Random Effects Logistic Regression Models Research Articles

The Association between Ambient Fine Particulate Matter and Oral Neoplasm among Smokers and Betel Quids Chewers

The association between fine particulate matter < 2.5 µm (PM2.5) and oral neoplasm has barely been addressed. The purpose of this study is to elucidate the association between PM2.5 and oral neoplasm, including oral potentially malignant disorder (OPMD) and oral cancer (OC), taking into account the geographical heterogeneity. Data for analysis were derived from nationwide OC screening program, targeting Taiwanese cigarette smokers and/or betel quid chewers, and the Taiwan Air Quality Monitoring Network between 2006 and 2016. Totally 3,864,045 smokers and/or betel quids chewers were enrolled in this study. Among them, 154,030 OPMD cases and 23,286 oral cancers were found during the study period. Information on age, gender, living area, personal oral habits, and monthly PM2.5 concentration in average were collected. We used the Bayesian random-effect logistic regression model to assess the association between PM2.5 and OPMD/OC. After adjusting for sex, age, and behavior of betel quid chewing and cigarette smoking, we found that subjects from areas of higher levels of PM2.5 (≥ 35 µg m–3) had an increased risk of OPMD/OC and OC by 11% (aRR = 1.11; 95% CI: 1.09–1.13) and 55% (aRR = 1.55; 95% CI: 1.49–1.60) respectively, compared to those from areas of lower PM2.5 (< 35 µg m–3). Such effect was further demonstrated in a concentration-dependent manner. Subjects from areas of higher PM2.5 levels were found to have greater risk of OPMD/OC in Taiwan. Future studies are warranted to investigate the effect of personal PM2.5 exposure on OPMD/OC risk.

Examining Hospital Variation on Multiple Indicators of Stroke Quality of Care.

Background:Provider profiling involves comparing the performance of hospitals on indicators of quality of care. Typically, provider profiling examines the performance of hospitals on each quality indicator in isolation. Consequently, one cannot formally examine whether hospitals that have poor performance on one indicator also have poor performance on a second indicator.Methods:We used Bayesian multivariate response random effects logistic regression model to simultaneously examine variation and covariation in multiple binary indicators across hospitals. We considered 7 binary patient-level indicators of quality of care for patients presenting to hospital with a diagnosis of acute stroke. We examined between-hospital variation in these 7 indicators across 86 hospitals in Ontario, Canada.Results:The number of patients eligible for each indicator ranged from 1321 to 14 079. There were 7 pairs of indicators for which there was a strong correlation between a hospital’s performance on each of the 2 indicators. Twenty-nine of the 86 hospitals had a probability higher than 0.90 of having worse performance than average on at least 4 of the 7 indicators. Seven of the 86 of hospitals had a probability higher than 0.90 of having worse performance than average on at least 5 indicators. Fourteen of the 86 of hospitals had a probability higher than 0.50 of having worse performance than average on at least 6 indicators. No hospitals had a probability higher than 0.50 of having worse performance than average on all 7 indicators.Conclusions:These findings suggest that there are a small number of hospitals that perform poorly on at least half of the quality indicators, and that certain indicators tend to cluster together. The described methods allow for targeting quality improvement initiatives at these hospitals.

Measuring urbanicity as a risk factor for childhood wheeze in a transitional area of coastal ecuador: a cross-sectional analysis

BackgroundThe urbanisation process has been associated with increases in asthma prevalence, an observation supported largely by studies comparing urban with rural populations. The nature of this association remains poorly understood,...

Elevated Fat Intake Increases Body Weight and the Risk of Overweight and Obesity among Chinese Adults: 1991-2015 Trends.

Studies on fat intake and obesity have been inconclusive. This study examined the associations between dietary fat intake and body weight and the risk of overweight and obesity in China. We used data from 23,859 adults aged 20–60 years who participated in the China Health and Nutrition Survey, an ongoing open-cohort study, from 1991 to 2015. We collected detailed dietary data by conducting three 24-h dietary recalls and weighing foods and condiments in household inventories. We examined the associations between fat intake and body weight, body mass index (BMI), and the risk of overweight and obesity with random-effects linear or logistic regression models for panel data. The Chinese population’s fat intake, percentage of energy intake from fat, and prevalence of high-fat diets (energy intake from fat > 30%) increased from 68.5 g per day (g/d), 23.1%, and 22.4%, respectively, in 1991 to 79.3 g/d, 35.6%, and 67.2%, respectively, in 2015. The prevalence of overweight and obesity increased from 12.3% to 37.3% during the same period. Fat intake, percentage of energy intake from fat, and a high-fat diet were positively associated with body weight, BMI, and the risk of overweight and obesity in both sexes (p < 0.001) after adjustment for nonfat energy intake, physical activity, and socioeconomic status. Increased fat intake and high-fat diets were associated with increased body weight, BMI, and risk of overweight and obesity. These findings could have a significant impact on Chinese policies and interventions to control overweight and obesity.

