#7 PH150A Feb3 Cohort Studies

#7 PH150A Feb3 Cohort Studies - Observational Study...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Observational Study Designs: Cohort Studies Barbara Abrams DrPH Barbara DrPH PH 150A: Introduction to Epidemiology Spring 2010 1 Optimal Maternal Weight Gain and Nutrition and Pregnancy Outcome: A Delicate Balance http://www.topnews.in/health/files/pregnancy-obese.jpg http://www.babble.com/CS/blogs/strollerderby/obese.jpg 2 Health Problems associated with Maternal Weight or Weight Gain Unhealthy baby size ?Preterm birth Neural tube defects Gestational Diabetes Pregnancy induced hypertension Complications during delivery Problems initiating successful breastfeeding Childhood obesity Long term obesity complicating the next pregnancy Long term obesity associated with numerous adverse health problems 3 Gestational Weight Gain and Five Outcomes, National Longitudinal Survey of Youth, 1979 Margerison C, Rehkopf D, Abrams B. in press, American Journal of Obstetrics and Gynecology, 2010 Distribution of Weight Gain during Pregnancy http://www.ottawa.ca/residents/health/families/planning/pregnanc y/healthy/weight_2_en-1.jpg Prevalence (%) of BMI ≥25 in U.S. Prevalence Women of Childbearing Age Over Time 50 45 40 35 30 25 20 15 10 5 0 age 20-29 age 30-39 NHE S 60-62 NHANE S 7174 NHANE S 2 76-80 NHANE S 3 88-94 6 http://blog.cleveland.com/health_i mpact/2009/05/medium_Weight_ guidelines_for_pregnant_women. jpg Objectives Introduce methods of prospective cohort studies and variations of this design When to use Pros and cons Ways to enhance study methods 8 Time Tunnel for Study Designs Survey: Cross Sectional Study PAST PRESENT 9 FUTURE Time Tunnel for Study Designs Survey: Cross Sectional Study Case Control Study PAST PRESENT FUTURE 10 Weight Retention in New Mothers: “After the Baby Comes” (ABC) Study School of Public Health, University of California, Berkeley and US Naval Medical Center, San Diego 11 Research Questions How much weight do women retain at various time points in the year after birth? What is the risk of becoming overweight or obese after pregnancy in a military population? Does active duty status modify the weight retention disparity previously seen in White and African American women? 12 Cohort Study Prospective cohort describes relationship between exposure and development of outcome over time exposure development Begin with study group of people without outcome Measure exposure Follow (can repeat exposure assessments ) Measure outcome Classify Compare 13 Time Tunnel for Study Designs Survey: Cross Sectional Study Prospective Cohort Study Case Control Study PAST PRESENT FUTURE 14 1. Select Study Group Source: Mothers of infants receiving well baby care at the Pediatric Clinic, Balboa Hospital, U.S. Naval Medical Center, San Diego, September 1997 to December 1999. Active duty women and wives of active duty servicemen. Consent form completed. 15 2. Measure exposure Baseline Postpartum Height, weight Active duty, rank Race, age, marital status, parity, education, income Recalled pre-pregnancy weight History of weight cycling Body image Weight Breast or bottle feeding Exercise Work Dieting practices Social support Stress Depression Financial insecurity Reported or recalled by woman on questionnaire Directly Measured 16 3. Follow the cohort Goal: to follow as many enrolled women as possible through the first year after birth 17 ABC Study Timeline and Data Collection Present Future Begin Measure Exposure Measure Outcome Clinic Questionnaires and Weight Baseline Questionnaire 3d 10d 12mo 2mo 4mo 6mo Exit Questionnaire 9mo Abstract prenatal records Birth Postpartum 18 Assess outcome: Measure weight at end of study to calculate BMI. by race and active duty Classify status. Compare outcome by exposure 19 Cohort Design Begin Measure/Classify Exposure Active Duty and Race ACTIVE DUTY: White African American Measure/Compare Outcome BMI ≥25 % BMI < 25 % BMI ≥ 25 % BMI < 25 % BMI ≥ 25 % BMI < 25 % BMI ≥ 25 % BMI < 25 % BMI ≥ 25 500 mothers enroll immediately after delivery all began pregnancy with BMI <25 NOT ACTIVE DUTY: White African American Present Future (12 months after birth) 20 Advantages of Cohort studies Studies disease development over time Investigator can plan/control methods of data collection Allows exposure to occur before disease. Allows before Provides better evidence for a cause-effect relationship than other observational designs 21 Disadvantages Can take a long time for results. Expensive. Not good for rare outcomes. 