D.G. Nel

A solution to the multivariate behrens-fisher problem

Some new algebra on pattern and transition matrices is used to determine the degrees of freedom and the parameter matrix, if the distribution of a linear sum of Wishart matrices is approximated by a single Wishart distribution. This approximation is then used to find a solution to the multivariate Behrens-Fisher problem similar to the Welch (1947) solution in the univariate case.

Obsessive-compulsive disorder and trichotillomania: a phenomenological comparison

BackgroundSimilarities between obsessive-compulsive disorder (OCD) and trichotillomania (TTM) have been widely recognized. Nevertheless, there is evidence of important differences between these two disorders. Some authors have conceptualized the disorders as lying on an OCD spectrum of conditions.MethodsTwo hundred and seventy eight OCD patients (n = 278: 148 male; 130 female) and 54 TTM patients (n = 54; 5 male; 49 female) of all ages were interviewed. Female patients were compared on select demographic and clinical variables, including comorbid axis I and II disorders, and temperament/character profiles.ResultsOCD patients reported significantly more lifetime disability, but fewer TTM patients reported response to treatment. OCD patients reported higher comorbidity, more harm avoidance and less novelty seeking, more maladaptive beliefs, and more sexual abuse. OCD and TTM symptoms were equally likely to worsen during menstruation, but OCD onset or worsening was more likely associated with pregnancy/puerperium.ConclusionsThese findings support previous work demonstrating significant differences between OCD and TTM. The classification of TTM as an impulse control disorder is also problematic, and TTM may have more in common with conditions characterized by stereotypical self-injurious symptoms, such as skin-picking. Differences between OCD and TTM may reflect differences in underlying psychobiology, and may necessitate contrasting treatment approaches.

A matrix derivation of the asymptotic covariance matrix of sample correlation coefficients

Abstract The asymptotic covariance matrix of the sample correlation matrix is derived in matrix form as an application of some new matrix theory in multivariate statistics.

Acute renal failure in the medical ICU still predictive of high mortality

BACKGROUND We aimed to determine the outcome and certain predictors of outcome for acute renal failure (ARF) in the medical intensive care unit (ICU) at Tygerberg Hospital. METHOD We conducted a retrospective, single-centre cohort study over 12 months comprising all patients admitted to the medical ICU with all causes of renal failure or who developed renal failure following admission to the ICU. RESULTS Of 198 medical patients admitted to the ICU, ARF occurred in 46 (23.2%). The leading cause of ARF was acute tubular necrosis. The ICU mortality for ARF patients was 47.8%, compared with 17.5% in ICU patients without ARF. Acute haemodialysis was performed in only 17.3% of the 46 ARF patients. Using Cox proportional hazard regression, we found that mean duration of stay (p<0.001), acute physiology and chronic health evaluation II (Apache II) score (p<0.001), mechanical ventilation (p<0.01), dialysis (p<0.04) and multi-organ failure (p<0.05) affected survival time. CONCLUSIONS We found that ARF is still associated with a high mortality rate and longer duration of stay, higher Apache II score, and need for mechanical ventilation; dialysis and presence of multi-organ failure were indicators of a higher mortality rate.

Probiotics, feeding tolerance, and growth: A comparison between HIV-exposed and unexposed very low birth weight infants

OBJECTIVE The aim of this study was to compare the effect of administration of probiotics on feeding tolerance and growth outcomes of HIV-exposed (but uninfected) versus HIV non-exposed preterm infants. The null hypothesis of this study states that there will be no difference in the feeding tolerance and growth outcomes for both probiotic-exposed and unexposed premature very low birth weight infants. METHODS A randomized, double-blind, placebo-controlled trial was conducted during the period from July 2011 to August 2012. HIV-exposed and non-exposed premature (<34 wk gestation) infants with a birth weight of ≥500 g and ≤1250 g were randomized to receive either a probiotic mixture or placebo. The multispecies probiotic mixture consisted of 1 × 10(9) CFU, Lactobacillus rhamnosus GG and Bifidobacterium infantis per day and was administered for 28 d. Anthropometrical parameters, daily intakes, and feeding tolerance were monitored. RESULTS Seventy-four HIV-exposed and 110 unexposed infants were enrolled and randomized (mean birth weight 987 g ± 160 g, range, 560-1244 g; mean gestational age 28.7 wk). In all 4227 probiotic doses were administered (mean 22.9/infant). There was no difference in the average daily weight gain for treatment groups or HIV exposure. The HIV-exposed group achieved significantly higher z scores for length and head circumference at day 28 than the unexposed group (P < 0.01 and P = 0.03, respectively). There were no differences in the incidence of any signs of feeding intolerance and abdominal distension between the groups. CONCLUSION Probiotic supplementation did not affect growth outcomes or the incidence of any signs of feeding intolerance in HIV exposure.

