It was found that the knowledge score was higher for physicians than for nurses and for personnel with previous training or greater professional experience. Demographic characteristics had only limited effects on the attitude and practice scores, while knowledge had a direct positive effect on attitude and indirect on practice. To our knowledge, this is the first survey evaluating the knowledge, attitudes and practices of PICU personnel regarding pediatric palliative care. This study provides new insights into the knowledge, attitudes and practices of physicians and nurses in Chinese PICUs with regard to pediatric palliative care. The results of this study may be used for designing and development of targeted education and training interventions to support PICU personnel providing pediatric palliative care.

Knowledge gaps identified after analysis of study results

In the present study, 7 of the 13 questions in the knowledge dimension were answered correctly by more than 85% of the surveyed PICU personnel, including items relating to the aims of pediatric palliative care (item K1), the importance of controlling pain/symptoms to make the patient comfortable and supporting parents/caregivers (item K4), the use of pain scales to assess pain (item K6), the importance of parents/caregivers being present during routine care (item K9), sleep deprivation as a major stressor in patients receiving palliative care (item K10), the importance of the early initiation of pediatric palliative care after diagnosis (item K11), and the benefits of pediatric palliative care on patient quality of life (item K12). In addition, more than two-thirds of the participants were aware that pediatric palliative care is provided for a wide range of non-malignant as well as malignant conditions (item K2), that patient comfort is an important aim, with unnecessary examinations and treatments avoided (item K8), and that active treatment to prolong survival should not be promoted if it compromises quality of life (item K13). However, nearly 40% of the medical personnel incorrectly believed that pediatric palliative care is only given to patients with life-threatening diseases after treatment has failed (item K3), and nearly half of the respondents incorrectly believed that opioid substitution therapy should be considered to reduce iatrogenic withdrawal syndrome regardless of the dose used, the duration of therapy or the drug utilized (item K7). Notably, the average knowledge score was 38.30 ± 3.80 points out of a possible maximum of 60 points. The above findings highlight knowledge gaps among the surveyed physicians and nurses, particularly with regard to the indications for palliative care and opioid use in children receiving palliative care.

Occupational, but not demographic factors influenced knowledge scores

The knowledge levels of the participants in the present study are broadly comparable to those reported previously by surveys of clinicians and nurses. For example, Zuniga-Villanueva et al. described a mean score of 6.8 out of 10 for pediatricians in Mexico [26]. Abuhammad et al. found that nurses in Jordan had a low score in knowledge of pediatric palliative care [27]. Zeru et al. determined that only 62.8% of surveyed nurses in Ethiopia had a good level of knowledge [28]. Similarly, Ghoshal et al. reported that less than half of the surveyed doctors and nurses in India had a good level of knowledge (defined as a knowledge score ≥ 70%) [22]. Detsyk et al. found that 25.3% of healthcare workers providing medical services to children (including nurses, general practitioners and pediatricians) did not know the meaning of pediatric palliative care, with 71.5% of the respondents believing it was mainly provided to patients with cancer, and only 54.8% of the respondents being aware that it was provided to children with incurable chronic diseases [29]. Additionally, only 59.7% of the respondents in the study of Detsyk et al. knew that palliative care should be initiated at the time of diagnosis of an incurable disease, while only 52.6% of the participants were aware that palliative care should also offer support to the relatives of seriously ill children [29]. In agreement with our data suggesting a deficiency in knowledge regarding opioid use, Stenekes et al. concluded that healthcare providers in Canada had knowledge gaps related to opioid use and the development of tolerance to opioids and sedatives [30]. Madden et al. also found variation in the level of comfort with different opioids among physicians in the USA [31].

In this study the subgroup analyses demonstrated that the knowledge score was higher for physicians than for nurses. Furthermore, a higher knowledge score was associated with previous training and working in a department where pediatric palliative care was available. Our findings are consistent with prior research concluding that a higher level of knowledge in pediatric palliative care was associated with training and greater experience in palliative care [26,27,28, 32, 33]. Occupation type has also been reported to influence knowledge level [29]. We suggest that the implementation of education and training programs, such as those described previously [22, 27, 32, 33], may help to improve the knowledge of PICU personnel in China regarding pediatric palliative care.

