This thesis aimed to examine service contact among children and young people with mental health problems, and has three complementary parts. The first is a secondary analysis of data from the British Child and Adolescent Mental Health Survey (BCAMHS) 2004, which explored mental health related service contact in relation to psychopathology over three years. The second and third parts focussed on young people with ADHD in transition from child services, which is a particularly challenging time. This involved a qualitative interview study of young peoples’ experiences, and an analysis of primary care prescribing of ADHD medication over the transition period using a cohort from the Clinical Practice Research Datalink from 2005-2013. Less than a third of children with a psychiatric disorder in BCAMHS reported contact with child mental health services. Instead, teachers were the most frequently used service, with two-thirds reporting mental health related contact. Interviews with young people with ADHD highlighted themes including concerns around medication management post transition and need for information. The prescribing analysis found that the majority of adolescents on ADHD medication at age 16 stopped during the transition period. This continuing disparity between estimates of symptom persistence and medication persistence suggests that many may be stopping medication from which they could still benefit; as various barriers have been identified to ongoing prescribing. In summary, the findings of these three linked studies suggest common themes in terms of unmet needs and gaps between policy and practice in mental health services for children and young people. One of the chief implications is the need for oversight and policy levers to ensure the implementation of best practice, accompanied by complementary efforts to better understand and overcome other barriers to providing optimal care, including research into knowledge and attitudes of different groups and the provision of targeted training.

L’hypertension artérielle (HTA) est considérée comme un facteur de risque modifiable de démence. Dans ce contexte, nous nous sommes intéressés à différents aspects du traitement de l’HTA indépendamment de la réduction de pression artérielle : d’une part aux effets des différentes classes d’antihypertenseurs sur le déclin cognitif et l’incidence de démence, en réalisant une méta-analyse en réseau ; et, d’autre part, sur la variabilité depression artérielle (PA), qui pourrait être impliquée dans la relation délétère entre l’HTA et la cognition, à partir d’une cohorte de sujets hypertendus traités. Nous avons également étudié le rôle de l’HTA dans la relation entre le syndrome métabolique et les troubles cognitifs, à partir d’une cohorte de patients avec facteurs de risque vasculaires. Nos résultats confirment la prévention de démence avec les antihypertenseurs et soutiennent l’hypothèse de bénéfices cognitifs supérieurs avec les antagonistes des récepteurs de l’angiotensine II par rapport aux autres classes d’antihypertenseurs, mais, ils suggèrent que la variabilité de pression artérielle ne constitue pas un mécanisme principal à l’origine de ces différences, puisque les inhibiteurs calciques offrent la meilleure réduction de variabilité de PA. Nos résultats ne supportent pas le rôle du syndrome métabolique en tant que tel dans le déclin cognitif, mais de celui de l’HTA et du diabète. L’ensemble de ces résultats démontrent qu’une approche multifactorielle doit être considérée dans le traitement de l’HTA, et offrent des perspectives dans le choix du traitement antihypertenseur pour prévenir le fardeau de la démence. / Hypertension is considered a modifiable risk factor of dementia. In this context, we studied different aspects of the treatment of hypertension independantly from the blood pressure reduction. We compared the effects of the different antihypertensive drug classes, on cognitive decline and incidence of dementia in a network meta-analysis on the one hand, and on short-term blood pressure variability, which is suspected to be related to cognitive decline,in a cohort of treated hypertensive patients, on the other hand. We also studied the contribution of hypertension in the relationship between metabolic syndrome and cognitive impairment, in a cohort of patients with vascular risk factors. Our results confirmed the benefits of antihypertensive treatment in the prevention of incident dementia, and support the hypothesis of superior benefits with angiotensin receptor blockers compared to the other drug classes. Our finding of calcium channel blockers being the class providing the lowest blood pressure variability does not support blood pressure variability being a primary mechanism involved in the differential effects of antihypertensive drug classes on cognition. We demonstrated that hypertension and diabetes, rather thanmetabolic syndrome in itself, were related to cognitive impairment. Overall, our results highlight the need to consider a multifactorial approach in hypertensiontreatment, and provide perspectives regarding the choice of antihypertensive treatment to prevent the burden of dementia.

