Accepted Manuscript

Use of the letter-based grading information disclosure system and its influence ondining establishment choice in Singapore: A cross-sectional study

Joel Aik, Anthony T. Newall, Lee-Ching Ng, Martyn D. Kirk, Anita E. Heywood

PII: S0956-7135(18)30084-7

DOI: 10.1016/j.foodcont.2018.02.038

Reference: JFCO 6001

To appear in: Food Control

Received Date: 3 November 2017

Revised Date: 25 January 2018

Accepted Date: 23 February 2018

Please cite this article as: Aik J., Newall A.T., Ng L.-C., Kirk M.D. & Heywood A.E., Use of the letter-based grading information disclosure system and its influence on dining establishment choice inSingapore: A cross-sectional study, Food Control (2018), doi: 10.1016/j.foodcont.2018.02.038.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service toour customers we are providing this early version of the manuscript. The manuscript will undergocopyediting, typesetting, and review of the resulting proof before it is published in its final form. Pleasenote that during the production process errors may be discovered which could affect the content, and alllegal disclaimers that apply to the journal pertain.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

1 | P a g e

Use of the letter-based grading information disclosure system and its influence on 1

dining establishment choice in Singapore: a cross-sectional study 2

3

Joel Aik a,b,*, Anthony T Newall a, Lee-Ching Ng b, Martyn D Kirk c, and Anita E Heywood a 4

a School of Public Health and Community Medicine, Faculty of Medicine, University of New South 5

Wales, Level 3, Samuels Building, Botany Road, Kensington, New South Wales 2052, Australia; 6

b National Environment Agency, 40 Scotts Road, #13-00, Singapore 228231 7

c National Centre for Epidemiology and Population Health, Australian National University, Canberra, 8

ACT 0200, Australia 9

10

*Corresponding author. 11

Email address: joel_aik@nea.gov.sg 12

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

2 | P a g e

Abstract 13

14

The aim of this study was to examine the consumer use of Singapore’s letter based grading 15

information disclosure system and its influence on dining establishment choice. We used 16

data from a national survey of 1,533 households collected from 2012 to 2013 in Singapore to 17

assess (i) the proportion of adults who refer to the letter grade before dining and (ii) the 18

impact of the letter grade on their willingness to dine at an establishment. We used 19

multivariable logistic regression to account for the independent effects of socio-demographic 20

factors. The proportion of respondents who referred to a letter grade before dining was 21

64.5% (95% confidence interval [CI] = 62.1%, 66.9%). Propensity for referral differed by 22

dining frequency, ethnicity and employment. Fewer respondents were willing to dine at a ‘C’ 23

(lower) graded establishment [10.3% (95% CI = 8.8%, 11.8%)] compared to a ‘B’ graded 24

establishment [85.3% (95% CI = 83.5%, 87.0%)]. Willingness to dine at a ‘C’ graded 25

establishment differed by dining frequency, housing type and citizenship. The letter based 26

grading information disclosure system in Singapore is commonly used among Singaporeans 27

and influences establishment choice. Our findings suggest that information disclosure 28

systems can be an effective tool in influencing consumer establishment choice and may be 29

useful to help improve food safety in retail food establishments. The implementation of such 30

information disclosure systems should be considered in other countries where it has yet to 31

be introduced and be periodically assessed for its effectiveness and to identify areas 32

requiring improvements. 33

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

3 | P a g e

1. Introduction 34

35

Food safety is becoming increasingly important. The World Health Organization estimates 36

that contaminated food causes 600 million individuals to fall ill (approximately 1 in 10 37

individuals globally), resulting in more than 420,000 deaths annually (World Health 38

Organization, 2015). The economic costs of food-borne illness are substantial. For example, 39

as one of the leading causes of food-borne illness, Salmonella is estimated to cost the 40

United States (US) and the European Union (EU) billions of dollars annually (Economic 41

Research Service (ERS), 2015; European Food Safety Authority, 2014). The consumption of 42

food produced outside homes has increased due to time scarcity (Jabs & Devine, 2006) and 43

as a result, the number of individuals who are more susceptible to food-borne illness is 44

expected to grow (Lund, 2015). 45

46

Several food safety strategies have been used to minimise the transmission of food-borne 47

pathogens. Food safety management systems (FSMS) aim to systematically identify and 48

eliminate physical, chemical and microbiological contamination in the production process to 49

ensure that food is safe for consumption (ISO, 2005). Legally mandating FSMS in food 50

establishments has been shown to improve food hygiene standards (Djekic et al., 2016). The 51

implementation of a food safety system can reduce microbiological contamination during the 52

food production process (Cusato et al., 2012). Dining establishment operators can achieve 53

higher hygiene standards with better food hygiene knowledge and more positive attitudes 54

towards food hygiene (Läikkö-Roto & Nevas, 2014). For example, in a hotel setting, food 55

handler training has been demonstrated to improve the safety of the food production process 56

(Gomes, Lemos, Silva, Hora, & Cruz, 2014). 57

58

One food safety strategy that has been introduced in several high-income countries is the 59

use of public information disclosure systems aimed at influencing where people eat. Such 60

systems aim to reduce the incidence of food-borne diseases by increasing consumer 61

demand for better hygiene standards in dining establishments. While not necessarily 62

mandatory, these information systems usually require dining establishments to publicly 63

display their assessed hygiene standards for consumers and are complementary to the 64

existing regulatory inspection regimes used by public health agencies. Studies have found 65

that posted hygiene grade cards or inspections scores can lead to improvements in the 66

hygiene standards of food establishments (Jin & Leslie, 2009; Vergeris, 2015; Waters et al., 67

2013; Wong et al., 2015), and that there is a positive consumer attitude towards food safety 68

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

4 | P a g e

certification in restaurants (Uggioni & Salay, 2014). Examples of such systems introduced 69

include the “Smiley Scheme” in Denmark in 2001 (The Danish Veterinary and Food 70

