (30.9) | 4.9 | 0.7 to 9.1 | 0.02 |
Screened negative | n = 371 | n = 370 | |||
Prescription/test/referral | 245 (66.2) | 243 (65.9) | −0.4 | −8.0 to 7.3 | 0.93 |
Recommended sick leave | 12 (3.2) | 13 (3.5) | 0.3 | −2.2 to 2.8 | 0.83 |
Planned follow up | 64 (17.3) | 73 (19.8) | 2.5 | −3.8 to 8.8 | 0.44 |
. | Blinding (%) . | Disclosure (%) . | Δ%b . | 95% CIc . | P valuec . |
---|---|---|---|---|---|
All patients | n = 885 | n = 900 | |||
Prescription/test/referral | 563 (63.7) | 582 (65.0) | 1.3 | −1.9 to 4.4 | 0.43 |
Recommended sick leave | 39 (4.4) | 41 (4.6) | 0.2 | −1.7 to 2.0 | 0.86 |
Planned follow up | 198 (22.4) | 236 (26.3) | 3.9 | 0.6 to 7.3 | 0.02 |
Screened positive | n = 514 | n = 530 | |||
Prescription/test/referral | 318 (61.9) | 339 (64.3) | 2.5 | −2.2 to 7.1 | 0.30 |
Recommended sick leave | 27 (5.3) | 28 (5.3) | 0.1 | −2.8 to 2.9 | 0.97 |
Planned follow up | 134 (26.1) | 163 (30.9) | 4.9 | 0.7 to 9.1 | 0.02 |
Screened negative | n = 371 | n = 370 | |||
Prescription/test/referral | 245 (66.2) | 243 (65.9) | −0.4 | −8.0 to 7.3 | 0.93 |
Recommended sick leave | 12 (3.2) | 13 (3.5) | 0.3 | −2.2 to 2.8 | 0.83 |
Planned follow up | 64 (17.3) | 73 (19.8) | 2.5 | −3.8 to 8.8 | 0.44 |
Missing data on actions taken 0.1% (n = 1) and 0.4% (n = 4) respectively.
The difference in actions taken is calculated by subtracting percentage of blinded from percentage in disclosed.
95% confidence intervals and P values obtained by general linear modelling adjusted for clustering at GP level.
Additional analyses
The GPs' self-reported benefit from SQ information with respect to identifying the actual problem was found to predict detection of any evaluated diagnoses, discussion of non-physical (psychological, social, or existential) issues, and follow-up appointments (Table 3).
Outcome . | Odds Ratiob . | CI 95% . | ||
---|---|---|---|---|
Diagnoses | ||||
Somatisation | 2.71 | 1.62–4.55 | ||
Anxiety | 2.89 | 1.45–4.74 | ||
Depression | 2.88 | 1.81–4.56 | ||
Alcohol abuse | 4.30 | 1.55–11.91 | ||
Subjects of conversation | ||||
Physical | 0.66 | 0.30–1.47 | ||
Psychological | 4.21 | 2.54–6.97 | ||
Social | 2.41 | 1.60–3.63 | ||
Existential | 3.35 | 2.11–5.33 | ||
Actions taken | ||||
Prescription/test/referral | 1.03 | 0.79–1.36 | ||
Recommended sick leave | 1.11 | 0.48–2.53 | ||
Planned follow up | 2.34 | 1.50–3.65 |
Outcome . | Odds Ratiob . | CI 95% . | ||
---|---|---|---|---|
Diagnoses | ||||
Somatisation | 2.71 | 1.62–4.55 | ||
Anxiety | 2.89 | 1.45–4.74 | ||
Depression | 2.88 | 1.81–4.56 | ||
Alcohol abuse | 4.30 | 1.55–11.91 | ||
Subjects of conversation | ||||
Physical | 0.66 | 0.30–1.47 | ||
Psychological | 4.21 | 2.54–6.97 | ||
Social | 2.41 | 1.60–3.63 | ||
Existential | 3.35 | 2.11–5.33 | ||
Actions taken | ||||
Prescription/test/referral | 1.03 | 0.79–1.36 | ||
Recommended sick leave | 1.11 | 0.48–2.53 | ||
Planned follow up | 2.34 | 1.50–3.65 |
Missing data on perceived benefit 81 (9%), somatisation 12 (1.3%) and remaining outcome variables 4 (0.4%). Self-reported benefit: None at all 326 (39.8%), a little 254 (31.0%), some 194 (23.7%), a lot 45 (5.5%), results were dichotomised between none at all and remaining categories.
