• An alternative classification to mixture modeling for longitudinal counts or binary measures

      Subtil, F; Boussari, O; Bastard, M; Etard, J-F; Ecochard, R; Génolini, C (SAGE Publications, 2014-09-01)
      Classifying patients according to longitudinal measures, or trajectory classification, has become frequent in clinical research. The k-means algorithm is increasingly used for this task in case of continuous variables with standard deviations that do not depend on the mean. One feature of count and binary data modeled by Poisson or logistic regression is that the variance depends on the mean; hence, the within-group variability changes from one group to another depending on the mean trajectory level. Mixture modeling could be used here for classification though its main purpose is to model the data. The results obtained may change according to the main objective. This article presents an extension of the k-means algorithm that takes into account the features of count and binary data by using the deviance as distance metric. This approach is justified by its analogy with the classification likelihood. Two applications are presented with binary and count data to show the differences between the classifications obtained with the usual Euclidean distance versus the deviance distance.
    • Association of prevalence of active transport to work and incidence of myocardial infarction: A nationwide ecological study

      Munyombwe, T; Lovelace, R; Green, M; Norman, P; Walpole, S; Hall, M; Timmis, A; Batin, P; Brownlee, A; Brownlee, J; et al. (SAGE Publications, 2020-12-18)
      Background There is a paucity of population-based geospatial data about the association between active transport and myocardial infarction. We investigated the association between active transport to work and incidence of myocardial infarction. Design This ecological study of 325 local authorities in England included 43,077,039 employed individuals aged 25–74 years (UK Census, 2011), and 117,521 individuals with myocardial infarction (Myocardial Ischaemia National Audit Project, 2011–2013). Methods Bayesian negative binomial regression models were used to investigate the association of active transport to work and incidence of myocardial infarction adjusting for local levels of deprivation, obesity, smoking, diabetes and physical activity. Results In 2011, the prevalence of active transportation to work for people in employment in England aged 25–74 years was 11.4% (4,531,182 active transporters; 8.6% walking and 2.8% cycling). Active transport in 2011 was associated with a reduced incidence of myocardial infarction in 2012 amongst men cycling to work (incidence rate ratio (95% credible interval) 0.983 (0.967–0.999); and women walking to work (0.983 (0.967–0.999)) after full adjustments. However, the prevalence of active transport for men and women was not significantly associated with the combined incidence of myocardial infarction between 2011–2013 after adjusting for physical activity, smoking and diabetes. Conclusions In England, the prevalence of active transportation was associated with a reduced incidence of myocardial infarction for women walking and men cycling to work in corresponding local geographic areas. The overall association of active transport with myocardial infarction was, however, explained by local area levels of smoking, diabetes and physical activity.
    • A Bayesian Response-Adaptive Trial in Tuberculosis: The endTB Trial

      Cellamare, M; Ventz, S; Baudin, E; Mitnick, CD; Trippa, L (SAGE Publications, 2016-08-23)
      To evaluate the use of Bayesian adaptive randomization for clinical trials of new treatments for multidrug-resistant tuberculosis.
    • The field is ever further: In search of the elusive space of fieldwork

      Stellmach, D (SAGE Publications, 2020-01-07)
      This short reflection considers how humanitarian workers conceptualize and practice “the field” as a site of action. Through the use of narrative ethnography, and drawing on comparisons with the practice of academic anthropology, it attempts to draw out disciplinary assumptions that govern how and where humanitarian action is undertaken. It demonstrates how the field is a central imaginary that underpins the principles and performance of both anthropology and humanitarian action. It highlights how the conceptualization of “the field” is itself a methodological tool in the practice of humanitarian intervention.
    • Pervasive refusal syndrome in child asylum seekers on Nauru.

      Newman, L; O'Connor, B; Reynolds, V; Newhouse, G (SAGE Publications, 2020-09-10)
      Objectives: Between 2013 and 2019, an estimated 200 children seeking asylum in Australia were detained on the island of Nauru. In 2018, 15 of these children developed the rare and life-threatening pervasive refusal syndrome (PRS). This paper describes the PRS case cluster, the complexities faced by clinicians managing these cases, and the lessons that can be learned from this outbreak. Conclusions: The emergence of PRS on Nauru highlighted the risks of long-term detention of children in settings that are unable to meet their physical and psycho-social needs. The case cluster also underscored (a) the difficulties faced by doctors working in conditions where their medical and legal obligations may be in direct conflict, and (b) the role of clinicians in patient advocacy.
    • Validation of mental health screening instruments in the Kashmir Valley, India

      Housen, T; Lenglet, A; Ariti, C; Ara, S; Shah, S; Dar, M; Hussain, A; Paul, A; Wagay, Z; Viney, K; et al. (SAGE Publications, 2018-03-19)
      The present study aimed to culturally adapt, translate, and validate the Hopkins Symptom Checklist-25 (HSCL-25) and the Harvard Trauma Questionnaire-Posttraumatic Stress Symptoms Checklist (HTQ-16) prior to use in a cross-sectional mental health population survey in the Kashmir Valley. Cultural adaptation and translation of the HSCL-25 and the HTQ-16 employed multiple forms of transcultural validity check. The HSCL-25 and HTQ-16 were compared against a "gold standard" structured psychiatric interview, the Mini International Neuropsychiatric Interview (MINI). Interviews were conducted with 290 respondents recruited using consecutive sampling from general medical outpatient departments in five districts of the Kashmir Valley. Receiver operating characteristics (ROC) analysis was used to estimate the cut point with optimal discriminatory power based on sensitivity and specificity. Internal reliability of the HSCL-25 was high, Cronbach's alpha (α) = .92, intraclass correlation coefficient (ICC) = 0.75, with an estimated optimal cut point of 1.50, lower than the conventional cut point of 1.75. Separation of the instruments into subscales demonstrated a difference in the estimated cut point for the anxiety subscale and the depression subscale, 1.75 and 1.57, respectively. Too few respondents were diagnosed with posttraumatic stress disorder (PTSD) during structured psychiatric interview, and therefore the HTQ-16 could not be validated despite the fact that high internal reliability was demonstrated (α = .90). This study verified the importance of culturally adapting and validating screening instruments in particular contexts. The use of the conventional cut point of 1.75 would likely have misclassified depression in our survey, leading to an underestimate of this condition.