James S Hodges

University of Minnesota Duluth, Duluth, Minnesota, United States

Are you James S Hodges?

Claim your profile

Publications (213)628.46 Total impact

  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Despite well-recognized recommendations to reduce human exposure to zoonotic pathogens, the use of personal and herd-level protective practices is inconsistent in communities where human interactions with animals are common. This study assessed household-level participation in rodent- (extermination, proper food storage, trash disposal), occupational- (preventive veterinary care, boot-wearing, glove-wearing), and garden-associated (restricting animal access, boot-wearing, glove-wearing) protective practices in farms, villages, and slums in the Los Rios region, Chile, where zoonotic pathogens are endemic. Questionnaires administered at 422 households across 12 communities recorded household-level socio-demographic characteristics and participation in nine protective practices. Household inclusion in the analysis of occupational practices required having livestock and a household member with occupational exposure to livestock (n = 127), and inclusion in analysis of garden practices required having a garden and at least one animal (n = 233). The proportion of households participating in each protective practice was compared across community types through chi-square analyses. Mixed effects logistic regression assessed household-level associations between socio-demographic characteristics and participation in each protective practice. Most households (95.3 %) reported participation in rodent control, and a positive association between the number of rodent signs in a household and rodent extermination was observed (OR: 1.75, 95 % CI: 1.41, 2.16). Occupational protective practices were reported in 61.8 % of eligible households; household size (OR: 1.63, 95 % CI: 1.17, 5.84) and having children (OR: 0.22, 95 % CI: 0.06, 0.78) were associated with preventive veterinary care. Among eligible households, 73.8 % engaged in protective practices when gardening, and species diversity was positively associated with wearing boots (OR: 1.27, 95 % CI: 1.03, 1.56). Household-level participation in all three protective practices within any exposure category was limited (<10.4 %) and participation in any individual protective practice varied considerably within and across community types. The levels of participation in protective practices reported in this study are consistent with descriptions in the literature of imperfect use of methods that reduce human exposure to zoonotic pathogens. The wide differences across communities in the proportion of households participating in protective practices against human exposure to zoonotic pathogens, suggests that future research should identify community-level characteristics that influence household participation in such practices.
    BMC Public Health 12/2015; 15(1):713. DOI:10.1186/s12889-015-1964-2 · 2.26 Impact Factor
  • Andrew G Huff · James S Hodges · Shaun P Kennedy · Amy Kircher ·
    [Show abstract] [Hide abstract]
    ABSTRACT: To protect and secure food resources for the United States, it is crucial to have a method to compare food systems' criticality. In 2007, the U.S. government funded development of the Food and Agriculture Sector Criticality Assessment Tool (FASCAT) to determine which food and agriculture systems were most critical to the nation. FASCAT was developed in a collaborative process involving government officials and food industry subject matter experts (SMEs). After development, data were collected using FASCAT to quantify threats, vulnerabilities, consequences, and the impacts on the United States from failure of evaluated food and agriculture systems. To examine FASCAT's utility, linear regression models were used to determine: (1) which groups of questions posed in FASCAT were better predictors of cumulative criticality scores; (2) whether the items included in FASCAT's criticality method or the smaller subset of FASCAT items included in DHS's risk analysis method predicted similar criticality scores. Akaike's information criterion was used to determine which regression models best described criticality, and a mixed linear model was used to shrink estimates of criticality for individual food and agriculture systems. The results indicated that: (1) some of the questions used in FASCAT strongly predicted food or agriculture system criticality; (2) the FASCAT criticality formula was a stronger predictor of criticality compared to the DHS risk formula; (3) the cumulative criticality formula predicted criticality more strongly than weighted criticality formula; and (4) the mixed linear regression model did not change the rank-order of food and agriculture system criticality to a large degree. © 2015 Society for Risk Analysis.