Odds Ratio obtained by logistic regression, corrected for patients' age and gender, GPs' participation in educational programme, and clustering at GP level.
Outcome . | Odds Ratiob . | CI 95% . | ||
---|---|---|---|---|
Diagnoses | ||||
Somatisation | 2.71 | 1.62–4.55 | ||
Anxiety | 2.89 | 1.45–4.74 | ||
Depression | 2.88 | 1.81–4.56 | ||
Alcohol abuse | 4.30 | 1.55–11.91 | ||
Subjects of conversation | ||||
Physical | 0.66 | 0.30–1.47 | ||
Psychological | 4.21 | 2.54–6.97 | ||
Social | 2.41 | 1.60–3.63 | ||
Existential | 3.35 | 2.11–5.33 | ||
Actions taken | ||||
Prescription/test/referral | 1.03 | 0.79–1.36 | ||
Recommended sick leave | 1.11 | 0.48–2.53 | ||
Planned follow up | 2.34 | 1.50–3.65 |
Outcome . | Odds Ratiob . | CI 95% . | ||
---|---|---|---|---|
Diagnoses | ||||
Somatisation | 2.71 | 1.62–4.55 | ||
Anxiety | 2.89 | 1.45–4.74 | ||
Depression | 2.88 | 1.81–4.56 | ||
Alcohol abuse | 4.30 | 1.55–11.91 | ||
Subjects of conversation | ||||
Physical | 0.66 | 0.30–1.47 | ||
Psychological | 4.21 | 2.54–6.97 | ||
Social | 2.41 | 1.60–3.63 | ||
Existential | 3.35 | 2.11–5.33 | ||
Actions taken | ||||
Prescription/test/referral | 1.03 | 0.79–1.36 | ||
Recommended sick leave | 1.11 | 0.48–2.53 | ||
Planned follow up | 2.34 | 1.50–3.65 |
Missing data on perceived benefit 81 (9%), somatisation 12 (1.3%) and remaining outcome variables 4 (0.4%). Self-reported benefit: None at all 326 (39.8%), a little 254 (31.0%), some 194 (23.7%), a lot 45 (5.5%), results were dichotomised between none at all and remaining categories.
Odds Ratio obtained by logistic regression, corrected for patients' age and gender, GPs' participation in educational programme, and clustering at GP level.
The GPs perceived SQ information to be useful in high-score patients, new patients, female patients, and patients whose problems had lasted more than 14 days (Table 4).
Characteristics . | Odds ratiob . | CI 95% . |
---|---|---|
High score on screening questionnaire | 3.08 | 2.17–4.38 |
New patient | 2.17 | 1.16–4.06 |
Duration of actual problem more than 14 days | 1.57 | 1.05–2.33 |
Female patient | 1.49 | 1.05–2.13 |
Patient's age | 0.99 | 0.98–1.01 |
Characteristics . | Odds ratiob . | CI 95% . |
---|---|---|
High score on screening questionnaire | 3.08 | 2.17–4.38 |
New patient | 2.17 | 1.16–4.06 |
Duration of actual problem more than 14 days | 1.57 | 1.05–2.33 |
Female patient | 1.49 | 1.05–2.13 |
Patient's age | 0.99 | 0.98–1.01 |
Incomplete data on 85 (9%): perceived benefit 81 (9%), new patient 9 (1%), and duration of actual problem 11 (1%). Self-reported benefit: None at all 326 (39.8%), a little 254 (31.0%), some 194 (23.7%), a lot 45 (5.5%), results were dichotomised between none at all and remaining categories.
Odds ratio obtained by logistic regression, adjusted for interactions and clustering at GP level and randomisation of GP to educational programme on somatisation. No additional explanatory variables were evaluated.
