The Science of Salt: A focused review on salt‐related knowledge, attitudes and behaviors, and gender differences

Abstract

The aim of the current review was to examine the scope of studies published in the Science of Salt Weekly that contained a measure of self‐reported knowledge, attitudes, and behavior (KAB) concerning salt. Specific objectives were to examine how KAB measures are used to evaluate salt reduction intervention studies, the questionnaires used, and whether any gender differences exist in self‐reported KAB. Studies were reviewed from the commencement of Science of Salt Weekly, June 2013 to the end of August 2017. Seventy‐five studies had relevant measures of KAB and were included in this review, 13 of these were salt‐reduction intervention‐evaluation studies, with the remainder (62) being descriptive KAB studies. The KAB questionnaires used were specific to the populations studied, without evidence of a best practice measure. 40% of studies used KAB alone as the primary outcome measure; the remaining studies used more quantitative measures of salt intake such as 24‐hour urine. Only half of the descriptive studies showed KAB outcomes disaggregated by gender, and of those, 73% showed women had more favorable KAB related to salt. None of the salt intervention‐evaluation studies showed disaggregated KAB data. Therefore, it is likely important that evaluation studies disaggregate, and are appropriately powered to disaggregate all outcomes by gender to address potential disparities.

1. INTRODUCTION

Cardiovascular disease (CVD) is a leading cause of mortality and morbidity globally.1 As high blood pressure is a leading risk factor for CVD,2 reducing blood pressure at a population level is vital if rates of CVD are to reduce. There is a large body of high‐quality evidence showing an increased risk of elevated blood pressure with high salt intake.3, 4 Therefore, the World Health Organisation of the United Nations established a voluntary target in 2013, to reduce mean population salt intake by 30% by 2025.5 Progress towards this target could potentially be threatened by research of variable quality that is sometimes contradictory.6 Because of this, Science of Salt Weekly was established to assist scientists, clinicians, and policymakers to keep up with evidence about the effects of dietary salt on disease and the implementation of salt reduction interventions. This assistance is done through communication of a MEDLINE search by a weekly email bulletin and regular reviews.7, 8

The growing database of studies included in Science of Salt Weekly provides the opportunity to examine different themes within population salt reduction more closely. This review is the first in the series to take a thematic approach and focuses on the studies reporting knowledge, attitudes, and behaviors (referred to as KAB) relating to salt. People's knowledge and attitudes influence their food choices and behaviors, all of which may also be influenced by education, socioeconomic status, and sex/gender.9, 10, 11, 12 At a population level KAB are targeted, generally through education interventions, to try and influence how individuals within the population act. Some guidelines on how to construct and conduct a KAB questionnaire have been published,13, 14, 15 yet there is no gold standard KAB measure for salt intake. Due to the range of eating and cooking practices, sources of salt in different diets, and dietary patterns around the world, it is unlikely that a single questionnaire will be suitable for all populations. Likewise, it is unlikely that all groups of the population would have the same level of KAB.

The importance of showing data disaggregated by sex and/or gender for all specified outcomes, acknowledging that sex/gender can influence outcomes and therefore hold implications for study interventions is also increasingly recognized.16, 17, 18 Men traditionally had higher rates of CVD than women. However, CVD is also a leading cause of death and disability for women globally.16, 19 Many salt intake studies report overall population intake despite the fact that in most countries men consume significantly more salt than women.20 While higher salt intake may be attributed to higher energy intake, actual differences between KAB in men and women might also impact intake levels, although the fact that women are more likely to report health‐promoting behaviors than men,21 also needs to be taken into account.

Therefore, the objectives of this review are 3‐fold:

• To understand the use of KAB measures in evaluation studies

• To describe the types of questionnaires used to assess KAB

• To assess if gender differences in KAB are evident

The results will strengthen our understanding of how KAB surveys could be better used to inform and monitor future salt reduction intervention studies.

2. METHODS

Science of Salt Weekly newsletter was established in 2013, publishing weekly summaries of articles related to dietary salt reduction in populations, found through a standardized MEDLINE literature search.7, 8 For this paper, studies published in Science of Salt Weekly from its initiation, June 26, 2013, until August 22, 2017, were screened by title and abstract by 2 reviewers against the following inclusion and exclusion criteria; all studies reporting and/or evaluating self‐reported salt‐related KAB for populations of all ages and health status were included. Questionnaire validation studies and review papers such as narrative reviews or systematic literature reviews were excluded. Further, studies without a KAB measure or that measured KAB, but did not present measures in a usable format (for example composite scores with other health factors, that cannot be individualized into specific KAB) were also excluded.