    Risk Analysis 04/2015; 35(8). DOI:10.1111/risa.12377 · 2.50 Impact Factor
  • P Chantaracherd · M T John · J S Hodges · E L Schiffman ·

    Journal of dental research 03/2015; 94(6). DOI:10.1177/0022034515578912 · 4.14 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Introduction: Epidermal nerve fiber (ENF) density, morphology, and epidermal innervation patterns were examined in children using 2 different techniques, punch biopsy and suction blister. Methods: Healthy children without symptoms or history of peripheral neuropathy and normal by neurologic examination were studied. Punch biopsy and suction blister specimens were collected from the lateral thigh and distal leg. ENFs were traced from confocal images of immunohistochemically stained samples. Statistical methods included repeated-measures analysis of covariance. Results: Blister and biopsy nerve counts were associated. ENF density in children was dense, lower for older children (P<0.01) and with no difference between boys and girls (P=0.92). Many ENFs appeared multibranched and elongated. Conclusions: Epidermal innervation in the pediatric population is dense and age-dependent. Blister specimens are less invasive and may provide an alternative to punch biopsy for determining ENF density in children at risk for neuropathy.
    Muscle & Nerve 03/2015; 51(3). DOI:10.1002/mus.24329 · 2.28 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: This study assessed the effects of cerebral malaria (CM) and severe malarial anemia (SMA) on individual neurocognitive domains. Eighty children with CM, 86 with SMA, and 61 community children (CC) were assessed for gross motor skills, fine motor skills, visual reception, receptive language, and expressive language a week after discharge (CM or SMA) or at enrolment (CC), and 6 and 12 months later. At 12-months follow-up, children with CM had significantly lower scores than CC for all outcomes. Children with SMA had significantly lower scores than CC for visual reception, receptive language, and expressive language, and scores that were lower but did not reach significance for gross and fine motor skills. Children with CM had significantly lower scores than children with SMA for fine motor skills. Children with SMA and CM have long-term impairment in multiple neurocognitive domains. Fine motor skills may be affected more profoundly in CM than SMA.
    Learning and Individual Differences 01/2015; DOI:10.1016/j.lindif.2015.01.010 · 1.62 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Fanconi anemia (FA) is an inherited DNA repair disorder associated with short stature and bone marrow failure usually requiring hematopoietic cell transplant (HCT). While low bone mineral density (BMD) has been reported in leukemia patients after HCT, little is known about BMD in FA children after HCT (FA HCT). This study's goals were to compare BMD in FA HCT to BMD in healthy controls, and in children who received HCT for hematologic malignancy (Cancer HCT), and to test for associations between BMD and risk factors for bone loss. This cross-sectional study included 20 FA HCT, 13 Cancer HCT, and 90 healthy controls, age-matched and <18 years old at evaluation. BMD Z-scores for total body (TBMD) and lumbar spine (LBMD) were measured by dual energy x-ray absorptiometry and adjusted for height-for-age Z-score (HAZ). FA HCT had lower mean TBMDHAZ Z-score (by 0.8 SD) and higher fraction with Z-score ≤ -1 than healthy controls (42% vs. 11%). No LBMD deficits were detected. FA HCT and Cancer HCT groups did not differ significantly in TBMD or LBMD Z-scores. In FA HCT patients, lower BMI and lower percent fat were associated with lower BMD. This study highlights the importance of monitoring BMD to optimize bone health in FA patients. Copyright © 2015 American Society for Blood and Marrow Transplantation. Published by Elsevier Inc. All rights reserved.