22 Must Follow Large Numbers of Healthy People for Sufficient Observations Example: Need to enroll and follow many women to get enough outcomes to reflect different weight outcomes in different raceethnic groups Only 1% of the Harvard cohort develops heart disease per year Remedy: Follow higher risk cohorts to reduce number of subjects needed Example: DES daughters, populations with high levels of exposure such as selected occupational groups 23 Error in Selecting Initial Sample (Response Bias) Example: Study participants different from non-participants (don’t reflect target population) Remedy: Make study participation easy Make multiple contacts Report response rate, compare non-responders to responders. The ABC staff recorded the relative bmi of women The bmi who said yes and no over a 2 week period 24 Loss to Follow-up Example: Exposure and disease differs in dropouts compared to those followed. Remedy: Select participants willing to complete study Obtain information to allow tracking later Keep in touch by phone/ mail Offer appropriate incentives Report % followed; compare characteristics to lost participants 25 Exposure Can Change Over Time Example: Diet, physical activity, sleep and smoking changes over time Women who report being on active duty may leave the military. Remedy: Reassess exposure status over time (if possible) 26 Definitions/Diagnosis Can Change Over Time Example: Definition of AIDS has changed several times since first cases identified Remedy: Use and maintain standard definitions during the study If choose to use new definition or criteria, reclassify the data for all study participants reclassify all 27 Surveillance Bias Study staff looks for the outcome more intensely in exposed than unexposed participants. Example: Were Staff more likely to make contact with and weigh women who appeared overweight or not overweight? Oral contraceptive users appeared to have more cervical cancer because they had more opportunity for screening. Remedy: Use identical and standard methods of assessing outcome in all participants. 28 Summary Advantages Studies outcome development over time Less bias in assessing exposure Best observational design for causeeffect Disadvantages Time consuming Expensive Need large numbers Exposures can change Exposures over time over Subject to problems Subject with response, loss to follow-up, and surveillance bias surveillance 29 Retrospective Cohort (Historical Cohort) Design Identify study group using existing records where data on exposure and outcome were prospectively recorded in the past Classify exposure based on records, and analyze subsequent outcome from records. 30 Time Tunnel for Study Designs Survey: Cross Sectional Study Prospective Cohort Study Retrospective Cohort Study Case Control Study PAST PRESENT FUTURE 31 Retrospective Cohort: Research Question Is pregnancy weight gain associated with overweight (BMI≥25) after pregnancy? Study Group: Participants in the ABC cohort. Exposure: prenatal weight gain recorded in the medical record. 32 Retrospective Cohort Design Measure Link to electronic perinatal database that collected data on gestational weight gain through each pregnancy Begin Classify Exposure Measure/Compare Outcome % BMI <25 % BMI ≥25 % BMI <25 % BMI ≥25 1999 PRESENT 33 Select all ABC mothers at first prenatal visit High prenatal weight gain Low prenatal weight gain 1997 1998 PAST FOLLOW-UP Retrospective Cohort Advantages: Because data already collected, less expensive, quicker Other advantages of cohort study Disadvantages Recorded data may not be of uniform quality or available at all 34 Nested Case Control Study The case control study is “nested” within a The cohort Identify the cases and controls from a cohort where data on exposure and outcome were prospectively collected Classify exposure of the cases and controls based on the already-collected cohort data 35 Time Tunnel for Study Designs Survey: Cross Sectional Study Prospective Cohort Study Retrospective Cohort Study [ Case Control Study Nested Case Control Study 36 PAST PRESENT FUTURE Nested Case Control: Research Question Is maternal plasma leptin (a polypeptide hormone Is leptin related to obesity) early in pregnancy associated with overweight (BMI≥25) after pregnancy? Study Group: Participants in the ABC cohort one year after birth Cases: women with BMI≥25 Controls: women with BMI <25 Exposure: analysis of maternal plasma samples frozen and stored at 28 weeks gestation 37 Nested Case-Control Design Obtain stored prenatal plasma samples; thaw, assay for leptin, Classify and Compare Exposure Begin with assessing outcome in ABC cohort Select Cases and Controls Leptin high Leptin low Controls: mothers with BMI <25 before and 1 year after pregnancy Cases: with BMI <25 before pregnancy and BMI ≥25 1 year after >25 pregnancy Leptin low Leptin high Past Present 38 Nested Case Control Study Same advantages of case control studies– quick, less expensive, good for rare outcomes Exposure data collected and recorded before the outcome occurred Know underlying population—possibly cases and controls more representative Depends on recorded data or stored data 39 Variations on Cohort Studies Prospective Cohort: enroll in present, follow into future Retrospective Cohort: same direction, but all data are already collected for another purpose Nested Case Control: a case control study performed using data collected for a prospective cohort study 40 http://www.birthunlimited.ca/worthWeight.jpg Websites for Cohort Studies The Coronary Artery Risk Development in Young Adults (CARDIA) Study http://www.cardia.dopm.uab.edu/o_brde.htm Havard Nurses Health study http://www.nurseshealthstudy.org Study of Women Across the Nation (SWAN) http://www.edc.gsph.pitt.edu/swan/ 42 ...
View Full Document

Ask a homework question - tutors are online