Studies were identified as either descriptive (self‐reported salt‐related KAB of a population [cross‐sectional and baseline studies]) or evaluation studies (studies evaluating the impact of a salt reduction intervention through self‐reported salt‐related KAB). Data were extracted following a data extraction table designed for this review and specific for study type. Notes were taken on the type of questionnaire used to measure KAB and the presence of data disaggregated by gender for KAB outcomes.

A descriptive checklist was developed to be able to compare characteristics of KAB measures used between studies and to describe how they were used as part of intervention evaluation (Appendix 1). This checklist was used for each evaluation study, assessed and discussed by 2 reviewers. The number of KAB questions were counted by a reviewer and discussed with a second reviewer; questions with subcategories were considered 1 question, and question assessment was only made by the information provided in the paper.

3. RESULTS

From the 626 studies published in Science of Salt Weekly between June 26, 2013, and August 22, 2017, 108 studies were identified by title and abstract search and assessed against the inclusion/exclusion criteria. Seventy‐five studies met the criteria and were included in this review (Figure 1). Thirteen studies were intervention studies that incorporated an element of salt‐related KAB in the evaluation of the intervention, 62 were descriptive studies (cross‐sectional surveys) measuring self‐reported salt‐related KAB. Key results from both sets of studies are described below.

Figure 1.

Studies included in the current review, June 2013‐August 2017

3.1. Evaluation studies

3.1.1. Study characteristics

Of the 13 intervention studies that used KAB as a measure of evaluation, 5 were conducted in low and middle‐income countries, and 6 were specifically in unwell populations (5 studies had hypertensive/pre‐hypertensive participants and one involved heart failure patients). The study population ranged from 11 to 1903 participants, with 2 specifically targeting older people (≥60 years).22, 23 The majority of the studies were conducted in 2016 (8 of 13; Table 1). Two studies included women only, and 1 study included men only. Of the remaining 10 studies, none reported data disaggregated by gender for the KAB outcome measures.

Table 1.

Summary of salt reduction evaluation studies with self‐reported knowledge, attitudes, and behavior (KAB) as an outcome measure

