uds_z: Calculate the z-scores and impairment indicators of UDS3...

In JiyueQin/eas: EAS(Einstein Aging Study)

Calculate the z-scores and impairment indicators of UDS3 cognitive tests

Description

uds_z generates demographically adjusted z-scores and impairment indicators of UDS3 cognitive tests. Users can specify whether to use EAS or NACC norms.

Usage

uds_z(dat, tests, norms = "eas", impair_sd = 1, out_mean_sd = F)

Arguments

dat	a dataframe or tibble which contains columns of demographic covariates and UDS3 cognitive tests in numeric format. It should have the four demographic variables named as female(binary, 1 indicates female), age(continuous), educyrs(continuous) and black_race(binary, 1 indicates non-hispanic black). Also, it should have at least one of the following 17 UDSNB 3.0 cognitive test variables(also see Details): mocascore,verbatimi,paraphrasei,verbatimd,paraphrased,bensonscorei,bensonscored,numspancorf, numspancorb,minttotal,fwords60sec,lwords60sec,flword,animals60sec,vegetables60sec,tr_a1,tr_b1. All the demographic variables and cognitive test variables need to be numeric.
tests	character or a character vector, the name of the cognitive test(s)
norms	specify if you want to use EAS norms(norms='eas') or NACC norms(norms='nacc'), default is 'eas'
impair_sd	a positive number, how many SD worse than the mean should be defined as impairment, default is 1.
out_mean_sd	logical, out_mean_sd=T additionally adds mean and sd estimates for each cognitive test in the output dataframe, default is F

Details

Below is the list of UDSNB 3.0 tests, the corresponding variables and item numbers on the UDSNB 3.0 form.
- Montreal Cognitive Assessment(MoCA): mocascore(1f)
- Craft Story 21 Recall(Immediate): verbatimi(3a), paraphrasei(3b)
- Benson Complex Figure Copy: bensonscorei(4a)
- Number Span Test-Forward: numspancorf(5a)
- Number Span Test(Backward):numspancorb(6a)
- Category Fluency:animals60sec(7a), vegetables60sec(7b)
- Trail Making Test: tr_a1(8a), tr_b1(8b)
- Craft Story 21 Recall(Delayed): verbatimd(9a), paraphrased(9b)
- Benson Complex Figure Recall: bensonscored(10a)
- Multilingual Naming Test (MINT): minttotal(11a)
- Verbal Fluency(Phonemic Test): fwords60sec(12a), lwords60sec(12d), flword(12g)

Regression estimates from normative data are used to generate gender, age, education, black race adjusted z-scores. Use ?norms_coef_eas and ?norms_coef_nacc to see the documentation of regression estimates from the two norms.

The mean is calculated as:

\hat{Y}=\hat{b0}+ \hat{b1}*female + \hat{b_2}*(age-77) + \hat{b3}*(educyrs-16)+ \hat{b4}*black

For trail A1 and trail B1, as a higher score indicates a worse performance, z-score is calculated as

-(Y - \hat{Y})/\hat{σ}

For all the other cognitive tests where a higher score indicates a better performance, z-score is calculated as

(Y - \hat{Y})/\hat{σ}

Thus, an impairment can be defined as a z-score below a certain cutoff. For example, using a cutoff of -1 means that a subject is defined as impaired on this test if their performance is 1SD worse than the population mean. The indicator of impairment is generated as:

impair =1 if z-score<=-1

You can specify how many SD worse than the population mean should be defined as impairment by providing impair_sd(default is 1). 1SD and 1.5SD are some common choices.

Value

a tibble(dataframe) with the demographically adjusted z-score and impairment indicator for each of the specified cognitive tests.

Author(s)

Jiyue Qin

See Also

https://github.com/JiyueQin/eas

Examples

# here is a sample datset used to calculate z-scores. You should prepare your dataset in this standard format.
head(sample_dat)
# calculate z-scores and the impairment indicators for tr_a1 and tr_b1 with NACC norms and 1.5 SD to define impairment.
uds_z(sample_dat, c('tr_a1','tr_b1'), norms = 'nacc', impair_sd = 1.5)
# calculate the z-score and the impaiment indicator for minttotal with EAS norms and 1SD to define impairment
# Also output mean and sd estimates in addition to the z-scores and the impairment indicators.
uds_z(sample_dat, 'minttotal', norms = 'eas', impair_sd = 1, out_mean_sd = T)

JiyueQin/eas documentation built on April 4, 2022, 1:51 a.m.