Estimating marginal causal effects in a secondary analysis of case‐control data

When an initial case‐control study is performed, data can be used in a secondary analysis to evaluate the effect of the case‐defining event on later outcomes. In this paper, we study the example in which the role of the event is changed from a response variable to a treatment of interest. If the aim is to estimate marginal effects, such as average effects in the population, the sampling scheme needs to be adjusted for. We study estimators of the average effect of the treatment in a secondary analysis of matched and unmatched case‐control data where the probability of being a case is known. For a general class of estimators, we show the components of the bias resulting from ignoring the sampling scheme and demonstrate a design‐weighted matching estimator of the average causal effect. In simulations, the finite sample properties of the design‐weighted matching estimator are studied. Using a Swedish diabetes incidence register with a matched case‐control design, we study the effect of childhood onset diabetes on the use of antidepressant medication as an adult. Copyright © 2017 John Wiley & Sons, Ltd.     

Pherotypes of pneumococcal strains co-existing in healthy children

Genetic diversity in the species Streptococcus pneumoniae is mainly driven by horizontal gene transfer. S. pneumoniae is naturally competent for transformation. Competence is induced by a pheromone termed competence stimulating peptide (CSP) by a quorum-sensing mechanism. Two CSP pherotypes predominate amongst clinical isolates of S. pneumoniae, CSP-1 and CSP-2, with ability to trigger competence in bacteria of the homologue pherotype. Opposing theories on the effect of pherotypes on speciation have been proposed, either as a barrier for intra-pherotype gene transfer, or as a mechanism for fratricide resulting in lysis of non-competent bacterial cells.The aim of the present study was to determine pherotype distribution in strains of S. pneumococci isolated from the nasopharynges of healthy children. We sequenced the locus encoding CSP, comC, in sets of strains obtained from children colonised by multiple pneumococcal strains simultaneously. The impact of pherotype on co-colonisation was determined by comparing the observed distribution of pherotypes in co-colonising strains with the estimated pair-wise probability based on the overall pherotype distribution in the sample set.Five distinct comC alleles were identified, encoding CSP belonging to the two dominating pherotypes, CSP-1 (62.7%) and CSP-2 (37.3%). The observed distribution of pherotypes in sets of co-colonising pneumococcal strains did not differ from the probability estimate. Thus, co-colonisation of S. pneumoniae in healthy children is not restricted by pherotype.     

Collaboration between general hospitals and community health services in the care of suicide attempters in Norway: a longitudinal study

Background  

The aim of this paper was to study the collaboration between emergency departments (EDs) in general hospitals and community health services (CHS) in Norway when providing psychosocial care and aftercare to patients treated in EDs following a suicide attempt. We wanted to explore the extent to which quality indicators at the hospital level measured in 1999 and 2006 could predict the presence or absence of a chain of care structure in the CHS in 2006.     

Service quality in hospital wards with different nursing organization: nurses' ratings

Title. Service quality in hospital wards with different nursing organization: nurses’ ratings. Aim. This paper is a report of a study to assess: (1) the relations between nursing organization models in hospital wards and nurses’ perception of the quality of patient care and dimensions of the practice environment, and (2) if these relations were modified by variations in local conditions at the ward level. Background. Previous literature is inconclusive concerning what model of nursing organization maximizes the quality of nursing services. Method. A cross‐sectional survey was carried out in a representative sample of Norwegian hospital wards in 2005. Intra‐ward organization models were classified as: (1) Team leader (n = 30), characterized by extensive responsibilities for team leaders, (2) Primary nurse (n = 18), with extensive responsibilities for named nurses, and (3) Hybrid (n = 37), (1) and (2) combined. We prepared multilevel regression models using scales describing quality of patient care, learning climate, job satisfaction, and relationships with physicians as dependent variables. As independent variables, we used variables representing local ward conditions. Results. Eighty‐seven wards and 1137 nurses (55% response rate) provided complete data. The ward level proportion of variance ranged from 0·10 (job satisfaction) to 0·22 (relationships with physicians). The univariate effect of organization models on quality ratings was not statistically significant. Introducing local ward conditions led to a statistically significant effect of primary nurse organization on relationships with physicians, and to a substantial proportional reduction in ward level variance, ranging from 32% (quality of patient care) to 24% (learning climate). Conclusion. Caution is needed about using service quality arguments when considering the possible benefits and drawbacks of different organizational models.     