Study (author, year)	Country	Aim and intervention	Theoretical framework of intervention	Design	Population	Outcome measures	Key results
Do et al. (2016)25	Vietnam	To evaluate the effects of a community‐based salt reduction intervention on mean population salt consumption, consumer knowledge and behaviors towards salt, and health outcomes. The intervention consisted of targeted communication through media, schools and community programs	COMBI	Cross‐sectional survey pre‐ and post‐uncontrolled community intervention	509 participants at baseline and 511 at follow‐up aged 25‐64 y, 51% female at baseline and 54% female at follow‐up	Primary; salt intake estimated by spot urine samples with a sub‐sample of 24 h urine collection. Secondary; blood pressure, hypertension, salt consumption‐related knowledge and behaviors	Community based intervention using the “COMBI tool.” Significant increase in positive knowledge and behaviors. Significant decrease in salt intake for the sub‐sample (n = 88 baseline, n = 73 follow‐up) with 24 h urine measures of 1.99 g/d
Land et al. (2016)26	Australia	To evaluate the effects of a community‐based salt reduction intervention and promotion of a smartphone application on mean population salt consumption, consumer knowledge and behaviors towards salt, and health outcomes. The intervention consisted of targeted communication through community channels, advertising, information sessions and “point‐of service” contact. Promotion of a smartphone application and salt substitute was emphasized	COMBI	Cross‐sectional survey pre‐ and post‐ uncontrolled community intervention	426 participants at baseline and 450 at follow‐up. 137 participants completed both baseline and follow‐up measures, aged 20 y or older, 58% female at both measured points	Primary; Salt intake measured by 24 h urine excretion. Secondary; salt‐related knowledge, attitudes and behavior	Increase in positive knowledge and behaviors (4, 2 k and 2 b). No change in 4 (1 k, 1 A, 2B). Decrease in 2 (2 B). 0.8 g/d decrease in salt intake
Seo et al. (2016)22	Republic of Korea	To evaluate effectiveness of a 1 mo salt reduction intervention in a meal centre targeted at an elderly population. The intervention consisted of a low salt meu and education sessions on salt reduction in a congregate meal service centre	Not stated	Uncontrolled community intervention	74 participants, low income elderly, 62% female, 41 participated in the education sessions, 33 did not	Survey of 74 participants and interviews with the foodservice providers post ‐intervention	One measure reported on positively; self‐reported reduction in soup and kimchi consumption post intervention
Lofthouse et al. (2016)27	New Zealand	To examine the feasibility of following a low sodium diet and evaluate its effects on other nutrients. Intervention consisting of an intensive low‐salt diet education, delivered by dietitians for 4 wk. The nutritional education sessions were one‐to‐one and covered health issues associated with excess dietary sodium intake, identification of foods typically high and low in sodium, training on how to read and interpret Nutrition Information Panel	Behavioral change theory	Mixed methods pilot study	11 healthy adults, aged 18‐65, 46% female, not already restricting their sodium intake	Primary; salt intake measured by 24 h urine excretion pre‐ and post‐ intervention. Secondary; 2 d weighed diet record and a semi‐structured questionnaire post intervention	Used in evaluating by summarising opinions and “how” you could actually make the intervention work (identifying barriers and facilitators). 18% decrease in sodium intake following intervention, main themes; trying to meet recommendations required a new way of eating, replacing foods, changes in taste perception, found eating out challenging, found NIP reading important but some struggled maintain this practice
Li et al. (2016)24	China	To evaluate whether a reduced‐sodium added‐potassium salt substitute intervention would reduce average population sodium intake in rural China. Intervention, community based health education delivered by township health educators with assistance from village council and village doctors‐ public lectures, distribution of promotional materials and other education sessions. Salt substitute was made available for purchase. Half the intervention villages were randomized to subsidisation of the price of salt substitute. Control villages received usual care	Not stated	Cluster‐randomized controlled trial	2566 participants in 119 villages (60 intervention and 59 control villages). 50% female	Primary; salt intake measured by 24 h urine excretion. Secondary; sodium to potassium ratio, salt related knowledge, attitude and behavior and blood pressure	Greater positive knowledge (3 questions) and attitudes (1 question) in intervention vs control group. Significantly less sodium excretion in intervention compared to control. Community based health education intervention
Cornélio et al. (2016)32	Brazil	To pilot and test the effect of an intervention focused on motivation and the behavior of limiting salt use during cooking. Two behavioral change techniques were used‐ consciousness raising and counterconditioning based on using 2 models: the transtheoretical model and social cognitive theory. Intervention was two 60‐90 min group (n = 10) sessions run by 2 nurses, each followed by a telephone call to reinforce session content. The control group received usual care	Transtheoretical model and Social Cognitive Theory	Randomized control trial	92 hypertensive women aged 40‐80 y, diagnosed for hypertension for at least 6 mo, responsible for meal preparation and ate at least 5 meals per week at home	Primary; Questionnaire‐ behavioral, self‐reported salt use iin cooking, questions about use of other seasonings, psychosocial variables, including intention, self‐efficacy and habit. Secondary; 24 h urine excretion	Positive change in salt‐related behaviors in the intervention vs control group. Urinary salt excretion decreases but difference between IG and CG not significantly different
