This research concerned the social experience of two groups of adults with intellectual disability (mental handicap), those with verbal skills who could describe their experience and those with profound disabilities whose needs were interpreted by carers. Conceptual issues have been discussed to provide a framework for understanding their relationships and also for understanding intellectual disability itself. Social constructionism, has influenced the investigations. Previous research shows that disabled people experience relatively little social contact regardless of their level of ability. Interactions and relationships were investigated so that practitioners aiming to improve intellectually disabled adults' social experience may have relevant information. It has been assumed that more interaction is better as it allows people to reflect upon themselves with regard to others, to construct themselves through their interactions. These investigations fall into two main sections, the first a discussion of attitudes toward self and others held by people who had, or did not have, a peer-group friend. Results suggested that further examination of the role of the self-concept in friendship formation would be fruitful and that people without a peer-group friend were similar to lonely non-disabled adults. The second section examines profound disability and contains a linked series of studies of adults' interactions with key workers. Results suggested that two interaction strategies used by staff were associated with increased positive responding from clients; that the developmental age of clients did not distinguish between them in their social responses to staff; that there was little evidence of turn-taking or the importance of staff making responses which follow the client's lead. Results were discussed in terms of their contribution to an alternative model for understanding profound disability. The project as a whole was discussed in terms of the interaction between method and conceptualisation and concludes with a number of recommendations for practitioners.

It has been reported that there is an over-representation of people with learning disabilities in prison populations. Various explanations have been offered for this including theories that people with disabilities are susceptible to criminal behaviour, and that they receive different treatment within the justice system. There is no evidence of over-representation of people with learning disabilities in the UK penal system although evidence does suggest that there are a significant number of prisoners with borderline learning disabilities who are psychologically vulnerable. This research set out to examine current levels of contact that psychiatrists have with various criminal justice agencies, as well as exploring the attitudes and beliefs of other criminal justice workers in order to assess current arrangements of treatment and care for offenders who have a learning disability. 791 consultant psychiatrists registered in four sections of the Royal College of Psychiatry were surveyed to detect current levels of contact with five criminal justice agencies, and also to assess their levels of involvement with the diversion of mentally disordered offenders into health care provision. Responses indicated that forensic and general psychiatrists had the highest levels of contact with criminal justice agencies, although there was also limited contact with child and learning disability psychiatrists. Forensic psychiatrists were also shown to be the group who were most likely to be involved with diversion schemes, and there was some evidence to suggest that respondents were unhappy with current arrangements for diversion because of bed and staff shortages. An attitude survey concerning offenders with learning disabilities was distributed to 100 criminal justice professionals (judges, magistrates, police and appropriate adults). 28 of these respondents also participated in semi-structured interviews which provided qualitative data to supplement the findings of the attitude survey. These studies provided evidence that criminal justice professionals were not eager to assign special rights to people because of their learning disability status, but instead were confident that the present legal system could accommodate the needs of offenders with learning disabilities in the present system. However, significant concern was raised by all respondents that specific training in needed to address learning disability issues before this can be fully achieved. There was little evidence found to support `susceptibility' theories of learning disability offending, although there was some evidence which gave support to the `different treatment' hypothesis.

