On Graphically Checking Goodness-of-fit of Binary Logistic Regression Models

OBJECTIVES: This paper is concerned with checking goodness-of-fit of binary logistic regression models. For the practitioners of data analysis, the broad classes of procedures for checking goodness-of-fit available in the literature are described. The challenges of model checking in the context of binary logistic regression are reviewed. As a viable solution, a simple graphical procedure for checking goodness-of-fit is proposed. METHODS: The graphical procedure proposed relies on pieces of information available from any logistic analysis; the focus is on combining and presenting these in an informative way. RESULTS: The information gained using this approach is presented with three examples. In the discussion, the proposed method is put into context and compared with other graphical procedures for checking goodness-of-fit of binary logistic models available in the literature. CONCLUSION: A simple graphical method can significantly improve the understanding of any logistic regression analysis and help to prevent faulty conclusions.

LOGITCPRPLOT: Stata module to graph component-plus-residual plot for logistic regression

logitcprplot can be used after logistic regression for graphing a component-plus-residual plot (a.k.a. partial residual plot) for a given predictor, including a lowess, local polynomial, restricted cubic spline, fractional polynomial, penalized spline, regression spline, running line, or adaptive variable span running line smooth

RRLOGIT: Stata module to estimate logistic regression for randomized response data

rrlogit fits a maximum-likelihood logistic regression for randomized response data.

Replicative phenotyping adds value to genotypic resistance testing in heavily pre-treated HIV-infected individuals--the Swiss HIV Cohort Study

Background Replicative phenotypic HIV resistance testing (rPRT) uses recombinant infectious virus to measure viral replication in the presence of antiretroviral drugs. Due to its high sensitivity of detection of viral minorities and its dissecting power for complex viral resistance patterns and mixed virus populations rPRT might help to improve HIV resistance diagnostics, particularly for patients with multiple drug failures. The aim was to investigate whether the addition of rPRT to genotypic resistance testing (GRT) compared to GRT alone is beneficial for obtaining a virological response in heavily pre-treated HIV-infected patients. Methods Patients with resistance tests between 2002 and 2006 were followed within the Swiss HIV Cohort Study (SHCS). We assessed patients' virological success after their antiretroviral therapy was switched following resistance testing. Multilevel logistic regression models with SHCS centre as a random effect were used to investigate the association between the type of resistance test and virological response (HIV-1 RNA <50 copies/mL or ≥1.5log reduction). Results Of 1158 individuals with resistance tests 221 with GRT+rPRT and 937 with GRT were eligible for analysis. Overall virological response rates were 85.1% for GRT+rPRT and 81.4% for GRT. In the subgroup of patients with >2 previous failures, the odds ratio (OR) for virological response of GRT+rPRT compared to GRT was 1.45 (95% CI 1.00-2.09). Multivariate analyses indicate a significant improvement with GRT+rPRT compared to GRT alone (OR 1.68, 95% CI 1.31-2.15). Conclusions In heavily pre-treated patients rPRT-based resistance information adds benefit, contributing to a higher rate of treatment success.

Influence of residual pockets on progression of periodontitis and tooth loss: results after 11 years of maintenance

BACKGROUND: Limited evidence exists on the significance of residual probing pocket depth (PPD) as a predictive parameter for periodontal disease progression and tooth loss. AIM: The aim of this study was to investigate the influence of residual PPD >or=5 mm and bleeding on probing (BOP) after active periodontal therapy (APT) on the progression of periodontitis and tooth loss. MATERIAL AND METHODS: In this retrospective cohort, 172 patients were examined after APT and supportive periodontal therapy (SPT) for 3-27 years (mean 11.3 years). Analyses were conducted using information at site, tooth and patient levels. The association of risk factors with tooth loss and progression of periodontitis was investigated using multilevel logistic regression analysis. RESULTS: The number of residual PPD increased during SPT. Compared with PPDor=7 mm 37.9 and 64.2, respectively. At patient level, heavy smoking, initial diagnosis, duration of SPT and PPD>or=6 mm were risk factors for disease progression, while PPD>or=6 mm and BOP>or=30% represented a risk for tooth loss. CONCLUSION: Residual PPD>or=6 mm represent an incomplete periodontal treatment outcome and require further therapy.

