Department of Biostatistics Faculty Papers

Permanent URI for this collectionhttps://digital.lib.washington.edu/handle/1773/11640

Browse

Now showing 1 - 14 of 14

Exon and junction microarrays detect widespread mouse strain- and sex-bias expression differences
(2008) Su, Wan-Lin; Modrek, Barmak; GuhaThakurta, Debraj; Edwards, Stephen; Shah, Jyothi, K.; Kulkarni, Amit V.; Russell, Archie; Schadt, Eric E.; Johnson, Jason M.; Castle, John C.
Background: Studies have shown that genetic and sex differences strongly influence gene expression in mice. Given the diversity and complexity of transcripts produced by alternative splicing, we sought to use microarrays to establish the extent of variation found in mouse strains and genders. Here, we surveyed the effect of strain and sex on liver gene and exon expression using male and female mice from three different inbred strains. Results: 71 liver RNA samples from three mouse strains - DBA/2J, C57BL/6J and C3H/HeJ - were profiled using a custom-designed microarray monitoring exon and exon-junction expression of 1,020 genes representing 9,406 exons. Gene expression was calculated via two different methods, using the 3'-most exon probe ("3' gene expression profiling") and using all probes associated with the gene ("whole-transcript gene expression profiling"), while exon expression was determined using exon probes and flanking junction probes that spanned across the neighboring exons ("exon expression profiling"). Widespread strain and sex influences were detected using a two-way Analysis of Variance (ANOVA) regardless of the profiling method used. However, over 90% of the genes identified in 3' gene expression profiling or whole transcript profiling were identified in exon profiling, along with 75% and 38% more genes, respectively, showing evidence of differential isoform expression. Overall, 55% and 32% of genes, respectively, exhibited strain- and sex-bias differential gene or exon expression. Conclusion: Exon expression profiling identifies significantly more variation than both 3' gene expression profiling and whole-transcript gene expression profiling. A large percentage of genes that are not differentially expressed at the gene level demonstrate exon expression variation suggesting an influence of strain and sex on alternative splicing and a need to profile expression changes at sub-gene resolution.
Association of kidney function with inflammatory and procoagulant markers in a diverse cohort: A cross-sectional analysis from the Multi-Ethnic Study of Atherosclerosis (MESA)
(2008) Keller, Christopher; Katz, Ronit; Cushman, Mary; Fried, Linda F.; Shlipak, Michael
Background: Prior studies using creatinine-based estimated glomerular filtration rate (eGFR) have found limited associations between kidney function and markers of inflammation. Using eGFR and cystatin C, a novel marker of kidney function, the authors investigated the association of kidney function with multiple biomarkers in a diverse cohort. Methods: The Multi-Ethnic Study of Atherosclerosis consists of 6,814 participants of white, African-American, Hispanic, and Chinese descent, enrolled from 2000-2002 from six U.S. communities. Measurements at the enrollment visit included serum creatinine, cystatin C, and six inflammatory and procoagulant biomarkers. Creatinine-based eGFR was estimated using the fourvariable Modification of Diet in Renal Disease equation, and chronic kidney disease was defined by an eGFR less than 60 mL/min/1.73 m2. Results: Adjusted partial correlations between cystatin C and all biomarkers were statistically significant: C-reactive protein (r = 0.08), interleukin-6 (r = 0.16), tumor necrosis factor-a soluble receptor 1 (TNF-aR1; r = 0.75), intercellular adhesion molecule-1 (r = 0.21), fibrinogen (r = 0.14), and factor VIII (r = 0.11; two-sided p less than 0.01 for all). In participants without chronic kidney disease, higher creatinine-based eGFR was associated only with higher TNF-aR1 levels. Conclusion: In a cohort characterized by ethnic diversity, cystatin C was directly associated with multiple procoagulant and inflammatory markers. Creatinine-based eGFR had similar associations with these biomarkers among subjects with chronic kidney disease.
