Critical analysis of information: an epidemiologic perspective

A slideshow presentation discussing methods for critically evaluating medical literature and observational studies

CRITICAL ANALYSIS OF INFORMATION: AN EPIDEMIOLOGIC PERSPECTIVE PERINATAL PROFESSIONAL CONFERENCE APRIL 21ST, 2017 PAST AND PRESENT PRACTICE: CARING FOR MODERN FAMILIES KAREN SCHLIEP, PHD, MSPH, DIVISION OF PUBLIC HEALTH, DEPARTMENT OF FAMILY AND PREVENTIVE MEDICINE ©UNIVERSITY OF UTAH HEALTH, 2017 OUTLINE • Brief personal introduction • Example: "Impact of stress on women and their offspring" • How to critically evaluate the scientific literature • Ways for you to get involved in research - PRAMs - Systematic Review - Secondary Data Analysis @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 Personal Introduction ©UNIVERSITY OF UTAH HEALTH, 2017 A LITTLE ABOUT ME… • BA, English, Carleton College (1984-1988) • United States Peace Corps, Mali, West Africa (1990-1992) A LITTLE MORE ABOUT ME… • BS, Biology (Minor, Chemistry), RA in Ecology Lab (1994-1997) • Research Coordinator: University of Utah School of Medicine (20022012) • MSPH/PhD, University of Utah School of Medicine (2005-2011) • Postdoctoral Fellow, Epidemiology Branch, NICHD (2012-2015) AND HERE I AM NOW • Assistant Professor, University of Utah; Asia Campus (2015-2016); SLC Campus, Fall 2016 MY RESEARCH PASSION Interplay between reproductive/pregnancy disorders, modifiable exogenous factors (exposome), non-modifiable endogenous factors (genome) and chronic disease pathogenesis. MOTIVATION FOR MY RESEARCH Impact of stress on women and their offspring ©UNIVERSITY OF UTAH HEALTH, 2017 Introduction Schliep KC, et al. Perceived stress, reproductive hormones, and ovulatory function among premenopausal women: a prospective cohort study. Epidemiology. 2015;26:177-84. ©UNIVERSITY OF UTAH HEALTH, 2017 PERCEIVED PSYCHOLOGICAL STRESS • Psychological stress: individual perceives that external demands tax or exceed his or her ability to cope. • 23% of US women reporting high levels of stress and 43% saying that their stress levels have increased over the last 5 years. - Stress in America Survey 2012 @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 ACUTE AND CHRONIC STRESS @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 STRESS AND REPRODUCTION Stress HPA SAM α-amylase Blood Flow @uupublichealth #uupublichealth HPG Sex hormones ©UNIVERSITY OF UTAH HEALTH, 2017 OBJECTIVE To examine the association between repeated measurements of perceived stress, reproductive hormone levels, and ovulatory function @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 Methods Schliep KC, et al. Perceived stress, reproductive hormones, and ovulatory function among premenopausal women: a prospective cohort study. Epidemiology. 2015;26:177-84. ©UNIVERSITY OF UTAH HEALTH, 2017 STUDY POPULATION • BioCycle Study (2005-2007): Followed 259 women from western NY region for up to 2 cycles • Inclusion: - Ages 18-44 - Regularly menstruating • Exclusion: - Reproductive disorders - Self-reported obesity or underweight @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 HORMONE ASSESSMENT @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 HORMONE & ANOVULATION ASSESSMENT • Total E2, Progesterone, LH, FSH and sex hormone binding globulin (SHBG) were assayed. • Anovulation: ≤ 5ng/mL luteal progesterone and no LH peak at the mid- or late luteal visits. • 94% of the 259 participants completed 7 or 8 clinic visits per cycle. @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 STRESS ASSESSMENT Please rate your level of stress today: 1 = not stressful 2 = a little stressful 3 = very stressful @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 COVARIATE ASSESSMENT • Age, race, income, education level, sexual and reproductive history, and depression (Center for Epidemiologic Studies Depression Scale) were obtained at baseline using standard questionnaires. • At the end of the follow-up period, total percent body fat was measured using duel energy x-ray absorptiometry scans. @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 POTENTIAL CONFOUNDERS: CHRONIC Depression Age Menstrual Cycle Function Stress Race Bodyfat Parity @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 POTENTIAL CONFOUNDERS: TIME VARYING Exercise Time 2 • sleep • sexual activity • pain or antibiotic medication use • caffeine/alcohol/tobacco • total energy and fiber intake • diet score Depression