The impact of mental and physical multimorbidity on healthcare utilization and health spending in China: A nationwide longitudinal population-based study.

In China, little evidence exists on the effect of mental and physical multimorbidity on individuals and the health system. This study aims to examine the prevalence of mental-physical multimorbidity and its impact on health service utilization and health expenditures. We conducted a panel study using two waves of data (in 2011 and 2015) from the China Health and Retirement Longitudinal Study, including 10,181 participants aged 45 years and older. Generalized linear regression models were used to assess the association of multimorbidity with total health expenditure and out-of-pocket expenditure (OOPE) on outpatient and inpatient care. Random-effects logistic regression models were used to examine the impact of multimorbidity on outpatient visits, admission to hospital and incidence of catastrophic health expenditure (CHE). Overall, 3210 participants (31.53% of 10,181) had mental-physical multimorbidity in 2015 in China. Compared to patients with a single physical disease, individuals with physical-mental multimorbidity had over 150% of the increase in the number of outpatient visits and days of hospitalization. The percentage change of OOPE for outpatient and inpatient care was 156.8% and 163.6%, respectively. Mental-physical multimorbidity was associated with an increased likelihood of experiencing CHE (OR = 2.205, 95% CI = 2.048, 2.051). Multimorbidity, particularly mental-physical multimorbidity, is associated with higher levels of health service use and a greater financial burden to individuals in China. Healthcare system needs to shift from single-disease models to new financing and service delivery models to more effectively manage mental-physical multimorbidity.

Contrasting painless and painful phenotypes of pediatric restless legs syndrome: a twin family study

ObjectiveThis study was designed to investigate painless and painful subsets of pediatric restless legs syndrome (RLS) for genetic influence and for associations with iron deficiency and common pediatric pain disorders. MethodsIn a twin family study, twins (3–18 years) and their oldest siblings, mothers and fathers completed questionnaires, assessing lifetime prevalence of RLS using current criteria, as well as history of iron deficiency and pediatric pain disorders. Subsets were categorized as RLS-Painless or RLS-Painful. Within twin pair analyses were conducted to assess familial and potential genetic effects for the defined subsets. Penalized maximum likelihood logistic regression was used to test familial associations. Random-effects logistic regression modeling was used in the total pediatric sample to investigate univariate and multivariate associations with the subsets. ResultsData were available for 2033 twin individuals (1007 monozygous (MZ), 1026 dizygous (DZ); 51.7% female), 688 siblings, 1013 mothers and 921 fathers. Odds ratios, correlations and casewise concordance were significantly higher in MZ than in DZ twins only for RLS-Painful. RLS-Painless, though familial (co-twin and mother), was not genetically influenced, but was independently associated with female sex (OR 0.52, p = 0.003), iron deficiency (OR 4.20, p < 0.001) and with persistent pain disorders (OR 2.28, p = 0.02). RLS-Painful was familial and was probably genetically influenced; was independently associated with non-migraine headaches (OR 2.70, p = 0.02) and recurrent abdominal pain (OR 2.07, p = 0.04). ConclusionsPediatric RLS was heterogeneous and was categorized into contrasting painless and painful phenotypes. RLS-Painless was associated with iron deficiency while RLS-Painful accounted for the heritability of RLS.

Effect of Case-Mix and Random Variation on Breast Cancer Care Quality Indicators and Their Rankability.

Hospital comparisons to improve quality of care require valid and reliable quality indicators. We aimed to test the validity and reliability of 6 breast cancer indicators by quantifying the influence of case-mix and random variation. The nationwide population-based database included 79 690 patients with breast cancer from 91 Dutch hospitals between 2011 and 2016. The indicator-scores calculated were: (1) irradical breast-conserving surgery (BCS) for invasive disease, (2) irradical BCS for ductal carcinoma-in-situ, (3) breast contour-preserving treatment, (4) magnetic resonance imaging (MRI) before neo-adjuvant chemotherapy, (5) radiotherapy for locally advanced disease, and (6) surgery within 5 weeks from diagnosis. Case-mix and random variation adjustments were performed by multivariable fixed and random effect logistic regression models. Rankability quantified the between-hospital variation, representing unexplained differences that might be the result of the level of quality of care, as low (<50%), moderate (50%-75%), or high (>75%). All of the indicators showed between-hospital variation with wide (interquartile) ranges. Case-mix adjustment reduced variation in indicators 1 and 3 to 5. Random variation adjustment (further) reduced the variation for all indicators. Case-mix and random variation adjustments influenced the indicator-scores of individual hospitals and their ranking. Rankability was poor for indicator 1, 2, and 5, and moderate for 3, 4, and6. The 6 indicators lacked validity and/or reliability to a certain extent. Although measuring quality indicators may stimulate quality improvement in general, comparisons and judgments of individual hospital performance should be made with caution if based on indicators that have not been tested or adjusted for validity and reliability, especially in benchmarking.

