P C O'Brien

Mayo Clinic - Rochester, Рочестер, Minnesota, United States

Are you P C O'Brien?

Claim your profile

Publications (356)2373.56 Total impact

  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Prior to a joint meeting of the Neurodiab Association and International Symposium on Diabetic Neuropathy held in Toronto, Ontario, Canada, 13-18 October 2009, Solomon Tesfaye, Sheffield, UK, convened a panel of neuromuscular experts to provide an update on polyneuropathies associated with diabetes (Toronto Consensus Panels on DPNs, 2009). Herein, we provide definitions of typical and atypical diabetic polyneuropathies (DPNs), diagnostic criteria, and approaches to diagnose sensorimotor polyneuropathy as well as to estimate severity. Diabetic sensorimotor polyneuropathy (DSPN), or typical DPN, usually develops on long-standing hyperglycaemia, consequent metabolic derangements and microvessel alterations. It is frequently associated with microvessel retinal and kidney disease—but other causes must be excluded. By contrast, atypical DPNs are intercurrent painful and autonomic small-fibre polyneuropathies. Recognizing that there is a need to detect and estimate severity of DSPN validly and reproducibly, we define subclinical DSPN using nerve conduction criteria and define possible, probable, and confirmed clinical levels of DSPN. For conduct of epidemiologic surveys and randomized controlled trials, it is necessary to pre-specify which attributes of nerve conduction are to be used, the criterion for diagnosis, reference values, correction for applicable variables, and the specific criterion for DSPN. Herein, we provide the performance characteristics of several criteria for the diagnosis of sensorimotor polyneuropathy in healthy subject- and diabetic subject cohorts. Also outlined here are staged and continuous approaches to estimate severity of DSPN. Copyright © 2011 John Wiley & Sons, Ltd.
    Full-text · Article · Oct 2011 · Diabetes/Metabolism Research and Reviews
  • [Show abstract] [Hide abstract]
    ABSTRACT: To establish a set of normal values for breast size in children up to two years of age, and to assess the effects of gender, gestational age, and type of feeding and growth parameters on breast size. Prospective cohort study over 20 months The Mayo Clinic Rochester, MN, neonatal nursery and Community Pediatrics Clinic. Well term infants and children through two years of age. Measurement of breast size based on previously described methods to yield a figure called the breast unit. Transformation of breast unit size into percentiles according to age. The 50(th) to 99(th) percentiles for 810 healthy term Caucasian infant visits were calculated. The 50th breast unit percentile was similar in males and females at birth and declined with age following a quadratic relationship for females and a cubic relationship for males. Breast tissue in female infants remained larger and persisted longer. Palpable breast tissue was still present in 45.2% of male and 61.6% of female visits after 10 months of age. At age 18 months, 5% of girls had a breast size unit greater than 2.88 cm(2) and 5% of boys had a breast size unit greater than 1.00 cm(2). These data allow creation of normal standards of breast size for age, which could provide a future clinical tool to assist clinicians in the evaluation of early childhood breast enlargement in similar populations.
    No preview · Article · Oct 2010 · Journal of pediatric and adolescent gynecology
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The purpose was to test whether physicians can validly and reproducibly diagnose diabetic sensorimotor polyneuropathy (DSPN). Twelve physicians assessed 24 patients with diabetes mellitus (DM) on consecutive days (576 examinations) with physical features and voice disguised. Results were compared to gold standard 75% group diagnosis (dx) and a nerve conduction score (Sigma5 NC nds). Masking of patients was achieved. Reproducibility measured by the kappa coefficient and compared to Sigma5 NC nd varied considerably among physicians: median and ranges: signs 0.8 (0.32-1.0); symptoms 0.79 (0.36-1.0), and diagnoses 0.47 (0.33-0.84), both low and high scores indicating poor performance. There was substantial agreement between 75% group dx and confirmed NC abnormality (abn). As compared to Sigma5 NC, individual physicians' clinical dx was excessively variable and frequently inaccurate. Study physician dx from signs and symptoms were excessively variable, often overestimating DSPN. Specific approaches to improving clinical proficiency should be tested.
    Full-text · Article · Aug 2010 · Muscle & Nerve
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Although age-related declines in dehydroepiandrosterone sulfate (DHEAS) and testosterone are associated with cardiovascular risk, it remains to be determined whether replacement of these hormones improves cardiovascular risk factors. This study sought to determine the effect of long-term replacement of dehydroepiandrosterone (DHEA) in elderly men and women and testosterone in elderly men on lipid and lipoprotein concentrations and particle sizes. A 2-yr randomized, placebo-controlled, double-blind study was conducted in 87 elderly men with low levels of DHEAS and bioavailable testosterone and 57 elderly women with low levels of DHEAS. Among elderly men, 29 received DHEA (75 mg/d), 27 received testosterone (5 mg/d), and 31 received placebo. Among the elderly women, 27 received DHEA (50 mg/d), and 30 received placebo. Baseline lipoprotein profiles in the elderly were compared to healthy younger participants. Low-density lipoprotein (LDL) and high-density lipoprotein (HDL) particle sizes and concentrations were quantified using nuclear magnetic resonance spectroscopy. The elderly had higher concentrations of total cholesterol, triglycerides, LDL cholesterol, total LDL particles, and small, dense LDL particles than the young. In men, neither DHEA nor testosterone affected LDL or HDL particle concentrations. In women, DHEA reduced HDL cholesterol [median difference (95% confidence intervals), -5.0 (-8.0, -2.0) mg/dl; P = 0.002] and the number of large HDL particles [-1.0 (-1.8, -0.2) micromol/liter; P = 0.003]. Long-term DHEA and testosterone had no significant effect on plasma lipoproteins in elderly men, but elderly women showed a lowering of the large HDL particles that may have potential adverse clinical implications.