Introduction: This thesis explored the feasibility of conducting a randomised control trial (RCT) of Acceptance and Commitment Therapy (ACT) within patient’s Spinal Cord Stimulation (SCS) treatment pathway. Previous ACT literature has reported that this is an effective intervention with persons with chronic pain, even when in a self-help format. Whilst this growing body of ACT literature is promising, chronic pain is a broad term used for a number and varying levels of disability. Chronic neuropathic pain (CNP) is usually more complex and resistant to treatment. According to NICE guidelines SCS should be the last treatment option for CNP sufferers. However, there is currently a dearth of literature exploring the effectiveness of the addition of a psychological interventions with the SCS population. Objectives: The aim of this thesis was to explore the parameters of interest in ACT with SCS patient population including recruitment, acceptability of intervention and treatment signals. Design: A mixed between-within group design with repeated measures. There were three conditions in the study. Methods: Ethical and NHS trust approval was obtained. SCS participants were recruited from one Neuromodulation clinic during their routine appointment. Fourteen SCS patients that consented to the study were randomised to either SCS combined with an ACT self-help intervention (SCS-ACT) or SCS and treatment as usual (SCS-TAU). A third arm of the study was included to gain additional information on the ACT self-help intervention. This arm had been assessed for SCS surgery at the same neuromodulation clinic but deemed to be not suitable for the surgery. They were invited to the study via a letter from the clinic. All participants completed outcome measures pre and post-intervention. The participants in the two ACT arms (SCS-ACT and ACT-only) were given an ACT self-help manual and received telephone support sessions over six consecutive weeks. These participants also completed an interview at the end of the intervention. Results: Recruitment to the study was lower than expected and the majority of the participants in the ACT conditions (77%) did not complete the self-help manual. All the participants in the SCS-ACT condition had reliable improvement on at least two outcome domains, however, due to the small sample size it was not possible to assess whether these improvements were due to the SCS surgery or the ACT intervention. The interviews identified a number of barriers that prevented participants completing the manual. Discussion: The study demonstrated that a number of amendments need to be made to the study design and the self-help manual before a full-scale RCT is justified. Therefore it is recommended that a number of alterations are implemented in another feasibility study to assess whether this improves recruitment, retention and outcomes.

Background. This topic was proposed by the Service User and Carer Advisory Panel (SUCAP) which informs and supports Clinical Psychology training at The University of Nottingham. The project developed due to their concern about the ambiguity of crisis. They suggested that their understandings may be different from that of professionals. The reconfiguration of acute mental health services influenced the nationwide implementation of community treatment alternatives. However, crisis has remained a poorly operationalised concept, which can lead to problems in clinical practice. An increase in empirical efforts has aimed to establish clarity and increase the clinical utility of definitions. Despite these efforts, little is known about how service users and carers understand crisis. As suggested by the SUCAP, it is unclear if there is commonality within understandings of crisis. Due to the multiple stakeholders, exploring understandings of crisis from multiple perspectives could be useful in clinical practice. Aim. This study aimed to use Q-methodology to investigate multiple perspectives of crisis. By including service users, carers and professionals the research aimed to highlight consensus and discrepancy within these perspectives. Method. Twenty-six participants were recruited from a National Health Service crisis team, a third sector crisis service and an independent carer support group. In Phase One, clinical interviews with 16 participants (taken from each group) were conducted. Qualitative analysis aided the process of constructing a Q-set of 78 statements that represented the understandings of a crisis experience. In Phase Two, 13 participants (some of whom took part in Phases One and Two) completed the Q-sort procedure, in which they rated the extent to which each statement was characteristic of crisis. To support analysis and interpretation, demographic information was obtained, and pre/post questions elicited reflection on their completed Q-sorts. An inverted factor analytic method was used to analyse the Q-sort data. Results. The results of each Phase are discussed. The rotated two factor solution accounted for fifty percent of variance in people’s understandings of crisis. Factor one related to the understanding of crisis as an experience which can be defined by changes in awareness and perception of reality. Factor two related to the understanding of crisis as intense emotional pain, which is largely related to fear, anxiety and despair. Contrary to existing definitions of crisis, risk and mental illness were not highlighted as central characteristics of crisis. Different pairs of ‘groups’ (service user/professional; carer/professional; service user/carer) had little impact on the amount of shared understanding. However, understandings of crisis were highly variable between individuals. Furthermore, service users’ understandings were more variable than those of professionals or carers. Discussion. Although the findings offer some support to current theories of crisis, they contrast with the dominating characteristics of crisis represented in policy and empirical literature. This suggests that over deterministic definitions of crisis, that focus on risk and mental illness, are insufficient. These findings have implications for both clinical practice and future research.

