ABSTRACT

In recent years, numerous state and local health departments have developed systems to disclose restaurant inspection results to consumers. Public disclosure of restaurant inspection results can reduce transmission of foodborne illness by driving improvements in sanitary conditions. In Minnesota, restaurant inspection results are not readily accessible for consumers to use to make decisions about where to eat. The objective of this study was to assess the consumer interest among Minnesota adults in having better access to restaurant inspection results and to identify preferred formats for disseminating this information. We conducted a survey among 1,188 Minnesota residents aged 18 years or older at the 2019 Minnesota State Fair. Overall, 94.4% of respondents wanted better access to restaurants' inspection information. More than three-quarters (77.1%) of respondents stated that they would use this information to decide where to eat. Respondents wanted to see inspection results online (71.6%) and at restaurants (62.1%). Increasing public access to inspection results could reinforce efforts by public health agencies and food service operators to improve the safety of foods prepared away from home.

HIGHLIGHTS
  • In Minnesota, restaurant inspection results are not readily accessible to consumers.

  • Public disclosure can improve sanitary conditions, reducing risk of foodborne illness.

  • There is interest in having better access to inspection results in Minnesota.

  • Minnesota adults want to see inspection reports online and in restaurants.

Over the past several decades, food prepared away from home has become an integral part of the American diet (18). There is evidence that restaurants play an important role in the transmission of foodborne illness (14, 16). Restaurants are frequent settings for sporadic (nonoutbreak) cases and outbreaks of foodborne illness (14, 16). In 2017, there were 841 foodborne disease outbreaks that resulted in 14,481 illnesses, 827 hospitalizations, and 20 deaths in the United States (3). Of these outbreaks, 64% for which a single location of preparation was reported occurred in restaurant settings (3). Our dependence on food prepared away from home emphasizes the critical importance of reducing the risk of foodborne illness transmission in restaurants.

Inspection of food service establishments to protect food safety is a core function of state and local regulatory agencies. Risk-based inspections for food service establishments seek to identify risk factors for foodborne illness and correct them to protect public health. Systems to communicate restaurant inspection results with the public can vary widely. In some locations, the information must be requested from regulatory agencies, which can take considerable time. In others, inspection results may be available online, upon request at restaurants, or visibly posted at restaurants. In recent years, an increasing number of jurisdictions have adopted or considered adopting letter grade placards or other disclosure programs to show restaurant inspection results; yet, the value of these postings has been the subject of considerable debate. Public disclosure systems can trigger a feedback loop that leads to improved sanitary conditions. By providing restaurant inspection results at the point of service, consumers can make informed decisions about potential risks and choose whether or not to eat at a restaurant. When restaurant operators identify and perceive these consumer decisions to affect their business, restaurant operators are incentivized to improve or maintain higher sanitary conditions (10). By improving sanitary conditions in restaurants, an expected outcome is a reduction in the risk of foodborne illness. Although there are multiple factors that affect dining decisions, public disclosure of restaurant inspections has been shown to have an impact on restaurant choice (1, 4, 12).

Studies to evaluate public disclosure programs have primarily assessed the impact on inspection results rather than a direct public health impact. A limited number of studies have attempted to quantify the impact of a public disclosure letter grade program on risk of foodborne illness. Studies conducted in Los Angeles County in California showed a decline in foodborne illness hospitalizations after the implementation of a letter grade program (13, 19). After New York City began posting letter grades at the point of service, the rate of Salmonella infections decreased 5.3% per year versus the rest of the state compared with the period before implementation (9). The New York City Department of Health & Mental Hygiene had previously shown that the letter grade program was associated with sustained improvements in sanitary conditions and that 88% of New Yorkers who had seen the letter grades used them in their dining decisions (22). These findings suggest that consumer behaviors helped drive the program goal of improving sanitary conditions in restaurants and reducing the risk of foodborne illness transmission in restaurants.

Minnesota is one of the least transparent states in the nation with respect to restaurant inspections. Inspection results are considered public data, but there is no routine release of inspection results and the information is not readily available at restaurants or online. Inspection results must be specifically requested, which in essence precludes their use by Minnesota consumers in the decision-making process. Given that Minnesota restaurant inspection reports are not readily accessible across the state, the objectives of this study were (i) to assess the consumer interest and support for public disclosure of inspection results among Minnesota adults and (ii) to identify the most useful formats and locations for disclosing these results. We hypothesized that there would be interest in public disclosure of restaurant inspection results in Minnesota and that interest would vary by average frequency of eating at a restaurant or other food service establishment.