    Biology of blood and marrow transplantation: journal of the American Society for Blood and Marrow Transplantation 01/2015; 21(5). DOI:10.1016/j.bbmt.2015.01.002 · 3.40 Impact Factor
  • P Chantaracherd · M T John · J S Hodges · E L Schiffman ·
    [Show abstract] [Hide abstract]
    ABSTRACT: The aim of this study was to determine the association between more advanced stages of temporomandibular joint (TMJ) intra-articular disorders ("TMJ intra-articular status"), representing a transition from normal joint structure to TMJ disc displacement with and without reduction (DDwR and DDwoR) to degenerative joint disease (DJD), and patient-reported outcomes of jaw pain, function, and disability ("TMD impact"). This cross-sectional study included 614 cases from the RDC/TMD Validation Project with at least one temporomandibular disorder (TMD) diagnosis. TMJ intra-articular status was determined by 3 blinded, calibrated radiologists using magnetic resonance imaging and computed tomography as one of normal joint structure, DDwR, DDwoR, or DJD, representing the subject's most advanced TMJ diagnosis. TMD impact was conceptualized as a latent variable consisting of 1) pain intensity (Characteristic Pain Index from the Graded Chronic Pain Scale [GCPS]), 2) jaw function (Jaw Functional Limitation Scale), and 3) disability (Disability Points from GCPS). A structural equation model estimated the association of TMJ intra-articular status with the latent measure TMD impact as a correlation coefficient in all TMD cases (n = 614) and in cases with a TMD pain diagnosis (n = 500). The correlations between TMJ intra-articular status and TMD impact were 0.05 (95% confidence interval [CI], -0.04 to 0.13) for all TMD cases and 0.07 (95% CI, -0.04 to 0.17) for cases with a pain diagnosis, which are neither statistically significant nor clinically relevant. Conceptualizing worsening of TMJ intra-articular disorders as 4 stages and characterizing impact from TMD as a composite of jaw pain, function, and disability, this cross-sectional study found no clinically significant association. Models of TMJ intra-articular status other than ours (normal structure → DDwR → DDwoR → DJD) should be explored. © International & American Associations for Dental Research 2015.
    Journal of Dental Research 01/2015; 94(3 Suppl). DOI:10.1177/0022034514565793 · 4.14 Impact Factor
  • J.O. Look · J.S. Hodges · E.L. Schiffman ·

    International Journal of Oral and Maxillofacial Surgery 10/2014; 43(10). DOI:10.1016/j.ijom.2014.07.002 · 1.57 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Background Childhood cancer survivors (CCS) are at high risk of developing treatment-related late effects, including cardiovascular disease and diabetes. Late effects can be exacerbated by low physical activity (PA) levels. Relationships between PA and cardiovascular risk factors during childhood have not been well described in CCS.ProcedurePA and cardiovascular risk factors were measured cross-sectionally in 319 CCS and 208 sibling controls aged 9–18 years. Comparisons between CCS and controls and associations of outcomes with PA (dichotomized at 60 min/day or treated as continuous) were performed with linear regression.ResultsAmong CCS, the high PA group had lower percent fat mass (24.4% vs. 29.8%, P < 0.0001), abdominal subcutaneous fat (67.9 vs. 97.3 cm3, P = 0.0004), and abdominal visceral fat (20.0 vs. 24.9 cm3, P = 0.007) and greater lean body mass (41.3 vs. 39.5 kg, P = 0.009) than the low PA group. Comparing CCS to controls, differences in waist circumference (Pinteraction = 0.04), percent fat mass (Pinteraction = 0.04), and abdominal subcutaneous (Pinteraction = 0.02) and visceral (Pinteraction = 0.004) fat between low and high PA groups were greater in CCS than controls, possibly due to greater overall adiposity in CCS.Conclusions High PA in CCS resulted in an improved cardiovascular profile, consisting primarily of lower fat mass and greater lean mass, similar to that observed in controls. This suggests interventions directed to increase PA in CCS may reduce the risk of future cardiovascular disease. Pediatr Blood Cancer © 2014 Wiley Periodicals, Inc.
