• Aucun résultat trouvé

Training General Practitioners to Detect Probable Mental Disorders in Young People During Health Risk Screening

N/A
N/A
Protected

Academic year: 2022

Partager "Training General Practitioners to Detect Probable Mental Disorders in Young People During Health Risk Screening"

Copied!
9
0
0

Texte intégral

(1)

Article

Reference

Training General Practitioners to Detect Probable Mental Disorders in Young People During Health Risk Screening

AMBRESIN, Anne-Emmanuelle, et al.

Abstract

The purpose of the study is to investigate whether a training intervention increases general practitioners' (GPs) detection sensitivity for probable mental disorders in young people.

AMBRESIN, Anne-Emmanuelle, et al . Training General Practitioners to Detect Probable Mental Disorders in Young People During Health Risk Screening. Journal of Adolescent Health , 2017, vol. 61, no. 3, p. 302-309

PMID : 28596103

DOI : 10.1016/j.jadohealth.2017.03.015

Available at:

http://archive-ouverte.unige.ch/unige:95819

Disclaimer: layout of this document may differ from the published version.

1 / 1

(2)

Original article

Training General Practitioners to Detect Probable Mental Disorders in Young People During Health Risk Screening

Anne-Emmanuelle Ambresin, M.D.

a,b,1

, Christiaan P. Otjes, M.D.

a,*,1

,

George C. Patton, M.B.B.S., M.R.C.Psych., M.D.

c,d,e

, Susan M. Sawyer, M.B.B.S., M.D.

c,d,e

, Sharmala Thuraisingam, M.Biostat.

a

, Dallas R. English, M.S., Ph.D.

f

,

Dagmar M. Haller, M.D., Ph.D.

a,f,g

, and Lena A. Sanci, M.B.B.S., Ph.D.

a

aDepartment of General Practice, University of Melbourne, Melbourne, Australia

bMultidisciplinary Unit for Adolescent Health, Lausanne University Hospital, Lausanne, Switzerland

cCentre for Adolescent Health, Royal Children’s Hospital, Melbourne, Australia

dDepartment of Paediatrics, University of Melbourne, Melbourne, Australia

eMurdoch Children’s Research Institute, Melbourne, Australia

fSchool of Population Health, University of Melbourne, Melbourne, Australia

gPrimary Care Unit, Department of Community Health and Medicine, Faculty of Medicine, University of Geneva, Geneva, Switzerland

Article history:Received November 3, 2016; Accepted March 23, 2017

Keywords: Adolescents; Young people; Mental disorder; Perceiving mental illness; Psychosocial health risk screening; Universal intervention; General practitioner; Primary care

A B S T R A C T

Purpose: The purpose of the study is to investigate whether a training intervention increases general practitioners’(GPs) detection sensitivity for probable mental disorders in young people.

Methods:Forty general practices were randomized to an intervention (29 GPs) or comparison arm (49 GPs). Intervention GPs participated in 9 hours of interactive training on youth-friendly care, psy- chosocial health risk screening, and responding to risk-taking behavior with motivational interviewing approaches, followed by practice visits assisting with integration of screening processes and tools.

Youth aged 14e24 years attending GPs underwent a computer-assisted telephone interview about their consultation and psychosocial health risks. Having a“probable mental disorder”was defined as either scoring high on Kessler’s scale of psychological distress (K10) or self-perceived mental illness.

Other definitions tested were high K10; self-perceived mental illness; and high K10 and self-perceived mental illness. Psychosocial health risk screening rates, detection sensitivity, and other accuracy pa- rameters (specificity, positive predictive value, and negative predictive value) were estimated.

Results: GPs’ detection sensitivity improved after the intervention if having probable mental disorder was defined as high K10 score and self-perceived mental illness (odds ratio: 2.81; 95%

confidence interval: 1.23e6.42). There was no significant difference in sensitivity of GPs’detection for our preferred definition, high K10 or self-perceived mental illness (.37 in both; odds ratio: .93;

95% confidence interval: .47e1.83), and detection accuracy was comparable (specificity: .84 vs. .87, positive predictive values: .54 vs. .60, and negative predictive values: .72 vs. .72).

Conclusions:Improving recognition of mental disorder among young people attending primary care is likely to require a multifaceted approach targeting young people and GPs.

Ó2017 Society for Adolescent Health and Medicine. All rights reserved.

IMPLICATIONS AND CONTRIBUTION

This study demonstrates the effectiveness of a complex intervention in improving general practi- tioners’ detection of probable mental disorders in psychologically dis- tressed young people self- perceiving a mental illness.

Thesefindings suggest that detection of mental disor- ders in youth could be improved utilizing a multi- faceted intervention tar- geting both youth and general practitioners.

Conflicts of Interest:The authors have no conflicts of interest to disclose.

Trial registration: ISRCTN.com ISRCTN16059206.

*Address correspondence to: Christiaan P. Otjes, M.D., Department of General Practice, University of Melbourne, 200 Berkeley Street, Carlton 3053 VIC, Australia.

E-mail address:chrisotjes@gmail.com(C.P. Otjes).

