Absolute risk:

Summary
This submission of SAS code demonstrates a macro for finding the absolute risk assuming there is only one competing risk. This method uses odds ratios from logistic regression to estimate hazards in order to find absolute risk. This macro uses the simple exponential model, which assumes that every hazard has a constant value in the time interval of interest.

Author(s)
Murphy, TE
Yale University

Code Language(s)
SAS

Keywords
absolute risk ,aging ,death ,disability ,

Analyzing Dyad Data with Additional Clustering:

Summary
This project provides a brief introduction and overview of approaches to dyadic analysis with references, and provides two approaches to performing dyadic analysis when dyads are nested in a higher level of clustering, such as dyads participating in cluster-randomized clinical trials. Mplus syntax is provided for estimating treatment effects on dyadic outcomes when dyads are cluster-randomized using Multilevel Structural Equation Modeling, and 3-Level Multilevel Modeling. The syntax may be used as an example and modified to fit the research question and data for a given study.

Author(s)
Katie Newkirk, PhD
Yale University

Collaborator(s)
Joan Monin, PhD; Heather allore PhD; Holly Laws, PhD

Code Language(s)
MPLUS

Keywords
Cluster Randomized Trials; Dyadic Analysis; Multilevel Structural Equation Model (MSEM); Multilevl Model

Link
https://doi.org/10.1017/S1041610222000898

Acknowledgement
This work is supported by the National Institute of Aging (NIA) of the National Institutes of Health (NIH) under Award Number U54AG063546, which funds NIA Imbedded Pragmatic Alzheimer’s Disease and AD-Related Dementias Clinical Trials Collaboratory (NIA IMPACT Collaboratory). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

Anticholinergic Medications Risk Assessment:

Summary
The purposes of the submission are two fold: 1) To provide a free-access tool of more than 300 medications with assigned anticholinergic scores for use by interested clinical, epidemiological and pharmacological researchers. 2) To describe the novel methodology and instruments that can be readily applied/adapted to new medications that are not rated for CR- ACHS by us, but are of interest to outside researchers.

Author(s)
Han, L
Yale University

Keywords
anticholinergic risk score ,risk assessment method ,

Assessing Probabilities of Harms and Benefits of Multiple Medication Use:

Summary
We describe a simple technique to assist patients and doctors in how to systematically add candidate medications to their existing regimen. It is based on a simple multiplicative combination of utility values for each medication. These medication-specific utility values are in turn based on a simple average of the probabilities of benefits and adverse effects as recorded in the clinical literature. This GRASP submission consists simply of this summary document which cites the related clinical reference article.

Author(s)
Murphy, TE
Yale University

Collaborator(s)
Agostini JV, Van Ness PH, Tinetti ME, Peduzzi PN, Allore HG

Keywords
adverse effects ,geometric mean ,multiple medications ,tradeoffs ,utility value ,weighting ,

Link
http://jah.sagepub.com/cgi/reprint/20/6/694

Acknowledgement
This study was supported by the Yale Claude D. Pepper Older Americans Independence Center (P30AG021342) from the National Institute on Aging. Dr. Agostini was supported by a Veterans Affairs Health Services Research Career Development Award. The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs.

Citations
Murphy TE, Agostini JV, Van Ness PH, Peduzzi P, Tinetti ME, Allore HG (2008). Assessing the probabilities of harms and benefits of multiple medication use. Journal of Aging and Health, 20(6); 694 - 709.

Calculation of Intraclass Correlation: continuous:

Summary
This submission provides SAS code for calculating intraclass correlation coefficients (ICC) and corresponding confidence intervals. The data, program, and output serve for demonstrational purposes only.

Author(s)
Van Ness, PH
Yale University

Code Language(s)
SAS

Keywords
Clustered Data ,Instrument Selection ,Intraclass Correlation Coefficient Confidence Intervals ,mixed models ,Reliability Testing ,SAS ,

Link
http://jah.sagepub.com/cgi/reprint/20/2/183

Acknowledgement
This study was supported by the Yale Claude D. Pepper Older Americans Independence Center (P30AG021342) from the National Institute on Aging.

