You are here:
Vanderbilt Biostatistics Wiki
>
Main Web
>
PedsGroup
>
PedsGroupActions
(19 Aug 2014,
MengXu
)
(raw view)
E
dit
A
ttach
-- Main.BenSaville - 06 Nov 2009 ---++ Actionable Items * Jackie Shopluck: additional analysis for Dex paper * John Soslow: manuscript review for TOF paper * Charles Phillips: residence paper ---++ On hold Items %RED% %GRAY% ---++ Completed Items * Arnold: Propensity matching for BiPAP project. (report sent on 08/19/2014) * Sarah Jaser: Analysis according to the hypothesis she sent over. * Meg and Soslow: redo the analyses: 1) use percent change of BMI in repeated measure analysis adjusting for baseline 2) use BMI Z score for Cox model * Amelia Keaton - manuscript: adding 2011-2012 data * Lindsay Roofe: remove mechnism_given==3 patients and rerun the analysis, change to row percent, try with/wo "social worker involved" in the model * Shuplock: Dex manuscript review, change the independent test to paired test after matching * Report sent: Meg McKane & Jon Soslow: assessing association between BMI and outcome of interest * Fleming: manuscript review * Fleming - additional analysis with more data * Mary Romano - Comparing two time points of data on the outcome of interest. Lost Kelly's old code need to start over. * Shopluck: Dex manuscript with new data * Jonathan Soslow: Analysis of "Obesity leads to underestimation of Ventricular Volumes and abnormal myocardial strain in repaired tetralogy of fallot as measured by Crdiac MRI" . Wait for low BMI category and will add spline of age and age*BMI in the model * John Williams: manuscript revisions * Kannankeril and Shuplock:working on the CIs of the difference on the outcomes before and after matching * Agarwal: Manuscript revision * Cary Fu: grant re-submission. Working on analysis plan and sample size calculation for grant proposal. * Kannankeril and Shuplock: Data analysis of "The effect of Peri-operative Eexmedetomidine on Arrhythmias after Surgery for congenital heart disease". Report was sent to investigator * Jill Simmons: questions for manuscript * James Gay and Charles Phillips: Analysis was done and already met with Charles to go over the results. * Abstract for Don Arnold: Spearman correlation coefficient and 95% CI at 3 time points between PEP and other outcomes * Redo the analysis for Linsay Roofe using the updated data (PEM variable was re-evaluated) * Roofe, Arnold, report was sent to investigator. Abstract was submitted. * Michael O'connor and Don Arnold: Inter-rater Reliability of Two Acute Asthma Severity Scores. Statistical Analysis Plan was sent to investigator. * Quick calculation of Weighted Kappa for Johnathan Soslow * Manuscript review for Dr. Simmons's "Prevalence of 25-Hydroxyvitamin D Deficiency in a Pediatric Population with Type 1 Diabetes Mellitus Compared to a Control Population" * Manuscript review for Dr. Agarwal's "<font face="Calibri,sans-serif" size="2">Risk factors and outcomes associated with excessive bleeding in pediatric cardiac surgery”</font> * Geoffrey Fleming - Aim: Assess if there is any improvement in the self-reported score on 3 areas and test whether the change is associated with the factor of interest. Report was sent to investigator * Agarwal-analysis for the blood products with bleeding. Done, report was sent to investigator * Rosen - check manuscript addition. Done, report was sent to investigator * Don Arnold - Analysis was done, report was sent to investigator. Aim: correlate PEP with other measurement of ASTHMA severity * Jill Simmons: Analysis was done and has been sent to investigator. * John Williams: Completed by Ben, MX was not added to the access of Redcap.Manuscript resubmission, wants to compare titers and positive rates for MPV vs. RSV viruses in prospective cohort of children. * Geoffrey Fleming - write AP * Dr. Mary Romano & Dr. Carrie Lind - passed to Kelly, Dec 2012 * SAP has been approved by PI's * Waiting for PI to send six months of data through Red Cap * Data exported from RedCap (07/31/2012) * Statistical Analysis in progress * Had to stop the analysis because I would need the post-prompt data which is not ready, I thought it was part of initial data loaded but is not.(08/07/2012) * PI's are including another variable on RedCap analysis will start soon. * Dr. Killen report is also done, Kelly and I kind of disagreed with the method used, I wanted to use Fisher Exact test and she wanted Chi-square test. * Dr. Thomsen * SAP has been sent to them waiting for approval from PI. * PI has approved and analysis should start any time soon with dummy data set from PI. * Analysis is in progress using dummy data set, as of 11/05/2012 * Original data set has just been received and has been run. * Statistical analysis report is ready 11/05/2012. * Analysis report sent to Ben for review. * Dr. Donald Arnold * Meeting with Ben possibly sometime this week to discuss SAP (Meeting date to be scheduled). * SAP partially done, sent to Ben for comments. * Need to update comments from Ben on SAP * SAP is done sent to Ben for comments (07.23.12) * SAP has been modified again and sent to Ben for comments (07.25.2012) * SAP sent to PI, we don't have the data yet. (07/31/2012). * We met to discuss SAP and SAP has been updated and sent to PI's.