Upload
jeffery-snow
View
212
Download
0
Embed Size (px)
Citation preview
STATEMENT OF THE PROBLEM ANDSTUDY DESIGN
Lu Ann Aday, Ph.D.
The University of Texas
School of Public Health
TYPE OF DESIGN: Observational – Cross-Sectional (One-Group) GROUPS
No. of groups: 1 Criteria for selection: Population of
interest
TIME PERIODS No. of time periods: 1 Reference periods: Present (and recall of
past)
TYPE OF DESIGN: Observational –Group-Comparison (Case-control) GROUPS
No. of groups: 2+ Criteria for selection: Population subgroups
with and without characteristic of interest
TIME PERIODS No. of time periods: 1 Reference periods: Present and recall of
past
TYPE OF DESIGN: Observational –Longitudinal (Prospective) GROUPS
No. of groups: 1 or 2+ Criteria for selection: Population or
subgroups that are and are not likely to develop characteristic of interest
TIME PERIODS No. of time periods: 2+ Reference periods: Present and future
TYPE OF DESIGN: Experimental –“True” Experiment (Randomized Clinical Trial)
GROUPS No. of groups: 2+ Criteria for selection: Randomly
determined subgroups of population
TIME PERIODS No. of time periods: 2+ Reference periods: Present and future
DESIGN DIMENSIONS
HEALTH SURVEY EXAMPLES
UNICEF MULTIPLE INDICATOR CLUSTER SURVEYS (MICS)
CALIFORNIA HEALTH INTERVIEW SURVEY (CHIS)
NATIONAL DENTAL MALPRACTICE SURVEY(NDMS)
Study Objectives 1. To estimate and monitor World Summit for Children and related World Fit for the Millennium Development Goals (MDGs) indicators of child survival and development for children and mothers in participating countries.2. To compare the indicators over time, as well as across countries.
1. To provide statewide estimates for the population of the State of California overall and local-level estimates for most counties in the State of California on a variety of public health topics, e.g., health status, health care access, insurance coverage.2. To compare estimates across local areas and between larger racial/ ethnic groups, and selected smaller ethnic groups.
1. To estimate dental malpractice insurance experience in a representative sample of U.S. dentists in 1991.2. To test hypotheses regarding the practice characteristics that are predictive of dental malpractice insurance experience.
Study Designs Descriptive longitudinal, comparative national surveys conducted in 1995, 2000, & 2005 in participating countries
Descriptive longitudinal, comparative state & local surveys conducted on a biennial basis, beginning in 2001, in the State of California
Analytical cross-sectional national survey of U.S. dentists, conducted in 1991
Research Questions
What? World Summit for Children & MDG indicators of child survival & development
health status, chronic conditions, health behaviors, health care access, insurance coverage, etc.
malpractice insurance experience & practice characteristics
Who? children & mothers state & county populations dentists
Where? participating countries State of California U.S.
When? 1995, 2000, 2005 2001, 2003, 2005 1991
Why? -- -- To test hypotheses regarding the practice characteristics that are predictive of dental malpractice insurance experience.
SAMPLE OBJECTIVE: Objective 1
1. TO ESTIMATE dental malpractice insurance experience in a representative sample of dentists in the U.S. in 1991.
SAMPLE OBJECTIVE: Cross-Sectional, Descriptive Design
OBJECTIVE ELEMENTSTO ESTIMATE STATISTICAL PROCEDURES:
Univariate, e.g., frequencies,
mode, median, mean
dental malpractice insurance experience WHAT?
in a representative sample of dentists WHO?
in the U.S. WHERE?
in 1991. WHEN?
SAMPLE OBJECTIVE: Objective 2
2. TO COMPARE dental malpractice insurance experience by the dentist’s demographic characteristics.
SAMPLE OBJECTIVE: Group-Comparison, Descriptive Design
OBJECTIVE ELEMENTSTO COMPARE STATISTICAL PROCEDURES:
Bivariate, e.g., chi-square, t-test, ANOVA, correlations
dental malpractice insurance experience WHAT? (dependent variable)
by the dentist’s demographic characteristics.
WHO? (one independent variable)
SAMPLE OBJECTIVE: Objective 3 3. TO ANALYZE the relative
importance of doctor-patient communication, practice characteristics, practice finances, and dentist’s demographic characteristics in predicting dental malpractice insurance experience.
SAMPLE OBJECTIVE: Objective 3—Hypothesis Doctor-patient communication is
more important than practice characteristics, practice finances, and dentist’s demographic characteristics in predicting dental malpractice insurance experience.
SAMPLE OBJECTIVE: Cross-Sectional, Analytical Design
OBJECTIVE ELEMENTSTO ANALYZE the relative importance of STATISTICAL PROCEDURES:
Multivariate, e.g., linear regression, logistic regression
• doctor-patient communication• practice characteristics• practice finances, and• dentist’s demographic characteristicsin predicting
WHO? (two or more independent/control variables)
WHY? (hypothesis re results)
dental malpractice insurance experience. WHAT? (dependent variable)
MEASUREMENT MATRIX:
National Dental Malpractice Survey (Aday & Cornelius, 2006, Table 15.1)
QUESTION CONCEPT LEVEL OBJECTIVE
3 Doctor-patient communication: characteristics of unsatisfactory patient encounters (Likert scale)
ordinal(interval)
3
10 Practice characteristics: avg. office waiting time for patient
ordinal 3
28 Practice finances: % patients with insurance coverage
interval 3
35 Malpractice insurance: no. of complaints
ordinal 1, 2, 3
54 Demographics: gender nominal 2, 3
SURVEY ERRORS: Matching the Survey Design to Survey Objectives
Systematic Errors Variable Errors
Poor internal validity
Poor external validity
Design specification ambiguity
Solutions to errors Use randomization, matching, or statistical controls to rule out other factors that may account for relationships between variables.
Clearly specify where, with whom and when the survey will be done in stating the study objectives and design the sample frame and survey sampling procedures accordingly.
Clearly specify the study objectives and related concepts to be measured in the survey, particularly in relationship to the underlying study design and data analysis plan for the study.