Development and validation of a sodium AnaLysis tool (SALT)

Abstract

Abstract Background Sodium is an essential nutrient; however, excess dietary sodium is associated with increased blood pressure levels. The 2004 Canadian Community Health Survey – Nutrition (CCHS 2.2) concluded that most Canadians exceeded the Tolerable Upper Intake Level (UL) of 2300 mg/day. The 2015 CCHS indicated that Canadians were still consuming above the UL. To assess population sodium intakes, a Sodium AnaLysis Tool (SALT) was developed. Methods We used data from CCHS 2.2 (2004) to group foods into types (e.g., popcorn, crackers) and general categories (e.g., snack foods) which formed the SALT questions. Portion sizes and sodium values were calculated for SALT questions. Over a one-month period, one hundred participants completed three, 24-h recalls (at beginning, middle, and end) and two SALT (SALT1 & SALT2) tools (at beginning and end). To assess both validity and reliability, statistical tests including Bland-Altman (B-A) plots, paired t-tests, differences between means, and correlations were conducted. The mean of the 3,24-h recalls (m24HR) was used for validation. Results Validity testing between SALT2 and the m24HR yielded variable results. A B-A plot between SALT2 and m24HR depicted a small bias of 7 mg/day of sodium. The sodium intake for m24HR (2742 ± 980 mg/day) (mean ± standard deviation) versus SALT2 (2735 ± 1174 mg/day) was not significantly different (p = 0.960). Pearson’s correlation between methods, although significant (p = 0.02) was poor (r = 0.202; de-attenuated r = 0.400). There was a fair, significant agreement (κ = 0.236, p = 0.02) for the classification of sodium intake into two categories (above or below the UL). Test-retest reliability results were also variable. There was moderate, significant agreement (κ = 0.488, p = 0.001) for classification of sodium intake into two categories between SALT1 and SALT2, a significant correlation (Pearson’s r = 0.785, p < 0.001), and the B-A plot depicted good agreement. However, the values for sodium intake for SALT1 (3185 ± 1424) vs SALT2 (2735 ± 1174) were significantly different (p = 0.005). Conclusions Results indicate that the SALT has the potential to be a valid and reliable tool for assessing dietary sodium intake of Canadian adult populations. Despite some classification issues, there may be some value in using the SALT to categorize sodium intakes. Further refinement of the SALT may be required.