national health and nutrition examination survey – oral health components laurie barker, msph...

29
National Health and Nutrition Examination Survey – Oral Health Components Laurie Barker, MSPH Mathematical Statistician Surveillance, Investigation and Research Team Division of Oral Health National Center for Chronic Disease Prevention and Health Promotion Office of Noncommunicable Diseases, Injury and Environmental Health [email protected]

Upload: kelley-lester

Post on 17-Dec-2015

216 views

Category:

Documents


1 download

TRANSCRIPT

National Health and Nutrition Examination Survey – Oral Health Components

Laurie Barker, MSPHMathematical Statistician

Surveillance, Investigation and Research TeamDivision of Oral Health

National Center for Chronic Disease Prevention and Health PromotionOffice of Noncommunicable Diseases, Injury and Environmental Health

[email protected]

Workshop Overview

• LectureI. IntroductionII. NHANES OverviewIII. NHANES Data ResourcesIV. Future Developments in NHANES

• Hands-on WorkshopV. Orientation to NCHS NHANES Web siteVI. Orientation to DRC tutorial materialsVII. Practice Examples (in SAS and SUDAAN)

I. Introduction

– CDC’s organization (oral health, NHANES) and relationship to NIDCR and other Federal agencies, and to universities and state governments

– Surveillance vs. Research

I. Introduction: CDC organization

– Within CDC• DOH

– located within the National Center for Chronic Disease Prevention and Health Promotion

– Funds 19 states to build oral health infrastructure– Funds part of NHANES, along with NIDCR

• Division of HANES– located within the National Center for Health Statistics– Designs, manages contract to operate NHANES– produces public and restricted data sets– Collaborates with other NCHS divisions to link NHANES to other

data sets– Publishes a wide variety of reports on NHANES

I. Introduction: CDC organization

– Outside of CDC• Department of Health and Human Services

– NIH/NIDCR– CDC– CMS– HRSA– IHS– SAMSHA– …

I. Introduction

– Surveillance vs. Research• Main purpose of surveillance is to monitor trends in

– known public health problems and related risk factors– Identify emerging problems– Broad picture of trends in national populations– Sentinel surveillance

» monitors rarer problems in higher risk populations» Or problems in populations representative of a smaller

area (county, city or specific population)

• Survey-based Surveillance at the national level is constrained by costs and burden to participants

I. Introduction

– Surveillance vs. Research• Changes in trends or new problems identified through

surveillance ideally trigger focused research to explain the observed change

• Surveillance data sets are often cross-sectional– Few truly longitudinal surveys– Limitations in the types of research that can be conducted with

surveillance data

– New developments• linkage among surveys, linkage with external data (CMS,

environmental) may support more research• Match rate could be a limiting issue

II. NHANES overview

– History of NHANES– NHANES compared to other national surveys– NHANES Oral Health Components• 2009-2010• Previous NHANES and precursor National Health

Examination Survey

– Survey design and analysis considerations• Statistical methods• Software

II. NHANES Overview: History of NHANES

• National Health Examination Surveys– 1959-62, 1963-65, 1966-1970

• National Health and Nutrition Examination Surveys– 1971-1974, 1988-1995– 1999+ Continuous NHANES

• Hispanic HANES 1986-1987

National Health Survey

NHES I 1959-1962 – 18-79y

NHES II 1963-1965 – 6-11y

NHES III 1966-1970 – 12-17y

NHANES I 1971-1974 – 1-74y

NIDR/OHSC I 1979-1980 5-17y

HHANES 1982-1984 6m-74y

NIDR/OHSC II 1986-1987 5-17y

NHANES III 1988-1994 2m+

NHANES 1999-2004 2y+

NHANES 2005-2008 5y+

NHANES 2009-2010 3-19y, 30y+

NHIS 1963-2010

NHAMCS 1992-2007

MEPS 1996-2008

State-based Health SurveyBSS - 3rd grade, K,1,2, HeadStart, older adults

BRFSS 1999, 2002, 2004, 2006, 2008, 2010 – 18y+Some states in other years from 1995

PRAMS 1990-1995, 1996-1999, 2000-2003, 2004-2008, 2009+

NSCH 2003, 2007, 2011 – 0-17yNSCSHCN 2001, 2005-2006, 2009-2010 – 0-17y

YRBSS 1991-2009, odd years

II. NHANES Overview: NHANES compared to other national surveys

• NHANES– National Household - questionnaire, exam, lab

• NHIS– National Household – questionnaire

• MEPS– National Household – questionnaire focused on costs

• BRFSS – State Telephone (multimodal now) – questionnaire– National estimates from combined states

• YRBSS– Paper survey– Separate National and State samples

II. NHANES Overview: NHANES Oral Health Components

– 2009-2010 – dental hygienist raters• Caries – tooth surface level, limited age range• Periodontal disease – full mouth

– 2005-2008 – medical technician raters• Caries – person level, limited age range

