Random forests for the analysis of matched case-control studies.

BACKGROUND: Conditional logistic regression trees have been proposed as a flexible alternative to the standard method of conditional logistic regression for the analysis of matched case-control studies. While they allow to avoid the strict assumption of linearity and automatically incorporate interactions, conditional logistic regression trees may suffer from a relatively high variability. Further machine learning methods for the analysis of matched case-control studies are missing because conventional machine learning methods cannot handle the matched structure of the data. RESULTS: A random forest method for the analysis of matched case-control studies based on conditional logistic regression trees is proposed, which overcomes the issue of high variability. It provides an accurate estimation of exposure effects while being more flexible in the functional form of covariate effects. The efficacy of the method is illustrated in a simulation study and within an application to real-world data from a matched case-control study on the effect of regular participation in cervical cancer screening on the development of cervical cancer. CONCLUSIONS: The proposed random forest method is a promising add-on to the toolbox for the analysis of matched case-control studies and addresses the need for machine-learning methods in this field. It provides a more flexible approach compared to the standard method of conditional logistic regression, but also compared to conditional logistic regression trees. It allows for non-linearity and the automatic inclusion of interaction effects and is suitable both for exploratory and explanatory analyses.

Lessons for Future Vaccination Policies: COVID-19 Vaccination Intention in People With and Without Chronic Diseases.

PURPOSE: This study compared COVID-19 vaccination intentions in those with and without chronic diseases (CDs and non-CDs) in South Korea. We hypothesized that the factors associated with COVID-19 vaccination intentions would differ between CDs and non-CDs in South Korea. METHODS: Using survey data collected through a Korean online panel in June 2021, we conducted a cross-sectional secondary data analysis. Of the 2292 participants, 411 had at least one chronic disease. To construct a comparable dataset, we selected non-CDs via a 1:1 case-control matching for age and gender. We then utilized a multivariable binary logistic regression model to explore the factors contributing to COVID-19 vaccination intentions in CDs and non-CDs. RESULTS: All told, over 75% of participants in both groups indicated that they intended to vaccinate against COVID-19. In both groups, those who mistrusted general vaccine benefits reported significantly lower COVID-19 vaccination intentions. Regarding factors associated with vaccination intentions, CDs identified anxiety regarding coronavirus and exposure to COVID-19 vaccination promotions at the community level, while non-CDs highlighted hesitancy regarding vaccines and confidence in government/health services. CONCLUSION: Improving vaccination acceptance will require the development and implementation of tailored approaches for CDs and non-CDs and efforts to minimize general vaccine mistrust.

Nosocomial outbreak caused by disinfectant-resistant Serratia marcescens in an adult intensive care unit, Hungary, February to March 2022.

In 2022, an outbreak with severe bloodstream infections caused by Serratia marcescens occurred in an adult intensive care unit (ICU) in Hungary. Eight cases, five of whom died, were detected. Initial control measures could not stop the outbreak. We conducted a matched case-control study. In univariable analysis, the cases were more likely to be located around one sink in the ICU and had more medical procedures and medications than the controls, however, the multivariable analysis was not conclusive. Isolates from blood cultures of the cases and the ICU environment were closely related by whole genome sequencing and resistant or tolerant against the quaternary ammonium compound surface disinfectant used in the ICU. Thus, S. marcescens was able to survive in the environment despite regular cleaning and disinfection. The hospital replaced the disinfectant with another one, tightened the cleaning protocol and strengthened hand hygiene compliance among the healthcare workers. Together, these control measures have proved effective to prevent new cases. Our results highlight the importance of multidisciplinary outbreak investigations, including environmental sampling, molecular typing and testing for disinfectant resistance.

Conditional concordance-assisted learning under matched case-control design for combining biomarkers for population screening.

Incorporating promising biomarkers into cancer screening practices for early-detection is increasingly appealing because of the unsatisfactory performance of current cancer screening strategies. The matched case-control design is commonly adopted in biomarker development studies to evaluate the discriminative power of biomarker candidates, with an intention to eliminate confounding effects. Data from matched case-control studies have been routinely analyzed by the conditional logistic regression, although the assumed logit link between biomarker combinations and disease risk may not always hold. We propose a conditional concordance-assisted learning method, which is distribution-free, for identifying an optimal combination of biomarkers to discriminate cases and controls. We are particularly interested in combinations with a clinically and practically meaningful specificity to prevent disease-free subjects from unnecessary and possibly intrusive diagnostic procedures, which is a top priority for cancer population screening. We establish asymptotic properties for the derived combination and confirm its favorable finite sample performance in simulations. We apply the proposed method to the prostate cancer data from the carotene and retinol efficacy trial (CARET).

