The implementation of variance analysis and analysis of interactive effect in multi-factor research design / 中国职业医学

Multi-factor research design is widely applied in scientific research. It can simultaneously explore the effects of multiple factors on outcome indicators. The consideration of the interactive effects of different factors is a critical issue when analyzing this type of data. The analytic strategy for main effects or simple effects depends on the significance of the interactive effect. However, many researchers tend to skip the analysis on interactive effects, or wrongly select statistical analysis method because of ignoring the test result. In this study, SPSS 20.0 and R 3.6.1 statistical software were used to simulate and illustrate how to analyze data from two most popular multi-factor design data——factorial design and repeated measurement design. The significance of evaluating interactive effect and corresponding key point analysis was explained. The possible consequences of ignoring the statistical significance of interactive effects were indicated, that include leading to low inspection efficiency, prone to draw wrong conclusions, loss of valuable information in the original data, or loss of practical significance of the analytic results. It is suggested that in the analysis of research data, we should first judge whether there are interactive effects, and then correctly choose main effect analysis or single effect analysis to avoid one-sided and wrong conclusions.

Analysis on determinants of infants overweight and obesity based on generalized estimation equation / 上海预防医学

Objective To explore the determinants of infants overweight and obesity using generalized estimation equation (GEE). Methods Data were extracted from 26 624 2-year old infants who completed required health examination from January 2017 to December 2019 in Minhang District of Shanghai.The weight for height at 6, 12, 18 and 24 months of age was calculated to evaluate nutrition status of infants.Ordinal multinomial GEE was fitted with malnutrition, normality, overweight and obesity as responding variables to explore the determinants of infants overweight and obesity. Results Incidence rate of overweight and obesity in infants of 6 to 24 months of age had a tendency of decline, and proportion of normality showed as increasing tendency along with the increase of months of age(P < 0.001).GEE showed that the following factors were more likely involved in infant overweight and obesity: male(OR=1.182), lower education level of mother (OR=1.399 for primary school, OR=1.124 for junior high school, and OR＝1.083 for senior high school, respectively), higher body weight of father(OR=1.003), higher pre-pregnant body weight of mother(OR=1.003), longer gestational age(OR=1.058), longer sleeping time at 6 month of age(OR=1.032), lower month of age (OR=2.911 for 6 months, OR=1.952 for 12 months, and OR=1.232 for 18 months, respectively).Those with low birth weight (OR=0.205 for < 2 500 g, and OR=0.410 for 2 500~3 999 g, respectively)and exclusive breastfeeding at 6 months of age(OR=0.946)had low likelihood of overweight and obesity. Conclusion Comprehensive measures should be taken to control infants overweight and obesity, including encouraging expectant parents to maintain normal body weight before pregnancy, strengthening the education of scientific feeding knowledge, and extending exclusive breastfeeding.

ROC Curve Based on Generalized Linear Mixed Effects Models in Repeated Diagnostic Tests / 中国卫生统计

Objective To investigate the impact of covariates on diagnostic test and assess the correlation between re-peated measurement data,this paper explores innovative modeling techniques of ROC curve.Methods We introduce the new ROC curve method based on generalized linear mixed effects model and apply Bayesian techniques to parameters estimation with Winbugs Software.Further,areas under the ROC curve(AUC)with different values of covariates could be calculated in terms of assessment.Results Cases analysis results indicate the proposed method efficiently explores the repeated measurement data and provides parameters with practical significance,serving as a golden reference.Conclusion The ROC curve based on generalized linear mixed effects models can be effectively used to solve the test accuracy evaluation problem of the repeated diagnostic trials.

