Phase 1: needs of parents
In order to develop a practical and useful decision aid on CAM use that is likely to be adopted by families affected by cancer, the project started with investigation of the needs of the parents.
Among those surveyed, 70 parents responded. Mothers made up the majority of respondents (93%). Among all respondents, most had a child with leukemia (55%) or a brain tumor (17%). Among the cancer-affected children of respondents, 31% were currently in treatment and 38% had finished treatment more than 1 year ago. More than half (56%) reported CAM use for their child. Among those who reported CAM, the use of food supplements/vitamins (32%), massage (22%), and homeopathy (22%) were most mentioned. One-third of those using CAM for their child had searched for information about the CAM treatment before use. Sources included the Internet (37%), family/friends (37%), and CAM practitioners (26%). Eighty-eight percent of respondents thought it was important to receive/find good quality information on CAM. Most respondents (79%) had a need for information on CAM following their child’s cancer diagnosis, including 38% just after diagnosis, 45% during the first year after cancer treatment, 34% more than a year following cancer treatment, and 14% in the palliative phase. Parents sought information on CAM for treatment of fatigue (62%), anxiety (47%), pain (46%), weakened immune system (44%), sleeping problems (42%), nausea/vomiting (35%), low mood (29%), decreased appetite (27%), intestinal problems (22%), concentration problems (16%), and weight loss (11%). Parents mentioned a need for a decision aid to support their discussions of CAM treatments with their physician (60%), to find reliable CAM practitioners (59%), for education regarding possible CAM use (45%), for support in asking questions of CAM experts (43%), and as a resource for reading about other parents’ experiences with CAM (40%). Preferences for the form of the decision aid were website (62%), app (23%), or informational brochure/leaflet (12%). Preference of most respondents was that the decision aid be accessible via the treating pediatric oncologist (88%) or VOKK (83%).
Direct content analysis was performed on five themes: (1) Search strategies: Parents usually start searching for CAM options after their child has been receiving treatment for some time as closer to time of diagnosis they have too much information to process. Parents report searching the Internet for CAM modalities effective for specific complaints or cancer types, for finding reliable CAM practitioners, and for information on whether particular CAM modalities can be combined with cancer treatment. (2) Decision-making: Decisions on the use of CAM are mostly based on the positive experiences of family/friends, on parent’s personal experiences with CAM, or on the availability of evidence for a particular CAM modality. (3) Format of the decision aid: A website with a search function for complaint, type of cancer, and CAM modality, along with a chat function allowing interaction with other parents and a CAM expert who can answer CAM-related questions. (4) Function of the decision aid: To allow the parents make their own decisions with regard to CAM, and also as a tool to discuss the outcome on CAM with the pediatric oncologist. (5) Content of the decision aid: Description of positive effects of particular CAM modalities, evidence for their use, potential side effects, and possible interference with cancer treatment.
Phase 2: evidence on CAM
Based on the results of phase 1, the project team needed to make a decision on the focus of the decision aid. Since there are more than 1800 CAM modalities  and performing systematic reviews to collect evidence-based information on all these modalities is time-consuming, it was not feasible to develop a decision aid on all childhood cancer-related complaints for which parents in phase 1 wanted information about. It was decided to focus on CAM use for pain during cancer treatment and post-treatment because (1) pain was in the top three complaints that parents wanted information about in phase 1 and (2) the Dutch Childhood Oncology Group (DCOG; www.skion.nl) was developing clinical guidelines for treatment of pain, so a joint venture was planned.