Background Alcohol-related deaths and hospital admissions have risen steeply in the United Kingdom and little is known about whether there are opportunities for preventive interventions or opportunities for the earlier detection of alcohol misuse. The purpose of this research was to investigate and describe patterns of healthcare utilisation in the pre-diagnosis period of alcoholic cirrhosis and alcoholic psychosis, two alcohol-related disorders which contribute significantly to alcohol related morbidity and mortality. This information is important as it could greatly facilitate a better understanding of opportunities for the earlier detection of alcohol misuse, and could also potentially encourage and assist practitioners in the timely identification of those who are at high risk of developing significant alcohol-related illnesses. Specific research objectives 1)To estimate rates of primary and secondary healthcare contacts before the diagnosis of alcoholic cirrhosis/psychosis and assess whether this differs from rates in the general population 2)To characterise the healthcare contacts and establish whether such consultations were for other conditions known to be related to alcohol. 3)As a secondary objective, to establish whether prior healthcare contacts led to recognition of alcohol problems and consequent interventions. Methods UK general practice and hospital admissions data were used for this research. A population of patients diagnosed with alcoholic cirrhosis and subsequently, a population of patients with alcoholic psychosis were identified. For each case, 10 patients without the disease who were of similar age and from the same general practice were selected as controls. For each disease group, a case-control study design was used to compare rates of primary care contacts and hospital admissions in cases to that of the control population. Further, the reasons for primary care visits and hospital admissions were assessed to describe the frequency of other alcohol-attributable morbidities in cases and controls, and to assess whether there were any morbidities which were strongly associated with the later development of alcoholic cirrhosis and psychosis. Finally, the proportions of patients who had a record of their alcohol use status or a record of an offer of brief alcohol intervention/referral were estimated. Results Alcoholic cirrhosis 2,479 patients with alcoholic cirrhosis and 24,790 controls matched on age and GP practice were available for analysis. Assessment of healthcare utilisation showed that people with alcoholic cirrhosis had 50% higher rate of GP visits and twice as many hospital admissions compared to the control population. This excess healthcare use was apparent as far as 10 years pre-diagnosis of alcoholic cirrhosis, but greater excesses were seen within 5 years of diagnosis. Cases of alcoholic cirrhosis did have a higher frequency of other conditions that are partly alcohol-attributable, especially diabetes, epileptic seizures, hypertensive diseases, injuries and digestive diseases. Of all these conditions, injuries, seizures and digestive diseases had the strongest association with the development of alcoholic cirrhosis. Overall, a high proportion of patients with alcoholic cirrhosis (84%) had their alcohol use documented on an average of 4 years before diagnosis and 58% of these records indicated harmful/hazardous drinking. However, only 1 in 5 patients (22%) had any record of a brief alcohol intervention or an alcohol referral. Alcoholic psychosis 1,731 patients with alcoholic psychosis and 17,310 matched controls were included in the study. As with alcoholic cirrhosis, there was a substantially higher rate of primary and secondary care use in cases which was apparent up to 10 years prior to disease diagnosis. Cases of alcoholic psychosis also had a higher frequency of partly alcohol attributable morbidities. However, the strongest associations were seen for intentional and unintentional injuries and seizures. 78% of cases of alcoholic psychosis did have their alcohol use recorded, 50% had a harmful/hazardous drinking record but only 1 in 5 (19%) had any intervention record. Conclusion and Implications: This research has shown that patients who misuse alcohol make frequent contacts with both primary and secondary care, suggesting that there are high number of opportunities within these healthcare settings to improve the early identification and treatment of alcohol use. Relatively few patients were identified to have received brief alcohol interventions/referral despite the large proportion of people with documented drinking problems, suggesting potential inadequacy in the treatment of alcohol problems. However, this may also partly be an artefact of recording biases. Considering that alcohol is one of the leading preventable causes of morbidity and mortality in the UK, these findings should encourage healthcare professionals to, at the very least, screen for alcohol misuse in patients presenting with any alcohol-attributable condition, and for policy makers to consider directing more effort towards integrating alcohol screening and brief intervention into routine clinical practice.

