Farrowing crates continue to be the most common system of housing farrowing and lactating sows, however continuous confinement is one of the most contentious welfare issues in livestock production. This review aims to critically evaluate the scientific literature on the implications of two important determinants of sow and piglet welfare: housing and human contact. While the evidence in sows is contradictory, there is consistent evidence of a greater short-term cortisol response in gilts introduced to farrowing crates than pens and there is limited evidence of changes in the HPA axis consistent with sustained stress in gilts during the fourth week of lactation. Confinement of sows during lactation increases stereotypic behaviour, reduces sow-piglet interactions, nursing duration and lying behaviour, and may contribute to leg and shoulder injuries in sows. Piglets reared by sows in pens display more play and less oral manipulative behaviours and generally have better growth rates than those reared in farrowing crates. However, there is increasing interest in developing alternate housing systems with minimal sow confinement and reduced piglet mortality risk. Recent research on temporary crating of sows suggests that confinement briefly around farrowing may be the best compromise between continuous housing in farrowing crates and pens, as it may reduce live-born piglet mortality, while providing the opportunity for the sow to move more freely prior to parturition. Together with housing, the behaviour of stockpeople is a key determinant of pig welfare. Recent evidence shows that positive handling of piglets during lactation reduces fear behaviour and physiological stress responses of pigs to humans, novelty and routine husbandry practices, and thus may be enriching for piglets. As a source of enrichment, positive human interactions provide several advantages: close interactions with piglets usually occur several times daily, positive interactions can be combined with routine checks, human interactions invariably provide variability in their predictability which will minimise habituation, and positive interactions may not require additional physical resources. There is a clear need for comprehensive research examining both the short- and long-term welfare implications and the practicality of less confinement of the sow, and positive handling of both sows and piglets during lactation.