Background: The use of antimicrobials has been presented as a significant risk factor for Clostridium difficile infection (CDI). Nevertheless, it remains unclear which antimicrobials are more likely to be associated with CDI. It is mandatory for acute trusts to report the numbers of diagnosed CDI cases to the Health Protection Agency (HPA). There is no national system to collect and analyse antimicrobial usage data from the trusts. The company IMS collects antimicrobial usage data from the trusts for creating marketing research statistics. Therefore, it was hypothesised that data collected from the HPA and from IMS could be used to undertake an ecological study about the association between CDI cases and antimicrobial use in English trusts. Methods: A trust-level Antimicrobial Usage Database provided by IMS and a database, including the numbers of CDI cases for patients aged 65 years old and above, provided by the HPA, were utilised in this work. These referred to 26 out of the 29 NW SHA trusts (that managed 64 hospitals) for the quarters between 2005 and 2008 inclusive. A sample of antimicrobial usage data collected directly from trusts was used to investigate potential limitations in using the Antimicrobial Usage Database for the purpose of this work. Multilevel models were used to study antimicrobial usage and the number of CDI cases over time. These models were also used to investigate the association between the CDI cases and antimicrobial usage in the trusts. The trends of trust antimicrobial usage over time were compared with DH recommendations for the prevention of CDI through antimicrobial prescribing published in 1994, 2005 and 2008. Results: Discrepancies between the antimicrobial usage recorded in the IMS database and in a sample of antimicrobial usage data collected from trusts were found for 31 out of 155 antimicrobial usage records; only 1 of these referred to an antimicrobial with high usage. Eight out of the 23 antimicrobial groups and 10 out of 63 antimicrobials were presented as having high usage. The antimicrobial usage over time increased significantly for 7 antimicrobial groups, decreased significantly for 2 groups and remained constant for 54 groups. The number of CDI cases reported for patients aged 65 years old and above decreased significantly over the time. Trust antimicrobial usage over time changed in the opposite direction compared to the DH recommendations published in 1994, 2004 and 2008, respectively, for 2 out of 11, 3 out of 12 and 3 out of 14 antimicrobial groups/antimicrobials. The increased usage of 5 antimicrobial groups was significantly associated with an increase in the number of CDI cases and an increased usage of 4 antimicrobial groups was significantly associated with a decreased number of CDI cases. Within the antimicrobial groups that were significantly associated with an increased number of CDI cases, the usage of 8 individual antimicrobials was significantly associated with the CDI cases. Discussion/Conclusion: Collecting antimicrobial usage over time for large groups of trusts is very time consuming and requires extensive data manipulation. The similarity of the results of this study with those of previously published studies suggest that HPA and IMS data may be used to investigate the association between CDI cases and antimicrobial usage in English trusts.

