Interviewing Presentation Assignment Instructions Overview As we have learned in this course, the clinical interview is a core component of a psychological assessment or evaluation. For this Inte
Chronic Stressors and Adolescents’Externalizing Problems: Genetic Moderation by Dopamine Receptor D4. The TRAILS Study Anna Roos E. Zandstra 1&Johan Ormel 1&Pieter J. Hoekstra 1&Catharina A. Hartman 1 Published online: 30 March 2017 # The Author(s) 2017. This article is published with open access at Springerlink.com Abstract The existing literature does not provide consistent evidence that carriers of the Dopamine D4 Receptor 7-repeat allele are more sensitive to adverse environmental influences, resulting in enhanced externalizing problems, compared to noncarriers. One explanation is that the adverse influences examined in prior studies were not severe, chronic, or distressing enough to reveal individual differences in sensitiv- ity reflected by DRD4–7R. This study examined whether the 7-repeat allele moderated the association between chronic stressors capturing multiple stressful aspects of individuals ’ lives and externalizing problems in adolescence. We expected that chronic stressor levels would be associated with external- izing levels only in 7-repeat carriers. Using Linear Mixed Models, we analyzed data fr om 1621 Dutch adolescents (52.2% boys), obtained in three measurement waves (mean age approximately 11, 13.5, and 16 years) from the TRacking Adolescents ’Individual Lives Survey (TRAILS) population-based birth cohort and the parallel clinic-referred cohort. Across informants, we found that higher levels of chronic stressors were related to higher externalizing levels in 7-repeat carriers but not in noncarriers, as hypothesized. Although previous studies on the 7-repeat allele as a modera- tor of environmental influences on adolescents ’externalizing problems have not convincingly demonstrated individual dif- ferences in sensitivity to adverse environmental influences, our findings suggest that adolescent carriers of the Dopamine D4 Receptor 7-repeat allele are more sensitive to chronic, multi-context stressors than noncarriers. Keywords Chronic stressors . Psychosocial adversity . Sensitivity to the environment . Dopamine D4 receptor 7-repeat allele ( DRD4–7R) . Externalizing problems . Adolescence Exposure to psychosocial stre ssors increases adolescents’risk of psychopathology (for an overview, see Grant et al. 2004), including rule-breaking and aggressive (externalizing) behavior as seen in oppositional defiant disorder (ODD) and conduct disorder (CD). However, individual differences in outcome are large (Jenkins 2008; Rutter 2005), suggesting that some individuals are more sensitive to their environment than others. A polymorphism in the third exon of the Dopamine D4 Receptor ( DRD4) gene encodes for a variable number of tan- dem repeats, ranging from 2 to 11 (Bakermans-Kranenburg and Van IJzendoorn 2011; Dmitrieva et al. 2011; Ptacek et al. 2011 ). The 7-repeat (7R) variant results in lower affinity for dopamine (Ptacek et al. 2011), one of the brain ’s chemical messengers that is of interest in relation to externalizing prob- lems, through its assumed role in reward mechanisms, moti- vation, and approach behavior (Dmitrieva et al. 2011). The global frequency of the 7R allele is about 20%, with consid- erable variation across populations (Chang et al. 1996). DRD4 –7R has been extensively examined as a moderator of the association between environmental influences and ex- ternalizing problems, based on the notion that the 7R allele may reflect sensitivity to the environment, for better and for worse. These studies included influential, relatively proximal environmental factors (and not, for example, exposure to media or video game violence, see Ferguson 2015;Savage Electronic supplementary material The online version of this article (doi:10.1007/s10802-017-0279-4) contains supplementary material, which is available to authorized users. * Catharina A. Hartman [email protected] 1 Department of Psychiatry, University of Groningen, University Medical Center Groningen, P.O. Box 30.001, 9700 RB Groningen, The Netherlands J Abnorm Child Psychol (2018) 46:73 –82 DOI 10.1007/s10802-017-0279-4 and Yancey2008). According to this Differential Susceptibility model (e.g., Ellis et al. 2011), sensitive individ- ualsarelikelytobepositivelyaffectedbybeneficialenviron- mental influences (e.g., peer acceptance) and negatively by adverse influences (e.g., psychosocial stressors), whereas less sensitive individuals are less affected by both. Most empirical support for this model comes from studies on laboratory-observed parenting factors in relation to exter- nalizing problems in toddlers and preschoolers. Specifically, 10 month-old 7R carriers exposed to low vs. high laboratory- observed maternal sensitivity showed high vs. low externaliz- ing levels, respectively, approximately 2.5 years later; whereas noncarriers appeared unaffected by maternal sensitivity (Bakermans-Kranenburg and Van IJzendoorn 2006). In addi- tion, an intervention aimed at reducing toddlers ’externalizing problems by promoting maternal positive discipline proved to be more effective in 7R carriers than in noncarriers at follow- up (mean age 27, 39, and 52 months at pretest, posttest, and follow-up, respectively, Bakermans-Kranenburg et al. 2008). However, one study showed that in European-Americans, the influence of warm-responsive and negative-intrusive parent- ing at 6 and 12 months on externalizing levels at 18, 24, and 30 months did not significantly differ between 7R carriers and noncarriers (Propper et al. 2007). Another study showed that maternal sensitivity at 14 months, but not at 36 or 48 months, interacted with DRD4–7R in predicting later externalizing problems (Windhorst et al. 2015). Specifically, higher mater- nal sensitivity at 14 months predicted lower externalizing levels at 18 months (as well as at 60 months, but only via indirect paths across time) in 7R carriers, but did not affect noncarriers. At 36 months, however, 7R carriers showed a similar response, but noncarriers showed the opposite (i.e., higher sensitivity predicted higher externalizing levels), rather than no response. One study in middle childhood showed higher sensitivity for better and for worse in 7R carriers com- pared to noncarriers, to verbal but not physical peer victimi- zation and with respect to self-reported but not parent-reported externalizing problems (DiLalla et al. 2015). Thus, these prior findings do not consistently support the Differential Susceptibility model. In samples with a broad age range that included adoles- cence findings showed no evidence that DRD4–7R moderated the influence of maternal expressed emotion (i.e., warmth, criticism) on conduct problems (mean age 11 years, range 5 –17 years, Sonuga-Barke et al. 2009) or on prosocial and antisocial behavior (mean age 17 years, range 7 –28 years, Richards et al. 2015). In studies that focused on externalizing problems in (pre)adolescents (between 11 and 20 years old), the notion that DRD4–7R may reflect sensitivity to the envi- ronment, for better and for worse, has received moderate but inconsistent support. One study showed higher sensitivity, for better and for worse, in 7R carriers, compared to noncarriers; to laboratory-obs erved early maternal stimulation and responsiveness, but not to parent-reported early family adver- sity, with respect to adolescents ’symptoms of CD/ODD (com- bined parent-report and self-report) and psychopathy (parent- report, Nikitopoulos et al. 2014). Another study showed rela- tively high sensitivity, for better and for worse, in 7R carriers to a broad range of intervention-targeted parenting behaviors (reported by parents), with respect to self-reported substance use, but not to parent-reported delinquency (Beach et al. 2010 ). In addition, recent findings showed higher sensitivity for better and for worse in 7R carriers compared to noncar- riers, to positive and negative social preference with respect to teacher-reported conduct problems (Buil et al. 2015). In con- trast, no moderating effect of DRD4genotype was found on the association between peer influence (self-reported peer rejection and acceptance) and several measures of externaliz- ing problems (parent-report and self-report, Janssens et al. 2015 ). Prior studies from our research group TRAILS (TRacking Ado lescents ’Individual Lives Survey) have, likewise, pro- duced inconsistent results. One of these (Nederhof et al. 2012a ) showed that the 7R allele moderated the association between parental separation and self-reported externalizing problems, although this effect pertained only to boys, not to girls, and only to the absence of parental separation, not to its presence. That is, externalizing levels of 7R –carrying boys compared to noncarriers were relatively low if their families were intact but did not differ if their parents had separated, suggesting sensitivity for better but not for worse. Other stud- ies from our research group showed no evidence that 7R car- riers are relatively sensitive to the influence of peers (teacher- reported peer victimization and self-reported social well-be- ing) on self-reported delinquency (Kretschmer et al. 2013)or of parenting (rejection, overprotection, and emotional warmth as reported by pre-adolescents) on delinquency and aggres- sion (combined parent-report and self-report, Marsman et al. 2013 ) or substance use (self-report, Creemers et al. 2011). Taken together, there are clearly many inconsistencies in the literature as to whether DRD4–7R may reflect individual differences in sensitivity to environmental influences. The in- consistencies in prior findings in adolescence, which is also the focus of the present study, do not appear to be driven by an informant effect, nor by differences in operationalization of externalizing problems (e.g., substance use vs. delinquency vs. broader externalizing measures) or environmental influ- ences (e.g., parent vs. peer influence; broad vs. narrow aspects of parenting). Rather, what seems to stand out in these prior findings is the lack of evidence that DRD4–7R reflects sensi- tivity to the detrimental effects of adverse environmental in- fluences. Of the few findings that did support high sensitivity not only for better but also for worse in 7R carriers (cf. Beach et al. 2010;Builetal. 