Male circumcision, religion, and infectious diseases: an ecologic analysis of 118 developing countries
(2006) Drain, Paul K.; Halperin, Daniel T.; Hughes, James P.; Klausner, Jeffrey D.; Bailey, Robert C.
Background: Both religious practices and male circumcision (MC) have been associated with HIV and other sexually-transmitted infectious diseases. Most studies have been limited in size and have not adequately controlled for religion, so these relationships remain unclear. Methods: We evaluated relationships between MC prevalence, Muslim and Christian religion, and 7 infectious diseases using country-specific data among 118 developing countries. We used multivariate linear regression to describe associations between MC and cervical cancer incidence, and between MC and HIV prevalence among countries with primarily sexual HIV transmission. Results: Fifty-three, 14, and 51 developing countries had a high (>80%), intermediate (20-80%), and low ( less than 20%) MC prevalence, respectively. In univariate analyses, MC was associated with lower HIV prevalence and lower cervical cancer incidence, but not with HSV-2, syphilis, nor, as expected, with Hepatitis C, tuberculosis, or malaria. In multivariate analysis after stratifying the countries by religious groups, each categorical increase of MC prevalence was associated with a 3.65/100,000 women (95% CI 0.54-6.76, p = 0.02) decrease in annual cervical cancer incidence, and a 1.84-fold (95% CI 1.36-2.48, p less than 0.001) decrease in the adult HIV prevalence among sub-Saharan African countries. In separate multivariate analyses among non-sub-Saharan African countries controlling for religion, higher MC prevalence was associated with a 8.94-fold (95% CI 4.30-18.60) decrease in the adult HIV prevalence among countries with primarily heterosexual HIV transmission, but not, as expected, among countries with primarily homosexual or injection drug use HIV transmission (p = 0.35). Conclusion: Male circumcision was significantly associated with lower cervical cancer incidence and lower HIV prevalence in sub-Saharan Africa, independent of Muslim and Christian religion. As predicted, male circumcision was also strongly associated with lower HIV prevalence among countries with primarily heterosexual HIV transmission, but not among countries with primarily homosexual or injection drug use HIV transmission. These findings strengthen the reported biological link between MC and some sexually transmitted infectious diseases, including HIV and cervical cancer.
Towards standardized measurement of adverse events in spine surgery: conceptual model and pilot evaluation
(2006) Mirza, Sohail K.; Deyo, Richard A.; Heagerty, Patrick J.; Turner, Judith A.; Lee, Lorri A.; Goodkin, Robert
Background: Independent of efficacy, information on safety of surgical procedures is essential for informed choices. We seek to develop standardized methodology for describing the safety of spinal operations and apply these methods to study lumbar surgery. We present a conceptual model for evaluating the safety of spine surgery and describe development of tools to measure principal components of this model: (1) specifying outcome by explicit criteria for adverse event definition, mode of ascertainment, cause, severity, or preventability, and (2) quantitatively measuring predictors such as patient factors, comorbidity, severity of degenerative spine disease, and invasiveness of spine surgery. Methods: We created operational definitions for 176 adverse occurrences and established multiple mechanisms for reporting them. We developed new methods to quantify the severity of adverse occurrences, degeneration of lumbar spine, and invasiveness of spinal procedures. Using kappa statistics and intra-class correlation coefficients, we assessed agreement for the following: four reviewers independently coding etiology, preventability, and severity for 141 adverse occurrences, two observers coding lumbar spine degenerative changes in 10 selected cases, and two researchers coding invasiveness of surgery for 50 initial cases. Results: During the first six months of prospective surveillance, rigorous daily medical record reviews identified 92.6% of the adverse occurrences we recorded, and voluntary reports by providers identified 38.5% (surgeons reported 18.3%, inpatient rounding team reported 23.1%, and conferences discussed 6.1%). Trained observers had fair agreement in classifying etiology of 141 adverse occurrences into 18 categories (kappa = 0.35), but agreement was substantial (kappa = 0.61) for 4 specific categories: technical error, failure in communication, systems failure, and no error. Preventability assessment had moderate agreement (mean weighted kappa = 0.44). Adverse occurrence severity rating had fair agreement (mean weighted kappa = 0.33) when using a scale based on the JCAHO Sentinel Event Policy, but agreement was substantial for severity ratings on a new 11-point numerical severity scale (ICC = 0.74). There was excellent interrater agreement for a lumbar degenerative disease severity score (ICC = 0.98) and an index of surgery invasiveness (ICC = 0.99). Conclusion: Composite measures of disease severity and surgery invasiveness may allow development of risk-adjusted predictive models for adverse events in spine surgery. Standard measures of adverse events and risk adjustment may also facilitate post-marketing surveillance of spinal devices, effectiveness research, and quality improvement.
