O transtorno de estresse pós-traumático (TEPT) tem um impacto negativo na vida de seus portadores. Conhecer os traços de personalidade mais preponderantes nas pessoas com TEPT pode auxiliar no sucesso do tratamento, tornando possível planejar intervenções mais adequadas. A terapia cognitivo-comportamental (TCC) é considerada um dos tratamentos mais eficazes para este transtorno; entretanto, nem todos aqueles que completam a terapia evoluem para uma melhora significativa. Uma das razões pode ser encontrada nas características individuais de personalidade de cada pessoa. Portanto, o presente estudo busca explorar a relação existente entre os domínios e traços de personalidade que estão associados à melhora do paciente que realizou a TCC. Método: 66 pacientes com diagnóstico de TEPT, segundo o DSM-IV-TR, com idade entre 18 e 60 anos participaram do estudo. Instrumentos: Para avaliar os traços de personalidade foi utilizado o instrumento NEO-PI-R. Para avaliar a gravidade e melhora da doença foi utilizada a escala de Impressão Clínica Global (CGI). Procedimento: Os pacientes passaram por avaliação psiquiátrica para assegurar o diagnóstico de TEPT. Após aceitarem participar do estudo, responderam ao NEO-PI-R e foram avaliados por médicos quanto a gravidade da doença, utilizando o CGI. Os participantes passaram por 13 sessões de TCC realizadas por profissionais devidamente treinados. Ao final, foram reavaliados para verificar se houve melhora após tratamento. Resultados: Quanto ao perfil de personalidade 71,2% apresentaram neuroticismo (N) alto, 75,8% relataram escore elevado em extroversão (E) e 45,5% eram baixos em conscienciosidade (C). Já os traços amabilidade (A) e abertura para experiência (O) apresentaram pontuações na média. As análises também demonstraram que os participantes que apresentavam o domínio de personalidade denominado consicienciosidade (C) foram associados ao resultado favorável do tratamento. Estima-se que a chance de melhora cresça 3,77 vezes se o paciente apresentar esse traço, quando comparado com os demais que não possuem essa característica. Duas facetas (assertividade e ações variadas) também foram correlacionadas com a melhora no tratamento. Conclusão: Embora a amostra do presente estudo seja limitada, os resultados apontam para a importância de se avaliar a personalidade do paciente. Acessar a personalidade é importante com a finalidade de tentar predizer qual o melhor tipo de tratamento terapêutico para cada um. As terapias breves (frequentemente administradas nos hospitais públicos) possuem um tempo limitado de tratamento, de modo que informações sobre as variáveis de personalidade podem ser particularmente muito útil / The Post-Traumatic Stress Disorder (PTSD) has a negative impact on the patients lives. Get to know their personality traits can help on the treatment success, by making possible to plan most appropriate interventions. Cognitive-Behavioral Therapy (CBT) is considered a first line treatment for PTSD. However, treatment response is not universal. One reason may be found in personality characteristics. The present study aims at investigating the association between personality dimensions and traits associated with improvement of patients who underwent CBT. Method - 66 PTSD patients diagnosed according to the DSM-IV-TR criteria were included in the study. The patients included were aged 18 to 60 years old. Instruments - We employed the NEO-PI-R instrument for the evaluation of personality dimensions and the Clinical Global Impressions Scale (CGI) for evaluation of clinical outcome. Procedure - Patients were assessed by Psychiatrists to ensure the diagnosis of PTSD. After accepting to participate of the study, they answered the NEO-PI-R Scale and were assessed by doctors to know the disease severity, using the CGI scale. Participants underwent 13 CBT sessions and were reassessed at the end of treatment. Results - The personality profile showed that 71.2% were high in neuroticism (N) and 75.8 reported low Extraversion. 45.5 were low in conscientiouness and the Agreableness (A) and openess (O) factors presented average scores. The analysis also showed that patients presenting the Conscientiousness (C) personality dimension showed a higher chance of improvement (OR=3.77). Two facets other dimensions (Assertiveness and Varied Actions) were also associated with better clinical outcome. Conclusion - determining predictors of outcome such as a patient\'s personality dimensions may point to the use of therapeutic treatment options with the higher odds of success, without too much therapeutic treatment experimentation. As therapies become briefer, information on personality variables may be particularly useful

