Food safety policies and practices in public spaces: The urban water, sanitation, and hygiene environment for fresh fish sold from individual vendors in Mzuzu, Malawi

Abstract In sub‐Saharan Africa, informal markets account for more than 80% of the total food selling. Fish is a major protein source for households in Malawi and is commonly purchased from individual vendors. The aim of this study was to review national acts and policies and local regulations focused on fresh fish sold at open‐air markets or by mobile vendors and to further examine the water, sanitation, and hygiene environment that may impact food safety in Mzuzu City, Malawi. The study used interviews, an observational checklist, and sampling of water and fish skin. In general, there was limited oversight of food safety where fresh fish are sold by vendors, and food safety guidance was inadequate. There was access to water in three of the four markets, but only two markets had safe water (0 cfu/100 ml for Escherichia coli). All vendors stored water in a container for use throughout the day to sprinkle over the fish with their bare hands to keep them from drying out. The mean washing water E. coli level was 700 cfu/100 ml. All fish skin samples (25/25) were positive for the presence of Salmonella spp., and most had high levels of E. coli. Sanitation facilities were available for vendors and customers in two of the four markets, but the use was limited. This research identified three key opportunities: (a) Regulatory framework including informal markets and mobile vendors; (b) Safe water, clean and functional toilets, and handwashing stations with soap at every market; and (c) Foodborne disease education for vendors.

supply, sanitation, and hygiene conditions have been linked to a high prevalence of undernutrition (Cumming & Cairncross, 2016;Dodos, Mattern, Lapegue, Altmann, & Aissa, 2017), the focus of previous studies has been on household access rather than where food is sold from urban public spaces. Morse, Masuku, Rippon, and Kubwalo (2018) noted that in Malawi "there is a significant threat to public health and {local food} market access due to uncoordinated, outdated or incomplete regulatory framework, poorly defined mandates, limited infrastructure, lack of equipment and skilled personnel, inadequate resources, and limited awareness and ability to comply with standards." Riley, Chilanga, Zuze, and Joynt (2018) noted that in Mzuzu, Malawi, fish is a main source of dietary protein, and fish was rated as more affordable by households than beef, pork, chicken, or other meat. Fresh fish is regularly consumed by 47% of households and is purchased at either area markets or from street sellers (Riley et al., 2018). Fish do not naturally harbor Escherichia coli (E. coli) or Salmonella spp. (Heinitz, Ruble, Wagner, & Tatini, 2000;Hansen, Clark, Ishii, Sadowsky, & Hicks, 2008;Food and Agriculture Organization of the United Nations [FAO], 2012). However, contamination can occur along the supply chain, as fish are handled by several intermediaries from the time of capture to sale at local markets (Samikwa, Kapute, Tembo, Phiri, & Holm, 2019). Heinitz et al. (2000) found a 12.2% incidence of Salmonella spp. in fresh fish imported into the United States.
The aim of this study was to review national acts and policies and local regulations focused on fresh fish sold at open-air markets or by mobile vendors and perform a detailed case study to further examine the water, sanitation, and hygiene infrastructure, vendor knowledge and behavior, and (as applicable) the presence of E. coli and Salmonella spp. that may impact food safety in Mzuzu City, Malawi.

| Study area
The study was conducted in Mzuzu City, Malawi, which covers 146 square kilometers and has a population of 220,000 according to the 2018 census (Malawi Government, 2018). There are 15 openair markets in the city, which provide products for daily consumer demands, including local fruits, vegetables, grains, and protein ( Figure 1). For 76% of households, a food market can be reached in less than a 30-min walk (Mzuzu City Council, 2014).