Citations
Van Ness PH, Towle VR, Juthani-Mehta M (2008). Testing measurement reliability in older populations: Methods for informed discrimination in instrument selection and application. Journal of Aging and Health; 20:183-197.

Calculation of Intracluster Correlation: binary:

Summary
This submission provides SAS code for creating an intraclass (or intracluster) correlation coefficient (ICC) for a binary outcome.

Author(s)
Van Ness, PH
Yale University

Code Language(s)
SAS 9.2

Keywords
Clustered Data ,Dependent Observations ,Recruiting ,SAS 9.2 ,

Acknowledgement
This work was supported in part by the Claude D. Pepper Older Americans Independence Center at Yale University School of Medicine

Calculation of Probabilistic Index:

Summary
This project provides SAS code to calculate the probabilistic index as proposed by Laura Acion et al. It consists of a SAS program that provides code for calculating the probabilistic index from the Wilcoxon Rank Sum Statistic or from the Somers' D measure of association.

Author(s)
Van Ness, PH
Yale University

Code Language(s)
SAS

Keywords
C statistic ,clinically significant effect size ,nonparametric ,SAS 9.2 ,Somers' D ,Wilcoxon Rank Sum ,

Link
http://www3.interscience.wiley.com/cgi-bin/fulltext/111089436/PDFSTART

Acknowledgement
This study was supported by the Yale Claude D. Pepper Older Americans Independence Center (P30AG021342) from the National Institute on Aging.

Citations
Acion L, Peterson JJ, Temple S, and Arndt S (2006). Probabilistic index: an intuitive, non-parametric approach to measuring the size of treatment effects. Statistics in Medicine; 25:591-602.

Covariance Structure Between Ligand-Specific TLR Cytokine Production:

Summary
This document ties together the five different files that makeup the GRASP example drawn from analysis of TLR-production of cytokines that is stimulated specifically from each of six different ligands per participant at one point in time, i.e., cross-sectional. Because the six ligands are clustered within participants, their associated outcome measurements are correlated. Specifically, we use an unstructured covariance structure that allows each participant to have his or her own specific correlation structure for their set of six ligands, thus accounting for the inherent variation of each participant. Data has been sanitized by removing identifying information and adding random noise.

Author(s)
Allore, HG
Yale University

Code Language(s)
SAS

Keywords
aging ,correlation structure ,cross-sectional ,cytokines ,immunology ,ligands ,mixed models ,multivariable ,SAS ,toll-like receptor (TLR) ,

Link
http://www.jimmunol.org/cgi/reprint/178/2/970

Acknowledgement
This work was supported by the Center of Excellence in Aging at Yale University, funded by the John A. Hartford Foundation, the Claude D. Pepper Older Americans Independence Center at Yale University (P30 AG021342) (to A.C.S. and D.V.D.), Yale School of Medicine Dean’s Pilot Project Grant in Translational Research (to A.C.S.), National Institute of Allergy and Infectious Diseases Grants NO1 AI 50031 (to E.F., R.R.M., and A.C.S.) and AI 053279 (to E.F.). D.V.D. is a Clinical Research Scholar in the Investigative Medicine Program, Yale University School of Medicine, supported by the Yale Mentored Clinical Research Scholars Program (National Institutes of Health Grant NCRR K12RR17594). R.M. is an Investigator of the Howard Hughes Medical Institute.