– 1999-2004 – dentist raters• Caries and sealants - Tooth surface level • Root caries, Fluorosis - Tooth level• Periodontal disease – 2 quadrants, facial/buccal sites

– Previous NHANES and precursor National Health Examination Survey – dentist raters• Caries, other measures of oral health, some out-of-date

II. NHANES Overview:Survey design and analysis considerations

• Statistical methods• Software

II. NHANES Overview:Survey design and analysis considerations

• Statistical methods– Complex sample survey design requires methods that

account for clustering due to survey design, in addition to any other clustering• Typically, design-based analyses have been recommended –

parameter estimation limited by PSUs (~15 per 2 years)• NCHS (and others) exploring other methods to account for

clustering, increase yield of information from survey data

– Sampling fraction, oversampling, non-response• NCHS provides weights that should be used• Little information to construct special weights due to disclosure

protections

II. NHANES Overview:Survey design and analysis considerations

• Software– SUDAAN (or WESVAR), historically the

recommended software due to limited procedures for complex samples in other software

– SAS, Stata, SPSS, R• All now have procedures to account for survey design • Point estimates should be the same from all packages• Variance estimates may differ due to choices each

package has made in methods to account for clustering– Differences probably less than 0.01?

II. NHANES Overview:Survey design and analysis considerations

• Key references– Heeringa, Steven G, Brady T. West and Patricia

Berglund. 2010. Applied Survey Data Analysis. Chapman & Hall, CRC, New York.

– Korn, Edward L. and Barry Graubard. 1999. Analysis of Health Surveys. John Wiley, New York.

– Groves, Robert M., Floyd J. Fowler, Jr., Mick P. Cooper, James M. Lepkowski, Eleanor Singer, Roger Tourangeau. 2004. Survey Methodology, 2nd edition. John Wiley, New York, 2009.

III. NHANES data resources

– NCHS NHANES Tutorial– Dental, Oral and Craniofacial Data Resource Center

(CDC and NIDCR funded project)• Current online resources• Tutorial materials in development

– Restricted data, linked data and Research Data Centers operated by NCHS or Census

III. NHANES data resources– NCHS NHANES Tutorial

http://www.cdc.gov/nchs/tutorials/index.htm • Excellent tutorial, materials being added regularly• Not specific to oral health, but good resources for

– Dietary interview data– Environmental chemical data

– Variable crosswalks• Cross year variable table

http://www.cdc.gov/nchs/nhanes/variable_tables/cross_year_variables_list.htm

• Not for oral health yet, but many others– Demographics, dietary recall, C-reactive protein, smoking, reproductive health

– Q-Bank – question evaluation reports• http://wwwn.cdc.gov/qbank/Home.aspx

III. NHANES data resources

– Dental, Oral and Craniofacial Data Resource Center (CDC and NIDCR funded project)• Current online resources

– Data Query System, some publications, survey questions inventory, Catalog of Surveys and Archive of Methods Related to Oral Health (400+ surveys listed)

• Tutorial materials in development– Will cover this in hands-on workshop

» Basic info on key data sets with oral health, linked to source» Expanded documentation of oral health components» Variable crosswalks» Code for creating analytic data sets» “clean” oral health data sets

III. NHANES data resources

– Restricted data, linked data and Research Data Centers operated by NCHS or Census

– Research Data Centers http://www.cdc.gov/rdc/ – Linked data – national surveys linked to CMS,

mortality, air quality, Social Security Benefit history• http://www.cdc.gov/nchs/data_access/data_linkage_ac

tivities.htm

Hands-on workshop

V. Orientation to NCHS NHANES Web siteVI. Orientation to DRC tutorial materialsVII.Practice examples (in SAS and SUDAAN)

Hands-on Workshop

V. Orientation to NCHS NHANES Web site– Locating and downloading files

• Demographics• Questionnaire

– Oral health– Oral health referral and recommendations– Household and other files

• Examination– Oral health examination– Dietary interview files

• Laboratory– Tobacco– C-Reactive Protein– Vitamin D

Hands-on workshop

VI. Orientation to DRC tutorial materials

Hands-on workshop

VII.Practice examples (in SAS and SUDAAN)– Setting up an analytic data set with

questionnaire items– Preparing an analytic data set with oral health

examination items and dental indices– Preparing an analytic data set with laboratory

items– Producing national estimates for Healthy People

objectives

Hands-on workshop

VII.Practice examples (in SAS and SUDAAN)– Setting up an analytic data set with

questionnaire items

Hands-on workshop

VII.Practice examples (in SAS and SUDAAN)– Preparing an analytic data set with oral health

examination items and dental indices

Hands-on workshop

VII.Practice examples (in SAS and SUDAAN)– Preparing an analytic data set with laboratory

items

Hands-on workshop

VII.Practice examples (in SAS and SUDAAN)– Producing national estimates for Healthy People

objectives• Tracking Healthy People 2010• CDC WONDER/Data2010