The loss of trust in the European Union during the great recession since 2007: The role of heuristics from the national political system

How can we explain the decline in support for the European Union (EU) and the idea of European integration after the onset of the great recession in the fall of 2007? Did the economic crisis and the austerity policies that the EU imposed—in tandem with the IMF—on several member countries help cause this drop? While there is some evidence for this direct effect of EU policies, we find that the most significant determinant of trust and support for the EU remains the level of trust in national governments. Based on cue theory and using concepts of diffuse and specific support, we find that support for the EU is derived from evaluations of national politics and policy, which Europeans know far better than the remote political system of the EU. This effect, however, is somewhat muted for those sophisticated Europeans that are more knowledgeable about the EU and are able to form opinions about it independently of the national contexts in which they live. We also find that the recent economic crisis has led to a discernible increase in the number of those who are disillusioned with politics both at the national and the supranational level. We analyze 133 national surveys from 27 EU countries by estimating a series of cross-classified multilevel logistic regression models.

The association between admission systolic blood pressure and mortality in significant traumatic brain injury: A multi-centre cohort study

INTRODUCTION Low systolic blood pressure (SBP) is an important secondary insult following traumatic brain injury (TBI), but its exact relationship with outcome is not well characterised. Although a SBP of <90mmHg represents the threshold for hypotension in consensus TBI treatment guidelines, recent studies suggest redefining hypotension at higher levels. This study therefore aimed to fully characterise the association between admission SBP and mortality to further inform resuscitation endpoints. METHODS We conducted a multicentre cohort study using data from the largest European trauma registry. Consecutive adult patients with AIS head scores >2 admitted directly to specialist neuroscience centres between 2005 and July 2012 were studied. Multilevel logistic regression models were developed to examine the association between admission SBP and 30 day inpatient mortality. Models were adjusted for confounders including age, severity of injury, and to account for differential quality of hospital care. RESULTS 5057 patients were included in complete case analyses. Admission SBP demonstrated a smooth u-shaped association with outcome in a bivariate analysis, with increasing mortality at both lower and higher values, and no evidence of any threshold effect. Adjusting for confounding slightly attenuated the association between mortality and SBP at levels <120mmHg, and abolished the relationship for higher SBP values. Case-mix adjusted odds of death were 1.5 times greater at <120mmHg, doubled at <100mmHg, tripled at <90mmHg, and six times greater at SBP<70mmHg, p<0.01. CONCLUSIONS These findings indicate that TBI studies should model SBP as a continuous variable and may suggest that current TBI treatment guidelines, using a cut-off for hypotension at SBP<90mmHg, should be reconsidered.

Risk factors associated with the longevity of multi-rooted teeth. Long-term outcomes after active and supportive periodontal therapy

AIM To investigate risk factors for the loss of multi-rooted teeth (MRT) in subjects treated for periodontitis and enrolled in supportive periodontal therapy (SPT). MATERIAL AND METHODS A total of 172 subjects were examined before (T0) and after active periodontal therapy (APT)(T1) and following a mean of 11.5 ± 5.2 (SD) years of SPT (T2). The association of risk factors with loss of MRT was analysed with multilevel logistic regression. The tooth was the unit of analysis. RESULTS Furcation involvement (FI) = 1 before APT was not a risk factor for tooth loss compared with FI = 0 (p = 0.37). Between T0 and T2, MRT with FI = 2 (OR: 2.92, 95% CI: 1.68, 5.06, p = 0.0001) and FI = 3 (OR: 6.85, 95% CI: 3.40, 13.83, p < 0.0001) were at a significantly higher risk to be lost compared with those with FI = 0. During SPT, smokers lost significantly more MRT compared with non-smokers (OR: 2.37, 95% CI: 1.05, 5.35, p = 0.04). Non-smoking and compliant subjects with FI = 0/1 at T1 lost significantly less MRT during SPT compared with non-compliant smokers with FI = 2 (OR: 10.11, 95% CI: 2.91, 35.11, p < 0.0001) and FI = 3 (OR: 17.18, 95% CI: 4.98, 59.28, p < 0.0001) respectively. CONCLUSIONS FI = 1 was not a risk factor for tooth loss compared with FI = 0. FI = 2/3, smoking and lack of compliance with regular SPT represented risk factors for the loss of MRT in subjects treated for periodontitis.