Citations
van Duin D, Mohanty S, Thomas V, Ginter S, Montgomery RR, Fikrig E, Allore HG, Medzhitov R, Shaw AC, (2007). Age-Associated Defect in Human TLR-1/2 Function. J Immunol, 178(2):970-5. PMID: 17202359 van Duin, Allore HG, Mohanty S, Ginter S, Newman FK, Belshe RB, Medzhitov R, Shaw AC Prevaccine determination of the expression of costimulatory b7 molecules in activated monocytes predicts influenza vaccine responses in young and older adults. J. Infectious Disease. 2007. 195(11):1590-7. PMID: 17471428 Panda A, Qian F, Mohanty S, van Duin D, Newman FK, Zhang L, Chen S, Towle V, Belshe RB, Fikrig E, Allore HG, Montgomery RR, Shaw AC. Age-associated Decrease in Toll-like Receptor Function in Primary Human Dendritic Cells Predicts Influenza Vaccine Response. J Immunology 2010. 184(5):2518-27. PMID: 20100933 Dunne, DW, Shaw AC, Bockenstedt LK, Allore H, Chen S, Malawista SE, Leng LL, Mizue Y, Piecychna M, Zhang L, Towle V, Bucala R, Montgomery RR, Fikrig E. Increased TLR4 expression and Downstream Cytokine Production in Immunosuppressed Adults compared to Non-immunosuppressed Adults. 2010. PLoS ONE. 5(6):e11343 PMID: 20596538. PMCID: PMC2893205

CRCoder-A SAS code helper for marginal and personalized concomitant risk modeling:

Summary
The SAS™ code helper for marginal and personalized concurrent risk modeling

Author(s)
Allore H, Charpentier P
Yale University

Collaborator(s)
Allore H, Charpentier P, McAvay G, Murphy T and Agogo G

Code Language(s)
SAS, PHP, MySQL

Keywords
Joint model, risk, longitudinal

Link
http://crcoder.phs.wakehealth.edu/

Acknowledgement
This work was supported by grants from the U.S. National Institute on Aging at the National Institutes of Health (R01 AG047891, P30AG021342-16S1, R33AG045050, U24AG059624).

Citations
McAvay GM, Murphy TE, Agogo GO, Allore HG. CRcoder: An Interactive Web Application and SAS Macro for Decision Support: Generation of Typical and Personalized Concurrent Risk. Permanente Journal Perm J. 2020;24. doi: 10.7812/TPP/19.078. Epub 2019 Dec 18.
https://doi.org/10.7812/TPP/19.078

Murphy TE, McAvay GJ, Agogo GO, Allore HG. Personalized and Typical Concurrent Risk of Limitations in Social Activity and Mobility in Older Persons with Multiple Chronic Conditions and Polypharmacy. Ann Epidemiol. 2019;S1047-2797(19)30222-4. doi: 10.1016/j.annepidem.2019.08.001. PMID: 31473124

Agogo GO, Murphy TE, McAvay GM, Allore HG. Joint modeling of concurrent binary outcomes for longitudinal observational studies using inverse probability of treatment weighting improved treatment effect estimation. Annals of Epidemiology 2019; pii: S1047-2797(18)30939-6. doi: 10.1016/j.annepidem.2019.04.008. PMID:31085069

Evaluation and Review of Imputation Methods for Multivariate Longitudinal data with Mixed-type Incomplete Variables:

Summary
This document ties together the three different files that make up the GRASP example for performing different multiple imputation methods to impute data with mixed types of incomplete longitudinal variables. The sample data used in this simulation study was extracted from the National Health and Aging Trend Study (NHATS), which includes four waves of observations on 5309 adults aged 65 years or older. The simulated datasets contain four time-varying incomplete variables: one binary variable, one continuous variable, and two count variables. In the simulation, we assume that the missing data mechanism is missing at random for the monotone missing data pattern and not missing at random for the intermittent missing data pattern. To simulate a monotone missing pattern, we generate the drop-out indicators for each subject, and the proportion of participants who start to drop out from the study at rounds 2, 3, and 4 are set to 20%, 15%, and 10%. To simulate an intermittent missing data pattern, we generate the missing indicators for each of the incomplete variables, and the average missing proportions of each of the incomplete variables at rounds 2, 3, and 4 are 20%, 35%, and 40%, respectively. For every simulated dataset, we conduct a multiple imputation procedure with 5 imputations using 10 imputation methods with different choices of imputation models. The imputation models include uni- or multivariate single- or multilevel generalized linear regressions, depending on implementation strategy (fully conditionals specification vs. joint modeling) and data format (wide vs. long). After imputation, we conduct three analyses: univariate generalized hierarchical model, latent growth model, and bi-variate generalized hierarchical model. We compare the estimates obtained from the multiple imputation procedure with those obtained from the original data with complete cases only. We summarize five metrics: (1) the relative bias, (2) the root of mean squared error (RMSE), (3) the 95% interval estimate width, (4) whether the interval estimate covers the estimate with complete data, and (5) the fraction of missing information (FMI) that measures the uncertainty in the imputed values for missing elements. Additionally, we record the computational time for a single imputation of each imputation model.

Author(s)
Cao Y, Allore HG, Vander Wyk B, & Gutman R
Brown University, Yale University

Code Language(s)
R

Keywords
Longitudinal study, aging study, Multivariate missing data, multiple imputation, simulation study

Citations
Cao Y, Allore HG, Vander Wyk B, & Gutman R. Evaluation and Review of Imputation Methods for Multivariate Longitudinal data with Mixed-type Incomplete Variables. Under Review.

Exploratory Meta Analysis on Uniform Assessment Battery across Aging Studies:

Summary
This summary file provides an overview of the exploratory meta analysis across different aging studies. Five components files plus this summary will be needed to run the SAS MACRO and the MATLAB function properly. The programs were originally written to analyze a set of "common assessment battery" of physical performance and biomarkers within Wake Forest University Health Sciences Pepper Center. The sample data used in this package is simulated data and for demonstration only. The relevant published work is referred if applicable. Using the SAS output, bubble plots with bubble size proportional to sample size can be produced in MATLAB (unfortunately with SAS 9.1 we could not figure out an efficient way to plot the same). Therefore MATLAB were used. You may also use any other favorite graphic software to make the plots.

Author(s)
Leng, X
Wake Forest University

Code Language(s)
SAS and Matlab

Keywords
aging ,bubble plot ,covariate adjustment ,cross-sectional ,functional disability ,MATLAB ,meta-analysis ,multiple linear regression ,physical performance ,report table ,SAS ,

Link
http://biomedgerontology.oxfordjournals.org/cgi/content/full/gln038

Longitudinal Extension of the Average Attributable Fraction (LE-AAF):

Summary
This is a SAS matrix language program to calculate the Longitudinal Extension of the Average Attributable Fraction (LE-AAF) measure. Detailed instructions, references and code are included in the documentation. Briefly, the LE-AAF is based on the concept of the average AF (AAF). The LE-AAF for a dichotomous risk factor estimates the marginal percentage contribution to the timed occurrence of an outcome that can be attributed to that risk factor and is symmetric, meaning the estimates are independent of the order in which the conditions occur. See documentation for further information and references.

Author(s)
Zhan Y, McAvay G
Yale University

Collaborator(s)
Allore HG, Lin H, Murphy TE.

Code Language(s)
SAS IML

Keywords
Attributable Fraction, Multimorbidity, Longitudinal

Acknowledgement
This work was supported by the National Institute on Aging at the National Institutes of Health (RF1AG058545 and R01AG055681 to ARQ, R01AG047891 to HGA who contributed from the Yale Claude D. Pepper Older Americans Independence Center P30AG021342, R33AG045050, P30AG066508).