1Both authors contributed equally to this work.

www.jahonline.org

1054-139X/Ó2017 Society for Adolescent Health and Medicine. All rights reserved.

http://dx.doi.org/10.1016/j.jadohealth.2017.03.015

(3)

Over the past two decades there has been growing emphasis on effective identification of psychosocial risks in young people attending primary care [1] partly due to shifts in causes of adolescent morbidity and mortality from infectious and somatic diseases to psychosocial and lifestyle problems [2]. The high prevalence (30%e40%) of emotional distress[3]is of particular concern because of its effects on education, relationships, and quality of life and its association with other health- compromising behaviors such as smoking and increased risk of suicide[4,5].

Globally, primary care clinicians provide most health care for common mental disorders[6]. In Australia, general practitioners (GPs) are the gateway to specialist mental health services. Yet, only 20%e60% of young people with mental disorder are iden- tified in primary care[3,7e9], with higher rates of underdiag- nosis and undertreatment of depression compared with adults [10], which contributes to significant unmet need[11]. Psycho- social assessment is a recommendation within policy and clinical practice guidelines that could be important in reducing the treatment gap were it regularly practiced[12].

Some studies have examined whether training, targeted on the identification of mental disorders, enhances GPs’capacity to identify mental disorders in youth[7,8,13]. A feasibility study of brief training for GPs focusing on systematically screening for mental health issues and intervening if depression was identi- fied, resulted in increased screening, and improved depression identification[13]. Training GPs in clinical skills with youth has also improved knowledge, skills, and self-perceived competency in working with young people[14].

The present article derives from a larger study [15] which aimed to assess the impact of a three-part intervention for GPs, including screening for a broad range of psychosocial health risks and counseling for identified risky behaviors, on young people’s engagement in risky behaviors (tobacco, alcohol and illicit drug use, road and driving risks, and sexual health risk taking).

This article examines whether this intervention also improved GPs’sensitivity in detecting young people’s probable mental disorder. We hypothesized that increased screening and discussion of a broad range of psychosocial health risks would lead to increased identification of mental disorders (increased detection sensitivity). Secondary aims, including screening rates for various psychosocial health risks, and other psychometric properties of detection accuracy (specificity, positive predictive values [PPVs], and negative predictive values [NPVs]), were also computed. This is the first study investigating whether GPs improve in their detection sensitivity of probable mental disor- ders after a training intervention aiming to improve discussion of psychosocial health risks in general between GPs and young people.

Methods

Ethics approval was obtained from the University of Melbourne Human Research Ethics Committee.

Study design

Data derived from a cluster randomized controlled trial (2007e2011), informed by CONSORT (Consolidated standards of reporting trials) guidelines[16], of screening young people for psychosocial health risks and responding to risky behaviors with motivational interviewing in the general practice setting (the

prevention access and risk taking in young people trial[15]). The detailed protocol and main results about impact on clinicians’ screening for risky behavior at baseline postintervention (T0), and young people’s engagement in risky behaviors 3 and 12 months postconsultation have been published elsewhere[15]. As the intervention affected practice systems, a cluster design was chosen where the practice was the unit of randomization. Prac- tices were stratified by postcode level advantage-disadvantage (Socio-Economic Indexes for Areas[17]) and practice type (pri- vate billing, bulk billing, and community health centers), form- ing six strata. Block randomization withfixed block sizes of two was used within strata. The allocation sequence was generated in Stata[17]by an independent statistician and remained concealed from researchers until completion of the 12-month follow-up.

The computer-assisted telephone interview (CATI) and the in- practice recruiting research assistants (RAs) were blind to the study arm status of the practices, and young people were not informed of the practice arm in any researcher communication.

Because of the nature of the intervention, clinicians (GPs and practice nurses) were not blind to study status. Data for this study derive from measures administered to youth (exit interview) and GPs (encounter form) in both trial arms at T0, after consultation, which was also after intervention.

Participants

General practices were recruited through a variety of methods including advertisements in General Practice Divisions and the Royal Australian College of General Practitioners’newsletters, direct mail outs, the Medicare Australia database, and the Victorian practice-based Research Network, encompassing urban and regional centers. All youth aged 14e24 years attending participating clinicians were eligible for inclusion unless clinicians judged them as too physically or mentally unwell to participate (e.g., vomiting, febrile, weak, cognitively impaired, or psychotic) or unable to give informed consent. Minors aged 14e17 years were eligible without parental consent if judged a mature minor by clinicians.

Intervention

In brief, the intervention was delivered at the practice level and included three components: (1) clinician training in youth- friendly care (6 hours) and motivational interviewing ap- proaches for management of risk-taking behaviors (3 hours)[18], (2) provision of a nonstandardized screening tool to assist in assessment and discussion of psychosocial health risks, and (3) two practice visits to feedback to clinicians the psychosocial health risks their young patients were experiencing; assist with developing office procedures to implement screening; train reception staff in youth-friendly care; and update practice specialist referral lists with youth-specific alternatives. Manage- ment of mental disorders was not a specific focus of the training but written resources for referral and management of common disorders accompanied the workshops. The psychosocial health risk screening tool was developed from the HEADSS mnemonic (Home; Education, eating, exercise; Activities and peers; Drugs, cigarettes and alcohol; Suicide, depression and other psychiatric symptoms; and Safety)[19], afterfinding that a similar HEADSS- based tool stimulated psychosocial health risk discussions by hospital-based clinicians [20]. The comparison arm clinicians received a 3-hour didactic seminar on best practice in adolescent health care, including psychosocial assessments.