A prospective cohort study of surgical treatment for back pain with degenerated discs; study protocol
(2005) Deyo, Richard A.; Mirza, Sohail K.; Heagerty, Patrick J.; Turner, Judith A.; Martin, Brook I.
Background: The diagnosis of discogenic back pain often leads to spinal fusion surgery and may partly explain the recent rapid increase in lumbar fusion operations in the United States. Little is known about how patients undergoing lumbar fusion compare in preoperative physical and psychological function to patients who have degenerative discs, but receive only non-surgical care. Methods: Our group is implementing a multi-center prospective cohort study to compare patients with presumed discogenic pain who undergo lumbar fusion with those who have nonsurgical care. We identify patients with predominant low back pain lasting at least six months, one or two-level disc degeneration confirmed by imaging, and a normal neurological exam. Patients are classified as surgical or non-surgical based on the treatment they receive during the six months following study enrollment. Results: Three hundred patients discogenic low back pain will be followed in a prospective cohort study for two years. The primary outcome measure is the Modified Roland-Morris Disability Questionnaire at 24-months. We also evaluate several other dimensions of outcome, including pain, functional status, psychological distress, general well-being, and role disability. Conclusion: The primary aim of this prospective cohort study is to better define the outcomes of lumbar fusion for discogenic back pain as it is practiced in the United States. We additionally aim to identify characteristics that result in better patient selection for surgery. Potential predictors include demographics, work and disability compensation status, initial symptom severity and duration, imaging results, functional status, and psychological distress.
A coarsened multinomial regression model for perinatal mother to child transmission of HIV
(2008) Gard, Charlotte C.; Brown, Elizabeth R.
Background: In trials designed to estimate rates of perinatal mother to child transmission of HIV, HIV assays are scheduled at multiple points in time. Still, infection status for some infants at some time points may be unknown, particularly when interim analyses are conducted. Methods: Logistic regression models are commonly used to estimate covariate-adjusted transmission rates, but their methods for handling missing data may be inadequate. Here we propose using coarsened multinomial regression models to estimate cumulative and conditional rates of HIV transmission. Through simulation, we compare the proposed models to standard logistic models in terms of bias, mean squared error, coverage probability, and power. We consider a range of treatment effect and visit process scenarios, while including imperfect sensitivity of the assay and contamination of the endpoint due to early breastfeeding transmission. We illustrate the approach through analysis of data from a clinical trial designed to prevent perinatal transmission. Results: The proposed cumulative and conditional models performed well when compared to their logistic counterparts. Performance of the proposed cumulative model was particularly strong under scenarios where treatment was assumed to increase the risk of in utero transmission but decrease the risk of intrapartum and overall perinatal transmission and under scenarios designed to represent interim analyses. Power to estimate intrapartum and perinatal transmission was consistently higher for the proposed models. Conclusion: Coarsened multinomial regression models are preferred to standard logistic models for estimation of perinatal mother to child transmission of HIV, particularly when assays are missing or occur off-schedule for some infants.
Mortality ascertainment of participants in the National Wilms Tumor Study using the National Death Index: comparison of active and passive follow-up results
(2007) Cotton, Cecilia A.; Peterson, Susan; Norkool, Patricia A.; Breslow, Norman E.
