NoPriorCVDBleedRisk: PREDICT Major Bleed (2018) Risk Score for People Without...

In billy-nz/PredictRiskScores: Functions for generating PREDICT CVD risk scores

Description

NoPriorCVDBleedRisk calculates the 5 year absolute risk of major bleeding (gastrointestinal, intracranial, and other bleeds), for people without a history of atherosclerotic CVD.

Usage

NoPriorCVDBleedRisk(dat, sex, age, eth, exsmoker, smoker, nzdep, diabetes,
                 familyhx, lld, bpl, cancer, gibleed, puddiag, alcohol, liver,
                 puddrug, nsaid, steroids, ssri, sbp, tchdl, ...)

Arguments

dat	an optional data.frame or data.table containing input data (see details)
sex	binary sex or gender
age	age in years (see details)
eth	ethnicity (see details)
exsmoker	ex-smoker or recently quit
smoker	currently smoking
nzdep	socio-economic deprivation (see details)
diabetes	diabetes status
familyhx	family history of premature CVD
lld	receiving lipid lowering medication
bpl	receiving at least one blood pressure lowering medication
cancer	prior primary malignancy excluding squamous and basal cell skin cancers
gibleed	prior gastrointestinal bleeding
puddiag	prior non-bleeding and non-perforated peptic ulcer disease
alcohol	chronic high use of alcohol
liver	chronic liver disease or pancreatitis
puddrug	receiving at lease one peptic ulcer disease medication
nsaid	receiving at least one nonsteroidal anti-inflammatary drug
steroids	receiving at least one corticosteroid medication
ssri	receiving at least one selective serotonine reuptake inhibitor
sbp	measured systolic blood pressure in mmHg
tchdl	most recent value of total:HDL cholesterol
...	further arguments (see values)

Details

When a dataset is supplied, a risk score is produced for each row of data, resulting in a numeric vector of the same length. Each argument requires the variable name from the dataset dat that corresponds with the parameter. If the parameter dat is not supplied, then each argument is assigned an individual's actual parameter value.

The risk prediction equations were developed from a cohort of people aged 30 to 74 years. Additional analyses indicate that the sex-specific risk equations performed adequately in those aged 75-79 years. People aged 18-29 years or 80 years and older, are outside the range used to derive the equation, and therefore risk will be even more of an approximation. The function will calculate ages 18-29 as 30; and ages 80-110 as 79. All other age inputs are invalid and will return NA.

The co-efficients for ethnicity apply only to the following groups: European, Maori, Pacific, Indian, and (non-Indian) Asian. To obtain a risk estimate, ensure that the ethnicity input is either labelled or encoded using one of the values listed below (see values). All other inputs are invalid and will return NA.

The scale for socioeconomic deprivation is derived from the New Zealand Index of Deprivation; a small area-based measure that combines census data relating to income, home ownership, employment, qualifications, family structure, housing, access to transport and communications. The equations require NZDep to be categorised as quintiles, with 1 being the least deprived and 5 being the most deprived.

Value

returns either a single 5-year major bleed risk estimate, or a numeric vector of risk estimates if dat is provided. Input values for each parameter must conform to the following convention:

sex	label or encode as one of the following: M, Male, 1 F, Female, 0
age	numeric value for years of age between 20 and 110
eth	label or encode as one of the following: NZ European, European, NZEO, Euro, E, 1, 10, 11, or 12 Maori, NZMaori, NZ Maori, M, 2, or 21 Pacific, Pacific Islander, PI, P, 3, 30, 31, 32, 33, 34, 35, 36, or 37 Indian, Fijian Indian, South Asian, IN, I, or 43 Asian, Other Asian, SE Asian, East Asian, Chinese, ASN, A, 4, 40, 41, 42, or 44 note: Other Asian includes non-Indian South Asian
nzdep	numeric value between 1 and 5
exsmoker	label or encode as one of the following: Y, Yes, Ex, Ex-smoker, Exsmoker, E, 1, T, TRUE N, No, Non-smoker, Non, 0, F, FALSE
smoker	label or encode as one of the following: Y, Yes, Smoker, Current, S, 1, T, TRUE N, No, Non-smoker, Non, 0, F, FALSE
diabetes, familyhx	label or encode as one of the following: Y, Yes, 1, T, TRUE N, No, 0, F, FALSE
bpl, lld	label or encode as one of the following: Y, Yes, 1, T, TRUE N, No, 0, F, FALSE
cancer, gibleed, puddiag, alcohol, liver	label or encode as one of the following: Y, Yes, 1, T, TRUE N, No, 0, F, FALSE
puddrug, nsaid, steroids, ssri	label or encode as one of the following: Y, Yes, 1, T, TRUE N, No, 0, F, FALSE
sbp, tchdl	numeric value of measured result. Note: SBP and total:HDL values must be available
...	further arguments: dp numeric value to set decimal place; default is 4 allow.age logical. Whether or not age range is extended outside of 30 - 74; default is TRUE. If set to FALSE, then NA is returned as risk estimate. allow.na logical. Whether or not missing values for binary variables and smoking status are treated as 0; default is TRUE. If set to FALSE, then NA is returned as risk estimate.

See Also

NoPriorCVDRisk
NoPriorCVDRisk_BMI
NoPriorCVDRisk_Policy
NoPriorCVDBleedRisk
NoPriorT2DRisk
PostCVDRisk
PostACSRisk

Author(s)

Billy Wu (R Developer) and Vanessa Selak (Principal Investigator)

References

Selak V, Jackson R, Poppe K, et al. Predicting Bleeding Risk to Guide Aspirin Use for the Primary Prevention of Cardiovascular Disease: A Cohort Study. Ann Intern Med. 2019;170:357-368. doi: https://doi.org/10.7326/M18-2808

Using the bleeding risk equation developed by Selak et al (2019), I developed a web-based risk calculator that provides clinicians with an individualised estimate of the CVD benefit and bleeding harms of aspirin for their patients without established CVD.

Full Article Online Calculator

Examples

# As calculator (dataset not provided)
NoPriorCVDBleedRisk(sex=0, age=55, eth=21, exsmoker=0, smoker=0, nzdep=5, diabetes=1,
                    familyhx=1, lld=1, bpl=1, cancer=1, gibleed=1, puddiag=1, alcohol=0,
                    liver=0, puddrug=0, nsaid=1, steroids=1, ssri=0, sbp=130, tchdl=5)

NoPriorCVDBleedRisk(sex=0, age=76, eth=21, exsmoker=0, smoker=N, nzdep=5, diabetes=1,
                    familyhx=1, lld=1, bpl=1, cancer=1, gibleed=1, puddiag=1, alcohol=0,
                    liver=0, puddrug=0, nsaid=1, steroids=NA, ssri=0, sbp=130, tchdl=5,
                    dp = 5, allow.age = F, allow.na = FALSE)

# As a vectoriser (dataset provided)
NoPriorCVDBleedRisk(dat=DT, sex=sex, age=index_age, eth=eth_vars, exsmoker=exsmoke,
                    smoker=smoking_current, nzdep=nzdep, familyhx=family_hx, diabetes=dm,
                    sbp=index_sbp, tchdl=tchdl, lld=lld, bpl=bpl, cancer=hx_cancer,
                    gibleed=hx_gibleed, puddiag=hx_hud, alcohol=alc, liver=hx_liver,
                    puddrug=pudmx, nsaid=nsaid, steroids=steroidmx, ssri=ssri)

billy-nz/PredictRiskScores documentation built on April 4, 2020, 6:23 p.m.