Studies of sexual murder are abundant, but an empirically supported understanding of this type of crime and its perpetrators is limited in important ways. The overall aim of this thesis was to work towards a theoretical model recognising that sexual murder is a heterogeneous concept. Also, the thesis aimed to understand sexual murder in the wider context of sexual aggression against adult women. First, a systematic review was conducted to clarify the existing typologies of sexual murder. Three consistently observed subtypes were found. These were labelled sexualised murder, grievance murder, and rape murder. A behaviour that could be used to differentiate groups was post mortem sexual interference (PMSI), indicative of sexualised murder. Subsequently, a study was conducted comparing subgroups of sexual murderers (those who had engaged in PMSI, and those who had not), and rapists, on developmental factors, adult lifestyle, and criminal career. The main finding was that the groups responded differently to early adversity. Difficulties with social and sexual relationships leading to the development of sadistic interests, as opposed to pervasive antisociality, distinguished PMSI sexual murderers from rapists, while non-PMSI sexual murderers were largely similar to rapists. However, although the role of psychopathy in sexual aggression is debated, no group was strongly characterised by psychopathy in the present study. A secondary study was conducted to verify that this finding was not associated with limitations in the assessment tool used (the Psychopathy Checklist: Screening Version [PCL: SV]; Hart, Cox, & Hare, 1995). Good psychometric properties were found. Next, a case study demonstrated the functional heterogeneity of the act of killing under the broadly defined term ‘sexual murder’, and highlighted the importance of this in forensic case formulation and intervention. Finally, drawing on the findings of the present research, a new model of sexual murder is proposed to address some of the issues discussed in this thesis.

Introduction: The construct of mindfulness, a non-judgemental awareness of present moment experience, has been increasingly recognised in recent years as being positively associated with psychological wellbeing. In light of accumulating evidence pointing to an inverse relationship between mindfulness and a wide range of psychological distress outcomes, including depression, mindfulness has been increasingly incorporated into modern psychotherapies and healthcare services, and the importance of psychometrically quantifying the construct of mindfulness has become paramount. One of the most reliable and valid instruments developed for the assessment of different aspects of dispositional mindfulness is the Five Facets Mindfulness Questionnaire (FFMQ; Baer et al., 2006), which measures the dimensions of ‘Nonreact’, ‘Observe’, ‘Actaware’, ‘Describe’ and ‘Nonjudge’. However, the psychometric properties of the FFMQ are yet to be tested in clinically depressed individuals with substantial levels of persistent depression. Moreover, there has been no attempt to date to qualitatively explore the experience of mindfulness in those naïve to mindfulness training in order to further determine its role in the management of depressive symptomatology. Methods: Using a convergent parallel mixed methods design, the present study investigated the psychometric properties of the FFMQ in a sample of 187 adults with moderate to severe, persistent depression recruited from a large National Institute for Health Research funded randomised controlled trial (Morriss et al., 2010). Internal consistency and test retest reliability (at six months) were assessed and construct validity was examined with confirmatory factor analyses (CFA) and by statistically correlating the FFMQ to measures of depression, and mindfulness-related constructs; self-compassion, rumination and experiential avoidance. In addition, using semi-structured interviews, a subset of 20 participants were interviewed to explore their experience of depression and perceived associated changes in dispositional mindfulness. Interview data were analysed using qualitative thematic analysis. Results: Results of psychometric testing supported the internal consistency and test-retest reliability of the FFMQ. CFA indicated that both a correlated and hierarchical model fit the data acceptably, with results slightly favouring the correlated model. Contrary to predictions however, CFA showed that the facet ‘Nonjudge’ did not load onto an overarching factor of mindfulness. ‘Nonjudge’ was further found to show a non-significant correlation with depression and only a weak correlation with experiential avoidance and rumination. Thematic analysis of the qualitative data indicated that participants’ ability to retain a non-judgmental awareness of present moment experience deteriorates with the onset of depressed mood. This seemed to occur automatically and deliberately as a strategy to avoid contact with painful internal and external experiences, hence indicating a self-inflicted process of awareness restriction that appears to be a contributing factor to the maintenance of depression. Conclusions: Contrary to what has been previously understood, albeit with different populations, the factor structure of the FFMQ alters in the face of moderate to severe, persistent depression, with the facet ‘Nonjudge’ no longer forming a component of this construct. Therefore, a four factor model (excluding Nonjudge) is proposed for use in this population. The qualitative data has provided possible explanations for the idiosyncratic behaviour of the facet Nonjudge in people experiencing moderate to severe, persistent depression. Both data sets converge to confirm an inverse relationship between mindfulness and depression. Moreover, the qualitative data suggests that deliberate efforts to restrict facets of mindfulness represents a conscious attempt to manage negative experiences that paradoxically maintains and aggravates depression. Future research following-up participants with substantial levels of depression into remission may shed further light on the role of mindfulness in moderate to severe, persistent depression.