Alignment between learning needs and learning goals of Mini-CEX in clerkships

Background: Defining learning goals (LG) in alignment with learning needs (LN) is one of the key purposes of formative workplace-based assessment, but studies about this topic are scarce. Summary of Work: We analysed quantitatively and qualitatively how often trainer-student pairs identified the same LN during Mini Clinical Evaluation Exercises (Mini-CEX) in clerkships and to what degree those LNs were in line with the recorded LGs. Multilevel logistic regression models were used to predict LGs by identified LNs, controlling for context variables. Summary of Results: 512 trainers and 165 students conducted 1783 Mini-CEX (98% completion rate). Concordantly, trainer-student pairs most often identified LNs in the domains ‘clinical reasoning’ (23% of 1167 complete forms), ‘organisation / efficiency’ (20%) and ‘physical examination’ (20%). At least one ‘defined’ LG was noted on 313 student forms (18% of 1710), with a total of 446 LGs. Of these, the most frequent LGs were ‘physical examination’ (49% of 446 LGs) and ‘history taking’ (21%); corresponding LNs as well as context variables (e.g. clinic size) were found to be predictors of these LGs. Discussion and Conclusions: Although trainer-student pairs often agreed in their identified LNs, many assessments did not result in an aligned LG or a LG at all. Interventions are needed to enhance the proportion of (aligned) LGs in Mini-CEX in order to tap into its full potential not only as a ‘diagnostic’ but also as an ‘educational tool’. Take-home messages: The sparseness of LGs, their dependency on context variables and their partial non-alignment with students’ LNs raise the question of how the effectiveness of Mini-CEX can be further enhanced.

Rhinovirus Infections and Associated Respiratory Morbidity in Infants: A Prospective Cohort Study.

BACKGROUND Risk factors promoting rhinovirus (RV) infections are inadequately described in healthy populations, especially infants. OBJECTIVES To determine the frequency of symptomatic and asymptomatic RV infections and identify possible risk factors from host and environment among otherwise healthy infants. METHODS In a prospective birth cohort, respiratory health was assessed in 41 term-born infants by weekly telephone interviews during the first year of life, and weekly nasal swabs were collected to determine RV prevalence. In a multilevel logistic regression model, associations between prevalence and respiratory symptoms during RV infections and host/environmental factors were determined. RESULTS 27% of nasal swabs in 41 infants tested positive for RVs. Risk factors for RV prevalence were autumn months (OR=1.71, p=0.01, 95% CI 1.13-2.61), outdoor temperatures between 5-10 °C (OR=2.33, p=0.001, 95% CI 1.41-3.86), older siblings (OR=2.60, p=0.001, 95% CI 1.50-4.51) and childcare attendance (OR=1.53, p=0.07, 95% CI 0.96-2.44). 51% of RV-positive samples were asymptomatic. Respiratory symptoms during RV infections were less likely during the first three months of life (OR=0.34, p=0.003, 95% CI 0.17-0.69) and in infants with atopic mothers (OR=0.44, p=0.008, 95% CI 0.24-0.80). Increased tidal volume (OR=1.67, p=0.03, 95% CI 1.04-2.68) and outdoor temperatures between 2-5 °C (OR=2.79, p=0.02, 95% CI 1.17-6.61) were associated with more symptoms. CONCLUSIONS RVs are highly prevalent during the first year of life, and most infections are asymptomatic. Frequency of RV infections is associated with environmental factors, while respiratory symptoms during RV infections are linked to host determinants like infant age, maternal atopy, or premorbid lung function.

