PT - JOURNAL ARTICLE AU - Crocker, JC AU - Beecham, EC AU - Kelly, P AU - Dinsdale, AD AU - Hemsley, J AU - Jones, L AU - Bluebond-Langner, M TI - RECRUITING PARENTS TO PAEDIATRIC PALLIATIVE CARE RESEARCH: IMPACT OF LOW INVITATION RATES ON SAMPLE BIAS AID - 10.1136/bmjspcare-2014-000653.17 DP - 2014 Mar 01 TA - BMJ Supportive & Palliative Care PG - 110--110 VI - 4 IP - 1 4099 - http://spcare.bmj.com/content/4/1/110.3.short 4100 - http://spcare.bmj.com/content/4/1/110.3.full SO - BMJ Support Palliat Care2014 Mar 01; 4 AB - Introduction Recruitment in paediatric palliative care is widely reported as challenging, with low rates of invitation by clinicians. The impact of this on sample bias is unknown. Aim(s) and method(s) We studied recruitment to a qualitative interview study about parental decision making for children cared for by a specialist palliative care (PC) team. PC clinicians were encouraged to introduce the study to parents over 12 months. With ethical approval, we used information from the PC team database and feedback from clinicians to explore the impact of low invitation rates on sample bias. Results The families of 519 living and 73 deceased patients were retrospectively identified as potentially eligible for recruitment. Clinicians invited parents of 28 (5.4%) living patients compared to 21 (28.8%) deceased patients (p=0.0001). On multivariable analyses, there was no association between patient demographics and invitation, but for living patients, total and out-of-hours contact time between family and PC team while eligible were independently associated with invitation (p<0.05). The most common reasons clinicians gave for not inviting parents of living and deceased patients were little or no contact with them and perceived burden. Conclusion(s) Invitation rates were especially low among parents of living patients. There was no evidence that this led to major demographic bias. However, the strong influence of family contact may have introduced bias potentially relevant to the project (e.g. patient stability and parent-clinician relationships) that we were unable to measure. We recommend that researchers consider levels of patient contact when planning studies requiring recruitment via clinicians.