Childhood secondhand smoke (SHS) exposure causes substantial ill health and mortality, and poses a significant economic and social burden. This thesis aimed to explore the prevalence and determinants of smoking in the home after childbirth, and to understand the experience and attitudes of mothers who stop smoking during pregnancy but relapse soon after delivery. In study one, the factors associated with child SHS exposure in the home were systematically reviewed. Parental smoking, low socioeconomic status and being less educated were all frequently and consistently independently associated with children’s SHS exposure in the home. Children whose parents held more negative attitudes towards SHS were less likely to be exposed. In study two, cohort data were analysed to estimate maternal self-reported prevalence of SHS exposure among young infants (≤3 months) of women who smoked just before or during pregnancy, and identify its associated factors. In 471 households, the prevalence of smoking in the home was 16.3% (95% CI 13.2-19.8%), and after multiple imputation controlling for non-response 18.2% (95% CI 14.0-22.5%). Mothers’ smoking ≥11 cigarettes per day were 8.2 times (95% CI 3.4-19.6) more likely to report smoking in the home. Significant associations were also observed for younger age, being of non-White ethnicity, increased deprivation and less negative attitudes towards SHS. In study three, semi-structured interviews with women who quit smoking during pregnancy, but relapsed ≤3 months postpartum were conducted. Central to mothers’ accounts of their smoking behaviours during pregnancy and postpartum was their desire to be a ‘responsible mother’. Mothers described using strategies to protect their infant from SHS exposure, and held strong negative attitudes towards other smoking parents. After relapsing, mothers repositioned themselves as ‘social/occasional’ smokers rather than ‘regular’ smokers. Taken together, these findings can be used to inform the development of future, targeted interventions to prevent or reduce infant and child SHS exposure.

This thesis explores the question, what are the important change processes in the treatment of personality disorder in a male forensic inpatient setting? A number of empirically supported therapies for personality disorder stress the importance of the therapist-client relationship in the change process. Therapist-patient relationships are therefore an important focus in this thesis. However, given the lack of research into change processes in this population, the focus is not limited to the therapist-patient relationship, but also considers other relationships and other aspects of the treatment milieu. A model of change processes for this patient group is developed through a series of studies. The first study, in chapter 3, is a qualitative investigation of patients’ perceptions of the process of change and the factors involved. Twelve patients completed a semi-structured interview and the results were analysed using thematic analysis. The study concludes that the cognitive dissonance between how patients expect to be treated and how they are actually treated is an important factor in motivating them to engage in treatment. It also concludes that the therapist-patient relationship and the wider interpersonal environment are both important to therapeutic change with this population. Chapters 4 and 5 describe the process of developing an appropriate dependent outcome measure for the thesis. Social functioning was selected as the dependent outcome variable. Chapter 4 is a systematic review of social functioning measures used with people with a diagnosis of personality disorder and concludes that there is a need to develop a new self-report measure specifically for people with a diagnosis of personality disorder in inpatient settings. Chapter 5 describes the development and validation of this new measure, the Hospital Social Functioning Questionnaire (HSFQ). Fifty-four patients completed a range of measures including the HSFQ. The HSFQ shows good internal consistency, test-retest reliability and concurrent validity with other measures. It appears to measure different aspects of social functioning from the Global Assessment of Functioning (GAF), the most widely used social functioning measure, and the two measures appear to complement each other. Chapter 6 is a quantitative study using the HSFQ and a self-report measure of patients’ perceptions of therapeutic change processes to test the initial model of change developed in chapter 3. Fifty patients completed a checklist about how they had changed during treatment and the factors that had