The Validity of Nutrition Screening Initiative DETERMINE Checklist Responses in Older Georgians

Abstract

The Nutrition Screening Initiative DETERMINE Checklist (NSI) is used throughout the United States to assess nutrition risk of those requesting the services of the Older Americans Act Nutrition Program (OAANP). This study examined the ability of the NSI to evaluate nutrition risk by comparing the responses between NSI and matched comparable survey questions using the self-administered mail survey data that were collected among 924 new OAANP applicants in Georgia (mean age 75.0 ± 9.2 years, 68.8% women, 26.1% Black). Ninety-four percent of our sample provided at least one discordant response (i.e., disagreement between responses to the NSI and matched questions). Questions regarding food intake most frequently yielded discordant responses. Black participants were more likely to provide discordant responses for the meal frequency question. Food insecure individuals were less likely to provide discordant responses for food intake questions. Those who lived alone were less likely to provide discordant responses for the dairy intake question. Some NSI items may have limited ability to reliably assess nutritional risk of older adults. Further efforts are warranted to improve nutritional assessment tools for use in vulnerable older adults.

KEYWORDS:

• Nutrition Screening Initiative
• Older Americans Act Nutrition Program

Notes

Note. Some survey data was modified (i.e., converted from 3–7 answer categories to 2) to match the answer categories of the NSI questions.

*GANSI = Georgia Nutrition Screening Initiative Checklist.

*FPL = federal poverty level.

*GANSI = Georgia Nutrition Screening Initiative Checklist.

†Sensitivity was calculated: (TP/(TP + FN)) × 100.

†Specificity was calculated: (TN/(TN + FP)) × 100.

^§True positives (TP) responded “yes” to GANSI and GANSI-match data. False positives (FP) responded “yes” to GANSI, but “no” to GANSI-match data.

^∥True negatives (TN) responded “no” GANSI and GANSI-match data. False negatives (FN) responded “no” to GANSI-match data, but “yes” to GANSI.

^¶Positive predictive value was calculated: (TP/(TP + FP)) × 100.

**Significant discordance based on McNemar's test at p < 0.05.

*GANSI = Georgia Nutrition Screening Initiative Checklist.

†F&V = fruits and vegetables.

‡OR = odds ratio.

^§CI = confidence interval.

^∥CM = congregate meals.

^¶HDM = home delivered meals.

This work was supported by the U.S. Administration on Aging Advanced Performance Outcomes Measures Project (grant number 427-93-08081298-99 to J.S.L.) and the United States Department of Agriculture/Economic Research Service/University of California Davis RIDGE program (grant number SUB07-000687-03 to J.S.L.).