Bovine tuberculosis (bTB) eradication, brushtail possum control, and causal inference

Abstract Bovine tuberculosis (bTB) in livestock (cattle and farmed deer) is the focus of an in‐progress eradication program in New Zealand. The program includes intensive brushtail possum population control as this species is a bTB host and could compromise eradication at a national level. Ten causal criteria (strength of association, experiment, temporality, plausible mechanism, biological gradient, consistency, coherence, specificity, analogy, validated causal predictions) were used to evaluate annual changes in bTB status in livestock herds relative to costs of possum population control and livestock bTB disease control. Hypotheses were evaluated using a model selection framework of the causal relationship between efforts (costs) and outcomes (bTB‐infected herds) following management. The 10 causal criteria were demonstrated empirically. Experiments used designs from the simplest pretest‐posttest comparison to a before‐after‐control‐impact design, though lacked random allocation of treatments to experimental units and some studies lacked statistical analyses. As hypothesised, the annual changes in bTB‐infected herds were positively related to annual costs. The predicted damped oscillations across years in annual changes in bTB‐infected herds were observed. The implications are that eradication of livestock bTB appears likely, though there is a risk that eradication may not be achieved by 2026, unless annual control expenditure is higher and/or more efficient in the future.