A.-E. Ambresin et al. / Journal of Adolescent Health xxx (2017) 1e8 2

(4)

Procedure

Young people arriving at the practice received a brochure outlining the major study steps and aims, including exploring how to help or prevent young people engaging in health risketaking behavior. Clinicians raised the prospect of research participation with eligible young people post- consultation, asking for permission to pass their contact details to CATI staff. Subsequently, to improve recruitment rates, in-practice RAs were employed to undertake this process. CATI staff called young people to fully explain the study, obtain con- sent, and conduct the exit interview. Young people were not informed in any researcher communication about which study arm their practice had been allocated to. Clinicians completed an encounter form postconsultation. The completed exit interview and encounter form were paired for the analysis.

For the current study, only the matched and completed GP encounter form-youth exit interview pairs were analyzed. This excluded the minority seen only by a practice nurse. At the time of this study, practice nurses were beginning to enter general practice (in around 40% of Australian practices) and had a limited scope of practice, including minor and preventive health issues, such as immunizations, pap smears, and wound dressings [21,22]; receptionists directly booked patients presenting for these issues with a practice nurse.

Measures

Psychosocial health risk screening. To assess GP screening behavior, the exit interview asked youth whether issues (emotions/feelings, tobacco use, alcohol use, illicit drug use, sexual health, road and driving safety, and fear or abuse in relationships) were raised or discussed during the consultation. If“raised”or“discussed,”the health risk was considered to be screened by the GP.

Mental health rating. Both the GPs’ encounter form and the young people’s exit interview had a similar question on mental health rating asking, “How do you rate your/your patient’s emotional or mental health?” [23]. There werefive response categories: (1) no mental illness, (2) mild symptoms, no illness, (3) minor illness, (4) moderate illness, and (5) severe mental illness. Categories 1 and 2 were collapsed into“no”and 3 to 5 were collapsed into “yes,” creating a binary variable for GPs’ assessment of the presence of a mental illness (yes/no) and for young people’s self-perception that they had a mental illness (yes/no). The collapsed categories enabled assessment of whether GPs were effective in distinguishing between mental illness and no mental illness.

K10 score. The Kessler’s scale of psychological distress (K10) is a 10-item scale designed to measure psychological distress, mostly anxiety or depressive states[24]. The K10 was administered to youth in the exit interview. The 10 items sum to give an overall score ranging from 10 to 50 points. We used stratum-specific likelihood ratios and the Bayesian method[25]to transform K10 scores into a binary variable (K10 score 20 ¼ “positive”), differentiating those with levels of psychological distress indica- tive of a high probability of having a mental disorder from the rest.

Definition of probable mental disorder. Probable mental disorder was defined when young people had either: (1) a positive K10 score or (2) self-perceived mental disorder. The term“probable”

indicates the K10’s status as a screening instrument indicating a probability of mental disorder, rather than a diagnostic tool. We justified the young people’s self-perception of having a mental disorder as a“case,”based on the perspective that self-perceived mental disorders, with or without a high K10 score, should be addressed by clinicians. Moreover, previous work identifies youth beliefs as a key factor driving clinician recognition and management of mental health issues[3]. Three other possible definitions of“probable mental disorder”were also explored to analyze the extent the chosen definition altered detection sensitivity. These alternative definitions are K10 positive; self- perceived mental illness positive; and K10 positive plus self- perceived mental illness positive.

Detection of probable mental disorder.GP detection of probable mental disorder was assumed if the GP rated the youth as having a mental illness on the encounter form, as described previously.

Detection sensitivity was defined as the proportion of youth with a probable mental disorder being rated correctly by GPs. In this article,“detection sensitivity” and“sensitivity”are used inter- changeably. Accuracy of detection refers to the psychometric properties of accuracy, which are specificity, PPV, and NPV in addition to sensitivity[26].

Statistical analyses. Young person, GP, and practice character- istics were described using frequencies for categorical mea- sures and means (standard deviation) for continuous measures.

Detection sensitivity was calculated in both arms. Odds ratios for the intervention effect on GPs’ detection sensitivity were estimated using logistic regression with generalized estimating equations and robust standard errors to account for correlated data within the practices (clusters). An odds ratio above one indicates a positive effect of the intervention on GP detection sensitivity compared with comparison GPs. The analyses were adjusted for trial stratifications (practice area socioeconomic index and billing type), recruitment method (GP or RA), and clinically relevant factors (youth age, sex, country of birth, GPs’ prior training in mental health, parent presence at the consultation, andfirst visit to the GP). Proportions of youth screened for psychosocial health risks were identified. Other psychometric values of detection accuracy (specificity, PPV, and NPV) were calculated for all four probable mental disorder definitions. Individuals were analyzed according to their arm allocation. All analyses were performed using Stata 13.1 soft- ware (Statacorp LP, College Station, TX)[17].

