SHEVLIN, READ J. 2015.Pdf
Total Page:16
File Type:pdf, Size:1020Kb
Journal of Criminal Psychology Testing complex hypotheses using secondary data analysis: is the association between sexual abuse and psychosis moderated by gender in a large prison sample? Mark Shevlin Jamie Murphy John Read Article information: To cite this document: Mark Shevlin Jamie Murphy John Read , (2015),"Testing complex hypotheses using secondary data analysis: is the association between sexual abuse and psychosis moderated by gender in a large prison sample?", Journal of Criminal Psychology, Vol. 5 Iss 2 pp. - Permanent link to this document: http://dx.doi.org/10.1108/JCP-02-2015-0009 Downloaded on: 26 May 2015, At: 18:29 (PT) References: this document contains references to 0 other documents. To copy this document: [email protected] Access to this document was granted through an Emerald subscription provided by 215423 [] For Authors If you would like to write for this, or any other Emerald publication, then please use our Emerald for Authors service information about how to choose which publication to write for and submission guidelines are available for all. Please visit www.emeraldinsight.com/authors for more information. About Emerald www.emeraldinsight.com Emerald is a global publisher linking research and practice to the benefit of society. The company manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as well as providing an extensive range of online products and additional customer resources and services. Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive preservation. *Related content and download information correct at time of download. Downloaded by Swinburne University of Technology At 18:29 26 May 2015 (PT) Testing complex hypotheses using secondary data analysis: is the association between sexual abuse and psychosis moderated by gender in a large prison sample? Prospective (Janssen et al. 2004), data linkage (Cutajar et al, 2010; Elklit Shevlin, 2011) general population, (Bebbington et al, 2004; Shevlin, Dorahy Adamson, 2007) and clinical studies (Lysaker, Davis, Gatton, ,erman, 2005) have repeatedly identi.ied links between trauma and psychosis. 0echanisms by which these e1periences may be linked have also been postulated (2ead, Perry, 0oskowit3, Connolly, 2001; Bentall 4ernyhough, 2008). To date, while most studies have estimated the relationship in samples o. males and .emales, or .emales only, .ew studies have e1plicitly tested i. the trauma7psychosis relationship is moderated by gender. Those studies that have reported gender di..erences have o..ered con.licting .indings. Shevlin, Dorahy and Adamson (2007) .or e1ample, reported a signi.icant rape by gender interaction in predicting psychosis among members o. the general population in the 8nited States o. America (National Comorbidity Survey), with Downloaded by Swinburne University of Technology At 18:29 26 May 2015 (PT) higher odds ratios .or males compared to .emales. Bebbington et al. (2011) however suggested that the link between se1ual abuse and psychosis was moderated by se1 among members o. the general population in the 8nited :ingdom (British Psychiatric 0orbidity Survey), and that this association was stronger .or .emales. 0oreover, a recent case control study by 4ischer et al. (2009) also suggested that CSA among .irst episode psychosis su..erers (Aetiology and Ethnicity in Schi3ophrenia and Other Psychoses (=SOP) study) represented a speci.ic threat .or .emales compared to males. ,owever, robust testing o. such moderated e..ects has been di..icult as the prevalence o. psychosis in population studies has been low, and reported rates o. se1ual trauma .or males have also been low. A .ailure to detect moderation may there.ore be due to a lack o. statistical power. While general population data may not a..ord suitably proportioned prevalence o. both phenomena across the se1es to .acilitate moderation based analyses, alternative population data may. This paper aimed to demonstrate that secondary data analysis can be a use.ul methodological approach to testing comple1 psychological hypotheses. In this case data .rom the Survey o. Psychiatric 0orbidity among Prisoners in England and Wales was used. This is a large scale psychiatric survey and due to the sample characteristics avoided many o. the methodological limitations o. previous studies. In particular the low base rates o. psychosis and male se1ual trauma has resulted in statistical tests o. association that have low power. ,owever previous research has indicated that the prevalence o. psychosis in prison populations is higher than .or the general population (Brugha et al, 2005; Birmingham, 0ason, Grubin, 1996) and that there are higher rates o. childhood and adult traumatic e1periences, in particular child abuse, .or both males and .emales (Carlson Sha.er, 2010; 0alacova et al, Downloaded by Swinburne