2015; Nikitopoulos et al. 2014)most were based on the absence of positive (beneficial) environ- mental influences. For example, while high levels of maternal 74 J Abnorm Child Psychol (2018) 46:73–82 stimulation and responsivity in the study by Nikitopoulos et al. (2014) were considered to be beneficial, low levels reflect an absence of beneficial influence, rather than presence of adverse influence (e.g., the presence of ma- ternal hostility). In contrast, of the prior findings in ado- lescence relating to actual ad verse influence (i.e., early family adversity, perceived p arental rejection or overpro- tection, parental divorce or separation, peer victimization, peer rejection, negative social preference) only one (Buil et al. 2015) suggested differences in externalizing levels between 7R carriers and noncarriers (Creemers et al. 2011; Janssens et al. 2015; Kretschmer et al. 2013; Marsman et al. 2013; Nikitopoulos et al. 2014). Thus, in adolescents, the Differential Susceptibility hypothesis, ex- tending the Diathesis-Stress theory (Zuckerman 1999)that some individuals are more vulnerable to the detrimental effects of adverse influences, has not received much sup- port from the data. One explanation could be that the adverse environmental influences examined in prior studies were not severe, chronic, or distressing enough to reveal individual differences in sen- sitivity reflected by DRD4–7R. Adverse environmental influ- ences may become more severe or distressing as they persist over time, taxing individuals ’physical and psychological cop- ing resources. In addition, subtle individual differences in sen- sitivity may be missed when the adverse environmental influ- ence is rather narrowly operationalized, capturing only one aspect of individuals ’lives. That is, the adverse environmental influences examined in prior studies generally reflected a spe- cific aspect of a single environmental domain (e.g., either family or peer group) while beneficial influences from other domains, if present, will compensate for their impact. Individual differences in sensitivity may thus be easier to de- tect by assessing environmental influences that are chronic and reflect multiple adverse aspects across multiple environ- mental domains (e.g., family, peers, school, and neighbor- hood). We hypothesize that if DRD4–7R truly reflects indi- vidual differences in sensitivity to the environment, not only for better, as some prior findings have shown, but also for worse, this may become evident in the presence of chronic, multi-context stressors, which may exceed sensitive individ- uals ’ability to cope. This study aimed to enhance our understanding of in- dividual differences in adolescents ’externalizing prob- lems following exposure to chronic stressors. To this end, we have examined whether DRD4–7R is a moderator of the association between chr onic stressors, operational- ized as number of long-term dif ficulties, and externalizing (CD and ODD) problems from preadolescence into ado- lescence. We expected that, only in 7R carriers, chronic stressor levels would be positively associated with exter- nalizing levels, whereas in noncarriers, we expected no influence of chronic stressor s on externalizing levels. Method Participants We obtained the data used in this study from the first three measurement waves of TRAILS (mean ages about 11, 13.5, and 16 years). TRAILS aims to contribute to the understand- ing of the etiology of mental health problems by following 10–12 year-old Dutch children biennially into adulthood. We pooled data from the TRAILS population-based birth cohort ( n= 2230) and the parallel clinic-referred cohort ( n = 543), to obtain a large sample with a wide range of problem severity and chronic stress. The sampling procedures, descriptive statistics, and response rates of both cohorts are well-documented (e.g., De Winter et al. 2005; Huisman et al. 2008 ; Ormel et al. 2012). In brief, TRAILS approached 135 primary schools in five muni cipalities in the Northern Netherlands to build the population cohort. Of these schools, 90.4% agreed to participate. TRAILS contacted eligible stu- dents and their parents (excluding individuals with mental retardation and individuals without a Dutch-speaking parent or parent surrogate), enrolling 76% ( n= 2230; 49.2% boys; 86.5% Dutch ancestry; mean age 11.11; SD0.56; range 10.01 –12.58) of those contacted in the study. The three data waves we included in this study ran from March 2001 to July 2002 (T1), September 2003 to December 2004 (T2), and September 2005 to August 2007 (T3); with response rates consistently above 80%. The smaller clinic-referred sample ( n= 543) consists of pre-adolescents who had been referred to the Groningen University Child and Adolesce nt Psychiatric Outpatient Clinic at any point in their life (20.8% ≤5years;66.1%6 – 9years;13.1%10 –12 years) for consultation or treatment. The first three data waves in the clinic-referred cohort ran two years behind those of the population cohort: From September2004toDecember2005(T1),September2006to November 2007 (T2), and September 2009 to February 2011 (T3). The measurement instruments and design for the clinic- referred cohort were the same as those of the population co- hort. Of the 1264 eligible pre-adolescents, 543 (65.9% boys; 98.2% Dutch ancestry; mean age 11.11; SD0.50; range 10.13 –12.40) enrolled in the study and finished baseline mea- surements (T1). Of these 543 baseline participants, 85.1% ( n = 462) participated in the second wave (T2). Of the T2 participants, 83.5% ( n= 386) also participated in the third wave (T3). Another 30 T2 dropouts agreed to participate in the third wave, resulting in a total T3 response rate of 76.6% ( n= 416) of the original sample. Selective attrition analyses have been described elsewhere (De Winter et al. 2005; Huisman et al. 2008; Nederhof et al. 2012b;Ormel et al. 2012). Importantly, baseline participants did not differ from non-participants with respect to externalizing problems. J Abnorm Child Psychol (2018) 46:73 –82 75 Compared to the population cohort, the clinic-referred co- hort had, on average, a higher socio-economic status, t (886.310) = 4.548, p< 0.001, consisted of more boys, t (859.668) = 7.274, p<0.001, and less individuals of non- Dutch ancestry, t(2209.170) = 12.563, p<0.001. Procedures Every measurement wave, adolescents and their parents (typi- cally the mother, >95%) filled out several questionnaires. Parents were assessed at home. Adolescents were assessed at school (population cohort) or at the Groningen University Child and Adolescent Outpatient Clinic (clinic-referred cohort), under the supervision of one or more well-trained assistants. For ad- olescents ’DNA analysis, blood or buccal cells were collected at T2 (clinic-referred cohort) or T3 (population cohort). Parents gave written informed consent prior to each assessment wave. Adolescents gave written informed assent at the second and third wave. TRAILS was approved by the National Dutch Medical Ethics Committee, in acco rdance with the ethical stan- dards laid down in the 1964 Declaration of Helsinki. Measures Externalizing Problems TRAILS used the Achenbach System of Empirically Based Assessment (ASEBA) family of measures of mental health problems (Achenbach and Rescorla 2001; Verhulst and Van der Ende 2013)ateachtime point. The Child Behavior Checklist (CBCL) and the Youth Self-Report (YSR) contain 120 items assessing behavioral and emotional problems in children over the past 6 months. These items can be rated as 0 (not true),1(somewhat or sometimes true), or 2 (very or often true) . We used DSM-IV-oriented subscales to define externalizing problems as the sum of the average scale scores of oppositional defiant problems ( k=5; Cronbach ’s α =0.81and α= 0.64 for parent-report and self- report, respectively) and conduct problems ( k=17, α=0.82 for parent-report; k=15, α= 0.75 for self-report). Then, sum scores were standardized. Externalizing problems correlated significantly ( p<0.001) with internalizing problems, at both T2 ( r= 0.52, parent-re- port, and r= 0.38, self-report) and T3 ( r= 0.54, parent-report, and r= 0.32, self-report). Since internalizing problems have no theoretical and empirical relevance in relation to DRD4–7R (e.g., Bakermans-Kranenburg and Van IJzendoorn 2006; DiLallaetal. 2015), potential interaction effects of DRD4 – 7R and stressors in predicting externalizing problems may be weakened by the presence of co-occurring internalizing problems. Therefore, we focused on externalizing problems adjusted for co-occurring internalizing problems (EXTadj). To that end, we computed the summed weighted average of anx- iety ( k=6; α=0.73and α= 0.61 for parent-report and self- report, respectively) and affective problems ( k=13; α=0.72 and α= 0.71 for parent-report and self-report, respectively), after which we computed residual externalizing scores ( M=0; SD = 1). Although our main focus is on residual externalizing problems (EXTadj), results are also described for unadjusted externalizing problems (EXT), with and without correcting for internalizing problems as a covariate. Chronic Stressors Preceding T2 and T3 We operationalized chronic stressor levels at T2 and T3 as the number of parent- reported long-term difficulties since the previous measure- ment. One of the parents, typically the mother, filled out a TRAILS questionnaire that listed long-term difficulties that were described in a broad way in an effort to capture multiple possible subtypes to which the adolescent might have been exposed since the previous interview (e.g., Oldehinkel et al. 2008 ;Zandstraetal. 