Les études observationnelles en pharmacoépidémiologie sont souvent mises en place pour évaluer un médicament mis sur le marché récemment ou concurrencé par de nombreuses alternatives thérapeutiques. Cette situation conduit à devoir évaluer l'effet d'un médicament dans une cohorte comprenant peu de sujets traités, c'est à dire une population où l'exposition d'intérêt est rare. Afin de prendre en compte les facteurs de confusion dans cette situation, certains auteurs déconseillent l'utilisation du score de propension au profit du score pronostique, mais cette recommandation ne s'appuie sur aucune étude évaluant spécifiquement les faibles prévalences de l'exposition, et ignore le type d'estimation, conditionnelle ou marginale, fournie par chaque méthode d'utilisation du score pronostique.La première partie de ce travail évalue les méthodes basées sur le score de propension pour l'estimation d'un effet marginal en situation d'exposition rare. La deuxième partie évalue les performances des méthodes basées sur le score pronostique rapportées dans la littérature, introduit de nouvelles méthodes basées sur le score pronostique adaptées à l'estimation d'effets conditionnels ou marginaux, et les compare aux performances des méthodes basées sur le score de propension. La dernière partie traite des estimateurs de la variance des effets du traitement. Nous présentons les conséquences liées à la non prise en compte de l'étape d'estimation du score de propension et du score pronostique dans le calcul de la variance. Nous proposons et évaluons de nouveaux estimateurs tenant compte de cette étape. / Pharmacoepidemiologic observational studies are often conducted to evaluate newly marketed drugs or drugs in competition with many alternatives. In such cohort studies, the exposure of interest is rare. To take into account confounding factors in such settings, some authors advise against the use of the propensity score in favor of the prognostic score, but this recommendation is not supported by any study especially focused on infrequent exposures and ignores the type of estimation provided by each prognostic score-based method.The first part of this work evaluates the use of propensity score-based methods to estimate the marginal effect of a rare exposure. The second part evaluates the performance of the prognostic score based methods already reported in the literature, compares them with the propensity score based methods, and introduces some new prognostic score-based methods intended to estimate conditional or marginal effects. The last part deals with variance estimators of the treatment effect. We present the opposite consequences of ignoring the estimation step of the propensity score and the prognostic score. We show some new variance estimators accounting for this step.

Nakaoka S, Ishizaki T, Urushihara H, Satoh T, Ikeda S, et al. (2014) Prescribing Pattern of Anti-Parkinson Drugs in Japan: A Trend Analysis from 2005 to 2010. PLoS ONE 9(6): e99021. doi:10.1371/journal.pone.0099021 / 京都大学 / 0048 / 新制・課程博士 / 博士(社会健康医学) / 甲第18649号 / 社医博第61号 / 新制||社医||8(附属図書館) / 31563 / 京都大学大学院医学研究科社会健康医学系専攻 / (主査)教授 髙橋 良輔, 教授 川上 浩司, 教授 松原 和夫 / 学位規則第4条第1項該当 / Doctor of Public Health / Kyoto University / DFAM

Ce travail porte sur l’étude du risque de démence chez les personnes âgées ayant consommé des benzodiazépines. Ces médicaments méritent une attention particulière du fait de (i) leur utilisation trop systématique et le plus souvent chronique contrairement aux recommandations préconisant des durées d’utilisation courtes (ii) leurs effets délétères sur la cognition demeurant mal évalués à long terme. La plupart des études conduites sur ce sujet ont conclu à une augmentation du risque de démence chez les sujets ayant utilisé des benzodiazépines. Un biais protopathique pouvait cependant, en partie du moins, avoir expliqué ces résultats : la prescription de benzodiazépines pouvait avoir été motivée par des prodromes souvent observés au cours des années précédant le diagnostic de la maladie. Afin de mieux prendre en considération ce biais, le projet BENZODEM a utilisé les ressources de la cohorte PAQUID (3777 sujets ≥ 65 ans tirés au sort sur les listes électorales de Dordogne et Gironde bénéficiant d’un suivi de plus de 20 ans). Ce projet, combinant deux études de cohorte et une étude cas-­‐témoins, a conclu à un risque de démence augmenté de 46 à 62% chez les utilisateurs de benzodiazépines et retardé de 5 à 15 