UK Food Data

gharpur

Picking up on a discussion in the old forums, I'm another person who would really love to see UK/Euro food data added. My daughter has recently started a ketogenic diet for her epilepsy, and we're finding Cronometer a fantastic tool to support this, BUT realised that there's a big problem with accurate tracking when we made her some celeriac mash yesterday... the macros were way off from the recipe we'd been given by her dietician, and it turns out that the USDA data in Cronometer has net carbs for celeriac at 4.7g/100g while the McCance and Widdowson (UK) dataset has them at 1.9g/100g. This kind of difference makes a HUGE impact on accurately managing a ketogenic diet, and my web research suggests that the UK data is likely the more accurate one due to better testing methods (the USDA data in many instances involve inferred/calculated values https://www.ars.usda.gov/ARSUserFiles/80400525/Articles/jfca10_102-114.pdf ). The accuracy issue is of course in addition to the obvious user benefits of having more international packaged foods available in the DB.

Accurate data is critical to the app's usefulness, and the McCance and Widdowson data is readily available in electronic form here: https://www.gov.uk/government/publications/composition-of-foods-integrated-dataset-cofid . I understand you're a small company, and as a computer scientist, I'd be more than happy to assist in manipulating this dataset into a format where it could be directly uploaded to Cronometer's tables, and am happy to be contacted by a developer if this offer is of any use.

Thanks again for a great app!

Aaron

@gharpur I'll leave the rest of the response to our nutrition data curation team, but I would like to point out that the net-carbs shown in celeriac will also depend on your cronometer setting for net-carbs to include or exclude sugar alcohols from the calculation, and that may explain most of the discrepancy you're seeing.

gharpur

Thanks for the quick reply Aaron! Changing the sugar alcohols certainly makes a difference - it takes net carbs from 7.4% down to 5.6%, but still leaves a big discrepancy to the McCance and Widdowson data which has 2.3%. I guess it also begs the question of whether celeriac truly has 1.8% sugar alcohols? I couldn't see where this number comes from, either within Cronometer, or on the raw USDA data here: https://ndb.nal.usda.gov/ndb/foods/show/2912?n1={Qv=1}&fgcd=&man=&lfacet=&count=&max=&sort=&qlookup=&offset=&format=Full&new=&measureby=&Qv=1&ds=Standard+Reference&qt=&qp=&qa=&qn=&q=&ing=

Karen_Cronometer

Sugar alcohols are naturally occurring in small amounts in the plants that we eat. The value for sugar alcohols comes from the Nutrition Coordinating Center - who pay particular attention to the analysis of carbohydrate components in foods.

A major project we are working on right now involves reviewing databases such as McCance and Widdowson to expand our database to include more commonly consumed foods for users outside of North America. We look forward to having many more foods available soon!

Aaron

@gharpur I should've mentioned, our two primary Celeriac entries do not come from USDA, they are the NCCDB entries, which are a higher quality analysis than USDA and include sugar alcohols in the data

gharpur

Karen/Aaron - thanks for the additional info, and it's great to hear that extra data will be available soon!

All this leaves me wondering: (a) whether I should trust NCCDB over McCance and Widdowson and (b) whether to include sugar alcohols or not, since opinions vary, and it seems that some are more ignorable than others!

Karen_Cronometer

I admire your attention to detail! While analytical data definitely provide more detailed nutrient data, no two celeriac are identical. The nutrient databases are tools we use to estimate our intake as best as possible. At the end of the day, you will have to evaluate the sources of information and decide for yourself.