Citations
1. Lin, H, Allore HG, McAvay G, Tinetti ME, Gill TM, Gross CP, Murphy TE.A Method for Partitioning the Attributable Fraction of Multiple Time-Dependent Coexisting Risk Factors for an Adverse Health Outcome. AJPH, 2012; 103(1):177-182. 2. Murphy T E, McAvay G, Carriero NJ, Gross CP,Tinetti ME, Allore HG, Lin H. Deaths observed in Medicare beneficiaries: average attributable fraction and its longitudinal extension for many diseases The Journals of Gerontology: Series A, Volume 71, Issue 8, August 2016, Pages 1113–1116. 3. Allore HG, Zhan Y, Cohen AB, Tinetti ME, Trantalange M, McAvay G. Methodology to Estimate the Longitudinal Average Attributable Fraction of Guideline-recommended Medications for Death in Older Adults With Multiple Chronic Conditions. Journals of Gerontology: J Gerontol A Biol Sci Med Sci, 2016, Vol. 00, No. 00, 1–5.

Multiple Imputation Methods for Longitudinal Studies:

Summary
Introduction: This document ties together the three different files that make up the GRASP example for performing different multiple imputation methods on the NHATS data. In this simulation study, we generated incomplete datasets following the two steps: (1) modeling the drop-out process using the original data under missing at random assumption; (2) predicting the drop-out status on data with complete cases only and removing observed values of individuals who are predicted to be drop-out at a certain time. We used two imputation strategies: fully conditional specification(FCS) and joint modeling (JM) for multiple imputation, and specified single-level or multilevel generalized linear regressions as imputation models depending on the data format and incomplete variables’ type. After imputation, we performed a statistical analysis using the multiple imputed data and compared the results with the estimates obtained from the complete dataset. We summarized four metrics: (1) the relative bias (qˆmeth??qˆ comp)=qˆ comp, where qˆmeth is the estimate obtained after multiple imputation using one of the imputation methods; (2) the root of mean squared error (RMSE); (3) the average 95% interval estimate width; (4) the empirical coverage of 95% interval estimates. The sample data used in this simulation study was extracted from the National Health and Aging Trend Study (NHATS), which included four waves of observations on 5309 adults aged 65 years or older.

Author(s)
Cao Y, Allore HG, Vanderwyk B, Gutman R.
Brown University and Yale University

Collaborator(s)
Cao Y, Allore HG, Vanderwyk B, Gutman R.

Code Language(s)
R

Keywords
Longitudinal study; Missing Data; Multiple Imputation; Simulation

Citations
R Core Team (2020). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Cao Y, Allore HG, Vander Wyk B, & Gutman R. Evaluation and Review of Imputation Methods for Multivariate Longitudinal data with Mixed-type Incomplete Variables. Unpublished manuscript.

Multiple Imputation Simulation:

Summary
This is a GRASP example for simulating missing data. In this simulation study, we made up pseudo ‘missing’ values by removing observed values from original datasets. The pseudo missing data were then imputed by two multiple imputation methods: sequential method vs. simultaneous method. We compared results between sequential and simultaneous methods using four criteria.

Author(s)
Ning, Y
Yale University

Keywords
Cardiovascular Health Study ,Longitudinal study ,missing data ,multiple imputation ,SAS ,simulation study ,

Multiple Testing for Correlated Outcomes in Clinical Trials:

Summary
When multiple outcomes are obtained from clinical trials, the most common approach in the medical literature is to analyze each outcome separately, presenting multiple p-values and an overall subject conclusion. However, it is well known that the excessive use of multiple significance tests can substantially increase the chance of false positive findings (type I error). On possible solution is to control a predefined error rate such as Family Error Rate (FER) in the strong sense, i.e., the probability of rejecting falsely at least one true individual null hypothesis, irrespective of which and how many of the individual null hypotheses are in fact true. Bonferroni method controls FER in the strong sense and is easy to apply. However when outcomes are correlated, Bonferroni method appears to be too conservative. We consider k normally distributed outcomes, each with known variance, for which all possible pairs have known correlation Ï쳌ij within each of two treatment groups. Then for given α , k, and Ï쳌ij we can derive numerically that the “nominalâ€쳌 value α* for individual tests which controls for FER at α under the null hypothesis. The MATLAB function multicompare calculates α* and the corresponding critical values as outputs. These can easily be extended to g (g >2) treatment groups and analysis of covariance for g (g ≥ 2) groups.