A tree-based modeling approach for matched case-control studies.

Conditional logistic regression (CLR) is the indisputable standard method for the analysis of matched case-control studies. However, CLR is strongly restricted with respect to the inclusion of non-linear effects and interactions of confounding variables. A novel tree-based modeling method is proposed which accounts for this issue and provides a flexible framework allowing for a more complex confounding structure. The proposed machine learning model is fitted within the framework of CLR and, therefore, allows to account for the matched strata in the data. A simulation study demonstrates the efficacy of the method. Furthermore, for illustration the method is applied to a matched case-control study on cervical cancer.

Cost-Efficient Multiply Matched Case-Control Study Designs.

In multiply matched case-control studies, a number of cases and controls may be included in each matched set. However, when per-participant costs between cases and controls differ, investigators should be aware of how the numbers of cases and controls per matched set affect the overall total study cost. Traditional statistical approaches to designing case-control studies do not account for study costs. Given an effect size, the power to detect differences is typically a function of the numbers of cases and controls within each matched set. Therefore, the same level of statistical power will be achieved based on various combinations of the numbers of cases and controls. Typical matched case-control studies match a case to a number of controls by levels of 1 or more known factors. Several authors have shown that for study designs with 1 case per matched set, the optimal number of controls within each matched set that minimizes the total study cost is the square root of the ratio of the cost of a case to the cost of a control. Herein, we extend this result to the setting of a multiply matched case-control study design, when 1 or more cases are matched to controls within each matched set. A Shiny web application implementation of the proposed methods is presented.

Relation of severe COVID-19 in Scotland to transmission-related factors and risk conditions eligible for shielding support: REACT-SCOT case-control study.

BACKGROUND: Clinically vulnerable individuals have been advised to shield themselves during the COVID-19 epidemic. The objectives of this study were to investigate (1) the rate ratio of severe COVID-19 associated with eligibility for the shielding programme in Scotland across the first and second waves of the epidemic and (2) the relation of severe COVID-19 to transmission-related factors in those in shielding and the general population. METHODS: In a matched case-control design, all 178,578 diagnosed cases of COVID-19 in Scotland from 1 March 2020 to 18 February 2021 were matched for age, sex and primary care practice to 1,744,283 controls from the general population. This dataset (REACT-SCOT) was linked to the list of 212,702 individuals identified as eligible for shielding. Severe COVID-19 was defined as cases that entered critical care or were fatal. Rate ratios were estimated by conditional logistic regression. RESULTS: With those without risk conditions as reference category, the univariate rate ratio for severe COVID-19 was 3.21 (95% CI 3.01 to 3.41) in those with moderate risk conditions and 6.3 (95% CI 5.8 to 6.8) in those eligible for shielding. The highest rate was in solid organ transplant recipients: rate ratio 13.4 (95% CI 9.6 to 18.8). Risk of severe COVID-19 increased with the number of adults but decreased with the number of school-age children in the household. Severe COVID-19 was strongly associated with recent exposure to hospital (defined as 5 to 14 days before presentation date): rate ratio 12.3 (95% CI 11.5 to 13.2) overall. The population attributable risk fraction for recent exposure to hospital peaked at 50% in May 2020 and again at 65% in December 2020. CONCLUSIONS: The effectiveness of shielding vulnerable individuals was limited by the inability to control transmission in hospital and from other adults in the household. Mitigating the impact of the epidemic requires control of nosocomial transmission.

Periodontitis as a Risk Factor for Rheumatoid Arthritis: a Matched-Cohort Study.

OBJECTIVE: The intent was to analyse the association of periodontitis with the development of rheumatoid arthritis (RA) using a representative population-based cohort and longitudinal matched-cohort design. METHODS: Participants were 40 years of age or older and had not been diagnosed with RA between 2002 and 2006. Among the participants, those who were newly diagnosed with periodontitis between 2004 and 2006 (excluding cases that had already been diagnosed with periodontitis between 2002 and 2003) were allotted to the periodontitis group. Among the participants, those who had never been diagnosed with periodontitis between 2002 and 2006 formed the control group, matched by sex, age, and household income at a 1:1 ratio. From 2007 to 2018, the 2 groups (n = 691,506) were followed to monitor the development of RA. The t-test and χ2 test compared the general characteristics and health-related variables of both groups. The Kaplan-Meier method with a log-rank test was conducted to compare the incidence of RA in both groups. The hazard ratio (HR) and adjusted hazard ratio (aHR) were calculated using a Cox proportional hazard regression analysis to evaluate the risk of subsequent RA. RESULTS: Univariate analysis revealed that the periodontitis group was more likely to develop RA than the control group (hazard ratio 1.10), and multivariate analysis also revealed a higher incidence risk of RA (adjusted hazard ratio 1.09) in the periodontitis group. CONCLUSIONS: Our findings demonstrate that periodontitis is associated with an increased risk of developing RA.