Irwan et al. (2016)23	Indonesia	To test the effects of 2 salt‐reduction training, and an efficacy‐maintenance program in improving knowledge, attitudes, self‐care practices and self‐efficacy in reducing salt intake. A salt reduction training (SRT) group received educational training to improve self‐care and self‐efficacy. A salt reduction and efficacy‐maintenance (SREM) group received 2‐d educational training in 1 wk and a 90‐min maintenance meeting after 1 mo of training. The control group received no intervention	Geragogy learning model/self‐efficacy theory	Randomized control trial	51 participants with hypertension/prehypertension aged 60 y and above. Patients with low cognition, hearing difficulties, sight problems, and other chronic illnesses were excluded	Knowledge of hypertension, attitudes towards self‐care and measure of self efficacy regarding reducing salt intake. Responses to open‐ended questions were analysed. Secondary; salt concentration in food, and urinary salt excretion	Increase in positive knowledge, attitudes and behaviors. Following intervention there was no difference in salt excretion (significantly different at baseline)
Park et al. (2016)33	South Korea	To examine the feasibility of a sodium reduction program implemented in restaurant owners and its effect on sodium intake and attitude towards providing consumers with healthy foods. The pilot intervention study involved nutrition education sessions which covered the health risks of a high sodium intake, and tips for lowering the sodium content in foods served in restaurant and home‐cooked foods. These sessions were based on a theoretical framework‐ health belief model and social cognitive theory. The education and counselling sessions were two 1‐h sessions	Health belief model and Social Cognitive Theory	Uncontrolled community intervention (pilot study)	48 restaurant owners and cooks aged between 29 and 63 y participated in the pilot study	Change in salt related behavior and spot urine dipstick tests to estimate sodium intake	Increase in positive behaviors (urine dipstick used to separate people into high salt intake vs low salt intake not to evaluate before/after)
De Keyzer et al. (2015)86	Belgium	To quantify sodium intake and evaluate the effects of the educational intervention with feasibility of this intervention. Intervention consisted of a printed educational materials and dietary advice to reduce their sodium intake. Additional. Personalized advice was received 3 times of the 28‐d study period	Not stated	Cohort intervention study	25 adult participants with hypertension, 60% female	Primary; salt intake measured by 24 h urine excretion, and blood pressure. Secondary; self‐reported salt consumption and use	No statistical analysis given for KAB measures, salt intake decreased by 1.6 g/salt/d
Enkhtungalag et al. (2015)31	Mongolia	To assess the “Pinch Salt Magnolia” campaign, relevant component aimed at reducing salt intake of employees at 3 main food factories. The intervention was multi‐factorial consisting of; staff training on the negative health impact of salt, how to consume a healthy diet and the provision of reduced salt foods in the company canteens/kitchens	Not stated	Pilot implementation study	240 participants, factory workers across 3 factory sites	Salt intake measured by 24 h urine excretion and self‐reported salt‐related knowledge attitudes and behaviors	Positive change in salt‐related knowledge and a decrease in salt intake by 3.06 g/d (over a 2 y time period)
de Freitas et al. (2014)29	Brazil	To assess the effectiveness of the Implementation Intentions strategy on reducing salt intake among hypertensive Brazilian women. Implementation intentions intervention; consisted of the intervention group receiving information on the benefits of aa low salt diet, 2 “in person” consults and 2 phone call consults discussing salt reduction and individual plans for salt reduction. The control group received usual health care	Implementations intentions theory	Randomized control study	98 hypertensive women aged, 18 or older with a hypertension diagnosis of at least 6 mo, responsible for meal preparation. 49 randomly assigned to the intervention and 49 to the control group	Primary; self‐reported salt consumption, specifically discretionary salt use and 24 h urinary sodium measure. Secondary endpoints; intention, self‐efficacy and habit of discretionary salt use	No change in KAB (high levels of KAB at baseline). 1.7 g/d reduction in salt intake
Welsh et al. (2013)28	America	To evaluate short and long‐term impacts of an educational intervention to reduce sodium intake in patients with heart failure. Educational intervention based on the Theory of Planned Behavior. The intervention group received instructions and support on a low sodium diet through home visits and phone calls, the control group did not receive dietary advice	Theory of planned behavior	Randomized control trial	52 participants aged 21 y or older with heart failure, 24 women and 28 men	Primary; dietary sodium intake measured by 3‐d food record. Secondary; Attitudes, subjective norm and perceived behavioral control measured through the Dietary Sodium Restriction Questionnaire. Both measured at baseline, 6 wk and 6 mo	Increase in positive salt‐related attitudes and perceived behavioral control, mean subscale scores at 6 wk positive correlated with dietary sodium intake at 6 mo Dietary sodium intake measured by food diaries decreased in intervention group at 6 mo (not 6 wk)
Kitaoka et al. (2013)34	Japan	To evaluate the effect of a dietary education program aimed at reducing dietary sodium intake and increasing dietary potassium intake. The intervention was education based, with a lecture and cooking session held on 5 occasions by a registered dietitian. The control group received usual care	Not stated	Non‐randomized controlled study	71 hypertensive men aged 40‐75 y with stage 1 or stage 2 hypertension. 39 men were assigned to the intervention group and 32 to the control group	Dietary habits questionnaire, food frequency questionnaire, spot urine test (sodium/potassium ratio) and blood pressure	Increase in positive behavior (1 question) and “salt restriction awareness” difference in sodium to potassium ratio; difference of 0.6 P = .03