A mobile health-facilitated behavioural intervention for community health workers improves exclusive breastfeeding and early infant HIV diagnosis in India: a cluster randomized trial.

IntroductionIndia’s national AIDS Control Organization implemented World Health Organization’s option B+ HIV prevention of mother‐to‐child transmission (PMTCT) guidelines in 2013. However, scalable strategies to improve uptake of new PMTCT guidelines to reduce new infection rates are needed. This study assessed impact of Mobile Health‐Facilitated Behavioral Intervention on the uptake of PMTCT services.MethodsA cluster‐randomized trial of a mobile health (mHealth)‐supported behavioural training intervention targeting outreach workers (ORWs) was conducted in four districts of Maharashtra, India. Clusters (one Integrated Counselling and Testing Center (ICTC, n = 119), all affiliated ORWs (n = 116) and their assigned HIV‐positive pregnant/postpartum clients (n = 1191)) were randomized to standard‐of‐care (SOC) ORW training vs. the COMmunity home Based INDia (COMBIND) intervention – specialized behavioural training plus a tablet‐based mHealth application to support ORW‐patient communication and patient engagement in HIV care. Impact on uptake of maternal antiretroviral therapy at delivery, exclusive breastfeeding at six months, infant nevirapine prophylaxis, and early infant diagnosis at six months was assessed using multi‐level random‐effects logistic regression models.ResultsOf 1191 HIV‐positive pregnant/postpartum women, 884 were eligible for primary outcome assessment; 487 were randomized to COMBIND. Multivariable analyses identified no statistically significant differences in any primary outcome by study arm. COMBIND was associated with higher uptake of exclusive breastfeeding at two months (adjusted Odds Ratio (aOR), 2.10; 95% CI 1.06 to 4.15) and early infant diagnosis at six weeks (aOR, 2.19; 95% CI 1.05 to 3.98) than SOC.ConclusionsThe COMBIND intervention was easily integrated into India’s existing PMTCT programme and improved early uptake of two PMTCT components that require self‐motivated health‐seeking behaviour, thus providing preliminary evidence to support COMBIND as a potentially scalable PMTCT strategy. Further study would identify modifications needed to optimize other PMTCT outcomes.

The effect of multimorbidity on functional limitations and depression amongst middle-aged and older population in China: a nationwide longitudinal study.

multimorbidity contributes to a large portion of the disease burden in low- and middle-income countries. However, limited research has been undertaken in China. This study has investigated the prevalence of multimorbidity and the associations of multimorbidity with activities of daily living (ADL), instrumental activities of daily living (IADL) and depression in China. the study participants included 10,055 adults aged 45years and older from three rounds of the China Health and Retirement Longitudinal Study 2011-2015. Random-effects logistic regression models were used to examine the association of multimorbidity with ADL limitation, IADL limitation and mental disease. the prevalence of multimorbidity amongst adults in China aged 45years and older was 62.1% in 2015. The prevalence of multimorbidity was increased with older age, among women, in a higher socio-economic group and in the most deprived regions. Multimorbidity is associated with an increased likelihood of experiencing ADL limitation (adjusted odds ratio [AOR] = 5.738, 95% confidence intervals (CI) = 5.733, 5.744) and IADL limitation (AOR = 2.590, 95% CI = 2.588, 2.592) and depression (AOR = 3.352, 95% CI = 3.350, 3.354). Rural-urban disparities in functional difficulties and depression were also found amongst patients with multimorbidity. the burden of multimorbidity is high in China, particularly amongst the older population. Multimorbidity is associated with higher levels of functional limitations and depression. China healthcare reforms should introduce integrated care models and patient-centred healthcare delivery. The increasing need for reorientation of healthcare resources considering the distribution of multimorbidity and its adverse effect requires more attention from health policymakers in China and other developing countries.

Measuring individual benefits of psychiatric treatment using longitudinal binary outcomes: Application to antipsychotic benefits in non-cannabis and cannabis users

ABSTRACT We present and evaluate a method for predicting individual treatment benefits based on random effects logistic regression models of binary outcomes that change over time. The method uses empirical Bayes predictors based on patients’ characteristics and responses to treatment. It is applicable to both 1-dimentional and 2-dimentional personalized medicine models. Comparisons between predicted and true benefits for simulated new patients using correlations, relative biases and mean-squared errors were used to evaluate prediction performance. The predicted benefits had relatively small biases and relatively high correlations with the true benefits in the simulated new patients. The predictors also captured estimated overall population trends in the evolution of individual benefits. The proposed approach can be used to retrospectively evaluate patients’ responses in a clinical trial, or to retrospectively or prospectively predict individual benefits of different treatments for new patients using patients’ characteristics and previous responses. The method is used to examine changes in the disorganized dimension of antipsychotic-naïve patients from an antipsychotic randomized clinical trial. Retrospective prediction of individual benefits revealed that more cannabis users had slower and lower responses to antipsychotic treatment as compared to non-cannabis users, revealing cannabis use as a negative prognostic factor for psychotic disorders in the disorganized dimension.