Author(s)
Leng, X
Wake Forest University

Code Language(s)
MATLAB

Keywords
aging ,Bonferroni ,correlated outcomes ,cross-sectional ,functional diability ,MATLAB ,multiple testing ,physical performance ,type I error ,

NLMIXED SAS Macro:

Summary
This submission provides SAS code for fitting a longitudinal logistic regression model in which a random intercept is included to induce a compound symmetry covariance structure for repeated measures on individual subjects.

Author(s)
Van Ness, PH
Yale University

Collaborator(s)
O'Leary JA Fried T Towle VR et al

Code Language(s)
SAS

Keywords
longitudinal models ,mixed models ,repeated measures ,SAS ,

Acknowledgement
This work was supported in part by grants from the Biostatistics Core of the Claude D. Pepper Older Americans Independence Center at Yale University School of Medicine (#2P30AG021342-06).

Citations
Van Ness, PH, J. O'Leary, et al. (2004). "Fitting longitudinal mixed effect logistic regression models with the NLMIXED procedure." Proceedings of the 29th Annual SAS® Users Group International Conference (SUGI 29) 29: 1-6 (www2.sas.com/proceedings/sugi29/190-29.pdf). Fried T, Byers, AL, Gallo, WT, Van Ness, PH, Towle, VR, O'Leary, JR, Dubin, JA. Prospective study of health status preferences and changes in preferences over time in older adults. Archives of Internal Medicine. 2006;166:890-895.

Sample size and power calculations for designing stepped wedge cluster randomized trials with subclusters:

Summary
The stepped wedge cluster randomized trial (SW-CRT) is an increasingly popular design for evaluating health service delivery or policy interventions. An essential consideration of this design is the need to account for both within-period and between-period correlations in sample size calculations. Especially when embedded in health care delivery systems, many SW-CRTs may have subclusters nested in clusters, within which outcomes are collected longitudinally. However, existing sample size methods that account for between-period correlations have not allowed for multiple levels of clustering. We present computationally efficient sample size procedures that properly differentiate within-period and between-period intracluster correlation coefficients in SW-CRTs in the presence of subclusters. We introduce an extended block exchangeable correlation matrix to characterize the complex dependencies of outcomes within clusters. For Gaussian outcomes, we derive a closed-form sample size expression that depends on the correlation structure only through two eigenvalues of the extended block exchangeable correlation structure. For non-Gaussian outcomes, we present a generic sample size algorithm based on linearization and elucidate simplifications under canonical link functions. For example, we show that the approximate sample size formula under a logistic linear mixed model depends on three eigenvalues of the extended block exchangeable correlation matrix. We provide an extension to accommodate unequal cluster sizes and validate the proposed methods via simulations. Finally, we illustrate our methods in two real SW-CRTs with subclusters.

Author(s)
Plourde K, Taljaard M, Li F
Yale University, Ottawa Hospital Research Institute, Yale University

Code Language(s)
R

Keywords
cluster randomized trial, eigenvalues, extended block exchangeable correlation structure, generalized linear mixed models, power analysis

Link
https://doi.org/10.1002/sim.9632

Acknowledgement
This work is supported by the National Institute of Aging (NIA) of the National Institutes of Health (NIH) under Award Number U54AG063546, which funds NIA Imbedded Pragmatic Alzheimer’s Disease and AD-Related Dementias Clinical Trials Collaboratory (NIA IMPACT Collaboratory). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

SAS Macro for Bivariate Analyses:

Summary
This submission consists of the four component files that makeup the GRASP example demonstrating a SAS-based macro that automatically calculates and tabulates the bivariate associations between continuous or categorical covariates and a dichotomous outcome. This example provides de-identified and sanitized data consisting of outcomes and 6 covariates from 50 pseudo subjects and serves for demonstrational purposes only.