(09/07/2012) * Waiting to hear from them * Received data today : Monday 10/15/2012. * PI has made significant changes to request , refer to PI's email 10/16/2012 and Ben's email: 10/21/2012 * Analysis in progress * Updates have been done sent to Ben for review last week. * Dr. Bremer wants us to do further analysis on two extra variables, data is in the dropbox folder. * Baseline values for this new data are all zeros so there is not need to adjust for baseline * This should be completely done by close of today : 11/12/12 * Dr Abramo wants us to reproduce the Statistical Analysis report on NIRS/DKA hypertonic Solution study with new data. * * Data is ready in Dropbox folder (Old data 31/new data 30) * Note every base excess variable should be negative it appears on data as positive * Remember each patient record is a sheet and read in one at a time in turns. * Updated and sent to Ben for review. * Dr. Abramo want Mean, SD, for variables in study , would give him mean, min, max, median for variables but some of these variables have issues and would still need further cleaning. Would be ready today. 05/11/2012 * Meeting with Ben/Rebecca/Dr. Abramo at 2pm wed(08/15/20/12) * SAP has been sent to Ben for review.(08/16/12) * SAP has been sent back to me for update * Updated SAP sent to Ben for approval (08/21/2012) * Has been sent to PIs, and its been approved * Data is received (08/27/2012)(Deadline in Dec 2012) * Statistical analysis in progress now...10/05/2012 * Analysis passed on to Mario *Dr. Yaa Kumah * * Would need to schedule a meeting with her again to modify the analysis plan. * Meeting at Ben's office 07/20/2012 Friday 3pm with Yaa Kuma. * SAP is complete sent to Ben for comments, (07.23.12). * Yaa is out of office, auto response email says, until 07/30/2012. (She has to confirm the data for this project). * SAP has been sent to Yaa (07.25.2012) * Waiting for data * PI wants to meet with me tomorrow ( 08/01/2012: 3pm) to make minor changes to statistical analysis plan. * Meeting :PI said data has two different variables that measure cell phone use: Mobile phone use and diabetes Mobile phone use each with scale 1- 27. * Research question : 1.Wants to know what extra explanation does each predictor contribute to the variation in the response variable in each full Model: (a) Adherence = prob.solv + barriers + MobilePhoneUse, (b) A1C =Adherence+prob.solv + barriers +MobilePhoneUse, (c) A1C =Adherence+prob.solv + barriers+Dia.MobilePhoneUse, (d) Adherence=prob.solv + barriers +Dia.MobilePhoneUse. * 2. Which of the independent variables has a greater effect in predicting the response variable, in the event that the independent variables have different units of scale? * Analysis Plan (Main Points) : 1. Fit a linear regression model ; A1C = prob.solv + barriers + MobilePhoneUse +Dia.MobilePhoneUse, and set contrast option in R to generate Type III Sums of Squares. eg. options(contrasts=c(unordered="contr.sum", ordered="contr.poly")), Anova(lm(Y ~ x1+x2+x3 , data=dat1, type=3)) or use SAS . Type III SS test the Null Hypothesis: for a chosen B = 0 , when all other predictors are in the model. (comments:Type III SS have received a lot of criticisms I said, but PI insisted she wants that) . * 2. Standardized all variables ( Y and X) so they all have a std of 1, and fit a similar model as above, and compare the standardized coefficients for the biggest in value. * PI wants to better put the data in shape before sending it (data not ready). * SAP is rewritten after meeting, has been sent to Ben for approval. * Data still not ready. * Data received (08/16/12) <span style="font-size: x-small;">Essentially columns I-M :(HbA1C | adherence | Barriers | problem solving |mobile phone use ).</span> * <span style="font-size: x-small;">In progress ..........over now</span> * <span style="font-size: x-small;">Analysis report sent to Ben (09/14/2012)<br /></span> * Dr. Amanda Berger * Meeting at 2:30pm at MCN 2220 to work on Abstract submission. 08/09/12 * SAP and part of the analysis has been done to meet abstract deadline, to be continued. (08/13/12) * Meeting with Dr. Amanda Berger and Dr. Franguol 2pm (08/14/2012) * Meeting to be schedule before the end of August to continue with project above. * Meeting on Monday 2:30pm (08/20/12) . Ben, me and PI's Dr. Berger , Dr Frangoul. * SAP is done and sent to Ben for approval.(08/21/2012) * SAP is sent to PI and its been approved. ( 08/23/2012) * Statistical Analysis Report is Ready (08/27/2012).- Will send to Ben today. * Sent back to me updates done and sent to Ben (08/31/2012) * Statistical Analysis Report sent to PI (09/05/2012) * Dr. Drew Bremer * Completed SAP, sent to Ben for approval. * Need to update comments from Ben on SAP. * Would meet Ben on last method on SAP for clarification, time not set yet. * SAP has been modified again and sent to Ben for comments (07.25.2012) * Statistical Analysis is in progress * Statistical Analysis Report is done and ready ( 08/03/2012). * Update of Statistical Report in progress (now). * Updates done and sent to Ben for approval (08/23/2012) * Sent back to me for updates, updates are done left with executive summary with referencing .(08/31/2012) * Updates are completely done sent to Ben for approval. (09/04/2012). * Statistical Analysis Report sent to PI ( 09/05/2012). * Dr. Jeff Lomenick * Meeting to discuss the results of Analysis Report on Thursday 2pm, 07.26.12, at Ben's Office. * Need to update Statistical Analysis report with comments from last meeting. * Updating SA report : in progress * Updated reports are done and ready for both TSH and Free.T4 ( 08/07/2012). * Report has been updated and sent to PI, PI is OK with it. * Dr. Hemant Agarwal. * This project will be handled by Yan Hu * SAP is ready in Wenli's Bulk Files. * Deadline first or second week in August. * Dr. Jeff Lomenick * * Currently cleaning Data * Meeting with Ben before 11:30am on Wed June 27, 2012 to discuss SAP * Analysis is completed sent to Ben for comments. * Analysis id done with. Old and Dead Meetings * Dr. Frangoul and Dr. Amanda Berger : Meeting set for Friday 14th, 2:30pm at Ben's Office . * Dr. Abramo : Meeting is set for September 11th 4pm at Bens Office :Confirmed * Dr. Stacy Killen : Meeting Thursday : Sept 06. 