COMBI, Communication for Behavioral Impact Framework.

3.1.2. Interventions and KAB measurement tools

Interventions evaluated varied, yet all focused on education—3 used a multi‐component intervention, 1 with the availability of a salt substitute and the testing of a subsidy24 and 2 using the “Communication for behavioral impact planning and implementation tool (COMBI).”25, 26 In addition to COMBI, Land et al26 promoted the smartphone application “Food Switch.” Other interventions included, individualized nutrition education and counselling,27, 28, 29, 30 group education,23, 31, 32, 33 group educational lectures with cooking classes,34 and group education with sodium reduced options from the foodservice.22 Eight of these interventions were designed using referenced theory (Table 1), with all 8 studies having a positive change in either KAB or salt intake measure.

The KAB tool used and the method for its use in evaluation are shown for each study in Table 2. While the majority were conducted pre‐ and post‐intervention 3 were only conducted post‐intervention. Seven studies used the KAB self‐report as the primary measure for intervention evaluation, yet only 1 study, Seo et al,22 used it as the only reported outcome measure. Eight studies provided information on the KAB tool used, and 5 provided information on tool pilot‐testing, of which 4 were tested for the predefined use. The method of answering questions varied, 5 used a Likert scale in combination with another method (open‐ended response, n = 2, yes/no answers, n = 3), and 3 used 3‐or‐more methods of answering. Studies reported on average 2 knowledge, 3 attitude, and 6 behavior questions. Examples of the types of questions used are provided in Table 3.

Table 2.

Comparison of the knowledge, attitudes, and behavior (KAB) questionnaires used

Study	How the Questionnaire was used			Questionnaire used						Reporting of outcomes
	Administration	Frequency of tool use	Primary evaluation measure (Yes,/No)	Questionnaire	Piloted testing	Number of Questions in each domain			Method of answering	Data disaggregated by gender
						Knowledge	Attitude	Behavior
Do et al.	Interview, in person	Pre‐ and Post‐intervention	N	NP	NP	2	‐	3	Mixed methods	NP
Land et al.	Interview, in person	Pre‐ and Post‐intervention	N	Adapted from the WHO tool	NP	4	2	3	Mixed methods	NP
Seo et al.	Interview, in person	Post intervention	Y	Referenced KAB tool from 3 previous studies in Korea	Y	2	8	4	Likert scale	NP
Lofthouse et al.	Interview, in person	Post intervention	N	NP	NP	‐	2	2	Open ended responses	NP
Li et al.	Interview, in person	Post intervention	N	NP	NP	3	1	4	Mixed methods	NP
Cornelio et al.	Interview, in person	Pre‐ and Post‐intervention	Y	Questions from referenced tools developed by authors in Brazil	Y	‐	2	4	Likert scale	All women
Irwan et al.	Interview, in person	Pre‐ and Post‐intervention	Y	Questions from referenced tools—not all related to salt	NP	12	7	10	Likert scale and open ended responses	NP
Park et al.	Self‐administered	Pre‐ and Post‐intervention	Y	Questionnaire adapted from previous referenced surveys and the Korea Community Health Survey	NP	‐	2	14	Likert scale and Yes/No answers	NP
De Keyzer et al.	Self‐administered, paper	Pre‐ and Post‐intervention	N	NP	NP	1	2	‐	NP	NP
Enkhtungalag et al.	Interview, in person	Pre‐ and Post‐intervention	Y with 24 h urine sample	Adapted from the WHO tool	NP	1	1	5	Likert scale and Yes/no answers	NP
De Freitas et al.	Interview, in person	Pre‐ and Post‐intervention	Y with 24 h urine sample	Questions from referenced tools developed by authors in Brazil	Y	‐	9	10	Likert scale	All women
Welsh et al.	Self‐administered, paper	Pre‐intervention, 6 wk and 6 mo	N	Referenced KAB tool	Y	‐	7	12	Likert scale and open ended responses	NP
Kitaoka et al.	Self‐administered, paper	Pre‐ and Post‐intervention	Y with spot urine samples	NP	NP	1	‐	5	Likert scale and Yes/no answers	All men

NP, not presented in the study, WHO, World Health Organization.

Table 3.

Examples of knowledge, attitude and behavior (KAB) questions used in salt reduction evaluation studies