Results

Overall, 40 of the 42 randomized practices (95.2%) completed the study, including 78 GPs. Two intervention practices declined participation before intervention because of time constraints. In total, 1,572 young people agreed to talk with the research team.

Excluding those unable to be contacted by the researchers (N¼390), and who declined (N¼281), left a study group of 901 (57.3%) in the main study. For this analysis, youth were also excluded if their exit interview had no matching encounter form (N ¼ 56) or if they consulted only with the practice nurse (N¼95). The latter differed little from the rest of the sample except that they were more likely to be older (18e24 years), female, and regular patients of the clinic. A total of 750 matched GP-young people pairs were included in the analysis with one to 42 youths per practice and one to 31 per GP (Figure 1).

(5)

Practices randomized

randomization

Figure 1. Trial Profile.*Young people were ineligible if they were: unable to speak English; physically or mentally unwell; or, under 18 years old, judged by clinician to be incompetent to make an informed decision for participation in minimal risk research and unable or unwilling to obtain parental consent.yNever had contact (wrong phone number, too many attempted phone calls, or responded more than 3 weeks after consultation and were therefore too delayed to participate). Adapted from PARTY paper by Sanci et al[15].

A.-E. Ambresin et al. / Journal of Adolescent Health xxx (2017) 1e8 4

(6)

Baseline characteristics

The baseline characteristics of practices, GPs, and young people (Table 1) reveal smaller proportions of intervention youth born in Australia and consulting with parents present. There were fewer female doctors, doctors with prior training in adolescent health, and younger doctors in the intervention arm, while recruitment using an RA was more frequent.

Detection sensitivity

GPs’detection sensitivity for the four definitions of probable mental disorder is presented inTable 2. In our study sample, 256 (34.2%) youth scored positively on either K10 or self-rating;

214 (28.6%) scored a positive K10; 141 (18.9%) self-perceived a mental illness; and 99 (13.2%) scored positive on both K10 and self-rating.

Intervention GPs showed better detection sensitivity than comparison GPs when both the K10 and young people’s self- perceived mental illness were positive, with odds ratio of 2.81 (95% confidence interval: 1.23e6.42). There was no difference between arms in GPs’ detection sensitivity for our preferred definition of probable mental disorder (positive K10 or self- perceived mental illness).

Screening for psychosocial health risks

Intervention GPs screened for psychosocial health risks more frequently than comparison GPs, particularly for tobacco, alcohol and illicit drug use, road and driving safety, and fear or abuse in relationships (Table 3). Mental health was screened equally frequently by intervention (33%) and comparison GPs (30%).

Accuracy of detection

GPs’detection accuracy for each definition of probable mental disorder is presented inTable 4. There were no major differences in the parameters of accuracy between intervention and com- parison arms. Sensitivity and NPVs were best for the definition of a case being K10 positive and self-perceived mental illness pos- itive. Regarding sensitivity, 61% and 52% of cases were correctly identified by intervention and comparison GPs, respectively, while the NPV was 93% and 92%, respectively, of those who tested negatively for not having a probable mental disorder.

Specificity was slightly lower in the intervention compared with the comparison arm across all four case definitions. PPV was best for the proposed case definition, and K10 positive with around 50% in both arms who tested positive actually being defined as a case.

Discussion

Using our proposed definition of probable mental disorder (K10 positive or self-rated positive), GPs participating in a multifaceted intervention designed to increase clinician Table 1

Baseline characteristics for young people (YP), general practitioner (GP), and cluster (practices) level

Intervention Comparison

YP’s characteristics n¼300 (%) n¼450 (%)

Age (years) 19.63 (3) 19.35 (3)

14e15 years 36 (12) 58 (13)

16e17 years 31 (10) 70 (16)

18e24 years 233 (78) 322 (72)

Gender (female) 222 (74) 336 (75)

Birth country (Australia) 228 (76) 404 (90)

Employment/educationa

Studying only 83 (28) 116 (26)

Working only 70 (23) 118 (26)

Studying and working 125 (42) 182 (40)

Neither studying nor working 21 (7) 32 (7)

First visit to this GP (yes)a 72 (24) 128 (28) Recruitment methods (recruited by GPs) 198 (66) 368 (82) YP accompanied by parents (yes)a 65 (22) 145 (32)

GP’s characteristics n¼29 (%) n¼49 (%)

Gender (female) 13 (45) 28 (57)

Country of graduation (Australia)a 21 (72) 35 (71) Prior training in adolescent health (yes)a 16 (55) 32 (65) Age group (years)a

25e44 9 (31) 28 (57)

45e54 14 (48) 14 (29)

55 5 (17) 7 (14)

Practice characteristics n¼19 (%) n¼21 (%)

Practice billing

Bulk billing 5 (26) 7 (33)

Private billing 11 (58) 13 (62)

Community health center 3 (16) 1 (5)

SES of the practice location

High 15 (79) 17 (81)

Low 4 (21) 4 (19)

Practice location (ARIA)

Metropolitan 17 (90) 15 (71)

Regionalb 2 (10) 6 (29)

ARIA ¼ Accessibility/Remoteness Index of Australia [27]; GP ¼ general practitioner; SEIFA¼Socio-Economic Indexes for Areas[17].

aDiscrepancies in total due to missing responses.

bRegional: inter and outer regional.