Submitted by Filipe dos Santos (fsantos@pucsp.br) on 2017-08-30T12:54:41Z No. of bitstreams: 1 Adriana Príncipe dos Reis Albuquerque Ramos.pdf: 1417205 bytes, checksum: 81d38f5cef551a305a49882ddf634251 (MD5) / Made available in DSpace on 2017-08-30T12:54:42Z (GMT). No. of bitstreams: 1 Adriana Príncipe dos Reis Albuquerque Ramos.pdf: 1417205 bytes, checksum: 81d38f5cef551a305a49882ddf634251 (MD5) Previous issue date: 2017-08-18 / Conselho Nacional de Pesquisa e Desenvolvimento Científico e Tecnológico - CNPq / There are evidences in the literature that thelearning of reading and writing is facilitated by phonological awareness and by the knowledge of the name of the letters. Current studies investigate which components of phonological awareness best predict reading and writing skills. The aim of this research was to analyze the relationship between the reading and writing of words with the variables of phonological awareness (measured by the knowledge of rhymes, syllabic segmentation and phonemic deletion) and knowledge of the name of the letters in children entering Primary School. For this, a brief longitudinal study was carried out accompanying the first year students of Elementary School during the school year. The study was composed of three data collection sessions, in which the students were evaluatedby a set of tasks of phonological awareness, knowledge of the alphabet, reading and writing of isolated words. The results showed a positive and significant correlation between letter knowledge, rhyme knowledge and phonemic deletion in the three collection moments with word reading. The syllabic segmentation obtained significant positive correlation only in the second collection. All components of phonological awareness, as well as letter knowledge, showed positive and significant correlations with word writing in three moments of the research. The results ofMultiple Regression tests(Stepwise)allowed the identification of knowledge of letter name and the knowledge of rhymes as predictors of reading and writing of words when the phonemic deletion did not obtain a measurable result. However, when the children started learning to read, the components of phonemic deletion and syllabic segmentation began to predict the reading of words with a high degree of significance, being the phonemic deletion also a predictor of word writing, accompanied by the knowledge of the names of the letters / Há evidências na literatura de que a aprendizagem da leitura e da escrita é facilitada pelaconsciência fonológica e pelo conhecimento do nome das letras. Estudos atuais investigam qual dos componentes da consciência fonológica melhor prediz as habilidades de leitura e escrita. Esta pesquisa teve como objetivoanalisaras relaçõesentre a leitura e escrita de palavras com as variáveis de consciência fonológica, (medida pelo conhecimento de rimas, segmentação silábica e deleção fonêmica) e conhecimentodo nome das letrasem crianças que ingressam no Ensino Fundamental. Paratal, foi realizado um estudo longitudinal breve acompanhando os alunos do primeiro ano do Ensino Fundamental durante o ano letivo. O estudo foi composto por três sessões de coleta de dados, nas quais os escolares foram avaliados por um conjunto de tarefas de consciência fonológica, conhecimento do alfabeto, leitura e escrita de palavras. Os resultados mostraram correlação positiva e significativa entre o conhecimento de letras, conhecimento de rimas e deleção fonêmica nos três momentos de coletas com a leitura de palavras. A segmentação silábica obteve correlação positiva significativa somente na segunda coleta. Todos os componentes da consciência fonológica, assim como o conhecimento de letras demonstraram correlações positivas e significativas com a escrita de palavras nos três momentos da pesquisa. Os resultados dos testes de Regressão Linear Múltipla Stepwise(por etapas) permitiram identificar o conhecimento do nome das letras e o conhecimento de rimas como preditores de leitura e escrita de palavras quando a deleção fonêmica não obteve resultado mensurável. No entanto, quando as crianças iniciaram a aprendizagem da leitura, os componentes deleção fonêmica e segmentação silábica passaram a predizer a leitura de palavras com um alto grau de significância, sendo a deleção fonêmica também preditora da escrita de palavras, acompanhada pelo conhecimento do nome das letras

