Clayton copula for survival data with dependent censoring: An application to a tuberculosis treatment adherence data.

Ignoring the presence of dependent censoring in data analysis can lead to biased estimates, for example, not considering the effect of abandonment of the tuberculosis treatment may influence inferences about the cure probability. In order to assess the relationship between cure and abandonment outcomes, we propose a copula Bayesian approach. Therefore, the main objective of this work is to introduce a Bayesian survival regression model, capable of taking into account the dependent censoring in the adjustment. So, this proposed approach is based on Clayton's copula, to provide the relation between survival and dependent censoring times. In addition, the Weibull and the piecewise exponential marginal distributions are considered in order to fit the times. A simulation study is carried out to perform comparisons between different scenarios of dependence, different specifications of prior distributions, and comparisons with the maximum likelihood inference. Finally, we apply the proposed approach to a tuberculosis treatment adherence dataset of an HIV cohort from Alvorada-RS, Brazil. Results show that cure and abandonment outcomes are negatively correlated, that is, as long as the chance of abandoning the treatment increases, the chance of tuberculosis cure decreases.

Marker events associated with adherence to HIV/AIDS treatment in a cohort study.

OBJECTIVE: To analyze how clinical and social events may impact adherence to antiretroviral treatment for HIV. METHODS: This is a historical cohort study with 528 patients who underwent treatment for HIV in a specialized care service in Alvorada, RS. A total of 3429 queries executed between the years 2004 and 2017 were analyzed. For each visit, data on treatment characteristics and the patients' clinical picture were collected. Adherence, as measured by patients' self-report, was the endpoint of the study. The logistic regression model via generalized estimating equations was used for estimating the associations. RESULTS: 67.8% of the patients analyzed have up to 8 years of education and 24.8% have a history of crack and/or cocaine use. Among men, being asymptomatic [odds ratio (OR) = 1.43; 95%CI 1.05-1.93], having more than 8 years of education (OR= 2.32; 95%CI 1.27-4.23), and never having used crack (RC = 2.35; 95%CI 1.20-4.57) were associated with adherence. For women, being older than 24 years (CR = 1.82; 95%CI 1.09-3.02), never having used cocaine (CR = 2.54; 95%CI 1.32-4.88) and being pregnant (RC = 3.28; 95%CI 1.83-5.89) increased the odds of adherence. CONCLUSIONS: In addition to defined sociodemographic characteristics, one-off events that may occur in the trajectory of patients on long treatment, such as starting a new pregnancy and not having symptoms, can impact patients' chances of treatment adherence.

Marker events associated with adherence to HIV/AIDS treatment in a cohort study / Eventos marcadores associados à adesão ao tratamento para HIV/aids em um estudo de coorte

ABSTRACT OBJECTIVE To analyze how clinical and social events may impact adherence to antiretroviral treatment for HIV. METHODS This is a historical cohort study with 528 patients who underwent treatment for HIV in a specialized care service in Alvorada, RS. A total of 3429 queries executed between the years 2004 and 2017 were analyzed. For each visit, data on treatment characteristics and the patients' clinical picture were collected. Adherence, as measured by patients' self-report, was the endpoint of the study. The logistic regression model via generalized estimating equations was used for estimating the associations. RESULTS 67.8% of the patients analyzed have up to 8 years of education and 24.8% have a history of crack and/or cocaine use. Among men, being asymptomatic [odds ratio (OR) = 1.43; 95%CI 1.05-1.93], having more than 8 years of education (OR= 2.32; 95%CI 1.27-4.23), and never having used crack (RC = 2.35; 95%CI 1.20-4.57) were associated with adherence. For women, being older than 24 years (CR = 1.82; 95%CI 1.09-3.02), never having used cocaine (CR = 2.54; 95%CI 1.32-4.88) and being pregnant (RC = 3.28; 95%CI 1.83-5.89) increased the odds of adherence. CONCLUSIONS In addition to defined sociodemographic characteristics, one-off events that may occur in the trajectory of patients on long treatment, such as starting a new pregnancy and not having symptoms, can impact patients' chances of treatment adherence.