    Pediatric Blood & Cancer 10/2014; 62(2). DOI:10.1002/pbc.25276 · 2.39 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Background: Elevated endogenous plasma erythropoietin (EPO) levels have been associated with protection from acute neurologic deficits in Kenyan children with cerebral malaria (CM). Based on these findings and animal studies, clinical trials of recombinant human EPO (rHuEPO) have been started in children with CM. Recent clinical trials in adults with acute ischemic stroke have demonstrated increased mortality with rHuEPO treatment. We conducted a study in children with CM to assess the relationship of endogenous plasma and cerebrospinal fluid (CSF) EPO levels with mortality and acute and long-term neurologic outcomes. Methods: A total of 210 children between 18 months and 12 years of age with a diagnosis of CM, were enrolled at Mulago Hospital, Kampala, Uganda. Plasma (n = 204) and CSF (n = 147) EPO levels at admission were measured by radioimmunoassay and compared with mortality and neurologic outcomes. Results: After adjustment for age and hemoglobin level, a 1-natural-log increase in plasma EPO level was associated with a 1.74-fold increase in mortality (95% confidence interval, 1.09-2.77, P = .02). Plasma and CSF EPO levels also correlated positively with coma duration (P = .05 and P = .02, respectively). Plasma and CSF EPO levels did not differ in children with vs those without acute or long-term neurologic deficits. Plasma EPO levels correlated positively with markers of endothelial and platelet activation and histidine-rich protein-2 levels, but remained associated with mortality after adjustment for these factors. Conclusions: High endogenous plasma EPO levels are associated with prolonged coma duration and increased mortality in children >18 months of age with CM.
    Clinical Infectious Diseases 09/2014; 60(1). DOI:10.1093/cid/ciu735 · 8.89 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Context. Vulvar pain level may fluctuate in women with vulvodynia even in the absence of therapy; however, there is little evidence suggesting which factors may be associated with variability. Objective. Determine the feasibility of using smartphones to collect prospective data on vulvar pain and factors that may influence vulvar pain level. Methods. 24 clinically confirmed women were enrolled from a population-based study and asked to answer five questions using their smartphones each week for one month. Questions assessed vulvar pain level (0-10), presence of pain upon wakening, pain elsewhere in their body, treatment use, and intercourse. Results. Women completed 100% of their scheduled surveys, with acceptability measures highly endorsed. Vulvar pain ratings had a standard deviation within women of 1.6, with greater variation on average among those with higher average pain levels (P < 0.001). On the weeks when a woman reported waking with pain, her vulvar pain level was higher by 1.82 on average (P < 0.001). Overall, average vulvar pain level was not significantly associated with the frequency of reporting other body pains (P = 0.64). Conclusion. Our smartphone tracking system promoted excellent compliance with weekly tracking of factors that are otherwise difficult to recall, some of which were highly associated with vulvar pain level.
    Pain Research and Treatment 06/2014; 2014(5):659863. DOI:10.1155/2014/659863
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Background Malaria in highland areas of Kenya affects children and adults. Local clinicians include symptoms other than fever when screening for malaria because they believe that fever alone does not capture all cases of malaria. Methods Individuals who presented to dispensaries in a highland Kenya site of low, unstable malaria transmission from 2007–2011 with 1 or more of 11 symptoms were tested by microscopy for malaria. Clinical malaria was defined as asexual Plasmodium falciparum infection on peripheral blood smear in an individual with any screening symptom. Asymptomatic P. falciparum infection was assessed in a cohort at ten time points to determine the extent to which symptomatic episodes with parasitaemia might be attributable to baseline (asymptomatic) parasitaemia in the community. Results 3,420 individuals were screened for malaria, 634 < 5 years of age and 2,786 ≥ 5 years of age. For the diagnosis of clinical malaria, the symptom of fever had a sensitivity and specificity of 88.9% and15.4% in children <5 years, and 55.8% and 54.4% in children ≥5 years, respectively. Adding the symptom of headache increased sensitivity to 94. 4% in children <5 years and 96.8% in individuals ≥5 years, but decreased specificity to 9.9% and 11.6%, respectively, and increased the number of individuals who would be tested by 6% and 92%, respectively. No combination of symptoms improved upon the presence fever or headache for detection of clinical malaria. In the cohort of asymptomatic individuals, P. falciparum parasitaemia was infrequent (0.1%). Conclusion In areas of low, unstable malaria transmission, fever is a sensitive indicator of clinical malaria in children <5 years, but not in older children and adults. Adding headache to fever as screening symptom increases sensitivity of detection in individuals ≥5 years old at the cost of decreased specificity. Screening for symptoms in addition to fever may be required to accurately capture all cases of clinical malaria in individuals ≥5 years old in areas of low malaria transmission.