MATERIALS AND METHODS

Target population and sampling strategy

We conducted a survey at the 2019 Minnesota State Fair in the University of Minnesota Driven to Discover (D2D) Research Facility during 22 to 27 August 2019. The D2D Research Facility has research bays where up to 10 studies can be conducted simultaneously in morning or afternoon shifts (7). The 2019 Minnesota State Fair had more than 2.1 million attendees across the 12 days (17). More than 60,000 attendees visited the D2D Research Facility, of which more than 27,700 participated in the 55 studies (7). Our study was conducted during four 5.5- to 6.5-h shifts. A sign was posted in the research bay that included the study name (“Dine or Dash”) and the eligibility criteria (Minnesota resident, aged 18+ years). Volunteer staff recruited fairgoers who chose to enter the D2D Research Facility and expressed interest in participating by providing a brief description of the survey, checking eligibility criteria, conducting informed consent, and providing iPads to eligible participants to complete the survey. The survey was conducted in English on iPads and completed individually by respondents.

(i) Informed consent

Eligible participants were given a brief overview of the informed consent verbally and provided with written informed consent on the first page of the survey. Interested participants selected on the survey that they agreed to participate before answering survey questions. Participants received a University of Minnesota drawstring backpack for participating. The study was reviewed and approved by the University of Minnesota Institutional Review Board.

(ii) Volunteer staff and training

There were 10 volunteer staff total and 4 per shift. Volunteer staff consisted of public health students and public health professionals. Volunteers were given a 1-h training by the student principal investigator that provided a brief overview of public disclosure systems, information about the D2D Facility, and an overview of recruitment and the informed consent process. All volunteers completed Collaborative Institutional Training Initiative Program Training for Human Subjects Research.

(iii) Data collection instrument

The 5- to 10-min survey was developed in REDCap software to assess consumer interest and support for public disclosure of restaurant inspection results in Minnesota and to identify the most useful formats and locations for public disclosure of restaurant inspection results. The survey was reviewed by public health professionals in academia and government partners and pilot tested by volunteer staff. The survey included questions about participant demographics; factors that are important to the respondent when choosing a place to eat; experience of foodborne illness (yes or no); confidence that Minnesota maintains its food safety standards (5-point Likert scale from “not at all confident” to “very confident”); perception of whether or not it is more common, less common, or equally common to get sick from food prepared in a restaurant compared with at home; interest in having better access to restaurant inspection information (yes or no); preferred locations (online webpage, mobile application, posted in a restaurant visible before entering, posted in a restaurant, visible upon entry); and whether or not the respondent would use restaurant inspection information to decide where to eat if it were posted at restaurants (yes, no, or not sure). The survey also included four pictures as examples of restaurant inspection placards and asked which placard they would most want to see if inspection results were posted at restaurants. The examples included a letter grade (blue, A; green, B; orange-yellow, C) based on New York City's system, a numerical score and letter grade (A/90 to 100, B/80 to 89, C/70 to 79) in black and white based on North Carolina's system, emojis and categories based on Washington's Seattle–King County system (green wide-mouth smile, excellent; chartreuse smile, good; lime small smile, okay; gray neutral face, needs improvement), and a stoplight color system with pass-level categories (green, pass; yellow, conditional pass; red, closed) similar to Sacramento County, California's system. Respondents could also select “none of the above.” For a respondent's preferred placard, they were then asked what the lowest rating for which they would feel comfortable eating at a restaurant. All survey questions were multiple choice or yes, no, or not sure, with the exception of zip code and age to allow for numerical input. The question of where a respondent wants to see restaurant inspection information allowed multiple items to be selected.