Evaluating the Impact of Intensive Case Management for Severe Vocational Injuries on Work Incapacity and Costs.

Purpose This study investigates the impact of an intensive case management program on sick leave days, permanent work incapacity levels and treatment costs for severe vocational injuries set up by the French National Insurance Fund in five health insurance districts. Methods The method employed relies on a four-step matching procedure combining Coarsened Exact Matching and Propensity Score Matching, based on an original administrative dataset. Average Treatment effects on the Treated were estimated using a parametric model with a large set of covariates. Results After one-year follow-up, workers in the treatment group had higher sickness absence rates, with 22 extra days, and the program led to 2.7 (95% CI 2.3-3.1) times more diagnoses of permanent work incapacity in the treatment group. With an estimated yearly operational cost of 2,722 € per treated worker, the average total extra treatment cost was 4,569 € for treated workers, which corresponds to a cost increase of 29.2% for the insurance fund. Conclusions The higher costs found for the treatment group are mainly due to longer sick leave duration for the moderate severity group, implying higher cash transfers in the form of one-off indemnities. Even though workers in the treated group have more diagnoses of permanent work incapacity, the difference of severity between groups is small. Our results on longer sick leave duration are partly to be explained by interactions between the case managers and the occupational physicians that encouraged patients to stay longer off-work for better recovery, despite the higher costs that this represented for the insurance fund and the well-documented adverse side effects of longer periods off-work.

Estimating the receiver operating characteristic curve in matched case control studies.

The matched case-control design is frequently used in the study of complex disorders and can result in significant gains in efficiency, especially in the context of measuring biomarkers; however, risk prediction in this setting is not straightforward. We propose an inverse-probability weighting approach to estimate the predictive ability associated with a set of covariates. In particular, we propose an algorithm for estimating the summary index, area under the curve corresponding to the Receiver Operating Characteristic curve associated with a set of pre-defined covariates for predicting a binary outcome. By combining data from the parent cohort with that generated in a matched case control study, we describe methods for estimation of the population parameters of interest and the corresponding area under the curve. We evaluate the bias associated with the proposed methods in simulations by considering a range of parameter settings. We illustrate the methods in two data applications: (1) a prospective cohort study of cardiovascular disease in women, the Women's Health Study, and (2) a matched case-control study nested within the Nurses' Health Study aimed at risk prediction of invasive breast cancer.

Oral hygiene and dentition status in children and adults with hemophilia: A case-control study.

BACKGROUND & OBJECTIVES: People with hemophilia constitute a significant proportion of the population and an oral health care professional faces a considerable challenge while treating them. This study aimed to assess the oral health and dentition status as well as fear of dental treatment in patients with hemophilia and compare it with age-matched healthy subjects. PATIENTS & METHODS: This single-center, case-control cross-sectional study was performed on 100 subjects with hemophilia and 100 age-matched healthy controls. Oral health and dentition status was recorded for all the subjects and scored using the simplified oral hygiene index (OHI-S), plaque index, and the dmft/DMFT index. RESULTS: There were a total of 41 children and 59 adults in the hemophilia group and 36 children and 64 adults in the healthy group. When compared to healthy subjects it was observed that individuals with hemophilia had higher debris and calculus scores which was indicative of poor orodental status. There was no significant difference observed in the DMFT scores among the study groups. INTERPRETATION & CONCLUSIONS: The oral hygiene of the hemophilics was poorer when compared to the healthy controls. The findings highlight the need for establishing interdisciplinary care for such individuals.

[Epigenetic research of cognitive deficit in schizophrenia: some methodological considerations]. / Épigeneticheskie issledovaniia kognitivnogo defitsita pri shizofrenii: nekotorye metodologicheskie problemy.