Long term studies of childhood cancer survivors are hampered by difficulties in tracking young adult participants. After performing a National Death Index (NDI) search we sought to identify which factors best predicted a match among known decedents from the National Wilms Tumor Study (NWTS) and to determine if record linkage could substitute for missing follow-up in a cohort of NWTS survivors. To our knowledge, this is the first study to compare passive mortality follow-up using the NDI to active follow-up of a childhood and young adult population. Records for 984 known decedents and 3,406 subjects whose January 1, 2002 vital status was unknown were sent to the NDI in June 2003. In April 2005 NWTS follow-up records were used to reassess January 1, 2002 vital status. Matches were established for 709 of 789 known decedents (sensitivity 89.9%) with a date of death between 1979 and 2001, the calendar period covered by the NDI at the time of the search. No matches were identified among 1,052 subjects known to be alive in 2002 (specificity 100%). Factors associated with decreased sensitivity were an unknown social security number (sensitivity 87.8%), Hispanic ethnicity (76.4%) and foreign birth (56.5%). For 2,351 subjects with 2002 vital status unknown who had 13,166 pre 2002 person-years of missing observation, only 18 deaths were ascertained by the NDI whereas 79.3 were expected based on NWTS mortality data. Mortality analyses based strictly on NDI search results and those based on NWTS follow-up augmented with NDI search results yielded inflated estimates of the 15 year survival rate when compared with estimates based on NWTS active follow-up. Match rates were comparable to those observed in adult populations. Since the same selection factors were likely associated with NDI failure to match and NWTS loss to follow-up, use of the NDI to fill in missing follow-up data appears unwarranted.
Evaluation of methods for oligonucleotide array data via quantitative real-time PCR
(2006) Qin, Li-Xuan; Beyer, Richard P.; Hudson, Francesca N.; Linford, Nancy J.; Morris, Daryl E.; Kerr, Kathleen F.
Background: There are currently many different methods for processing and summarizing probelevel data from Affymetrix oligonucleotide arrays. It is of great interest to validate these methods and identify those that are most effective. There is no single best way to do this validation, and a variety of approaches is needed. Moreover, gene expression data are collected to answer a variety of scientific questions, and the same method may not be best for all questions. Only a handful of validation studies have been done so far, most of which rely on spike-in datasets and focus on the question of detecting differential expression. Here we seek methods that excel at estimating relative expression. We evaluate methods by identifying those that give the strongest linear association between expression measurements by array and the "gold-standard" assay. Quantitative reverse-transcription polymerase chain reaction (qRT-PCR) is generally considered the "gold-standard" assay for measuring gene expression by biologists and is often used to confirm findings from microarray data. Here we use qRT-PCR measurements to validate methods for the components of processing oligo array data: background adjustment, normalization, mismatch adjustment, and probeset summary. An advantage of our approach over spike-in studies is that methods are validated on a real dataset that was collected to address a scientific question. Results: We initially identify three of six popular methods that consistently produced the best agreement between oligo array and RT-PCR data for medium- and high-intensity genes. The three methods are generally known as MAS5, gcRMA, and the dChip mismatch mode. For medium- and high-intensity genes, we identified use of data from mismatch probes (as in MAS5 and dChip mismatch) and a sequence-based method of background adjustment (as in gcRMA) as the most important factors in methods' performances. However, we found poor reliability for methods using mismatch probes for low-intensity genes, which is in agreement with previous studies. Conclusion: We advocate use of sequence-based background adjustment in lieu of mismatch adjustment to achieve the best results across the intensity spectrum. No method of normalization or probeset summary showed any consistent advantages.
Transformations for cDNA Microarray Data
(2003) Cui, Xiangqin; Kerr, M. Kathleen; Churchill, Gary A.
