Objective To evaluate the extent to which parents are satisfied with and understand the information they are given when their consent is sought for their child to participate in a phase III randomised clinical trial and the reasons for their decision.
Patients and method The authors carried out a prospective study. The authors included all parents whose consent was sought for their child to participate in the FRALLE 2000A protocol (acute lymphoblastic leukaemia) at two centres. The parents were questioned twice by a qualified psychologist using a semidirected interview, 1 and 6 months after consent was sought.
Results 43 first interviews were carried out. All the parents declared they were satisfied with the explanations provided by the physician. 35 (81%) parents felt that the information provided with the request for consent was appropriate. Eight (19%) parents did not realise that their child had been included in a research protocol. 16 (39%) parents did not understand the concept of randomisation. Half the parents could explain neither the aim of the clinical trial nor the potential benefit of inclusion to their child. Only one third of the parents were aware that they had an alternative. The principal factor underlying their decision, as stated by 29 parents (67%), was confidence in the medical team.
Conclusions The parents signed consent forms without having fully understood all the elements specific to the experimental protocol. Rather, the parents based their decision on their confidence in the medical team, even when their child's life was at risk.
Statistics from Altmetric.com
Acute leukaemia accounts for 30%–35% of cases of childhood cancer. In France, the incidence of acute leukaemia is four cases per year per 100 000 children aged <15 years, with a peak between the ages of 2 and 5 years.1 Improved treatment has considerably modified the course of this disease, whether involving the acute treatment of an invariably fatal disease or long-term treatment of a disease that often has a favourable outcome. The inclusion of a child in a clinical trial requires the consent of both parents. Requesting such consent for the inclusion of a child with leukaemia in a research protocol immediately after diagnosis presents various difficulties. In particular, there is a risk that information relating to the diagnosis and information associated with the protocol will be confused. We carried out a prospective study of the extent to which parents understand the information they receive when their consent is sought for the participation of their child in a paediatric hemato-oncology clinical trial. The trial concerned was the phase III randomised multicentre FRALLE 2000A study, which includes 60% of children with acute lymphoblastic leukaemia (ALL; non-T, standard-risk category) in France. We also assessed the parents' opinion of the consent process, identified elements likely to affect parental comprehension and investigated the reasons behind the parents' decision.
What is already known on this topic
▶. Problems of informed consent for paediatric cancer randomised trials have been observed in the USA and the UK.
▶. Understanding the information provided differs greatly between parents involved in clinical trials.
What this study adds
▶. Our study shows the apparent universality of poor comprehension by some parents.
▶. It is surprising how poor the process of comprehension remains in randomised clinical trials.
Patients and methods
This prospective study was carried out over a period of 20 months in the pediatric hematology departments of two Parisian hospitals (Armand Trousseau and Saint Louis) participating in the FRALLE 2000A protocol. We included all the parents from whom consent was sought for the participation of their children in this protocol at these two centres. The FRALLE 2000A trial is a phase III multicentric randomised study involving 50%–60% of children in France aged 1–10 years and treated for ALL with a white blood cell count at diagnosis below 50 000/µl.2 The principal aim of this trial was to determine whether there is any benefit of event-free survival from the administration of anthracyclines, daunorubicin in particular, during the induction phase of treatment for ALL with standard chemosusceptibility risk. The patients were randomised to two groups: one given anthracyclines and the other not given anthracyclines on day 21. The information document describes the following risks: aplasia, hair loss, digestive problems such as nausea, vomiting, mucitis and diarrhea, pancreatic problems, cardiac problems, bone and joint problems and neurological problems. The children were treated for approximately 3 years and participated in the study for a total of 5 years, including the follow-up period.
Physicians provided the parents with an information sheet and consent form 1 month before randomisation, together with a diagram showing the different phases of the treatment.
With the agreement of physicians participating in the study, the physician seeking parental consent suggested participation in our study to the parents. The parents were interviewed during a consultation scheduled as part of routine follow-up or during hospitalisation of their child. The choice of interview time was designed to minimise the additional constraints (time and travel) imposed on the parents. The parents were interviewed twice by a qualified psychologist devoted to our study 1 and 6 months after they were asked to take part in the FRALLE 2000A trial. The interviews took place on site and were recorded and fully transcribed. The parents were allowed to express themselves freely in a semidirected interview in which they were asked to respond to standardised questions. These questions were based on legal texts describing the elements required for consent to be valid3 4 and on previous national5 and international6 7 studies. The basic structure of the semidirected interviews was determined in collaboration with psychologists, a member of a parents' association for children with seriously illnesses and the investigating physicians from the centres participating in this study.