Author(s)
Murphy, TE
Yale University

Code Language(s)
SAS

Keywords
admission characteristics ,bivariate analyses ,Confusion Assessment Method (CAM) ,delirium ,ICU ,logistic regression ,SAS ,

Link
http://archinte.ama-assn.org/cgi/content/full/167/15/1629

Acknowledgement
This work was supported in part by the American Lung Association and Connecticut Thoracic Society (ID CG-002-N), Claude D. Pepper Older Americans Independence Center at Yale University School of Medicine (P30AG21342), and the Franklin T. Williams Geriatric Development Initiative through The CHEST Foundation, Association of Subspecialty Professors, Hartford Foundation. Dr Pisani is a recipient of a National Institutes of Health K23 Mentored Career Development Award (K23 AG 23023-01A1). Dr Inouye is supported in part by grants R21AG025193 and K24AG000949 from the National Institute on Aging. Dr Inouye holds the Milton and Shirley F. Levy Family Chair at Harvard University.

Citations
Pisani MA, Murphy TE, Van Ness PH, Araujo KLB, Inouye SK (2007). Admission characteristics associated with delirium in older patients in a medical intensive care unit. Archives of Internal Medicine; 167(15); 1629-1634.

SAS MIXCORR Macro for data with repeated measures and clumping at zero:

Summary
A mixed-effects, mixed-distribution model for data with repeated measures and clumping at zero.

Author(s)
Tooze, JA
Wake Forest University

Collaborator(s)
Light, LS

Code Language(s)
SAS

Keywords
clumping at zero ,excess zero ,food intake ,mixed distribution ,physical activity ,repeated measures ,SAS ,walking speed ,

Simple Bootstrap with Proc Surveyselect:

Summary
This submission ties together the four component files that demonstrate a simple SAS-based method of generating bootstrap datasets using Proc Surveyselect. In the past we would write a page or two of code to generate these datasets which is done here in a few lines by Proc Surveyselect. This example provides de-identified and sanitized data consisting of gender and a single outcome from 100 pseudo subjects and serves for demonstrational purposes only.

Author(s)
Murphy, TE
Yale University

Code Language(s)
SAS

Keywords
bootstrap ,confidence interval ,median ,

Acknowledgement
This work was supported in part by the Claude D. Pepper Older Americans Independence Center at Yale University School of Medicine (P30AG021342), Dr. Pisani is a recipient of a NIH K23 Mentored Career Development Award (K23 AG 23023-01A1) and the Chest Foundation and Boehringer Ingelheim Pharmaceuticals, Inc. Clinical Research Award in Women’s Pulmonary Health.

Citations
Akgun K, Murphy TE, Araujo KLB, Van Ness PH, Pisani MA. Equivalent treatment of older medical intensive care unit patients based on gender. ''Under Review''.

SpatioTemporal Modeling in Longitudinal Analyses of Areal Units:

Summary
This document ties together six different files that demonstrate spatiotemporal modeling in the longitudinal analyses of areal units. Clinically, this is the statistical analysis underlying the Connecticut Collaboration for Fall Prevention (CCFP) at the Yale Pepper Center. This example takes a sample of 40 zip code tabulation areas (ZCTAs) from the total of 111 used in the formal, published analysis and serves for demonstrational purposes only. Data has been sanitized by removing identifying information and adding random noise.

Author(s)
Murphy, TE
Yale University

Collaborator(s)
Tinetti ME, Allore HG

Code Language(s)
WinBUGS

Keywords
Bayesian methods ,CMS data ,ecological ,falls ,hierarchical models ,longitudinal models ,non-randomized design ,spatio-temporal correlation ,T2 ,translational research ,WinBUGS ,

Link
http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B7P72-4R0CKRG-2&_user=4...