    Preview · Article · Feb 2010 · The Journal of Clinical Endocrinology and Metabolism
  • [Show abstract] [Hide abstract]
    ABSTRACT: It is unclear whether adipocyte size or body fat distribution is most strongly linked to the metabolic complications of obesity. Our objective was to test whether adipocyte size better predicts metabolic characteristics of obesity than body composition. We analyzed the relationship between metabolic and anthropometric data collected from 432 largely Caucasian research volunteers (264 women) participating in studies conducted in the Mayo General Clinical Research Center between 1995 and 2008. Metabolic variables included fasting plasma glucose, insulin, and triglyceride concentrations. Anthropometric variables included body composition, fat distribution, and sc abdominal and femoral adipocyte size. Using both univariate and multivariate regression analysis, fasting triglyceride in both men and women was best predicted by computed tomography of visceral fat area. Fasting insulin concentrations were best predicted by sc abdominal fat area in women (r(2) = 0.40; P < 0.01) and body mass index in men (r(2) = 0.53; P < 0.0001); adipocyte size did not contribute independently. In men, fasting glucose concentrations were predicted by femoral adipocyte size (partial r(2) = 0.07; P = 0.002), body mass index (partial r(2) = 0.03; P = 0.07), and age (partial r(2) = 0.02; P = 0.06). In women, fasting glucose was predicted by abdominal sc fat area (partial r(2) = 0.12; P < 0.0001) and age (partial r(2) = 0.03; P = 0.01). Our hypothesis that adipocyte size is the best predictor of metabolic characteristics was not supported in this population. The alternative explanation is that fat mass and body fat distribution have more influence on metabolic responses than adipocyte size.
    No preview · Article · Nov 2009 · The Journal of Clinical Endocrinology and Metabolism
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Reductions in insulin sensitivity in conjunction with muscle mitochondrial dysfunction have been reported to occur in many conditions including aging. The objective was to determine whether insulin resistance and mitochondrial dysfunction are directly related to chronological age or are related to age-related changes in body composition. Twelve young lean, 12 young obese, 12 elderly lean, and 12 elderly obese sedentary adults were studied. Insulin sensitivity was measured by a hyperinsulinemic-euglycemic clamp, and skeletal muscle mitochondrial ATP production rates (MAPRs) were measured in freshly isolated mitochondria obtained from vastus lateralis biopsy samples using the luciferase reaction. Obese participants, independent of age, had reduced insulin sensitivity based on lower rates of glucose infusion during a hyperinsulinemic-euglycemic clamp. In contrast, age had no independent effect on insulin sensitivity. However, the elderly participants had lower muscle MAPRs than the young participants, independent of obesity. Elderly participants also had higher levels inflammatory cytokines and total adiponectin. In addition, higher muscle MAPRs were also noted in men than in women, whereas glucose infusion rates were higher in women. The results demonstrate that age-related reductions in insulin sensitivity are likely due to an age-related increase in adiposity rather than a consequence of advanced chronological age. The results also indicate that an age-related decrease in muscle mitochondrial function is neither related to adiposity nor insulin sensitivity. Of interest, a higher mitochondrial ATP production capacity was noted in the men, whereas the women were more insulin sensitive, demonstrating further dissociation between insulin sensitivity and muscle mitochondrial function.
    Full-text · Article · Oct 2009 · Diabetes
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Aging, low dehydroepiandrosterone (DHEA), and testosterone are associated with increased adiposity and metabolic risk. Treatment with these hormones may improve these abnormalities. The objective of the study was to determine effects of aging, DHEA, or testosterone replacement on adiposity, meal fat partitioning, and postabsorptive lipolysis. This was a cross-sectional, 2-yr, double-blind, randomized, placebo-controlled trial. The study was conducted in the general community. Elderly women and men (>or=60 yr) with low DHEA sulfate (women and men) and bioavailable testosterone (men) concentrations and young adults. Thirty elderly women each received 50 mg DHEA or placebo daily for 2 yr. Thirty elderly men received 75 mg DHEA, 29 received 5 mg testosterone (patch), and 32 received placebo daily for 2 yr. Thirty young women and 32 young men served as controls. In vivo measures of meal fat storage into sc fat, postabsorptive lipolysis, and regional adiposity at baseline and after treatment. At baseline, the elderly had more body fat, greater systemic lipolysis (women, P = 0.0003; men, P < 0.0001) adjusted for resting energy expenditure, greater meal fat oxidation (women, P = 0.026; men, P = 0.0025), and less meal fat storage in sc fat (women, P = 0.0139; men, P= 0.0006). Although testosterone treatment increased meal fat storage into upper- vs. lower-body fat in elderly men, neither hormone affected regional adiposity, meal fat oxidation, or systemic lipolysis. Aging, in the context of low DHEA sulfate (women and men) and bioavailable testosterone (men) concentrations, is associated with changes in meal fat partitioning and postabsorptive lipolysis that are not corrected by DHEA and only partly corrected by testosterone replacement.