Age Stress Time 1 Stress Time 3 Race Menstrual Cycle Function (e.g., Estrogen) Time 4 Bodyfat Final models were adjusted for age, race, percent body fat, depression and time-varying vigorous exercise. Parity @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 STATISTICAL ANALYSIS • Mixed models: to assess association between average daily stress, hormones and anovulation. • Case crossover analysis on women who experienced both an ovulatory and anovulatory cycle (n=24). - "The ultimate form of statistical adjustment for confounding by constant subject characteristics." ID 101, Cycle 1 Stressed? @uupublichealth #uupublichealth ID 101, Cycle 2 X Stressed? ©UNIVERSITY OF UTAH HEALTH, 2017 Results Schliep KC, et al. Perceived stress, reproductive hormones, and ovulatory function among premenopausal women: a prospective cohort study. Epidemiology. 2015;26:177-84. ©UNIVERSITY OF UTAH HEALTH, 2017 Daily Stress Tertile Total Population Low Moderate High 259 (100%) 87 (33.6%) 87 (33.6%) 85 (32.8%) 27.3 ± 8.2 28.3 ± 8.4 27.0 ± 8.0 26.5 ± 8.2 Caucasian 154 (59) 51 (33) 47 (31) 56 (36) African American 51 (20) 20 (39) 22 (43) 9 (18) Asian 37 (14) 10 (27) 12 (32) 15 (41) 29.5 ± 5.9 29.4 ± 5.9 29.1 ± 6.1 30.0 ± 5.8 0.95 5 (1, 9) 3 (0, 6) 4 (1, 8) 6 (3, 11) <0.001 9.9 (2.3, 20.0) 8.2 (0.7, 20.0) 10.7 (3.2, 17.5) 10.6 (2.8, 22.3) 0.46 0.08 (0.02, 3.5) 53.6 (17.8, 147.7) 0.08 (0.02, 3.9) 51.7 (16.4, 147.7) 0.07 (0.02, 2.3) 71.1 (23.3, 156.2) 0.14 (0.02, 4.2) 49.7 (16.7, 134.1) Participants [n (%)] Age (years) P-value 0.34 Race[n (%)] Percent body fat Depression score [median (IQR)] Vigorous exercise (min/day) [median (IQR)] Alcohol (g) [median (IQR)] Caffeine (mg/day) [median (IQR)] 0.27 0.49 0.69 PSS-14 (One Baseline) PSS-14 (One Baseline) PSS-4 (Four/Cycle x 2) 25.0 PSS-14 (One Baseline) PSS-4 (Four/Cycle x 2) Daily (Across 2 Cycles) 20.0 Percent Difference: High versus Low Stress (95% Confidence Interval) 15.0 10.0 5.0 0.0 -5.0 -10.0 -15.0 -20.0 -25.0 Estradiol Free Estradiol Luteal Progesterone LH FSH ANOVULATION Adjusted Odds Ratio: High vs Low Stress (95% Confidence Interval) 10 1 0.1 PSS-14 (One Baseline) PSS-4 (Four/Cycle x 2) Daily (Across 2 Cycles) STRESS AND SPORADIC ANOVULATION: CASE CROSSOVER ANALYSES Stressed/Stressed OR = 3.0; P = 0.24 Stressed/Anovulatory 8 6 Stressed/Ovulatory Not Stressed/Not Stressed 2 6 22 women with one ovulatory and one anovulatory cycle with complete daily stress measures Discussion Schliep KC, et al. Perceived stress, reproductive hormones, and ovulatory function among premenopausal women: a prospective cohort study. Epidemiology. 2015;26:177-84. ©UNIVERSITY OF UTAH HEALTH, 2017 SUMMARY • High daily perceived stress was significantly associated with lower total and free E2, LH, higher FSH, and an increased risk of sporadic anovulation @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 CONCLUSION • Recent, but not chronic, perceived stress is associated with alterations in reproductive hormone concentrations & sporadic anovulation • Potential for healthy interventions. @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 LONG-TERM AFFECTS OF TRAUMA/STRESS @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 How to critically evaluate the scientific literature ©UNIVERSITY OF UTAH HEALTH, 2017 https://theebmproject.wordpress.com/ @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 WHAT TO LOOK FOR IN OBSERVATIONAL STUDIES https://theebmproject.wordpress.com/ @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 WHAT TO LOOK FOR IN OBSERVATIONAL STUDIES https://theebmproject.wordpress.com/ @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 WHAT TO LOOK FOR IN OBSERVATIONAL STUDIES https://theebmproject.wordpress.com/ @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 WHAT TO LOOK FOR IN OBSERVATIONAL STUDIES Is selection bias present? - In a cohort study, are participants in the exposed and unexposed groups similar in all important respects except for the exposure? - In a case-control study, are cases and controls similar in all important respects except for the disease in question? Grimes and Schulz, Bias and causal associations in observational research. Lancet. 2002, 359(9302):248-52. @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 WHAT TO LOOK FOR IN OBSERVATIONAL STUDIES Is information bias present? - In a cohort study, is information about outcome obtained in the same way for those exposed and unexposed? - In a case-control study, is information about exposure gathered in the same way for cases and controls? Grimes and Schulz, Bias and causal associations in observational research. Lancet. 