SO087RISK OF INTRADIALYTIC HYPOTENSION BY DAY OF THE WEEK IN MAINTENANCE HEMODIALYSIS

Abstract Background and Aims Intradialytic hypotension (IDH) affects up to 50% of maintenance hemodialysis (HD) patients and is associated with a higher risk of cardiovascular (CV) events and mortality. CV events are more common on the days of dialysis and are most pronounced following the longer interdialytic interval. We wished to investigate if the rates of IDH differed according to dialysis day. Method We examined data from adult patients undergoing in-center thrice-weekly maintenance HD (N= 975 patients/n=59,004 sessions from a large dialysis organization [LDO] and N=1,838 patients/n=64,407 sessions from the Hemodialysis [HEMO] Study). We fit random effects logistic regression models to determine the association of day of HD (Mon/Tue [HD1]; Wed/Thu [HD2]; Fri/Sat [HD3]) with various definitions of IDH. Models were adjusted for age, sex, race, Kt/V, membrane flux, heart failure, peripheral vascular disease, diabetes, access type (catheter vs. fistula vs. graft), pre-dialysis blood urea nitrogen, ultrafiltration rate, pre-HD systolic blood pressure (SBP). IDH definitions included Nadir90 (nadir intra-HD SBP &amp;lt;90mmHg), Nadir90/100 (nadir intra-HD SBP &amp;lt;90mmHg if pre-HD SBP &amp;lt;160mmHg or &amp;lt;100mmHg if pre-HD SBP ≥160mmHg), Fall20 (pre-SBP minus nadir SBP ≥20mmHg) and Fall30 (pre-SBP minus nadir SBP ≥30mmHg). Results Mean age was 60 years, 44% were female and 38% were black in the LDO cohort, while mean age was 58 years, 56% were female and 63% were black in HEMO. Nadir90 occurred in 15% of LDO and 11% of HEMO sessions; Nadir90/100 occurred in 18% of LDO and 14% of HEMO sessions; Fall20 occurred in 76% of LDO and 68% of HEMO sessions; Fall30 occurred in 60% of LDO and 50% of HEMO sessions. Overall, a monotonic increase in the risk of IDH was observed for HD2 and HD3 compared with HD1 for all IDH definitions in both cohorts (see Table and Figure). Conclusion We observed a consistent monotonic increase in the risk of IDH for HD2 and HD3, compared with HD1 in two separate cohorts of maintenance HD patients. Potential explanations may relate to progressive reduction in post-HD weight during the dialytic week, or perhaps more aggressive ultrafiltration in advance of the longer interval. Further research to determine the underling mechanisms is necessary to guide practicing nephrologists when individualizing the HD prescription.

1237-P: The Type 1 Diabetes Identification Initiative: Improving the Care of Inpatients with Type 1 Diabetes

Background: Specialist care may improve outcomes for inpatients with diabetes. However, research specific to type 1 diabetes (T1D) is lacking. We aimed to investigate whether automatic alerts from the Electronic Medical Record (EMR) enable prompt review of inpatients with T1D by a specialist diabetes team. Methods: The type 1 diabetes identification (“T1DI”) initiative was established to improve identification and review of inpatients with T1D by a specialist diabetes team. A specific T1DI code was entered into the EMR of patients with T1D. On subsequent admission, the specialist diabetes team received an automatic message and generated a list of inpatients requiring review. “T1DI” admissions of more than 1-day were analysed over a 1-year period (1 January-31 December 2018). The control group was T1D admissions as coded by health information services following discharge. Analysis was performed for patients admitted to units other than endocrinology to investigate review outcomes. Data was analysed using Chi2 (categorical), Wilcoxon rank sum (continuous) and random effects logistic or negative binomial regression models. Results: There were 136 admissions of adults with T1D over the study period (T1DI n=67, control n=69). The median age was 49 (IQR 33-64) years; 53% were females. Specialist diabetes team review was more common in the T1DI cohort compared to the control group (endocrinology review AOR 10.27, CI 2.26-46.55, P=0.003; diabetes education review AOR 7.22, CI 2.73-19.10, P&amp;lt;0.001). Prompt endocrinology review (no later than day 2 of admission) was more likely in T1DI inpatients (AOR 2.77, CI 1.15-6.65, P=0.023). Conclusion: The T1DI initiative was associated with improvements in occurrence and timing of specialist diabetes team review for inpatients with T1D. This initiative could be implemented in other hospitals with EMR systems, enabling more effective identification and management of T1D in the inpatient setting. Disclosure R. Loveridge: None. N. Torkamani: None. S.L. Patterson: None. L. Churilov: None. E.I. Ekinci: None.