contributed to that change, as well as measures of social functioning. Self-reported levels of change were highly correlated with measures of patient functioning, though significant levels of change did not occur until the latter stages of treatment. The behaviour of therapists was particularly important throughout treatment, though participants in the final stage of therapy reported that the behaviour of other staff was as important as that of therapists, suggesting that, by this stage of treatment they are able to extend their range of supportive and therapeutic relationships. The results support a limited reparenting attachment-based model of therapeutic change. Chapter 7 is a pattern matching study that tests and refines the model of change. Ten patients completed a semi-structured interview about their interactions with their therapist. Their responses were analysed using a modified version of pattern matching to test hypotheses generated by the limited reparenting attachment-based model of change. The results support the limited reparenting model and suggest that patients’ attachment relationships with their therapists are an important change process for this population, particularly in the earlier stages of treatment. Chapter 8 presents a three-stage model of change based on the results of this thesis. On first admission, patients enter the orienting/ cognitive dissonance phase, in which they start to engage in treatment after perceiving a consistent improvement in how they are currently regarded and treated compared to how they have been regarded and treated previously, particularly in prison. Next, they enter the reparenting phase, during which their relationship with their therapist is the most important factor affecting change. Many features of the therapist-patient relationship during this phase parallel attachment processes between children and caregivers. Finally, patients enter the exploration/ generalisation phase in which they are able to explore from the secure base of their relationship with their therapist and develop supportive and therapeutic relationships with other staff members. This model provides a useful framework for working therapeutically with this patient group.

Background Smoking during pregnancy is the leading preventable cause of poor birth outcomes for mothers and babies. In 2010, 26% of women smoked during or in the 12 months prior to pregnancy, and 12% smoked throughout pregnancy. 120 trials of nicotine replacement therapy (NRT) have shown that it is effective for smoking cessation in non-pregnant smokers. However, the 9 trials conducted in pregnant smokers, provide no evidence that NRT helps pregnant smokers to stop. In 2012, a large randomised controlled trial (RCT) investigating NRT used in pregnancy for smoking cessation reported that it initially doubled smoking cessation rates, but found no evidence that it enhanced cessation throughout pregnancy. Trial participants reported low adherence to NRT. One possible reason for this is that nicotine metabolism is much faster in pregnancy and so the doses of NRT which have been trialled in pregnancy may not sufficiently treat women’s nicotine withdrawal symptoms. This study not only found that children of mothers in the intervention arm had better developmental outcomes at 2 years of age, but also that a dose-response relationship existed between reported higher NRT use and the latter. Using data from this trial, this thesis aims to further explore factors driving/associated with both higher adherence to NRT in pregnancy and improved birth outcomes, and to raise hypotheses about ways in which NRT may be used effectively to help pregnant smokers attempting to quit. Methods The studies reported in this thesis used data from 1,050 pregnant women recruited to the Smoking, Nicotine and Pregnancy (SNAP) trial. Linear and logistic regressions, mediation analysis, factor analysis and structural equation modelling techniques were employed to answer the following questions: (1) which factors are associated with smoking cessation in pregnancy?; (2) which factors, if any, mediate the negative influence of social disadvantage on cessation in pregnancy?; (3) is greater NRT use is associated with poor birth outcomes?; (4) is adherence to NRT associated with greater odds of cessation?; (5) what are the characteristics associated with the rate of nicotine metabolism and does more rapid metabolism of nicotine reduce a woman’s chances of achieving cessation? Results (1) Within a trial of NRT used for smoking cessation, pregnant women who were more educated (Odds Ratio (OR) at one month post-quit date (one month): 1.82, 95%CI: 