| Data collection methods
Study data were collected using an observational infrastructure checklist, interviews, sampling of water, and sampling of fresh fish skin from October 2018 to February 2019.
Following a review of national acts and policies, and local regulations, four agencies found to be present in Mzuzu and able to provide monitoring or evaluation were interviewed using a structured interview in English. They were asked what they knew of the regulatory framework covering water, sanitation, hygiene, and food safety supporting fresh fish sold at markets or by mobile vendors. In addition, they were asked the frequency and type of monitoring and evaluation performed. At each permanent open-air market (n = 4), the use of the nearest sanitation facility (including access for both vendors and customers) was observed, and the same researcher (the first author) counted the number of users who went in/out of the sanitation facility over a period of 1 hr between the hours of 8 a.m. and 4 p.m. for five consecutive days. Immediately following this, the number of general market users (including vendors and customers) was counted over a 1-hr period. An observational checklist that covered water access, handwashing station with soap, and sanitation facility infrastructure was completed at each market (n = 4).
When a water source was available at the market, a 100 ml water sample was collected (n = 3; Mzuzu Central Market, Chibavi Market, and Zolozolo Market). Two fresh fish were purchased, and a 100 ml sample of water used for washing the fresh fish was also collected from the 25 vendors. Water samples were collected in Whirl-Pak ® F I G U R E 1 Fresh fish being sold at an open-air market, Mzuzu, Malawi bags containing sodium thiosulfate (Nasco). Fish samples were collected in Whirl-Pak ® plastic bags. Samples were transported and analyzed at an academic laboratory at the department of Fisheries and Aquatic Science, Mzuzu University.
Escherichia coli and Salmonella spp. were selected as nonindigenous bacteria of public health significance introduced in fish and fishery products through environmental contamination by domestic and/or industrial waste (FAO, 2012). The E. coli analysis was performed using a field-based plate method for contamination monitoring (Thomas, Andrés, Borja-Vega, & Sturzenegger, 2018). The Salmonella spp. analysis methodology was based on Heinitz et al. (2000), but the product sample was reduced to a 1.5 g portion and the procedure adapted for a field-based method based on prestudy trials in our laboratory.
Water samples (n = 28; 25 samples of water used by individual vendors for washing the fresh fish onsite and three market water source samples) and fresh fish skin samples (n = 25) were analyzed for E. coli using 3M™ Petrifilm™ E. coli/Coliform Count Plates (3M™).
For market water source samples (n = 3), a 100 ml volume was sampled. Fresh fish samples (n = 25) were prepared by laterally scraping the skin surface (approximately 1.5 g) using an aseptic blade and adding the skin to 10 ml previously boiled and cooled water in a Whirl-Pak ® plastic bag. The mixture was hand shaken for 2 min, and then, 1 ml of this water was analyzed for E. coli.
All samples for E. coli analysis were incubated within 6 hr of sample collection. Samples were incubated at 35°C for 24 hr, and the results are presented as colony forming units per 100 ml (cfu/100 ml) for water and cfu/g for fish skin. An equipment blank, as a negative control, of boiled and cooled water was analyzed each day (n = 5), and a value of zero for E. coli was found for each. No positive controls for E. coli or total coliforms were conducted.
A second fish sample was prepared by combining approximately 1.5 g of skin and 10 ml of prepared 3M™ Salmonella Enrichment Base and 3M™ Salmonella Enrichment Supplement in a Whirl-Pak ® plastic bag.
The mixture was incubated at 41.5°C for 24 hr. Then, a cotton swab was dipped in the sample, and it was spread on a 3M™ Petrifilm™ Salmonella Express Plate. Samples were incubated a second time at 41.5°C for 24 hr, and presumptive positives were identified. Five equipment blanks using boiled and cooled water, but the same methodology, were analyzed as a negative control, and all were found to be negative for Salmonella spp. No positive controls for Salmonella spp.
were conducted.

| Data analysis
Laboratory samples were analyzed in duplicate, and the mean results were recorded. Statistical analyses were conducted using the R Project 3.5.1 statistical package (Vienna, Austria). Water source results for E. coli were compared to national and WHO guidelines for drinking water (Malawi Bureau of Standards [MBS], 2013;WHO, 2017). If a p value was less than a .05 significance level, we concluded there was a significant difference.

| Ethics
This study and its informed consent procedures were approved by the Republic of Malawi National Commission for Science and Technology (Protocol Number P.10/18/327). Informed, verbal consent was obtained from the study participants.

| National acts, policies, and local regulations
The key national acts, policies, and local regulations covering water, sanitation, and hygiene infrastructure standards, and food safety evaluation criteria for fresh fish being sold at urban openair markets or by mobile vendors are presented in Table 1. While some of these mention markets, none address mobile vendors.
There were no specific criteria in national acts or policies regarding standards for water, sanitation, and hygiene infrastructure, or monitoring and evaluation guidelines in public spaces where fresh fish are being sold. There were, however, some local regulations There's currently no specific food safety policy governing the sale of fresh fish by mobile vendors. But in general terms, the Public Health Act gives powers to health inspectors to inspect food and food premises to ensure that food is wholesome and safe for human consumption. Monitoring of this has worked in formal food premises but has been haphazard in the informal food sector as we do not have policies that specifically address informal food markets. Existing The respondent from the MBS knew there was a frozen fish standard but was not aware of the fresh fish specification from their same organization (MBS, 2007).
TA B L E 1 Regulatory framework supporting the link between water, sanitation, and hygiene infrastructure standards and food safety evaluation criteria where fresh fish are sold, Mzuzu, Malawi