Study	Knowledge	Attitudes	Behaviors
Do et al.	Knowledge on consequnces of high salt intake (High salt intake can cause; hypertension, stroke, heart attack)	NA	Discuss salt reduction with others (never, several times, many times)
Land et al.	Can high salt cause serious health problems? (Y/N)	How important is lowering salt in your diet? (Y/N)	Do you avoid processed foods? (Y/N)
Seo et al.	“I think it is necessary to make an effort to reduce salt intake”	“I am satisfied with the taste of dishes that I've received in congregate meal service center during the healthy eating program”	“I ate less Kimchi than I used to during the healthy eating program”
Lofthouse et al.	NA	“How did you find trying to consume a low sodium diet?”	Did you make any other changes to your food choices over this time?
Li et al.	Knowledge that excess salt intake is harmful	Concerned about salt in diet	Household use of salt substitute
Cornelio et al.	NA	“I intend to use <4 g/d of salt during cooking over the next two months”	“In the last two months, what better describes your behavior of using <4 g of salt per day (corresponding to one “flat’ teaspoon of salt) during the cooking of all daily meals (ie, breakfast, lunch, dinner and snacks)”
Irwan et al.	12 point questionnaire on knowledge related to salt and hypertension	Belief that salt reduction can help control blood pressure	Add salt to meals (Y/N)
Park et al.	NA	“I think it is necessary to use healthy cooking methods for customers”	“I use less salt when cooking and let customers use additional salt if necessary”
De Keyzer et al.	Knowledge of their own sodium consumption	Felt that their were barriers to eating a low salt diet (eg, consuming less cold‐ cut meats)	NA
Enkhtungalag et al.	“Do you think that too much salt in your diet could cause a serious health problem?”	“How important to you is lowering the salt in your diet?”	“How often do you add salt to your food before you eat it or as you are eating it?”
De Freitas et al.	NA	“I trust in my ability to add less than one teaspoon of salt a day when cooking all my meals: definitely not… definitely yes”	“adding more than one teaspoon of salt a day when cooking all my meals is something I do frequently: definitely not… definitely yes”
Welsh et al.	NA	“It is important for me to follow my low‐salt diet”	Instructed to rate how much items such as “the cost of low salt food” and “the taste of low‐salt foods” kept them from following a low sodium diet
Kitaoka et al.	Salt restriction awareness; usually, sometimes, rarely	NA	Use of soy sauce and salt at thek; often, sometimes, rarely

NA, not applicable, questions on this construct (knowledge, attitude, or behavior) were not used or presented by authors of the study.

3.1.3. KAB outcomes

Outcomes from the interventions varied (Table 1), 10 of 13 studies reported a significant positive change in at least one of their predefined measures of KAB, of these 10 studies, 6 also reported a significant decrease in measured salt intake (measured by 24‐hour urine, spot urine test, or diet recall). Seven studies reported findings where the KAB outcome was not consistent with measured salt intake. Three reported a decrease in measured salt intake, but no significant change in KAB (or without having quantified the change in salt‐related KAB) and 4 studies reported positive changes in KAB without a significant difference in measured salt intake. The 8 studies that were designed using referenced theory showed a positive change in KAB,23, 32, 33 a reduction in salt intake,27, 29 or both.25, 26, 28 Three of the 5 studies conducted in low‐middle income countries showed significant reductions in salt intake as measured by 24‐hour urine (1.99, 3.06, and 1.70 g/d in Vietnam,25 Mongolia,31 and Brazil,29 respectively), 2 of which also showed positive changes in KAB.25, 31 The outcomes from the 2 studies conducted only in women were discordant with 1 showing positive changes in KAB, yet no change in salt intake,32 and the other having a significant reduction in salt intake, but no change in KAB.29 The study conducted only in men found a positive change in the salt‐related behavior question, with a change in sodium to potassium ratio.34

3.2. Descriptive studies

3.2.1. Study characteristics

The majority of studies containing an element of self‐reported salt‐related KAB within the defined period came from descriptive studies (cross‐sectional and baseline studies), Table S1 (with references 52–111). Of these, only 20% included populations from low‐and‐middle‐income countries (13 studies), 21 studies focused on unwell populations (those with hypertension, diabetes, chronic kidney disease, cardiovascular disease), 2 focused on older populations (>50 years), and 2 included children and adolescents. There was a range of study population size from 16 to 173 778 participants.

3.2.2. KAB tool used and gender differences in outcomes

As with the evaluation studies, tools used to measure KAB differed: 4 studies specifically stated using an adapted version of the World Health Organization tool35 and 17 referenced a specific alternative KAB tool or multiple tools. The majority developed their own tool, with only 8 stating being pilot‐tested.

Of the 62 studies, 58 included data on men and women, of these less than half (45%, 26 studies) presented data disaggregated by gender for the predefined knowledge, attitude, and behavior outcomes. Twenty‐two studies reported a difference between men and women, with 16 of these identifying more positive study specific salt‐related KABs in women compared to men, 3 studies showing more positive KABs in men, and 3 stating general differences (eg, levels of KAB differed between men and women for different questions within the study [Table S1]).

4. DISCUSSION

In this first science of salt review that takes a thematic approach, we examined studies published in Science of Salt Weekly that contained a measure of salt‐related, self‐reported KAB. We found a wide range of tools used across 75 studies, including 62 descriptive studies and 13 evaluation studies. Although over a third of the descriptive studies showed gender differences in KAB outcomes, KAB data were not disaggregated by gender in the evaluation studies.