    Malaria Journal 04/2014; 13(1):163. DOI:10.1186/1475-2875-13-163 · 3.11 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Background: Cerebral malaria (CM) is associated with long-term neurocognitive impairment in children ≥5 years of age. No prospective studies to date have assessed neurocognitive impairment in children with CM <5 years of age, or in children with severe malarial anemia (SMA), a form of severe malaria estimated to affect as many as 5 million children annually. Methods: Children <5 years of age presenting to Mulago Hospital, Kampala, Uganda, with CM (n = 80) or SMA (n = 86) were assessed for overall cognitive ability, attention, and associative memory 1 week after discharge and 6 and 12 months later. The z scores for each domain were computed based on scores of 61 healthy community children (CC), who were also tested at enrollment and 6 and 12 months later. Groups were compared using mixed linear models, adjusted for age, weight for age, and child's education. Results: At 12 months, children with CM had lower adjusted scores than CC in cognitive ability (P < .001), attention (P = .02), and associative memory, (P = .002). Children with SMA had lower scores than CC in cognitive ability (P = .01) but not attention or associative memory. Cognitive ability scores in children with CM and SMA did not differ significantly. Conclusions: In children <5 years of age, SMA is associated with long-term impairment in cognitive ability, whereas CM is associated with additional impairment in the areas of attention and associative memory. SMA may be a major contributor to long-term neurocognitive impairment in children in sub-Saharan Africa.
    Clinical Infectious Diseases 04/2014; 59(3). DOI:10.1093/cid/ciu293 · 8.89 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Background: Tools that estimate recent and long-term malaria transmission in a population would be highly useful for malaria elimination programs. Methods: The prevalence of antibodies to 11 Plasmodium falciparum antigens was assessed by cytometric bead assay or enzyme-linked immunosorbent assay in 1000 people in a highland area of Kenya over 14 months, during a period of interrupted malaria transmission. Results: Antibodies differed by antigen in acquisition with age: rapid (>80% antibody positive by age 20 years, 5 antigens), moderate (>40% positive by age 20 years, 3 antigens), or slow (<40% positive by age 20 years, 3 antigens). Antibody seroreversion rates in the 14 months between samples decreased with age rapidly (7 antigens), slowly (3 antigens), or remained high at all ages (schizont extract). Estimated antibody half-lives in individuals >10 years of age were long (40 to >80 years) for 5 antigens, moderate (5-20 years) for 3 antigens, and short (<1 year) for 3 antigens. Conclusions: Antibodies to P. falciparum antigens in malaria-endemic areas vary by age, antigen, and time since last exposure to P. falciparum. Multiplex P. falciparum antibody testing could provide estimates of long-term and recent malaria transmission and potentially of a population's susceptibility to future clinical malaria.
    The Journal of Infectious Diseases 04/2014; 210(7). DOI:10.1093/infdis/jiu225 · 6.00 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Evaluate the association of abdominal visceral and subcutaneous fat, independent of total body fat, with cardiometabolic risk factors and insulin resistance among youth. Visceral and subcutaneous fat, percent total body fat, insulin resistance (adjusted for lean body mass: Mlbm), blood pressure, glucose, insulin, and lipids were obtained in 472 youth ages 6-18 years. Linear regression, adjusted for age, sex, race, Tanner stage, and percent total body fat, was used to evaluate associations of visceral and subcutaneous fat with cardiometabolic risk factors. Visceral fat was associated inversely with Mlbm (p=0.003) and positively with fasting insulin (p=0.002) and triglycerides (p=0.002). Visceral fat levels above the mean were associated inversely with HDL cholesterol (p=0.002), and positively systolic blood pressure (p<0.0001) and non-HDL cholesterol (p<0.0001). Subcutaneous fat was associated inversely with Mlbm (p=0.003) and HDL cholesterol (p<0.05), and positively with fasting glucose (p<0.05), fasting insulin (p=0.0003), systolic blood pressure (p=0.005), and triglycerides (p=0.003). Subcutaneous fat levels above the mean were associated with non-HDL cholesterol (p=0.0002). These findings suggest that there may be a threshold level of visceral and subcutaneous fat (regardless of total body fat), that when exceeded in childhood, is more likely to be associated with many cardiometabolic risk factors. Triglycerides and insulin resistance appear to be associated with these fat depots at even lower thresholds of abdominal adiposity.