Two channel microarray data often contain systematic variations that can be minimized by data transformation prior to further analysis. The most commonly observed effects are revealed by viewing scatter plots of the logarithm of the ratio by the average logarithmic intensity of the two color channels (RI plots). In this paper we present a general model for signal intensity data with multiple error sources. We demonstrate how these sources of error influence the shape of an RI plot. We then compare some currently available transformation strategies in terms of their mechanism and performance on both simulated and real microarray data. A linlog transformation is proposed to stabilize the variance of the log ratios. We also propose a regional smoothing method to remove variation in log ratios due to spatial heterogeneity on the microarray surface. The discussed transformations represent an important initial step in microarray data analysis for both ratio-based and ANOVA methods.
Synchrony of change in depressive symptoms, health status, and quality of life in persons with clinical depression
(2006) Diehr, Paula; Derleth, Ann M.; McKenna, Stephen P.; Martin, Mona L.; Bushnell, Donald M.; Simon, Gregory; Patrick, Donald L.
Background: Little is known about longitudinal associations among measures of depression, mental and physical health, and quality of life (QOL). We followed 982 clinically depressed persons to determine which measures changed and whether the change was synchronous with change in depressive symptoms. Methods: Data were from the Longitudinal Investigation of Depression Outcomes (LIDO). Depressive symptoms, physical and mental health, and quality of life were measured at baseline, 6 weeks, 3 months, and 9 months. Change in the measures was examined over time and for persons with different levels of change in depressive symptoms. Results: On average, all of the measures improved significantly over time, and most were synchronous with change in depressive symptoms. Measures of mental health changed the most, and physical health the least. The measures of change in QOL were intermediate. The 6-week change in QOL could be explained completely by change in depressive symptoms. The instruments varied in sensitivity to changes in depressive symptoms. Conclusion: In clinically depressed persons, measures of physical health, mental health, and quality of life showed consistent longitudinal associations with measures of depressive symptoms.
Managed care and patient ratings of the quality of specialty care among patients with pain or depressive symptoms
(2007) Grembowski, David; Paschane, David; Diehr, Paula; Katon, Wayne; Martin, Diane; Patrick. Donald L.
Background: Managed care efforts to regulate access to specialists and reduce costs may lower quality of care. Few studies have examined whether managed care is associated with patient perceptions of the quality of care provided by physician and non-physician specialists. Aim is to determine whether associations exist between managed care controls and patient ratings of the quality of specialty care among primary care patients with pain and depressive symptoms who received specialty care for those conditions. Methods: A prospective cohort study design was conducted in the offices of 261 primary physicians in private practice in Seattle in 1997. Patients (N = 17,187) were screened in waiting rooms, yielding a sample of 1,514 patients with pain only, 575 patients with depressive symptoms only, and 761 patients with pain and depressive symptoms. Patients (n = 1,995) completed a 6-month follow-up survey. Of these, 691 patients received specialty care for pain, and 356 patients saw mental health specialists. For each patient, managed care was measured by the intensity of managed care controls in the patient's health plan and primary care office. Quality of specialty care at follow-up was measured by patient rating of care provided by the specialists. Outcomes were pain interference and bothersomeness, Symptom Checklist for Depression, and restricted activity days. Results: The intensity of managed care controls in health plans and primary care offices was generally not associated with patient ratings of the quality of specialty care. However, pain patients in more-managed primary care offices had lower ratings of the quality of specialty care from physician specialists and ancillary providers. Conclusion: For primary care patients with pain or depressive symptoms and who see specialists, managed care controls may influence ratings of specialty care for patients with pain but not patients with depressive symptoms.
The effect of different public health interventions on longevity, morbidity, and years of healthy life
(2007) Diehr, Paula; Derleth, Ann; Cai, Liming; Newman, Anne B.