Abstract C070: Racial disparities in body mass index and ovarian cancer risk in the OCWAA Consortium

Abstract Objective: Obesity disproportionately affects African American (AA) women, and there is some suggestion that its association with ovarian cancer risk may be stronger among AA compared to white women, but no study to date has been adequately powered to compare risk between the two populations. The Ovarian Cancer in Women of African Ancestry (OCWAA) Consortium provided a unique opportunity to evaluate the association between body mass index (BMI) and invasive epithelial ovarian cancer (EOC) risk in AA and white women and to estimate the contribution of obesity to ovarian cancer risk in both racial groups. Methods: The OCWAA Consortium is a collaboration of six of the largest epidemiologic studies of ovarian cancer in the United States that include AA women: four case-control and two case-control studies nested within cohort studies. The six studies are the Chicago Case-Control Study, the North Carolina Ovarian Cancer Study, the Los Angeles Ovarian Cancer Case-Control Studies, the African American Cancer Epidemiology Study, the Black Women’s Health Study, and the Multiethnic Cohort. BMI before diagnosis was available in all the studies and data on BMI and relevant confounders were harmonized for analyses. There were 1,144 AA cases, 2,910 AA controls, 3,174 white cases, and 9,160 white controls included. Adjusted odds ratios (ORs) and 95% confidence intervals (CIs) were computed using random-effects multi-level logistic regression models separately by race and histotype with control for relevant covariates, which included study, age, education, parity, oral contraceptive use, age at menarche and family history of breast/ovarian cancer. Multinomial regression was used for histotype analyses. Population attributable risk (PAR) estimates were computed by race and histotype. Results: The prevalence of obesity (BMI≥30 kg/m2) was higher in AAs compared to white women for both cases (53.2% vs 21.4%) and controls (45.7% vs. 18.4%). For EOC, there was little evidence of an association with obesity for white women. For AAs, risk was elevated for BMI≥30 kg/m2 (ORBMI 30-34=1.37, 95% CI: 1.12, 1.69; ORBMI≥35 = 1.26, 95% CI: 0.74, 2.15) relative to BMI 18.5-&amp;lt;25 kg/m2. There was an elevated risk for non-high-grade serous EOC for both AA and white women (for AA women, ORBMI 30-34=1.67, 95% CI: 1.14, 2.44 and ORBMI≥35 = 1.76; 95% CI: 0.84, 3.69; for white women, ORBMI ≥35 = 1.39; 95% CI: 1.06, 1.82), but not for high-grade serous EOC. The PAR for BMI≥30 was 7.1% (95% CI: 2.4, 15.6) for AAs and 2.6% (95% CI: -0.3, 5.4) for whites for all EOC. For non-high-grade serous EOC, the PAR was 16.5% (95% CI: 3.4, 28.2) for AAs and 6.4% (95% CI: 2.2, 10.3) for whites. Conclusions: Obesity was a contributor to overall EOC risk among AA women but not among white women. Obesity did not contribute to the risk of high-grade serous EOC for AAs or whites, but it was associated with non-high-grade serous cases among both AA and white women. The association with non-high-grade serous EOC was greater among AAs than whites. Citation Format: Elisa V Bandera, Fabian Camacho, Deanna Chyn, Emily K Cloyd, Traci N Bethea, Alicia Beeghly-Fadiel, Charlotte E Joslin, Evan Myers, Patricia G Moorman, Heather M Ochs-Balcom, Holly R Harris, Lauren C Peres, Veronica Wendy Setiawan, Anna H Wu, Lynn Rosenberg, Joellen M Schildkraut. Racial disparities in body mass index and ovarian cancer risk in the OCWAA Consortium [abstract]. In: Proceedings of the Twelfth AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2019 Sep 20-23; San Francisco, CA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2020;29(6 Suppl_2):Abstract nr C070.

Aggressive care at end-of-life in the Veteran’s Health Administration versus fee-for-service Medicare among patients with advanced lung cancer.