    Preview · Article · Jun 2009 · The Journal of Clinical Endocrinology and Metabolism
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: To investigate the combined ability of hippocampal volumes, 1H magnetic resonance spectroscopy (MRS) metabolites, and cerebrovascular disease to predict the risk of progression to dementia in mild cognitive impairment (MCI). We identified 151 consecutively recruited subjects with MCI from the Mayo Clinic Alzheimer's Disease Research Center and Patient Registry who underwent MRI and 1H MRS studies at baseline and were followed up with approximately annual clinical examinations. A multivariable proportional hazards model that considered all imaging predictors simultaneously was used to determine whether hippocampal volumes, posterior cingulate gyrus 1H MRS metabolites, white matter hyperintensity load, and presence of cortical and subcortical infarctions are complementary in predicting the risk of progression from MCI to dementia. Seventy-five subjects with MCI progressed to dementia by last follow-up. The model that best predicted progression to dementia included age, sex, hippocampal volumes, N-acetylaspartate (NAA)/creatine (Cr) on 1H MRS, and cortical infarctions. Based on age- and sex-adjusted Kaplan-Meier plots, we estimated that by 3 years, 26% of the MCI patients with normal hippocampal volumes, NAA/Cr ratios >1 SD, and no cortical infarctions will progress to dementia, compared with 78% of the MCI patients with hippocampal atrophy, low NAA/Cr (< or =1 SD), and cortical infarction. Multiple magnetic resonance (MR) markers of underlying dementia pathologies improve the ability to identify patients with prodromal dementia over a single MR marker, supporting the concept that individuals with multiple brain pathologies have increased odds of dementia compared with individuals with a single pathology.
    Full-text · Article · Apr 2009 · Neurology
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: We investigated age and sex effects and determined whether androgen replacement in elderly individuals (> or = 60 yr) could augment protein synthesis. Thirty young men and 32 young women (18-31 yr) were studied once, whereas 87 elderly men were studied before and after 1 yr of treatment with 5 mg/day testosterone (T), 75 mg/day dehydroepiandrosterone (DHEA), or placebo (P); and 57 elderly women were studied before and after 1 yr of treatment with 50 mg/day DHEA or P. [(15)N]Phenylalanine and [(2)H(4)]tyrosine tracers were infused, with measurements in plasma and vastus lateralis muscle. Whole-body protein synthesis per fat-free mass and muscle protein fractional synthesis rate (FSR) were lower in elderly than in young individuals (P<0.001), not significantly affected by hormone treatments, and higher in women than in men (P<0.0001), with no sex x age interaction. In regression analyses, peak O2 consumption (VO2peak), resting energy expenditure (REE), and sex were independently associated with muscle FSR, as were VO2peak, REE, and interactions of sex with insulin-like growth factor-II and insulin for whole-body protein synthesis. Women maintain higher protein synthesis than men across the lifespan as rates decline in both sexes, and neither full replacement of DHEA (in elderly men and women) nor partial replacement of bioavailable T (in elderly men) is able to amend the age-related declines.
    Preview · Article · Oct 2008 · The FASEB Journal
  • Source
    Peng Huang · Robert F Woolson · Peter C O'Brien
    [Show abstract] [Hide abstract]
    ABSTRACT: O'Brien (Biometrics 1984; 40:1079-1087) introduced a rank-sum-type global statistical test to summarize treatment's effect on multiple outcomes and to determine whether a treatment is better than others. This paper presents a sample size computation method for clinical trial design with multiple primary outcomes, and O'Brien's test or its modified test (Biometrics 2005; 61:532-539) is used for the primary analysis. A new measure, the global treatment effect (GTE), is introduced to summarize treatment's efficacy from multiple primary outcomes. Computation of the GTE under various settings is provided. Sample size methods are presented based on prespecified GTE both when pilot data are available and when no pilot data are available. The optimal randomization ratio is given for both cases. We compare our sample size method with the Bonferroni adjustment for multiple tests. Since ranks are used in our derivation, sample size formulas derived here are invariant to any monotone transformation of the data and are robust to outliers and skewed distributions. When all outcomes are binary, we show how sample size is affected by the success probabilities of outcomes. Simulation shows that these sample size formulas provide good control of type I error and statistical power. An application to a Parkinson's disease clinical trial design is demonstrated. Splus codes to compute sample size and the test statistic are provided.
    Full-text · Article · Jul 2008 · Statistics in Medicine

