Limits...
Evaluation of an Innovative Method for Calculating Energy Intake of Hospitalized Patients

View Article: PubMed Central - PubMed

ABSTRACT

The purpose of this study was to evaluate a multi-component method for capturing nutrient intake, which used observation, photography, and an innovative computer program. To assess reliability and accuracy, multiple responsible employees (REs) independently conducted nutrient intake assessments on simulated meals; each RE’s results relating to energy intake were compared to those from the other REs and to those obtained by pre- and post-meal weighing of the food items. System efficiency was assessed by having REs perform independent assessments on the same set of simulated meals using either the new or traditional hospital method for which the REs had to document each food item served and then find the items in a computer database–steps that were automated in the new method. Interrater reliability for energy intake estimated on clinic wards was excellent (intraclass correlation coefficient = 0.975, 95% CI 0.958 to 0.992) and there was a high level of agreement between the REs’ estimates and the true values determined by food weighing; per the method of Bland and Altman the mean difference between the two types of estimates was 0.3 kcal (95% CI, −8.1 to 8.7 kcal) with limits of agreement of −79.5 kcal to 80.1 kcal. Compared to the traditional method, energy intake assessments could be completed using the multi-component method in less than a third of the time. These results indicate the multi-component method is an accurate, reliable, and efficient method of obtaining energy intake assessments for hospitalized patients.

No MeSH data available.


Agreement between the two methods of estimating energy consumption (RE and food weighing, n = 90) as depicted using the method of Bland and Altman [32]. Linear correlation between the differences and the means indicated by solid line (y = 0.02x − 10.6, r2 = 0.02, p = 0.20). Limits of Agreement (i.e., the mean difference in estimates ± 2SD) indicated by dotted lines.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC5037542&req=5

nutrients-08-00557-f001: Agreement between the two methods of estimating energy consumption (RE and food weighing, n = 90) as depicted using the method of Bland and Altman [32]. Linear correlation between the differences and the means indicated by solid line (y = 0.02x − 10.6, r2 = 0.02, p = 0.20). Limits of Agreement (i.e., the mean difference in estimates ± 2SD) indicated by dotted lines.

Mentions: The level of agreement between RE estimates of energy consumption and estimates based on weighing the food items is graphically depicted in a Bland-Altman plot (Figure 1) [32]; the differences between the estimate pairs were plotted against the mean of the two estimates. The mean difference between the two types of estimates was 0.3 kcal (95% confidence interval −8.1 kcal to 8.7 kcal), p = 0.948), indicating no significant bias. The Limits of Agreement (i.e., the mean difference in estimates ± 2SD) were −79.5 kcal to 80.1 kcal, which were narrow enough to indicate that the RE estimates are sufficiently accurate for clinical use. The differences ranged from 89 kcal to a minus 90 kcal and only five observations were outside the limits of agreement, which is consistent with the 5% that would be expected to fall outside the limits based on a normal distribution. There was no significant correlation between the differences and the means (r = 0.14, p = 0.20) indicating that the agreement of the estimates was independent of the percentage of food calories consumed.


Evaluation of an Innovative Method for Calculating Energy Intake of Hospitalized Patients
Agreement between the two methods of estimating energy consumption (RE and food weighing, n = 90) as depicted using the method of Bland and Altman [32]. Linear correlation between the differences and the means indicated by solid line (y = 0.02x − 10.6, r2 = 0.02, p = 0.20). Limits of Agreement (i.e., the mean difference in estimates ± 2SD) indicated by dotted lines.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC5037542&req=5

nutrients-08-00557-f001: Agreement between the two methods of estimating energy consumption (RE and food weighing, n = 90) as depicted using the method of Bland and Altman [32]. Linear correlation between the differences and the means indicated by solid line (y = 0.02x − 10.6, r2 = 0.02, p = 0.20). Limits of Agreement (i.e., the mean difference in estimates ± 2SD) indicated by dotted lines.
Mentions: The level of agreement between RE estimates of energy consumption and estimates based on weighing the food items is graphically depicted in a Bland-Altman plot (Figure 1) [32]; the differences between the estimate pairs were plotted against the mean of the two estimates. The mean difference between the two types of estimates was 0.3 kcal (95% confidence interval −8.1 kcal to 8.7 kcal), p = 0.948), indicating no significant bias. The Limits of Agreement (i.e., the mean difference in estimates ± 2SD) were −79.5 kcal to 80.1 kcal, which were narrow enough to indicate that the RE estimates are sufficiently accurate for clinical use. The differences ranged from 89 kcal to a minus 90 kcal and only five observations were outside the limits of agreement, which is consistent with the 5% that would be expected to fall outside the limits based on a normal distribution. There was no significant correlation between the differences and the means (r = 0.14, p = 0.20) indicating that the agreement of the estimates was independent of the percentage of food calories consumed.

View Article: PubMed Central - PubMed

ABSTRACT

The purpose of this study was to evaluate a multi-component method for capturing nutrient intake, which used observation, photography, and an innovative computer program. To assess reliability and accuracy, multiple responsible employees (REs) independently conducted nutrient intake assessments on simulated meals; each RE’s results relating to energy intake were compared to those from the other REs and to those obtained by pre- and post-meal weighing of the food items. System efficiency was assessed by having REs perform independent assessments on the same set of simulated meals using either the new or traditional hospital method for which the REs had to document each food item served and then find the items in a computer database–steps that were automated in the new method. Interrater reliability for energy intake estimated on clinic wards was excellent (intraclass correlation coefficient = 0.975, 95% CI 0.958 to 0.992) and there was a high level of agreement between the REs’ estimates and the true values determined by food weighing; per the method of Bland and Altman the mean difference between the two types of estimates was 0.3 kcal (95% CI, −8.1 to 8.7 kcal) with limits of agreement of −79.5 kcal to 80.1 kcal. Compared to the traditional method, energy intake assessments could be completed using the multi-component method in less than a third of the time. These results indicate the multi-component method is an accurate, reliable, and efficient method of obtaining energy intake assessments for hospitalized patients.

No MeSH data available.