Many of the studies used theories and frameworks to inform interventions. This review is consistent with the findings of a previous review that demonstrated that behavior change intervention based on theoretical models or frameworks may be effective in changing behavior.36 Of the 8 studies that developed interventions based on theoretical models or frameworks, all reported improvements, either in KAB, measured salt intake, or both. Some theoretical models require an understanding of the target populations KAB as formative research, and as such may be able to use more tailored questions, likely making the association between measured KAB and actual behavior change closer. For example, Park et al33 conducted a study specifically in restaurant owners, with the intervention based on the Health Belief Model (also known as the Social Cognitive Theory).33, 37 This model aims to build self‐efficacy in a given environment, in this case, a Chefs’ perceived ability to produce low‐sodium food for their customers. The intervention comprised of education sessions with lessons on reducing sodium in foods produced in their restaurants and daily lives. Based on this intervention, specific questions regarding perceived self‐efficacy and behavior were asked (eg, “I check sodium contents frequently using a salimeter when serving foods.”) that showed a significant increase in reported behavior at follow‐up. Studies that used frameworks also had positive outcomes, although the direct link with KAB is less clear. Do et al25 and Land et al26 both employed COMBI frameworks, which provide guidance for communication areas for intervention.38 Interventions were multifaceted and altered for the study locations, regions within Vietnam and Australia, respectively. Both found reductions in measured salt intake and improvements in the generalized KAB questions asked, yet authors could only hypothesize which aspects of the intervention were responsible for the positive outcomes. Therefore, the usefulness of KAB questionnaires as an outcome measures of the effectiveness of interventions is likely study‐dependent.

The majority of the descriptive studies that presented data disaggregated by gender showed a difference in KAB between men and women, yet disaggregated data were not presented in the evaluation studies. While this may be because KAB is reported as a secondary outcome in many of the evaluation studies, others have relatively small numbers and lack the power to examine differences. A difference in the self‐report of dietary KAB in a variety of nutrition‐related contexts between men and women is documented in the literature, with women reporting more health‐conscious knowledge, attitudes, and behaviors,9, 21, 39 along with studies showing a link between self‐reported and actual intake.40, 41 Therefore, the fact that gender disparities are not being reported is potentially problematic and further emphasizes the need for appropriately powered studies to enable disaggregation of data.16

One of the limitations of KAB surveys is the potential difference between knowledge, attitudes, and actual behavior, and the potential difference between self‐reported behavior and actual behavior.42 While ideally behavior should be observed or objectively measured rather than self‐reported, in practice this is very difficult to achieve. Previous studies have investigated the association between self‐reported diet‐related knowledge, attitudes, and behavior with more objective measures of behavior. A review of the relationship between nutrition knowledge and dietary intake by Spronk et al40 had similar limitations to our study, with the heterogeneity of questionnaires used, which limited generalizability. They found that most studies showed some positive, yet weak, relationship between self‐reported knowledge and dietary intake, with higher rates of dietary knowledge in women.40 Attitudes are assessed as a key construct to behavior change. A study of 1256 Irish adults conducted by Hearty et al41 found the majority of individuals had positive attitudes towards healthy eating and perceived healthy eating behavior. Diet quality was assessed through a 7‐day food diary, and it was found that individuals who perceived their diets as healthy were more likely to comply to dietary guidelines, these individuals were also more likely to be older, women and of higher socioeconomic status.41 When assessing dietary intakes, biomarkers can be used to assess the difference between actual and self‐reported dietary intakes, with bias in self‐reported measures often reported.43 Other diet‐related behaviors are also prone to bias. A few studies have shown that people over‐estimate use of nutrition labels, comparing self‐reported and observed behavior.44, 45 Therefore, it is important to include measures of behavior such as observed behavior or intake biomarkers and not just knowledge and attitudes, and understand that self‐reported behavior may differ from actual behavior in interpreting KAB data.

Targeting of the KAB questionnaire to the main dietary sources of salt that you are trying to reduce within the population, and behaviors associated with that source of salt is important. Where discretionary salt is the major source, this may result in a stronger association between self‐reported and actual behavior. Johnson et al46 illustrated a strong relationship between salt intake and KAB in a large population study conducted in India, hypothesizing that this is due to the main contributor of salt to their diet being discretionary salt, similar to other Asian and low‐middle income countries.20 By contrast, Land et al47 in a smaller Australian population found no association between KAB and measured salt intake. In Western‐style diets, most of the salt consumed is in processed food making it difficult to self‐regulate.47 Foods that contribute substantially to salt intake in Western diets, including bread, cereals, and cakes are not obviously salty, which contributes to a lack of awareness of salt intakes in this context and requires a different range of interventions and KAB questions.