Depression and anxiety in multisomatoform disorder: Prevalence and clinical predictors in primary care

OBJECTIVE Multisomatoform disorder (MSD) is characterised by > or = 3 medically inexplicable, troublesome physical symptoms, together with a > or = 2-year history of somatisation. The aim of this study was to evaluate the prevalence of depressive and anxiety disorders in a South African sample MSD, and to compare demographic and clinical outcomes in those patients with and without co-morbidity. METHODS Fifty-one adult outpatients with MSD were recruited from primary care clinics in the Cape Town metropolitan area. Participants were assessed for the presence of co-morbid depressive and anxiety disorders using the Mini Neuropsychiatric Interview-Plus (MINI-Plus). Outcomes included somatic symptom severity, disability, reported sick days and health care visits, pain experience, patient satisfaction with health services, and clinician-experienced difficulty. RESULTS A current co-morbid depressive disorder was present in 29.4% (N = 15) of patients, and a current co-morbid anxiety disorder in 52.9% (N = 27). MSD patients with a co-morbid depressive disorder (current or lifetime) had significantly higher physical symptom counts, greater functional impairment, higher unemployment rates, more clinician-reported difficulties, and more dissatisfaction with health care services than those without the disorder. A larger number of co-morbid disorders was associated with greater overall disability. CONCLUSION High rates of co-morbid depressive and anxiety disorders were present in a South African sample of primary care patients with MSD. Not all patients had co-morbidity, which is consistent with the view that MSD should be viewed as an independent disorder. However, co-morbid depressive disorders were associated with increased symptom severity and functional impairment, consistent with previous reports from developing countries, emphasising the importance of comorbidity in MSD.

Probiotics Reduce Necrotizing Enterocolitis Severity in HIV-exposed Premature Infants

OBJECTIVE To assess the effect of probiotics on the incidence of necrotizing enterocolitis (NEC) in premature infants born to human immunodeficiency virus (HIV)-positive and HIV-negative women. PATIENTS AND METHODS HIV-exposed and HIV-unexposed premature infants were randomized to either the probiotic or the placebo group. The probiotic consisted of 1 × 10(9) colony-forming units, Lactobacillus rhamnosus GG and Bifidobacterium infantis per day. RESULTS In total, 74 HIV-exposed and 110 HIV-unexposed infants were enrolled and randomized. The incidence of death [4 (5.4%) vs. 7 (6%); p = 0.79] and NEC [4 (5%) vs. 5 (5%); p = 0.76] did not differ significantly between the HIV-exposed and HIV-unexposed groups. A significant difference was found for total NEC incidence between the study and control groups [3 (3%) vs. 6 (6%); p = 0.029]. The incidence of NEC in the HIV-exposed group differed significantly [Bells I 2 (5%) vs. Bells III 2 (5%); p = 0.045). CONCLUSION Probiotic supplementation reduced the incidence of NEC in the premature very low birth weight infants; however, results failed to show a lower incidence of NEC in HIV-exposed premature infants. A reduction in the severity of disease was found in the HIV-exposed study group.

Fetal heart rate patterns at 20 to 24 weeks gestation as recorded by fetal electrocardiography

Abstract Introduction: With advancing technology it has become possible to accurately record and assess fetal heart rate (FHR) patterns from gestations as early as 20 weeks. The aim of our study was to describe early patterns of FHR, as recorded by transabdominal fetal electrocardiogram according to the Dawes–Redman criteria. Accordingly, short-term variability, basal heart rate, accelerations and decelerations were quantified at 20–24 weeks gestation among women with uncomplicated pregnancies. Methods: This study was conducted in a subset of participants enrolled in a large prospective pregnancy cohort study. Our final data set consisted of 281 recordings of women with good perinatal outcomes who had undergone fetal electrocardiographic assessment as part of the Safe Passage Study. Results: The success rate of the recordings was 95.4%. The mean frequency of small and large accelerations was 0.5 and 0.1 per 10 min, respectively and that of small and large decelerations 0.3 and 0.008 per 10 min, respectively. The mean and basal heart rates were both equal to 148.0 bpm at a median gestation of 161 days. The mean short term variation was 6.2 (SD 1.4) ms and mean minute range 35.1 (SD 7.1) ms. Conclusion: The 20–24-week fetus demonstrates FHR patterns with more accelerations and decelerations, as well as higher baseline variability than was anticipated. Information from this study provides an important foundation for further, more detailed, studies of early FHR patterns.

Predicting mortality and length-of-stay for neonatal admissions to private hospital neonatal intensive care units: a Southern African retrospective study

OBJECTIVES To predict neonatal mortality and length of stay (LOS) from readily available perinatal data for neonatal intensive care unit (NICU) admissions in Southern African private hospitals. METHODS Retrospective observational study using perinatal data from a large multicentre sample. Fifteen participating NICU centres in the Medi-Clinic private hospital group in Southern Africa. We used 2376 infants born between 1 January - 31 December 2008 to build the regression models, and a further 1 578 infants born between 1 January - 31 December 2007 to test the models. Outcome measures were mortality and length of hospital stay for NICU admissions. RESULTS Of the infants included in the 2008 dataset, ninety-one (3.8%) died after being admitted to NICU centres. The median LOS for non-transferred survivors was 11 days. An analysis of the structural peculiarities of the data showed high correlations between groups of the perinatal variables pertaining to the size and Apgar scores of the newborn infants, respectively. The logistic regression model to predict neonatal mortality had a good fit (AUC: 0.8507, misclassification rate: 13.6%), but the low positive predictive value of this model reduces its usefulness. The poisson log-linear model to predict LOS had a good fit (predicted R(2): 0.7027). CONCLUSIONS Apgar score at one minute, birth weight, and delivery mode significantly influence the odds of neonatal death and are associated with significant effects on LOS.

A predictive approach to the detection of additional information in a multivariate regression model

A predictive approach for the detection of additional information in a multivariate linear regression model is considered for the case of known and unknown error covariance matrices. The predictive density of future Observations on the additional variables under the model that they carry no information has been compared with the predictive density under the model that they do carry information. The Kullback-Leibler measure of divergence is used as a measure of comparison between the models.