4pm. at Ben's Office...status : over * Dr. Michael Glenn O'Connor: Meeting on Friday , Sept. 7th --12:45pm to 1:30 pm over * Meeting on Tuesday Oct 2, at 3pm at Ben's office : Bill Herman, Ben, Evan Crossfield, Steve * Dr. Ettinger * Rerun the analysis report using 5 level acuity level * Dr. Yaa Kumah (Steve) * Research questions (grant) * To determine the sample size of exploring the association between A1C values and general phone use, social net work and diabetes related phone use * To doing the sample size justification of exploring the association between barrier questions, problem solving questions, and adherence questions and general phone use, social net work and diabetes related phone use * Research questions (Analysis 96 patients) * To explore the association between barrier questions, problem solving questions, and adherence questions and general phone use, social net work and diabetes related phone use * Issues: Outcomes are in 5 point likert scale and she'd like to look at 50 questions as outcome and 20 questions as predictors, not enough sample size and Type I error inflation. * Analysis plan * Using the summary scores for each questionnaire, fit a linear regression model on Adherence, as a function of barrier, problem solving questions, mobile use, A1C. * Using the summary scores for each questionnaire, fit a linear regression model on A1C, as a function of barrier, problem solving questions, mobile use, Adherence. * Using the summary scores for each questionnaire, fit a linear regression model on Adherence, as a function of barrier, problem solving questions, diabetic mobile use, A1C. * Using the summary scores for each questionnaire, fit a linear regression model on A1C, as a function of barrier, problem solving questions, diabetic mobile use, Adherence. * Dr. Yaa Kumah-Crystal (the same analysis but using the new data to adjust for more covariates) * Analysis plan * Fit a linear regression model on A1C values as a function of regiment status (Pre/Post), age, gender, status*age, status*gender with a sandwich estimator accounting for repeated measures [age being categorized into 3 interval] * Provide predicted plot of A1C vs. time for each age group * Provide predicted plot of A1C vs. time for each gender group * Perform wilcoxon signed rank test to weight and height before and after regimen chang, paired t test if normality assumption is satisfied. * Perform wilcoxon signed rank test to compare # of hospitalizations before and after regimen chang, paired t test if normality assumption is satisfied. * Dr. Agarwal * Review manuscript * Revisit the cpb data to determine whether cpb during time can predict cardiac or non-cardiac complications, LOI, LPICU, and Death, with adjustment for other confounders. * Prepare the analysis plan. * Katherine Kudyba * Provide descriptive stats for 15 pcr results * Use the new time window and rerun the report * Include the p values in Table 1 * Dr. Jonathan Soslow * Fit a linear regression on average T1 time, with the predictor of group, adjusting for age * Assess the relationship of region and average t1 time (possibly variance) at each region of MRI image with the outcome--"LVEF", adjusting for age, medication, etc. * Write the analysis plan * Dr. Shoemaker * Provide the plot in R * Dr. Puzanovova * Provide two figures at baseline * Dr. Tom Abramo * Perform paired t test for the baseline and three post treatment time points(5,10,15) * Conduct the linear regression to determine the correlation between the changes pCO2, PH, and GCS and the delta changes in outcome * Fit the linear mixed effects model to explore the time trend in NIR readings. * Dr. Jessica Hebert Mouledoux * Create demographic table * Create fisher exact results table * Dr. Don Arnold * Bayesian model averaging method to select variables in the dataset * Figure out confidence intervals for the parameters * Include FEV1 just for sensitivity test * Dr. Revi and Dr. Phillips * Sample size justification for matched cased control study to explore the differences in number of repeats in estrogen receptor * Analysis plan to compare case control group. t test on 150 patients in each group * Dr. Mary Romano * Send an email to attend the clinic * Dr. Bondi * Add one more time point * Describe time * Set up an meeting to discuss * Dr. Ettinger * Provide publication quality plot * Dr. Bondi * Longitudinal study on diabetic patients with and without intervention * Provide descriptive stats * Provide histograms of A1C and number of observations * Create a table showing the number of observations * Fit linear regression model using main effects only, with cubic spline function and GEE sandwich estimator * Provide contrast table for the differences between two time points * Provide Predicted plot vs time * Fit linear regression model using main effects and one interaction term, with cubic spline function and GEE sandwich estimator * Provide contrast table for the differences between two time points * Provide Predicted plot vs time by three age level * Dr. Jim Gay * Comparison of sample of chart reviews (n=200) vs. all readmissions for manuscript of readmission preventability * Provide CIs for the comparisons between reviewed and un-reviewed patients on a list of predictors * Dr. DeBaun * Table 1 of demographics (see previous manuscript) with 7 columns * Give Dionna first 25 and last 25 patients IDs in blood pressure, age, height, weight, hemoglobin. * Frequencies for 3 types of events (silent infarct, CIA, over stroke) * Total patient years (overall, by those with and without events) * Dr. Shoe maker * Provide graphs * Review manuscript * Bootstrap