Table 2

A comparison of intervention and comparison general practitioner’s (GP) detection sensitivity for probable mental health disorder in young people (YP) attending primary care for each case definition of a probable mental health disorder

Definitions Intervention Comparison Unadjusted Adjusteda

N n (%) N n (%) ORb(95% CI) pvalue ORb(95% CI) pvalue

Either K10 or self-perceived 102 38 (37) 154 57 (37) .98 (.56e1.72) .94 .93 (.47e1.83) .83

K10 positive 91 35 (38) 123 49 (40) .93 (.50e1.71) .80 .82 (.40e1.68) .59

Self-perceived positive 49 26 (53) 92 40 (43) 1.46 (.76e2.80) .25 1.58 (.75e3.32) .23

Both K10 and self-perceived 38 23 (61) 61 32 (52) 1.40 (.64e3.05) .40 2.81 (1.23e6.42) .01

Positive OR indicates the effect of the intervention.

CI¼confidence interval; OR¼odds ratio.

aAdjusted for YP’s age, YP’s sex, YP’s country of birth, YP’sfirst visit to GP, YP recruitment method, YP accompanied by parent, GP prior training in adolescent health, practice billing type, and socioeconomic status of the practice location.

bOR calculated using generalized estimation equations with robust standard errors to adjust for clustering at the clinic level.

(7)

screening for a broad range of psychosocial health risks did not increase their detection sensitivity for probable mental disorder.

In both arms, GPs failed to detect a probable mental disorder in two out of three young people (sensitivity 37%). However, they correctly ruled out a probable mental disorder in four out offive (specificity 84% and 87% in intervention and comparison arms), suggesting that they were more accurate at excluding than recognizing probable mental disorder. Sensitivity of GP probable mental disorder assessments is clearly lower than the results of standardized screening tests (Patient Health Questionnaire, Beck Depression Inventory) [28]. The background prevalence of mental disorders suggests that there may be an added value of using standardized psychosocial screening instruments in pri- mary care for all young people.

We observed a trend toward increased detection sensitivity in the intervention arm compared with the comparison arm when young people perceived that they had a mental disorder. More- over, detection sensitivity significantly increased in the inter- vention arm for the K10 and self-perceived mental illness positive group, without loss of specificity. Although thesefind- ings need to be treated cautiously, since testing multiple defi- nitions increases the probability of a significant result by chance, it is conceivable that because intervention GPs discussed a range of health issues more than comparison GPs, they were more likely to detect self-perceived concerns. Prior work shows GPs are better at detecting mental disorders in young people who perceive that they have a mental disorder [3]. However, dis- crepancies have been identified in the concordance of young people’s and GPs’assessment of having a mental disorder[29].

It has been suggested that strategies to influence young people’s recognition of mental health issues might be beneficial alongside strategies to lift GPs’discussion of psychosocial issues[3]; our findings support that these two factors combined lift detection sensitivity.

Failure to observe a difference between detection in the intervention and the comparison arms for our preferred case definition may be due to a true absence of the effect or may reflect methodological issues. Other studies demonstrate that training a small highly motivated group of GPs in mental health improved their detection of depression[7,13]. However, these studies solely focused on mental disorders rather than the breadth of issues in our study and used a“before-after”design without a comparison arm which is known to overestimate the effect size[30].

The intervention may have been insufficiently intensive to influence GPs’behaviors; prevention access and risk taking in young people was based on a previously effective 15-hour GP intervention delivered over 6 weeks [14]. In today’s time- pressured climate, GPs find difficulty committing to long periods of training, so our intervention was less time intensive.

Furthermore, the training received in this intervention was not specifically focused on mental disorders but rather on health risk behaviors, notwithstanding provision of written resources about management and referral for specialized mental health treat- ments. The lack of focus on mental health screening and man- agement arose because management skills have been the focus of new nationally funded schemes to better train and remunerate general practice management of mental disorders[31], whereas Table 3

Psychosocial health risks screened by the general practitioner (GP) in their consultation with the young person (YP)

Psychosocial risks Intervention, N¼300 Comparison, N¼450 Unadjusted Adjusteda

n (%) n (%) ORb(95% CI) pvalue ORb95% CI pvalue

Tobacco use 94 (31.5) 106 (23.7) 1.48 (.95e2.29) .08 1.62 (1.05e2.49) .03

Alcohol use 91 (30.5) 81 (18.2) 1.98 (1.23e3.20) .005 2.10 (1.30e3.39) .002

Illicit drug use 41 (13.8) 37 (8.3) 1.76 (1.05e2.95) .03 1.77 (1.07e2.94) .03

Sexual healthc 116 (38.8) 182 (40.6) .92 (.58e1.47) .03 .89 (.57e1.40) .62

Road and driving safetyd 25 (8.4) 5 (1.1) e e e e

Emotional distress 97 (32.8) 134 (29.9) 1.11 (.70e1.78) .65 1.18 (.74e1.88) .49

Fear or abuse in relationshipsd,e 13 (5.2) 4 (1.1) e e e e

Totals vary due to missing responses.