1.24-2.67, p=0.002/OR at delivery: 1.89, 95%CI: 1.16-3.07, p=0.010) and less nicotine dependent (OR for baseline cotinine at one month: 0.94, 95%CI: 0.91-0.96, p<0.001/OR for baseline cotinine at delivery: 0.96, 95%CI: 0.92-0.99, p=0.010) / (OR for HSI at one month: 0.50, 95%CI: 0.32-0.76, p=0.001/OR for HSI at delivery: 0.43, 95%CI: 0.23-0.79, P=0.006) were more likely to achieve cessation. (2) There was evidence that the observed relationship between social disadvantage and cessation was mediated by women’s nicotine dependence (17.1% of the effect of social disadvantage on cessation), but not by them living with partners who smoked. (3) In an analysis investigating the relationship between use of NRT in pregnancy and birth outcomes, greater reported use of NRT was not associated with either better or worse birth outcomes, either in a-priori analyses using birth weight (β: -0.46, 95%CI: -3.58 to 2.66, p=0.773) and being born small for gestational age (SGA) (OR: 1.01, 95%CI: 0.99 to 1.03, p=0.184), or in exploratory analyses using all other SNAP trial birth outcomes. (4) In the first ever detailed analysis of the phenomenon of adherence with NRT in pregnancy, women who adhered more completely had lower pre-treatment cotinine concentrations (β: -0.08, 95%CI: -0.15- -0.01, p=0.020), lower heaviness of smoking index (β: -0.27, 95%CI: -0.50- -0.05, p<0.001) and were more likely to have been assigned to active rather than placebo NRT (β: 0.51, 95%CI: 0.29-0.72, p<0.001). Greater adherence with NRT was positively and significantly associated with increased odds of smoking cessation (Adjusted OR at one month: 1.11, 95%CI: 1.08-1.13, p<0.001/Adjusted OR at delivery: 1.06, 95%CI: 1.03-1.09, p<0.001), but there was no treatment allocation (i.e. nicotine or placebo)-adherence interaction with respect to cessation (LRT p=0.151). Analyses could not exclude the possibility that the observed adherence-cessation relationship was caused by women who relapsed to smoking consequently also stopping use of NRT. It also remains unclear whether the adherence-cessation relationship may be due the characteristics of individuals who are more likely to adhere (i.e. women who are prone to following instructions to use NRT, might also be more likely to follow instructions to not smoke). (5) In another novel analysis, pregnant women who metabolised nicotine more quickly had reduced odds of cessation (OR at one month: 0.87, 95%CI: 0.76-0.99, p=0.043/OR at delivery: 0.79, 95%CI: 0.66-0.94, p=0.008); however, there was no evidence that NRT was more effective in slower metabolisers. Conclusions Having higher levels of education and lower levels of nicotine dependence were associated with cessation. Nicotine dependence mediates the negative effect social disadvantage has on achieving smoking cessation in pregnancy. Greater reported use of NRT in pregnancy neither harms the foetus nor protects it from the harms of smoking, with respect to birth weight and SGA, but there is insufficient evidence to draw similar inferences for all birth outcomes. Adherence is associated with better smoking cessation outcomes, and women who exhibit greater adherence have differing characteristics compared to those that adhere less. It was not possible to determine whether or not these better outcomes were a function of women’s characteristics or if they were attributable to the treatment they received. The rate of nicotine metabolism appears to be important in determining whether or not pregnant smokers will manage to successfully stop smoking, but there was no evidence that NRT effectiveness was influenced by nicotine metabolism. Quitting smoking during pregnancy is important for both the mother and the child’s health outcomes. Socially disadvantaged women are less likely to achieve cessation but this thesis provides the first evidence to suggest that these odds may be improved by addressing nicotine dependence; NRT provides an eminently treatable way of doing this. This thesis’s findings suggest that use of NRT is at least as safe as continued smoking with respect to birth outcomes, although probably safer. This thesis also provides the first evidence on the associations between individual pieces of NRT and cessation in pregnancy, as well as the first evidence on nicotine metabolism’s association with cessation in pregnancy. Overall, NRT provides a promising way to increase the odds of cessation for disadvantaged pregnant smokers, however, there is a need for further research and the work in this thesis should provide a spur to investigate the effects that adherence to NRT and an individuals’ metabolism may have on NRT effectiveness in pregnancy.