Administration, 2017), the Food Hygiene Rating Scheme in the United Kingdom (UK) in 2010 71

(Food Standards Agency (UK), 2015b), the restaurant letter grading program in New York 72

City, US in 2010 (McKelvey, Wong, & Matis, 2015) and “Scores on Doors” in New South 73

Wales, Australia in 2011 (New South Wales Food Authority, 2017). 74

75

The incidence of reported non-travel related food-borne illnesses in Singapore rose by 76

almost 200% from 15.9 to 47.0 per 100,000 population over a 13-year period (Ministry of 77

Health (Singapore), 2015). Singaporeans are consuming meals prepared away from homes 78

more than before; almost 2 in 3 dine out at least 4 times a week compared to 1 in 2 more 79

than a decade ago (Health Promotion Board Singapore, 2010). The exact disease incidence 80

attributable to dining establishments is unknown. However, dining establishments are known 81

to be an important source of foodborne diseases (Gormley et al., 2011; Gould et al., 2013; 82

Jones & Angulo, 2006; Park, Kwak, & Chang, 2010) and consumers have to rely on 83

establishments to ensure that the meals that they prepare and sell are safe for consumption. 84

85

Singapore introduced its food hygiene grading system in 1997. Public health inspectors from 86

its National Environment Agency (NEA) periodically assess the hygiene standards of each 87

licensed dining establishment. Establishment operators are usually not informed of their 88

visits. The outcome of one of these regulatory inspections is then displayed in the form of a 89

colour coded certificate containing the grade ‘A’, ‘B’, ‘C’ or ‘D’. The prominent display of the 90

assessed grade is legally mandated. An ‘A’ grading is the highest that can be awarded, while 91

‘D’ is the lowest and signifies the minimum standard of food hygiene for business continuity. 92

Since 2012, no establishments have been graded ‘D’. 93

94

Despite being one of the primary measures implemented to ensure food safety, the 95

consumer use of Singapore’s food hygiene grading system has not been examined. In this 96

analysis of survey data collected from 2012 and 2013, we examined consumers’ use of the 97

information disclosure system in Singapore and its influence on their choice of dining 98

establishments. We also assessed consumer attitudes, perceptions and practices of dining 99

out and food safety. 100

101

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

5 | P a g e

2. Materials and methods 102

103

2.1 Study population 104

105

Singapore is a city-state with an estimated multi-ethnic population of 5.6 million, of which 106

approximately 3.4 million are Singaporean citizens and 0.5 million are permanent residents 107

(Department of Statistics Singapore, 2017). There are more than 37,000 licenced food 108

establishments (Ministry of Environment and Water Resources Singapore, 2017), an 109

average of more than 48 per square kilometre. Licensed establishments include individual 110

food stalls in hawker centres, food courts and coffeeshops, restaurants, takeaway food 111

kiosks and caterers. 112

113

2.2 Study design 114

115

We obtained data from a household survey that the NEA conducted on the knowledge, 116

attitudes, beliefs and practices of food safety and hygiene of Singaporeans from 2012 to 117

2013. The aim of the survey was to collect household information to identify opportunities for 118

improvement in food hygiene and safety programmes. The NEA obtained a random sample 119

of 1,700 household addresses from the Government Department of Statistics Singapore and 120

sent a letter of notification to each of them in the four main languages used in Singapore 121

(English, Chinese, Malay and Tamil) to inform them of the purpose of the survey and to 122

provide assurance on data confidentiality. The NEA outsourced the data collection to a 123

market research company and both were responsible for training bilingual (English and one 124

other main language) interviewers to conduct the face-to-face surveys using structured 125

questionnaires that had been pretested in a pilot face-to-face survey on 50 individuals. 126

Survey respondents were randomly selected from each household by choosing the adult 127

with the nearest upcoming birthday. Respondents were given a S$5 voucher for their time. 128

The questionnaire comprised 14 close-ended questions on attitudes, perceptions and 129

practices of dining out and food safety, including factors related to dining establishment 130

choice and dining frequencies; and 10 socio-demographic questions: age, citizenship, 131

gender, ethnicity (according to the ethnic groups defined by the government), education, 132

marital status, number of children, employment status, income level and housing type 133

(categorised as public or private housing). Respondents were asked to rank 7 choice factors: 134

(i) cleanliness, (ii) taste, (iii) food hygiene practices of staff, (iv) price, (v) service quality, (vi) 135

recommendations and (vii) hygiene grading, in order of importance for each of the categories 136

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

6 | P a g e

of dining establishments. Categories of dining establishment included: (i) hawker centres, (ii) 137

coffeeshops and canteens, (iii) food courts, (iv) restaurants and cafes, (v) food kiosks, (vi) 138

bakeries and cake shops, and (vii) caterers. A rank score of ‘1’ signified the most important 139

factor while ‘7’ signified the least important factor, except for caterers in which the 140

cleanliness and food hygiene practices of staff could not be observed and only 5 factors 141

were ranked. 142

143

2.3 Study Measures 144

145

We analysed two main outcome measures: (1) referencing the food hygiene grade before 146

patronising a dining establishment; and (2) willingness to patronise a ‘C’ graded dining 147

establishment (‘C’ being the lowest grade currently given). For the first outcome measure, 148

respondents who either “agreed” or “strongly agreed” with the statement “I always look at the 149

grading of the food establishment before patronising it” were considered as those who 150

referenced the assessed grade. For the second outcome measure, it was assumed that all 151

respondents would dine at the highest ‘A’ graded establishments. For each of the 7 152

categories of food establishments, respondents were first asked if they would dine at a ‘B’ 153

graded establishment. Those who would were then asked if they would also dine at a ‘C’ 154

graded establishment. Binary variables were created to reflect the proportion of respondents 155

who were willing to dine at (i) any ‘B’ and (ii) any ‘B’ or ‘C’ and a test for difference in 156

proportions carried out if applicable. 157

158

2.4 Statistical Analysis 159

160

We present descriptive statistics for each independent demographic variable and the 161

responses to questions about food safety perceptions. We also present mean rank scores 162

for factors associated with establishment choice. We used housing type and employment 163

status as proxies for socioeconomic status as there was inadequate data on income (38% of 164

respondents did not provide this). Where appropriate, categorical variables were collapsed 165

for reporting and inclusion in the models (see Supplemental File). We used multivariable 166

logistic regression, which is appropriate for assessing the relationship between a dependent 167

categorical variable and multiple independent categorical or continuous variables (Wiest, 168