Background: Choosing cost-effective strategies for improving the health of the public is difficult because the relative effects of different types of interventions are not well understood. The benefits of one-shot interventions may be different from the benefits of interventions that permanently change the probability of getting sick, recovering, or dying. Here, we compare the benefits of such types of public health interventions. Methods: We used multi-state life table methods to estimate the impact of five types of interventions on mortality, morbidity (years of life in fair or poor health), and years of healthy life (years in excellent, very good, or good health). Results: A one-shot intervention that makes all the sick persons healthy at baseline would increase life expectancy by 3 months and increase years of healthy life by 6 months, in a cohort beginning at age 65. An equivalent amount of improvement can be obtained from an intervention that either decreases the probability of getting sick each year by 12%, increases the probability of a sick person recovering by 16%, decreases the probability that a sick person dies by 15%, or decreases the probability that a healthy person dies by 14%. Interventions aimed at keeping persons healthy increased longevity and years of healthy life, while decreasing morbidity and medical expenditures. Interventions focused on preventing mortality had a greater effect on longevity, but had higher future morbidity and medical expenditures. Results differed for older and younger cohorts and depended on the value to society of an additional year of sick life. Conclusion: Interventions that promote health and prevent disease performed well, but other types of intervention were sometimes better. The value to society of interventions that increase longevity but also increase morbidity needs further research. More comprehensive screening and treatment of new Medicare enrollees might improve their health and longevity without increasing future medical expenditures.
Weight-modification trials in older adults: what should the outcome measure be?
(2002) Diehr, Paula; Newman, Anne B.; Jackson, Sharon A.; Kuller, Lewis; Powe, Neil
Background: Overweight older adults are often counseled to lose weight, even though there is little evidence of excess mortality in that age group. Overweight and underweight may be more associated with health status than with mortality, but few clinical trials of any kind have been based on maximizing years of healthy life (YHL), as opposed to years of life (YOL). Objective: This paper examines the relationship of body mass index (BMI) to both YHL and YOL. Results were used to determine whether clinical trials of weight-modification based on improving YHL would be more powerful than studies based on survival. Design: We used data from a cohort of 4,878 non-smoking men and women aged 65–100 at baseline (mean age 73) and followed 7 years. We estimated mean YHL and YOL in four categories of BMI: underweight, normal, overweight, and obese. Results: Subjects averaged 6.3 YOL and 4.6 YHL of a possible 7 years. Both measures were higher for women and whites. For men, none of the BMI groups was significantly different from the normal group on either YOL or YHL. For women, the obese had significantly lower YHL (but not YOL) than the normals, and the underweight had significantly lower YOL and YHL. The overweight group was not significantly different from the normal group on either measure. Conclusions: Clinical trials of weight loss interventions for obese older women would require fewer participants if YHL rather than YOL was the outcome measure. Interventions for obese men or for the merely overweight are not likely to achieve differences in either YOL or YHL. Evaluations of interventions for the underweight (which would presumably address the causes of their low weight) may be conducted efficiently using either outcome measure.
Missing Confounding Data in Marginal Structural Models: A Comparison of Inverse Probability Weighting and Multiple Imputation
(2008) Moodie, Erica E.M.; Delaney, Joseph A.; Lefebvre, Genevieve; Platt, Robert W.
Standard statistical analyses of observational data often exclude valuable information from individuals with incomplete measurements. This may lead to biased estimates of the treatment effect and loss of precision. The issue of missing data for inverse probability of treatment weighted estimation of marginal structural models (MSMs) has often been addressed, though little has been done to compare different missing data techniques in this relatively new method of analysis. We propose a method for systematically dealing with missingness in MSMs by treating missingness as a cause for censoring and weighting subjects by the inverse probability of missingness. We developed a series of simulations to systematically compare the effect of using case deletion, our inverse weighting approach, and multiple imputation in a MSM when there is missing information on an important confounder. We found that multiple imputation was slightly less biased and considerably less variable than the inverse probability approach. Thus, the lower variability achieved through multiple imputation makes it desirable in most practical cases where the missing data are strongly predicted by the available data. Inverse probability weighting is, however, a superior alternative to naive approaches such as complete-case analysis.

Browse

Recent Submissions