12025 Background: The Veteran’s Health Administration (VHA) allows simultaneous receipt of cancer treatment and hospice care, termed concurrent care, while fee-for-service Medicare does not. Although many physicians who care for patients in the VHA also care for private sector patients, it is unclear whether there is a “spillover” relation between end of life (EOL) care in the VHA and Medicare systems at the regional level. We examined temporal trends, as well as regional-level associations between Medicare and VHA EOL practice for patients with advanced lung cancer. Methods: We conducted a retrospective study on VHA and SEER-Medicare (SM) decedents from 2006-2012 with stage IV non-small cell lung cancer (NSCLC) who received any lung cancer care. Aggressive care (AC) at EOL was defined as any of the following within 30 days of death– intensive care unit (ICU) admission, no-hospice care, cardiopulmonary resuscitation(CPR), mechanical ventilation (MV), &gt; 1 inpatient admission and receipt of chemotherapy. Descriptive statistics were used to compare outcomes. We also analyzed the association between Medicare hospital referral region (HRR) hospice admissions, Medicare HRR EOL spending, and VHA AC use adjusted for patient’s characteristics using a random intercept mixed effect logistic regression model after matching VHA facilities with Medicare facilities in a particular HRR. Results: AC use significantly decreased during the study period, from 46% to 31% among 18,371 Veterans and from 42% to 38% among 25,283 in the SM cohort, (t-test P &lt; .05). Hospice use significantly increased within both cohorts (p &lt; .001). The receipt of chemotherapy at EOL was similar for both cohorts throughout the study period. Veterans who received care in regions with higher hospice admissions among Medicare beneficiaries were significantly less likely to receive AC at EOL (adjusted Odds Ratio (aOR): 0.13 95%CI: 0.08-0.23, P &lt; .001) than veterans in regions with lower Medicare hospice use. Medicare HRR spending at the EOL was not associated with receipt of AC among Medicare beneficiaries (aOR): 1.004 95%CI: 1.00-1.009, P = 0.07). Conclusions: Perhaps due to availability of concurrent care, VHA patients received less aggressive care at EOL as compared to SM patients. At the regional level, greater hospice use among Medicare beneficiaries was significantly associated with reduced AC within the VHA.

The role of education and other transition milestones in the progression of social inequalities in cigarette smoking between the ages of 18 and 25: Evidence from the Canadian National Population Health Survey

Young adulthood is now considered a sensitive period in the progression of health inequalities over the life-course. This age group experiences highly dynamic and socially patterned life-course events that require nuanced modelling choices compared to those commonly used in public health sciences. To illustrate this, we estimate changes in the risk of smoking according to student status, employment status, living arrangements, and relationship status at different ages across education categories. We used longitudinal data in 1,243 young adults followed every two years between the ages of 18–19 and 24–25 in the Canadian National Population Health Survey (1994–95 to 2010–11). We examined the age-graded associations of occasional and daily smoking with educational attainment and transition statuses using random-effects multinomial logistic regression models with interaction terms. Post-secondary education, living with parents, studying, and being in a relationship were associated with a lower risk of daily smoking in main models, with some of these associations significantly varying in magnitude by age. The risk associated with living without parents at ages 18–19 disappeared by ages 24–25. Being single was both a protective factor at ages 18–19 and a risk factor at ages 24–25. Finally, the risk associated with being single was also stronger among those who did not pursue post-secondary education. These findings support the argument that the simple conceptualization of young adults’ circumstances – independent from age and social context – provides a limited understanding of the progression of health inequalities over the life-course. Research needs to consider the dynamic and intersecting nature of transition milestones during this intensive life period.