    04/2014; 4(2):101-107. DOI:10.1111/cob.12044
  • [Show abstract] [Hide abstract]
    ABSTRACT: Cancer is the leading cause of death by disease among U.S. children ages 1-14 years; however, over the last few decades there has been a dramatic increase in survival in these individuals. Although free of cancer, survivors are faced with a variety of chronic health problems including an increased risk of cardiovascular and metabolic abnormalities. The effect of cancer treatments on vascular structure and function in childhood cancer survivors (CCS) has not been examined. Measures of carotid artery stiffness (compliance and distensibility) and thickness (IMT), brachial artery endothelial-dependent dilation (EDD), and endothelial-independent dilation (EID) were obtained from ultrasound imaging in 319 CCS (age: 14.6 ± 0.1 years; male/female: 112/96) who were >5 years from diagnosis and 208 (age: 13.6 ± 0.2 years; male/female: 171/148) siblings who had never been diagnosed with cancer. Participants were 9-18 years of age at examination. Survivors of leukemia had lower carotid distensibility and compliance, indicating increased arterial stiffness, when compared to controls. There were no significant differences in measures of carotid stiffness or EDD in survivors of solid tumors and central nervous system (CNS) tumors compared to controls. EDD was lower in leukemia survivors than in controls, and EID was greater in survivors of CNS tumors than in controls. These results demonstrate that early in life, CCS have arterial changes indicating increased risk for premature atherosclerosis and cardiovascular disease. Therefore, it is reasonable to advocate that efforts should be directed at monitoring and managing cardiovascular risk factors in CCS. Pediatr Blood Cancer © 2013 Wiley Periodicals, Inc.
    Pediatric Blood & Cancer 03/2014; 61(3). DOI:10.1002/pbc.24829 · 2.39 Impact Factor
  • Source
    Lisa Henn · James S. Hodges ·
    [Show abstract] [Hide abstract]
    ABSTRACT: Scattered reports of multiple maxima in posterior distributions or likelihoods for mixed linear models appear throughout the literature. Less scrutinised is the restricted likelihood, which is the posterior distribution for a specific prior distribution. This paper surveys existing literature and proposes a unifying framework for understanding multiple maxima. For those problems with covariance structures that are diagonalisable in a specific sense, the restricted likelihood can be viewed as a generalised linear model with gamma errors, identity link and a prior distribution on the error variance. The generalised linear model portion of the restricted likelihood can be made to conflict with the portion of the restricted likelihood that functions like a prior distribution on the error variance, giving two local maxima in the restricted likelihood. Applying in addition an explicit conjugate prior distribution to variance parameters permits a second local maximum in the marginal posterior distribution even if the likelihood contribution has a single maximum. Moreover, reparameterisation from variance to precision can change the posterior modality; the converse also is true. Modellers should beware of these potential pitfalls when selecting prior distributions or using peak-finding algorithms to estimate parameters.