The influence of the rearing period on intramammary infections in Swiss dairy heifers: A cross-sectional study

Healthy replacement heifers are one of the foundations of a healthy dairy herd. Farm management andrearing systems in Switzerland provide a wide variety of factors that could potentially be associated withintramammary infections (IMI) in early lactating dairy heifers. In this study, IMI with minor mastitispathogens such as coagulase-negative staphylococci (CNS), contagious pathogens, and environmentalmajor pathogens were identified. Fifty-four dairy farms were enrolled in the study. A questionnaire wasused to collect herd level data on housing, management and welfare of young stock during farm isitsand interviews with the farmers. Cow-level data such as breed, age at first calving, udder condition andswelling, and calving ease were also recorded. Data was also collected about young stock that spent aperiod of at least 3 months on an external rearing farm or on a seasonal alpine farm. At the quarterlevel, teat conditions such as teat lesions, teat dysfunction, presence of a papilloma and teat lengthwere recorded. Within 24 h after parturition, samples of colostral milk from 1564 quarters (391 heifers)were collected aseptically for bacterial culture. Positive bacteriological culture results were found in 49%of quarter samples. Potential risk factors for IMI were identified at the quarter, animal and herd levelusing multivariable and multilevel logistic regression analysis. At the herd level tie-stalls, and at cow-level the breed category “Brown cattle” were risk factors for IMI caused by contagious major pathogenssuch as Staphylococcus aureus (S. aureus). At the quarter-level, teat swelling and teat lesions were highlyassociated with IMI caused by environmental major pathogens. At the herd level heifer rearing at externalfarms was associated with less IMI caused by major environmental pathogens. Keeping pregnant heifersin a separate group was negatively associated with IMI caused by CNS. The odds of IMI with coagulase-negative staphylococci increased if weaning age was less than 4 months and if concentrates were fed tocalves younger than 2 weeks. This study identified herd, cow- and quarter-level risk factors that may beimportant for IMI prevention in the future.

Variation in partner notification outcomes for chlamydia in UK genitourinary medicine clinics: multilevel study

Objectives To compare different ways of measuring partner notification (PN) outcomes with published audit standards, examine variability between clinics and examine factors contributing to variation in PN outcomes in genitourinary medicine (GUM) clinics in the UK. Methods Reanalysis of the 2007 BASHH national chlamydia audit. The primary outcome was the number of partners per index case tested for chlamydia, as verified by a healthcare worker or, if missing, reported by the patient. Control charts were used to examine variation between clinics considering missing values as zero or excluding missing values. Hierarchical logistic regression was used to investigate factors contributing to variation in outcomes. Results Data from 4616 individuals in 169 genitourinary medicine clinics were analysed. There was no information about the primary outcome in 41% of records. The mean number of partners tested for chlamydia ranged from 0 to 1.5 per index case per clinic. The median across all clinics was 0.47 when missing values were assumed to be zero and 0.92 per index case when missing values were excluded. Men who have sex with men were less likely than heterosexual men and patients with symptoms (4-week look-back period) were less likely than asymptomatic patients (6-month look-back) to report having one or more partners tested for chlamydia. There was no association between the primary outcome and the type of the health professional giving the PN advice. Conclusions The completeness of PN outcomes recorded in clinical notes needs to improve. Further research is needed to identify auditable measures that are associated with successful PN that prevents repeated chlamydia in index cases.