2015). The stressors included: (1) chron- ic illnesses or physical handicaps of the child or (2) a family member; (3) high work pressure at school; (4) housing prob- lems; (5) neighborhood problems, such as violence or discrim- ination; (6) financial problems; (7) lack of friends; (8) being bullied; (9) long-lasting conflicts with family members or (10) others; and (11) long-lasting conflicts between family mem- bers. On an open item, parents could also disclose additional long-term difficulties. We coded these additional problems either as a long-term difficulty or dismissed them according to well-defined rules —in particular whether the described sit- uation is typically considered stressful and enduring. For ex- ample, we coded a turbulent home environment, such as mov- ing frequently from house to house or parents having an on/off relationship, as long-term difficulties. Situations that we rejected as long-term difficulty included normative or non- enduring situations such as the transition to middle school, puberty, and quarrels with siblings. The number of reported difficulties ranged from 0 to 10. To reduce the influence of extreme and rare scores, we truncated 3 to 10 long-term diffi- culties as 3 or more, based on the frequency distribution (see Online Resource 1). DRD4 Genotyping DNA was extracted from blood samples or buccal swabs (Cytobrush®) using a manual salting out procedure (Miller et al. 1988). The 48 bp direct repeat poly- morphism in exon 3 of DRD4was genotyped on the Illumina BeadStation 500 platform (Illumina Inc., San Diego, CA, USA), described in detail elsewhere (Nederhof et al. 2012a). The genotyping assay was carried out in a CCKL quality- certified laboratory and has been validated in earlier tests. Three percent blanks as well as duplicates between plates were processed as quality controls during genotyping. Determination of the length of the alleles was performed by direct analysis on an automated capillary sequencer (ABI3730, Applied Biosystems, Nieuwerkerk, The Netherlands) using standard conditions. We formed two 76 J Abnorm Child Psychol (2018) 46:73–82 groups according to the presence of at least one 7R allele (1 = 7R carrier; 0 = noncarrier, i.e., all others). Data Analysis Data Preparation and Preliminary AnalysesFor this study, our statistical analysis method required at least one value for each predictor on T1-T3 and at least T2 or T3 externalizing problems. Thus, we needed T2 and/or T3 parent-reported and/ or self-reported externalizing problems, T2 and/or T3 chronic stressors, and DRD4. Participants not from Dutch ancestry were excluded, since genetic effects and gene-environment interaction effects are not necessarily generalizable across ra- cial populations (Bakermans-Kranenburg and Van IJzendoorn 2011 ; also see Propper et al. 2007). Of each sibling pair, we excluded one participant at random. We performed indepen- dent samples t-tests to check whether included and excluded subjects differed with respect to our study variables. Main Analyses We computed correlation coefficients be- tween the predictors and T2 and T3 externalizing problems. The possible presence of gene-environment correlations (i.e., DRD4 genotype is associated with exposure to chronic stressors) may drive gene-env ironment interaction effects and therefore needs to be ruled out. We used Linear Mixed Modeling (LMM) to investigate the effects of chronic stressors, DRD4–7R, and their hypothesized interaction in predicting subsequent EXTadj. LMM allows for missing data at different measurement waves, which is an important advantage for a longitudinal design (Kwok et al. 2008 ). Using PASW Statistics 18, we conducted LMM anal- yses (T2 and T3 in a single analysis), separately for parent- reported and self-reported EXTadj. We included the indepen- dent variables of age (time-variant), sex (0 = female; 1 = male), initial EXTadj at T1, chronic stressors (time-variant), and DRD4 –7R, as well as an interaction between chronic stressors and DRD4 –7R. All non-dichotomous variables were centered prior to analysis. For interpretation of interaction effects we plotted EXTadj levels based on the estimated regression coef- ficients, for different levels of each predictor. We used the Maximum Likelihood estimation procedure and considered a p-value < 0.05 to be statistically significant. For post-hoc probing of statistically significant interaction effects, we computed simple slopes, which reflect the slopes of regression lines in a plot, and regions of significance, indi- cating the range of values of a predictor at which the interac- tion effect is statistically significant (Preacher et al. 2006). Regions of significance result from separate analyses that may produce values of a predictor that fall outside the true data range. To examine the potential influence of sex on sig- nificant interaction effects, we repeated our main analyses (in which we controlled only for a main effect of sex on EXTadj) by adding sex by DRD4–7R and sex by stressors interaction terms to the model, as has recently been recommended in the literature (Keller 2014). If findings showed significant sex by predictor interaction effects, we tested for an additional three- way interaction effect of chronic stressors, DRD4–7R, and sex, in predicting EXTadj. To check the influence of co- occurring internalizing problems, we repeated the analysis replacing the outcome variable EXTadj with EXT; that is, externalizing problems unadjusted for co-occurring internal- izing problems. Additionally, we added internalizing prob- lems as a covariate to the model, to rule out the possibility that our main findings are driven by the use of residual exter- nalizing problems. Results Results of Preliminary Analyses Three hundred and nine participants had missing data for both measurements of chronic stressors, and 137 for parent- reported as well as self-reported externalizing problems. Of the 1861 participants with available DRD4data, those not from Dutch ancestry ( n= 166) were excluded. Of the sibling pairs in the remaining groups, one of each was excluded ( n = 22). Altogether, we excluded a total of 1152 participants ( n = 1005 population cohort; n= 147 clinic-referred cohort) from this study, resulting in a final sample of 1621 subjects ( n = 1225 population cohort; n= 396 clinic-referred cohort). We compared the final study sample (mean age 11.09; SD 0.55; range 10.01 –12.58; 52.2% boys; 75.6% population co- hort) with those who were not included. We found that partic- ipants were somewhat younger, t(2769) = −2.309, p=0.021, and had higher T2 chronic stressor levels, t(1570.583) = 2.248, p = 0.025. There were no significant differences between the groups with respect to sex, p= 0.692, DRD4–7R, p=0.478, and parent- and self-reporte d externalizing problems,p=0.538 and 0.252, respectively. DRD4 –7R and Chronic Stressors Table 1shows descriptive statistics and frequencies of the final sample and Table 2Pearson correlations between predic- tors and parent-reported and self-reported EXTadj. There was no indication of gene-environment correlations as DRD4–7R was not significantly associated with chronic stressors at T2 (Spearman r= −0.03, p=0.243)orT3( r=0.00, p=0.932). See Online Resource 1for number of chronic stressors report- ed and frequencies per chronic stressor. As shown in Table 3, parent-reported and self-reported EXTadj problems were significantly predicted by a two-way interaction effect of chronic stressors and DRD4–7R ( p = 0.023 and p= 0.024, respectively). We plotted the levels of EXTadj for low, average, high, and very high levels of the J Abnorm Child Psychol (2018) 46:73 –82 77 truncated chronic stressor variable (corresponding to 0, 1, 2, and 3 or more long-term difficulties, respectively), separately for 7R carriers and noncarriers. Figure1shows that higher chronic stressor levels were related to higher EXTadj in 7R carriers, while EXTadj of noncarriers was stable across chron- ic stressor levels. Post-hoc probing of these interaction effects resulted in simple slopes and regions of significance. The increase in EXTadj with chronic stress level was statistically significant for 7R carriers, t(1557.022) = 3.85, p< 0.001 for parent- report, and t(1572.308) = 3.74, p< 0.001 for self-report, while the slope of EXTadj across chronic stress groups did not sig- nificantly differ from zero for noncarriers, t(1557.022) = 1.06, p = 0.288 for parent-report, and t(1572.308) = 1.41, p=0.159 for self-report. Regions of significance showed that the interaction effect between chronic stressors and DRD4–7R in predicting EXTadj was statistically significant below −1.34 and −0.32 chronic stressors for parent-report and self-report, respectively, both non-existent values, and above 1.85 and 4.80 chronic stressors for parent-report and s elf-report, respectively. We conclude that our findings apply to the upper end of the chronic stressor range, not to the lower end, and that the effect is stronger based on parent-report of externalizing problems than self-report. Controlling for potential interaction effects of sex with chronic stressors or DRD4–7R, parent-reported and self- reported EXTadj problems were still significantly predicted by a two-way interaction effect of chronic stressors and DRD4 –7R, p= 0.037 and p= 0.029, respectively. In these models, sex did not interact with DRD4–7R in predicting Ta b l e 1 Descriptive statistics (left) and frequencies (right) of the variables used in this study Va r i a b l e NMean (SD)Range 0( N)1( N)2( N)3+( N) Age T1 1621 11.09 (0.55) 10.01_12.58 T2 1620 13.35 (0.61) 11.58_15.08 T3 1576 16.14 (0.68) 14.42_18.48 CBCL EXT a T1 1573 6.03 (5.12) 0_31 T2 1586 4.62 (4.84) 0_29 T3 1451 4.46 (5.12) 0_34 YSR EXT a T1 1597 6.03 (4.41) 0_28 T2 1608 5.91 (4.20) 0_29 T3 1542 6.14 (4.52) 0_31 Stressors b T2 1587 1.27 (1.50) 0_10 657 389 256 285 T3 1458 1.34 (1.57) 0_10 559 392 229 278 DRD4 –7R c 1052 569 CBCL Child Behavior Checklist, YSRYouth Self-Report, EXTExternalizing problems (DSM-oriented subscales oppositional defiant problems and conduct problems), DRD4–7R Dopamine D4 Receptor 7-repeat allele, T measurement wave aSum of 22 item scores for parent-report and 20 items for self-report; range per item 0 –2bNumber of long-term difficulties experienced since previous measurementcCoded as 0 = noncarrier; 1 = carrier Ta b l e 2 Pearson correlation matrix of predictors and outcome variables, with parent-reported externalizing problems below and self-reported exter- nalizing problems above diagonal Self-report Variables T2Stressors T3Stressors DRD4–7R a Sex a T2EXTadj T3EXTadj T2Stressors 1 0.57*** -0.03 0.05 0.07** 0.06* T3Stressors 0.57*** 1 0.00 -0.00 0.06* 0.09*** DRD4 –7R a -0.03 0.0010.07** -0.01 -0.02 Sex a 0.05 -0.000.07**10.12*** 0.16*** Parent-report T2EXTadj 0.15*** 0.15*** -0.02 0.09*** 0.39*** 0.31*** T3EXTadj 0.13*** 0.13*** -0.02 0.09*** 0.30*** 0.46*** DRD4 –7R Dopamine D4 Receptor 7-repeat allele, EXTadjExternalizing problems adjusted for co-occurring internalizing problems, Tmeasurement wave. DRD4–7R was coded as 0 = noncarrier; 1 = carrier. Sex was coded as 0 = female; 1 = male. aSpearman rank order correlation, *** p<0.001, ** p<0.01,* p<0.05 78 J Abnorm Child Psychol (2018) 46:73–82 parent-reported or self-reported EXTadj,p=0.524and p = 0.241, respectively, nor with chronic stressors in predicting self-reported EXTadj, p= 0.150. However, sex did significantly interact with chronic stressors in predicting parent-reported EXTadj, p= 0.017, which may explain why the interaction effect of chronic stressors and DRD4–7R was somewhat weak- er compared to our main results. Visual inspection showed that the association between chronic stressor levels and EXTadj was stronger in 7R carriers than in noncarriers (both boys and girls), as in Fig. 1, and stronger in boys than in girls (both 7R carriers and noncarriers). However, we found no evidence of a