Lee, & Carlin, 2015). The outcome measures in our study were dichotomous dependent 169

variables while all the sociodemographic and establishment specific dining frequency 170

variables that were assessed were categorical. We coded the first dependent variable that 171

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

7 | P a g e

represented referencing the food hygiene grade before patronising an establishment as ‘1’ if 172

a respondent referenced the grade and ‘0’ otherwise. We coded the second dependent 173

variable that represented willingness to dine at a ‘C’ graded establishment as ‘1’ if a 174

respondent was willing to dine and ‘0’ otherwise. We used the Likelihood Ratio Test (LRT) to 175

determine the associations between each outcome measure (dependent variable) and the 176

independent variables representing the socio-demographic factors and establishment 177

specific dining frequencies. The measure of effect for each independent variable on the 178

dependent variable in the multivariable model was expressed as an adjusted odds ratio 179

(AOR). Statistical significance was evaluated at the 5% level. We retained independent 180

variables in the final model for each outcome measure only if they were significantly 181

associated. LRT p-values, stratum specific AORs and 95% CIs for the effects of independent 182

variables were presented. All analyses were performed using STATA 12.1 software 183

(StataCorp, USA). 184

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

8 | P a g e

3. Results 185

186

3.1 Participant characteristics 187

188

A total of 1,533 respondents participated in the study (90.2% response rate). The majority of 189

participants were Singaporean (89.9%), of Chinese ethnicity (77.4%), lived in public housing 190

(83.0%) and had secondary education or higher (92.6%). Slightly less than half of them were 191

male while about a third of them were not employed. The characteristics of the study 192

population are summarised in Table 1. 193

194

195

Table 1. Socio-Demographic Characteristics of Study Population in Singapore, 2012-2013 196

197

Hawker centres and coffeeshops/canteens were frequented most, with more than 85% of 198

respondents dining there at least 1-3 times a week (see Figure 1). Bakeries/cake shops were 199

the least frequented, with more than 99% of respondents reporting a dining frequency of 2-3 200

times a month or less. 201

202

203

Fig. 1. Establishment Dining Frequencies of Study Population in Singapore, 2012-2013 204

205

206

3.2 Food hygiene perceptions and establishment choice factors 207

208

Few respondents (0.9%; n=14) perceived the food hygiene standards in Singapore to be 209

“poor” or “very poor”. Approximately 42% (n=639) perceived the standards to be “good” or 210

“very good”, while the remaining 57.4% (n=880) perceived them to be “average”. The 211

majority of respondents (96.4%; n=1,479) either agreed or strongly agreed with the 212

statement “I am confident that eating out is safe as the Government licences all the food 213

establishments”, while the remaining 3.5% (n=54) disagreed or strongly disagreed with it. 214

When asked about reporting unhygienic practices to the NEA, 83.7% (n=1,283) disagreed or 215

strongly disagreed that they would do so, while the remaining 16.4% (n=250) agreed or 216

strongly agreed that they would do so. Approximately two-thirds (66.0%; n=1,012) reported 217

that they were willing to visit a food establishment even if they received negative comments 218

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

9 | P a g e

about the cleanliness and hygiene standards from people they knew. A summary table of the 219

results is available in the Supplemental File. 220

221

Dining establishment cleanliness (mean rank score: 2.31, SD: 0.69), food taste (mean rank 222

score: 2.72, SD: 1.04) and food hygiene practices of staff (mean rank score: 2.83, SD: 0.86) 223

were the top three establishment choice factors for respondents across all categories of 224

establishments. Comparatively, food price (mean rank score: 3.78, SD: 1.27), establishment 225

service quality (mean rank score: 4.82, SD: 0.89) and recommendations (mean rank score: 226

5.62, SD 0.79) were ranked lower. The assessed hygiene grading of a dining establishment 227

was the lowest ranked factor (mean rank score: 5.90, SD: 1.00). 228

229

3.3 Referencing food hygiene grades 230

231

Sixty-five percent [64.5% (95% CI = 62.1%, 66.9%); n=989] of respondents indicated that 232

they would always refer to the assessed hygiene grade before patronising an eating 233

establishment. After multivariable adjustment, the strongest associations that remained were 234

related to ethnicity and dining frequency (see Figure 2). Malay ethnicity was strongly 235

associated with higher odds of referencing grades compared to Chinese ethnicity [AOR: 236

2.59, (95% CI = 1.65, 4.06)]. We also found that respondents had lower odds of referencing 237

grades if they reported dining daily at hawker centres [AOR: 0.50, (95% CI 0.35, 0.72)] and 238

food courts [AOR: 0.42, (95% CI = 0.20, 0.89)] compared to those that dined there 1-3 times 239

a week (summary of results in Supplemental File). 240

241

242 Fig. 2. Adjusted ORs for factors associated with referencing assessed hygiene grades before patronizing a food 243 establishment in Singapore, 2012 to 2013. The diamonds indicate the point estimates and the horizontal navy blue lines 244 indicate the 95% confidence intervals for those estimates. The vertical grey line indicates the null value of 1.00. 245 Reference categories are indicated with a value of 1.00. 246

