Dairy production in Uganda has recently shown steady growth. Development and intensification of dairy production may bring about the issue of negative energy balance (NEB) followed by ketosis in cattle. However, the current dairy herd health and management status in the southwestern region of Uganda has not been reported. The objective of this study was to identify current herd management status in this area by describing herd management practices and production status, and by investigating relationships between feeding management practices, nutritional status, and daily milk yield of dairy cows.

Thirty farms participated in this study. Herd attributes, management practices, nutritional and production status of the cows were collected by interviews and inspections from October 2016 to March 2017. In order to estimate the total effects of feeding management on blood β-hydroxybutyrate (BHB) and milk yield, a causal diagram was created. Multivariable analyses were performed using linear mixed-effects models, setting BHB of cows within 21 days after calving and milk yield as response variables, feeding management factors as exposure variables, potential confounders as covariates, and herd as a random effect variable.

The mean herd size of adult cows on participating farms (n = 30) was 35.5 and average milk yield 9.8L/cow/day. The proportion of exotic breeds was 74.5% of 506 adult cows. Supplementary concentrates and fodder were used in 40% and 70% of farms, respectively; grazing was conducted at 93.3%. The prevalence of ketosis including subclinical ketosis (SCK) for cows within 21 days after calving was 10.8% (4/37, 95% confidence interval (CI): 3.0-25.4%). From the multivariable models estimating the total effect, cows fed concentrates had higher milk yield (9.20L/cow/day) than cows not (5.95 L/cow/day, ratio between groups: 1.55, 95% CI: 1.02-2.34, p＝0.041). Cows in the farm where rotational grazing was conducted had higher milk yield (5.78L) than those in the farms where rotational grazing was not (3.46L, ratio between groups: 1.67, 95% CI: 1.11-2.51, p＝0.017). No significant effect of feeding management on BHB was estimated (p＝0.092).

Exotic dairy cattle breeds are dominant on dairy farms in southwestern Uganda. This study revealed that exotic and cross-breed cows have the potential to produce higher milk yields, given sufficient nutrition. SCK cows were observed in this study area. Cows with high ketone concentrations require special precautions. In order to increase milk yield without nutritional disorders in Uganda where the dairy industry is fast developing, introduction of adequate feeding management is important.

The long time needed for the elimination of Salmonella in dairy cattle can cause significant production losses. Ruminal fermentation (RF) disorders in Salmonella-infected dairy cattle may complicate the process of Salmonella eradication. Thus, this study aims to statistically evaluate any association between negative conversion rate (NCR) of Salmonella in dairy cattle feces and RF conditions using a Hidden Markov Model. Epidemiological data were obtained from 21 dairy cattle belonging to a Salmonella-infected dairy herd. Fecal samples were collected and cultured 23 times every two weeks for approximately one year. RF conditions were evaluated based on the milk fat to protein ratio (FPR) and categorized as optimum condition (OC; FPR: 1.1—1.5), negative energy balance (NEB; FPR >1.5) or rumen acidosis (RA; FPR <1.1). The odds ratios for the NCR in the cattle with NEB and RA were 0.0141 (95% credible interval: 1.08×10^－26, 2.75) and 0.0429 (95% credible interval: 2.00×10^－14, 0.628), respectively, compared with that in the cattle with OC. These results suggest that it is difficult to eliminate Salmonella in dairy cattle with RF disorders.

On 5th January 2020, World Health Organization notified all member states about the new outbreak of pneumonia of unknown cause in Wuhan City, Hubei Province of China. It was a beginning of pandemic of the novel transmissible disease, which was later named as COVID-19. The causative agent was soon searched using whole genome analysis and a new coronavirus that is very similar to the virus causing SARS was identified. The virus was named SARS-CoV-2. COVID-19 rapidly spread inside China and the Government of China decided to shut all public transportation at the end of January. In Japan, a man who travelled back from Wuhan city in early January was the first case. To rescue Japanese nationals trapped in Wuhan, the Japanese Government chartered five flights. All passengers were put under quarantine after arrival and were subjected to PCR testings. On 1^st February, a passenger on a cruise ship ‘Diamond Princess’ that had departed from Yokohama was confirmed to be infected with SARS-CoV-2 at Hong Kong. Therefore, all passengers and crews on the ship were put under quarantine offshore Yokohama, the final destination of the cruise. PCR testings were performed on symptomatic persons and persons that had close contact with cases. The number of cases increased day by day from the start of the quarantine. Epidemiological research was conducted worldwide from an early stage and demonstrate good indices to understand epidemiology of COVID-19. Case reports from China showed presence of many asymptomatic cases especially in younger generations. Mathematical modelling studies estimated the length of incubation period of COVID-19 and its serial interval and the possibility of transmission during asymptomatic period was also highlighted to give warning. Effectiveness of the quarantine measures on the Diamond Princess was evaluated in another study and drastic reduction of the predicted number of future cases was proven. At the end of February, a specialist advisory group was organized under the Ministry of Health, Labor and Welfare. Advised by this group, Japan started to tackle COVID-19 by applying a unique strategy in the world.

Regression modelling is a popular approach for analysing epidemiological data. While its analytical techniques are widely covered by standard textbooks, modelling approaches vary between disciplines, potentially confusing many users. Identifying the motivation behind epidemiological analyses, particularly understanding of causal inference, would naturally guide users. Nevertheless, this subject has not been taught systematically in many of the veterinary epidemiology curriculums. This series of articles aim to showcase the minimum standard for regression modelling, highlighting basic concepts such as confounding. In this first article, a regression modelling approach to making causal inference is explained.