Parents were asked whether they were satisfied with the infomation they received from the investigator (“yes” or “no”), whether they judged the amount of information received to be enough (“sufficient”, “too much” or “not enough”) and whether they found the quality of the information to be adequate (“adequate” or “too complex”).
We evaluated the extent to which the parents had understood the 11 elements included in the information they were given concerning the FRALLE 2000A protocol by asking them about the content of the information sheets provided. Understanding was classified as complete, partial or null for each element if all, some or no words, respectively, of the information sheet were expressed by the parents. The questions asked are listed in table 1.
We then tried to identify factors predictive of sufficient understanding of the information received at the time at which consent was sought:
▶. whether the parents read the information sheet (yes/no)
▶. whether the parents sought additional information (yes/no)
▶. parents' profession
▶. native language (French or other).
To test the influence of covariates on comprehension by logistic regression, partial and complete comprehension were pooled for each of the 11 elements and compared with the absence of comprehension.
We also assessed whether parental comprehension remained stable between the two interviews (1 and 6 months after participation was first proposed).
We identified the reasons underlying the parents' decision to give their consent, using the following questions: “How difficult was it to take the decision you took concerning the participation of your child in this protocol?”, “What were the principal elements underlying your decision?”, “Who do you feel you took the final decision?”, “In your opinion, what role did your relationship with the physician who asked you to participate in the protocol play in influencing your decision?”, “What do you expect from the physician?”.
Each interview was coded by two investigators (psychologists) and was reviewed by a third (a pediatrician) in cases of disagreement. The collected data were entered into an Access database by H Chappuy, Paris, France. The results are expressed as percentages, means and SD, and medians and range. Qualitative values were compared using χ2 tests, quantitative values were compared using t tests and the influence of covariates was assessed by logistic regression analysis. Agreement between the two psychologists was assessed by a κ reliability test: for each item, the value of κ was above 0.9.
We met 37 mothers and 14 fathers for first interview (in eight cases, both parents were present for the interview). The second interview was attended by 29 mothers and 10 fathers (in seven cases, both parents were present). The children, 26 boys and 17 girls, had a mean age of 3.68 years (SD 2.26; range 1.61–9.8 years). French was not the native language for nine (18%) of the parents. Thirty-seven (72%) parents were married, four (8%) were divorced and two (4%) were single parents. In most cases, the family included two children (n=30, 70%). The profession was known for 47 of the parents: 20 (43%) had a high socioprofessional standing (senior management, company director, etc), 18 (38%) had intermediate status (employees and labourers) and nine (19%) had no profession.
Characteristics of the interviews
We carried out 75 interviews: 43 first interviews (M1) and 32 second interviews (M6). Six of those who attended the first interview refused to participate in the second, and six patients were lost to follow-up (medical management continued at another centre). The mean duration of the first interviews was 52.3 min (15–120 min; SD 19.8) and that of the second interviews was 39.4 min (15–75 min; SD 16.6). The mean time between the request for consent and the first interview was 33 days (25–64 days; SD 5.94), and the mean time between the request for consent and the second interview was 8.65 months (5–14 months; SD 2.2).
Satisfaction with the information provided
In most cases, consent for FRALLE 2000A was sought from both parents seen together (n=38, 88%). In five cases, only one of the parents received the information (geographical separation or single-parent families). The mean time taken to give a response was 4.56 days (1–21 days; SD 4.37). Thirty-one (72%) parents had read all or part of the information provided.
All the parents were satisfied with the explanations provided by the physician. Thirty-five (81%) parents felt that the information provided during the discussion of consent was sufficient, seven (16%) parents felt that they had received too much information and one (2%) parent did not consider there to have been enough information. Forty parents expressed an opinion concerning the quality of the information. Sixteen (40%) of these parents felt that the information provided was too complex (use of medical terms).
We obtained 41 responses concerning personal research for additional information: 19 (46%) parents had not sought additional information. Most (19/22) of the other parents had sought additional information about the disease rather than about the protocol (3/22). Their principal source of information was the internet.
Eight (19%) parents confused consent for care with consent for participation in research when they consented to the inclusion of their child in the FRALLE 2000A research protocol. Ten (23%) parents did not know that randomisation had been used to assign treatments, and six (14%) knew that randomisation existed but could not explain it.
Figure 1 summarises parental comprehension of the various items at M1 and M6. Comprehension of the tested items was similar at M1 and M6. Only the comprehension of the course of the protocol was higher after 6 months; comprehension of randomisation and individual benefits was lower.