AIM: To highlights the problems of assessing cognitive deficits in schizophrenia, relevant to the epigenetic, as well as a wide range of other approaches to the search for biological bases of cognition. MATERIAL AND METHODS: The literature on the weaknesses in the evaluation of cognitive functions in patients with schizophrenia are summarized and discussed. The analysis is illustrated by our experience in developing a cognitive battery and a sample to examine relationships between DNA methylation in blood cells and cognitive deficits in schizophrenia. RESULTS AND CONCLUSION: It has been shown that to assess cognitive deficits in patients and to reduce the influence of confounders in epigenetic analysis it is necessary (1) to use a battery with the existing co-normative data in the target population, which allows to evaluate representativeness of control and patients included in the study sample, (2) to verify the theoretically driven battery structure using normative population and a cohort of patients, (3) to balance groups of cases and controls on the number, age and sex, for which an individual matching of cases and controls is best suited, (4) to conduct an additional statistical analysis controlling for education and smoking.

A case-control study of the association between psychosocial factors and the occurrence of laryngeal cancer.

The aim of the present study was to investigate the association between social psychological factors and the occurrence of laryngeal cancer. A 1:1 matched case-control study was conducted. The participants completed questionnaires that included general information, such as the Life Event Scale, the Social Support Rating Scale, and the Eysenck Personality Questionnaire (EPQ). Scores were compared between the groups using paired t-tests and Wilcoxon's signed-rank tests. No significant difference in the psychoticism scale of the EPQ was observed between the two groups (P>0.05). However, significant differences were observed in scores on the life events and social support scales and in the remaining dimensions of the EPQ (all P-values <0.05). Positive life events may be protective factors for laryngeal cancer, whereas reduced utilization of social support may be a risk factor for laryngeal cancer.

Cow's Milk Allergy in Infancy and Later Development of Juvenile Idiopathic Arthritis: A Register-Based Case-Control Study.

We examined the association between cow's milk allergy (CMA) and juvenile idiopathic arthritis (JIA). The material for this case-control study was collected from national registers of all children born in Finland between 2000 and 2010 and diagnosed with JIA (n = 1,298) and age-, sex-, and place-matched controls (n = 5,179). We identified 235 children with CMA; 66 of these children also had JIA. A conditional logistic regression analysis was performed to evaluate the association between CMA and JIA and to test whether exposure to antibiotics would be a covariate for this association. In boys (but not in girls), a diagnosis of CMA and the use of hypoallergenic formula in infancy were associated with the later development of JIA (odds ratio = 2.4, 95% confidence interval: 1.6, 3.6). The association was most evident in boys who were diagnosed with JIA before age 3 years or diagnosed with CMA with predominantly gastrointestinal symptoms. There was no statistically significant additive interaction between CMA and antibiotic exposure in the later development of JIA. These associations may reflect impaired maturation of intestinal immunity and integrity in boys with a risk of JIA. Predisposing factors related to JIA pathogenesis seem to display a sex-linked disparity.

Bayesian Variable Selection Methods for Matched Case-Control Studies.

Matched case-control designs are currently used in many biomedical applications. To ensure high efficiency and statistical power in identifying features that best discriminate cases from controls, it is important to account for the use of matched designs. However, in the setting of high dimensional data, few variable selection methods account for matching. Bayesian approaches to variable selection have several advantages, including the fact that such approaches visit a wider range of model subsets. In this paper, we propose a variable selection method to account for case-control matching in a Bayesian context and apply it using simulation studies, a matched brain imaging study conducted at Massachusetts General Hospital, and a matched cardiovascular biomarker study conducted by the High Risk Plaque Initiative.

Self-reported oral behaviour habits in hyperdivergent and normodivergent facial types.

The relationship between facial morphology and jaw function remains controversial. The purpose of this study was to investigate differences in self-reported oral behaviour habits between individuals with normodivergent and hyperdivergent facial types. Some 80 cases and controls were individually matched on age, sex ethnicity and treatment stage. The participants were recruited from an orthodontic clinic, and included both adolescents and adults. Habitual oral activity was assessed using the Oral Behaviour Checklist (OBC) based on their experiences in the past 4 weeks. Univariate and bivariate analyses were performed. The sample had a mean age of 17·2 years (SD = 4·6; range = 12-49 years), and was predominantly female (65·0%) and of New Zealand European origin (91·3%). The prevalence of reporting one or more frequently performed habitual muscular behaviour in either study group was over 85% (P > 0·05). There was no difference in total OBC score between the hyperdivergent (25·6; SD: 9·0) and normodivergent group (25·3; SD: 9·9). Moreover, there was no difference in the prevalence of either nocturnal or daytime oral behaviours between the two groups. While this study did not include any objective measures of functional or habitual activity, we found no differences in self-reported oral behaviour habits between normodivergent and hyperdivergent individuals. The findings do not support an association between vertical facial form and habitual muscular activity.