247

3.4 Willingness to dine at establishments of different assessed grades 248

249

Approximately 85% (95% CI 83.5%, 87.0%; n=1,303) of respondents reported willingness to 250

dine at any ‘B’ graded establishment. When ‘C’ graded establishments were considered, only 251

10.3% (95% CI = 8.8%, 11.8%, n=158) reported willingness to do so. The difference in 252

proportions was statistically significant (p<0.01). After multivariable analysis, the strongest 253

associations that remained were related to dining frequency (see Figure 3). Compared to 254

those who reported dining 1-3 times a week, the odds of reporting willingness to dine at any 255

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

10 | P a g e

‘C’ graded establishment were higher in those who reported dining daily at hawker centres 256

[aOR: 2.75 (95% CI = 1.68, 4.48)] and coffeeshops/canteens [AOR: 2.44, (95% CI = 1.44, 257

17.20)]. Those who reported dining 2-3 times a month or less frequently at food kiosks were 258

also more likely to report willingness to dine at a ‘C’ graded establishment [AOR: 5.04 (95% 259

CI = 3.21, 7.90)]. Besides establishment dining frequencies, we found that the odds of 260

reporting willingness to dine at any ‘C’ graded establishment were higher in those that 261

reported living in Public (1 – 2, 3, 4 or 5 Room Flat) housing [AOR: 2.33 (95% CI = 1.27, 262

4.27)] compared to those that did not, and also higher in those that had 3 or more children 263

[AOR: 1.66 (95% CI = 1.01, 2.72)] compared to those who had none. Respondents who had 264

lower odds of reporting the willingness to dine at a ‘C’ graded establishment reported 265

secondary education as the highest qualification [AOR: 0.43 (95% CI = 0.24, 0.74)], being a 266

permanent resident in Singapore [AOR: 0.42 (95% CI = 0.21, 0.87)] and dining 2-3 times a 267

month or less frequently at bakeries/cake shops [AOR: 0.04 (95% CI = 0.01, 0.27)] 268

(summary of results in Supplemental File). 269

270

Fig. 3. Adjusted ORs for factors associated with willingness to dine at a 'C' graded food establishment in Singapore, 2012 271 to 2013. The diamonds indicate the point estimates and the horizontal navy blue lines indicate the 95% confidence 272 intervals for those estimates. The vertical grey line indicates the null value of 1.00. Reference categories are indicated 273 with a value of 1.00. 274

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

11 | P a g e

4. Discussion 275

276

In this cross-sectional study, we examined the consumer use of the letter based grading 277

information disclosure system and its influence on dining establishment choice in Singapore. 278

Our study suggests that the majority of Singaporeans use the letter based grading 279

information disclosure system in Singapore and discriminate their patronage of dining 280

establishments according to the assessed letter grading. We also found that population 281

socio-demographic factors accounted for differences in its use and influence. Public 282

information disclosure systems have been used by health authorities with the aim of 283

influencing consumer dining establishment choice and consequently induce consumer 284

demand for better food hygiene standards. Our study findings add to the growing body of 285

evidence which supports the complementary use of non-regulatory measures in reassuring 286

diners of food safety in dining establishments (Uggioni & Salay, 2014; Wong et al., 2015) 287

and improving food hygiene standards (Jin & Leslie, 2009; Vergeris, 2015; Waters et al., 288

2013; Wong et al., 2015). 289

290

To date, few studies have assessed the consumer use of information disclosure systems 291

aimed at influencing establishment choice. Evidence from such studies can inform the 292

development and evaluation of interventions designed to improve food hygiene standards in 293

dining establishments. In a cross-sectional survey of adults in the UK in 2012, 10% of 294

respondents used an information disclosure scheme in deciding whether to patronise an 295

establishment (Taylor Nelson Sofres-British Market Research Bureau Limited, Policy 296

Institute Studies (UK), University of Westminster, & Food Standards Agency (UK), 2013). 297

Another study in the US found that only 6.6% of respondents reported using a restaurant 298

inspection report every time or almost every time in establishment choice (Jooho, Jing, & 299

Almanza, 2017). In our study, we estimated that the proportion of respondents who would 300

always check the letter grade before patronising a dining establishment was 64.5%, 301

consistent with a previous Singaporean study (60.7%) (Choi, MacLaurin, Cho, & Hahm, 302

2010). The larger estimate in referrals in our study compared to those in other countries may 303

be attributed to the level of awareness of such information schemes, or cultural differences in 304

establishment choice considerations, or differences in the level of any bias in self-reporting 305

between studies. Compared to respondents of Chinese ethnicity, those of Malay ethnicity 306

were more likely to reference the assessed grade before dining at an eating establishment. 307

Greater focus on messaging in both English and Chinese languages could be considered for 308

the non-Malay ethnic population of consumers. Respondents who were not employed were 309

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

12 | P a g e

less likely to report that they would reference the assessed grade. This may be because the 310

importance of dining at a more affordable establishment exceeds that of its assessed 311

hygiene standards. 312

313

Similar to other studies in the UK, the US and Turkey (Aksoydan, 2007; Harrington, 314

Ottenbacher, & Way, 2013; Lee, Niode, Simonne, & Bruhn, 2012; Taylor Nelson Sofres-315

British Market Research Bureau Limited et al., 2013), we found that cleanliness was one of 316

the most important factors in consumers’ choice of the majority of establishments. The 317

importance of cleanliness to participants suggests that their perceived risk of food poisoning 318

acquisition from unsanitary premises was an important factor affecting their choice. 319

Surprisingly, the assessed hygiene grading was ranked as the least important factor in the 320

present study but was consistent with the finding from another study where the assessed 321

hygiene grade/score is among the lower ranked factors (Taylor Nelson Sofres-British Market 322

