EBPP MOD 3

Journal Quality Databases

Pubmed, Cochane Library, CINAHL

Directory of Open Acess Journal

Transparency and best practices in scholarly publishing

Higher JCR

Greater significance within discipline

Scrimago Journal and Country Rank

Visibility of Journals in Scopus Database

Journal Quality

Databases

Peer Review

Impact Factor

Predatory Journal

Predatory Journal

More open access journals

Dilution of scientific credibility

Poor peer review

Lower impact factor

Publication Bias

Quantitative studies not reporting significant difference that are not published or timely.

Bias against the qualitative research

Not all journals

Trustworthiness Factors

Internal Validity

External Validity

Reliability

Objectivity

Internal Validity

Assessing methods to determine if the study is actually measuring the intended primary outcome

Criterion validity ‘

Concurrent validity

External Validity

Generalize the findings back to the population from which the sample was derived and to generalize the results to a realistic context

• How was sample selected

• How was sample assigned

• Is intervention realistic

Inverse

Internal and external validity have a _____ relationship

Extraneous variables

Trustworthiness is influenced by _____ ____

This also introduces error to the study

Affect response variable

Confounding variable

Cohen’s Kapppa

correlation coefficient 0.8+ = trustworthiness

Objectivity

Unbiased, honesty, precise,

Control for unconscious bias

Blinding

Conflict of Interest

Reliability

Were the data measures consistently?

Objectivity

Was the conduct of the study unbiased?

Sample

Drawn from the population

Descriptive statistics

To organize summarize and display data to make them more understandable

Inferential statistics

TO provide predictions about population characteristics based on on information from a sample drawn from population

Nominal Data Scale

Unordered mutually exclusive categories

EX: Gender , marital status

Ordinal Data Scale

Mutually exclusive categories that are ordered in some meaningful way

Ex: Course Grades

Numerical

Quantitative data with finite numbers or counts (discrete) or infinitely values ( continuous )

Ordinal Data

As nominal/ numerical at times

Mood on scale of 1-20

Numerical Scale

Interval: Values without true zero point

Ex: Degrees F/C

Ratio: Data with true zero point

EX: Weight and blood pressure

Nominal

Frequency ( count and percent)

Categorical

Ordinal

Frequency or Central tendency and spread

Catergorical and Continuous

Numerical

Central tendency and spread (SD and variance)

Continuous

Statistical Inference

Estimation ( point estimate/ CI)

Hypothesis Test

Hypothesis Test

Procedure for testing a claim about a property of a population

H0: Null Hypothesis

Standard theory which we want to prove wrong

HA: Alternative Hypothesis

Working Theory

Level of Significance

It gives the probability of incorrectly rejecting the null hypothesis when it is actually true

Type 1 error

Reject null hypothesis when it is true

AVOID

Difference exists

Type 2 error

Accept null hypothesis when it is false

No difference

Reject H0

P value < 0.05 (any other confidence interval)

Fail to reject H0

P value> 0.05

Confidence interval (Estimation)

At a certain degree of confidence, an interval contains true population mean

Higher degree of confidence

( lower alpha level)

Wider the interval

Small sample size , Large SD

Wider interval

Statistic significant

No 0 included within the mean difference test

Statistic significant

DOES NOT include 1 with RATIO TEST

Confounding variable

A mixing or blurring of effects hen a researcher attempts to relate an exposure to an outcome but actually masseurs the effect of a third factor, the _____

Confounding variable

A variable that changes the relationship between an independent and DV because it is related to both

NOT intermediate link in chain of causation

Decreases internal validity

Confounded Criteria

Must be a risk to outcome

Must be associated with exposure

Must not be an intermediate step in causal pathway

Must not be surrogate of exposure

How to control Confounding

Randomization, Restriction, Matching, Stratification, Multivariable Modeling

Randomization

Process of assigning individual to treat pr control based on accepted mechanism

Everyone has assume probability

Restriction

Inclusion and exclusion criteria

Potential confounders are prohibited or restricted from very and producing the confounding effect

Matching

Indentifying a characteristic that is alleged to be a source of bias and matching cases with controls or exposed patients with unexposed

Constrained by sample size

Stratification

Control for variation drive by a confounded at the analysis level.

Analyze each stratum defined by the level of the confounder

Multivariable Modeling

Efficient way to control confounding at analysis levels

Provide information on the contribution of each confounder to occurrence of the outcome

Interval

Temperature

Year

Sample Size

Ratio

Height

Age

Heart Rate

Nominal

Sex

Race

Diagnosis

Ordinal

Level of pain

Level of agreement

Standard deviation

Meah for a population

Confidence Intervals

Used to describe a variable (single measurement or comparison)

Can be used instead of HT

Range of reasonable values

Estimate magnitude, direction, and certainty of measurement

1, 0

Magic number for ratios, Magic number for difference

Parametric

Ratio, continuous, or interval scale

Normal shape

Stuff sample size

Non parametric

Nominal or ordinal scale

Non-normal shape

Small sample size

Non equivalent variability

Chi Square

Nominal Data

Independent sample

2 or more treatments and 2 or more outcomes (2×2) table

Fishers Exact Test

Nominal Data

Independent

Used for SMALL Samples/ Infrequent Outcomes

T test

Continuous dependent

2 groups

ANOVA

Continuous level of measurement

F= t squared

NO confounder - 1 way ANOVA

1 confounder -2 way ANOVA

Multiple t tests

Increase chance of Type I error

Bonferron method of correction ( DIVIDE ALPHA BY # OF TESTS)

LOWER

MCPS ______ the alpha to control for Type 1 error ( balance with type 2)

Multivariate Statistic

Multiple linear regression

Multiple logistic regression

Cox proportional hazard models

Appropriate Statistical Test

Study Design

Level of measurement

Confounders

Assumptions

Prevalence

Proportion of a defined population currently living with a given condition during a specific end time period or on a specified date

Incidence

Number of new cases within a defined population at a defined point/ period of time