Characteristics . | Odds ratiob . | CI 95% . |
---|---|---|
High score on screening questionnaire | 3.08 | 2.17–4.38 |
New patient | 2.17 | 1.16–4.06 |
Duration of actual problem more than 14 days | 1.57 | 1.05–2.33 |
Female patient | 1.49 | 1.05–2.13 |
Patient's age | 0.99 | 0.98–1.01 |
Characteristics . | Odds ratiob . | CI 95% . |
---|---|---|
High score on screening questionnaire | 3.08 | 2.17–4.38 |
New patient | 2.17 | 1.16–4.06 |
Duration of actual problem more than 14 days | 1.57 | 1.05–2.33 |
Female patient | 1.49 | 1.05–2.13 |
Patient's age | 0.99 | 0.98–1.01 |
Incomplete data on 85 (9%): perceived benefit 81 (9%), new patient 9 (1%), and duration of actual problem 11 (1%). Self-reported benefit: None at all 326 (39.8%), a little 254 (31.0%), some 194 (23.7%), a lot 45 (5.5%), results were dichotomised between none at all and remaining categories.
Odds ratio obtained by logistic regression, adjusted for interactions and clustering at GP level and randomisation of GP to educational programme on somatisation. No additional explanatory variables were evaluated.
Predictors of unrecognised high SQ scores were patients who 1) reported fair or poor self-rated health, 2) perceived their presenting problem as not simply physical, 3) presented with many physical symptoms, and 4) made the GPs feel uncertain, annoyed or insufficient (Table 5).
Characteristics . | Odds Ratioc . | CI 95% . |
---|---|---|
GP's recognition of patient as being ‘mentally distressed’b | 2.67 | 1.79–3.97 |
Patients' self-rated health fair or poor | 2.72 | 1.94–3.82 |
Patients' perceiving actual problem as not simply physical | 2.30 | 1.47–3.61 |
Patients presenting with many physical symptoms | 2.33 | 0.99–5.45 |
GP's feeling of uncertainty, annoyance, or insufficiency during encounter | 4.09 | 0.92–18.17 |
Characteristics . | Odds Ratioc . | CI 95% . |
---|---|---|
GP's recognition of patient as being ‘mentally distressed’b | 2.67 | 1.79–3.97 |
Patients' self-rated health fair or poor | 2.72 | 1.94–3.82 |
Patients' perceiving actual problem as not simply physical | 2.30 | 1.47–3.61 |
Patients presenting with many physical symptoms | 2.33 | 0.99–5.45 |
GP's feeling of uncertainty, annoyance, or insufficiency during encounter | 4.09 | 0.92–18.17 |
Incomplete data on 279 (32%): GPs' recognition 38 (4.3%), fair or poor self rated health 137 (15.5%), patient's perception of actual problem 149 (16.8%), patients presenting with many physical symptoms 10 (1.1%), GP's feeling of uncertainty, annoyance, or insufficiency 7 (0.8%).
Actual problem presented being due to ‘somatisation’, or patient having a significant psychiatric disorder.
Odds ratio obtained by logistic regression, adjusted for GPs' participation in educational programme, patients' age and gender, interactions between recognition and GP-variables, and clustering at GP level. Excluded from this model (using LR test) were GP ratings ‘the patient is being chronically ill’, ‘the main problem is hard to define’, and ‘the patient seems unreasonally worried’.
Characteristics . | Odds Ratioc . | CI 95% . |
---|---|---|
GP's recognition of patient as being ‘mentally distressed’b | 2.67 | 1.79–3.97 |
Patients' self-rated health fair or poor | 2.72 | 1.94–3.82 |
Patients' perceiving actual problem as not simply physical | 2.30 | 1.47–3.61 |
Patients presenting with many physical symptoms | 2.33 | 0.99–5.45 |
GP's feeling of uncertainty, annoyance, or insufficiency during encounter | 4.09 | 0.92–18.17 |
Characteristics . | Odds Ratioc . | CI 95% . |
---|---|---|
GP's recognition of patient as being ‘mentally distressed’b | 2.67 | 1.79–3.97 |
Patients' self-rated health fair or poor | 2.72 | 1.94–3.82 |
Patients' perceiving actual problem as not simply physical | 2.30 | 1.47–3.61 |
Patients presenting with many physical symptoms | 2.33 | 0.99–5.45 |
GP's feeling of uncertainty, annoyance, or insufficiency during encounter | 4.09 | 0.92–18.17 |
Incomplete data on 279 (32%): GPs' recognition 38 (4.3%), fair or poor self rated health 137 (15.5%), patient's perception of actual problem 149 (16.8%), patients presenting with many physical symptoms 10 (1.1%), GP's feeling of uncertainty, annoyance, or insufficiency 7 (0.8%).
Actual problem presented being due to ‘somatisation’, or patient having a significant psychiatric disorder.