  • No preview · Article · Jul 2008 · Alzheimer's and Dementia
  • Barbara Tilley · Peng Huang · Peter C. O'Brien

    No preview · Chapter · Jun 2008
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: We tested if rates of brain atrophy accelerate in individuals with amnestic mild cognitive impairment (aMCI) as they progress to typical late onset Alzheimer disease (AD). We included comparisons to subjects with aMCI who did not progress (labeled aMCI-S) and also to cognitively normal elderly subjects (CN). We studied 46 subjects with aMCI who progressed to AD (labeled aMCI-P), 46 CN, and 23 aMCI-S. All subjects must have had three or more serial MRI scans. Rates of brain shrinkage and ventricular expansion were measured across all available serial MRI scans in each subject. Change in volumes relative to the point at which subjects progressed to a clinical diagnosis of AD (the index date) was modeled in aMCI-P. Change in volumes relative to age was modeled in all three clinical groups. In aMCI-P the change in pre to post index rate (i.e., acceleration) of ventricular expansion was 1.7 cm(3)/year, and acceleration in brain shrinkage was 5.3 cm(3)/year. Brain volume declined and ventricular volume increased in all three groups with age. Volume changes decelerated with increasing age in aMCI-P, and to a lesser extent in aMCI-S, but were linear in the matched CN. Among all subjects with aMCI, rates of atrophy were greater in apolipoprotein E epsilon 4 carriers than noncarriers. Rates of atrophy accelerate as individuals progress from amnestic mild cognitive impairment (aMCI) to typical late onset Alzheimer disease (AD). Rates of atrophy are greater in younger than older subjects with aMCI who progressed to AD and subjects with aMCI who did not progress. We did not find that atrophy rates varied with age in 70- to 90-year-old cognitively normal subjects.
    Full-text · Article · Jun 2008 · Neurology
  • Peter C. O'Brien · Barbara C. Tilley · Peter J. Dyck