Research Bureau Limited et al., 2013). A study found that consumers in Canada use a range 323

of observable indicators as proxies for the experience and/or credence characteristics 324

associated with food safety hazards (Henson et al., 2006). These indicators include the 325

consumer's own observation on standards of hygiene, broader concepts of restaurant 326

quality, levels of patronage and external sources of information, of which official inspection 327

certificates form one part. It is plausible that Singaporean consumers use visual cues to 328

assess the standard of hygiene in an establishment as opposed to using the assessed grade 329

which may not be entirely reflective of the standard at the time of patronage. 330

331

A public information disclosure system can be considered an effective one if it discriminates 332

consumer patronage at establishments with relatively lower food hygiene standards. Our 333

study results are reassuring and reflect the ability of the letter based grading system in 334

Singapore to influence dining establishment choice - we found that most respondents 335

favoured well graded establishments over the more poorly graded ones. Only a small 336

proportion of respondents were willing to dine at poorly graded establishments. Our results 337

are consistent with an earlier study carried out in California which showed that respondents 338

were least likely to dine at a ‘C’ graded restaurant compared to better graded ones (Seiver & 339

Hatfield, 2002). These findings were also in agreement with those from a population-based 340

cross-sectional survey on the use of the food hygiene rating scheme carried out in the United 341

Kingdom in 2015 (Food Standards Agency (UK), 2015a). 342

343

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

13 | P a g e

Respondents who lived in public housing were more likely to report that they were willing to 344

dine at a ‘C’ graded establishment. This may suggest that respondents with relatively lower 345

disposable income levels value other factors such as affordability more highly. Respondents 346

who reported having three or more children were also more likely to report that they were 347

willing to dine at a ‘C’ graded establishment. It is possible that larger families may accord 348

convenience as a more important factor in establishment choice. Educational initiatives 349

aimed at reinforcing the use of the system could be prioritized in housing estates associated 350

with relatively lower income levels and at larger families. Respondents with secondary 351

education were less likely to dine at a ‘C’ graded establishment compared to those with 352

primary education. Individuals with higher levels of education have been associated with 353

better food hygiene knowledge (Mullan, Wong, Todd, Davis, & Kothe, 2015). However, this 354

association was not straightforward as it was not found for higher levels of education in the 355

present study. Indeed, people with higher education have also sometimes reported poorer 356

food safety practices than those with lower or no education (RØSsvoll et al., 2013). 357

358

In several categories of food establishments, daily consumer dining was associated with a 359

lower propensity to reference the assessed grade and a greater willingness to patronise any 360

poorly graded establishment. These findings were not surprising, given that frequent diners 361

of restaurants have been found to be less concerned about food safety compared to those 362

who dined infrequently (Knight, Worosz, & Todd, 2009). However, the present study also 363

found that infrequent diners (2-3 times a month or less frequent) of food courts and catering 364

establishments were less likely to refer to the assessed hygiene grade, and infrequent diners 365

of food kiosks were also more likely to dine at ‘C’ graded ones, indicating that some 366

infrequent diners may also either be less concerned or knowledgeable about food safety. 367

Initiatives to reinforce the awareness and use of the information system could be prioritised 368

at hawker centres and coffeeshops/canteens which are more frequent choice venues for 369

daily diners. 370

371

Despite the demonstrated utility of such information disclosure systems, they may be limited 372

in their influence over the hygiene standards of food establishments that have strong 373

patronage linked to other important consumer choice factors such as price and taste. The 374

internet has also been used effectively as a tool to spread consumer dissatisfaction (Aguiar 375

et al., 2018), demonstrating its potential to improve food safety and quality. Where 376

appropriate, other tools such as social media can play an important role in communicating 377

and influencing food safety (Peng, Li, Xia, Qi, & Li, 2015). Negative social media comments 378

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

14 | P a g e

have been found to significantly increase food safety risk perceptions and reduce behavioral 379

intentions of consumers compared with mixed or positive comments (Frewer et al., 2015). 380

Given that almost 80% of the Singaporean population use smart devices to access the 381

internet (Infocomm Development Authority of Singapore, 2017), we suggest developing an 382

app which features information on recent critical hygiene violations committed by food 383

establishments and allows consumers to enter their establishment specific views on the 384

observed cleanliness and hygiene for other app users to view. 385

386

Our study has several strengths. The sample was large and nationally-representative 387

(n=1,700). The response rate was in excess of 90%, indicating very low potential for 388

selection bias. The use of standardized questionnaires and training of all interviewers would 389

have minimised observer bias. However, common to all cross-sectional studies, participants 390

may have provided socially desirable answers when answering questions (i.e. they may be 391

more willing to dine at a relatively poorer graded restaurant and may look less often at the 392

food hygiene grading that indicated). For future studies, one way to attempt to address this 393

could be through the use of a smartphone based food diary app that facilitates the capture of 394

the assessed grades of eating establishments and helps keep records of establishments at 395

which respondents dined. 396

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

15 | P a g e

5. Conclusions 397

398

Public information disclosure systems can play an important role in ensuring food safety in 399

dining establishments. While it could be expected that consumers prefer to eat in 400

establishments with better hygiene all else being equal, the use and role of the grading 401

system as a way of consumers making this decision had not previously been assessed in 402

Singapore. Ours is the first nationally representative study to provide evidence that 403

consumers refer to the information system in Singapore and that the letter-grading 404

information disclosure system positively influences their self-reported choice in eating 405

establishments. Our findings suggest that such national public information disclosure 406

systems can be an effective means of improving food safety in dining establishments and 407

complement the regulatory measures of public health authorities. Several socio-demographic 408

factors were important influencers in the use of the information disclosure system and should 409

be taken into account in the design and implementation of such systems. The 410

implementation of such information disclosure systems should be considered in other 411

countries where it has yet to be introduced. However, such systems should also be 412

periodically assessed for its effectiveness and to identify areas requiring improvements. 413