CI * Dr. DeBaun * Give separate document for descriptives of variables in analysis * Dr. Ettinger * Review the draft of the manuscript * Dr. Natasha Halasa (Grant proposal) * Read about sample size calculation for non-inferiority study * Review reponses from Natasha * Write power calculation part in the grant * Dr Halasa * Clean up the new dataset * Rerun the analysis being restricted in the different time range * Dr. Arnold * Bootstrap CIs in Table 1 * Dr. Arnold * Review research letter of cPRAM * Sample size calculation * Katherine Kudyba and Dr. Halasa * Send Katie figures file * Describe 4 patients with ROS but died * List x-ray reasons for abnormal patients * Dr. Jim Gay * Conduct comparisons between A vs. C+D and B vs. C+D * Dr. Natasha Halasa * Create a graph of raw counts of bacteria across month * Rerun the analysis for Lindsey Lawrence using the complete datset * Dr. Jim Gay (hospitalization dataset) * Provide histograms for 4 continuous outcomes * Fit ols() model for LOS, charges, relative weight and age regressed on group (A, B, and D) * Use sandwich estimator to adjust for repeated measures * Fit ordinal logistic regression model for CRG group regressed on group, with sandwich estimator to adjust for repeated measures. * Dr. Ashley Shoemaker * Rerun the analysis with the new data * Katherine Kudyba and Dr. Halasa * Add two variables in the comparison list * Group those patient with both ROS and other diagnosis into ROS=N * Use categorical version of variables * Dr. Puzanovova * Replace all log variables in Table 15 * Remove all the w-variables in Table 15 * Provide descriptive tables for baseline psychological variables, with CIs and P values * Provide correlation coefficients among all 12 variables * Dr. Creech * Manuscript review * Dr. Ettinger * Provide three graphs with publication quality * Dr. Puzanovova * Add rmssd as the outcome in the secondary report * Conduct comparisons between IBS and healthy patient groups in those outcomes of interest at baseline. * Explore the correlation between physiological factors( Heart rate) and psychological factors (severity ?) * Dr. Jim Gay * Perform Kendall's concordance W to assess the inter-rater reliability * Perform GEE Poisson model on the consensus ratings without covariates * Perform GEE Poisson model on the consensus ratings regressed on a set of risk factors * Dr. Shoemaker * Fitting linear regression model adjusted by robcov() * Provide predicted means plot * Dr. Ettinger * Fit a logistic regression model on the binary outcome--admission Y/N with individual acuity level and interaction term * See how the regression model fitting results look like to decide whether need to remove average Acuity from the current model * Create a descriptive for demographic variables * Dr. Don Arnold(PRAM) * Doing research on CI of R square (bootstrap) * Doing comparisons among three scores * Standardize the cPRAM and mPRAM score to 12 points scale * Provide two plots for PRAM vs. cPRAM or mPRAM separately, with R square (different colors) on the same panel * Dr. Ettinger * Calculate the average time to admission for patients with bed requests * Provide the graph of predicted rate versus across DOW * Katherine Kudyba (waiting on the list of variables of interest) * Summary table between two groups with or without positive ROS * Summary table between two groups with or without positive virus * Summary table between two groups by age groups (<28 days versus >28 days and < 6 months) * Dr. Agarwal * Review manuscript * Dr. Jim Gay * Write analysis plan * Doing research on how to assess agreement among ordinal ratings * Dr. Ettinger * Merge the admission and visit dataset * Cluster patients into half hour interval * Count # of bed requests, # of visits, take average of acuity level * Fit a negative or poisson model on the number of bed requests, offset by number of visits, adjusted by day, month and acuity level * Provide a summary table of estimated rates * Provide a contrast test on shift intervals vs. all other intervals * Provide a predicted plot of rate across intervals * Dr. Revi * Review manuscript * Dr. Revi Mathew * Rerun the analysis on subset data * Add one variable in the model * Dr. Arnorld * Double check normalization * Add spearman correlation test * Dr. Jim Gay * Make a graph in R * Dr. Don Arnold (cPRAM) * Remove duplicate records * Provide histograms of two outcomes * Verify whether standardized variables are correct * Fit linear restricted cubic spline model on tow outcomes on each predictor (6 models) * Provide ANOVA table, summary table and predicted plots * Provide Bland Altman plots to assess the agreement of three standardized measurements * Calculate ICC with CI * Fit linear restricted cubic models with baseline measurement of FEV and PRAM * Provide ANOVA table, summary table and predicted plots * Replicate the analysis on Rios * Dr.Warolin * Created a summary variable * Provide descriptive stats for overall and by x ray status * Provide histogram of outcomes, considering log transformation if needed * Perform negative binomial regression model on the number of episodes at each position as a function of reflux status * Perform t test on three continuous outcome. Be cautious about the normality of three outcomes. * Figure out the prior distribution for BMA method * Dr.Puzanovova (07/26/11) * Generate two pdf, one with coding flipped, and word count removed; * The other one with coding flipped, and word count removed, baseline interval included in the outcome and 5 contrast comparison for outcomes mean RRI, SD RRI, and HF * Provide 5 contrasts at each interval and across all intervals, respectively * Replicate it on original scale for log transformed