University of Technology At 18:29 26 May 2015 (PT) 2012). There.ore this study aims to test i. the relationship between se1ual abuse and psychosis is moderated by gender using a large sample o. male and .emale prisoners .rom England and Wales. Method Sample The data .or the current study was based on the Survey o. Psychiatric 0orbidity among Prisoners in England and Wales (Singleton, 0elt3er, Gatward, Coid Deasey, 1998) which was conducted to estimate the prevalence o. mental disorders in the prison population. The sample comprised prisoners .rom all 1A1 prisons in England and Wales. Participants were randomly selected .rom three groups using sampling .ractions (male remand 1 in 8, male sentenced 1 in A4, and .emale remand and sentenced prisoners 1 in A). A total o. A,56A prisoners were recruited and A,142 (88B) completed the survey o. whom 75.5B were male. 0ale prisoners and .emale prisoners had the same modal age group (25729 years). Slightly over hal. o. the sample (54.AB) was sentenced with the remainder on remand, and the percentage o. males (65.6B) who were sentenced was higher than .emales (A4.ABC χ2D 189.96, d. D 1, p E .001). 59.2B o. the sample had been in prison be.ore, and this was more common .or males (χ2D 257.28, d. D 1, p E .001). There were more .emales (25.0B) than males (21.6B) who had been in prison .or more than 1 year (χ2D 8.64, d. D 1, p E .05). Details o. the sample and survey methods can be .ound in Singleton et al (1998). Psychosis and Sexual Victimization A random subsample o. 1 in 5 participants were interviewed with the Schedule .or Assessment in Neuropsychiatry (World ,ealth Organi3ation, 1999) by psychiatrists Downloaded by Swinburne University of Technology At 18:29 26 May 2015 (PT) or psychologists .or assessment o. a psychotic disorder. 4rom this data an algorithm was used to identi.y those .actors which were predictive o. a diagnosis o. Fprobable psychosisG .or the non7selected participants based on the .ollowing criteriaC (1) sel.7 reported use o. anti7psychotic medication, (2) admissions to hospital .or mental health reasons, (A) sel.7reported diagnosis or symptoms o. psychosis and (4) endorsement o. the probe and secondary item relating to auditory hallucinations in the Psychosis Screening Questionnaire (Bebbington Nayani, 1995). This combination o. responses was .ound to be predictive o. the presence o. psychosis in those also assessed with the SCAN. Psychosis presence in the current analysis was coded as F1G and psychosis absence was coded as F0G. The Iuestion relating to se1ual victimi3ation was taken .rom the F:ey Li.e Events and Post Traumatic StressG section o. the Iuestionnaire. Participants were presented with cards with lists o. stress.ul li.e e1periences and were asked, JLooking at the card, could you tell me i. you have ever su..ered .rom any o. the problems or events shown on the cardK. The .irst card included JSe1ual abuseK and participant responses to this were coded FLesG (1) or FNoG (0). Control variables A range o. demographic variables, potential con.ounding variables, and clinical variables were also entered in the analysis as control variables. These wereC Gender: 4emale (0), 0ale (1). AgeC Age was collapsed into 7 categoriesC (1) 16720, (2) 21724, (A) 25729, (4) A07A4, (5) A57A9, (6) 40744, (7) 45M years. Downloaded by Swinburne University of Technology At 18:29 26 May 2015 (PT) Living statusC Participants were asked JJust be.ore you came into prison were you living...K; responses were coded as FAloneG (1) and FOtherG (0). Employment: Participants were asked JJust be.ore you came into prison did you have any kind o. paid workNK and responses were coded FNoG (0) or FLesG (1) . Education: Participants were asked J,ave you got any Iuali.ications o. any sortNK and responses were coded FNoG (0) or FLesG (1) . Immigration: Participants were asked which country they were born in and the responses were coded F8nited :ingdomG (1) or FOtherG (0) . Common mental disorders The Clinical Interview Schedule (CIS72) was used to produce speci.ic ICD710 diagnoses .or si1 FneuroticG disorders. This was recoded to indicate i. each participant had at least one FneuroticG disorder (1) or no FneuroticG disorder (0). Problem drinking Identi.ication o. ha3ardous alcohol use was based on a score greater than 8 on the Alcohol 8se Disorders Identi.ication Test (Saunders et al, 199A). Cannabis use: The FDrug 8seG section o. the Iuestionnaire included Iuestions on the .reIuency o. drug use. A derived variable was J8sed cannabis in year be.ore prisonK and this was coded FLesG (1) or FNoG (0). Other traumas. Participants were presented with a list o. stress.ul li.e events .rom the JList o. Li.e Threatening E1periencesK (Brugha, Bebbington, Tennant, ,urry, 1985). Participants were asked i. they had e1perienced JBullyingK, JOiolence at workK, JOiolence in the homeK, and J,omelessnessK.