Incorporating recent information on establishment cleanliness and its recent critical hygiene 414

violations may increase its relevance and thus improve its use. Further research is needed to 415

assess which strategies are most likely to improve the use of such information disclosure 416

systems in consumer dining choices. 417

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

16 | P a g e

Contributors 418

JA conceptualized the study, led the writing of the article, developed the analytic plan, 419

analysed the data and interpreted it. 420

AEH, ATN & MDK provided feedback on data analysis, contributed significantly to the 421

interpretation of results and revising drafts of the article. 422

LCN contributed significantly to the analysis, interpretation of results and to revising drafts of 423

the article. 424

All authors critically revised the article for important intellectual content and approved all 425

components of the final draft. 426

427

Acknowledgements 428

Note. The data utilized in this project were made available by the National Environment 429

Agency, Singapore. The information and opinions expressed reflect solely those of the 430

authors. 431

432

Ethics Approval 433

This secondary analysis was granted ethical approval by the Human Research Ethics 434

Advisory Panel of the University of New South Wales, Australia (HC15800). 435

436

Conflicts of Interest 437

None. 438

439

Funding Source 440

This research did not receive any specific grant from funding agencies in the public, 441

commercial, or not-for-profit sectors. 442

443

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

17 | P a g e

References 444

Aguiar, R. S., Esmerino, E. A., Rocha, R. S., Pimentel, T. C., Alvarenga, V. O., Freitas, M. Q., . . . Cruz, A. G. 445

(2018). Physical hazards in dairy products: Incidence in a consumer complaint website in Brazil. Food 446

Control, 86, 66-70. doi:https://doi.org/10.1016/j.foodcont.2017.11.020 447

Aksoydan, E. (2007). Hygiene factors influencing customers' choice of dining-out units: findings from a study 448

of university academic staff. Journal of Food Safety, 27(3), 300-316. doi:10.1111/j.1745-449

4565.2007.00081.x 450

Choi, H. C., MacLaurin, T., Cho, J.-E., & Hahm, S.-P. (2010). Food Hygiene Standard Satisfaction of Singaporean 451

Diners. Journal of Foodservice Business Research, 13(3), 156-177. doi:10.1080/15378020.2010.500225 452

Cusato, S., Gameiro, A. H., Corassin, C. H., Sant'Ana, A. S., Cruz, A. G., Faria, J. d. A. F., & de Oliveira, C. A. F. 453

(2012). Food Safety Systems in a Small Dairy Factory: Implementation, Major Challenges, and 454

Assessment of Systems' Performances. Foodborne Pathogens and Disease, 10(1), 6-12. 455

doi:10.1089/fpd.2012.1286 456

Department of Statistics Singapore. (2017). Statistics - Latest Data. Retrieved from 457

http://www.singstat.gov.sg/statistics/latest-data#16 Accessed 30 October 2017 458

Djekic, I., Kuzmanović, J., Anđelković, A., Saračević, M., Stojanović, M. M., & Tomašević, I. (2016). Effects of 459

HACCP on process hygiene in different types of Serbian food establishments. Food Control, 60, 131-460

137. doi:http://dx.doi.org/10.1016/j.foodcont.2015.07.028 461

Economic Research Service (ERS), U. S. D. o. A. U. (2015). Cost Estimates of Foodborne Illnesses. Retrieved 462

from http://ers.usda.gov/data-products/cost-estimates-of-foodborne-illnesses.aspx Accessed 16 463

January 2018 464

European Food Safety Authority. (2014). EFSA explains zoonotic diseases: Salmonella. Retrieved from 465

https://www.efsa.europa.eu/en/topics/topic/salmonella Accessed 16 January 2018 466

Food Standards Agency (UK). (2015a). Food hygiene rating scheme - biannual public attitudes tracker (Wave 2 467

- 2015). Retrieved from https://www.food.gov.uk/sites/default/files/fs244011wave2.pdf Accessed 30 468

October 2017 469

Food Standards Agency (UK). (2015b). Food Hygiene Ratings. Retrieved from http://ratings.food.gov.uk/ 470

Accessed 30 October 2017 471

Frewer, L. J., Fischer, A. R. H., Brennan, M., Bánáti, D., Lion, R., Meertens, R. M., . . . Vereijken, C. M. J. L. 472

(2015). Risk/Benefit Communication about Food – A Systematic Review of the Literature. Critical 473

Reviews in Food Science and Nutrition, 1728-1745. doi:10.1080/10408398.2013.801337 474

Gomes, C. C. B., Lemos, G. F. C., Silva, M. C., Hora, I. M. C., & Cruz, A. G. (2014). Training of Food Handlers in a 475

Hotel: Tool for Promotion of the Food Safety. Journal of Food Safety, 34(3), 218-223. 476

doi:10.1111/jfs.12116 477

Gormley, F. J., Little, C. L., Rawal, N., Gillespie, I. A., Lebaigue, S., & Adak, G. K. (2011). A 17-year review of 478

foodborne outbreaks: describing the continuing decline in England and Wales (1992–2008). 479

Epidemiology & Infection, 139(05), 688-699. doi:doi:10.1017/S0950268810001858 480

Gould, L. H., Walsh, K. A., Vieira, A. R., Herman, K., Williams, I. T., Hall, A. J., & Cole, D. (2013). Surveillance for 481

Foodborne Disease Outbreaks — United States, 1998–2008. Retrieved from United States: 482

http://www.cdc.gov/mmwr/preview/mmwrhtml/ss6202a1.htm Accessed 30 October 2017 483

Harrington, R. J., Ottenbacher, M. C., & Way, K. A. (2013). QSR Choice: Key Restaurant Attributes and the 484

Roles of Gender, Age and Dining Frequency. Journal of Quality Assurance in Hospitality & Tourism, 485

14(1), 81-100. doi:10.1080/1528008X.2013.749380 486

Health Promotion Board Singapore. (2010). National Nutrition Survey 2010. Retrieved from 487

https://www.hpb.gov.sg/dpcs/default-source/pdf/nns-2010-report.pdf?sfvrsn=18e3f172_2 Accessed 488