    No preview · Chapter · Dec 2007
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Magnetic resonance (MR)-based volume measurements of atrophy are potential markers of disease progression in patients with amnestic mild cognitive impairment (MCI) and Alzheimer's disease (AD). Longitudinal changes in (1)H MR spectroscopy ((1)H MRS) metabolite markers have not been characterized in MCI subjects. Our objective was to determine the longitudinal (1)H MRS metabolite changes in patients with MCI, and AD, and to compare (1)H MRS metabolite ratios and ventricular volumes in tracking clinical disease progression in AD. The neuronal integrity marker N-acetylaspartate/creatine ratio declined in MCI and AD patients compared to cognitively normal elderly. The change in (1)H MRS metabolite ratios correlated with clinical progression about as strongly as the rate of ventricular expansion, suggesting that (1)H MRS metabolite ratios may be useful markers for the progression of AD. Choline/creatine ratio declined in stable MCI, compared to converter MCI patients and cognitively normal elderly, which may be related to a compensatory mechanism in MCI patients who did not to progress to AD.
    Full-text · Article · Oct 2007 · Neurobiology of aging
  • [Show abstract] [Hide abstract]
    ABSTRACT: Assessing clinimetric performance of diabetic sensorimotor polyneuropathy (DSPN) end points in single and multicenter trials. Assessed were placebo-treated patients with DSPN in the Viatris and Eli Lilly trials and an epidemiologic cohort. Test reproducibility in clinical trial cohorts (r(I) approximately 0.7-0.85) approached that in the epidemiologic cohort (r(I) approximately 0.85-0.95). Associations between pairs of end points explained <10% of the variability of data (sometimes 15-35%), being higher in the epidemiologic cohort and the Viatris trial than in the Lilly trial. Most end points did not show monotonic worsening over 4 years. However, sural nerve amplitude and peroneal motor conduction velocity did. A nerve conduction score (Sigma 5 NC nds [5 attributes of nerve conduction expressed as normal deviates]) did not show monotonic worsening in established DSPN. In the epidemiologic cohort followed for 9.5 years, monotonic worsening of small magnitude occurred for sural amplitude, vibration detection threshold, and especially for composite quantitative sensation. The main reason why it is difficult to demonstrate monotonic worsening of neuropathic end points appears to be a very slow worsening of DSPN, a placebo effect for symptoms and signs, and measurement noise. Demonstrating disease progression in controlled trials of DSPN is more likely when 1) patients with developing rather than established DSPN are selected, 2) type 1 diabetic patients are preferentially recruited, 3) patients are selected who cannot or will not achieve ideal glycemic control, 4) end points chosen are known to show monotonic worsening, and 5) a restricted number of centers and expert examiners (trained, certified, using standard approaches, and reference values and interactive surveillance of tests) are used.
    No preview · Article · Oct 2007 · Diabetes care
  • F John Service · Peter C O'Brien
    [Show abstract] [Hide abstract]
    ABSTRACT: To assess the influences of a wide variety of glucose variables on hemoglobin A1c (A1C). The Diabetes Control and Complications Trial database, restricted to volunteers whose 7-point daily capillary glucose profiles were complete in >or=80% of quarterly collections and who were in the study for >or=4 years, was used for analysis. Regression analyses were done to develop an equation for estimating A1C based on concurrent and prior mean blood glucose (MBG) values. The multivariate coefficient of determination (R2) was calculated for MBG, mean postprandial blood glucose, mean preprandial blood glucose, digestive glycemia, interdigestive glycemia, individual time points of the 7-point glucose profile, range of blood glucose, SD of blood glucose, M-value, and mean amplitude of glycemic excursions in relationship to A1C. By using regression analysis, the correlation between A1C and MBG within each study subject was determined. The most accurate prediction of A1C was obtained from the concurrent MBG. With use of univariate analysis, all glucose variables correlated significantly with concurrent A1C, the strongest correlation occurring with MBG. In multivariate analysis, the primary predictor of A1C was MBG; all other glucose variables added nothing to the models. Within-subject correlations between MBG and A1C showed considerable variation. A1C correlates best with MBG derived from 7-point-daily capillary glucose profiles. The influences of glucose measured at specific time points during the day or various measures of glucose variability on A1C are less than that of MBG. Within the limitations of the intermittent glucose determinations, wide variations in the relationship of MBG to A1C among and within patients with type 1 diabetes remain unexplained.
    No preview · Article · Jul 2007 · Endocrine Practice
  • Yasmin Jayasinghe MBBS · Ruth Cha · Peter O'Brien

