Livestock Research for Rural Development 25 (2) 2013 Guide for preparation of papers LRRD Newsletter

Citation of this paper

Application of Hazard Analysis Critical Control Point (HACCP) practice to the milk production and marketing chain: the case of Tanga region, Tanzania

E S Swai and L Schooman*

Veterinary Investigation Centre, PO Box 1068 Arusha, Tanzania
* Tanga Dairy Trust, PO Box 1720, Tanga, Tanzania


To reduce the risk of milk-borne zoonotic infections in Tanzania, control should take place at all levels of the marketing chain from the farm to the consumer. Hazard Analysis-Critical Control Point (HACCP) practice has been applied in this review to identify and describe potential hazards along the milk production and marketing chain with a view to extrapolating its applicability to the dairy industry of Tanzania.

The HACCP may be considered relevant in some production systems but the review revealed that it requires involvement of actors at all levels, from producers to processors, animal health service providers, regulators, retailers and consumers. Information, Education and Communication (IEC) combined with improved diagnosis and testing at different levels along the milk production and marketing chain should further contribute to the reduction of risk hazards. Given the complexity of integrating HACCP concepts into a chain that is dominated by informal operations the HACCP concept should be seen as part of an efficient total hygiene practice in farm units, animal health services, milk collecting centres, kiosks, processing plants and distribution points. This emphasizes the need for the adoption of “good farming practices, good veterinary practices, good hygienic practices”, which ultimately lead to foundational practices allowing for the implementation of on-farm, processing plant, storage and retail point HACCP programmes. 

Key words: dairy sector, food safety, unpasteurized milk, zoonoses


Milk is the most perishable of all farm produce. It is an ideal culture medium for the growth of a variety of micro-organisms due to its nutritive value (Chambers 2002) and is vulnerable to microbial invasion and food poisoning. Unhygienic practices in farm units, at milk collecting centres (MCC), processing plants and during post processing handling are associated with a potential health risks to consumers due to the presence of pathogens in the milk and to environmental contamination. Contamination during the retail selling process can further be attributed to handling equipment (cans, tins, measuring jars), air, personnel and even consumers (Swai and Schoonman 2011). Microbial contamination of milk and milk products must not exceed levels that could adversely affect shelf life and, if it does, it renders the milk unwholesome and hence unfit for human consumption (ILRI 2005). 

Milk in Tanga is marketed through a combination of formal and informal milk marketing channels. Raw or traditional (often called “informal”) milk markets account for over 70% of milk sales in Tanzania (Staal 2000; ILRI 2005; URT 2010; Schoonman and Swai 2011). The market agents involved include farmer dairy co-operatives, small traders using bicycles and public or private transport and small retail outlets such as kiosks and shops. Unprocessed milk carries the risk of infection by several milk-borne diseases, including zoonoses. Recent studies on microbial quality of raw marketed milk in Tanga have indicated high levels of coliforms (Swai and Schoonman 2011). Informally marketed milk poses a high risk as it is not quality checked. The quality check in formal marketing channels does not, however, include testing for coliforms, for brucellosis and for other zoonotic diseases. The most important means of acquiring milk-borne infections is through consumption of raw milk. The risk of infection is the reason for public health regulations which should discourage the informal milk markets that sell unpasteurized milk (Mwangi et al 2000) but these regulations are rarely enforced in many developing countries and especially in Tanzania. 

An important step in developing targeted policies more supportive of market participation by the majority of players is to conduct risk analysis (risk-assessment, risk-management and risk-communication) on milk-borne health risks applied to the various production and marketing systems. There is a need to promote and increase implementation of Hazard Analysis of Critical Control Points (HACCP) and consumer food safety education efforts at all stages of the milk production and marketing chain. The HACCP can be defined as a system which identifies, evaluates and controls hazards that are significant for food safety (FAO 1997). If HACCP is applied to milk marketing it should underpin the approach to milk safety at all stages of the chain. Controlling the risk of human infection with zoonotic disease can then be controlled by Critical Control Points (CCPs) at different stages, resulting in a sustained reduction of pathogens at relevant points. 

