Palliative Care

EHR predictors modestly discriminate mortality in nursing home residents with dementia

In 922 Australian nursing home residents with moderate–severe dementia, five EHR-derived predictors were retained: age, male sex, oral intake assistance, baseline weight, and 6-month weight change. The model had modest discrimination (apparent c-statistic 0.64, optimism-corrected 0.61) and acceptable calibration (optimism-corrected slope 0.82). Authors propose these routinely collected variables could support anticipatory care planning if externally validated. Apply findings only to similar nursing home populations and avoid overreliance on a single model without local validation. 1

PaCT-ILD: a co-designed one-page referral tool for fibrosing ILD

Using international surveys, interviews, and co-design workshops, investigators developed PaCT-ILD, a one-page clinician tool with five domains and three referral thresholds. Prioritized triggers include distressing symptoms, worsening quality of life, psychosocial distress, patient or caregiver request, and recurrent hospitalizations. The tool documents reasons for non-referral and prompts revisiting palliative care at future visits. Effectiveness, feasibility, and acceptability remain untested and should be evaluated before routine implementation. 2

Families report high involvement and satisfaction after precision-medicine treatment recommendations

Among families in the PRISM pediatric precision medicine trial, most expected benefits and a recommendation at enrollment. Seventy percent of parents received a treatment recommendation, half recalled it, and parents reported high involvement (93/100) and satisfaction (95/100) in decisions. Receiving a recommendation was not associated with greater trial regret (p>0.05). These findings apply to families of children with poor-prognosis cancers in similar trial settings, not general pediatric oncology care. 3