Multilevel modeling versus cross-sectional analysis for assessing the longitudinal tracking of cardiovascular risk factors over time

Department of Biostatistics, Boston University School of Public Health, Boston, MA, U.S.A.
Statistics in Medicine (Impact Factor: 2.04). 12/2013; 32(28). DOI: 10.1002/sim.5880
Source: PubMed

ABSTRACT Correlated data are obtained in longitudinal epidemiological studies, where repeated measurements are taken on individuals or groups over time. Such longitudinal data are ideally analyzed using multilevel modeling approaches, which appropriately account for the correlations in repeated responses in the same individual. Commonly used regression models are inappropriate as they assume that measurements are independent. In this tutorial, we use multilevel modeling to demonstrate its use for analysis of correlated data obtained from serial examinations on individuals. We focus on cardiovascular epidemiological research where investigators are often interested in quantifying the relations between clinical risk factors and outcome measures (X and Y, respectively), where X and Y are measured repeatedly over time, for example, using serial observations on participants attending multiple examinations in a longitudinal cohort study. For instance, it may be of interest to evaluate the relations between serial measures of left ventricular mass (outcome) and of its potential determinants (i.e., body mass index and blood pressure), both of which are measured over time. In this tutorial, we describe the application of multilevel modeling to cardiovascular risk factors and outcome data (using serial echocardiographic data as an example of an outcome). We suggest an analytical approach that can be implemented to evaluate relations between any potential outcome of interest and risk factors, including assessment of random effects and nonlinear relations. We illustrate these steps using echocardiographic data from the Framingham Heart Study with SAS PROC MIXED. Copyright © 2013 John Wiley & Sons, Ltd.

1 Follower
  • [Show abstract] [Hide abstract]
    ABSTRACT: To elucidate the hospital characteristics associated with hospital performance and time trends in acute myocardial infarction (AMI) care using multilevel multivariable analysis of longitudinal data. Retrospective longitudinal study. One hundred and fourteen hospitals in Japan. A total of 26 210 AMI patients admitted between 2008 and 2011. A composite score was calculated from five AMI process measures. Hospital performances and time trends were then investigated based on this composite score. Using generalized linear mixed models with random intercepts (indicating hospital baseline performance) and random slopes (indicating trends in improvement), we analyzed the associations between performance and the following factors: hospital ownership, AMI case volume, number of cardiovascular specialists per AMI patient and participation in a public disclosure program. Hospitals that demonstrated high performance in the composite score were significantly associated with high AMI case volume, municipal ownership and agreement to named disclosure of hospital performance. The following factors were significantly associated with time trends of improvement in performance: public and private ownership, AMI case volume and number of cardiovascular specialists per AMI patient. In addition, higher performances were associated with diminished improvement. Time trends in improvement were related to baseline performance and several hospital characteristics. Furthermore, hospitals that had agreed to named disclosure of performance were more likely to have better quality of care at the initial point of public disclosure. These findings can inform the decision-making process for quality improvement, and allow a greater understanding and interpretation of disclosed performances in quality measures.
    International Journal for Quality in Health Care 08/2014; 26(5). DOI:10.1093/intqhc/mzu073 · 1.58 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The development of donor-specific HLA antibodies (DSA) is associated with worse renal allograft survival in adult patients. This study assessed the natural history of de novo DSA, and its impact on renal function in pediatric renal transplant recipients (RTR). HLA antibodies were measured prospectively using single-antigen-bead assays at 1, 3, 6 and 12 months posttransplant followed by 12-monthly intervals and during episodes of allograft dysfunction. Of 215 patients with HLA antibody monitoring, 75 (35%) developed DSA at median of 0.25 years posttransplant with a high prevalence of Class II (70%) and HLA-DQ (45%) DSA. DSA resolved in 35 (47%) patients and was associated with earlier detection (median, inter-quartile range 0.14, 0.09–0.33 vs. 0.84, 0.15–2.37 years) and lower mean fluorescence intensity (MFI) (2658, 1573–3819 vs. 7820, 5166–11 990). Overall, DSA positive patients had more rapid GFR decline with a 50% reduction in GFR at mean 5.3 (CI: 4.7–5.8) years versus 6.1 (5.7–6.4) years in DSA negative patients (p = 0.02). GFR decreased by a magnitude of 1 mL/min/1.73 m2 per log10 increase in Class II DSA MFI (p < 0.01). Using Cox regression, independent factors predicting poorer renal allograft outcome were older age at transplant (hazard ratio 1.1, CI: 1.0–1.2 per year), tubulitis (1.5, 1.3–1.8) and microvasculature injury (2.9, 1.4–5.7). In conclusion, pediatric RTR with de novo DSA and microvasculature injury were at risk of allograft failure.
    American Journal of Transplantation 10/2014; 14(10):2350-2358. DOI:10.1111/ajt.12859 · 6.19 Impact Factor