outcomes * Dr Agarwal * Through "Bayesian model averaging" to select predictors * Fit proportional odds logistic model to compare group differences * Review the manuscript * Dr. Creech Buddy * Review manuscript * Dr Jim Gay * Review manuscript * Fill in the stats * Check the numbers * Improve figures * Dr.Puzanovova and Dr. Walker * Provide descriptive tables * Provide histograms * Provide a graph of raw means across time points * For baseline data, fit a linear regression model with restricted cubic spine function to assess the group effect * Provide Anova table, summary table, table of broken down effects, table of predicted values, dot chart of predicted value * For longitudinal data, fit a linear regression model with restricted cubic spine function, including 3-way and 2-way interactions. * Provide Anova table, summary table, table of broken down effects, table of predicted values, line graph of predicted value * Figure out how to get effects averaging out another variable in summary table * Replace summary table stats with these averaging estimates * Dr. Revi * Provide data * Do comparison * Ben Deschner * Review abstract * Dr Joycy Granger * Fit cox regression model * Review manuscript * Dr Mathew Revi * rerun the analysis with data with new metrix * Compare the outcome with old report * Dr. Ben Deschner * Provide spaghetti plot by group * Label the data points by group in predicted plot * Add interaction term into the model * Add VSD into the propensity score model * Estimate outcome change at three time points * Help Natalia Jimenez with her numbers (manuscript) * Dr. Locklair * Provide trajectory plot of asthma score over time * Perform wilcoxon test between initial measurement and last measurement * Dr. Natasha Halasa * Restrict the date in the defined range * Rerun the analysis with the new data * Dr. Locklair * Reshape the data set from wide to long * Transform outcomes in log scale * Provide histograms and spaghetti plot for three outcomes * Calculate ICC * Fit linear model with restricted cubic spline, adjusting for repeated measures * Provide table of estimates, ANOVA tables and predicted means plot * Dr. Ben Deschner * Provide descriptive stats * Calculate propensity score * Fit a linear regression model with sandwich estimator for longitudinal outcomes * Fit cox proportional hazards model for time to event outcome * Revi * redo analysis with new data (note new cutoffs for positive vs. negative test results) and send ID #'s * Fit cox model in the report * Sara Rippel, manuscript preparation (waiting on data), Clinical outcomes of former pediatric patients with histologic reflux esophagitis * Has about 10 outcomes, mostly dichotomous, a couple continuous (depression scales), wants to compare 3 groups (GER, Normal Endoscopy, Controls) * Provide 2*2 tables and histograms * For binary outcomes, use logistic regression * Perform logistic regression models for ordinal outcomes * Perform linear regression model with continuous outcomes * In Subgroup analysis, we will fit logistic regression to calculate propensity scores * Fit a second logistic regression model to incorporate both group and propensity scores * Dr. Yaa Kumah-Crystal (wait on the data) * Add a table of predicted values at -400,0,400 * Add more predictors in the model * Provide predicted plots * Dr. Sara Horst * Perform logistic regression with restricted cubic spline on each continuous variable * Provide predicted probability plots for each continuous variable * Provide table of estimates and anova table * If depression is not significant, remove Abd symptom score from the model and rerun it. only table needed. * Power analysis * Calculate sample size for two outcomes using chi-square test for two proportions * Calculate effect size using t test * Dr. Bremer (monkey data) * Manuscript review * Dr.Revi Mathew (11/12/10) * Manuscript review * Patrick Drayna (fellow) and Dr. Don Arnold * Provide estimates and CIs for differences in predicted means between baseline and follow-up time points * Perform linear mixed effects model and provide table of estimates * Provide a predicted means plot for lme model * Add dose, age and gender to restricted cubic spline model * Calculate the standard deviation of differences between baseline and follow-up * Calculate the standard deviation within each time point * Calculate the correlation between baseline and follow-up * Dr. Bremer (monkey data) * Provides methods and summary section * Provide estimates and CI for each outcome * Provide predicted plots for each outcome, log and original scale * Dr. Arnold (Asthma patients) * Provide descriptive stats and histograms of outcomes * Test for correlation among predictors * Perform logistic and linear regression models on three outcomes (Delta or 2hr adjusted for baseline value) * Check whether sample size can support the model, if not, need to choose which one to use * Patrick Drayna (fellow), Don Arnold PAS abstract/manuscript requiring repeated measures. * Restructure data so that it's tall, not wide * Provide histograms of Isop by time to check normality assumptions * If normality satisfied, fit linear model with spline for time: y = time + dose, use Robcov for repeated measures * Dr. Bremer and Dr. Mathew * Hochberg correction for all the p value * Add two columns in the table, the new alpha level and test result * Take out all the variables with small sample size * Create a table for Spearman test to show the corrected p value and results * Dr. Bremer (grant: pilot data) * Provide histograms for original and log scale variables * Provide sample size calculation using paired t test between 12 months and baseline * Fit linear mixed effect model on each outcome for the pilot