Submitted by PPG Ecologia e Evolu??o da Biodiversidade (eebpg.ciencias@pucrs.br) on 2018-10-16T17:56:11Z No. of bitstreams: 1 Tese - Pereira - Joana.pdf: 10961538 bytes, checksum: b96f5fee6d02e741124bc5a31abdb086 (MD5) / Approved for entry into archive by Sheila Dias (sheila.dias@pucrs.br) on 2018-10-18T12:48:41Z (GMT) No. of bitstreams: 1 Tese - Pereira - Joana.pdf: 10961538 bytes, checksum: b96f5fee6d02e741124bc5a31abdb086 (MD5) / Made available in DSpace on 2018-10-18T13:37:00Z (GMT). No. of bitstreams: 1 Tese - Pereira - Joana.pdf: 10961538 bytes, checksum: b96f5fee6d02e741124bc5a31abdb086 (MD5) Previous issue date: 2018-09-24 / Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior - CAPES / Distribution models are considered important tools in biogeography and ecology studies as they allow spatial and temporal extrapolation of species distribution from a set of occurrence data points as function of environmental predictors. In this study, we aimed to propose logistic distribution models for 50 fish species along the Sinos River Basin (Rio Grande do Sul, Brazil), by using altitude and basin area as geomorphological descriptors. Upstream-downstream richness model for entire basin by using multiple and logistic regression was also presented. The dataset derived from several sampling programs performed from 1998 to 2014, comprising 86 sampled locations with fish collections with gillnets, electric fishing or both. Two logistic models were constructed using environmental data of altitude and upstream basin area, testing linear (Model 1) and nonlinear (Model 2) factor responses for each species. Performance of the two models was evaluated using sensitivity, specificity, AUC (Area Under Curve) and TSS (True Skill Statistical). The adjusted models presented sensitivity values ranging from 47,1 to 99,8, specificity from 49,94 to 98,10, AUC from 0,75 to 0,99 and TSS from 0,26 to 0,98. The linear response model, although simpler, proved to be effective in predicting species distribution, as observed in previous studies. The basin area had a positive effect on the distribution of most species according to the linear model, although this was not verified for nonlinear models due to the interaction between variables. In general, the nonlinear model presented higher performance values for the evaluated metrics for the majority of species, but suffered from overfitting and a patchy distribution estimation. The species richness increased along a longitudinal gradient, presenting its maximum value downstream of the basin. Of the four estimated richness models, models with six variables were the ones with the lowest residual variation. / Modelos de distribui??o s?o considerados como ferramentas importantes em estudos de biogeografia e ecologia, pois permitem a extrapola??o espacial e temporal da distribui??o de esp?cies a partir de um conjunto de dados de ocorr?ncia em fun??o de preditores ambientais. Neste estudo, objetivamos propor modelos de distribui??o log?stica para 50 esp?cies de peixes ao longo da bacia do rio dos Sinos (Rio Grande do Sul, Brasil), utilizando altitude e ?rea da bacia como descritores geomorfol?gicos. Modelos de riqueza montante-jusante para toda a bacia usando regress?o m?ltipla e log?stica tamb?m foram propostos. O conjunto de dados derivou de v?rios programas de amostragem realizados de 1998 a 2014, compreendendo 86 locais amostrados com coleta de peixes com redes de emalhar, pesca el?trica ou ambos. Dois modelos log?sticos foram constru?dos utilizando dados ambientais de altitude e ?rea de bacia a montante, testando respostas lineares (Modelo 1) e n?o lineares (Modelo 2) para cada esp?cie. O desempenho dos modelos foi avaliado