There are few studies conducted to characterize milk-borne hazards in Tanzania despite increased consumption over the last few decades of raw, unpasteurized milk (Shirima et al 2003; Mdegela et al 2004; Kivaria et al 2006). In this study, we use HACCP principles to explore hazard points for milk along the production chain to retail milk outlets using the Tanga milk shed as an example area. 

Materials and Methods

Study area 

This survey formed part of a broader research on epidemiology of leptospirosis in cattle and relative risk to public health which was carried out between December 2001 and May 2005. Tanga region is situated in the northern coastal area of Tanzania and lies between latitudes 4 and 6 South and longitudes 37 and 39 East. The climate varies from hot humid coastal lowlands in the east to the cool Usambara Mountains in the north and semi-arid plains in the southwest. There are two rainy seasons in most years (March-May and September-November) and annual rainfall ranges from 250 to 1,400 mm. Day time temperatures vary from 23-28 C during the cool season (May-September) to 30-33 C during the hot season (December-March). The mean relative humidity is 75% (range 60-90%). 


Cattle production system


Cattle in the area are kept either in traditional herds or at smallholder dairy farms. Cattle types kept under smallholder intensively managed dairy herds (≤10 cross-bred animals of all ages, breeds and sexes) include Bos taurus (Friesian, Ayrshire, Jersey, Simmental) and crosses of these with B.indicus (Tanzania shorthorn zebu (TSZ), Boran and Sahiwal). The level of taurine blood varies from 50 to 85%. Traditional extensively managed herds (≥30 indigenous cattle of all ages and sexes) comprise mainly TSZ and Boran. The systems can be divided into urban, peri-urban and rural. Traditional herds are found mainly in the rural lowlands in areas that are less suitable for crop production. Traditional cattle are confined in an enclosure with dirt flooring at night and are kept for meat, milk and for financial security. Smallholder dairy farmers manage their cattle under zero-grazing when they are kept in sheds that range from simple structures with a sand floor and a thatched roof to those with a concrete floor and iron sheets as roofing material. Other smallholder farmers graze their cattle by day and house them at night. Smallholder production in the past was mainly for subsistence but is now increasingly market orientated. Other livestock species of economic importance are goats, sheep, pigs and poultry.


Milk production, handling and marketing system 

Hand milking at smallholder dairy farms is mostly done in a crush or in a cowshed with a concrete floor. In traditional herds milking takes place in the open enclosure with a dirt floor. Detailed procedures on milking techniques, pre- and post-udder preparation, storage, transport, marketing channels and characteristics of milk marketing agents are described in detail by Swai et al (2005), Schoonman and Swai (2011) and Schoonman et al (2011). 

Risk pathways along the milk marketing chain 

Identification of risk pathways along the food production chain elaborates where risk factors occur and possible actions to reduce or control the risks of introduction and transmission of zoonoses. Risk pathways for transmission to the human population depend on the type of zoonosis and might be by direct contact, consumption of animal products or by vectors/fomites. The risk pathways along the chain for milk and milk products in Tanga are shown in Figure 1. The time after production (including preservation methods, processing, quality checks and inspection) differ along the various pathways, and between product handling agents and consumers. Human infection through consumption of the product can occur on farm or further along the chain in rural communities or in consumers in urban areas. The main difference between the two is that the risk pathway for the urban group is normally longer which can result in increased contamination of the product, either by cross contamination (e.g. in bulking milk from different sources) or by multiplication of the organism in the marketing chain.


Figure 1: Risk pathways for food borne zoonoses along the milk marketing chain in Tanga

HACCP framework/ principles  

HACCP identifies the points in the process that are hazardous, their risk factors and potential level of risk, so that CCPs for remedial action can be identified. The CCPs should be associated with the hazard and should lead to reduction in or control of the risk. Primary CCPs (CCP-1) are those that give 100% control, whereas secondary CCPs (CCP-2) provide only partial control of the risk. If, however, two CCPs each reducing the risk by half are applied in series the overall risk would be reduced to a quarter. The application of a number of partial controls along the chain and acting together can reduce risk to an acceptable level. The application of HACCP principles to milk-borne disease agents along the production to marketing chain is shown schematically in Figure 2 and involves:  


i.         Construction of a flow diagram showing the livestock production process and the food production chain.  