ans par rapport à l’initiation du traitement. La seconde partie du programme (BENZODEM2) a consisté en une étude cas-­‐témoins conduite sur un large échantillon de sujets de plus de 65 ans enregistrés sur la base de données de la Régie de l’Assurance Maladie du Québec (RAMQ). Ce programme a permis (1) de valider les précédents résultats (risque augmenté de 30 à 80% en fonction de la dose, la durée du traitement et la nature des molécules) (2) d’identifier les profils de consommation associés à un excès de risque : consommateurs de plus de 3 mois avec une relation dose-­‐effet marquée et molécules à longue demi-­‐ vie d’élimination. Des explorations complémentaires ont permis de conclure que cet excès de risque n’était pas expliqué par une mortalité différentielle entre groupes comparés ni par la prescription d’autres médicaments psychotropes. Une autre étude menée sur PAQUID montrait une absence de différence entre consommateurs et non consommateurs de benzodiazépines vis-­‐à-­‐vis de l’évolution des scores mesurant les fonctions cognitives. Ces résultats ont permis d’émettre des hypothèses concernant le mécanisme de l’association entre utilisation de benzodiazépines et démence: (1) les benzodiazépines pourraient constituer des marqueurs précoces de la maladie ; (2) les benzodiazépines pourraient aussi diminuer les capacités de recours à la réserve cognitive en réponses aux lésions précoces de la maladie au stade préclinique ; (3) il est aussi possible que ces deux explications soient combinées. / This work deals with the risk of dementia in elderly individuals who have used benzodiazepines. These drugs deserve particular attention because (i) their use appears to be too systematic and most often chronic despite good practice guidelines recommending short durations of use (ii) their deleterious effects on cognition remain underevaluated for the long-­‐term. Most of the studies conducted concluded that there was an increased risk of dementia among benzodiazepine users. In fact, a protopathic bias could, at least in part, have explained these results. Indeed, the prescription of benzodiazepines could have been motivated by the prodromes often observed several years before the clinical diagnosis of a dementia. With the aim of better controlling for this bias, the BENZODEM project used the resources of the PAQUID cohort (3777 subjects ≥65 years randomly sampled from electoral lists in South-­‐West France, with a 20-­‐ year follow-­‐up). This project combined two cohort studies and one case-­‐control. These studies concluded in a risk of dementia increased by 46 to 62% in benzodiazepine users and delayed by 5 to 15 years after treatment initiation. The second part of the programme (BENZODEM2) consisted of a case-­‐control study conducted in a large sample of subjects >65 years registered in the Quebec Health care database (Régie de l’Assurance Maladie du Québec, RAMQ). It was thus possible(1) to validate the previous results by using a different population (the risk was found to be increased by 30 to 80% depending on the patterns of use regarding dose, duration and type of molecule), (2) to identify the patterns of use which appeared to be at risk; excess risk was only apparent for uses of more than three months with a marked dose-­‐effect relationship, and was higher for molecules with a long elimination half-­‐life. Complementary explorations using the PAQUID cohort indicated that the excess risk in exposed was not explained by a differential mortality rate between the groups compared. Other studies suggested that the link found remained independently of the prescription of other psychotropics. Another analysis in the PAQUID cohort showed that, in the absence of dementia, no difference was observed between benzodiazepine users and non-­‐users with regards to the evolution of scores evaluating cognitive functions. These results led to several assumptions about the putative mechanism explaining the relationship found between benzodiazepine use and dementia: (1) benzodiazepines could be early markers of symptoms such as anxiety, depression or insomnia, which are potential prodromes or risk factors for this disease, (2) these drugs could also reduce the ability to use cognitive reserve in order to cope with early lesions of the disease during the preclinical stage, (3) the association found could also result from these two mechanisms.

No description available.

No description available.