three- way interaction effect of chronic stressors, DRD4–7R, and sex, in predicting parent-reported or self-reported EXTadj, p=0.610 and p= 0.251, respectively. These posthoc findings suggest that the association between chronic stressors level and EXTadj (at least parent-report) is moderated by DRD4–7R as well as by sex, but independent of each other. Without adjusting externalizing f or co-occurring internaliz- ing problems, a two-way interact ion effect of chronic stressors and DRD4 –7R did not hold in predicting parent-reported EXT, p = 0.161, but still significantly predicted self-reported EXT, p = 0.045. Visual inspection showed that the association be- tween chronic stressor levels and parent-reported EXT was strong overall with negligible differences between 7R carriers and noncarriers (albeit in the same direction as our main re- sults). The association between chronic stressors level and self-reported EXT was similarly strong for 7R carriers but was attenuated in noncarriers, as in Fig. 1but less pronounced. This weakening of effects due to co-occurring internalizing problems may suggest that our mai n findings pertain especially to Bpure ^externalizing problems and less to internalizing or comorbid externalizing and internalizing problems. In accor- dance, when we added internalizing problems as a covariate to the model, effects regained strength. Namely, a two-way in- teraction effect of chronic stressors and DRD4–7R was not sig- nificant in predicting parent-reported EXT, p= 0.099, but signif- icantinpredictingself-reportedEXT, p= 0.029, approaching our original findings on externalizing problems adjusted for Ta b l e 3 TheDRD4 –7-repeat allele significantly interacted with chronic stressors level in predicting parent-reported and self-reported externalizing prob- lems controlling for baseline ex- ternalizing problems Parent-reported EXTadj Self-reported EXTadj Parameter Estimate a SEa p Estimate a SEa p Intercept b 15.04 27.83 0.588 -50.36 31.34 0.108 Age 3.16 7.99 0.693 -5.96 9.26 0.520 Sex -15.42 35.72 0.666 102.05 40.35 0.012 T1 EXTadj 594.31 18.07 <0.001 372.42 20.22 <0.001 Stressors 11.92 16.81 0.479 21.04 19.04 0.266 DRD4 –7R 22.29 36.80 0.545 -14.54 40.93 0.725 DRD4 –7R*Stressors 65.18 28.65 0.023 73.12 32.43 0.024 EXTadj Externalizing problems corrected for internalizing problems, DRD4–7R Dopamine D4 Receptor 7-repeat allele, Tmeasurement wave. Variables were mean-centered except for DRD4–7R (0 = noncarrier; 1 = carrier) and sex (0 = female; 1 = male) aValues multiplied by 1000 for ease of interpretationbParticipants varied significantly ( p<0.01) in intercept for parent-reported EXTadj, var.(u0j) = 263.97 a,chi- square (1) = 240.66, and self-reported EXTadj, var.(u0j) = 316.20 a, chi-square (1) = 207.23 Fig. 1 Adjusted externalizing problems reported by parents (left panel) and adolescents (right panel) increased significantly with number of chronic stressors in DRD4–7R carriers but did not change in noncarriers. EXTadjExternalizing problems adjusted for co-occurring internalizing problems, 7RDopamine D4 Receptor 7-repeat allele.Levels of chronic stressors refer to the number of long-term difficulties. According to post-hoc probing, the interaction effect is statistically significant on the right of the dashed line for parent-report (above 1.85) and outside our data range for self-report (above 4.80 stressors) J Abnorm Child Psychol (2018) 46:73 –82 79 internalizing problems and with visual inspection showing sim- ilar plots as those depicted in Fig.1. Thus, our main findings do not appear to be driven by the method we used to correct for internalizing problems. These post-hoc analyses show the robustness and specific- ity of our main results. Tables and figures from these analyses are available upon request. Discussion This study aimed to contribute to the literature by examining whether DRD4–7R moderated the association between chronic stressors and externalizing probl ems. As hypothesized, higher chronic stressor levels were related to higher externalizing levels in 7R carriers but not in noncarriers, suggesting high vs. low sensitivity, respectively, to adverse environments. These results were consistent across informants and were not driven by ado- lescents ’gender. Although it has been posited that the 7R allele reflects sensitivity to adverse as well as to beneficial environmen- tal influences on externalizing problems (e.g., Bakermans- Kranenburg et al. 2008), this theory has received inconsistent support in adolescence (cf. Beach et al. 2010;Builetal.2015; Creemers et al. 2011;Janssensetal. 2015;Kretschmeretal. 2013 ; Marsman et al. 2013; Nederhof et al. 2012a; Nikitopoulos et al. 2014;Richardsetal. 2015; Sonuga-Barke et al. 2009). In particular, with only one exception (Buil et al. 2015), none of these prior studies have convincingly demonstrat- ed sensitivity to environmental influences in the adverse range. The present study thus adds to the literature by showing this sensitivity to adverse circumstanc es, at least to chronic stressors. Our results contrast with prior findings that DRD4–7R did not moderate the association between early family adversity (a relatively broad measure of environmental influence, as the one used in the present study), and adolescents ’symptoms of CD/ODD and psychopathy (Nikitopoulos et al. 2014). These findings were based on data from a parent-interview, conducted when participants were 3 months old, assessing which of eleven family adversity factors (e.g., low educational level, marital discord) were present in the year prior to the child ’s birth. One obvious explanation for the difference in findings would be the early age at which the environmental influence was assessed and, consequently, the large amount of time and contextual influences that passed between assess- ments of the environmental predictor and the behavioral out- come (i.e., 15 years), in contrast to our study, which assessed more recent environmental influences. However, given that the same study did show a moderating effect of DRD4–7R on the influence of laboratory-observed early maternal stimu- lation and responsiveness, also assessed at 3 months, we can- not conclude that DRD4–7R only moderates recent and not early environmental influences. The null finding may be due to prenatal family difficulties that are resolved before birth or that do not have longlasting effects on children. Moderating effects of DRD4–7R may be easier to detect when focusing on ongoing or chronic environmental difficulties, which presum- ably have a major impact on sensitive individuals, taxing their ability to cope, but not on less sensitive individuals. Other prior findings that the 7R allele did not moderate ad- verse environmental influences on adolescents ’externalizing problems came from our own research group (Creemers et al. 2011;Kretschmeretal. 2013; Marsman et al. 2013; Nederhof et al. 2012a ). These prior TRAILS findings have shown that the influences of parental rejection and overprotection, as perceived by preadolescents at T1 (mean age 11 years) on delinquency and aggression at T2 (mean age 13.5 years, combined parent-report and self-report; Marsman et al. 2013) and on substance use at T3 (mean age 16 years, self-report; Creemers et al. 2011)werenot moderated by DRD4–7R (7R carriers vs. noncarriers). Furthermore, the influence of parental separation, assessed at T1 and T3, on self-reported externalizing levels at T3 did not differ between 7R carriers vs. noncarriers (Nederhof et al. 2012a). Finally, teacher-reported peer victimization at T2 did not influence self-reported delinquency at T4 (mean age 19 years) in 7R carriers, in contrast to 4R carriers (Kretschmer et al. 2013). These findings have led our colleagues to suggest that moderat- ing effects of DRD4–7R on environmental influences on exter- nalizing problems apply less to adolescence than to childhood (Kretschmer et al. 2013; Marsman et al. 2013), less to peer in- fluence than to other environmental factors (Kretschmer et al. 2013), or may differ according to the operationalization of exter- nalizing problems (Creemers et al. 2011). Given that samples, age range, and genetic and outcome measures used in these studies partially overlap with ou rs, it is likely that our findings differ due to the way in which we have operationalized environ- mental adversity. Whereas some of the previously addressed ad- versities may be ongoing, as were the difficulties we have assessed, our study appears to stand alone in its measurement of chronic difficulties that collectively capture many different aspects of individuals ’lives (e.g., both family and peer contexts). Thus, our findings suggest that moderating effects of DRD4–7R on the association between adverse environmental influences and externalizing problems do extend to adolescence when focusing on chronic multi-context stresso rs. However, this finding will need to be replicated by future research. Although internalizing probl ems were not directly investi- gated in the current study, post- hoc findings suggest that inter- action effects of DRD4–7R and stressors in predicting external- izing problems may be weakened by the presence of co- occurring internalizing problem s. This is according to expecta- tion, given that internalizing (unlike externalizing) problems lack a clear theoretical connection to DRD4–7R and given that prior studies found no evidence that DRD4–7R moderated the association between environmenta l influences and internalizing problems (Bakermans-Kranenburg and Van IJzendoorn 2006; maternal sensitivity; DiLalla et al. 2015; peer victimization). 