2002, 359(9302):248-52. @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 WHAT TO LOOK FOR IN OBSERVATIONAL STUDIES Is confounding present? - Could the results be accounted for by the presence of a factor-eg, age, smoking, sexual behaviour, diet- associated with both the exposure and the outcome but not directly involved in the causal pathway? Grimes and Schulz, Bias and causal associations in observational research. Lancet. 2002, 359(9302):248-52. @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 WHAT TO LOOK FOR IN OBSERVATIONAL STUDIES If the results cannot be explained by these three biases, could they be the result of chance? - What are the relative risk or odds ratio and 95% CI? - Is the difference statistically significant, and, if not, did the study have adequate power to find a clinically important difference? Grimes and Schulz, Bias and causal associations in observational research. Lancet. 2002, 359(9302):248-52. @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 WHAT TO LOOK FOR IN OBSERVATIONAL STUDIES If the results still cannot be explained away, then (and only then) might the findings be real and worthy of note. Grimes and Schulz, Bias and causal associations in observational research. Lancet. 2002, 359(9302):248-52. @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 Practice Time Abheiden et al. Hypertensive disorders of pregnancy appear not to be associated with Alzheimer's Disease Dementia and Cognitive Disorders. 2015;5:375-385. ©UNIVERSITY OF UTAH HEALTH, 2017 WHAT TO LOOK FOR IN OBSERVATIONAL STUDIES - What type of study is this? - Is selection bias present? - Is information bias present? - Is confounding present? - If the results cannot be explained by these 3 biases, could they be the result of chance? - What would you conclude re: evidence? @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 IS THERE ANY EPIDEMIOLOGIC EVIDENCE? • Only two studies, both conducted in the past year: Case control study Abheiden, Dement Geriatr Cogn Disord Extra 2015;5:375-385 IS THERE ANY EPIDEMIOLOGIC EVIDENCE? • Only two studies, both conducted in the past year: Theilen, Obstet Gynecol. 2016;128(2):238-244 Retrospective cohort of UPDB death records. 60,850 exposed (PIH) women matched to 123,140 unexposed women (no PIH) by age, year of childbirth (cohort effect), and parity. Risks of All-Cause and Cause-Specific Mortality Associated with Preeclampsia All Causes All Cancers Breast Cancer Colon Cancer Endocrine, nutritional and metabolic disease Causes of Death Diabetes Nervous System Parkinson's Disease 14.99 Alzheimer's Disease Circulatory System IHD Stroke External Causes Suicide -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 Mortality Hazard Rates • Hazard Ratios and 95% Confidence Intervals Estimated with Cox Regressions 8 9 10 11 Risk of Mortality due to Alzheimer's Disease by Hypertensive Pregnancy Status Exposure n (%) Unexposed 19 Eclampsia 1 Preeclampsia 16 HR (95% CI) 1.00 (0.03, 30.40) 4.02 (1.08, 14.99) Perfect Partnerships ©UNIVERSITY OF UTAH HEALTH, 2017 PUBLICALLY AVAILABLE DATASETS • Link between self-reported life stress(partner-related, traumatic, financial and emotiona development of preterm birth, small for gestational age, and preeclampsia. • Maternal Oral Health Prior to Conception and Adverse Birth Outcomes • How hospital interventions can impact breastfeeding initiation and success • Predictors of post partum depression in representative sample of Utah mothers. @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 SYSTEMATIC REVIEWS: PRISMA AND PROSPERO • PRISMA is an evidence-based minimum set of items for reporting in systematic reviews and meta-analyses. • PROSPERO is an international database of prospectively registered systematic reviews in health and social care, welfare, public health, education, crime, justice, and international development, where there is a health related outcome. http://www.prisma-statement.org/ https://www.crd.york.ac.uk/PROSPERO @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 SECONDARY DATA ANALYSIS OF EXISTING STUDIES WITH UTAH MOTHERS AND CHILDREN >80% enrollees from UT in both the EAGeR and ENDO Studies @uupublichealth #uupublichealth ©UNIVERSITY OF UTAH HEALTH, 2017 Questions? [e-mail] @schliepy ©UNIVERSITY OF UTAH HEALTH, 2017