Data analyses

Descriptive statistics included calculating means and standard errors and proportions for demographic and categorical variables. Responses were compared by how frequently a respondent ate out on average per week (zero, one to two, three to four, five or more times per week) using chi-square tests for independence. Logistic regression was used to compare whether or not respondents wanted better access to restaurant inspection results (yes versus no) and whether they would use these results to decide where to eat (yes versus no, not sure) across categories of how frequently a respondent ate out on average per week. Models were adjusted for potential confounding variables that were selected a priori. These variables were age group (18 to 24, 25 to 44, 45 to 64, 65+ years), education (high school diploma/General Education Diploma or less, some college/associate's degree, bachelor's/4-year degree, graduate/professional degree), seven-county Minneapolis–St. Paul metropolitan area residence (yes or no), and previous experience of foodborne illness (yes or no). Statistical significance was determined at P < 0.05. Analyses were conducted using Stata 14.2 (StataCorp LLC, College Station, TX).

RESULTS

Participants

Overall, 1,204 surveys were conducted. Of these respondents, 1,200 completed the survey, for a response rate of 99.7%. Upon review of the data, 12 respondents were excluded because they did not meet the eligibility requirement. The final analytic sample included 1,188 respondents.

Demographics

The mean age of the 1,188 participants was 48.5 ± 0.5 years (Table 1). More than half (61.4%) of respondents had a bachelor's degree or higher. Overall, 88.6% of respondents reported their race as white or Caucasian and 95.1% were not Hispanic. Most respondents (76.9%) lived in the seven-county Minneapolis–St. Paul metropolitan area.

TABLE 1

Demographics of study participants overall and by frequency of eating out in an average week, Minnesota, 2019

Demographics of study participants overall and by frequency of eating out in an average week, Minnesota, 2019
Demographics of study participants overall and by frequency of eating out in an average week, Minnesota, 2019

Dining behaviors, attitudes, and perceptions

More than half (56.1%) of respondents reported eating out, on average, one to two times per week, 28.7% reported eating out three to four times per week, and 11.0% reported eating out five or more times per week (Table 2). Overall, 56.7% (n = 674) of respondents reported that they believe they have had food poisoning or experienced a foodborne illness; of these respondents, 80.7% (n = 544) believe they got sick from food at a restaurant, coffee shop, or other food service establishment.

TABLE 2

Dining behaviors, attitudes, and perceptions overall and by frequency of eating out in an average week, compared using chi-squared tests for independence, Minnesota, 2019

Dining behaviors, attitudes, and perceptions overall and by frequency of eating out in an average week, compared using chi-squared tests for independence, Minnesota, 2019
Dining behaviors, attitudes, and perceptions overall and by frequency of eating out in an average week, compared using chi-squared tests for independence, Minnesota, 2019

Overall, 54.5% of respondents reported that they believe it is more common for people in the United States to get food poisoning from a restaurant compared with food prepared at home. Furthermore, 72.6% of respondents reported that they have avoided eating at a restaurant or other food service establishment because they, someone they know, or someone they have read or heard about had gotten sick from consuming food at the establishment. However, respondents were confident that Minnesota restaurants maintain high food safety standards, with 19.6% reporting they were very confident and 55.3% reporting that they were moderately confident.

Interest in public disclosure and preferred disclosure formats

There was widespread support for having better access to restaurant inspection results in Minnesota. Overall, 94.4% of respondents supported having better access to inspection results (Table 3). Those who ate out one to two times per week on average (odds ratio [OR]: 3.3; 95% confidence interval [CI]: 1.7 to 6.4; P < 0.01) and three to four times per week on average (OR: 2.9; 95% CI: 1.4 to 5.9; P = 0.01) were significantly more likely to want better access to restaurant inspection results compared with those who ate out five or more times per week (Table 4). This was also true after adjustment for age, metropolitan residence, education, and previous experience of foodborne illness. Among those who wanted better access, 79.7% said that they would use inspection results to decide whether or not to eat at a restaurant if they were posted onsite, with significant variations by how often a respondent ate out (P = 0.01). The odds of reporting that they would use inspection results to decide where to eat if they were available at restaurants was significantly higher among those who ate out one to two times per week compared with those who ate out five or more times per week on average (OR: 1.7; 95% CI: 1.1 to 2.5; P = 0.02). Among those who did not want better access to inspection results, 35.4% reported that they would use them to decide where to eat.