80 J Abnorm Child Psychol (2018) 46:73–82 Our study included a number of limitations. First, we collect- ed parent-reports, not self-repor ts, of long-term difficulties be- cause we assumed that parents are better and more stable judges of the difficulties that put chronic strain on family life. The stressors we examined included issues such as chronic housing problems and neighborhood problems. The drawback may be that such factors may be less stressful for adolescents than parents assume, thus overestimating stressor exposure. Another draw- back of parent-report can be that parents may not have full insight into the other chronic stressors we measured, such as bullying, that weigh heavily on adolescents ’life, leading to underestimating stressor exposure. Although shared method var- iance may strengthen the interaction effect between chronic stressors and the 7-repeat allele in predicting externalizing prob- lems reported by parents, it does not explain our similar findings on self-reported externalizing problems. Second, we focused on chronic adversities and did not study sensitivity to positive chron- ic conditions. A formal test of Differential Susceptibility includes both beneficial and adverse aspects of the environment, while we have only addressed the latter Diathesis-Stress model. Although sensitivity to beneficial environmental influences has been exam- ined relatively frequently in relation to DRD4–7R, as outlined in the introduction, it certainly would have complemented our find- ings, had we been able to incorporate this. Strengths of the study include the large sample of longitu- dinal, multi-informant data from pre-adolescence well into adolescence, large inter-individual differences in levels of ex- ternalizing problems and chronic stressors, and the use of Linear Mixed Modeling that allowed for optimal use of all available data from multiple measurements. In sum, whereas previous studies on DRD4–7R as a moder- ator of environmental influences on adolescents ’externalizing problems have not convincingly d emonstrated sensitivity to en- vironmental influences in the adverse range, we were able to do so by focusing on chronic multi-context stressors. Our finding that higher levels of chronic stressors were associated with higher externalizing levels in 7R carriers but not in noncarriers suggests high vs. low sensitivity, respectively, to adverse environments. We encourage further studies of environmental influences that reflect multiple adverse aspects across multiple environmental domains (e.g., family, peers, school, and neighborhood). Acknowledgments This research is part of the TRacking Adolescents ’ Individual Lives Survey (TRAILS). Participating centers of TRAILS in- clude various departments of the University Medical Center and University of Groningen, the Erasmus University Medical Center Rotterdam, the University of Utrecht, the Radboud Medical Center Nijmegen, and the Parnassia Bavo group, all in the Netherlands. TRAILS has been financially supported by various grants from the Netherlands Organization for Scientific Research (NWO), ZonMW, GB-MaGW, the Dutch Ministry of Justice, the European Science Foundation, BBMRI-NL, the participating universities, and Accare Center for Child and Adolescent Psychiatry. We are grateful to all ado- lescents, their parents, and teachers who participated in this research, and to everyone who worked on this project and made it possible. Compliance with Ethical Standards Conflict of Interest The authors declare that they have no conflict of interest. Ethical Approval TRAILS was approved by the National Dutch Medical Ethics Committee and has therefore been performed in accor- dance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. Informed Consent Parents gave written informed consent prior to each assessment wave. Adolescents gave written informed assent at the second and third waves. This manuscript contains no information that discloses the identity of our participants or violates their privacy. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. References Achenbach, T. M., & Rescorla, L. A. (2001). Manual for the ASEBA school-age forms & profiles . Burlington, VT: University of Vermont, Research Center for Children, Youth & Families. Bakermans-Kranenburg, M. J., & Van IJzendoorn, M. H. (2006). Gene- environment interaction of the dopamine D4 receptor (DRD4) and observed maternal insensitivity predicting externalizing behavior in preschoolers. Developmental Psychobiology, 48, 406– 409. doi: 10.1002/dev.20152 . Bakermans-Kranenburg, M. J., & Van IJzendoorn, M. H. (2011). Differential susceptibility to rearing environment depending on dopamine-related genes: new evi dence and a meta-analysis. Development and Psychopathology, 23,39 –52. doi: 10.1017 /S0954579410000635 . Bakermans-Kranenburg, M. J., Van IJzendoorn, M. H., Pijlman, F. T. A., Mesman, J., & Juffer, F. (2008). Experimental evidence for differ- ential susceptibility: dopamine D4 receptor polymorphism (DRD4 VNTR) moderates intervention effects on toddlers’ externalizing behavior in a randomi zedcontrolledtrial.Developmental Psychology, 44 ,293–300. doi: 10.1037/0012-1649.44.1.293 . Beach, S. R. H., Brody, G. H., Lei, M. K., & Philibert, R. A. (2010). Differential susceptibility to parenting among African American youths: testing the DRD4 hypothesis. Journal of Family Psychology, 24 ,513–521. doi: 10.1037/a0020835 . Buil, J. M., Koot, H. M., Olthof, T., Nelson, K. A., & Van Lier, P. A. C. (2015). DRD4 genotype and the developmental link of peer social preference with conduct problems and prosocial behavior across ages 9-12 years. Journal of Youth and Adolescence, 44 , 1360– 1378. doi: 10.1007/s10964-015-0289-x . Chang, F. M., Kidd, J. R., Livak, K. J., Pakstis, A. J., & Kidd, K. K. (1996). The world-wide distribution of allele frequencies at the hu- man dopamine D4 receptor locus. Human Genetics, 98,91–101. doi: 10.1007/s004390050166 . Creemers, H. E., Harakeh, Z., Dick, D. M., Meyers, J., Vollebergh, W. A. M., Ormel, J., et al. (2011). DRD2 and DRD4 in relation to regular alcohol and cannabis use among adolescents: does parenting modify the impact of genetic vulnerability? The TRAILS study. Drug and Alcohol Dependence, 115 ,35–42. doi: 10.1016/j.drugalcdep.2010.10.008 . J Abnorm Child Psychol (2018) 46:73 –82 81 De Winter, A. F., Oldehinkel, A. J., Veenstra, R., Brunnekreef, J. A.,Verhulst, F. C., & Ormel, J. (2005). Evaluation of non-response bias in mental health determinants and outcomes in a large sample of pre- adolescents. European Journal of Epidemiology, 20 , 173–181. doi: 10.1007/s10654-004-4948-6 . DiLalla, L. F., Bersted, K., & John, S. G. (2015). Peer victimization and DRD4 genotype influence problem behaviors in young children. Journal of Youth and Adolescence, 44 , 1478–1493. doi: 10.1007 /s10964-015-0282-4 . Dmitrieva, J., Chen, C., Greenberger, E., Ogunseitan, O., & Ding, Y. C. (2011). Gender-specific expression of the DRD4 gene on adolescent delinquency, anger and thrill seeking. Social Cognitive and Affective Neuroscience, 6 ,82–89. doi: 10.1093/scan/nsq020 . Ellis, B. J., Boyce, W. T., Belsky, J., Bakermans-Kranenburg, M. J., & Van IJzendoorn, M. H. (2011). Differentia l susceptibility to the environment: an evolutionary-neurodevelopmental theory. Development and Psychopathology, 23 ,7–28. doi: 10.1017/S0954579410000611 . Ferguson, C. J. (2015). Do angry birds make for angry children? A meta- analysis of video game influences on children’s and adolescents’ aggression, mental health, prosocial behavior, and academic perfor- mance. Perspectives on Psychological Science, 10 ,646–666. doi: 10.1177/1745691615592234 . Grant, K. E., Compas, B. E., Thurm, A. E., McMahon, S. D., & Gipson, P. Y. (2004). Stressors and child and adolescent psychopathology: measurement issues and prospective effects. Journal of Clinical Child and Adolescent Psychology, 33 , 412–425. doi: 10.1207 /s15374424jccp3302_23 . Huisman, M., Oldehinkel, A. J., De Winter, A., Minderaa, R. B., De Bildt, A., Huizink, A. C., et al. (2008). Cohort profile: the Dutch ‘TRacking Adolescents ’Individual Lives ’Survey ’; TRAILS. International Journal of Epidemiology, 37,1227 –1235. doi: 10.1093/ije/dym273. Janssens, A., Van den Noortgate, W., Goossens, L., Verschueren, K., Colpin, H., De Laet, S., et al. (2015). Externalizing problem behav- ior in adolescence: dopaminergic genes in interaction with peer ac- ceptance and rejection. Journal of Youth and Adolescence, 44 , 1441 –1456. doi: 10.1007/s10964-015-0304-2 . Jenkins, J. M. (2008). Psychosocial adversity and resilience. In M. Rutter, D. Bishop, D. Pine, S. Scott, J. Stevenson, E. A. Taylor, & A. Thapar (Eds.), Rutter’s handbook of child and adolescent psychiatry (5th ed., pp. 377 –391). Oxford: Blackwell. Keller, M. C. (2014). Gene x environment interaction studies have not properly controlled for potential confounders: the problem and the (simple) solution. Biological Psychiatry, 75 ,18–24. doi: 10.1016/j. bi opsych.2013.09.006 . Kretschmer, T., Dijkstra, J. K., Ormel, J., Verhulst, F. C., & Veenstra, R. (2013). Dopamine receptor D4 gene moderates the effect of positive and negative peer experiences on later delinquency: the TRacking Adolescents ’Individual Lives Survey study. Development and Psychopathology, 25 ,1107–1117. doi: 10.1017/S0954579413000400 . Kwok, O. M., Underhill, A. T., Berry, J. W., Luo, W., Elliott, T. R., & Yoon, M. (2008). Analyzing longitudinal data with multilevel models: an example with individuals living with lower extremity intra-articular fractures. Rehabilitation Psychology, 53 ,370–386. doi: 10.1037/a0012765 . Marsman, R., Oldehinkel, A. J., Ormel, J., & Buitelaar, J. K. (2013). The dopamine receptor D4 gene and familial loading interact with per- ceived parenting in predicting externalizing behavior problems in early adolescence: the TRacking Adolescents’ individual lives sur- vey (TRAILS). Psychiatry Research, 209 ,66–73. doi: 10.1016/j. psychres.2012.10.022 . Miller, S. A., Dykes, D. D., & Polesky, H. F. (1988). A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Research, 16 ,1215–1215. doi: 10.1093/nar/16.3.1215 . Nederhof, E., Belsky, J., Ormel, J., & Oldehinkel, A. J. (2012a). Effects of divorce on Dutch boys ’and girls ’externalizing behavior in gene x environment perspective: diathesis stress or differential susceptibility in the Dutch TRacking Adolescents ’Individual Lives Survey study? Development and Psychopathology, 24 ,929– 939. doi: 10.1017/S0954579412000454 . Nederhof, E., Jorg, F., Raven, D., Veenstra, R., Verhulst, F. C., Ormel, J., & Oldehinkel, A. J. (2012b). Benefits of extensive recruitment effort persist during follow-ups and are consistent across age group and survey method. The TRAILS study. BMC Medical Research Methodology, 12 , 93. doi:10.1186/1471-2288-12-93 . Nikitopoulos, J., Zohsel, K., Blomeyer, D., Buchmann, A. F., Schmid, B., Jennen-Steinmetz, C., et al. (2014). Are infants differentially sensi- tive to parenting? Early maternal care, DRD4 genotype and exter- nalizing behavior during adolescence. Journal of Psychiatric Research, 59 ,53–59. doi: 10.1016/j.jpsychires.2014.08.012 . Oldehinkel, A. J., Verhulst, F. C., & Ormel, J. (2008). Low heart rate: a marker of stress resilience. The TRAILS study. Biological Psychiatry, 63 , 1141–1146. doi: 10.1016/j.biopsych.2007.12.006 . Ormel,J.,Oldehinkel,A.J.,Sijtsema,J.,VanOort,F.,Raven,D., Veenstra, R., et al. (2012). The TRacking Adolescents ’Individual Lives Survey (TRAILS): design, current status, and selected find- ings. Journal of the American Academy of Child and Adolescent Psychiatry, 51 ,1020–1036. doi: 10 .1016/j.jaac.2012.08.004 . Preacher,K.J.,Curran,P.J.,&Bauer,D.J.