Les formes orales de traitements anticancéreux se sont considérablement développées récemment. La question de l’adhésion au traitement devient donc un nouvel enjeu en oncologie. Cette thèse aborde de façon générale le problème de l’adhésion aux traitements oraux en oncologie, et plus spécifiquement celui de la mesure en bases de données médico-administrativesde l’observance et de la persistance à l’hormonothérapie, traitement oral majeur dans le cancer du sein. Le point de vue retenu est celui de la pharmacoépidémiologie, à savoir l’étude des consommations pharmaceutiques en contexte clinique. La première partie de cette thèse fait le point sur les connaissances actuelles concernant l’adhésion à l’hormonothérapie - tamoxifène et inhibiteurs de l’aromatase (IA) - dans le cancer du sein. La seconde partie, reposant sur l’analyse de cohortes de patientes atteintes de cancer du sein sélectionnées à partir de (1) la UK General Practice Research Database et (2) des données de l’Assurance Maladie, a montré que : - Plus de la moitié des femmes de moins de 40 ans au diagnostic ne reçoivent plus de tamoxifène à 5 ans. Il s’agit du groupe de femmes le plus à risque d’arrêt prématuré de traitement. - Chez les femmes âgées de plus de 50 ans au diagnostic, les arrêts de traitement sont moins fréquents pour les IA que pour le tamoxifène. - Les déterminants associés à la non-persistance sont un faible soutien social et la déclaration précoce de non-prise de traitement par la patiente chez les femmes jeunes. Chez les femmes âgées, l’utilisation de médecines complémentaires et alternatives, la présence de comorbidités sont associées à une augmentation du risque d’arrêt de traitement. A contrario, la présence d’une poly-médication est associée à une diminution du risque d’arrêt.- Dans les études précédentes, une proportion importante de femmes reprend son traitement au moins une fois après l’avoir arrêté de façon prolongée. Les arrêts transitoires de traitements ont été pris en compte à l’aide de modèles multi-états. La probabilité d’arrêt de traitement estimée à partir de ces modèles est plus faible que celle mesurée par la méthode de Kaplan-Meier, après la première année de traitement. La non-adhésion à l’hormonothérapie est fréquente. Certains de ses déterminants sont modifiables ou peuvent servir à identifier précocement les patientes à risque de non-observance. La prise en compte des arrêts transitoires de traitement est importante dans la mesure de la persistance. L’adhésion est l’élément clé faisant le lien entre l’efficacité d’un traitement mesuréedans un essai clinique et son impact dans la vraie vie. Il est urgent de prendre conscience de l’importance de la non-adhésion des formes orales en oncologie / The use of oral anticancer therapies has significantly increased in recent years. Adherence to these therapies has therefore become a major issue in the field of oncology. This thesis focuses on the question of treatment adherence in oncology, and more specifically on the use of medical records and administrative databases to estimate adherence and persistence to hormonal therapy—now a major form of oral breast cancer therapy. Our perspective is based on pharmacoepidemiology, i.e. the study of drugs in a clinical setting. The first part of this thesis synthesizes current knowledge on adherence and persistence to hormonal therapy for BC – i.e. tamoxifen and aromatase inhibitor therapies.The second part, which is based on the study of two cohorts constituted (1) from the UK General Practice Research Database and (2) from the French National Health Insurance System, demonstrates that - More than half of women younger than 40 at diagnosis do not receive any tamoxifen at 5years of follow-up. This group of women presents the highest rates of treatmentinterruption. - Among women over 50 at diagnosis, those receiving some form of AI therapy discontinue less frequently than those on tamoxifen treatment. - Determinants of non-persistence identified in the studies under review include low social support and self-reporting of non-compliance among younger women. Among older women, those using complementary or alternative medicine or suffering from comorbidities are more likely to discontinue their treatment, whereas women usingpolypharmacy are less likely to discontinue. - In previous studies, a large proportion of women who discontinued their treatment resumed after a prolonged gap. To account for these temporary treatment discontinuations, we used multi-state models. The probability of being off treatment estimated from these models is lower than that estimated from Kaplan-Meier estimates, after the 1st year of treatment. Adherence to hormonal therapy is largely suboptimal. Some of its determinants are modifiablefactors, while others can be used to identify sub-groups of patients at high risk of non-adherence. Accounting for temporary treatment discontinuation is important when measuring nonpersistence. Adherence is a key element for the translation of efficacy measured in clinical trials into effectiveness in real life. There is an urgent need to acknowledge the problem of nonadherence to oral therapy in oncology