All the parents were aware of at least one risk associated with participation. Only one (2%) parent was able to list all the risks involved. Twenty-five (60%) parents were able to list at least half the risks, whereas 6 (38%) parents were able to list fewer than half the risks. The risks most often identified were those of hair loss and cardiac problems.
Among the tested covariates, only the French native language was predictive of the comprehension of randomisation (p<0.001).
Factors affecting the decision
In response to the question “How difficult was it to take the decision you took concerning the participation of your child in this protocol?”, 35 (81%) parents said that the decision was easy either because they felt they had no choice (n=22) or because they found the decision logical and rational (n=13). Eight (19%) parents said that they had found the decision difficult.
In response to the question “What were the principal elements underlying your decision?”, the most frequent answers given by parents were confidence in the medical team (n=29; 67%), possible benefits for other children (n=16; 37%) and access to the best possible treatment for their own child (n=13; 30%).
Most parents (n=30; 71%) felt that they had control over the decision, whereas others felt they made the decision with the physician (n=3; 8%) and some felt that the physician was responsible for the decision (n=9; 21%). The patients expected the following qualities from the physician seeking informed consent: sincerity and transparency (n=22; 51%), competence (n=19, 44%), availability (n=14; 33%) and humaneness (n=9; 21%).
Understanding of the information provided differed greatly between parents; this seemed inconsistent with the satisfaction of almost all parents with the information received from the investigator. Half the parents had felt the need to look for additional information, although this mostly concerned information about the disease rather than the FRALLE 2000A protocol. The disparate parental understanding we observed is consistent with what has been reported by other researchers,8,–,10 showing the apparent universality of poor comprehension by at least some parents involved in clinical trials. Parental satisfaction with the information received does not, therefore, guarantee good comprehension. Ideally, the parents giving consent should have completely understood all the elements included on the consent form. In reality, comprehension differs according to the element considered. Thus, most of the parents understood the possibility of potential benefit to other children and the notion of voluntary participation, but fewer understood the basic elements of the protocol to which they had consented. The results of our study reflect not only parental comprehension but also information retention. Information retention is probably dependent on the parents' interest in the different elements of information. A significant proportion of parents did not understand that their children had participated in a clinical trial. This phenomenon has been previously described in clinical research involving both adults11 12 and children13 (“therapeutic misconception”). Indeed, Angiolillo et al14 suggested that a multistep process should be used for obtaining consent for studies addressing treatment of acute leukaemia. Other factors have been identified as being associated with poor parental comprehension. These factors include socioeconomic level and native language.9 15,–,18 We only found an effect of native language on the comprehension of randomisation in our study. Very few studies have investigated changes in parental comprehension over time in a paediatric setting. Greenley et al19 showed that the comprehension of most parents remains stable between two interviews at different times (day 2 and month 6), consistent with the findings of our study. The context of paediatrics may influence comprehension, as shown by Simon et al15: they found that adult patients tended to be better informed, interacted more closely with their physician and had a better understanding than parents of children participating in the study. The authors suggested that this difference arose from the greater emotional and psychological burdens on parents consenting to paediatric studies.
It is particularly surprising that in our study, many parents who gave their consent for the participation of their child in a clinical trial said that they felt they did not have a choice; this was not because they were coerced by the physician but because they felt that the FRALLE 2000A protocol was clearly the best treatment for their child. The possible benefit for other children had a considerable effect on the decision of the parents.20 21 Confidence in the physician was also a key factor affecting their decision, much more so than their knowledge of the protocol. The responses of the parents clearly show that they expect the physician to demonstrate transparency and sincerity.
Informed consent is never perfect, and perhaps, the process is more flawed when operating in the context of the emotions associated with the diagnosis of cancer. However, it remains surprising how poor the parental comprehension actually was. Parental understanding of the information provided could be improved in several ways. First, efforts should be made to distinguish as clearly as possible between care and research. If possible, information about care and research should be given at different places, at different times and by different people. A two-step consent process could be used, providing parents with the possibility of asking more detailed questions and allowing the investigating physician to explain the study again. Investigators should also systematically ask the parents to reformulate the information they have been given to verify that they have understood; this would also encourage active parental participation in a two-way information exchange process.
The authors thank the parents who agreed to participate in this study, making this work possible. The authors also thank the investigating physicians of the French Childhood Cancer Society (la Société Française des Cancers de l'Enfant).
Funding This work received financial support in the form of a Contrat d'Initiation à la Recherche Clinique (Assistance Publique Hôpitaux de Paris).
Competing interests None.
Ethics approval This study was approved by the ethics committee of the Hôpital Tarnier-Cochin, France.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.