Tissue coverage of a hydrophilic polymer-coated zotarolimus-eluting stent vs. a fluoropolymer-coated everolimus-eluting stent at 13-month follow-up: an optical coherence tomography substudy from the RESOLUTE All Comers trial

Aims To compare the tissue coverage of a hydrophilic polymer-coated zotarolimus-eluting stent (ZES) vs. a fluoropolymer-coated everolimus-eluting stent (EES) at 13 months, using optical coherence tomography (OCT) in an ‘all-comers' population of patients, in order to clarify the mechanism of eventual differences in the biocompatibility and thrombogenicity of the devices. Methods and results Patients randomized to angiographic follow-up in the RESOLUTE All Comers trial (NCT00617084) at pre-specified OCT sites underwent OCT follow-up at 13 months. Tissue coverage and apposition were assessed strut by strut, and the results in both treatment groups were compared using multilevel logistic or linear regression, as appropriate, with clustering at three different levels: patient, lesion, and stent. Fifty-eight patients (30 ZES and 28 EES), 72 lesions, 107 stents, and 23 197 struts were analysed. Eight hundred and eighty-seven and 654 uncovered struts (7.4 and 5.8%, P= 0.378), and 216 and 161 malapposed struts (1.8 and 1.4%, P= 0.569) were found in the ZES and EES groups, respectively. The mean thickness of coverage was 116 ± 99 µm in ZES and 142 ± 113 µm in EES (P= 0.466). No differences in per cent neointimal volume obstruction (12.5 ± 7.9 vs. 15.0 ± 10.7%) or other areas–volumetric parameters were found between ZES and EES, respectively. Conclusion No significant differences in tissue coverage, malapposition, or lumen/stent areas and volumes were detected by OCT between the hydrophilic polymer-coated ZES and the fluoropolymer-coated EES at 13-month follow-up.

Epidemiology and predictors of spinal injury in adult major trauma patients: European cohort study

This is a European cohort study on predictors of spinal injury in adult (≥16 years) major trauma patients, using prospectively collected data of the Trauma Audit and Research Network from 1988 to 2009. Predictors for spinal fractures/dislocations or spinal cord injury were determined using univariate and multivariate logistic regression analysis. 250,584 patients were analysed. 24,000 patients (9.6%) sustained spinal fractures/dislocations alone and 4,489 (1.8%) sustained spinal cord injury with or without fractures/dislocations. Spinal injury patients had a median age of 44.5 years (IQR = 28.8-64.0) and Injury Severity Score of 9 (IQR = 4-17). 64.9% were male. 45% of patients suffered associated injuries to other body regions. Age <45 years (≥45 years OR 0.83-0.94), Glasgow Coma Score (GCS) 3-8 (OR 1.10, 95% CI 1.02-1.19), falls >2 m (OR 4.17, 95% CI 3.98-4.37), sports injuries (OR 2.79, 95% CI 2.41-3.23) and road traffic collisions (RTCs) (OR 1.91, 95% CI 1.83-2.00) were predictors for spinal fractures/dislocations. Age <45 years (≥45 years OR 0.78-0.90), male gender (female OR 0.78, 95% CI 0.72-0.85), GCS <15 (OR 1.36-1.93), associated chest injury (OR 1.10, 95% CI 1.01-1.20), sports injuries (OR 3.98, 95% CI 3.04-5.21), falls >2 m (OR 3.60, 95% CI 3.21-4.04), RTCs (OR 2.20, 95% CI 1.96-2.46) and shooting (OR 1.91, 95% CI 1.21-3.00) were predictors for spinal cord injury. Multilevel injury was found in 10.4% of fractures/dislocations and in 1.3% of cord injury patients. As spinal trauma occurred in >10% of major trauma patients, aggressive evaluation of the spine is warranted, especially, in males, patients <45 years, with a GCS <15, concomitant chest injury and/or dangerous injury mechanisms (falls >2 m, sports injuries, RTCs and shooting). Diagnostic imaging of the whole spine and a diligent search for associated injuries are substantial.