Specific believes formed the mostly positive attitude

The mean attitude score of 38.30 ± 3.80 points out of maximum 60 points indicates that, overall, the participants in this study had a moderately positive attitude toward pediatric palliative care. The majority of respondents (> 69%) gave “very positive” or “positive” answers to 6 of the first 7 questions in the attitude dimension (A1–A3, A5–A7), whereas the responses for item A4 varied. More than 60% of the respondents believed that pediatric palliative care was influenced by economic (item A8), personnel-related (item A9), family-related (item A10), social (item A11) and implementation-related (item A12) factors. In addition, structural equation modelling confirmed that knowledge had a direct positive effect on attitude, although influence of knowledge or attitude on practice was not significant. Previous studies have also reported a variety of factors believed to affect the implementation of effective pediatric palliative care, including family preference for life-sustaining treatment, family not ready to acknowledge an incurable condition, parent discomfort with the possibility of hastening death [11], inadequate training [12], clinician misperceptions and emotional burden, prognostic uncertainty about treatment options [13], socio-cultural factors, nature of the patient and disease, insufficient training, regulatory/political issues [14], lack of adequate funding, lack of palliative care programs, difficulty integrating palliative care into existing pediatric care at the organizational level, and lack of knowledge [15].

New education and training programs are needed to further improve practice

The average practice score was 35.48 ± 5.72 points (possible range, 9–45 points), which suggests that there was room for improvement in the practices of the PICU personnel enrolled in this study. Most respondents (> 71%) indicated that, when required, they and their team always/often took care of analgesia, sedation, environment optimization, symptom management and communication, as well as screening and interventions to prevent delirium, or mitigate against iatrogenic withdrawal syndrome. However, only around half of the participants reported to have experience using neuromuscular blockers when indicated. A previous study in Holland reported that neuromuscular blockers were administered in 16% of cases at the time of withholding/withdrawing life-sustaining treatment [34]. Less than half of the respondents in the present study stated that they or their team withdrew life-sustaining treatment when necessary, and less than two-thirds provided support to the parents/caregivers of patients receiving end-of-life care. Overall, our findings concur with those of other studies reporting suboptimal practices [22, 35, 36].

Somewhat unexpectedly, despite being associated with higher knowledge scores, previous training appeared to have little or no effect on the attitudes and practices of PICU personnel regarding pediatric palliative care; practice was influenced by knowledge mostly indirectly, via attitude. This emphasizes the need for improved education and training programs that better target suboptimal attitudes and practices. In line with previous studies in various settings, some crucial aspects for the implementation of a pediatric palliative care program could be supported by the KAP approach, in particular collecting data to characterize the need for pediatric palliative care [22, 23, 37]. Additionally, KAP assessment would help to raise awareness among hospital administration and clinical personnel about pediatric palliative care, and potentially aid in forming a balanced multidisciplinary palliative care team. Thus, above results may be used for the providing education and training to clinical personnel to support continuous provision of pediatric palliative care in PICU.

Strengths and limitations of the study

This multicenter study used KAP methodology and validated questionnaire to address the issue of the availability and barriers in implementation of pediatric palliative care; the results were additionally strengthened by application of SEM model which allowed to explore the potential influence of other factors on practice scores. Study population included a wide variety of ICU professionals, with different age and experience; still the sample size may be small to detect some specific differences between groups. Participants from five big cities in China answered the questionnaire, allowing to increase the generalizability of the results, but local peculiarities should still be taken into account. Although the KAP questionnaire was developed based on published recommendations and demonstrated good reliability, it may have limitations with regard to its ability to assess perceptions of pediatric palliative care. Finally, this study did not evaluate whether education/training programs would enhance the questionnaire scores, which is of interest for the future research.