    International Statistical Review 03/2014; 82(1). DOI:10.1111/insr.12046 · 1.20 Impact Factor
  • Andrew G. Huff · Amy Kircher · James S. Hodges · Shaun Kennedy ·
    [Show abstract] [Hide abstract]
    ABSTRACT: Determining food system criticality is necessary to mitigate risks to the nation’s food supply and prioritize and allocate funding. The Food and Agriculture Sector Criticality Assessment Tool (FASCAT) is a tool used broadly by state governments to determine the criticality of food systems throughout the US State officials (SOs) responsible for food defense (n=32) were surveyed to determine whether FASCAT is of value to food defense and to determine SOs’ security beliefs, values, and practices related to food defense. Results indicated that: (1) SOs believe FASCAT is easier to use than other forms of risk assessment; (2) FASCAT training may have introduced bias into assessment of probability, threat, vulnerability, and consequences; (3) FASCAT is valuable to SOs; (4) SOs do not routinely follow security management best practices; (5) SOs believe that intentional biological threats to the food system are the most probable threats, though without supporting evidence; and (6) SOs believe food defense risk mitigation is not adequately funded by state or federal governments. These findings indicate that even though bias was potentially introduced to FASCAT assessments, SOs believe FASCAT has been useful to them in determining food system criticality. SOs indicate that more funding is needed from state and federal governments to adequately mitigate and manage food defense risks, and that they require more comprehensive training from food defense subject matter experts in threat assessment, risk mitigation, and security management to reduce the possibility of bias from FASCAT training.
    Journal of Homeland Security and Emergency Management 01/2014; 11(1). DOI:10.1515/jhsem-2013-0063 · 0.55 Impact Factor
  • Laura A. Hatfield · James S. Hodges · Bradley P. Carlin ·
    [Show abstract] [Hide abstract]
    ABSTRACT: Treatments affect many aspects of disease; for example, a drug may improve symptoms, prolong survival, and cause serious side effects. A broader perspective on clinical effectiveness, considering multiple outcomes, requires analyses that account for relationships among outcomes. So-called joint modeling induces such relationships via shared parameters. Practical questions arise, including "When do we require a joint model?" and "How much do we gain by its use?" Motivated by these questions, we compare Gaussian joint models with shared latent parameters to separate models for each outcome individually. When we assume a single longitudinal measurement, known error variances, and no censoring, joint and separate treatment effect posteriors converge as the priors become improper. This result still holds when we add multiple longitudinal measurements and unknown error variance, but not when we make the prior informative for at least one treatment effect (longitudinal or survival). Joint models also improve inference under some censoring scenarios. Our results suggest that joint models are most useful when an information imbalance allows abundant information in one outcome to compensate for a paucity of information in another.
    Statistics and its interface 01/2014; 7(4):439-453. DOI:10.4310/SII.2014.v7.n4.a2 · 2.93 Impact Factor
  • James S. Hodges ·

    Environmetrics 12/2013; 24(8). DOI:10.1002/env.2244 · 1.51 Impact Factor

Publication Stats

4k Citations
628.46 Total Impact Points


  • 1997-2015
    • University of Minnesota Duluth
      • • Department of Mechanical and Industrial Engineering
      • • Medical School
      Duluth, Minnesota, United States
  • 2013
    • Technische Universität München
      München, Bavaria, Germany
  • 2010-2012
    • Abbott Northwestern Hospital
      Minneapolis, Minnesota, United States
  • 2011
    • Saint Catherine University
      Minneapolis, Minnesota, United States
  • 2009
    • University of North Carolina at Chapel Hill
      North Carolina, United States
    • Centenary Institute
      Camperdown, New South Wales, Australia
  • 2008
    • University of Kentucky
      Lexington, Kentucky, United States
    • American Hospital Dubai
      Dubayy, Dubai, United Arab Emirates
  • 2006
    • North Carolina State University
      • Department of Statistics
      Raleigh, North Carolina, United States
  • 2000-2002
    • University of Minnesota Twin Cities
      • • School of Dentistry
      • • Rural Health Research Center (RHRC)
      Minneapolis, Minnesota, United States
  • 1999
    • University of California, San Francisco
      San Francisco, California, United States