Correlation between ocular pulse amplitude measured by dynamic contour tonometer and visual field defects

Purpose To investigate the correlation between ocular pulse amplitude and visual field defects in patients with glaucoma, ocular hypertension, and glaucoma suspicion when measured with the Pascal Dynamic Contour Tonometer, and to verify if the ocular pulse amplitude is an independent predictor for visual field parameters. Methods Seventy-seven eyes (42 patients) with glaucoma, ocular hypertension or glaucoma suspicion were examined. Ocular pulse amplitude was measured with the dynamic contour tonometer by one investigator masked to the visual field data. Visual fields were performed within three months of ocular pulse amplitude measurement by the Octopus or Humphrey Field Analyser, and were analysed with Peridata Software. Mean defect, pattern standard deviation (√ Loss Variance) and regression analysis of those parameters (Trend Indices) were correlated with the ocular pulse amplitude for each eye. Results Forty-nine eyes had glaucoma, 14 had ocular hypertension, and 14 were glaucoma suspects. The mean follow-up was 46.5 (range 6–96) months. There was a significant correlation between OPA and MD and OPA and PSD, even after correction for IOP (and diagnostic group and eye): the estimated slope equals 2.68 (S.E. = 0.82, p = 0.003) and −0.86 (S.E. = 0.33, p = 0.014), respectively. There was even a weak correlation between OPA and the evolution of MD (dB/year). The slope estimate for OPA equals 0.070 (S.E. = 0.033), p = 0.037. However, after correction for IOP (and diagnostic group and eye), the strength of the relationship is reduced and the evidence disappears: the slope estimate for OPA now equals 0.039 (S.E. = 0.041), p = 0.34. There is no evidence for an association between OPA and the evolution of Trend-PSD. Conclusion A small ocular pulse amplitude, as measured with a dynamic contour tonometer, is correlated with moderate to severe glaucomatous visual field loss and might be a risk factor for the development of glaucomatous visual field defects.     

Treatment of hypercholesterolemia in children and adolescents

In patients with severe genetic hypercholesterolemia, therapeutic reduction of elevated serum total cholesterol and LDL cholesterol should begin in early childhood to lower the risks of cardiovascular disease later in life. We evaluated the effects of outpatient therapy with diet alone and with combined diet and drug therapy in children and adolescents with hypercholesterolemia of apparent dominant inheritance. Serum lipid values before and during dietary treatment were available in 35 patients (mean age at start of treatment 7.9 years, range 2.0-17.6 years) followed for an average duration of 17.5 months (range 4-70 months). A comparison between untreated state and combined therapy with diet and cholestyramine was possible in 14 patients (mean age 8.6 years, range 2.4-17.0 years) followed for 27.9 months (range 4-97 months). Dietary modification achieved by repeated counseling and training lowered serum total cholesterol by mean (+/- SE) 11.7 +/- 1.9% (p < 0.0001) and LDL cholesterol by 17.3 +/- 3.5% (p < 0.0001). However, five of 35 patients did not show an appreciable effect of therapy (cholesterol reduction < 5%), possibly because of non-compliance. Diet combined with cholestyramine in an average dose of 0.36 g/kg body weight/day reduced total cholesterol by 33.0 +/- 2.4% (p < 0.0001) and LDL cholesterol by 37.5 +/- 4.3% (p < 0.0001) and was effective in all patients. Both forms of treatment had no effect on serum triglycerides and HLD cholesterol. No serious side effects were noted, and percentile values for weight and height remained unchanged in all but three obese children.(ABSTRACT TRUNCATED AT 250 WORDS)     

Determinants of outcome in ANCA-associated glomerulonephritis: a prospective clinico-histopathological analysis of 96 patients

BACKGROUND: The predictive value of clinical and renal histological features for renal outcome in patients with anti-neutrophil cytoplasmic autoantibody (ANCA)-associated glomerulonephritis was investigated in a prospective analysis of 96 patients with ANCA-associated vasculitis, and moderate renal involvement (creatinine <500 micromol/L). METHODS: The extent of 39 histological features in 96 biopsies (performed at entry in a clinical trial) was scored by two independent observers, according to a standardized protocol. Age, gender, diagnosis, glomerular filtration rate at entry (GFR0), ANCA-specificity, proteinuria, and treatment of these 96 patients were also taken into account. Treatment was standardized and started after the biopsy was performed. End-points included renal function at 18 months (GFR18), GFR18 corrected for GFR0 (CORGFR18), and the occurrence of relapse or death. RESULTS: Parameters that most strongly correlated with GFR18 were GFR0 (r = 0.67), interstitial fibrosis (r = -0.45), glomerulosclerosis (r = -0.37), and tubular atrophy (r = -0.36). Parameters that most strongly correlated with CORGFR18 were segmental (r = 0.45) and cellular (r = 0.30) crescents, and fibrinoid necrosis (r = 0.46). None of the clinical and histological features predicted the occurrence of relapse or death. By applying a stepwise linear multiple regression analysis, we designed a formula for the estimation of renal function at 18 months: GFR18 (mL/min) = 17 + 0.71 x GFR0 (mL/min) + 0.34 x fibrinoid necrosis (%) + 0.33 x segmental crescents (%), (r2 = 0.60; standard deviation = 19 mL/min). Our results were independent of diagnosis, ANCA-specificity, and treatment limb. CONCLUSIONS: These data suggest that in ANCA-associated glomerulonephritis, GFR0 and predominantly chronic renal lesions are potent predictors of GFR18. Active lesions are associated with renal function recovery and may be reversible. The formula for the estimation of GFR18 shows that a combination of GFR0 and renal histology is a better predictor for GFR18 than GFR0 only.