Why family medicine requires a paradigm shift from linear guidelines to complexity science

Evidence-based medicine (EBM) has standardised care by reducing variance, yet its reliance on linear, reductionist models often fails the patient in primary care, especially the multimorbid patient. Current clinical guidelines assume that biological inputs lead to predictable outputs, effectively treating the human body as a machine. However, the patient, especially the multimorbid patient, functions as a complex adaptive system, characterised by emergent behaviours where small perturbations can lead to disproportionate outcomes. This paper argues that the 'straight-line thinking' of current EBM is mathematically insufficient for general practice. Through clinical vignettes, we demonstrate how single-disease guidelines create unsustainable treatment burdens. Furthermore, we present a rationale for how non-linear analyses of physiological signals can predict frailty years before clinical diagnosis. We propose that family medicine must integrate soft computing methodologies-such as fuzzy logic, fractal analysis and causal inference-not as technological novelties but as the theoretical necessity required to map the non-linear topology of human health. This perspective reframes family medicine not as an exception to evidence-based practice but as the clinical setting where complexity-aware evidence is most critically needed.