O10.3. EARLY CHILDHOOD DEVELOPMENT AND LATER PSYCHOTIC EXPERIENCES: FINDINGS FROM THE RAINE STUDY

BackgroundEarly childhood development is considered to be the most critical developmental phase throughout the lifespan. Cognitive and motor dysfunction are hallmark features found along the psychosis spectrum and have also been shown in young people who report subclinical psychotic experiences (PE). Whether such deficits can be detected in infancy is not yet known. The aim of the present research was to examine early childhood developmental skills and their association with PE in childhood and adolescence.MethodsData was taken from the Western Australia Pregnancy Cohort (Raine) Study, a pregnancy cohort study in which 2,868 infants were enrolled. Early childhood development was measured at age 1, 2 and 3 using the Infant Monitoring Questionnaire (IMQ; now the Ages and Stages questionnaire. Outcome of interest: PE was measured at age 10 via parent-report, and at age 14 and 17 via self-report by the participants using the 2 items (‘hearing things’ and ‘seeing things’) from the Child Behaviour Checklist (CBCL)/Youth Self-Report (YSR) thought problems subscale. The PE group included any participant who endorsed either or both of the items at any of the 3 time-points. The PE group was further subdivided into those who endorsed transient PE (one time-point) or recurring PE (2 or more time-points). Random effects logistic regression models were performed to investigate the relationship between early childhood development and later risk. Developmental time specific investigations (at age 1, 2 and 3 years of age separately) were also conducted using logistic regressions. Cumulative risk based on category specific deficits at age 3 was also calculated. This analysis was also applied to compare the PE group vs. controls, the transient PE group vs. controls and the recurring PE group vs. controls.ResultsIn the first 3 years of life, lower scores in communication (adjusted OR = 1.32, 95% CI = 1.01–1.72, p = 0.05) and adaptive (problem-solving) (adjusted OR = 1.26, 95% CI = 1.01–1.59, p = 0.048) skills were found to be predictive of PE in childhood and/or adolescence. For the age specific analysis, adaptive skill deficits specifically at 1 year of age predicted PE (adjusted OR = 1.18, 95% CI = 1.03–1.35, p = 0.017), while at 2 years of age, deficits in communication (adjusted OR = 1.35, 95% CI = 1.18–1.54, p<0.001) and adaptive skills (adjusted OR = 1.25, 95% CI = 1.08–1.45, p<0.005) predicted PE. Importantly, at 3 years of age, deficits in any of the 5 categories (communication, gross motor, fine motor, adaptive and personal social) were predictive of PE. Deficits (lowest 10th percentile) in 1–2 categories at 3 years of age led to an almost a 2-fold increase risk of having a PE (adjusted OR = 1.94, 95% CI = 1.38–2.71, p < 0.001), while deficits in 3–5 categories led to a 3-fold increased risk of having a PE (adjusted OR = 3.02, 95% CI = 1.65–5.50, p = <0.001). When subdividing PE into those with transient and recurring PE, different patterns emerged, in which motor deficits in the first 3 years of life were more associated with the recurring PE group.DiscussionThe present research suggests that lower scores in early childhood developmental skills in the first 3 years of life are strongly associated with childhood and adolescent PE. Results showed that deviances in communication and problem-solving abilities during the first 3 years of life are particularly associated. At age 3 specifically, deficits in all domains of development are predictive of PE, and the greater the number of category deficits, the greater the risk. Differing patterns of development emerge for the transient PE and recurring PE groups. The findings enlighten the understanding of the neurodevelopmental origins of early onset PE for a proportion of individuals.

Safety and Efficacy of Tumor Necrosis Factor Antagonists in Older Patients With Ulcerative Colitis: Patient-Level Pooled Analysis of Data From Randomized Trials

Treatment of older patients (more than 60 years) with ulcerative colitis (UC) can be a challenge, because they might be more vulnerable to adverse events (AEs). We determined the effects of age on the safety and efficacy of anti-tumor necrosis factor (TNF) therapy in a pooled analysis of data from randomized trials. We obtained individual patient-level data from 4 trials of anti-TNF therapy for patients with UC from the Yale Open Data Access Project. Participants were assigned to groups of older age (60 years or older) and younger age (younger than 60 years). The primary outcome was difference in serious AEs (SAEs), defined as death, life-threatening event, hospitalization, and/or significant disability. Secondary outcomes were severe infections, non-severe infections, neoplasms, and achievement of clinical remission, defined by trial investigators as Mayo score ≤ 2 with no sub-score >1 at the end of induction or maintenance therapy. A random effects logistic regression model was fitted to estimate the effect of anti-TNF therapy on safety and efficacy by age, adjusting for confounders and trial-level effects. The study cohort included 2257 patients (231 60 years or older). Higher proportions of older patients receiving anti-TNF therapy had SAEs (20%) and hospitalizations (14.4%), compared with younger patients (10.2% had SAEs and 5.2% were hospitalized); there were no significant differences between groups in proportions with severe or non-severe infections. Compared with placebo, there was no significant difference in safety risks associated with anti-TNF therapy (SAEs reduced by 5.4% in older patients vs reduction of 2.4% in younger patients; hospitalizations reduced by 6.7% in older patients vs reduction of 2.5% in younger patients; severe infections reduced by 3.1% vs increase of 0.7% in younger patients). There was no significant difference in between older vs younger patients in efficacy of anti-TNF therapy in inducing remission (odds risk ratio, 1.05, 95% CI, 0.33-3.39) or in maintaining remission (odds risk ratio, 0.49; 95% CI, 0.18-1.33). In a pooled analysis of data from randomized trials, we found that older patients with UC have an increased baseline increased risk of SAEs, but no increase in risk can be attributed to anti-TNF therapy in older vs younger patients.

Predictors of Intradialytic Symptoms: An Analysis of Data From the Hemodialysis Study