data * Provide estimates and CI for multiple comparison * Provide predicted plots for each outcome * Dr. Yaa Kumah-Crystal * Reshape the dataset from wide to long * Provide descriptive stats and histograms * Perform linear mixed effect model to compare pre vs. post A1C values * Perform linear model with spline on time variable * Provide linear graph based on the model * Dr. Sarah Horst * Perform logistic regression with restricted cubic spline * Provide predicted probability plot based on the model output * Dr. Matthew Locklair(12/2/10) * Combine individual csv file to one dataset * Seek for matches in two datasets * Provide descriptive stats and histograms of matched records * Perform correlation test for each pair * Transform three outcomes * Perform linear mixed model to test for association for athma data * Dr. Agarwal * Provide forest plot and boxplot for each table * Adjust scale of each plot * Perform proportional odds ratio test and logistic regression on morbidity after surgery, adjusted for bypass time and bleeding status * Dr. James Gay (manuscript) (11/10/10) * Review stats in manuscript * Dr. Buddy Creech (11/12/10) * Provide figures and tables for manuscript * Dr.Revi Mathew (11/12/10) * Provide descriptive tables and histograms * Compare case and control groups with Wilcoxon signed rank test * Perform spearman correlation test * Dr.Arnold (11/15/10) * Provide spaghetti plots and median plots for two outcome groups * Provide boxplots * Dr.Claudia Florez (11/12/10) * Provide boxplots for groups comparisons * Dr. Natasha Halasa (11/12/10) * Provide comparisons for the variables of interest between 6 groups * Dr. James Gay (manuscript) * Provide histograms and 2*2 tables with updated data * Perform Wilcoxon rank sum test for 0-7 days vs. 8-15 days and for for acute vs. chronic * Provide stats for two groups * Dr. Buddy Creech (10/18/10) * Provide a matrix of Spearman correlations * Perform GEE logistic regression model on children colonization * Evaluate effect of maternal colonization at birth, discharge, two months and 4 months through contrast matrix * Perform ANOVA table to give overall p value * Provide table of predicted probability for each combination of maternal colonization and time * Refit the model with the predictor Mcol at enrollment * Repeat both model with Vandy data only (be careful of the sample size * Dr. Mathew Revi * Provide descriptive stats for subgroups of patients * Provide contingency tables for two tests * Dr. Puzanovova * Provide means graph of four groups for each outcome * Dr. Shari Barkin and Dr. Rachel Biber (Manuscript revision) * Add a table of contents * Delete Sections 2.1 and 2.2 * For sections 2.4 and 2.5, use BMI z score (variable \bmizscore") instead of BMI percentile. * Perform restricted cubic splines for the predictor instead of polynomials * Give summary tables along with p-values and R-squared * Provide corresponding graph for the continuous predictor of interest. * Dr.Puzanovova (second plan) (9/29/10) * Subset data with time intervals 1-4 * Provide spaghetti plot for all patients * Calculate the difference between time interval 1 and 2, 2 and 3, 3 and 4 * Fit linear model with baseline values and predictors * Provide table of parameters and anova output * Provide predicted means for differences for each combination * Perform interaction test on two way interaction term and refit the model if not significant * Provide another table of parameters if not significant * Subset data with time intervals 1-3 * Perform interaction test on three way interaction term and refit the model if not significant * Provide table of parameters, ANOVA table and table of predicted means only for final model * Provide predicted means graphs for 4 groups vs. time intervals * Dr. Agarwal Hemant: bleeding study (9/17/10) * Provide descriptive stats * Perform logistic regression on bleeding outcome with predictors of interest * Provide predicted probability plot for continuous variables * Perform Wilcox Rank Sum test for the difference in the bleeding group vs. non-bleeding group on blood product transfusion outcomes * Provide descriptive stats for blood product use by bleeding status * Perform logistic regression on outcome bleeding with PICU variables * Provide predicted graph for delta time * Perform the Wilcox Rank Sum test and Fisher exact test to test for the differences of distribution in morbidity variables by bleeding status * Dr. Agarwal Hemant: CBP study (9/17/10) * Provide descriptive stats * Perform logistic regression on death with adjusting for severity of surgery, bypass time and their interaction * Provide predicted probability graph for bypass time * Provide descriptive table for bypass time by death * Perform proportional adds logistic regression model on ordinal outcomes * Dr. Agarwal Hemant: CBP study 3.5 years (9/17/10) * Provide descriptive stats * Perform proportional adds logistic regression model on ordinal outcome and test for the interaction term to decide the final model * For final model with interaction term, evaluate the effect of the continuous predictor at each level of rachs * Provide table of parameters for the final model * Perform logistic regression model on death * Provide table of parameters and predicted probability plot for bypass time * Dr. Mathew Revi (first plan 8/26/10) * Provide descriptive stats for overall data and subsets * Provide counts and proportions for two successive tests by resutls * Perform logistic regression on outcome "steroids treatment" to test the effect of initial GHST test and ACTH test 1cortisol value * Provide predicted probability graphs for these two continuous variables * Dichotomize the initial GHST test and ACTH test 1 