Acknowledgement
The Connecticut Collaboration for Fall Prevention is supported by the Donoghue Medical Research Foundation (DF#00-206) and the Claude D. Pepper OAIC at Yale University School of Medicine (#P30AG21342).

Citations
1) Murphy TE, Tinetti ME, and Allore HG (2008). Hierarchical models to evaluate translational research: Connecticut collaboration for fall prevention. Contemporary Clinical Trials, 29:343-350. 2) Tinetti, M. E., Baker, D. I., King, M., Gottschalk, M., Murphy, T. E., Acampora, D., et al (2008). Effect of dissemination of evidence in reducing injuries from falls. New England Journal of Medicine, 359(3), 252-261.

Statistical Power Calculator:

Summary
This submission provides a link to statistical power calculator (SPC) provided by Department of Biostatistics and Data Science within the Division of Public Health Sciences at Wake Forest University School of Medicine. The SPC is based on SAS PROC POWER. Currently available calculators include: correlation, one sample proportion, one sample mean, paired proportion, paired means, two sample proportion and two sample means. There is also a guide for choosing which calculator to use.

Author(s)
Hepler J, Lovato J
Wake Forest University

Collaborator(s)
Chen H, Lang W, Babcock D, Robertson J, Leng X

Code Language(s)
SAS, ColdFusion, SQL

Keywords
Correlation, Cross-Sectional, Mean, Power, Proportion, Sample Size, Paired, Calculation, Computational, Alpha, Variance

Link
http://power.phs.wakehealth.edu/

Acknowledgement
This calculator application is a labor of love supported by Department of Biostatistics and Data Science within the Division of Public Health Sciences at Wake Forest School of Medicine. Special thanks to everyone who has helped over the years.

Temporal Configuration Analysis:

Summary
Temporal Configuration Analysis (TCA) is a stand-alone application that analyzes the multivariate discrete longitudinal data in gerontology studies, such as ADL, physical activities (FAST23) and other social science studies. The method is based on the Hidden Markov Model (HMM) backbone to separate subjects into hidden classes whose interpretations depend on the estimated conditional probabilities. TCA expands the HMM into a more general Bayesian Network (BN) framework and also incorporates the generalize linear regression models into the BN framework to study covariates' impact on the BN parameters. For example, in an aging study, we like to see how BMI affects subjects' physical performance in transitioning from worse to better states.

Author(s)
Zhang, Q and Ip, E
Wake Forest University

Code Language(s)
MATLAB

Keywords
activities of daily living (ADLs) ,Bayesian methods ,generalized linear models ,intervention studies ,longitudinal models ,MATLAB ,mixed models ,physical activity ,physical performance ,repeated measures ,walking speed ,

Temporally Successive Multiple Imputation of Longitudinal Data:

Summary
This document ties together the five component files that makeup the GRASP example demonstrating the technique of temporally successive multiple imputation as applied to data from the Precipitating Events Project (PEP) at the Yale Pepper Center. This example takes a sample of 100 pseudo subjects with 8 covariates measured over six waves of follow-up. The procedure was used to impute missing observations for risk factors measured at face-to-face interviews performed at 18-month intervals. This example serves for demonstrational purposes only.

Author(s)
Han, L
Yale University

Code Language(s)
SAS

Keywords
activities of daily living ,longitudinal models ,missing at random (MAR) ,missing data ,multiple imputation ,

Link
http://www3.interscience.wiley.com/cgi-bin/fulltext/117995771/PDFSTART

Acknowledgement
The work for this report was funded by grants from the National Institute on Aging (R01AG022993, R37AG17560). The study was conducted at the Yale Claude D. Pepper Older Americans Independence Center (P30AG21342). Dr. Gill is the recipient of a Midcareer Investigator Award in Patient-Oriented Research (K24AG021507) from the National Institute on Aging.

Citations
Gill TM, Han L, Allore HG. Predisposing factors and precipitants for bathing disability in older persons. Journal of the American Geriatrics Society. 2007;55:534-40.