Elevated zero maze as an animal model of state anxiety in Kunming mice：test-retest reliability / 中国药理学通报

Aim To explore the test-retest reliability of elevated zero maze (EZM), as an animal model of state anxiety in Kun-ming mice. Methods The behavior of adult, male and female, Kunming mice in EZM was recorded twice, for five minutes, with a one-week inter-trial interval. The following parameters were evaluated: percentage of time on EZM open arms (Otime% ), time on EZM open arms ( Otime), time on EZM closed arms (Ctime), percentage of entries into EZM open arms (Oentries% ), entries into EZM open arms (Oentries), entries into EZM closed arms (Centries), total entries into EZM open and closed arms (Entries). Subsequently, the intraclass corre-lation coefficient ( ICC) and the kappa agreement index ( κ) were calculated for these parameters. Results ICC for Otime%(♂ ICC = 0. 753,P < 0. 01;♂ + ♀ ICC = 0. 535,P < 0. 05), Otime (♂ ICC = 0. 753, P < 0. 01; ♂ + ♀ ICC = 0. 535, P <0. 05), Ctime ( ♂ ICC = 0. 753, P < 0. 01; ♂ + ♀ ICC =0. 535, P < 0. 05), and Oentries (♂ ICC = 0. 719, P < 0. 01;♀ ICC = 0. 494, P < 0. 05; ♂ + ♀ ICC = 0. 583, P < 0. 01), Centries (♀ ICC = 0. 658, P < 0. 01; ♀ ICC = 0. 508, P <0. 05; ♂ + ♀ ICC = 0. 562, P < 0. 01), Entries ( ♂ ICC =0. 691, P < 0. 01; ♀ ICC = 0. 502, P < 0. 05; ♂ + ♀ ICC =0. 574, P < 0. 05) showed a good and fair test-retest reliability, respectively. Only κ for Otime% (♂ κ = 0. 393, P < 0. 05), Ctime (♂ κ = 0. 393, P < 0. 05), Oentries (♂ κ = 0. 308, P< 0. 05; ♂ + ♀ κ = 0. 256, P < 0. 05), Centries ( ♂ κ =0. 427, P < 0. 01; ♂ + ♀ κ = 0. 238, P < 0. 05), Entries (♂κ = 0. 469, P < 0. 01) displayed a significant but poor reliability between the test-retest trials. Conclusions The parameters Otime% and Entries for Kunming mice in EZM are stable over time, and indicate an acceptable reliability with a one-week in-ter-trial interval. Meanwhile, more attention should be paid not to repeated measure effect, but to the sex effect (Male animals recommended to EZM, especially).

Elevated plus maze as an animal model of state anxiety in Kunming mice:test-retest reliability / 中华行为医学与脑科学杂志

Objective To explore the test-retest reliability of elevated plus maze (EPM) which is taken as an animal model of state anxiety in Kunming mice.Methods The behavior of adult,male and female,Kunming mice in EPM was recorded twice with one-week inter-trial interval for lasting five minutes.The following parameters were evaluated:percentage of time on EPM open arms (Otime％),percentage of time on EPM closed arms (Ctime％),time on EPM open arms (Otime),time on EPM closed arms (Ctime),percentage of entries into EPM open arms (Oentries％),entries into EPM open arms (Oentries),entries into EPM closed arms (Centries),total entries into EPM open and closed arms (Entries).Subsequently,the intraclass correlation coefficient (ICC) and the kappa agreement index (κ) were calculated for these parameters.Results ICC for Otime％ (♂ ICC=0.506,P＜0.05;♀ ICC=0.456,P＜0.05;♂ + ♀ ICC=0.492,P ＜0.01),Oentries (♂ ICC=0.673,P＜0.01;♀ ICC=0.648,P＜0.01;♂ + ♀ ICC=0.665,P＜0.01),Centries (♂ ICC=0.704,P＜0.01;♀ ICC=0.661,P＜0.01;♂ + ♀ ICC=0.686,P＜0.01),Entries (5 ICC=0.701,P＜0.01;♀ ICC=0.661,P＜0.01;5 + ♀ ICC=0.685,P＜0.01) showed a fair test-retest reliability,respectively.Only κ for Entries (♂ κ=0.253,P＜0.05;♀ κ=0.339,P＜0.05;♂ + ♀ κ=0.298,P＜0.01)displayed a significant but poor reliability between the test-retest trials.Conclusion The Otime％ Entries parameters as an acceptable index for Kunming mice in EPM can be used for exploratory and locomotion activity.Sex and repeated measurement have different effects on the Otime％ and Entries parameters,especially the more on the former.Therefore,Otime％ and Entries can be considered when EPM as an animal model of state anxiety in Kunming mice;meanwhile,more attention should be paid to both the sex (male animals recommended to EPM,especially) and repeated measurement effects.

