Table 2.
Mean daily energy intake among 3728 adults with complete portion size data (reference), compared with energy intakes calculated with portion sizes derived from four imputation methods, Danish Health Examination Survey 2007–2008
| Men | Women | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Energy | 95 % CI | RMSE | 95 % CI | Bias | 95 % CI | Energy | 95 % CI | RMSE | 95 % CI | Bias | 95 % CI | |||
| % | MJ | MJ | kJ | kJ | kJ | kJ | % | MJ | MJ | kJ | kJ | kJ | kJ | |
| Total energy | ||||||||||||||
| Reference | – | 10·97 | 10·81, 11·13 | Ref. | – | Ref. | – | – | 8·81 | 8·69, 8·92 | Ref. | – | Ref. | – |
| Median | – | 10·45 | 10·37, 10·50 | 1118 | 1098, 1139 | 579 | 563, 596 | – | 8·28 | 8·26, 8·30 | 1061 | 1011, 1111 | 469 | 455, 482 |
| KNN | – | 11·37 | 11·32, 11·41 | 1281 | 1262, 1299 | −340 | −365, −315 | – | 8·53 | 8·51, 8·56 | 1181 | 1129, 1234 | 218 | 191, 244 |
| MLR | – | 10·78 | 10·73, 10·83 | 1060 | 1028, 1092 | 248 | 223, 274 | – | 8·57 | 8·55, 8·60 | 1051 | 997, 1105 | 178 | 161, 195 |
| Coca | – | 10·80 | 10·75, 10·83 | 1230 | 1196, 1264 | 234 | 207, 261 | – | 8·56 | 8·54, 8·59 | 1146 | 1087, 1205 | 188 | 166, 210 |
| Fat | ||||||||||||||
| Reference | 31·2 | 3·43 | 3·36, 3·49 | Ref. | – | Ref. | – | 29·9 | 2·64 | 2·60, 2·68 | Ref. | – | Ref. | – |
| Median | 31·8 | 3·32 | 3·31, 3·34 | 375 | 364, 386 | 124 | 119, 130 | 30·8 | 2·55 | 2·54, 2·56 | 305 | 292, 317 | 67 | 65, 70 |
| KNN | 31·7 | 3·61 | 3·59, 3·63 | 502 | 491, 513 | −161 | −175, −146 | 30·0 | 2·56 | 2·56, 2·57 | 395 | 387, 404 | 56 | 47, 64 |
| MLR | 31·2 | 3·37 | 3·36, 3·39 | 392 | 381, 403 | 75 | 68, 82 | 30·0 | 2·57 | 2·56, 2·58 | 345 | 330, 361 | 45 | 39, 51 |
| Coca | 31·3 | 3·38 | 3·36, 3·39 | 473 | 458, 489 | 70 | 59, 81 | 30·0 | 2·57 | 2·56, 2·58 | 392 | 377, 407 | 49 | 43, 54 |
| Protein | ||||||||||||||
| Reference | 16·1 | 1·77 | 1·74, 1·80 | Ref. | – | Ref. | – | 16·3 | 1·44 | 1·42, 1·45 | Ref. | – | Ref. | – |
| Median | 16·5 | 1·72 | 1·71, 1·73 | 210 | 205, 215 | 57 | 54, 60 | 16·7 | 1·38 | 1·38, 1·38 | 191 | 188, 193 | 49 | 47, 50 |
| KNN | 16·3 | 1·86 | 1·84, 1·87 | 273 | 267, 279 | −78 | −87, −69 | 16·1 | 1·38 | 1·37, 1·38 | 251 | 246, 257 | 53 | 48, 58 |
| MLR | 16·2 | 1·74 | 1·73, 1·75 | 220 | 214, 225 | 37 | 32, 42 | 16·4 | 1·41 | 1·40, 1·41 | 211 | 205, 216 | 21 | 18, 25 |
| Coca | 16·2 | 1·75 | 1·73, 1·76 | 271 | 263, 278 | 34 | 27, 40 | 16·4 | 1·41 | 1·40, 1·41 | 249 | 243, 256 | 23 | 20, 26 |
| Carbohydrates | ||||||||||||||
| Reference | 42·2 | 4·63 | 4·57, 4·68 | Ref. | – | Ref. | – | 44·0 | 3·88 | 3·83, 3·92 | Ref. | – | Ref. | – |
| Median | 41·0 | 4·28 | 4·26, 4·30 | 613 | 598, 627 | 362 | 354, 371 | 42·7 | 3·54 | 3·52, 3·55 | 675 | 616, 733 | 319 | 307, 330 |
| KNN | 41·6 | 4·73 | 4·71, 4·76 | 672 | 656, 688 | −92 | −111, −73 | 44·1 | 3·77 | 3·75, 3·78 | 693 | 636, 750 | 88 | 70, 105 |
| MLR | 41·9 | 4·52 | 4·49, 4·55 | 580 | 560, 599 | 122 | 106, 138 | 43·8 | 3·75 | 3·77, 3·77 | 640 | 576, 704 | 100 | 86, 114 |
| Coca | 41·9 | 4·53 | 4·50, 4·55 | 602 | 585, 621 | 116 | 104, 128 | 43·8 | 3·75 | 3·73, 3·77 | 652 | 595, 708 | 105 | 89, 121 |
RMSE, root-mean-square error; bias, mean error; median, sex-specific median imputation which is equivalent to using sex-specific standard portion sizes; Coca, ‘comparable categories’; KNN, k-nearest neighbours; MLR, multinomial logistic regression; Ref., referent category.
The four methods were compared by their ability to predict the reference. The reference energy intakes were computed with a set of complete reported portion sizes. The results presented are mean values of ten imputations with each method (on random splits of the data). Note that a positive bias indicates an underestimation of the reference and a negative bias indicates an overestimation.