I am trying to estimate the Usual intake, Estimated Average Rate (EAR), and UL of Magnesium by using NCI method. Since the recommended EAR and UL level of Magnesium are the same for most of the age groups ( 350 mg), I decided to account "total magnesium intake (food + supplement)" to estimate the EAR and use " magnesium intake only from supplement " to estimate the UL. I was successful in estimating the EAR percentage, but the syntax did not work for the UL (I just replaced the variable which represents the dietary supplement intake of Magnesium instead of total". Here is the syntax I used for the UL:
"%nci2015(data = master.nci, response = vsddmag, WTS = WTS_M, BSW=BSW,nBSW= 500,
foodtype = MAG, subject = sampleid, repeat = suppid, covars_prob=, covars_amt = FSDDEKC,
outlib = work,pred = _pred_unc_magnesium, param = _param_unc_magnesium, modeltype = amount, lambda = , seq = suppid, weekend = ADMFW, vargroup =, numvargroups = subgroup=DHHDDRI, start_val1 = , start_val2 = , start_val3 = , vcontrol = , nloptions = ,titles = , printlevel = 1, cutpts = 500 1000, ncutpts = 2, nsim_mc = 100, add_da = nci_supp ,EAR = mag_EAR, UL = mag_UL, seed = 0, mcsimda = usual_intake_mag, wkend_prop = 3/7, byvar = , final = final, address = P:\NCI\practice\);
I would really appreciate it if anyone can help me with this issue.
Thanks,
Elsa
Hi Elsa,
I am not sure about the actions the macro performs, so it is hard to know what exactly goes wrong. I suspect the parameter estimates the proportion of individuals below the EAR value: Pr(Xmag_UL). A couple of possibilities:
Perhaps you could share the %nci2015; macro so it is easier to find the problem.
However, I think the reason for failure is much simpler: usual (30-d) supplement intake from CCHS does not need to be measurement error corrected. I guess there are no repeated measurement for the vsddmag variable so the NCI method will not work anyway. In other words, you don’t have to apply the NCI method to estimate Pr(X>UL) of intake from supplement only! You can use the “raw” intake from supplement directly. Please note that this is different than if you were considering both intake from foods and supplement.
Also, a couple of unsolicited advices:
Finally, I recommend you look up these references:
The SIMPLE macro (allows you to use the NCI method without having to code complex steps): Luo, H., Dodd, K. W., Arnold, C. D., & Engle-Stone, R. (2021). Introduction to the SIMPLE Macro, a Tool to Increase the Accessibility of 24-Hour Dietary Recall Analysis and Modeling. J Nutr, 151(5), 1329-1340. doi:10.1093/jn/nxaa440
Detailed description of the univariate method: Luo, H., Dodd, K. W., Arnold, C. D., & Engle-Stone, R. (2022). Advanced dietary analysis and modeling: a deep dive into the National Cancer Institute method. J Nutr. doi:10.1093/jn/nxac144
Assumptions, statistical techniques, and considerations for measurement error methods: Kirkpatrick, S. I., Guenther, P. M., Subar, A. F., Krebs-Smith, S. M., Herrick, K. A., Freedman, L. S., & Dodd, K. W. (2022). Using short-term dietary intake data to address research questions related to usual dietary intake among populations and subpopulations: Assumptions, statistical techniques, and considerations. J Acad Nutr Diet, 122(7), 1246-1262. doi:10.1016/j.jand.2022.03.010
Method to assess intake from foods + supplement: Bailey, R. L., Dodd, K. W., Gahche, J. J., Dwyer, J. T., Cowan, A. E., Jun, S., . . . Tooze, J. A. (2019). Best Practices for Dietary Supplement Assessment and Estimation of Total Usual Nutrient Intakes in Population-Level Research and Monitoring. J Nutr, 149(2), 181-197. doi:10.1093/jn/nxy264
- Badges
- Science topic