An approach towards improving safety and hygiene control practices in emerging dairy chains : the case of Tanzania

The underdeveloped state of emerging dairy chains have been implicated as the underlying reasons for the constant quality and safety challenges. Interventions have been implemented, usually led by dairy companies, developmental agencies, and governments, covering training, technical assistance and infrastructure development. However, it is not clear if existing interventions have translated into an improvement in safety and hygiene control practices. Moreover, there is doubt whether these interventions have been specific enough to trigger the improvement needed in practices, limit the constant milk quality and safety problems, and safeguard the consumer. This thesis therefore, aimed to gain insights into the underdeveloped state of safety and hygiene control practices in emerging dairy chains and identify how to improve these practices through a behaviour-based training intervention.Using in-depth interviews, focus group discussions and systematic on-site observations, the possible causes of persistent challenges in safety and hygiene control practices in an emerging dairy chain, using&nbsp; Tanzania as an example, was investigated. We found a lack of uniform organisation of control activities and enforcement on milk quality and safety along the chain. The formal chain was mainly targeted regarding setting and enforcement of requirements on quality and safety, while this was completely absent in the informal chain. This was reflected in basic preventive and monitoring safety and hygiene control practices at the farm such that teat dipping, milking cooling, mastitis control practices and proper feed storage facilities were commonly lacking. Similarly, basic practices were observed in the operation of milk traders and local retail shops, exacerbating public health risks. However, milk collection centres and dairy processing companies seem to have improved preventive and monitoring control practices reflected in hygienically designed cooling tanks and standard tests for milk quality. We also investigated dairy farmers’ perception of two existing interventions, commercial and non-commercial, &nbsp;on the extent of support toward on-farm safety and hygiene control practices. We found that the non-commercial intervention, which combined training and provision of milk storage containers, was perceived by farmers to be more supportive of practices than the commercial intervention, which only offered training on demand. However, both interventions did not directly reflect in better safety and hygiene control practices.&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;Furthermore,&nbsp; we developed a customised tool to assess and differentiate performance levels of safety and hygiene practices crucial for the control of microbial and chemical (i.e., aflatoxin) safety of fresh milk along the chain. The principles for diagnostic tools was used to identify crucial control practices and indicators linked to microbial load and aflatoxin M1 (AFM1) in fresh milk. Poor, basic, intermediate and standard level differentiation grids were developed based on the type of equipment used, the degree of performance of actual practices, and the extent of documentation of procedures and data recording. Small and large scale dairy farms were selected to test the customised tool in a pilot study in Tanzania. Three distinct clusters were identified based on practice performance. For all the three clusters, basic level performance was observed for crucial control practices of dairy farmers of both large and small-scale farms such as milk safety monitoring method, udder and teat care, and personal hygiene.We investigated whether differences in safety and hygiene control practices translated into distinctions in milk quality and safety at the farm, and the implication for actors further in the Tanzania chain. Safety and hygiene control practices along the chain were differentiated using the customised assessment tool, concurrently with milk sampling to analyse for AFM1 at the farm and selected microbial parameters (i.e., total bacteria count, coliforms and Staphylococcus aureus). The microbial analysis of fresh milk at the farm indicated that the milk already exceeded maximum limits to the extent that no further growth was observed along the chain, irrespective of the level of practice performance, while between poor and basic level performances, there was no clear difference in milk quality and safety. High levels of AFM1 were observed among farmers who stored feed longer with limited measures to control mould growth.Based on the previous findings, we &nbsp;developed a training intervention based on the theory of planned behaviour (TPB) and analysed its effectiveness to influence behaviour drivers of dairy farmers towards safety and hygiene control practices. The training program targeted personal hygiene, udder and teat care, shed and floor sanitation, and milk cooling and storage practices. A three-day training scheme was designed involving three teaching and learning methods: teaching with slides presentation and group discussions, videos, pictures and story analysis, and practical demonstration to cover the four practices for each day. The training program in Tanzania, which covered 107 dairy farmers, demonstrated that dairy farmers significantly progressed in knowledge above pre-training levels significantly. The improved knowledge triggered better attitudes, improved perception of the social pressure that others have on them to perform the practices, better awareness of the control they possessed, and positive intentions to perform on-farm safety and hygiene control practices. A multiple linear regression analysis was performed to test the predictive influence of the components of the TPB on farmers’ intention to perform the practices. The results showed that the subjective norm was the most consistent predictor of the intention to perform personal hygiene (β=0.30, p<0.002), udder and teat care (β=0.36, p<0.001), and shed and floor sanitation practices (β=0.34, p