ii.       Carrying out a hazard analysis that starts with a risk assessment. Introduction of a disease agent into the herd, transmission of a disease agent within the herd and transmission of a disease agent to the human population are all different hazards in this context. The risk assessment includes hazard identification, exposure assessment, hazard characterization and risk characterization of the disease agent of concern. This should provide an estimate of the occurrence of the disease pathogen and the severity of infection in humans and costs in human and financial terms of a case and hence the importance of the pathogen as a risk for the population. This risk assessment depends on information produced by epidemiological studies. All the risk pathways for the hazards to occur have to be identified. A disease agent can have multiple pathways either for introduction into a herd or for transmission to the human population. As a last step of the hazard analysis, risk factors along the risk pathways have to be identified and ranked. The risk factors are identified either by epidemiological studies where Odds Ratios (or Relative Risks) are produced for these factors through univariate or multivariate analysis or from literature and conjoint analysis where experts rank the factors.


Figure 2: Conceptual inventory of critical and control points along the milk production to consumer chain


iii.      Selecting possible CCPs where these should have a relationship with the hazard (e.g. infection of the herd), be in balance with human and financial resources and be feasible in the system. Ideally, CCPs should be quantifiable in relation to the hazard. This might not always be possible, but it is important to know if a CCP is of the CCP1 or CCP2 type (Figure 2). Identification of a CCP may be aided by application of a decision tree. 

The decision where to select a CCP, at an earlier or later stage along the risk pathway depends not only on the nature of the disease agent and its consequences, management practices along the risk pathway, human and financial resources available and consumer practices but also on information of the presence of the disease agents in the animal production system and along the animal production chain. If, for instance, a certain disease agent is not yet present in the herd but present in the environment, the hazard of introduction into the herd should be controlled. If the disease agent is already present in the herd, or if there is no proper control available to prevent introduction into the herd, control should focus more on transmission within the herd and to the human population. The transmission from old animals to replacement animals needs to be controlled if reduction/eradication of the disease is to be achieved. In considering the risk of human infection, it is important to know the division of labour on the farm to identify which persons are particularly at risk. 

Figure 3 shows a decision tree for identification of CCPs along the marketing chain. The decision tree can be applied at all steps of the marketing chain. When identifying CCPs at a certain step in the marketing chain, the possibility of an increase of the hazard and further control measures down the chain, needs to be taken into consideration. Cooling of milk, for example, might not eliminate or reduce the hazard of Vero-cytotoxigenic Escherichia coli (VTEC) as such but it stops the multiplication of the organism and hence the increase of the hazard. If it is not known if downstream processing takes place to eliminate or reduce this hazard, then cooling of milk should still be considered a CCP. Depending on the information available and the hazard to be controlled, CCPs might be selected earlier or later along the chain. For hazards which cannot be eliminated completely by one CCP, a combination of CCPs at different stages should be selected. Even if one CCP along the marketing chain could eliminate the hazard completely, e.g. pasteurization of milk in the case of tuberculosis and brucellosis, it can break down. Selection of more than one CCP along the marketing chain is therefore a valid step to reduce the risk. In the case of TB and brucellosis it might be better to aim for eradication at farm level as well, rather than relying on pasteurization alone.


Figure 3: Decision tree for identification of CCPs along the marketing chain

Applicability of HACCP 

The HACCP was designed to guarantee the quality of animal products. As such it was mainly designed for a linear pathway, for example a meat processing plant. The introduction of diseases into a herd, and in particular disease agents capable of infecting human populations, can follow multiple pathways. Noordhuizen et al. (1997) showed that HACCP could be applied at farm level to control disease on the farm. Animal health control is then seen as another feature of quality management. Infected animals or contaminated products such as milk or meat should then be seen as product quality deviations. Many bio-security measures, as part of “Good Farming Practice” (GFP), aim at disease control at herd level but are not specific to certain human-related diseases causing agents. The advantage of the application HACCP methodology over the application of GFP is that it is simple, highly farm (or situation) specific, provides self-management through a network of control points, requires low labour input and corrective actions can be identified (Noordhuizen et al 1997). 

