Table 1. Factors associated with adherence to hospital guidelines regarding undernutrition monitoring (first step: Checking prealbumin level), department of internal medicine of the Lausanne university hospital, 2013 and 2014.
| Characteristics | Bivariate analysis | Multivariate analysis | ||
|---|---|---|---|---|
| Unadjusted OR (95% CI) | P-value | Adjusted OR (95% CI) |
P-value | |
| Admission year | ||||
| 2013 | 1 (ref.) | 1 (ref.) | ||
| 2014 | 1.24 (0.98–1.58) | 0.073 | 1.45 (1.11–1.88) | 0.006 |
| Gender | ||||
| Men | 1 (ref.) | 1 (ref.) | ||
| Women | 1.03 (0.83–1.27) | 0.79 | 1.00 (0.79–1.26) | 0.97 |
| Age category | ||||
| 18–59 | 1 (ref.) | 1 (ref.) | ||
| 60–79 | 1.70 (1.12–2.60) | 0.013 | 1.63 (1.03–2.58) | 0.036 |
| 80+ | 2.05 (1.37–3.06) | <0.001 | 2.03 (1.29–3.18) | 0.002 |
| p-value for trend* | <0.001 | 0.002 | ||
| Main diagnosis | ||||
| Miscellaneous | 1 (ref.) | 1 (ref.) | ||
| Circulatory system diseases | 1.37 (1.01–1.87) | 0.046 | 1.23 (0.86–1.76) | 0.24 |
| Cancer | 0.82 (0.57–1.19) | 0.30 | 0.51 (0.32–0.79) | 0.003 |
| Digestive system diseases | 0.64 (0.39–1.02) | 0.062 | 0.55 (0.33–0.92) | 0.023 |
| Infectious diseases | 1.43 (0.94–2.19) | 0.093 | 1.21 (0.76–1.93) | 0.42 |
| Mental & behavioral disorder/ Nervous system disease | 1.64 (1.10–2.44) | 0.015 | 1.74 (1.12–2.69) | 0.013 |
| Respiratory system diseases | 1.31 (0.96–1.79) | 0.092 | 1.24 (0.87–1.76) | 0.22 |
| NRS-2002 categories | ||||
| Medium (3–4) | 1 (ref.) | 1 (ref.) | ||
| High (>4) | 1.5 (1.17–1.91) | 0.001 | 1.11 (0.84–1.46) | 0.45 |
| Charlson comorbidity index | ||||
| Low (CCI<2) | 1 (ref.) | 1 (ref.) | ||
| High (CCI≥2) | 1.30 (1.06–1.60) | 0.013 | 1.36 (1.05–1.77) | 0.020 |
| Any nutritional management | ||||
| No | 1 (ref.) | 1 (ref.) | ||
| Yes | 4.95 (3.96–6.20) | <0.001 | 5.57 (4.38–7.07) | <0.001 |
Abbreviations: ref., reference category; CCI, Charlson comorbidity index; NRS-2002, nutrition risk screening 2002. Analyses were performed among patients with NRS-2002≥3 (n = 1,605). Bivariate analysis performed using logistic regression in which adherence to the guideline (yes or no) was the outcome and each characteristic was the predictor; multivariable analysis performed using logistic regression adjusting for all variables in the table.
* P-values for trend are computed using orthogonal polynomial contrasts (command contrast p. of Stata).