Background Most indications and doses of medicines used in paediatric palliative medicine are derived from anecdotal evidence and involve scaled down adult doses.

Aims To review current prescribing practices and identify priority areas for research and prescribing guidance.

Methods Questionnaire survey of Association of Paediatric Palliative Medicine (APPM) members and non-medical prescribers working in paediatric palliative care.

Results Twenty-seven medical and nine non-medical prescribers working in hospice (49%), hospital (57%) and community (66%) completed the survey. Priorities for further systematic review evidence and consensus guidance to support prescribing were identified as (% reporting as important or priority): 1 Equivalent starting doses of strong opioids administered PO, SC or IV in babies and young children (89%) 2 ketamine (81%) 3 buccal or intranasal fentanyl (77%) 4 methadone (74%) 5 dexamethasone doses for different indications (71%) Experience of prescribing these options ranged from 63% prescribing strong opioids in babies and young children several times a year to 47% never having prescribed methadone. Respondents reported prescribing ketamine regularly several times a year (31%) or not at all (28%). Respondents reported never prescribing buccal or intranasal fentanyl (39%) or prescribing regularly several times a year (33%). 53% prescribed dexamethasone regularly several times a year. 61% prescribers would start wherever a dose range (eg, 200–300 mg/kg/dose) seemed most appropriate, the remainder at the lowest dose. When calculating doses 71% respondents would round up or down to the nearest doses that is practical to administer as long as the actual dose administered falls within the dose range given but 14% would always round down.

Conclusions Respondents identified clear priorities for further systematic review evidence and consensus guidance to support prescribing. Prescribing practices appear to vary considerably from individual to individual: this requires further exploration. Doses given as a range can be ambiguous.