SPADE.RIVM-package {SPADE.RIVM}R Documentation

SPADE.RIVM : SPADE, a Statistical Program to Assess Dietary Exposure

Description

SPADE is developped at the RIVM (National Institute for Public Health and the Environment of the Netherlands) and contains various programs for estimating the habitual (usual) intake distribution of micronutrients, foods, supplements and combinations of these compounds.
The R package SPADE.RIVM together with a detailed tutorial/manual is available at www.spade.nl

SPADE is intended for the estimation of habitual intake distributions of
1. daily intakes of e.g. micronutrients (1-part model)
2. episodical intakes of e.g. foods, where all persons are supposed to be potential consumers of the food (2-part model)
3. intakes of dietary supplements (daily or episodical)
Supplements have often a non-continuous density but a spicky one. A special model, developed by the SPADE team, is used. Information on the frequency of the dietary supplements can be taken into account.
4. episodical intakes, containing zeros from real never-consumers
External information on real never-consumers is needed, e.g. information from a Food Frequency Questionnaire (FFQ), which indicates the zeroes of the real never-consumers of e.g. the food.
5. intakes from one food source and one dietary supplement (3-part model)
where external information is available on the users and non-users of the supplement. For this case the first-shrink-then-add approach is implemented and also the user and non-user parts of the sample are modelled separately.
6. intakes of micronutrients from several food sources, one fortified food with known non-consumers, and one supplement (multi-part model)
Also here the first-shrink-then-add approach is used, but in contrast with 5., the supplement user part and non-user part of the supplements are modelled at once.
7. intakes from daily or episodical intakes for various subgroups (1- and 2-part model)
The function divides the original data into the user-defined subgroups, (e.g. social economic status (SES), gender or region) and carries out for the whole sample and for each subgroup the 1- or 2-part model.

Options in SPADE
SPADE is still in development, but many options are already available.
We mention the most important ones.

SPADE models the intakes after a Box-Cox transformation as a continuous function of age, in years or another time unit. The intake amounts are modelled with a fractional polynomial of age and in case of episodical intakes also a second model part is needed in which the intake frequencies are modelled with a Beta-Binomial distribution and again as a function of age. For more flexibility, cubic splines are used. Several (fast) approaches for the back-transformation are implemented.
The resulting usual intake distributions are reported in tables and graphically in plots for every age unit and also diagnostic reports and graphics are provided.
The user can define easily age classes and the final results are reported per age class in tables and graphically in plots.
Survey weights (including sampling strata) are allowed and handled in all models.
Comparison with dietary reference values. The estimated usual intake distributions can be simultaneously compared with the user provided dietary reference values: allowed are 2 lower cut-off point, 2 upper cut-off point and one Adequate Intake (AI).
The uncertainty in the final results (percentiles per age or age-class, proportions below or above the cut-off points) are provided as confidence intervals obtained by dedicated bootstrap functions, which provide also several diagnostic plots.
Diagnostic and reporting graphics are saved automatically in one of the following formats: jpeg, wmf or jpg in colour or gray-scale plots.
Final tables are saved as csv-files which can be used easily in Excel or ASCII editors.
A detailed user-manual/tutorial is available at www.spade.nl.

Advantages of SPADE

- the latest version of SPADE is always available at www.spade.nl
- easily to install in the free available statistical program R
- detailed manual-tutorial available
- SPADE is not web-based, so no need to upload data to an external website
- open source, changes and extensions are possible
the SPADE team appreciates the communications of these changes to add them, if useful, to SPADE.

Details

Package: SPADE.RIVM
Type: Package
Version: 3.2.03
Date: 2017-11-20 GPL-3
LazyLoad: no

~~ An overview of how to use the package, including the most ~~ ~~ important functions ~~

Author(s)

Arnold Dekkers PhD
Janneke Verkaik-Kloosterman PhD
Marga Ocké PhD
National Institute of Public Health and the Environment (RIVM)
Bilthoven, the Netherlands
email: spade@rivm.nl (the SPADE TEAM)

Maintainer: Arnold Dekkers <Arnold.Dekkers@rivm.nl>

References

Dekkers, Arnold LM, Verkaik-Kloosterman, Janneke, van Rossum, Caroline T.M., Ocke, Marga C.
SPADE, a new statistical program to estimate habitual dietary intake from multiple food sources and dietary supplements
J. Nutr., 2014, 144 (12), to appear,
http://dx.doi.org/10.3945/jn.114.191288

Arnold Dekkers, Janneke Verkaik-Kloosterman, Caroline van Rossum, Marga Ocke
SPADE, Statistical Program to Assess habitual Dietary Exposure
Manual 2.0
December 2014, RIVM, Bilthoven, The Netherlands
freely available at www.spade.nl

See Also

~~ Optional links to other man pages, e.g. ~~ ~~ <pkg> ~~

Examples

1-part model for daily intakes
 f.spade(frml.ia=folate~fp(age,frml.if="no.if",data=DNFCS,min.age=7,
 max.age=69,sex="male")
 

2-part model for episodically consumed foods
set first the random seed to get identical results
 set.seed(10)
 f.spade(frml.ia=syn_fol~fp(age,frml.if=syn_fol~cs(age),data=DNFCS,
 min.age=7,max.age=69,sex="male")
 

2-part model for intakes from dietary supplements
 set.seed(20140325)
 f.spade.supplement(frml.ia=s_syn_fol~age, frml.if=s_syn_fol~cs(age), 
  nsusu.var = "user", data=DNFCS, min.age=7, max.age = 69, sex.lab ="female", 
  dt.pop = population2008, age.classes = c(6,9,13,19,30,50,69))
 

2-part with known never-aclohol users
 set.seed(20140424)
 f.spade.with.nonconsumers(frml.ia=alcohol~fp(age),
  frml.if=alcohol~cs(age), data=DNFCS, min.age=18, max.age=69,sex="female", 
  prb=c(seq(0.25,by=0.05,to=0.50),.90,.95,.99), dgts.distr=1, 
  age.classes=c(17,19,30,50,69), dt.pop=population2008, 
  nonuser.name="user_alcohol",nonuser.index=0)
  

3-part model for intakes from one food source and one dietary supplement
 set.seed(20140325)
 f.spade3 (
  frml.ia.nonsu = syn_fol ~ fp(age), 
  frml.if.nonsu = syn_fol ~ cs(age),
  frml.ia.su = syn_fol ~ fp(age), 
  frml.if.su = syn_fol ~ cs(age), 
  frml.ia.supp = s_syn_fol ~ age + vitb_freq_win + fol_freq_win +
    mult_freq_win + multmin_freq_win + vitb_freq_rest + 
    fol_freq_rest + mult_freq_rest,
  frml.if.supp = s_syn_fol ~ cs(age),  
  nsusu.var = "user", 
  data = DNFCS, 
  min.age = 7, 
  max.age = 69, 
  sex.lab = "female", 
  age.classes = c(6,13,30,69), 
  weights.name = "w_demoseas_comb_nas",
  dt.pop = population2008)


[Package SPADE.RIVM version 3.2.07 Index]