result * Provide the 2*2 table and perform fisher exact test to test for the associat<script type="text/javascript" src="http://biostat.mc.vanderbilt.edu/wiki/pub/System/TinyMCEPlugin/tinymce/jscripts/tiny_mce/themes/advanced/langs/en.js?v=%24Rev:%205860%20%282009-12-28%29%20%24"></script><script type="text/javascript" src="http://biostat.mc.vanderbilt.edu/wiki/pub/System/TinyMCEPlugin/tinymce/jscripts/tiny_mce/plugins/foswikibuttons/langs/en.js?v=%24Rev:%205860%20%282009-12-28%29%20%24"></script><script type="text/javascript" src="http://biostat.mc.vanderbilt.edu/wiki/pub/System/TinyMCEPlugin/tinymce/jscripts/tiny_mce/plugins/foswikiimage/langs/en.js?v=%24Rev:%205860%20%282009-12-28%29%20%24"></script>e between binary predictors and outcome "steroids treatment" * Subset the dataset for whose deficiency status are available * Repeat the logistic regression on outcome deficiency status with initial GHST test and ACTH test 1 cortisol value * Repeat the Fisher exact test on outcome deficiency status * Provide predicted probability plots for initial GHST test and ACTH test 1 cortisol value * Dr.Abramo (pre tapped study) (9/10/10) * Updated the dataset with new files (including 4 new pts) * Rerun the pre analysis on the new dataset * Perform linear regression model on means and standard deviations with the same predictors * Dr.Abramo (pre and post tapped study) (9/10/10) * Updated the dataset with new files (including 4 new pts) * Rerun the pre and post analysis on the new dataset * Perform linear mixed effect model on left and right NIRS readings with tap status variables ( pre, during and post) and other predictors of interest * Split time period after getting tapped into 8 small intervals and refit the model with these interval dummy variables * Dr. Hemant (survey1) * Provide descriptive stats * Merge two dataset * Provide 2*2 tables for each survey questions for both pre and post surveys * Provide summary statistics and perform Wilcoxon test betweeen pre and post surveys * Repeat the tables and analysis for post and handoff surveys * Provide summary statistics for 4 Likert-scale questions * Perfrom Kruskal-Wallis test to compare 5 role groups * Dr.Abramo (post) * Provide 6*6 matrix plots for pre and post-tapped patients * Provide Bland Altman test for left and right measurement agreement * Linear mixed analysis for NIRS pre tap readings in comprehensive model * Dr.Puzanovova (second plan) * Provide ICC table and effective sample size for all the outcome variables on first 3 time points only * Fit a restricted cubic spline model for one outcome with interaction term and adjusted with robcov() * Provide ICC for time points 1 and 2, 2 and 3 * Fit a linear mix model with the same predictors and outcome * Provide ICC table for all the outcome variables on first 3 time points only * Dr. Matthew Revi (08/26/10) * Provide descriptive stats * Label all the variables * Provide descriptive tables by variable and by tests * Clean data by switching rows and columns * Replace Metryrapone test variable in the descriptive tables * Provide descriptive stats for subgroup patients (151, 18, 13 and 34 separately) * Create a new outcome named "Deficiency" * Combine the two steroids treatment use (Maintenance and Stress) * Rerun the two models with two new outcome (continuous and dichotomized) * Provide predicted probability plots for each model * Dr. James Gay (hospitalization dataset) (8/13/10) * Lump together the APR-DRG MDC levels * Replace severity with MDCs in descriptive table and GEE model * Provide predicted probability tables to all categorical variables * Provide graphs to depict predicted probabilities * Dr. Abramo (8/11/10) * Rescale all the matrix plots * Bland Altman test for left and right measurement agreement * Provide comprehensive and individual plots * Dr. Dalabih * Provide a survival analysis graph * Convert the odds ratio in survival table * Provide analysis for time to mild status * Produce graphs for time to mild status * Dr. Abramo * Provide graphs of readings for each patient * Create a matrix of graphs of readings * Provide matrix graphs for 18 patients (128 pre-tap data) * Provide descriptive stats for standard deviations for 128 patients * Dr. Halasa Natasha * Provide 3 descriptive graphs * Dr.Abramo * Data manipulation to create tap marker * Create labels for during-tap * Making descriptive tables for pre and post tap data * Making histograms * ICC analysis * Study differences in distributions * Linear regression model analysis * Dr. James Gay (hospitalization dataset) * Create descriptive tables * Create outcome variables and fitting logistic regression GEE model * ANOVA for GEE * Predictive probability * Dr. Dalabih 7/19/10 * Update all analyses with new data set * Adding race and gender to the model for infusion status * Dr.Granger(waiting for input) 7/15/10 * Produce tables and figures for Etco2 manuscript * Dr. James Gay (patients dataset) 7/12/10 * Provide descriptive statistics on overall data, and by group * Create summary table for age distributions * Create summary table for age distributions by 7 categories * Detect differences in age and gender distributions between Group C and other inpatients * Detect group difference in total number of hospital days and hospitalizations * Dr.Abramo * Check and update stats in the NIR paper * Check and update stats in ETco2 paper * Dr. Erin Alber * Review manuscript stats * Dr.Granger * Create table and figures for ATV manuscript * Look at the correlation of the new variable for Etoco2 project * Dr.Abramo * Clean the dataset and rerun the analysis * Create two types of time variables * Plot NIRS reading across time using both types of variables * Dr. Romano * Provide