RESUMO OBJETIVO Analisar como eventos clínicos e sociais podem impactar na adesão ao tratamento antirretroviral para o HIV. MÉTODOS Trata-se de um estudo de coorte histórica com 528 pacientes que realizaram o tratamento para o HIV em um serviço de assistência especializada em Alvorada, RS. Foram analisadas 3429 consultas executadas entre os anos de 2004 e 2017. Para cada consulta, foram coletados dados de características do tratamento e do quadro clínico dos pacientes. A adesão, aferida pelo autorrelato dos pacientes, foi o desfecho do estudo. O modelo de regressão logística via equações de estimação generalizadas foi utilizado para estimação das associações. RESULTADOS 67,8% dos pacientes analisados possuem até 8 anos de estudos e 24,8% têm histórico de uso de crack e/ou cocaína. Entre os homens, estar assintomático [razão de chances (RC) = 1,43; IC95% 1,05-1,93], possuir mais de 8 anos de estudo (RC = 2,32; IC95% 1,27-4,23) e nunca ter usado crack (RC = 2,35; IC95% 1,20-4,57) estiveram associados à adesão. Para as mulheres, possuir mais de 24 anos (RC = 1,82; IC95% 1,09-3,02), nunca ter usado cocaína (RC = 2,54; IC95% 1,32-4,88) e estar em gestação (RC = 3,28; IC95% 1,83-5,89) aumentaram as chances de adesão. CONCLUSÕES Além de características sociodemográficas definidas, eventos pontuais que podem ocorrer na trajetória de pacientes em tratamentos longos, como início de uma nova gestação e não apresentar sintomas, podem impactar nas chances de adesão dos pacientes ao tratamento.

Dual and Triple Infections With Influenza A and B Viruses: A Case-Control Study in Southern Brazil.

Influenza surveillance is important for disease control and should consider possible coinfection with different viruses, which can be associated with disease severity. This study analyzed 34 459 patients with respiratory infection from 2009 to 2018, of whom 8011 were positive for influenza A virus (IAV) or influenza B virus (IBV). We found 18 cases of dual influenza virus infection, including coinfection with 2009 pandemic influenza A(H1N1) virus (A[H1N1]pdm09) and influenza A(H3N2) virus (1 case), A(H1N1)pdm09 and IBV (6 cases), A(H3N2) and IBV (8 cases), and nonsubtyped IAV and IBV (3 cases); and 1 case of triple infection with A(H3N2), A(H1N1)pdm09, and IBV. Compared with 76 monoinfected patients, coinfection was significantly associated with cardiopathy and death. Besides demographic characteristics and clinical symptoms, we assessed vaccination status, antiviral treatment, timeliness of antiviral use, hospitalization, and intensive care unit admission, but no significant differences were found between coinfected and monoinfected cases. Our findings indicate that influenza virus coinfection occurs more often than previously reported and that it can lead to a worse disease outcome.

Incidence of hepatocellular carcinoma in patients with chronic liver disease due to hepatitis B or C and coinfected with the human immunodeficiency virus: A retrospective cohort study.

AIM: To assess the incidence of hepatocellular carcinoma (HCC) in chronic liver disease due to hepatitis B virus (HBV) or hepatitis C virus (HCV) coinfected with human immunodeficiency virus (HIV). METHODS: A retrospective cohort study was performed, including patients with chronic liver disease due to HBV or HCV, with and without HIV coinfection. Patients were selected in the largest tertiary public hospital complex in southern Brazil between January 2007 and June 2014. We assessed demographic and clinical data, including lifestyle habits such as illicit drug use or alcohol abuse, in addition to frequency and reasons for hospital admissions via medical records review. RESULTS: Of 804 patients were included (399 with HIV coinfection and 405 monoinfected with HBV or HCV). Coinfected patients were younger (36.7 ± 10 vs 46.3 ± 12.5, P < 0.001). Liver cirrhosis was observed in 31.3% of HIV-negative patients and in 16.5% of coinfected (P < 0.001). HCC was diagnosed in 36 patients (10 HIV coinfected and 26 monoinfected). The incidence density of HCC in coinfected and monoinfected patients was 0.25 and 0.72 cases per 100 patient-years (95%CI: 0.12-0.46 vs 0.47-1.05) (long-rank P = 0.002), respectively. The ratio for the HCC incidence rate was 2.98 for HIV-negative. However, when adjusting for age or when only cirrhotic are analyzed, the absence of HIV lost statistical significance for the development of HCC. CONCLUSION: In this study, the presence of HIV coinfection in chronic liver disease due to HBV or HCV showed no relation to the increase of HCC incidence.