Systematic review on CAM for pain
Figure 2 shows the study selection process. A total of 6936 records were identified from searches, of which11 RCTs met the criteria for GRADE analysis [25,26,27,28,29,30,31,32,33,34,35]. An additional 18 articles served as input for phase 3 (see online resource 5). The characteristics of the included RCTs are summarized in Table 1. RCTs were found for five CAM modalities: hypnotherapy (N = 5), massage (N = 3), healing touch (N = 1), music therapy (N = 1), and mind-body intervention (N = 1). Ten out of 11 studies investigated the changes in pain related to procedures such as venapuncture, lumbar puncture, or bone marrow aspiration. The 11 studies included children with all types of cancer. The quality of studies was moderate to high for the RCTs on hypnotherapy and low to moderate for the other CAM modalities (Table 1). Since the studies differed with respect to pain scales, control groups, and number of RCTs per CAM modality, pooled effects on the primary outcome could only be analyzed for hypnotherapy and massage (Table 2). There was high-quality evidence that hypnotherapy significantly reduces cancer-related procedural pain in children; the pooled effect was statistically significant (MD, − 1.37; 95% CI, − 1.60, − 1.15; P < 0.00001) compared with standard care [27,28,29], and statistically significant (MD, − 1.13; 95% CI, − 1.34, − 0.94; P < 0.00001) compared with an attention-control group [26,27,28,29] (Table 2). Both analyses showed significant heterogeneities (P < 0.001), with I2 values of 87% in comparison with standard care and 86% in comparison with attention-control. One high-quality study on hypnotherapy by Smith et al.  was not included in the meta-analysis because it used a different VAS scale and outcome data were only reported for children with high hypnotizability and children with low hypnotizability. There was low-quality evidence for no effect of massage (Swedish/acupressure and effleurage/petrissage techniques) on cancer-related procedural pain compared with standard care; the pooled effect of two studies was not significantly different (MD, − 0.77; 95% CI, − 1.82, 0.28; P = 0.15, heterogeneity I2 = 0%, P = 0.80) (Table 2). No evidence was found that the five CAM modalities were unsafe to treat pain in children with cancer.
Since it is more challenging to conduct trials in children than in adults , especially with regard to cancer, the number of pediatric patients in studies is still limited . Therefore, in order to collect more information on the appropriate use of specific CAM modalities for pain management in children with cancer, it was decided to investigate the clinical experiences of pediatric oncologists with expertise in CAM. Experts had positive clinical experience with all five CAM modalities that were found in the systematic review (see online resource 6).
Phase 3: development decision aid
Data of the systematic review and the clinical experience from pediatric oncologist, as obtained in phase 2, formed the main input for the development of the decision aid in phase 3. A crucial step in phase 3 was to reach consensus among stakeholders with respect to recommendations on CAM use. Five CAM modalities (hypnotherapy, mind-body techniques, massage, healing touch, and music therapy) were included in the decision aid based on the results from the systematic review. The clinical expert opinions of pediatric oncologists in Integrative Oncology as well as patients’ needs and preferences were considered important supplementary sources of information for these five CAM modalities. Recommendations were given for each CAM modality (see Table 3). The project team processed the information from the phase 2 literature analysis and expert meeting into a concept structure for a website-based decision aid, and the content was written by three text writers.
Phase 4: evaluation and implementation of decision aid
The content and structure of the decision aid as developed in phase 3 was evaluated by the VOKK, parents, and pediatric oncologists. Feedback included comments on the inclusion of too much scientific terminology that some information was too general in nature, to improve clarity with regard to which CAM modalities are supported by evidence and with regard to the information on which the content was based. According to the preference of parents (see “Phase 1: needs of parents” under the “Results” section), the decision aid was linked to the VOKK, the Dutch organization for parents of children with cancer: http://www.complementairezorg-vokk.nl/.
The overall structure of the decision aid is depicted in Fig. 3. To facilitate its use, the VOKK announced the launch of the decision aid via the newsletter, website, and Facebook page and via an article in their printed news magazine. A “business” card of the decision aid was developed for the Princess Máxima Center for pediatric oncology, the recently established high-complex care center for childhood cancer in the Netherlands , to make parents aware of the existence of the decision aid. Awareness among healthcare professionals was created via DCOG, and the decision aid was introduced in education programs at the Princess Máxima Center and at the University Medical Center in Utrecht. The Louis Bolk Institute assumed ownership and responsibility for maintenance and updating of the decision aid.