usando sensibilidade, especificidade, AUC (Area Under Curve) e TSS (True Skill Statistical). Os modelos ajustados apresentaram valores de sensibilidade variando de 47,1 a 99,8, especificidade de 49,94 a 98,10, AUC de 0,75 a 0,99 e TSS de 0,26 a 0,98. O modelo de resposta linear, embora mais simples, mostrou-se eficaz na predi??o da distribui??o de esp?cies, como observado em estudos anteriores. A ?rea da bacia teve um efeito positivo na distribui??o da maioria das esp?cies de acordo com o modelo linear, embora isso n?o tenha sido verificado para os modelos n?o lineares devido ? intera??o entre as vari?veis. Em geral, o modelo n?o linear apresentou valores de desempenho mais altos para as m?tricas avaliadas para a maioria das esp?cies, mas sofreu de hiperajuste e uma estimativa de distribui??o fragmentada. A riqueza de esp?cies aumentou ao longo de um gradiente longitudinal, apresentando seu valor m?ximo a jusante da bacia. Dos quatro modelos estimados de riqueza, os modelos com seis vari?veis foram os que apresentaram a menor amplitude de varia??o dos res?duos.

Every year more than 72,000 adolescents and young adults (AYAs) in the United States (US) aged 15-39 years are diagnosed with cancer. AYAs represent a population that falls into a care gap between pediatric and adult medical services. Additionally, AYAs have experienced increased cancer incidence and decreased five-year survival rates compared to other age groups. The spectrum of tumors seen in AYAs differs from children and older adults, with 90% of the tumors stemming from ten cancer types. Melanoma of the skin, characterized by the uncontrolled growth of pigment-producing cells, is the third most common cancer diagnosed among AYAs in the US. Overexposure to ultraviolet (UV) radiation from sunlight or artificial sources is the greatest risk factor for melanoma. AYAs seem to be particularly at risk for developing melanoma due to increased UV exposure early in life. This study’s objectives were to understand the unique characteristics of melanoma in AYAs in Tennessee and identify the predictors of late- stage diagnosis. The sample for this study includes all incident melanoma cancer cases (N=1109) in AYAs from the Tennessee Cancer Registry (TCR) for the years 2004-2008, inclusive. AYA cases were defined as cancer cases that were diagnosed in individuals ages 15-39 years, inclusive. Melanoma cases were defined according to the International Classification of Diseases- Oncology (ICD-O-3) site codes C440-C449. Melanoma cases that had a specified stage at diagnosis were included for final analysis (N= 315). Stage of diagnosis was determined through the SEER Summary Stage 2000 variable and coded into in situ, localized, and combined regional & distant stage. Univariate and multivariate analyses were performed for the following predictor variables: insurance status (private insurance vs. other), age group (5- year groups), and sex (male vs. female). The majority of the sample was white (96.5%), female (63.8%), had private insurance (85%) and was diagnosed with localized stage melanoma (69.4%). Individuals with government insurance were eight times more likely to be diagnosed with late stage melanoma compared to individuals with private insurance (OR 8.4, CI 3.0-23.3, p < 0.01). AYAs in the 15-19 year old age group were six times more likely to be diagnosed with late stage melanoma compared to 35-39 year olds (OR 6.3, CI 1.7-22.9, p=0.01). Females were 57% less likely to be diagnosed with late stage melanoma compared to males (OR 0.53, CI 0.30-0.93, p < 0.05). These findings indicate that individuals with government insurance may not receive adequate melanoma screening and preventative care compared to individuals with private insurance. While females were less likely to be diagnosed with late stage melanoma, females have a much greater risk of being diagnosed with melanoma at any stage. Finally, the increased risk of late stage diagnosis in the 15-19 year old age group may be associated with greater UV exposure from indoor and outdoor tanning. These data suggest the need for targeted cancer awareness and control activities specific to AYAs. Future studies are needed to explore the variations in late stage diagnosis of melanoma in AYAs in Tennessee.