Risk factors associated with acute/subacute cerebral infarction in HIV-negative patients with cryptococcal meningitis.

OBJECTIVE: To explore the risk factors associated with acute/subacute cerebral infarction (ASCI) in HIV-negative patients with cryptococcal meningitis (CM). METHODS: This case-control study included 10 HIV-negative CM patients with ASCI and 30 age- and sex-matched HIV-negative control (1:3) CM patients without ASCI. The clinical manifestations and neuroimaging findings were collected. Risk factors for ASCI in the HIV-negative CM patients were confirmed by conditional logistic regression analysis. RESULTS: Among the 10 HIV-negative CM patients with ASCI, all cases had lacunar infarctions. Single infarctions were found in 6 patients, and multiple infarctions in 4. Hydrocephalus (p=0.020, OR=23.77, 95% CI, 1.67-339.33) and smoking (p=0.039, OR=11.63, 95% CI, 1.14-118.96) were found to be independently associated with the occurrence of ASCI. CONCLUSIONS: Hydrocephalus and smoking may increase the risk of ASCI in HIV-negative CM patients. In the clinical course, cerebral infarction should be strongly suspected in CM patients with hydrocephalus or smoking histories.

Repeated exposure to antibiotics in infancy: a predisposing factor for juvenile idiopathic arthritis or a sign of this group's greater susceptibility to infections?

OBJECTIVE: Previous exposure to antibiotics has been associated with the pathogenesis of several autoimmune diseases. Our objective was to explore whether childhood exposure to antibiotics would be associated with the risk of developing juvenile idiopathic arthritis (JIA). METHODS: The material was collected from national registers containing all children born in 2000-2010 in Finland and diagnosed with JIA by the end of December 2012 (n = 1298) and appropriate controls (n = 5179) matched for age, sex, and place of birth. All purchases of antibiotics were collected from birth until the index date (i.e., the date of special reimbursement for JIA medications). A conditional logistic regression was performed to evaluate the association between the exposure to antibiotics and the risk of JIA. RESULTS: The risk of JIA increased with the number of antibiotic purchases from birth to the index date: for ≥ 1 purchases versus none, OR 1.6, 95% CI 1.3-1.9 with an upward trend in OR (p < 0.001). Antibiotic groups lincosamides and cephalosporins showed the strongest association with JIA (OR 6.6, 95% CI 3.7-11.7, and OR 1.6, 95% CI 1.4-1.8, respectively). Overall exposure to antibiotics before 2 years of age was associated with an increased risk of JIA (OR 1.4, 95% CI 1.2-1.6), with the trend test of OR (p < 0.001). CONCLUSION: Previous early and repeated exposure to antibiotics may predispose individuals to develop JIA. Alternatively, the apparent association may reflect shared susceptibility to infections and JIA.

Effectiveness of one dose of mumps vaccine against clinically diagnosed mumps in Guangzhou, China, 2006-2012.

Although mumps-containing vaccines were introduced in China in 1990s, mumps continues to be a public health concern due to the lack of decline in reported mumps cases. To assess the mumps vaccine effectiveness (VE) in Guangzhou, China, we performed a 1:1 matched case-control study. Among children in Guangzhou aged 8 mo to 12 y during 2006 to 2012, we matched one healthy child to each child with clinically diagnosed mumps. Cases with clinically diagnosed mumps were identified from surveillance sites system and healthy controls were randomly sampled from the Children's Expanded Programmed Immunization Administrative Computerized System in Guangzhou. Conditional logistic regression was used to calculate VE. We analyzed the vaccination information for 1983 mumps case subjects and 1983 matched controls and found that the overall VE for 1 dose of mumps vaccine, irrespective of the manufacture, was 53.6% (95% confidence interval [CI], 41.0-63.5%) to children aged 8 mo to 12 y. This post-marketing mumps VE study found that immunization with one dose of the mumps vaccine confers partial protection against mumps disease. Evaluation of the VE for the current mumps vaccines, introduction of a second dose of mumps vaccine, and assessment of modifications to childhood immunization schedules is essential.