CI¼confidence interval; OR¼odds ratio.

aAdjusted for YP’s age, YP’s sex, socioeconomic status of the practice location, practice billing type, and recruitment method of young people.

bOR calculated using generalized estimation equations with robust standard errors to adjust for clustering at the clinic level.

cDiscussed at least one of the two sexual health risks (contraception and/or protection from STIs).

dModels notfitted due to insufficient frequencies for reliable estimates.

eFor young people aged 17 years or above (N¼251 in intervention arm and N¼352 in comparison arm).

Table 4

GPs’detection accuracy for probable mental health disorder in young people attending primary care by trial arm and by case definition for probable mental health disorder

Either K10 or self-perceived K10 positive only Self-perceived positive only Both K10 and self-perceived Intervention,

n¼300 (95% CI)

Comparisona, n¼448 (95% CI)

Intervention, n¼300 (95% CI)

Comparisona, n¼448 (95% CI)

Intervention, n¼300 (95% CI)

Comparisona, n¼448 (95% CI)

Intervention, n¼300 (95% CI)

Comparisona, n¼448 (95 %CI) Sensitivity .37 (.27e.49) .37 (.28e.47) .38 (.26e.52) .40 (.30e.51) .53 (.40e.65) .43 (.32e.56) .61 (.45e.74) .52 (.37e.67) Specificity .84 (.75e.90) .87 (.81e.91) .83 (.76e.89) .86 (.80e.90) .82 (.75e.88) .84 (.79e.88) .82 (.75e.87) .84 (.78e.88) PPV .53 (.40e.65) .43 (.32e.56) .50 (.42e.58) .52 (.39e.64) .37 (.27e.48) .42 (.35e.49) .33 (.24e.43) .34 (.27e.42) NPV .61 (.45e.74) .52 (.37e.67) .76 (.69e.81) .79 (.72e.84) .90 (.84e.94) .85 (.80e.89) .93 (.89e.96) .92 (.88e.94) CI¼confidence interval; NPV¼negative predictive value; PPV¼positive predictive value.

aTwo K10 and young persons’self-perceived ratings missing (1% of total 450 observations) in comparison arm.

A.-E. Ambresin et al. / Journal of Adolescent Health xxx (2017) 1e8 6

(8)

no such incentives support broader psychosocial health risk screening. A different form of GP training focusing on mental disorder detection and management, including the use of stan- dardized screening tools, may be superior in improving detection of probable mental disorders.

Comparison GPs were offered standard talks on adolescent health, including the role of psychosocial health risk screening.

This potentially acted as a“refresher,”given that over two thirds of the GPs had prior training in adolescent health, and thereby contributed to the lack of impact of the intervention. While the ideal comparison arm would have received no training, brief upskilling in adolescent health was offered to enhance recruit- ment and retention of GPs in the study.

There are also patient factors blocking detection of mental disorders. Rates of depression underdiagnosis and undertreat- ment are higher in adolescents than in adults, mainly because of the different clinical presentation with more fluctuating and unexplained physical symptoms[10]. Moreover, depressed youth reluctantly disclose mental health issues due to health literacy factors including beliefs about mental health, stigma, low awareness of mental disorders and the help GPs may provide, and negative views of pharmacological interventions[32,33]. In this study, potential clinically relevant factors were identified before the analyses so they could be accounted for; however, it is possible that other important factors were not measured.

Detection accuracy may have been compromised because our definition of probable mental disorder was not based on a diagnostic tool; the K10 is a screening instrument indicative of a high probability of having a mental disorder, and self-perceived mental illness is a subjective measure. It is likely that our defi- nition was more inclusive compared with receiving a mental health diagnosis, which could have contributed to lower sensi- tivity of detection. Screening rates were estimated based on youth self-report, since we were unable to otherwise measure disclosure of psychosocial health risks because of screening by the GP. Limited implementation of the screening intervention is likely to have contributed to low detection sensitivity. As this was an effectiveness rather than efficacy trial, young people were analyzed according to the treatment arm they were allocated to, rather than to whether they actually received the screening or not. This highlights barriers to the implementation of in- terventions as intended in “real-life” practice, which warrants further exploration.

The currentfindings sit within the complex debate around mental health screening in primary care and difficulties defining a mental disorder case. There is little evidence that screening improves mental health outcomes, and concerns exist about stigma and increased strain on health care resources potentially arising from incorrectly pathologizing normal adolescent developmental changes[34]. However, the 2016 U.S. preventive taskforce systematic review on the efficacy of adolescent depression screening in primary care recommended screening because effective interventions are available[28]. They found no evidence of harm from screening. Future work might test whether screening for depression in clinical practice is accept- able and feasible for patients and doctors and results in better recognition and clinical outcomes.