Suicide represents a major problem for all members of society. Findings from studies have shown that suicide risk is among the highest in young adult populations. Although a variety of factors have been identified in trying to predict suicide risk in young adults, one variable that has received considerable attention has been social disconnectedness. For example, according to Joiner’s (2005) theory of suicide, being socially disconnected from others represents a critical factor that heightens a person’s risk for suicide. Consistent with this view, findings from numerous studies have shown a reliable association between loneliness and suicide risk (Hatcher & Stubbersfield, 2013). For example, Chang and his colleagues (Chang et al., 2015; Chang, Muyan, & Hirsch, 2015; Muyan & Chang, 2015) have found a reliable association between loneliness and greater suicide risk (e.g., hopelessness, suicidal behaviors) in diverse cultural groups. Indeed, in one study, Chang, Hirsch, Sanna, Jeglic, and Fabian (2011) found that loneliness was an important predictor of suicide risk in Latina college students. Although these findings are important and point to the potential value of loneliness as a predictor of suicide risk in diverse adult groups, they do not take into account more culture- specific variables. Accordingly, we conducted the present study to examine the extent to which culture- specific (emic) variables might play an important role in predicting suicide risk, and to also determine if the inclusion of a general (etic) explanatory variable would add to the prediction of suicide risk, namely, hopelessness and suicidal behaviors, in a sample of 155 Latino college students. We conducted a pair of regression analyses including demographic variables (viz., age & sex) in the First Step, ethnic identity variables (ethnic affirmation, ethnic identity achievement, ethnic behaviors, & other group orientation; Phinney, 1992) in the Second Step, and loneliness (Russell, Peplau, & Curtrona, 1980) in the final Third Step, in predicting hopelessness and suicidal behaviors. Results of these analyses for both outcomes indicated that demographic variables did not account for significant variance in suicide risk. However, ethnic identity variables, as a set, were found to account for a significant 11% of the variance in hopelessness, and a significant 10% of the variance in suicide behaviors. Noteworthy, the inclusion of loneliness was found to account for an additional 19% of unique variance in hopelessness, and an additional 13% of unique variance in suicidal behaviors. Taken together, our findings underscore the importance of considering both emic and etic factors in trying to identify variables that may place Latinos at heightened risk for suicide. Additionally, our findings also point to the potential value of targeting specific ethnic identity variables (e.g., ethnic identity achievement) as well as loneliness when working with Latinos at risk for suicide.

Le paradigme actuel de pleine participation sociale place la lecture en compétence fondamentale. Dans ce contexte, notre thèse cherche à comprendre le développement de la lecture dans la déficience intellectuelle et d’approcher des interrelations possibles entre les différents domaines cognitifs et la lecture. Notre étude longitudinale suit le développement de la lecture d’une trentaine d’enfants non-lecteurs bénéficiant d’un dispositif Ulis Ecole dans l’Académie de Rouen. Le protocole expérimental évalue la lecture et des compétences générales et spécifiques impliquées en lecture durant trois années scolaires. Pour comparaison, une trentaine d’enfants non-lecteurs ne présentant pas de déficience intellectuelle est évaluée avec le même protocole de la grande section au CE1.Les séquences développementales de la lecture observées mettent en évidence l’importance de la connaissance du nom et du son des lettres dans le développement de la lecture dans la déficience intellectuelle, tandis que les compétences phonologiques se développeraient en appui sur ces connaissances de la langue écrite. Toutefois, ces enfants peuvent présenter des profils de lecteurs en développement identiques à ceux d’enfants au développement typique. Ces résultats questionnent la définition et le développement de la conscience phonologique, ses interrelations avec la mémoire phonologique et son impact sur le niveau de lecture ultérieur. Les retombées sociales de notre recherche visent un changement de représentation du potentiel de développement de la lecture dans la déficience intellectuelle pour une école inclusive où les besoins doivent guider la réponse à l’intervention et non le diagnostic. / Today, full social participation paradigm places reading in fundamental competence. In this context, our thesis seeks to understand reading development in intellectual disability and to approach possible interrelations between different cognitive domains and reading. Our longitudinal study follows reading development of thirty non-reading children in Ulis School in Rouen Academy. The experimental protocol measures reading skills and general and specific skills involved to read during three school years. For comparison, 30 non-readers children without intellectual disability are evaluated with the same protocol from preschool third year to primary school second year. The developmental sequences of reading observed further highlight the importance of knowledge of the name and sound of letters in the development of reading in intellectual disability, while phonological skills would develop due to this written language knowledge. However, these children may have similar developing reader profiles to those of typically developing children. These results question the definition and development of phonological awareness, its interrelations with phonological memory and its impact on the future level of reading. The social benefits of our research aim to change the representation of the reading developmental potential in intellectual disability for an inclusive school where needs must guide the response to the intervention and not the diagnosis.