descriptive stats summary tables for questions with scores, single and multiple choices, respectively * Provide descriptive stats and histograms for survey results * Dr. Dalabih * Rerun the analysis with corrected data * Detect raw drug level effect * Dr.Puzanovova (first plan) * Describe variables of interest * Provide mean tables and plots for 13 primary outcome variables (one plot for each variable) * Provide mean tables and plots for each of 13 primary outcome variables under 4 subgroups (one plot for each variable) * Detect group effect and gender effect on primary outcomes using lme model * Produce interaction plots for variables of interest * Determine correlation among variables of interest with linear mixed model * Compare GEE packages in R * Run GEE model to test for association of 6 variables with outcomes * Dr.Abramo * Combine 130 csv files into one dataset * Rerun the analysis with new dataset * Dr. Dalabih * Describe time to mild status * Detect drug dose effect * Spearman correlation test on q2hs time and time to mild status * Granger * Survival analysis on association between outcomes and Etco2 values * Erin Albers * Rerun analysis without #83 and edit report * UNC.MP * Run descriptive stats * Dr. Dalabih * Run descriptive stats and provide summaries for variables of interest * Fit cox proportional hazards model with either binary or continuous predictor to assess drug influence * Erin Albers * Check normality assumptions * Run repeated measures ANOVA to determine condition effect * Pairewise comparison("Tukey" HSD) * Dr.Abramo * Combine 130 csv files into one dataset * Change variable labels * VCH readmission study * Run descriptive stats * Dr. Abramo * Remove one file from dataset * Burn a data cd * MPV project * Produce descriptive statistics tables * Use summary.formula to produce 2*2 table * Run Fisher exact test for subject with small expected cell count * Run univariate logistic regression on continuous variables * Barkin (12 hr) * Create format for gender, BMI category * Run regressions * Produce graphs * Produce demographic table (age, sex bmi percentile, WC, BIA) * Granger * Get ICC for Delta, forced and unforced data * Include FEV% to the descriptive table * Convert and format outcome variables * Descriptive Statistics on main variables in wide data set (delta's, Asthma score, %FEV, 10 outcome variables) * Make tall data set (1 row per observation) * Detect time effect in deltas * Detect baseline athma score effect * Detect FEV effect in deltas * Produce boxplots for non-forced,forced and deltas at various time points * Abramo * Update Abramo datasets with new csv file * Use Abramo data to get ICC using book formula and lme() * get CI using book formula * compare icc() and ICC.CI() function using both textbook data and Abramo data. * Based on the distribution of readings, calculate ICC and R˛ * Mark Reiderer analysis * Describe scores for each item in questionnaire by residence group * Plot histograms of scores for each question by group * Describe three outcomes by year and group and perform Wilcoxon test * Plot histograms of outcomes by year and group * ATV (Granger) * Investigate predicted means using Mean() function of rms package <p> </p> * ATV (Granger) * Produce vertical boxplots by year for the outcomes of interest, with year (2000,2001,etc.) on the x-axis * Have one figure for each outcome * Outcomes (y-axis): hospital charges, age, ISS, GCS, ICU, Hosp days, * Have another plot for year(x-axis) vs. total number of accidents(y-axis) * Formal test for trend over time * Perform test of counts vs. time * Edit report * Abramo * Import data from Abramo study into one R data file * Create tapped variable * Remove observation at and after "tapped" * Describe left and right column data and plot histograms separately * Create new data set with just left and right means by patients'ID * Barkin * Produce requested table - keep track of hours (12 hr) * Add "Obese group" to previous table (1hr) * Hermant * Produce one table for Death vs. all predictors * Produce separate tables for predictors vs. three outcome variables * Miranda Rainese * Describe data set * Summarize outcomes by predictor * Perform ordinal logistic regression on outcomes vs. the predictor * Edit the report %BLUE% ---++ Dr. Halasa UBS funding * Dr. Najwa khuri * Request to provide descriptive stats for a set of predictor on patients all the cases of RSV infection * Compare the patients to those with other virus individually or collectively * Vaccine clinical trial * Prepare the HAI dataset * Write an analysis plan * Provide descriptive stats by day and by severity * Change the denominator in the barchart as the number of patients in each group * Provide bar charts for local and systemic AE * Provide 4 tables comparing 0-3 and 4-7 by group( 0-3[HD vs. SD], 4-7[HD vs. SD], HD[0-3 vs 4-6], SD[0-3 vs. 4-6] * Provide spheghathi plot for temperature across day * Perfom chi-square test for all types of AEs * SOT * Prepare the HAI dataset * Perfom chi-square test for all types of AEs * PIV groups comparisons * Provide descriptive stats by PIV status group * rerun the report after fixing the errors <!-- ---+++ Page Preferences -->
E
dit
|
A
ttach
|
P
rint version
|
H
istory
: r346
<
r345
<
r344
<
r343
|
B
acklinks
|
V
iew topic
|
Edit
w
iki text
|
M
ore topic actions
Topic revision: r346 - 19 Aug 2014,
MengXu
Main
Department Home Page
Biostatistics Graduate Program
Vanderbilt University Medical Center
Main Web
Main Web Home
Search
Recent Changes
Changes
Topic list
Biostatistics Webs
Archive
Main
Sandbox
System
Register
|
Log In
Copyright © 2013-2022 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Vanderbilt Biostatistics Wiki?
Send feedback