Are the HACCP principles relevant to the Tanga milk marketing chain? 

The HACCP could also be applied to the introduction of disease agents into the herd and to the transmission of disease agents to the human population, either through the milk marketing chain or through direct contact. If HACCP is to be applied at farm level, the multiple risk pathways for introduction into the herd, intra-herd transmission and transmission to humans need to be considered and CCPs along all these pathways must be identified. The first three steps in applying the HACCP concept, starting with the construction of a flow diagram, the hazard analysis and the identification of CCPs, might be easy to apply under Tanga (and the general Tanzanian) circumstances, both on farm and along the marketing chain. They will help to develop a holistic approach to the control of milk-borne disease, inclusive of zoonoses.  The first three steps alone without active ‘Good Farming Practice’ (GFP), ‘Good Hygienic Practices’ (GHP), ‘Good Animal Health Service Practices’ (GAHSP) and other supportive instruments (efficient diagnostic and testing infrastructures, policies and legislation) including establishing critical limits, corrective actions, verification procedures, monitoring and documentation may not, however, ensure the proper application of controls. The latter will assist in control of the safety and quality of animal products and demonstrate this control to consumers. The advantage of applying HACCP is that it is flexible and that it can be tailor-made for a particular farm or system. The CCPs will not be the same in all situations, as disease prevalence differs between farms, as do risk factors for zoonoses. Separate CCPs might have to be identified for the different routes along the multiple pathways but the number of CCPs should be kept to a minimum in order to keep the system manageable (Noordhuizen et al 1997). 

The CCPs to control high coliform counts in milk, for example, will reduce the risk of consumers being infected with milk-borne zoonoses due to faecal contamination. The high coliform counts found in milk marketed in Tanga (Swai and Schoonman 2011) indicate that CCPs should start on the farm and be applied all along the chain to guarantee safety. This is especially useful for pathogens causing zoonoses which are characterized by latent (i.e. clinically inapparent brucellosis) herd infections that do not result in visible tissue damage. Some of these inapparent infections are very important in developing countries where a large number of people have a compromised immune system due to HIV infection. 

The HACCP applied along the milk marketing chain should identify CCPs all along the chain. The introduction and transmission of zoonoses such as brucellosis and tuberculosis need to be controlled at herd level. Proper hygiene during milking, proper cleaning of utensils and the use of metal cans for storage and transport can reduce contamination and thus the risk of zoonoses such as VTEC. Proper handling, cooling and pasteurization, especially of bulk milk along the chain can further control the risk of zoonoses. None of these controls, except pasteurization, will eliminate 100% of the risk but the effect of multiple CCPs in series can reduce risk to an acceptable level. Even if pasteurisation is already practised somewhere along the chain, the other CCPs are still important as the system might break down. A CCP should be feasible to apply and appropriate to the production system. It might not be feasible, for instance, for a traditional herd to stop communal grazing, or adopt zero-grazing as part of a CCP. 

Controlling the hazard of human zoonosis infection may be controlled by applying CCPs at different stages, causing a progressive reduction of pathogens at vulnerable points along the marketing chain. Table 1 shows an example of a HACCP application for tuberculosis. 

In order for HACCP to work, education is needed at all levels. Livestock keepers need to be educated on the risk of zoonoses in general. If information on a particular zoonosis is not available then they should at least be informed of the risk involved in certain activities and husbandry and management practices. Workers along the marketing chain must be educated, not only to reduce their own risk but also to reduce the risk to consumers. At the end of the chain consumers need to be educated on the risks involved in eating animal products that might not be safe for consumption. Involvement at all levels -- livestock keepers, people along the marketing chain and consumers -- is important to reduce the risk of zoonoses. This requires decentralization of the control capacity and information to be available along the marketing chain but this might be easier to achieve in formal as compared to informal marketing systems. In Tanzania a considerable proportion of marketed animal products reaches consumers through informal channels where risk control information may not be so easily disseminated, understood or complied with. In this situation, advocating the control of zoonoses at consumer level still remains a very important practice.