| Water, sanitation, and hygiene environment for fresh fish vendors
Mzuzu City Assembly (2001)  When water was available, the market vendors (15/15) reported using it; however, at the one market without a water supply (Area 1B Market), vendors reported either obtaining water from their home (4/5) or buying it from a nearby shop (1/5).
All vendors (25/25) used stored onsite water to sprinkle over the fish throughout the day to keep them from drying out; they called this "washing water." Mobile vendors stored water in a 1-or 2-L plastic bottle. Market-based vendors stored it in a 5-to 20-L metal or plastic bucket. Water was splashed from the water storage container onto the fish with bare hands.  Where there were sanitation facilities present, the vendors (8/10) generally reported using them, though this was contrary to our researcher observations of a mean of one person per hour using each sanitation facility. The sanitation facilities were designed for a higher F I G U R E 2 Escherichia coli in fresh fish skin, stored washing water, and water sources where fresh fish are sold, Mzuzu, Malawi level of use than was observed, as one market had five stalls and the other had two stalls.
The eight vendors who reported using their market's sanitation facilities said they paid MK50-100 (USD 0.07-0.14) as a pay-peruse system. These fees may be a barrier. In order to pay for the use of the sanitation facilities, a vendor in this study would have to sell a small pile of E. sardella containing 4-8 fishes or one individual O.

microcephalus.
The Chibavi market sanitation facilities were cleaned by the community and had the lowest cost per use. Furthermore, all of these vendors (5/5) indicated that they were being managed well and they were satisfied with the safety and cleanliness. It was stated that cleaning was performed daily. In contrast, the second market with a sanitation facility, the Mzuzu Central Market, was cleaned by local government employees (from the City Council) and had fewer (2/5) vendors who indicated that it was being managed well and fewer (2/5) vendors who were satisfied with the safety and cleanliness.
When vendors were asked "How does the condition of the sanitary facility present here compare to your home?", all of the Chibavi market vendors (5/5) reported that it was better than that of their home but fewer of those (2/5) at the Mzuzu Central Market said it was better.
Solid waste was also a problem at each market area, with all mar-  When using ice, vendors reported getting it from a shop within the market area; they did not make their own ice. For the mobile vendors, fish were not covered with block ice, a sunlight barrier, or a dust barrier.

| D ISCUSS I ON
Handwashing with soap and water by vendors while at work and keeping food at a proper temperature appear to be the most effective ways of reducing E. coli and Salmonella spp. contamination in fresh fish sold by vendors in Mzuzu, Malawi. Still, knowledge of foodborne diseases is low among vendors. To accelerate these behaviors, more training is needed along with proper infrastructure to facilitate the practice. Although the vendors' handwashing practices were not directly observed, the lack of soap was inferred to contribute to the presence of E. coli and Salmonella spp., as the vendors sprinkled water onto the fish throughout the day. While the FAO (2009) guidelines state that potable water is required when there is contact with fish, only three of the four markets had water available for fish vendors, and this water was potable at only two markets.