    No preview · Article · Apr 2007 · Journal of Pediatric and Adolescent Gynecology
  • Source

    Preview · Article · Feb 2007 · Diabetes Care
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: It is not known whether rigorous intraoperative glycemic control reduces death and morbidity in cardiac surgery patients. To compare outcomes of intensive insulin therapy during cardiac surgery with those of conventional intraoperative glucose management. A randomized, open-label, controlled trial with blinded end point assessment. Tertiary care center. Adults with and without diabetes who were undergoing on-pump cardiac surgery. The primary outcome was a composite of death, sternal infections, prolonged ventilation, cardiac arrhythmias, stroke, and renal failure within 30 days after surgery. Secondary outcome measures were length of stay in the intensive care unit and hospital. Patients were randomly assigned to receive continuous insulin infusion to maintain intraoperative glucose levels between 4.4 (80 mg/dL) and 5.6 mmol/L (100 mg/dL) (n = 199) or conventional treatment (n = 201). Patients in the conventional treatment group were not given insulin during surgery unless glucose levels were greater than 11.1 mmol/L (>200 mg/dL). Both groups were treated with insulin infusion to maintain normoglycemia after surgery. Mean glucose concentrations were statistically significantly lower in the intensive treatment group at the end of surgery (6.3 mmol/L [SD, 1.6] [114 mg/dL {SD, 29}] in the intensive treatment group vs. 8.7 mmol/L [SD, 2.3] [157 mg/dL {SD, 42}] in the conventional treatment group; difference, -2.4 mmol/L [95% CI, -2.8 to -1.9 mmol/L] [-43 mg/dL {CI, -50 to -35 mg/dL}]). Eighty two of 185 patients (44%) in the intensive treatment group and 86 of 186 patients (46%) in the conventional treatment group had an event (risk ratio, 1.0 [CI, 0.8 to 1.2]). More deaths (4 deaths vs. 0 deaths; P = 0.061) and strokes (8 strokes vs. 1 strokes; P = 0.020) occurred in the intensive treatment group. Length of stay in the intensive care unit (mean, 2 days [SD, 2] vs. 2 days [SD, 3]; difference, 0 days [CI, -1 to 1 days]) and in the hospital (mean, 8 days [SD, 4] vs. 8 days [SD, 5]; difference, 0 days [CI, -1 to 0 days]) was similar for both groups. This single-center study used a composite end point and could not examine whether outcomes differed by diabetes status. Intensive insulin therapy during cardiac surgery does not reduce perioperative death or morbidity. The increased incidence of death and stroke in the intensive treatment group raises concern about routine implementation of this intervention.
    Full-text · Article · Feb 2007 · Annals of internal medicine

Publication Stats

27k Citations
2,373.56 Total Impact Points


  • 1976-2011
    • Mayo Clinic - Rochester
      • • Department of Surgery
      • • Department of Health Science Research
      • • Department of Neurology
      • • Department of Hematology
      • • Department of Urology
      • • Department of Psychiatry & Psychology
      Рочестер, Minnesota, United States
  • 1982-2008
    • Mayo Foundation for Medical Education and Research
      • • Department of Neurology
      • • Department of Medicine
      Рочестер, Michigan, United States
  • 2007
    • Medical University of South Carolina
      • Division of Biostatistics and Epidemiology
      Charleston, South Carolina, United States
  • 1999
    • University of Vermont
      Burlington, Vermont, United States
  • 1994-1999
    • University of Guam
      Mangilao Village, Mangilao, Guam
  • 1998
    • Tri-Service General Hospital
      T’ai-pei, Taipei, Taiwan
  • 1991
    • John Hunter Hospital
      New Lambton, New South Wales, Australia
  • 1985
    • Dana-Farber Cancer Institute
      Boston, Massachusetts, United States
    • Rutgers New Jersey Medical School
      Newark, New Jersey, United States
  • 1980-1984
    • Baylor College of Medicine
      • Department of Obstetrics and Gynecology
      Houston, Texas, United States
    • Washington University in St. Louis
      • Department of Psychiatry
      San Luis, Missouri, United States
  • 1979
    • National Cancer Institute (USA)
      베서스다, Maryland, United States
  • 1974
    • University of Minnesota Rochester
      Рочестер, Minnesota, United States