Table 1: Tuberculosis along the milk marketing chain

Hazard analysis

Tuberculosis is endemic in Tanga. Prevalence is higher in smallholder herds (around 6%) compared to traditional herds (<0.2%) (Swai and Schoonman 2012). It is not known if tuberculosis prevalence is the same in other areas of Tanga region which deliver milk to the Tanga market. Humans can be infected by ingestion of raw milk products but pasteurized products are normally safe. Human infections caused by M. tuberculosis and M. bovis are clinically and radiologically indistinguishable. The proportion of human tuberculosis cases caused by M. bovis in Tanzania is not known. Mfinanga et al (2004) found M. bovis in 10% of human mycobacterial adenitis cases in Arusha. Tuberculosis, including M. bovis, is a serious risk in Tanzania with the high number of immuno-compromised individuals due to the high HIV prevalence (overall 7%) (URT 2012).

Assessment: Although the infection level is probably low, tuberculosis is present in the Tanga region, especially in the smallholder system which produces most of the marketed milk. The disease in man is serious and, considering the high prevalence of HIV, tuberculosis should be controlled along the milk marketing chain.


Hazard Pathway

Risk factors

Possible CCPs

Applicability/ Remarks

Transmission of the zoonosis to the human population

Food borne:

  • on farm
  • public

  consuming raw milk

  immuno compromised individuals

   Process milk before consumption (boiling)

Only sufficient processing will eliminate the hazard

Challenges associated with the use HACCP principles 

Information along the food chain, such as the origin of livestock products (traceability), the disease status of production units and on the processing steps before consumption, as well as, consumer consumption preferences, are important. This type of information is often incomplete or lacking in many developing countries including Tanzania. Secondly, important information is often lacking or insufficient as there are no or few epidemiological data available to identify zoonosis hazards, infected herds cannot be diagnosed and products cannot be traced to these herds. In such an uncertain situation, where it is not clear if an animal or herd is infected or if a product is contaminated, CCPs should focus on the potential zoonoses hazards inherent to certain activities or practices. The CCPs which can control more than one zoonoses simultaneously will be more important in this case. Public health promotion on education and inter-disciplinary one-health collaboration between animal health service providers, public health practitioners and policy makers should also result in a more efficient and effective joint approach to the operationalization of HACCP programmes in Tanzania. 

Limitation on the wider use of HACCP principles 

The application of HACCP in developing countries is, however, a major challenge and will take time. First, a major part of the milk market is informal. Second, it requires the involvement of actors, and compliance with CCPs, at all levels. Education is needed for livestock keepers, workers along the marketing chain and consumers on the risk of milk-borne zoonoses in general and the importance of CCPs. People might be more committed to comply with CCPs if the hazard of contracting zoonoses becomes more visible. Improved diagnostics and testing for these milk-borne zoonoses in health facilities can help in this regard. Showing the burden of milk-borne zoonoses by estimating Disability-Adjusted Life Years (DALYs) might be an important method of demonstrating the importance of zoonoses in Tanzania, which  are currently a under-reported and neglected(Kunda et al 2007).



The authors are very grateful to the livestock keepers, animal health practitioners, milk marketing agents who gave their time to this research. This paper is published with the permission from the Director of Veterinary Services in Tanzania.


Chambers J V 2002 The microbiology of raw milk. In: Dairy Microbiology Handbook, 3rd edn (ed. R.K. Robinson), John Wiley & Sons, New York.  Pp. 39-90 

FAO (Food Agriucture Organisation) 1997 Codex Alimentarius. Food Hygiene Basic Text, FAO, Rome, pp 58. 

ILRI (International Livestock Research Institute) 2005 Bridging the regulatory gap for small-scale milk traders. ILRI Briefing Paper. Accessed on 15/11/ 2012 from  

Kivaria F M, Noordhuizen J P T M and Kapaga A M 2006 Evaluation of the Hygienic Quality and Associated Public Health Hazards of Raw Milk Marketed by Smallholder Dairy Producers in the Dar es Salaam Region, Tanzania, Tropical Animal Health and Production 38: 185-194. 

