Jason Roy

Where People Die: A Multilevel Approach to Understanding Influences on Site of Death in America

Despite documented preferences for home death, the majority of deaths from terminal illness occur in hospital. To better understand variation in place of death, we conducted a systematic literature review and a multilevel analysis in which we linked death certificates with county and state data. The results of both components revealed that opportunities for home death are disproportionately found in certain groups of Americans; more specifically, those who are White, have greater access to resources and social support, and die of cancer. From the multilevel analysis, the higher the proportion minority and the lower the level of educational attainment, the higher the probability of hospital death while investment in institutional long-term care, measured by regional density of nursing home beds and state Medicaid payment rate, was associated with higher probability of nursing home death. These results reinforce the importance of both social and structural characteristics in shaping the end-of-life experience.

Daily Pain That Was Excruciating at Some Time in the Previous Week: Prevalence, Characteristics, and Outcomes in Nursing Home Residents

Objectives: To examine the prevalence, correlates, and consequences of nursing home (NH) staff reports of “excruciating” level of pain at some time in the previous week in persons with daily pain reported on the Minimum Data Set (MDS).

Bereaved Family Member Perceptions of Quality of End‐of‐Life Care in U.S. Regions with High and Low Usage of Intensive Care Unit Care

Objectives: To compare the quality of end‐of‐life care of persons dying in regions of differing practice intensity.

Government Expenditures at the End of Life for Short- and Long-Stay Nursing Home Residents: Differences by Hospice Enrollment Status

Objectives: To examine end‐of‐life government expenditures for short‐ and long‐stay Medicare‐ and Medicaid‐eligible (dual‐eligible) nursing home (NH) hospice and nonhospice residents.

Impact of cognitive function on assessments of nursing home residents' pain.

Objectives:We sought to examine the impact of residents’ cognitive function on the quality of Minimum Data Set (MDS) pain data using the latent variable approach. Research Design:Using the Resident Assessment Instrument (RAI) protocol, nursing home (NH) staff and well-trained study nurses independently assessed 3736 NH residents. Measures:Inter-rater agreement of pain ratings between NH staff and study nurses was quantified by weighted kappas and polychoric correlations and compared among groups of residents with no/mild, moderate, and severe cognitive impairment. Probit models were built to examine the effect of residents’ cognitive function on thresholds raters used to rate pain. Results:Of 3736 residents, 40.4% had no/mild, 35.9% moderate, and 23.7% severe cognitive impairment. Both NH staff and study nurses recorded less frequent and less severe pain for residents with more severe cognitive impairment. The inter-rater agreement on pain ratings between NH staff and study nurses was good—weighted kappas were greater than 0.5 and polychoric correlations greater than 0.7. The thresholds raters used to record pain were similar for NH staff and study nurses and progressively increased when raters recorded pain for residents with more severe cognitive impairment. Conclusions:Given the RAI protocols, the quality of MDS pain data collected by NH staff was similar to that of well-trained nurses regardless of residents’ cognitive function. Our results strongly support the notion that specialized pain assessment instruments are needed to adequately detect pain for the large proportion of cognitive impaired NH residents.

Risk Factors for Umbilical Venous Catheter-Associated Thrombosis in Very Low Birth Weight Infants

Thrombosis in neonates is a rare but serious occurrence, usually associated with central catheterization. The objective of this study was to investigate the risk factors associated with catheter related thrombosis in very low birth weight (VLBW) infants.

Resident, nursing home, and state factors affecting the reliability of Minimum Data Set quality measures.

Nursing home quality measures impact policy decisions such as reimbursement or consumer choice. Quality indicators in the United States are collected through the federally mandated Minimum Data Set (MDS). Bias in MDS data collection or coding can thus have a negative impact on policy applications. To understand whether bias was present in coding, the authors studied 5174 pairs of MDS assessments that were independently collected by nursing home staff and study nurses from 206 nursing homes. The authors developed multivariate multilevel models to identify nursing home and resident characteristics that were significantly associated with the data quality of multiple MDS measures of nursing home quality. The outcomes were coding differences between nursing home staff and study nurses. Resident characteristics explained little of the variation in coding differences among facilities, while facilities characteristics explained 4% to 20% of the variation and state location further explained 13% to 34% of the variation. A generalized effect of nursing home state location tended to be consistent across measures. States that overidentified problems also tended to have worse quality indicators and vice versa. Comparisons of MDS-based quality indicators reflect differences in assessment practices at least as much as true quality differences. Efforts to standardize assessment practices across states are needed.

Latent class models and their application to missing-data patterns in longitudinal studies.

Latent class models have been developed as a flexible way of modeling the correlation of multivariate data, as a method for discovering subpopulations with similar response profiles and as a dimension reduction tool. In this manuscript, we provide a review of some of this literature and describe specific developments in several statistical and substantive areas. We then describe latent class models that could be used for characterizing missing-data patterns in longitudinal studies with regularly spaced observation times, where there is a large amount of intermittent missing data. We illustrate by analyzing data from a longitudinal study of depression, where there were 379 unique missing-data patterns.

SAS for Mixed Models, Second Edition. R. C.Littell, G. A. Milliken, W. W. Stroup, R. D. Wolfinger, and O. Schabenberger

This book was designed to demonstrate, with applications, how to use the SAS System to analyze data using many different types of mixed models. To fully benefit from the book, readers should have familiarity with SAS and some knowledge of generalized linear models and mixed-effects models. The second edition of this book comes some ten years after publication of the first edition of this popular book. While there are updates throughout the book (and some reorganization), the primary new material includes diagnostics for mixed models, power calculations, and Bayesian approaches to mixed modeling in SAS. The second edition also includes examples using PROC NLMIXED and PROC GLIMMIX, which were not part of the available SAS procedures when the first edition was written. The second edition also demonstrates the use of the output delivery system (ODS). Finally, new, interesting examples were added to their already long list of case studies. Chapter 1 provides an introduction to statistical modeling, as well as to the notation and terminology that will be used throughout the book. Models described in the chapter include analysis of variance (ANOVA), linear regression and random effects models. The chapter ends with a flow chart that shows which SAS procedures can be used for a wide variety of scenarios (normal or non-normal data; independent or dependent errors; fixed or random effects; linear or nonlinear model). Chapters 2–9 and 11 describe a wide variety of models and illustrate how to use SAS to fit each type of model. The topics covered in these chapters include split-plot designs, random effects models, multifactor treatment designs, repeated measures data, analysis of covariance, heterogeneous variance models and models with spatial variability. Each type of design/model is illustrated with an example. The authors provide the SAS code and interpret each line of output. In cases where a model can be fitted using more than one SAS procedure, they provide the code for each and compare the results. In some situations one procedure is clearly better than the others. For example, in Chapter 4, they illustrate why PROC GLM should not be used for split-plot experiments. For a variety of problems, the “lsmeans” and “estimate” statements are used within the corresponding procedure to produce the specific estimates and tests that are of interest. These chapters cover both very simple and basic problems (e.g., one-way random effects treatment structure in section 3.2) and much more complicated problems (e.g., nested model with unequal random effect variances in section 9.4). Chapter 10 describes some diagnostics for mixed models and illustrates the ideas using SAS. It is demonstrated how influence plots can easily be obtained from