Balancing the rota: registered nurses’ preferences for shift patterns – a discrete choice experiment

OBJECTIVE: To elicit and quantify registered nurses' (RNs') preferences for key shift attributes and to estimate the trade-offs between shift pattern characteristics and pay. DESIGN: Cross-sectional online discrete choice experiment with d-efficient design; 12 choice tasks per respondent. Preferences were estimated using multinomial logit models with clustered standard errors; willingness to accept (WTA) values were derived as ratios of attribute coefficients to the pay coefficient. SETTING: United Kingdom. PARTICIPANTS: 1449 registered nurses (RNs) recruited via the Nursing and Midwifery Council distribution list. PRIMARY AND SECONDARY OUTCOME MEASURES: Utility coefficients (β) for shift length, schedule predictability, shift request flexibility, weekend working frequency, pay level and night-shift frequency; WTA derived from the pay attribute; subgroup and interaction analyses by age, pay band, current shift status, caring responsibilities, sector and satisfaction. RESULTS: All six attributes were statistically significant predictors of shift pattern choices. The strongest aversions were to night-only working (β=-1.113) and having no whole weekends off (β=-0.790), which required an estimated 29% additional compensation in the primary model and 21% additional compensation, respectively. Unpredictable rotas (β=-0.220), restricted shift-choice flexibility (β=-0.189 for only swaps) and absence of pay increase (β=-0.525) also reduced utility. Nine-to-ten-hour shifts were preferred over 8-hour shifts (β=0.144), whereas 12.5-hour shifts were disliked (β=-0.179). Subgroup analyses showed significant heterogeneity by age and current shift status, with more limited evidence of pay-band differences. Interaction analyses suggested broadly consistent aversion to night work and weekend loss across the nursing workforce. CONCLUSIONS: Nurses strongly prefer rota designs that avoid exclusive night work and protect whole weekends off. Predictability and meaningful scope to request shifts also matter. These preference estimates provide policy-relevant benchmarks for rota design and unsocial-hours compensation.