Kunda J, Fitzpatrick J, Kazwala R, French N P, Shirima G, Macmillan A, Kambarage D, Bronsvoort M and Cleaveland S 2007 Health-seeking behaviour of human brucellosis cases in rural Tanzania, BMC Public Health 3: 315. 

Mdegela R H, Kusiluka L J, Kapaga A M, Karimuribo E D, Turuka F M, Bundala A, Kivaria F, Kabula B, Manjurano A, Loken T and Kambarage 2004 Prevalence and determinants of mastitis and milk- borne zoonoses in smallholder dairy farming sector in Kibaha and Morogoro districts in eastern Tanzania, Zoonoses and Public health 51(3): 123-128 

Mfinanga S G M, Morkve O, Kazwala R R, Cleaveland S, Sharp M J,Shirima G and Nilsen R 2003 The role of livestock keeping in tuberculosis trends in Arusha, Tanzania, International Journal of Tuberculosis and Lung Disease 7(7): 695-704.  

Mwangi A, Arimi S M, Mbuguna S, Kangethe E K, Ouma E A, Omore A O, McDermott J J and Staal S J 2000 Application of HACCP to improve the safety of informally marketed raw milk in Kenya, Proceedings of the 9th symposium of the International Society for Veterinary Epidemiology and Economics, Colorado. 

Noordhuizen J P T M, Frankena K, Welpelo H J 1997 Part II: Applying HACCP principles to animal health care at farm level. Food borne diseases. Paper presented at the 49th seminar of the European Association of Agricultural Economists. 

Schooman L, Swai E S and Daborn C J 2011: Differences in socio-economic demography, herd parameters and hygienic management practices amongst livestock keepers in two contrasting production systems of Tanga region, Tanzania. Livestock Research for Rural Development. Volume 23, Article #237. Retrieved November 16, 2012, from

Schooman L and Swai E S 2011: Marketing, handling and physical quality of raw marketed milk in Tanga region of Tanzania. Livestock Research for Rural Development. Volume 23, Article #191. Retrieved November 16, 2012, from 

Shirima G M, Fitzpatrick J, Cleaveland S, Kambarage D M, Kazwala R R, Kunda J and French N P 2003 Participatory Survey on Zoonoses Affecting Livestock Keeping Communities in Tanzania, Journal of Animal and Veterinary Advances 4: 253–258.  

Staal S J, Delgado C, Baltenweck I and Krusha R 2000 Spatial aspects of producer milk price formation in Kenya: A joint household-GIS approach. A contributed paper at the International Association of Agricultural Economists meeting, Berlin, Germany, August 2000. Smallholder Dairy (Research and Development) Project Research Report. 

Swai E S, Karimuribo E D, Schoonman L, French N P, Fitzpatrick J, Kambarage D and Bryant M J  2005: Description, socio-economic characteristics, disease managements and mortality dynamics in smallholder's dairy production system in coastal humid region of Tanga, Tanzania. Livestock Research for Rural Development. Volume 17, Article #41. Retrieved November 16, 2012, from

Swai E S and Schoonman L 2011 Microbial quality and associated health risks of raw milk marketed in the Tanga region of Tanzania, Asia Pacific Journal of Tropical Biomedicine 1(3): 217-222. 

Swai E S and Schoonman L 2012 Differences in prevalence of tuberculosis in indigenous and crossbred cattle under extensive and intensive management systems in Tanga Region of Tanzania, Tropical Animal Health and Production 44:459–465 

URT (United Republic of Tanzania) 2010. Ministry of Livestock and Fisheries Development, Budget Speech, 2010/211. Government Printer, Dar-es-Salaam, pp 14-58 

URT (United Republic of Tanzania) 2012 The Prime Minister's Office Tanzania Commission for AIDS, Facts and Statistics, downloaded on 22/11/2012 and accessed from 

Received 21 December 2012; Accepted 14 January 2013; Published 5 February 2013

Go to top