30 October 2017 489

Henson, S., Majowicz, S., Masakure, O., Sockett, P., Jones, A., Hart, R., . . . Knowles, L. (2006). Consumer 490

assessment of the safety of restaurants: the role of inspection notices and other information cues. 491

Journal of Food Safety, 26(4), 275-301. doi:10.1111/j.1745-4565.2006.00049.x 492

Infocomm Development Authority of Singapore. (2017). Annual Survey on Infocomm Usage in Households and 493

by Individuals for 2015. Retrieved from https://www.imda.gov.sg/industry-development/facts-and-494

figures/survey-reports Accessed 30 October 2017 495

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

18 | P a g e

ISO. (2005). ISO 22000 Food safety management systems - requirements for any organization in the food 496

chain. 497

Jabs, J., & Devine, C. M. (2006). Time scarcity and food choices: An overview. Appetite, 47(2), 196-204. 498

doi:http://dx.doi.org/10.1016/j.appet.2006.02.014 499

Jin, G. Z., & Leslie, P. (2009). Reputational Incentives for Restaurant Hygiene. American Economic Journal: 500

Microeconomics, 1(1), 237-267. doi:doi: 10.1257/mic.1.1.237 501

Jones, T. F., & Angulo, F. J. (2006). Eating in Restaurants: A Risk Factor for Foodbome Disease? Clinical 502

Infectious Diseases, 43(10), 1324-1328. 503

Jooho, K., Jing, M., & Almanza, B. (2017). Consumer Perception of the Food and Drug Administration's Newest 504

Recommended Food Facility Inspection Format: Words Matter. Journal of Environmental Health, 505

79(10), 20-25. 506

Knight, A. J., Worosz, M. R., & Todd, E. C. D. (2009). Dining for Safety: Consumer Perceptions of Food Safety 507

and Eating Out. Journal of Hospitality & Tourism Research, 33(4), 471-486. 508

doi:10.1177/1096348009344211 509

Läikkö-Roto, T., & Nevas, M. (2014). Restaurant business operators' knowledge of food hygiene and their 510

attitudes toward official food control affect the hygiene in their restaurants. Food Control, 43, 65-73. 511

doi:http://dx.doi.org/10.1016/j.foodcont.2014.02.043 512

Lee, L. E., Niode, O., Simonne, A. H., & Bruhn, C. M. (2012). Consumer perceptions on food safety in Asian and 513

Mexican restaurants. Food Control, 26(2), 531-538. 514

doi:https://doi.org/10.1016/j.foodcont.2012.02.010 515

Lockis, V. R., Cruz, A. G., Walter, E. H. M., Faria, J. A. F., Granato, D., & Sant'Ana, A. S. (2010). Prerequisite 516

Programs at Schools: Diagnosis and Economic Evaluation. Foodborne Pathogens and Disease, 8(2), 517

213-220. doi:10.1089/fpd.2010.0645 518

Lund, B. (2015). Microbiological Food Safety for Vulnerable People. International Journal of Environmental 519

Research and Public Health, 12(8), 10117. Retrieved from http://www.mdpi.com/1660-520

4601/12/8/10117 521

McKelvey, W., Wong, M. R., & Matis, B. (2015). Letter Grading and Transparency Promote Restaurant Food 522

Safety in New York City. Journal of Environmental Health, 78(2), 46-48. 523

Ministry of Environment and Water Resources Singapore. (2017). Key Environmental Statistics 2017. Retrieved 524

from http://www.mewr.gov.sg/docs/default-source/default-document-library/grab-our-research/kes-525

2017.pdf Accessed 30 October 2017 526

Ministry of Health (Singapore). (2015). Communicable Diseases Surveillance in Singapore 2015. Retrieved from 527

https://www.moh.gov.sg/content/moh_web/home/Publications/Reports/2016/communicable-528

diseases-surveillance-in-singapore-2015.html Accessed 30 October 2017 529

Mullan, B., Wong, C., Todd, J., Davis, E., & Kothe, E. J. (2015). Food hygiene knowledge in adolescents and 530

young adults. British Food Journal, 117(1), 50-61. doi:doi:10.1108/BFJ-03-2013-0060 531

New South Wales Food Authority. (2017). Scores on Doors. Retrieved from 532

http://www.foodauthority.nsw.gov.au/rp/scoresondoors Accessed 30 October 2017 533

Park, S.-H., Kwak, T.-K., & Chang, H.-J. (2010). Evaluation of the food safety training for food handlers in 534

restaurant operations. Nutrition Research and Practice, 4(1), 58-68. doi:10.4162/nrp.2010.4.1.58 535

Peng, Y., Li, J., Xia, H., Qi, S., & Li, J. (2015). The effects of food safety issues released by we media on 536

consumers’ awareness and purchasing behavior: A case study in China. Food Policy, 51, 44-52. 537

RØSsvoll, E. H., Lavik, R., Ueland, Ø., Jacobsen, E., Hagtvedt, T., & Langsrud, S. (2013). Food Safety Practices 538

among Norwegian Consumers. Journal of Food Protection, 76(11), 1939-1947. doi:10.4315/0362-539

028X.JFP-12-269 540

Seiver, O. H., & Hatfield, T. H. (2002). Grading Systems for Retail Food Facilities: Preference Reversals of 541

Environmental Health Professionals. Journal of Environmental Health, 64(10), 8. 542

Taylor Nelson Sofres-British Market Research Bureau Limited, Policy Institute Studies (UK), University of 543

Westminster, & Food Standards Agency (UK). (2013). Exploring food attitudes and behaviours in the 544

UK: Findings from the Food and You Survey 2012. Retrieved from United Kingdom: 545

http://www.food.gov.uk/sites/default/files/food-and-you-2012-main-report.pdf Accessed 30 October 546