Important questions remain about optimal ways to detect young people’s mental disorders in primary care, including providing GPs with appropriate resources to improve case detection. Training GPs is clearly only part of the answer. Our study adds to other work suggesting that young people’s

self-recognition enhances detection, along with continuity of care with the provider [3], destigmatizing mental health and help seeking[35], and adjustments in care organization such as through the use of technological assessments[36].

Finally, the complexity of time barriers and financial incentives that limit or promote screening needs consideration, given that routine screening is challenging within current 10- to 15-minute appointments. Recent research highlights the poten- tial of previsit screening and e-health initiatives to overcome this barrier[37,38]. Further research should focus on approaches to enable young people, clinicians, and practice systems to better engage with discussions about mental health; connect clinicians to resources that might assist with case-confirmation and evidence-based care options, and connect young people with appropriate resources to improve mental health literacy. With participation of young people, clinicians, practice staff, and par- ents, more effective solutions for accurately detecting mental disorder in young people presenting to primary care may become a reality[38].

Acknowledgments

The authors gratefully acknowledge the PARTY research team for their work in implementing the trial. They also thank Brenda Grabsch for her assistance with the PARTY methodology and to Patty Chondros for her expert consultation on data analyses.

They are most grateful to the clinicians, practices, and young people who participated in the study.

Funding Sources

This study was supported by grants from the Australian Health Ministers’ Advisory Council (AHMAC PDR 2005/06);

Australian Primary Health Care Research Institute (AHPRI); Na- tional Health and Medical Research Council (566849). The funding bodies had no role in the study design data collection, analysis, interpretation, or manuscript writing.

References

[1]Ozer EM, Adams SH, Lustig JL, et al. Increasing the screening and counseling of adolescents for risky health behaviors: A primary care intervention.

Pediatrics 2005;115:960e8.

[2]Tylee A, Haller DM, Graham T, et al. Youth-friendly primary-care services:

How are we doing and what more needs to be done? Lancet 2007;369:

1565e73.

[3]Haller DM, Sanci LA, Sawyer SM, Patton GC. The identification of young people’s emotional distress: A study in primary care. Br J Gen Pract 2009;

59:e61e70.

[4]Gould MS, Greenberg T, Velting DM, Shaffer D. Youth suicide risk and preventive interventions: A review of the past 10 years. J Am Acad Child Adolesc Psychiatry 2003;42:386e405.

[5]Fergusson DM, Woodward LJ. Mental health, educational, and social role outcomes of adolescents with depression. Arch Gen Psychiatry 2002;59:

225e31.

[6]Wang PS, Aguilar-Gaxiola S, Alonso J, et al. Use of mental health services for anxiety, mood, and substance disorders in 17 countries in the WHO world mental health surveys. Lancet 2007;370:841e50.

[7]Gledhill J, Kramer T, Iliffe S, Garralda ME. Training general practitioners in the identification and management of adolescent depression within the consultation: A feasibility study. J Adolesc 2003;26:245e50.

[8]Pfaff JJ, Acres JG, McKelvey RS. Training general practitioners to recognise and respond to psychological distress and suicidal ideation in young people. Med J Aust 2001;174:222e6.

[9]Martinez R, Reynolds S, Howe A. Factors that influence the detection of psychological problems in adolescents attending general practices. Br J Gen Pract 2006;56:594e9.

(9)

[10] Thapar A, Collishaw S, Pine DS, Thapar AK. Depression in adolescence.

Lancet 2012;379:1056e67.

[11] Merikangas KR, He J-P, Burstein M, et al. Service utilization for lifetime mental disorders in U.S. Adolescents: Results of the National Comorbidity Survey-Adolescent Supplement (NCS-A). J Am Acad Child Adolesc Psychi- atry 2011;50:32e45.

[12] Ozer EM, Zahnd EG, Adams SH, et al. Are adolescents being screened for emotional distress in primary care? J Adolesc Health 2009;44:520e7.

[13] Kramer T, Iliffe S, Bye A, et al. Testing the feasibility of therapeutic iden- tification of depression in young people in British general practice.

J Adolesc Health 2013;52:539e45.

[14] Sanci LA, Coffey CM, Veit FC, et al. Evaluation of the effectiveness of an educational intervention for general practitioners in adolescent health care: Randomised controlled trial. BMJ 2000;320:224e30.

[15] Sanci L, Chondros P, Sawyer S, et al. Responding to young people’s health risks in primary care: A cluster randomised trial of training clinicians in screening and motivational interviewing. PLoS ONE 2015;

10:e0137581.

[16] Campbell MK, Elbourne DR, Altman DG, CONSORT group. CONSORT statement: Extension to cluster randomised trials. BMJ 2004;328:702e8.

[17] Australian Bureau of Statistics (ABS). SEIFA: Socio-Economic Indexes for areas. Canberra: Australian Bureau of Statistics; 2006.

[18] McCambridge J, Strang J. The efficacy of single-session motivational interviewing in reducing drug consumption and perceptions of drug- related risk and harm among young people: Results from a multi-site cluster randomized trial. Addiction 2004;99:39e52.