(2006).Computationaltoolsfor probing interactions in multiple linear regression, multilevel model- ing, and latent curve analysis. Journal of Educational and Behavioral Statistics, 31 , 437–448. doi: 10.3102/10769986031004437 . Propper, C., Willoughby, M., Halpern, C. T., Carbone, M. A., & Cox, M. (2007). Parenting quality, DRD4, and the prediction of externalizing and internalizing behaviors in early childhood. Developmental Psychobiology, 49 ,619–632. doi: 10.1002/dev.20249 . Ptacek, R., Kuzelova, H., & Stefano, G. B. (2011). Dopamine D4 receptor gene DRD4 and its association with psychiatric disorders. Medical Science Monitor, 17 , RA215–RA220. Richards, J. S., Hartman, C. A., Franke, B., Hoekstra, P. J., Heslenfeld, D. J., Oosterlaan, J., et al. (2015). Differential susceptibility to maternal expressed emotion in children with ADHD and their siblings? Investigating plasticity genes, prosocial and antisocial behaviour. European Child & Adolescent Psychiatry, 24 , 209–217. doi: 10.1007/s00787-014-0567-2 . Rutter, M. (2005). Environmentally mediated risks for psychopathology: research strategies and findings. Journal of the American Academy of Child and Adolescent Psychiatry, 44 ,3–18. doi: 10.1097/01. chi.0000145374.45992.c9 . Savage, J., & Yancey, C. (2008). The effects of media violence exposure on criminal aggression – a meta-analysis. Criminal Justice and Behavior, 35 ,772–791. doi: 10.1177/0093854808316487 . Sonuga-Barke, E. J. S., Oades, R. D., Psychogiou, L., Chen, W., Franke, B., Buitelaar, J., et al. (2009). Dopamine and serotonin transporter genotypes moderate sensitivity to maternal expressed emotion: the case of conduct and emotional problems in attention deficit/ hyperactivity disorder. Journal of Child Psychology and Psychiatry, 50 ,1052–1063. doi: 10.1111/j.1469-7610.2009.02095.x . Verhulst, F. C., & Van der Ende, J. (2013). Handleiding ASEBA. Vragenlijsten voor leeftijden 6 tot en met 18 jaar [ASEBA manual: questionnaires for ages 6 through 18 years] . Rotterdam: ASEBA Nederland. Windhorst, D. A., Mileva-Seitz, V. R., Linting, M., Hofman, A., Jaddoe, V. W. V., Verhulst, F. C., et al. (2015). Differential susceptibility in a developmental perspective: DRD4 and maternal sensitivity predicting externalizing behavior. Developmental Psychobiology, 57 ,35 –49. doi: 10.1002/dev.21257 . Zandstra, A. R. E., Hartman, C. A., Nederhof, E., Van den Heuvel, E. R., Dietrich, A., Hoekstra, P. J., & Ormel, J. (2015). Chronic stress and adolescents ’mental health: modifying effects of basal cortisol and parental psychiatric history. The TRAILS study. Journal of Abnormal Child Psychology, 43 ,1119–1130. doi: 10.1007/s10802-014-9970-x . Zuckerman, M. (1999). Vulnerability to psychopathology: a biosocial model . Washington DC: American Psychological Association. 82 J Abnorm Child Psychol (2018) 46:73–82 Journal ofAbnormal ChildPsychology isacopyright ofSpringer, 2018.AllRights Reserved.
Interviewing Presentation Assignment Instructions Overview As we have learned in this course, the clinical interview is a core component of a psychological assessment or evaluation. For this Inte
595 Original article How events in emergency medicine impact doctors’ psychological well-being Laura Howard, 1,2 Christopher Wibberley, 2 Liz Crowe, 3,4 Richard Body 1,5 To cite: Howard L, Wibberley C, Crowe L, et al. Emerg Med J 2018;35:595–599. 1Emergency Department, Manchester University Foundation NHS Trust, Manchester Academic Health Science Centre, Manchester, UK 2Faculty of Health, Psychology and Social Care, Manchester Metropolitan University, Manchester, UK 3School of Medicine, The University of Queensland, Herston, Queensland, Australia 4Paediatric Intensive Care Unit, Lady Cilento Children’s Hospital, South Brisbane, Queensland, Australia 5Cardiovascular Sciences Research Group, The Innovation Centre, The University of Manchester, Manchester, UK Correspondence to Dr Laura Howard, Emergency Department, Manchester University Foundation NHS Trust, Manchester M13 9WL, UK; [email protected] doctors. org. uk Received 3 October 2017 Revised 29 July 2018 Accepted 10 August 2018 Published Online First 21 August 2018 © Author(s) (or their employer(s)) 2018. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ. Abs TrACT b ackground Emergency medicine is a high-pressured specialty with exposure to disturbing events and risk. We conducted a qualitative study to identify which clinical events resulted in emotional disruption and the impact of these events on the well-being of physicians working in an ED. Methods W e used the principles of naturalistic inquiry to conduct narrative interviews with physicians working in the ED at Central Manchester University Hospitals NHS Foundation Trust, between September and October 2016. Participants were asked, ’Could you tell me about a time when an event at work has continued to play on your mind after the shift in which it occurred was over?’ Data were analysed using framework analysis. The study had three a priori themes reported here. Other emergent themes were analysed separately. r esults W e interviewed 17 participants. Within the first a priori theme (’clinical events’) factors associated with emotional disruption included young or traumatic deaths, patients or situations that physicians could relate to, witnessing the impact of death on relatives, the burden of responsibility (including medical error) and conflict in the workplace. Under theme 2 (psychological and physical effects), participants reported substantial upset leading to substance misuse, sleep disruption and neglecting their own physical needs through preoccupation with caring. Within theme 3 (impact on relationships), many interviewees described becoming withdrawn from personal relationships following clinical events, while others described feeling isolated because friends and family were non-medical. Conclusions Clinical events encountered in the ED can affect a physician’ s psychological and physical well- being. For many participants these effects were negative and long lasting. b AC kgr O und Emergency physicians (EP) work in a fast-paced and high-pressured environment, with constantly changing teams. The EP must be prepared to manage and process whatever occurs on a shift. This includes the potential of exposure to death, serious illness, trauma and suffering, often with no clear solution to the problems they face. 1 This environment can lead to emotional disruption in the physician, which may be short or long term in effect. Previous research shows that EPs are at high risk of compassion fatigue, secondary trauma and burnout, and this may change behaviour in the workplace or at home. 2 3 Despite growing literature reporting the potential stresses on EPs, there are no known qualitative studies exploring the meaning of emotionally disruptive events to EPs. There has been little discussion of the nature and effect of these events as reported by the physi- cians themselves. Qualitative research can provide a rich and deep insight into the psychological and physical disruption caused by events in the ED and could allow us to create meaningful steps to developing targeted interventions to protect the well-being of EPs. The aim of this study was to explore the impact of clinical events on the well- being of EPs, and to identify which events result in emotional disruption. Me ThOds Our study design was founded in the principles of naturalistic inquiry. 4–6 The research design drew on five elements of such enquiry listed by Lincoln and Guba 4: 1. It used tacit knowledge gained by working in the setting in which the research was carried out. 2. A qualitative method was adopted. 3. Inductive data analysis was applied. 4. The results reflected a case study reporting mode. 5. The implications were considered tentative, es – pecially with respect to transferability. key messages What is already known on this subject ► The prevalence of compassion fatigue, secondary trauma and burnout has been studied in emergency physicians, but there is little research on how different clinical events impact on doctors psychologically, physically and emotionally. What this study adds ► In this qualitative research study of 17 emergency physicians we begin to understand how clinical events in emergency medicine have potential to impact on doctors’ well-being in terms of individual psychological and physical effects, and effects on relationships. ► This study highlights the need for intervention after clinical events in the ED and further research into how best to support a physician during these times. on December 9, 2021 by guest. Protected by copyright. http://emj.bmj.com/ Emerg Med J: first published as 10.1136/emermed-2017-207218 on 21 August 2018. Downloaded from Howard L, et al. emerg Med J 2018;35:595–599. doi:10.1136/emermed-2017-207218 596 Original article setting and participants Participants were recruited from Central Manchester Univer- sity Hospitals NHS Foundation Trust ED. Written information sheet and given an opportunity to ‘opt in’ to the study were distributed around the department, during key times such as morning handover, junior doctor teaching and consultant meet- ings over a week’s period in September 2016. There were 69 potential participants comprising 26 consultants, 8 non-training registrars, 11 non-training senior house officers, 5 training regis- trars, 15 foundation doctors and 4 general practitioner trainees. P articipants were made aware that: selection from those ‘opting in’ may occur to ensure representation of doctors at all levels of training, and that selection would continue until data satura- tion. 7 8 Written consent was obtained from all participants prior to the interviews and it was confirmed with all participants that they could withdraw at any time. All participants were informed that our staff support team were available for anyone experi- encing distress following the interview. data collection Narrative style interviews were conducted with EPs using the prompt: ‘Could you tell me about a time when an event at work has continued to play on your mind after the shift in which it occurred was over?’ Follow-up questions were then based on the information provided by the interviewee. All participants were interviewed by a single investigator (LH) who was also an EP, working within the organisation. Interviews were audio recorded and transcribed verbatim for analysis. Pseudonyms were used in all transcripts and are used throughout this report. Interview recordings and transcripts were stored on an encrypted, pass- word-protected hard drive. data analysis Data analysis was carried out using a process similar to that of framework analysis, allowing the inclusion of both a priori and emergent themes. 