Odds ratio obtained by logistic regression, adjusted for GPs' participation in educational programme, patients' age and gender, interactions between recognition and GP-variables, and clustering at GP level. Excluded from this model (using LR test) were GP ratings ‘the patient is being chronically ill’, ‘the main problem is hard to define’, and ‘the patient seems unreasonally worried’.
Supplementary analyses
The educational programme on somatisation had no statistically significant influence on the main outcome measures. However, GPs randomised to the educational programme did report increased benefit from the screening information (OR 2.66, 95% CI 1.25 to 5.68).
Discussion
Results from this trial suggest that the GPs' recognition and management of common mental disorders can be influenced by means of a brief SQ. GPs with low or moderate recognition rates seem to benefit most. We found that GPs' self-reported benefit from the SQ-information may serve as a simple and useful predictor of diagnosis and care. We identified some simple predictors of benefit from psychiatric screening that may prove useful as case-finding criteria in routine clinical practice.
The results suggest unbiased randomisation and successful blinding, but the correctness of the inclusion procedure depended on the GPs' secretaries. Our analyses were based on random allocation and not ‘intention to diagnose/treat’ as 138 patients were registered as ‘not participating for other reasons’. The screening benefit may have been reduced due to: 1) the lack of a ‘run-in period’ of the screening procedure; 2) a potential Hawthorne bias due to focus on recognition; 3) cross-contamination of the GPs' own diagnostic approaches; 4) sensitisation of participating patients regardless of randomisation status; 5) better communication skills and psychotherapeutic training status of the participating GPs compared to non-participating GPs; 6) introduction of an educational programme mainly focusing on somatisation; 7) the slightly younger age of participating than non-participating patients, given the observed increase in psychiatric morbidity with age.24 Unfortunately, we did not ask GPs to specify their choice when asked about actions taken in terms of prescriptions, tests, and referrals. However, one would expect major differences to have emerged during the follow-up consultations.
To our knowledge, this is the first RCT examining the impact of routine screening for common mental disorders on recognition and care. Results from non-randomised trials on screening for various mental disorders suggest improved recognition, but only limited effect on rates of new management actions.25,26 Our findings suggest that screening for typically comorbid mental disorders (especially depression and alcohol abuse) may increase recognition rates and actions taken in primary care. Disclosure of SQs revealed no significant impact on GP recognition of somatisation and anxiety whereas their self-reported benefit from the SQ information was found to predict both diagnoses. This inconsistency calls for further validation of the screening instrument. Non-recognition at first consultation may simply reflect the GPs' reluctance to make unfamiliar diagnoses. The finding of an increased rate of planned follow-ups when SQs were disclosed supports this assumption. Increases in rates of planned follow-ups may be of particular importance as patients are more likely to be diagnosed and treated during follow-ups.27 Additional analyses revealed that screening disclosure was effective among GPs with moderate and low recognition rates whereas disclosure showed no impact on GPs with high recognition rates. This finding clearly demonstrates the ceiling effect on recognition when highly skilled GPs are recruited.
Though screening may improve recognition and care, the impact on patient health outcomes still remains to be assessed. Future RCTs on mental health screening in primary care should more actively address problems with recruitment of GPs, selection of patients, and problems with biased increases in baseline detection. Targeting interventions at GPs with low and moderate recognition rates, making use of pragmatic case-finding criteria in the selection of patients, and randomisation at practice level28 may prove more effective.
The results from this trial suggest that screening for typically comorbid mental disorders may influence GPs' recognition and provision of care. GPs' self-reported benefit from the screening information may, at least theoretically, serve as a useful predictor of improvements at patient level. Our findings support the need for a feasible and effective case-finding approach in order to make GPs benefit from mental health screening in routine clinical practice. Case-finding among new patients, patients who perceive their presenting problem as not simply physical, and patients with fair or poor self-rated health may be especially beneficial.
We hope that our results will encourage others to carry out randomised screening trials for common mental disorders, adopting the outlined case-finding approach. We suggest that the first health care professional meeting the patient should identify candidates for completing the SQ by routinely asking simple questions like; “what is the reason for your visit?” and “how are you feeling in general?”
Declaration
Funding: the project was financed by the Interdisciplinary Research Programme of the Danish National Research Council: “Sundhedsfremme og forebyggelsesforskning” (grant number 9801278). The GPs' training participation, data collection and use of SQs was financed by The Regional Health Assurance in Aarhus County through a local pay agreement (project number 0871).