[19] Goldenring JM, Rosen DS. Getting into adolescent heads: An essential up- date. Contemp Pediatr 2004;21:6490.

[20] Yeo M. Psychosocial risk screening in adolescent inpatients. Doctorate.

Melbourne, Australia: University of Melbourne, Department of Pediatrics;

2008.

[21] Hegney D, Price K, Patterson E, et al. Australian consumers’expectations for expanded nursing roles in general practiceechoice not gatekeeping. Aust Fam Physician 2004;33:845e9.

[22] Porritt J. Australian Divisions of General Practice. Nurses in general practicedimproving access and quality health care for consumers.

Canberra: Australian Division of General Practice; 2004.

[23] Üstün TB, Sartorius N. Mental illness in general health care: An interna- tional study. Hoboken, NJ: John Wiley & Sons; 1995.

[24] Kessler RC, Barker PR, Colpe LJ, et al. Screening for serious mental illness in the general population. Arch Gen Psychiatry 2003;60:184e9.

[25] Furukawa TA, Kessler RC, Slade T, Andrews G. The performance of the K6 and K10 screening scales for psychological distress in the Australian

National Survey of Mental Health and Well-being. Psychol Med 2003;33:

357e62.

[26] Peat J, Barton B. Medical statistics: A guide to data analysis and critical appraisal. Hoboken, NJ: John Wiley & Sons; 2008.

[27] Australian Bureau of Statistics (ABS). ARIA: Accessibility/Remoteness Index of Australia. Available at:http://www.health.gov.au/internet/main/

publishing.nsf/Content/health-historicpubs-hfsocc-ocpanew14a.htm.

Accessed September 16, 2017.

[28] Forman-Hoffman V, McClure E, McKeeman J, et al. Screening for major depressive disorder in children and adolescents: A systematic review for the U.S. Preventive Services Task Force. Ann Intern Med 2016;164:

342e9.

[29] Haller DM. Illness beliefs and the identification of emotional distress in youth: A study in primary care; Thesis Chapter 9.6. Melbourne, Australia:

University of Melbourne; 2008.

[30] McGuire W, Clerihew L. The effect offluconazole prophylaxis in very low birthweight infants is overestimated by before-after studies. Arch Dis Child Fetal Neonatal Ed 2010;95:F182e7.

[31] Howard JW. Better mental health services for Australia, media release, better access. 2015. Available at:http://pandora.nla.gov.au/pan/10052/20061221 -0000/www.pm.gov.au/news/media_releases/media_Release1858.html.

Accessed October 5, 2017.

[32] Haller DM, Sanci LA, Sawyer SM, Patton G. Do young people’s illness beliefs affect healthcare? A systematic review. J Adolesc Health 2008;42:

436e49.

[33] Tait L. To disclose or not to disclose psychological problems to GPs. Br J Gen Pract 2009;59:638e9.

[34] Sanci L, Lewis D, Patton G. Detecting emotional disorder in young people in primary care. Curr Opin Psychiatry 2010;23:318e23.

[35] Clement S, Schauman O, Graham T, et al. What is the impact of mental health-related stigma on help-seeking? A systematic review of quantita- tive and qualitative studies. Psychol Med 2015;45:11e27.

[36] Olson AL, Gaffney CA, Hedberg VA, Gladstone GR. Use of inexpensive technology to enhance adolescent health screening and counseling. Arch Pediatr Adolesc Med 2009;163:172e7.

[37] Gadomski AM, Fothergill KE, Larson S, et al. Integrating mental health into adolescent annual visits: Impact of previsit comprehensive screening on within-visit processes. J Adolesc Health 2015;56:267e73.

[38] Webb M, Sanci L, Kauer S, Wadley G. Designing a health screening tool to help young people communicate with their general practi- tioner. OzCHI’15 Proceedings of the Annual Meeting of the Australian Special Interest Group for Computer Human Interactions 2015:

124e133.

A.-E. Ambresin et al. / Journal of Adolescent Health xxx (2017) 1e8 8

Références

Documents relatifs

The Quality Mental Health Care Framework builds on the Health Standards Organization/Canadian Patient Safety Institute Canadian Quality and Patient Safety Framework for Health

Source: Nanos Research online survey, September 15-17, 2019, n=1,004 people in Canada, 18 years of age and older?. No margin of

To overcome barriers to closing the gap between resources and the need for treatment of mental disorders, and to reduce the number of years lived with disability and deaths

Seek psychosocial support from available services in your community, if stress, worry, fear or sadness get in the way of your daily activities for several days in a row, or if

Given that psychological impacts of emergencies and of emergency protective actions implemented during response are often greater than the actual physical impact of radiation, it is

In the aftermath of a population’s exposure to severe stressors, it is preferable to focus on medium-and long-term development of community-based and primary mental health

• Равномерное распределение самой трудной инеблагодарной работы среди всех членов коллектива для того, чтобы одному человеку или небольшой группе лиц не

During observational period, the superintendent must examine the patient and discharge him if he is not mentally ill; or he may authorize detention and