The big data era poses great challenges as well as opportunities for researchers to develop efficient statistical approaches to analyze massive data. Sufficient dimension reduction is such an important tool in modern data analysis and has received extensive attention in both academia and industry. In this dissertation, we introduce inverse regression estimators using Fourier transforms, which is superior to the existing SDR methods in two folds, (1) it avoids the slicing of the response variable, (2) it can be readily extended to solve the high dimensional data problem. For the ultra-high dimensional problem, we investigate both eigenvalue decomposition and minimum discrepancy approaches to achieve optimal solutions and also develop a novel and efficient optimization algorithm to obtain the sparse estimates. We derive asymptotic properties of the proposed estimators and demonstrate its efficiency gains compared to the traditional estimators. The oracle properties of the sparse estimates are derived. Simulation studies and real data examples are used to illustrate the effectiveness of the proposed methods. Wavelet transform is another tool that effectively detects information from time-localization of high frequency. Parallel to our proposed Fourier transform methods, we also develop a wavelet transform version approach and derive the asymptotic properties of the resulting estimators.

Background: Interest in alternatives to hormone therapy (HT) for menopausal symptoms increased following its association with serious health risks. In terms of phytotherapeutic interventions, while traditional use supports a range of herbs for treating menopausal symptoms, evidence from rigorous scientific trials is limited, and has largely focused on the phytoestrogenic plants. Because of some safety concerns over long-term use of isoflavones, the present study focused on two non-estrogenic herbs, Hypericum perforatum and Vitex agnus-castus, also employed in this context in the Anglo-American and European traditions. Both herbs have shown effectiveness for the alleviation of symptoms of premenstrual syndrome (PMS), which is reported to be more severe during the perimenopause, and may account for many of the so-called 'menopausal symptoms' at this time. Research on menopausal vasomotor symptoms is prone to substantial placebo responses. There has been much interest in increasing our understanding of the placebo response with a view to controlling it in clinical research and harnessing it in clinical practice. Methods & Results: A double-blind, randomised controlled trial (RCT), with a 16-week treatment phase, was conducted on 100 late-perimenopause and early postmenopause women. The herbal combination (Hypericum and Vitex) was not found to be superior to placebo for any of the endpoints - daily weighted flushing scores, overall menopausal symptoms (on the Greene Climacteric Scale) and depression (on the Hamilton Depression Inventory). However, significant improvements across the treatment phase were observed in both arms for all of these outcome measures. No significant change was found for either group on the Utian quality of life scale. The effects of the herbal combination were also examined on PMS-like symptoms in the small sub-population of late-perimenopausal women, and found to be superior to placebo for total PMS-like symptoms and the sub-clusters, PMS-D (depression) and PMS-C (cravings). The active treatment group also showed significant improvements on PMS-A (anxiety) and PMS-H (hydration), although these effects were not superior to placebo. Predictors of the placebo response were investigated and found to include study-entry anxiety for the outcome measures of flushing, depression and overall menopausal symptoms, and improvement during non-treatment run-in for depression and overall symptoms. Because no difference had been found between 'active' and placebo groups in the menopause RCT, it was hypothesised that the same predictors would predict the response to the study treatment. However, low anxiety was significantly associated with improvement in this group. None of the other variables that predicted the placebo response was relevant to the study treatment response. This finding is discussed with reference to the possibility that 'drug' effects and placebo effects are not necessarily additive, and that the same magnitude of effect in both arms might not necessarily imply activity via the same pathways. Conclusions: This research contributes to the growing body of scientific knowledge about evidence-based complementary therapies that informs the community, health-care providers and regulatory authorities. The findings may facilitate identification of potential placebo responders in future research. The need for more research in the area of mechanisms of placebo versus active responses is supported.