TABLE 3

Accessibility and preferred disclosure format overall and by frequency of eating out in an average week, compared using chi-square tests for independence, Minnesota, 2019

Accessibility and preferred disclosure format overall and by frequency of eating out in an average week, compared using chi-square tests for independence, Minnesota, 2019
Accessibility and preferred disclosure format overall and by frequency of eating out in an average week, compared using chi-square tests for independence, Minnesota, 2019
TABLE 4

Differences in wanting restaurant inspection results to be more accessible and use in dining decisions compared using logistic regression, Minnesota 2019

Differences in wanting restaurant inspection results to be more accessible and use in dining decisions compared using logistic regression, Minnesota 2019
Differences in wanting restaurant inspection results to be more accessible and use in dining decisions compared using logistic regression, Minnesota 2019

When asked where they would like to see inspection results reported, respondents selected a median of two of the four options. The most frequently selected location was an online webpage (71.6%), 62.1% wanted to see inspection results at restaurants, and 34.3% wanted to see these results on a mobile application. Overall, 25.2% wanted to see inspection results upon entering, 18.6% wanted to see them before entering (e.g., in the window, visible from outside the restaurant), and 18.4% wanted to see them in both locations. There were no significant differences for where respondents wanted to see inspection results by how often they ate out on average.

The preferred format for disclosure of restaurant inspection results at restaurants was a numerical score and letter grade (34.0%) followed by stoplight colors and categories (25.3%), a letter grade (22.3%), and emojis and categories (16.8%). When asked what the lowest rating for their preferred system a respondent would feel comfortable eating at a restaurant, the most commonly selected choice varied by how many levels the placard system had (Table 5). For the numerical and letter grade placard, which had three levels, 30% reported feeling comfortable at the highest rating (A/90 to 100) and 54% felt comfortable at the second-highest rating (B/80 to 89). By contrast, for those who preferred the stoplight system, 66% reported that the lowest rating for which they would feel comfortable eating at a restaurant was the higher of the two ratings a restaurant could receive and still be open (green, pass).

TABLE 5

For a respondent's preferred posting type, the lowest rating for which they would still feel comfortable eating at a restaurant, Minnesota 2019

For a respondent's preferred posting type, the lowest rating for which they would still feel comfortable eating at a restaurant, Minnesota 2019
For a respondent's preferred posting type, the lowest rating for which they would still feel comfortable eating at a restaurant, Minnesota 2019

DISCUSSION

This survey showed that there was widespread support for improved access to restaurant inspection results in Minnesota. Furthermore, the majority of respondents reported that if inspection results were posted at restaurants, they would use them to decide where to eat. Given these findings and that previous studies on public disclosure of inspection results have shown improvements in sanitary conditions in restaurants (5, 22) and reductions in foodborne illness rates (9, 13, 19), adoption of public disclosure of restaurant inspection results at the point of service may be a useful strategy for Minnesota to help reduce the burden of foodborne illness.

Nearly three-fourths of respondents wanted restaurant inspection reports to be available online, and more than half of respondents wanted to see results available at restaurants. Although the preferred format for disclosure at restaurants was a letter grade and numerical score, this appeared to be only a moderate preference over the other responses, and there was variation by how often a respondent ate out on average. In a study to inform a disclosure policy for New Zealand that compared a letter grade placard and speedometer-type gauge, the letter grade was preferred and was perceived to convey the information in a more interesting manner (8). Interestingly, our study found that respondents who currently work in food service preferred just a letter grade over the other options, which was also consistent with food service workers in New Zealand who preferred the letter grade over the speedometer gauge (8). Respondents who ate out on average one to two times or three to four times per week were the majority of those sampled and appeared to have similar responses. By contrast, respondents who rarely ate out or more frequently ate out did not appear to value access to inspection results as much. This has implications for developing a disclosure system that will be most useful for consumers. Additional research is needed to determine the most appropriate and useful method and audience for disclosure at restaurants.