2017 547

The Danish Veterinary and Food Administration. (2017). About the Danish Smiley Scheme. Retrieved from 548

http://www.findsmiley.dk/en-US/Forside.htm Accessed 30 October 2017 549

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

19 | P a g e

Uggioni, P. L., & Salay, E. (2014). Sociodemographic and knowledge influence on attitudes towards food safety 550

certification in restaurants. International Journal of Consumer Studies, 38(4), 318-325. 551

doi:10.1111/ijcs.12101 552

Vergeris, S. (2015). The Food Hygiene Rating Scheme and the Food Hygiene Information Scheme: Evaluation 553

findings 2011-2014. Retrieved from United Kingdom: http://www.psi.org.uk/pdf/2015/fhrs-fhis-554

eval2011-14.pdf Accessed 30 October 2017 555

Waters, A. B., VanDerslice, J., Porucznik, C., Kim, J., DeLegge, R., & Durrant, L. (2013). Impact of Internet 556

Posting of Restaurant Inspection Scores on Critical Violations. Journal of Environmental Health, 75(10), 557

8-12. 558

Wiest, M. M., Lee, K. J., & Carlin, J. B. (2015). Statistics for clinicians: An introduction to logistic regression. 559

Journal of Paediatrics and Child Health, 51(7), 670-673. doi:10.1111/jpc.12895 560

Wong, M. R., McKelvey, W., Ito, K., Schiff, C., Jacobson, J. B., & Kass, D. (2015). Impact of a Letter-Grade 561

Program on Restaurant Sanitary Conditions and Diner Behavior in New York City. American Journal of 562

Public Health, 105(3), e81-e87. doi:10.2105/AJPH.2014.302404 563

World Health Organization. (2015). Estimates of the global burden of foodborne diseases. Retrieved from 564

565

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPTTable 1

Socio-Demographic Characteristics of Study Population in Singapore, 2012-2013

Description of Variables and Categories Number (%)

Socio-demographic characteristics

Citizenship

Singaporean 1,378 (89.9%)

Permanent Resident 155 (10.2%)

Age in Years

18-29 271 (17.7%)

30-39 307 (20.0%)

40-49 358 (23.5%)

50-59 328 (21.4%)

60-65 269 (17.6%)

Gender

Female 823 (53.7%)

Male 710 (46.3%)

Ethnicity

Chinese 1,187 (77.4%)

Malay 133 (8.7%)

Indian 155 (10.1%)

Others 58 (3.8%)

Housing Type

Public Executive/Non-Public 260 (17.0%)

Public (1 – 2, 3, 4 or 5 Room Flat) 1,273 (83.0%)

Marital Status

Single 428 (27.9%)

Married 1,056 (68.9%)

Separated, divorced or widowed 49 (3.2%)

Number of Children

None 557 (36.3%)

1 221 (14.4%)

2 529 (34.5%)

≥3 226 (14.7%)

Highest Educational Qualification

Primary or Lower 114 (7.4%)

Secondary 624 (41.4)

Post-Secondary, Diploma or Professional Qualification 494 (32.2%)

Degree or Higher 291 (19.0%)

Employment Status

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPTTable 1

Socio-Demographic Characteristics of Study Population in Singapore, 2012-2013

Description of Variables and Categories Number (%)

Not Employed 534 (34.8%)

Employed 999 (65.2%)

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

34.6%

99.6%

50.4%

49.1%

25.2%

11.5%

11.5%

63.1%

0.4%

49.1%

50.9%

72.5%

78.9%

77.7%

2.3%

0.0%

0.5%

0.0%

2.2%

9.6%

10.8%

0.0% 20.0% 40.0% 60.0% 80.0% 100.0%

Caterers

Bakeries/Cake shops

Food Kiosks

Restaurants/Cafes

Food Courts

Coffeeshops/Canteens

Hawker Centres

Proportion of respondents

Daily 1-3 times a week 2-3 times a month or less

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

1.000.70

1.002.59

1.271.19

0.501.00

0.74

0.421.00

0.75

4.611.00

0.80

0.821.00

0.73

Socio

-Demo

graphi

cFa

ctors

Establ

ishme

ntDin

ing Fr

equenc

iesEmployed

Not Employed

Chinese EthnicityMalay EthnicityIndian EthnicityOther Ethnicity

Daily1-3 times a week or less

2-3 times per month or less

Daily1-3 times a week or less

2-3 times per month or less

Daily1-3 times a week or less

2-3 times per month or less

Daily1-3 times a week or less

2-3 times per month or less

Employment Status

Ethnicity

Hawker Centres

Food Courts

Food Kiosks

Caterers

0.20 1.00 2.50 10.00 40.00AORs

More likely to refer Less likely to refer

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

1.000.42

1.000.41

1.001.29

0.811.66

1.000.43

0.801.03

2.751.00

0.66

2.441.00

1.45

1.821.00

5.04

1.000.04

Socio

-Demo

graphi

c Fact

orsEst

ablish

ment D

ining F

requen

cies

SingaporeanPermanent Resident

Public 1 to 5 RoomPublic Executive or Private

012

>2

Primary or LowerSecondary

Post Secondary or Professional QualificationDegree or Higher

Daily1-3 times a week or less

2-3 times per month or less

Daily1-3 times a week or less

2-3 times per month or less

Daily1-3 times a week or less

2-3 times per month or less

1-3 times a week or less2-3 times per month or less

Nationality

Housing Type

No. of Children

Highest Educational Qualification

Hawker Centres

Coffee shops/Canteens

Food Kiosks

Bakeries/Cake shops

0.01 1.00 20.00AORs

More likely to dine Less likely to dine

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Highlights:

• 64.5% of respondents referred to the letter grade before dining at an

establishment

• Only 10.3% of respondents would dine at a ‘C’ (poorly) graded establishment

• Letter-based grading information disclosure system influences establishment

choice.

• Sociodemographic factors were associated with information disclosure system

use.