Doctor of Philosophy (Phd) / Despite general agreement that severely emotionally disturbed children and adolescents are an "at risk" group, and that ongoing evaluation and research into the effectiveness of services provided for them is important, very little outcome evaluation actually takes place. The absence of well-conducted and appropriately interpreted studies is particularly notable for day or residential treatment programs, which cater for the most severely emotionally disturbed youths. This thesis outlines the main areas of conceptual, pragmatic and methodological confusion and neglect which impede progress in research in this area. It argues for plurality of data analytic strategies and research designs. It then critically reviews the reported findings about the effectiveness of day and residential treatment in specialist facilities, and the predictors of good outcomes for this treatment type. This review confirms that there is very little to guide practice. Having argued for the legitimacy of its methods and the necessity to address basic questions, the thesis reports the results of a naturalistic study based on data accumulated during a decade-long evaluative research program taking place at Arndell Child and Adolescent Unit, Sydney. The study addresses the question of what child, family and treatment variables predict outcome for 159 children and adolescents treated at this facility from 1990 to 1999. Statistically significant results with large effect size were obtained. Among the most disturbed subgroup of forty three children, (a) psychodynamic milieu-based treatment was shown to be more effective than the “empirically-validated” cognitive-behavioural treatment which superseded it in 1996, and (b) children from step-families showed better outcome than those from other family structures. Furthermore, it was found for the study sample as a whole that severe school-based problem behaviours were associated with a limited trajectory of improvement in home-based problem behaviour. These results are discussed with regard to implications for treatment, research methodology, policy and further studies.

<p>This thesis investigates health trends among very old people and the allocation of public home-help services. A further aim is to examine methodological issues in mortality analysis. Three data sources are used: (1) The Tierp study of community-dwelling persons (n=421, ages 75+), (2) the SWEOLD nationally representative samples (n=537 in 1992 and n=561 in 2002, ages 77+), and 3) SNAC-K comprised of home-help recipients in a district of Stockholm (n=1108, ages 65+).</p><p>Study I suggests that the length of the follow-up period may explain some of the differences found in predictor strength when comparing mortality studies. Predictors that can change rapidly (e.g., health) were found to be strongest for the short term, with a lower average mortality risk for longer follow-ups. Stable variables (e.g., gender) were less affected by length of follow-up.</p><p>Studies II and III present a measure of complex health problems based on serious problems in at least two of three health domains. These were diseases/symptoms, mobility, and cognition/communication. Prevalence of complex health problems increased significantly between 1992 and 2002. Older age, female gender, and lower education increased the odds of having complex problems. Complex problems strongly predicted 4-year mortality. Controlled for age, gender, health, and education, mortality decreased by 20% between 1992 and 2002. Men with complex problems accounted for this decrease. Thus, in 2002 the gender difference in mortality risk was almost eliminated among the most vulnerable adults.</p><p>Study IV revealed that physical and cognitive limitations, higher age, and living alone were significantly related to home-help allocation, with physical and cognitive limitations dominating. Psychiatric symptoms did not affect the assessment.</p><p>The increased prevalence of complex health problems and increased survival among people with complex needs have important implications concerning the need for collaboration among service providers.</p>