More than half of survey respondents reported that they believe it is more common to get food poisoning from restaurants compared with food prepared at home, which is consistent with findings from the 2016 U.S. Food and Drug Administration Food Safety Survey (15). Furthermore, nearly three-fourths of respondents stated that they have avoided eating at a food service establishment because they, someone they know, or someone they have heard of has gotten sick from eating there. Despite this, survey respondents expressed confidence that Minnesota restaurants maintain high food safety standards. In a population-based sample to understand perceptions of New York City's public disclosure program, 76% of respondents felt more confident in a restaurant's food safety when an A grade was posted (22). In our study, when the preferred disclosure format included more than two levels (e.g., A, B, or C), respondents most frequently reported still feeling comfortable eating at the second-highest rating (e.g., B) than the highest. When there were only two options, however (e.g., pass or conditional pass), respondents were more likely to feel comfortable eating only at the highest level. This is consistent with a Canadian study where the majority of respondents expressed that they would be unlikely to eat at a restaurant that had a “conditional pass” placard (12). These findings not only provide useful information that Minnesota adults would use inspection results to inform their dining decisions but also inform how they would use them. Studies have shown, however, that the decision to dine at lower rated restaurants is affected by factors including dining frequency (1) and whether or not the restaurant had recently been closed due to a public health hazard (11). Future studies are needed to further understand how consumers use these information cues in their dining decisions to develop effective public disclosure programs.

Although this study was conducted among a large sample size, there are several notable limitations. The study population was not a random sample of Minnesota residents, which may affect the generalizability of the results. Respondents were those that chose to visit the University of Minnesota D2D Research Facility and may not be representative of all Minnesota residents. Our survey sample included a high proportion of highly educated respondents, a large proportion of older adults, and a greater proportion of respondents who lived in the seven-county metro area. Among survey respondents aged 25 years or older, 66.4% had a bachelor's degree or higher compared with 34.8% of all Minnesotans (20). Similarly, 41.2% of respondents were aged 45 to 64 years compared with 26.7% of Minnesotans (20). Overall, 76.9% of respondents lived in the seven-county metro area, compared with 55.0% of the total population (21). However, even after adjustment for demographic variables, results showed respondents who ate out one to two times per week and three to four times per week were more likely to report wanting better access to restaurant inspection information. Future studies should be conducted that include a population-based sample. Strengths of this study include the large sample size and the intentional recruitment and design of survey questions to not bias results. Study advertising materials did not reference restaurants, inspections, or food safety by name in an attempt to reduce self-selection among participants who may have strong opinions about restaurant food safety.

Overall, this study shows that there is interest in having better access to restaurant inspection results in Minnesota, both online and at restaurants. Coincidentally, on the last day of data collection, 27 August 2019, the city of Minneapolis announced that health inspection reports will soon be posted online (6). Our study provides evidence that there is widespread support for this in Minneapolis and across the state. Increasing public access to inspection results could reinforce efforts by environmental health agencies and foodservice operators to improve the safety of foods eaten outside the home.

ACKNOWLEDGMENTS

We thank the following people for reviewing the survey and/or assisting with data collection: Patrick Balius, Jeff Bender, Kim Carlton, Logan Ebeling, Nicole Hedeen, Elizabeth Johnson, Sara Lammert, Ross A. McKenzie, Paige Mushinski, Nancy Rolstad, Robert L. Scharff, Joseph Servadio, Kirk Smith, Katie Tastad, and Madhura S. Vachon. We also thank the following D2D Research Facility staff: Brian Bania, Lauren Duval, Annie Hotop, and Atticus M. Wolfe. This research was supported by the National Institutes of Health's National Center for Advancing Translational Sciences Grant UL1TR002494. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health's National Center for Advancing Translational Sciences. Additional support was provided by the University of Minnesota Graduate School Doctoral Dissertation Fellowship.

