We thank Cheung et al. for their interest in our manuscript describing the frequency, nature, predictors and associated outcomes of family presence for 26,886 patient admissions to 15 medical-surgical intensive care units (ICUs) in Alberta, Canada [1]. We agree with their thoughtful assessment that our study raises questions about the optimal frequency and nature of family presence [2]. Using mixed-effects negative binomial regression models, we observed that family presence episodes were strongly associated with patient demographic characteristics, and daily time-dependent measures of both severity of illness and administration of life-sustaining technologies. Our data also show that families are more likely to be present on days when patients have delirium and less likely to be present once the delirium resolves. As Cheung et al. highlight, this intuitively makes sense as families are present when the patient is sickest and treatment decisions and support are needed. While open visitation policies provide families with important freedom, work is needed to understand how to optimize presence for patient and family experiences and outcomes [3]. Consideration is also needed as to how family presence continues after ICU discharge when patients might be in greater need of support [4]. Is there a risk that families exhaust themselves with extensive presence while the patient is critically ill to only then be less available when the patient leaves the ICU?

Perhaps, the most surprising finding from our study is that almost one in ten critically ill patients admitted to the ICU had no documented family presence during their ICU stay. While our focus on documented family presence almost certainly underestimates actual family presence, the data suggest that a small number of patients had no family presence during their stay. What does it tell us when no one calls or visits when a patient is critically ill? Perhaps, family presence in the ICU is a surrogate measure for social connection. While the primary focus of critical care is to preserve life and limb, critical illness may bring to light long standing patterns of social relationships, identify those who are socially isolated and potentially provide opportunities for intervention and secondary prevention.

How do we optimize family presence in the ICU? A starting point would be to standardize how family presence is recorded so that ICUs can reliably track the incidence, nature and consequences. Incorporating minimalist time stamped tick box charting of family presence (who, how, what) during routine bedside assessments into electronic health records would limit the work involved and provide a minimal data set for basic quality indicators (Fig. 1). For example, during the coronavirus disease 2019 (COVID-19) pandemic, many hospitals implemented restrictive visitation policies [5]. A system that reliably and prospectively tracked family presence would allow the impacts of such policy changes to be assessed. If we truly believe that families are important members of the care team, measuring their presence in the ICU would be a step towards better understanding how we currently engage with families and how we might better engage in the future.

Fig. 1
figure 1

Minimum family presence dataset variables to develop quality indicators