Misunderstandings of statistics in medical research / 中华医学科研管理杂志

To explore the misunderstandings of statistics in medical research,we surveyed related papers from CNKI published in the first quarter of 2012 and analysed the statistical application in those papers.Misunderstandings were identified and solutions were proposed to avoid them.

Using repeated measurement of 'analysis of variance' on hypertension patients to evaluate the self-management program in the communities / 中华流行病学杂志

Objective To explore the application of repeated measurement‘analysis of variance'(ANOVA)in evaluating the effectiveness of 'community-based hypertension self-management program'.Methods A community-based parallel controlled trial was conducted among 3 communities.169 patients in intervention group took part in the course on hypertension self-management program once a week and 204 patients in control group received routine hypertension management services.Data collected through questionnaire at baseline and 6 months,12 months after intervention and were compared through repeated ANOVA measurement.Results Subjects in the intervention group showed statistical significance and linear trends in health self-evaluation,distress,in lOW spirit,self-efficacy in managing symptoms(SEMS),self-efficacy to managing diseases in general(SEMDG),systolic blood pressure(SBP)and diastolic blood pressure(DBP)over time by univariate test of repeated measurement ANOVA.The score of SEMS increased from 6.84±2.53 at baseline to 8.20±1.44 at 12 months after intervention while SEMDG from 7.28±2.45 to 8.89±1.05,and SBP decreased from 137.66±7.30 inln Hg(1 mm Hg=0.133 kPa)to 130.41±7.71 mm Hg.DBP decreased from 84.13±6.70 mm Hg to 81.04±5.98 mm Hg respectively.Only tow spirit and SBP changed over time were seen in the control group.Self-evaluation,distress,in low spirit,caused by diseases,SEMS,SEMDG and SBP were statistically different between control and intervention groups,and the effect of interaction between groups and time span were statmtically significanton indicators as self-evaluation,low spirit,self-efficacy in managing symptoms,seLf-efficacy tO manage diseases and sBP etc,by multivariate test of repeated measurement ANOVA. Conclusion Repeated measurement ANOVA not only could be used to analyze group-effect,but could also explain the effect and the interaction among groups and time,to make the results more reliable.The self-management approach could improve the health status and self-efficacy of the patients,so as to reduce the blood pressure.Our result showed that it was effective for hypertensive patients to be on the chronic diseases selfmanagement program.

Acute Caffeine Effect on Repeatedly Measured P300

The acute effect of a single−dose of caffeine on the P300 event−related brain potential (ERP) was assessed in a study using a repeatedly presented auditory oddball button−press task. A dose (5mg/kg body−weight) of either caffeine or placebo lactose, dissolved in a cup of decaffeinated coffee, was administered double−blindly to coffee drinkers who had abstained from coffee for 24hrs, with the presentation order of the sessions counterbalanced and separated by 2−4 weeks. The caffeine−treatment condition demonstrated a smaller P300 amplitude and a shorter latency overall than the placebo treatment condition. The mean P300 amplitude value difference (caffeine minus placebo) increased with the successive trial blocks. Caffeine ingestion appears to yield a lower resource−consumption and a net increase in allocating attention resources for task performance across repeated measurements.

An empirical research and estimate of bepainted membrance method for measuring expanded area in skin and soft tissue expansion / 重庆医科大学学报

0.91,no significant different was found between BMM and golden standard;t2,3=19.150,P