REFERENCES

REFERENCES
1.
Aik,
J.
,
Newall
A. T.
,
Ng
L.-C.
,
Kirk
M. D.
, and
Heywood
A. E.
.
2018
.
Use of the letter-based grading information disclosure system and its influence on dining establishment choice in Singapore: a cross-sectional study
.
Food Control
90
:
105
112
.
2.
Centers for Disease Control and Prevention
.
2019
.
FoodNet fast
.
U.S. Department of Health and Human Services
,
Centers for Disease Control and Prevention, Atlanta
.
3.
Centers for Disease Control and Prevention
.
2019
.
Surveillance for foodborne disease outbreaks, United States, 2017, annual report
.
U.S. Department of Health and Human Services
,
Centers for Disease Control and Prevention, Atlanta
.
Available at: https://www.cdc.gov/fdoss/pdf/2017_FoodBorneOutbreaks_508.pdf. Accessed 15 October 2019.
4.
Choi,
J.
,
Miao
L.
,
Almanza
B. A.
, and
Nelson
D. C.
.
2013
.
Consumers' responses to restaurant inspection reports: the effects of information source and message style
.
J. Foodserv. Bus. Res
.
16
:
255
275
.
5.
Choi,
J.
, and
Scharff
R. L.
.
2017
.
Effect of a publicly accessible disclosure system on food safety inspection scores in retail and food service establishments
.
J. Food Prot
.
80
:
1188
1192
.
6.
City of Minneapolis
.
2019
.
Online inspection reports, priority 3 violation point change, Somali CFPM class - environmental health eNews
.
Accessed 27 August 2019.
7.
Driven to Discover
.
2019
.
D2D: the Driven to Discover Research Facility
.
Available at: http://d2d.umn.edu/. Accessed 10 December 2019.
8.
Filion,
K.
, and
Powell
D.
.
2011
.
Designing a national restaurant inspection disclosure system for New Zealand
.
J. Food Prot
.
74
:
1869
1874
.
9.
Firestone,
M. J.
, and
Hedberg
C. W.
.
2018
.
Restaurant inspection letter grades and Salmonella infections, New York, New York, USA
.
Emerg. Infect. Dis
.
24
:
2164
2168
.
10.
Fung,
A.
,
Graham
M.
, and
Weil
D.
.
2007
.
Full disclosure: the perils and promise of transparency
.
Cambridge University Press
,
Cambridge
.
11.
Hatfield,
T. H.
, and
Seiver
O. H.
.
2001
.
Preference reversals in grading systems for retail food facilities
.
J. Environ. Health
63
:
19
25
.
12.
Henson,
S.
,
Majowicz
S.
,
Masakure
O.
,
Sockett
P.
,
Jones
A.
,
Hart
R.
,
Carr
D.
, and
Knowles
L.
.
2006
.
Consumer assessment of the safety of restaurants: the role of inspection notices and other information cues
.
J. Food Saf
.
26
:
275
301
.
13.
Jin,
G. Z.
, and
Leslie
P.
.
2003
.
The effect of information on product quality: evidence from restaurant hygiene grade cards
.
Q. J. Econ
.
118
:
409
451
.
14.
Jones,
J. L.
, and
Angulo
F. J.
.
2006
.
Eating in restaurants: a risk factor for foodborne disease?
Clin. Infect. Dis
.
43
:
1324
1328
.
15.
Lando,
A.
,
Verrill
L.
,
Liu
S.
, and
Smith
E.
.
2017
.
2016 Food safety survey report. U.S
.
Food and Drug Administration
,
Washington, DC
.
Accessed 14 October 2019.
16.
Lee,
P.
, and
Hedberg
C. W.
.
2016
.
Understanding the relationships between inspection results and risk of foodborne illness in restaurants
.
Foodborne Pathog. Dis
.
13
:
582
586
.
17.
Minnesota State Fair
.
2019
.
Attendance
.
Available at: https://www.mnstatefair.org/about-the-fair/attendance/. Accessed 2 October 2019.
18.
Okrent,
A. M.
,
Elitzak
H.
,
Park
T.
, and
Rehkamp
S.
.
2018
.
Measuring the value of the U.S. food system: revisions to the Food Expenditure Series
.
Technical Bulletin 1948
.
USDA, Economic Research Service
.
19.
Simon,
P. A.
,
Leslie
P.
,
Run
G.
,
Jin
G. Z.
,
Reporter
R.
,
Aguirre
A.
, and
Fielding
J. E.
.
2005
.
Impact of restaurant hygiene grade cards on foodborne-disease hospitalizations in Los Angeles County
.
J. Environ. Health
67
:
32
36
.
20.
U.S. Census Bureau
.
2013–2017
.
American Community Survey 5-year estimates
.
U.S. Census Bureau
,
Washington, DC
.
21.
U.S. Census Bureau
.
2019
.
Annual estimates of resident population: April 1, 2010 to July 1, 2018
.
U.S. Census Bureau
,
Washington, DC
.
22.
Wong,
M. R.
,
McKelvey
W.
,
Ito
K.
,
Schiff
C.
,
Jacobson
J. B.
, and
Kass
D.
.
2015
.
Impact of a letter-grade program on restaurant sanitary conditions and diner behavior in New York City
.
Am. J. Public Health
105
:
e81
e87
.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)