Most patients receiving maintenance hemodialysis (HD) experience adverse symptoms, which are associated with decreased quality of life. Despite decades of experience, our understanding of causes of HD symptoms remains limited. We aimed to identify modifiable patient- and HD-related predictors of intradialytic symptoms. Prospective cohort. We leveraged patient-level (n=1,838) and HD session-level (n=64,797) data from the Hemodialysis Trial. Pre-HD serum urea nitrogen (SUN) level, pre-HD systolic blood pressure (SBP), intradialytic SBP decline, and ultrafiltration rate (UFR). Intra-HD symptoms, including cramps, nausea, chest pain, headache, and lightheadedness. Random-effects logistic regression models. Overall, symptoms occurred in 10.7% of HD sessions. Higher pre-HD SUN level (per 10 mg/dL) was associated with higher adjusted odds of muscle cramping and lightheadedness (adjusted ORs [aORs] of 1.20 [95% CI, 1.17-1.22] and 1.13 [95% CI, 1.08-1.18], respectively). SBP decline (from the predialysis value to the dialysis session nadir, per each 10-mm Hg decrease) was associated with greater risk for muscle cramping, headache, chest pain, vomiting, and lightheadedness (the largest aORs were for the 2 latter symptoms: 1.24 [95% CI, 1.20-1.28] and 1.37 [95% CI, 1.33-1.42], respectively). Higher UFR (per 1 mL/kg/h) was associated with greater odds of cramping (aOR, 1.03; 95% CI, 1.02-1.03). Conversely, higher pre-HD SBP (per 10 mm Hg) was associated with reduced risk for vomiting (aOR, 0.88; 95% CI, 0.85-0.92) and lightheadedness (aOR, 0.82; 95% CI, 0.80-0.85). Measured osmolality, dialysate prescription data, and time stamps for symptom occurrence were not available. Clinical trial data may not be broadly generalizable. Higher pre-HD SUN level, UFR, pre-HD SBP, and SBP decline are independently associated with different patterns of adverse intradialytic symptoms. Recognition that different symptoms may have variable causes may allow tailoring of personalized treatments in future interventional studies.

Time-varying Analysis of Traffic Conflicts at the Upstream Approach of Toll Plaza

This study investigates the traffic conflict risks at the upstream approach of toll plaza during the vehicles’ diverging period from the time of arrival at the diverging area to that of entering the tollbooths. Based on the vehicle’s trajectory data extracted from unmanned aerial vehicle (UAV) videos using an automated video analysis system, vehicles’ collision risk is computed by extended time to collision (TTC). Then, two time-varying mixed logit models including time-varying random effects logistic regression (T-RELR) model and time-varying random parameters logistic regression (T-RPLR) are developed to examine the time varying effects of influencing factors on vehicle collision risk, and four models including the standard random effects logistic regression (S-RELR) model, standard random parameters logistic regression (S-RPLR) model, distance-varying random effects logistic regression (D-RELR) model and distance-varying random parameters logistic regression (D-RPLR) are developed for model performance comparison. The results indicate that the T-RPLR model has the highest prediction accuracy. Eight influencing factors including following vehicle’s travel distance, following vehicle’s initial lane, following vehicle’s toll collection type, leading vehicle’s toll collection type, distance between two vehicles’ centroids, and following vehicle’s speed, are found to have time-varying effects on collision risk. Meanwhile, the first six factors are found to exhibit heterogeneous effects over the travel time. Another important finding is that the vehicle that comes from the innermost lane has an increasing trend to be involved in traffic conflicts, whereas the collision risks of other vehicles decrease as the travel time increases. Moreover, vehicles with higher speed have a decreasing probability to be involved in crashes over the travel time. Interestingly, the results of D-RPLR model are similar with that of T-RPLR model. These findings provide helpful information for accurate assessment of collision risk, which is a key step toward improving safety performance of the toll plazas’ diverging areas.

Association between childhood conditions and arthritis among middle-aged and older adults in China: the China Health and Retirement Longitudinal Study

AbstractThis study examined the association between childhood conditions and arthritis among middle-aged and older adults in China. The data were derived from the 2015 wave and the life-history module of the China Health and Retirement Longitudinal Study. Face-to-face interviews were conducted with respondents age 45 and over across China. Multiple imputation was used to handle the missing data, generating a final analytic sample of 19,800. Doctor-diagnosed arthritis was the main outcome variable. Random-effects logistic regression models were used to test the proposed models. Approximately 8 per cent of the respondents had better family financial status in childhood than their neighbours. Close to 8 per cent had been hospitalised or encountered similar conditions (e.g. confined to bed or home) for at least one month in childhood. Around one-third reported better subjective health in childhood than their peers. The majority of the respondents (80%) reported that they had stable health resources, and that their mothers were illiterate during their childhood. Childhood family financial status, subjective health, mother's education, access to health care and medical catastrophic events were found to be significant factors associated with arthritis in later life, after controlling for adulthood and older-age conditions (family financial status: odds ratio (OR) = 0.885, 95 per cent confidence interval (95% CI) = 0.848–0.924; subjective health: OR = 0.924, 95% CI = 0.889–0.960; mother's education: OR = 0.863, 95% CI = 0.750–0.992; access to health care: OR = 0.729, 95% CI = 0.552–0.964; medical catastrophic events: OR = 1.266, 95% CI = 1.108–1.446). The study results highlight an important role that childhood conditions play in affecting the onset of arthritis in late life in China. Health-care providers may consider childhood conditions as a valuable screening criterion to identify risk populations, which could be used to guide health promotion and prevention programmes, and promote healthy ageing.