BACKGROUND: Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) disease (COVID-19) has high morbidity and mortality in older adults and people with dementia. Infection control and prevention measures potentially reduce transmission within hospitals. AIMS: We aimed to replicate our earlier study of London mental health in-patients to examine changes in clinical guidance and practice and associated COVID-19 prevalence and outcomes between COVID-19 waves 1 and 2 (1 March to 30 April 2020 and 14 December 2020 to 15 February 2021). METHOD: We collected the 2 month period prevalence of wave 2 of COVID-19 in older (≥65 years) in-patients and those with dementia, as well as patients' characteristics, management and outcomes, including vaccinations. We compared these results with those of our wave 1 study. RESULTS: Sites reported that routine testing and personal protective equipment were available, and routine patient isolation on admission occurred throughout wave 2. COVID-19 infection occurred in 91/358 (25%; 95% CI 21-30%) v. 131/344, (38%; 95% CI 33-43%) P < 0.001 in wave 1. Hospitals identified more asymptomatic carriers (26/91; 29% v. 16/130; 12%) and fewer deaths (12/91; 13% v. 19/131; 15%; odds ratio = 0.92; 0.37-1.81) compared with wave 1. The patient vaccination uptake rate was 49/58 (85%). CONCLUSIONS: Patients in psychiatric in-patient settings, mostly admitted without known SARS-CoV-2 infection, had a high risk of infection compared with people in the community but lower than that during wave 1. Availability of infection control measures in line with a policy of parity of esteem between mental and physical health appears to have lowered within-hospital COVID-19 infections and deaths. Cautious management of vulnerable patient groups including mental health patients may reduce the future impact of COVID-19.