Skip to main content

Table 4 Final multivariable Cox-regression model examining the association of FMD with clinical mastitis and culling

From: Impact of foot-and-mouth disease on mastitis and culling on a large-scale dairy farm in Kenya

Variable Clinical mastitis Culling
  HR (95%CI) P -value   HR (95% CI) P -value
FMD
Case   2.9 (0.97, 8.9) 0.057   1.7 (0.88, 3.1) 0.12
Non-case   Baseline -   Reference -
Age aAge category Age category
  1.5- < 2.0 y Baseline - −28- <227d Baseline -
  2.0- < 2.8 y 0.61 (0.12, 3.0) 0.54 227- < 577d 0.29 (0.07, 1.2) 0.090
  2.8- < 3.5 y 0.77 (0.16, 3.7) 0.75 577- < 974d 0.93 (0.32, 2.7) 0.89
  3.5- < 4.3 y 0.64 (0.12, 3.3) 0.60 974- < 1363d 0.76 (0.21, 2.7) 0.76
  4.3-9.7 y 2.6 (0.45, 15.1) 0.29 1364-3543d 1.5 (0.34, 6.3) 0.60
Lactation stage 0.45 (0.24, 0.83) 0.010
Non-lactating - -  
Early lactation (<0-100d) - -  
Mid lactation (101-250d) - -  
Late lactation (>250d) - -  
Dry - -  
Breed b 1.4 0.031   - -
Tick borne disease (last 12 months)
Yes - -   3.2 (1.8, 5.9) <0.0001
No - -   Baseline  
Parity
0 Baseline -   Baseline  
1 11.3 (2.5, 51.2) 0.002   2.8 (0.91, 8.4) 0.074
2 9.8 (1.7, 55.9) 0.010   0.66 (0.16, 2.6) 0.55
3 4.6 (0.65, 32.6) 0.13   2.4(0.60, 9.8) 0.21
≥4 8.1 (1.2, 55.8) 0.033   0.88 (0.16, 5.0) 0.89
Time varying interactions
FMD 0.43 0.016   - -
Lactation stage - -   2.0 (1.4, 2.6) <0.0001
  1. aAge categories based on quintiles.
  2. bIncluded as a linear effect. Categories are 100% exotic breed, <25% indigenous, 25% indigenous, 50% indigenous. Animals were included in the analysis if present on the farm during the outbreak period (31st August-28th September 2012) and were followed for 12 months after the commencement of the outbreak. For the analysis of clinical mastitis, cattle were included if over the age of 18 months at the start of outbreak, considered as the age when clinical mastitis becomes a possibility. Culling was defined as exiting the herd due to any disease, death of low production. Hazard ratios (HR) incorporate time varying effects with logarithmic multiplier functions to account for non-proportional hazards. Final model for mastitis included age, breed and parity with FMD incorporated as a time varying effect. For culling the final model included age, lactation stage, tick-borne disease in the 12 months preceding the outbreak and parity with lactation stage as a time varying effect.