Local (Mzuzu City Council, 2014) and national (National Statistical
Office [Malawi] and ICF, 2017) documents track the water supply and sanitation infrastructure for households, but data are limited for public spaces and mainly comprise data from schools and health facilities.
In our study, we found that the national food safety acts and policies and local regulations for fresh fish that are in place are more applicable to formal retail settings, which primarily serve higher income customers, than to open-air markets. The largest gap was for mobile vendors, where deficiencies are not being addressed at all.
Even for the limited national guidelines that are in place, we found that they were not followed. For example, none of the vendors were observed to have any way of determining whether they were in com- Globally, as shown by Cumming and Cairncross (2016), poor water, sanitation, and hygiene access have in some cases been linked to poor childhood growth and development. Despite these findings, this is a complex issue because interventions in low-and middle-income countries have dominantly focused on household access rather than public spaces where food is present and purchased.
For weaning children, even the safety of cooked foods is important.
There is an important link between our study results and the work performed by Touré, Coulibaly, Arby, Maiga, and Cairncross (2013) in Mali on home-prepared soup cooked with local fresh fish and vegetables commonly eaten as a weaning food; they reported that after cooking, 68% of fish soup samples had greater than 100 thermotolerant coliforms/g and that after further storage at room temperature, 90% of fish soup samples had greater than 100 thermotolerant coliforms/g. However, the prevalence of thermotolerant coliforms was effectively reduced after the mothers improved their hygiene practices during cooking and storage by using potable water, washing dishes, washing hands, and reheating leftovers. Importantly, Touré et al. (2013)  Council also has several implementing partners working on water supply, sanitation, and hygiene promotion in public spaces. One specific example is the European Union, which invested approximately €3,000,000 (USD 3,400,000)  The timing of these projects provides an opportunity to envision a food safety chain as part of a comprehensive urban planning approach that combines safe water together with sanitation access for public places, including food markets. The results of our study raise the question of whether these projects actually improved the living standards in urban public spaces or improved food safety for the people of Mzuzu City.
For just these three projects, the total investment in Mzuzu was about USD 15 per person in the city. Though the Mzuzu Central Market had new sanitation facilities that were built under one of these projects, the existing facilities adjacent to the fish market included in our study were still being used because of the convenience of being within 30 m of the fish vendors' selling point, as opposed to the new sanitation facility 160 m away and across a busy highway (the newer facility was not included in our study because it was not commonly used by fresh fish vendors). In addition to sanitation facilities being in good condition, a vendor or customer must also want to use them. Similar to the findings by Afacan and Gurel (2015) on public sanitation facilities in Turkey, our sanitation facilities had issues with cleanliness. However, in our study, even when payment was required to use public sanitation facilities, it did not necessarily mean they were clean. Based on their reported working hours, vendors would likely need to use a sanitation facility multiple times during the day, incurring a cost of MK 50-100 (USD 0.07-0.14) for each use. Thus, sanitation facilities that are not in working order or that people prefer not to use may lead to increased open urination or defecation or the use of other unimproved facilities that are farther away from their place of work. The situation is worse for the mobile vendors who would need to find sanitation facilities on their route and leave their bike and merchandise. The lack of handwashing with soap in all the markets is a critical gap from a food safety standpoint. Based on our study results, at least for the two markets with working sanitation facilities, the sanitation facilities were probably sufficient for users, whereas at the two other markets, this infrastructure should be immediately prioritized. McGinnis, Marini, Amatya, and Murphy (2019) noted that well-maintained privately owned, pay-per-use community sanitation facilities in Nepal were as clean or cleaner than household sanitation facilities. Based on the swab sample results for bacterial concentrations in that study, in community sanitation facilities, the seat/floor, tap/handle/bucket used for anal cleansing, and door handles were the surfaces to focus on when cleaning.

| LI M ITATI O N S
The number of markets and vendors included in this study was small but did include a range of geographic areas and market sizes.
This study did not involve market owners and operators nor did it consider the profit margin of vendors compared to food safety practices or consumer food safety awareness. Intermediaries as a source of introducing contamination into the fresh fish supply chain were not addressed in this study. The fish in this study would traditionally be cooked prior to consumption. Further research to include more of the mobile fresh fish sellers operating in the city is needed.

| CON CLUS ION
In conclusion, there are food safety gaps for fresh fish sold from individual vendors in Mzuzu, Malawi, resulting from urban water, sanitation, and hygiene infrastructure, the lack of vendor knowledge and behavior, and inadequate oversight. This study on four open-air markets and mobile fish vendors found high E. coli levels in the water used for washing the fresh fish onsite and on the fish skin, and only two markets had a water source free of E. coli for vendors to use.
By using E. coli and Salmonella spp. analyses alongside sanitary surveys and interviews, our study provides some preliminary evidence that indicates a need for critical control points to improve food safety

ACK N OWLED G M ENTS
Texas Tech University is acknowledged for providing consumable laboratory supplies. The authors also appreciate the technical support from Amanda Joynt and editing by Kip McGilliard.

CO N FLI C T O F I NTE R E S T
The authors declare that they do not have any conflict of interest.

E TH I C A L A PPROVA L
This study was approved by the Republic of Malawi National Commission for Science and Technology (Protocol Number P.10/18/327).

I N FO R M E D CO N S E NT
Verbal informed consent was obtained from all study participants.