Os inibidores da enzima conversora da angiotensina são uma classe de medicamentos freqüentemente prescrita pelos médicos e importante para o tratamento da Hipertensão Arterial Sistêmica (HAS) e da Insuficiência Cardíaca Congestiva (ICC). Os dois primeiros protótipos desta classe, o captopril e o maleato de enalapril, constam na Relação Nacional de Medicamentos Essenciais (RENAME) devido à importância terapêutica, eficácia clínica e segurança comprovada, além de seu custo-efetividade. Para cumprir o objetivo de estudar a utilização desta classe terapêutica foi realizado um levantamento no banco de dados da Secretaria Municipal de Saúde de Ribeirão Preto (SMS-RP) do estado de São Paulo (SP) para identificar os usuários do Sistema Único de Saúde (SUS) que receberam a dispensação de captopril e maleato de enalapril pelas farmácias das Unidades Básicas de Saúde (UBS) e Distritais de Saúde (UBDS) no período compreendido entre 01/03/2006 e 28/02/2007. Identificou-se que 9.560 pacientes utilizaram os inibidores da ECA, sendo que destes, 46,57% utilizaram captopril, 45,74% enalapril e 7,69% os dois fármacos simultaneamente ou não. A idade média dos usuários foi de 61 anos e houve um aumento progressivo da utilização desses agentes com o incremento da faixa etária e houve predominância para o gênero feminino. A aderência ao tratamento dos usuários das unidades de saúde do Distrito Oeste de Ribeirão Preto foi estimada em 80,6%. Enquanto que a dispensação única dos inibidores da ECA foi encontrada para 8,6% dos indivíduos, com idade média de 53,5 anos e as doses médias prescritas de captopril e de enalapril foram de 63,8 mg e 19,8 mg/dia respectivamente. Por outro lado, as doses médias prescritas e mantidas de captopril foram de 69,9 mg/dia e de enalapril foram de 21,35 mg/dia. Aproximadamente 0,3% dos pacientes utilizaram captopril em doses médias prescritas e mantidas iguais ou superiores a 150 mg e 0,65% dos pacientes receberam doses de enalapril acima de 40 mg, porém não foram encontradas doses subterapêuticas prescritas para esses medicamentos. Além disso, 20,21% dos pacientes analisados neste estudo tiveram seus esquemas terapêuticos alterados, sendo que a maioria teve apenas uma alteração. Um total de 92,69% dos usuários utilizou mais de um medicamento além dos inibidores da ECA e o incremento desse valor foi diretamente proporcional a faixa etária. O número de pacientes com risco de apresentar interação medicamentosa foi de 3.974 (41,57%), sendo que a maioria dos pacientes utilizou apenas um medicamento com essa possibilidade. / The angiotensin-converting enzyme inhibitors is a class of drugs often prescribed by pharmacian and important for the treatment of systemic arterial hypertension (SAH) and the Congestive Heart Failure (CHF). The first two prototypes of this class, enalapril maleate and captopril, are in the National Essential Drugs (RENAME) because of their therapeutic importance, clinical efficacy and safety established, and its cost-effectiveness. The aim of this study was reached through a survey database of the Municipal Health Secretary of Ribeirão Preto (SMS-RP) of São Paulo (SP) state to identify the users of the Unified Health System (SUS) that received dispensation of captopril and enalapril maleate by the basics health units (UBS) and districts health units (UBDS) for the period between 01/03/2006 and 28/02/2007. It was identified that 9,560 patients used ACE inhibitors, of which, 46.57% used captopril, 45.74% enalapril and 7.69% these two drugs simultaneously or not. The average age of users was 61 years and there was a progressive increase in the use of these agents with increasing age. The treatment adherence was estimated at 80.6% in users of health care units in the Western District of Ribeirão Preto. A single dispensing of ACE inhibitors was found for 8.6% of individuals with a mean age of 53,5 years and mean dose of 63.8 mg and the average prescribed doses of captopril were 69.9 and 19.8 mg/day respectively. Furthermore, the mean doses prescribed and maintained for captopril was 69.9 mg/day and enalapril were 21.35 mg/day. Around 0.3% of patients used average prescribed and maintained captopril doses equal or greater than 150 mg and 0.65% of patients received enalapril doses above 40 mg, but there were no prescribed subtherapeutic doses of these drugs. Moreover, 20.21% of patients analyzed in this study had their treatment regimens modified, and the majority had only one change. A total of 92.69% of users used more than one drug than the ACE inhibitors and the increase of this value was directly proportional to age. The number of patients with potential risk of a potential drug interaction was 3,974 (41.57%), and the most patients used only one drug with possibility.