Ethical approval: this study was approved by The Scientific Ethics Committee in the County of Aarhus, the Danish Data Protection Agency and the Scientific Research Evaluation Committee of the Danish College of General Practitioners.
Conflicts of interest: none.
References
Fink P, Sorensen L, Engberg M, Holm M, Munk-Jørgensen P. Somatization in primary care. Prevalence, health care utilization, and general practitioner recognition.
Ustun TB, Sartorius N. Mental illness in General Health Care, An International Study. Chichester: John Wiley & Sons;
Ford DE, Kamerow DB. Screening for psychiatric and substance abuse disorders in clinical practice.
Screening and behavioral counseling interventions in primary care to reduce alcohol misuse: recommendation statement.
Pignone MP, Gaynes BN, Rushton JL, Burchell CM, Orleans CT, Mulrow CD et al. Screening for depression in adults: a summary of the evidence for the US Preventive Services Task Force.
Hickie IB, Davenport TA, Ricci CS. Screening for depression in general practice and related medical settings.
Christensen KS, Toft T, Frostholm L, Ornbol E, Fink P, Olesen F. The FIP study: a randomised, controlled trial of screening and recognition of psychiatric disorders.
Gilbody SM, OHouse A, Sheldon TA. Routinely administered questionnaires for depression and anxiety: systematic review.
Hickie IB, Davenport TA, Scott EM, Hadzi-Pavlovic D, Naismith SL, Koschera A. Unmet need for recognition of common mental disorders in Australian general practice.
Leon AC, Olfson M, Broadhead WE, Barrett JE, Blacklow RS, Keller MB et al. Prevalence of mental disorders in primary care. Implications for screening.
Kroenke K, Spitzer RL, Williams JB, Linzer M, Hahn SR, deGruy FV et al. Physical symptoms in primary care. Predictors of psychiatric disorders and functional impairment.
Beich A, Gannik D, Malterud K. Screening and brief intervention for excessive alcohol use: qualitative interview study of the experiences of general practitioners.
Pignone MP, Gaynes BN, Rushton JL, Burchell CM, Orleans CT, Mulrow CD et al. Screening for depression in adults: a summary of the evidence for the US Preventive Services Task Force.
Fink P, Rosendal M, Toft T. Assessment and Treatment of Functional Disorders in General Practice: The Extended Reattribution and Management Model—An Advanced Educational Program for Nonpsychiatric Doctors.
Derogatis LR, Cleary PA. Confirmation of the dimensional structure of the SCL-90: A study in construct validation.
Fink P, Ewald H, Jensen J, Sorensen L, Engberg M, Holm M et al. Screening for somatization and hypochondriasis in primary care and neurological in-patients: a seven-item scale for hypochondriasis and somatization.
Fink P, Oernboel E, Hansen MS, Sondergaard L, De Jonge P. Detecting mental disorders in general hospitals by the SCL-8 scale.
Fink P, Jensen J, Borgquist L, Brevik JI, Dalgard OS, Sandager I et al. Psychiatric morbidity in primary public health care. A Nordic multicenter investigation. Part I: Method and prevalence of psychiatric morbidity.
Ware Jr JE, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection.
The Functional Illness in Primary Care Study. www.auh.dk/cl_psych/dk/forsk.htm.
Fink P, Sorensen L, Engberg M, Holm M, Munk-Jørgensen P. Somatization in primary care. Prevalence, health care utilization, and general practitioner recognition.
Toft T, Christensen KS, Frostholm L, Oernboel E, Fink P, Olesen F. Prevalence and Comorbidity among Mental Disorders in Primary Care. Results from the FIP study. Psychol Med
Spitzer RL, Williams JB, Kroenke K, Linzer M, deGruy FV, Hahn SR et al. Utility of a new procedure for diagnosing mental disorders in primary care. The PRIME-MD 1000 study.
Spitzer RL, Kroenke K, Williams JB. Validation and utility of a self-report version of PRIME-MD: the PHQ primary care study. Primary Care Evaluation of Mental Disorders. Patient Health Questionnaire.
Kessler D, Bennewith O, Lewis G, Sharp D. Detection of depression and anxiety in primary care: follow up study.
Author notes
aThe Research Unit for Functional Disorders, Aarhus University Hospital and bThe Research Unit for General Practice, University of Aarhus, Denmark