9 This process involved data being read and re-read and emerging themes noted alongside a priori themes using a constant comparative approach. Phases of familiarisation and indexing led to the development of a thematic framework, discussed and agreed between the authors. This framework was used for data extraction, mapping and interpretation. A priori themes arising from the aims were ‘the clinical events that resulted in emotional disruption’; ‘the psychological and physical effect of these events’; and ‘work events intruding on personal relationships’. Emergent themes (not reported here) were ‘positive experiences within the ED’; ‘cultural influences on experiences’; and ‘professional help and support’. This pape r focuses on the a priori themes (see figure 1): a second paper will focus on emergent themes. resul T s In total, 17 physicians opted in, and all were interviewed. The authors agreed that data saturation had been reached after 15 interviews. Two further interviews had been scheduled and were therefore completed Of the 17 physicians recruited to the study, there were 10 men and seven women, with various levels of experience ( table 1). Clinical events We identified several factors relating to clinical events that seemed to cause emotional disruption, often for many years after the event. Participants could recall the clinical encounters in vivid detail and clearly reported the compounding factors that they believed made these events memorable. Young or traumatic deaths Several participants reported being affected by witnessing patients dying at either a young age or a perceived young age, or in a traumatic manner. The detail of these memories remained vivid to participants. Some of these included doctors recalling events where highly invasive procedures had been carried out on young babies and children and the child had still died. Other Figure 1 A priori themes and subthemes . Table 1 Grades of participating doctors grade of doctor P articipants (n) Consultant 7 Training registrar 3 Non-training registrar level 2 Non-training senior house officer level 3 Foundation doctor 2 on December 9, 2021 by guest. Protected by copyright. http://emj.bmj.com/ Emerg Med J: first published as 10.1136/emermed-2017-207218 on 21 August 2018. Downloaded from Howard L, et al. emerg Med J 2018;35:595–599. doi:10.1136/emermed-2017-207218 597 Original article confronting events included a young woman who had been sexually assaulted and left to die, where entire families had died leaving only one surviving member and having to tell a pregnant mother and young child that their husband/father had died.Our data demonstrated that the term ‘young’ was used broadly to describe children and ‘younger’ adults up to 50 years of age. There was a repeated theme among participants of the tragedy for adults who had died ‘before their time’. For example: I just saw a guy with heart failure, he was a young guy and he was waiting for a heart transplant… I went to see him the next day and he had died. He had just arrested in the night…it’s just so sad th at someone would die so young, so quickly and so tragically. Events EPs can relate to their own lives When participants were able to identify with the patient on any level, it appeared to have a powerful impact on them. Dr C reported being upset by an event that could relate to their own child, saying: Things can be very upsetting. Particularly I remember, a couple of times, you know having a child in… who was the same age as my child… who had been killed in a road traffic accident. I found that really hard because he was a similar size and age and… if you didn’t look at his face you would assume it was the same. Similarly, Dr B reported the negative impact of treating a ‘Young guy that was killed on his bike, because I like cycling, I cycle to work, I cycle all the time….’ Bearing witness to the consequences of death on relatives Participants reported bearing witness to the pain and distress of the surviving family and friends as more distressing than being involved in the death of a patient. Dr H reported that witnessing family bereavement is ‘the thing [they find] hardest’ and described being affected by a ‘little boy, only 3 or 4, saying goodbye to his daddy.’ Dr N noted the negative impact of witnessing sudden and unexpected death, saying, ‘Now that fella is there whose wife just had a cough. He is on his own now for the rest of his life… I think about that more than actual death.’ The burden of responsibility Participants spoke of the pervasive responsibility and fear of making a mistake at work that disturbed their personal life. Some participants expressed overt regret following a scenario, which they perceived as their own mistake. Dr R said: I shouldn’t have sent that patient home or I should have got them to come and see them coz I think regardless of whether you have written down on a bit of paper and discussed with so and so, it is still your patient and you have to take responsibility. Some had witnessed a negative outcome for a patient and worried that it may have been their fault, even though there was no evidence of a personal error. Doctors worried that if it had been a different physician or they had done something differ – ently the outcome could have changed. Dr R said, ‘I came in the next day and unfortunately, he had died in the department, and that was quite a big thing to happen. Obviously, I just thought I had done something wrong.’ While several physicians spoke about specific scenarios, for other participants there was a cumulative impact of the burden of taking responsibility within emergency medicine (EM). Fear of hearing of an error or missing a diagnosis resulted in constant reflection and fear for many participants. Dr R said, ‘I am going to go home at the end of a shift and I am going to worry and I am going to feel like… loose ends haven’t been tied up.’ Dr O said, ‘I was dreaming about patients and thinking about… have I checked her bloods? What if her potassium is high? What if her amylase is this? I used to phone people in the middle of the night.’ It was also apparent that legal proceedings and coronial cases created stress for participants. Dr E described profound effects from a ‘big critical incident type thing and lots of enquiries and there was going to be an inquest but it took three and half years to go to inquest.’ Dr N reported that one of the many difficulties regarding the legal system was EPs ‘didn’t have control of any of the process.’ Conflict in the workplace Encountering conflict in the workplace was a source of emotional distress for a small number of participants. Dr G described how ‘conversations with other specialties at times can be fraught and there are certain times where that has had a real impact on me.’ Psychological and physical symptoms of events A range of both psychological and physical effects, along with sleep disruption, were experienced by those reporting emotion- ally disruptive events within the ED. Psychological symptoms The emotional responses described by participants varied from emotional distress, crying, active avoidance of emotions, a reduc- tion of self-confidence and becoming withdrawn. Dr H recalled, ‘Driving to work in tears every day, driving home in tears every day.’ Dr E reported, ‘Whenever I started thinking about it, it was a weight just you know, it was just there all the time.’ These emotional responses had a significant effect on many partici- pants. Dr L stated that these events caused him to ‘just get to the lowest point you can possibly get to, or the lowest point where you could not function.’ The participants reported how the emotions of these cases had created a reduction in self-confidence and self-esteem. Dr H said, ‘It shook my confidence at work and a bit at home. That made me really question myself.’ Similarly, Dr J said, ‘It had a massive impact on my kind of self-esteem, my self-confidence… how worthwhile I felt I was… and I got kind of in a really bad way.’ Dr O told how colleagues had commented that their ‘spark had gone out.’ Participants described actively trying to block feelings related to negative emotions. Dr L said, ‘Drink, you know, use other substances, you know they are not healthy things to manage the feelings that you have but they are things that we do as a way of getting away from real life or work.’ For others, such as Dr O, emotions were avoided by ‘de-personalis[ing] a lot of it… like… they are humans and… they… a lot of the time it’s like they are broken machines that need fixing.’ Physical symptoms Participants reported experiencing physical symptoms following an event. For example, Dr Q reported, ‘Weight loss… not eating.’ This physical manifestation of stress or anxiety was also described by Dr O who said that they had ‘not got space for anything else including your own physical needs. You are just so weighed down by someone’s amylase and someone’s lactate and someone’s drain fluid that you’re not even thinking about… hang on a second I haven’t eaten.’ on December 9, 2021 by guest. Protected by copyright. http://emj.bmj.com/ Emerg Med J: first published as 10.1136/emermed-2017-207218 on 21 August 2018. Downloaded from Howard L, et al. emerg Med J 2018;35:595–599. doi:10.1136/emermed-2017-207218 598 Original article Sleep disruption Participants described how events at work had a negative effect on their sleep. Disruption and concerns about sleep hygiene, sleep routine and patterns were mentioned in the majority of interviews. Dr J said, ‘At 3 o’clock in the morning I would be lying awake going over every patient that I had seen that night or that day, and I would be second guessing myself, I would be stressing myself out, worrying myself sick.’ Others, such as Dr B, would awake in the middle of the night with intrusive thoughts: ‘Sometimes, all of a sudden, you wake up in a cold sweat thinking about things, worrying about things, things you might have missed.’ For some sleep was disrupted for long periods. Dr E said, ‘I certainly didn’t sleep properly for three and a half years. I thought about it probably every day, except for a few days on holiday, and looking back it had a huge impact on my life.’ Work events intruding on personal relationships Participants reported that work events regularly had a nega- tive impact on their personal relationships. These relationships also seemed to affect the way they processed the event that had occurred. Non-medical friends and family unable to identify and support EPs Participants who had significant others working outside the healthcare community identified this as isolating. These partic- ipants felt they were unable to share their work and fears. Dr A stated: None of the people in my life outside of work have any kind of frame of reference for any of this. This is all really weird and they don’t really know what to say and it’s making everybody else feel very uncomfortable so I kind of need to package this up and deal with it in a way that I can deal with it, rather than expecting other people to be able to. This inability to share with non-medical loved ones was described as creating a source of conflict in relationships when EPs had to process work events alone. Dr L stated that this ‘created a friction I guess at home.’ Similarly, Dr K reported, ‘I become a little bit short tempered.’ Participants with non-medical partners felt they should not share events at work with their loved ones. This is explained by Dr F, ‘Because of the nature of emergency medicine we have, we see, we do, we feel, we experience stuff which other people shouldn’t, so I don’t impose those on people who are not in this particular club. In fact, I don’t impose them on anybody.’ Becoming withdrawn In order to process events at work many interviewees described becoming withdrawn. Dr M said, ‘I stew on things, I just go very, very indrawn.’ This was an experience not just felt in the home but socially as well. For example, Dr H said, ‘When something nasty has happened at work you would notice it at home by me being quieter. Less likely to be going out with friends.’ Dr J said, ‘I was becoming more withdrawn and I started to feel anxious at social events and gatherings amongst my closest friends and family. I had to kind of retreat in.’ disCussi On There is increasing evidence of the risk of potential harm to EPs in relation to their work. Burnout and trainee retention are recognised problems within EM, which may be a result of emotional disruption. 10 However, to our knowledge this work is the first to use qualitative methods to allow an in-depth exploration of the impact of events on the well-being of EPs. Our findings demonstrate that events occurring at work have a profound impact on EPs. This may occur as a reaction to routine events, rather than being restricted to those that are particularly traumatic or related to medical error. A wide range of events triggered distress in EPs, which fitted into the broad themes presented. How EPs will respond to events is difficult to predict, as they are deeply personal to them and their circumstances. Cases at work affect EPs both emotionally and physically and extend to disruption of personal relationships. Our participants often reported experiencing difficulty with sleeping due to intru- sive thoughts about events at work. These effects can last for long periods of time, for some several years. These events often affected physicians’ relationships with their family and friends. Several participants reported becoming more withdrawn, feeling unable to talk to friends and relatives (particularly if they were from non-medical backgrounds) and reported that this could cause conflict and frustration at home. Reported negative effects, following cases that continued to intrude on the well-being of EPs, were not associated with the level of experience. Concerning and long-lasting impacts of these events were described equally among participants. The results exemplified that you cannot predict the impact of a particular case across a cohort of participants. What may be an ‘ordinary’ case with no physical or emotional impact for one EP may have a profound and particularly damaging outcome for another EP. The cases that one physician may identify with may not have the same resonance with another physician. Additionally, the burden of responsibility felt by one physician may not be carried by another. These results indicate the need for physicians to have a deep sense of self-awareness to their own reactions to cases and a strong investment in their own well-being. Leaders in EM ought to provide a multifaceted approach to EPs’ well-being, and be able to account for these individual responses and requirements, in order to facilitate recovery and resilience. strengths and limitations All participants opted into this study after reading a participant information leaflet and were therefore a self-selected group who were eager to share their narrative, creating an inherently biased group. EPs who remained resilient irrespective of the clinical case or those who felt deeply shamed or impacted by a case may have opted not to participate. All the participants spoke of past events in the interviews, which may mean the study misses an important group of physicians who were being affected by events at the time of data collection. To maintain participant confidentiality, the age, gender and years of clinical experience of participants have not been disclosed in the results. While this was a single- centre study, the training doctors interviewed and consultants (who had all been through EM training) were often recounting events that occurred while working in EDs in different hospitals. In this study, 41% (7 of 17) of participants were consultants in EM, meaning they had many years of experience in this field. Due to this, they may have developed mechanisms for coping with events they are exposed to in EM. Conversely, sustained exposure to such events may have resulted in a greater negative impact on their well-being, when emotionally disruptive cases were encountered. All of the interviews were conducted by LH, who is an EP. This allowed for consistency in the style and conduct of the interviews. As LH is an EP, interviewees may have felt prepared to share experiences more openly, removing the obstacle of the interviewees feeling that they would have to explain the medicine on December 9, 2021 by guest. Protected by copyright. http://emj.bmj.com/ Emerg Med J: first published as 10.1136/emermed-2017-207218 on 21 August 2018. Downloaded from Howard L, et al. emerg Med J 2018;35:595–599. doi:10.1136/emermed-2017-207218 599 Original article as well as their feelings. Conversely, as LH was a colleague of the staff the responses may have been more guarded, reducing openness of interviewees. If an investigator with a non-med- ical background had conducted the interviews, they may have needed to ask more questions about the cases in order to gain an understanding that was assumed by LH. These advantages and disadvantages parallel those discussed in the literature relating to insider-outsider research. 11 implications This study demonstrates that events at work may affect a physi- cian, and in some cases, have a negative impact on all aspects of their lives; emotionally, physically and on personal relationships. The research brings up many questions, mainly about how we can help physicians when these events do happen. Our findings lead us to ask whether we require a culture change in EM, so that physicians have a safe environment to reflect and share concerns with a goal to minimising long-term impact. They should also lead us to ask specific questions about matters that seemed to particularly affect EPs. For example, we should ask how we best provide support to EPs undergoing legal proceedings such as coroner’s cases, and how we can help physicians process the responsibility they carry. COnClusi On This research gives insight into how the cases physicians can encounter in the ED can affect a physician’s psychological and physical well-being. In some interviews, these effects are deeply moving showing the need for further research in EM well-being, especially researching tools to increase resilience. Contributors LH designed the research study, collected the data, analysed the results and led on writing the manuscript. CW and RB were involved in input and advice on research design, analysed the results and critically reviewed the manuscript. LC was involved in input on research design and critically reviewed the manuscript. Funding T he authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors. Competing interests None declared. P atient consent Not required. e thics approval Health Research Authority (IRAS ID 209236) and Manchester Metropolitan University Ethics Committee. Provenance and peer review Not commissioned; externally peer reviewed. d ata sharing statement Unpublished data from this study (the emergent themes) will be presented in a second paper which will be submitted for consider ation of publication. Open access T his is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http:// creativecommons. org/ licenses/ by- nc/ 4. 0/. Refe Rences 1 Crowe L. Identifying the risk of compassion fatigue, improving compassion satisfaction and building resilience in emergency medicine. Emerg Med Australas 2016;28:106–8. 2 Br agard I, Dupuis G, Fleet R. Quality of work life, burnout, and stress in emergency department physicians: a qualitative review. Eur J Emerg Med 2015;22:227–34. 3 Dasan S , Gohil P, Cornelius V, et al. Prevalence, causes and consequences of compassion satisfaction and compassion fatigue in emergency care: a mixed-methods study of UK NHS Consultants. Emerg Med J 2015;32:588–94. 4 Lincoln YS, Guba EG. Naturalistic inquiry: Sage, 1985. 5 Gubrium JF , Holstein JA. The new language of qualitative method. Oxford: Oxford University Press, 1997. 6 Bowen GA. Naturalistic inquiry and the saturation concept: a research note. Qualitative Research 2008;8:137–52. 7 Nelson J . Using conceptual depth criteria: addressing the challenge of reaching saturation in qualitative research. Qualitative Research. doi: 10.1177/1468794116679873. [Epub ahead of print 14 Dec 2016]. 8 Ritchie J , Spencer L. Qualitative data analysis for applied policy research. The qualitative researcher’s companion. 2002;573:305–29. 9 Spencer L, Ritchie J, O’Connor W. Analysis: practices, principles and processes. Qualitative research practice: a guide for social science students and researchers, 2003:199–218. 10 Clifford M, Ian C, 2013. Emergency Medicine- background to HEE proposals to address workforce shortages. https:// hee. nhs. uk/ sites/ default/ files/ documents/ Emergency% 20medicine-% 20background% 20to% 20HEE% 20proposals% 20to% 20address% 20workforce% 20shortages. pdf 11 Dwyer SC , Buckle JL. The space between: on being an insider-outsider in qualitative research. Int J Qual Methods 2009;8:54–63. on December 9, 2021 by guest. Protected by copyright. http://emj.bmj.com/ Emerg Med J: first published as 10.1136/emermed-2017-207218 on 21 August 2018. Downloaded from Howard L, et al. emerg Med J 2018;35:595–599. doi:10.1136/emermed-2017-207218