A range of different KAB tools is noted in this review based on population, including health status and sources of salt in local diet and intervention. Studies implementing interventions in unwell populations (hypertension or heart failure) had questions more targeted to the health condition and related KAB,23, 28, 29, 30, 32, 34 whereas studies in healthy populations were more likely to use broader questionnaires, such as the WHO KAB tool.24, 25, 26, 31 While this has made it difficult to generalize results in this review, it seems appropriate that the tool be tailor‐made to fit the specific population and context.

Our review is descriptive due to the heterogeneity of studies included, and we were not able to compare questions from specific questionnaires due to a lack of information provided by some papers. This further limited our ability to make recommendations on KAB measures and investigate how responses to certain questions are related to salt intake at a population level. In the current review, we did not assess quality as the majority of the salt reduction evaluation studies included here have been reviewed in previous Science of Salt reviews.48, 49, 50, 51 Also, given the focus was on KAB and there is no gold standard KAB measure with which to compare, we did not feel we could assign quality. As KAB is a subjective measure, there is a risk of bias, including recall and social desirability bias. Acknowledging this limitation we developed a tool (Appendix 1) to describe the KAB tool using factors thought to be important in KAB development,13, 14, 15 we propose that this may be a useful way to evaluate KAB questionnaires.

5. CONCLUSIONS

In this thematic review, we have explored the use of self‐reported, salt‐related KAB measures in salt‐reduction intervention‐evaluation studies and descriptive studies between June 2013 and August 2017. We have identified a variety of KAB questionnaires used, and without evidence of a best practice measure, conclude that this is likely appropriate given the differences in populations studied. While this is the case, we have developed a tool to evaluate KAB questionnaires and propose that it could be referenced when developing KAB surveys. Additionally, given the demonstrated differences in KAB between men and women in the descriptive studies, disaggregation of outcome variables in evaluation studies by gender is likely important to inform future salt‐reduction programs and policies.

DISCLOSURES

JW is the Director of the World Health Organization (WHO) Collaborating Centre on Population Salt Reduction. Of the studies reviewed JW was an author on Grimes et al. (2017), Land et al. (2016), Enkhtungalag et al. (2015), Nowson et al. (2015), and Land et al. (2014); RM on Lofthouse et al. (2016) and McLean et al. (2014); JAS, KT and JW on Do et al. (2016); CJ, TSR and JW on Johnson et al. (2017); and KT, CJ and JW on Webster et al. (2016). We declare no further conflicts of interest.

Supporting information

 

APPENDIX 1. Knowledge, attitudes, and behavior (KAB) questionnaire evaluation tool

1.1.

Criteria		Comments
How the tool was used
Administration	Self‐administered (paper or online)
	Interview via telephone
	Interview in person
	Other (specify)
When was the KAB administered?	Post intervention
	Pre‐ and Post‐intervention or more
Was the KAB the primary tool used for evaluation of intervention?	Yes (if secondary measures please specify)
	No (please specify)
KAB questionnaire
Questionnaire used	Information on questions not provided
	1 or 2 Knowledge, Attitudes, and Behaviors questions not referenced
	1 or 2 Knowledge, Attitudes, and Behaviors questions referenced
	Unreferenced questionnaire
	Referenced KAB questionnaire (please specify reference)
	Tool based on the World Health Organization KAB Questionnaire
Piloting of questionnaire used	Measure not pilot‐tested in sample population/without convincing text regarding testing in the sample population
	Measure pilot‐tested in sample population/with convincing text regarding testing in the sample population
	Not enough information provided to assess
Were the questions/tool based on a model/framework	Yes
	No
	Unclear
Number of questions in each domain	Knowledge
	Attitudes
	Behavior
How were the questions answered?	Yes/no, True/false
	Multi‐choice
	Likert scale (strongly agree to strongly disagree)
	Mixture of the above
	Open ended response
Reporting of outcomes
Are all specified KAB measures reported in a complete and usable format	Yes
	No
Are KAB results reported disaggregated by gender	Yes
	No
Other comments on KAB use

McKenzie B, Santos JA, Trieu K, et al. The Science of Salt: A focused review on salt‐related knowledge, attitudes and behaviors, and gender differences. J Clin Hypertens. 2018;20:850–866. 10.1111/jch.13289