Introduction

Consent to participate in research is obligatory, as has been clearly recognized by the World Medical Association in the Declaration of Helsinki [1]. When the subject is a minor, permission by a responsible relative can replace that of the subject. For parental or proxy consent to be valid, four essential components of the informed consent process must be fulfilled: the person granting permission must be mentally competent, have received appropriate information about the purpose and duration of the study and its risk and benefits, understand the information, and give consent voluntarily without coercion [2]. Previous studies have addressed these components [3, 4, 5] and factors associated with the decision to consent [3, 6, 7, 8, 9, 10].

To our knowledge, no studies have yet addressed the relationship between obtaining informed consent and factors associated with the severity of illness among a group of critically ill children admitted to a neonatal and pediatric intensive care (ICU). Our study examined whether parental authorization of involvement in a clinical study is influenced by the child’s severity of illness at the time of the consent decision. Part of this work has previously been published as an abstract in the 43rd Annual Congress of the ESPR in Utrecht, The Netherlands, on 4–9 September 2002 [11].

Materials and methods

The study population included parents or legal representatives of children who were approached for consent to enroll their child in a nutritional assessment study. Children from the age of 12 years were asked for consent themselves at the same time as their parents, provided their condition permitted this. Informed consent was requested within 24 h following ICU admission and was obtained by any of the members of the research team according to international guidelines [12]. Parents received the outline of the study in an informative document which adhered to the guidelines of the Central Committee on Research Involving Human Subjects [13]and had been approved by the institutional review board of our hospital. Eligible children were all those admitted to our ICU during 2001, from preterm neonates to 18-year-olds. Exclusion criteria were treatment with extracorporeal membrane oxygenation, withholding/withdrawing of treatment, and inclusion into an ongoing nutritional intervention study. Of the 421 cases informed consent was provided in 333 (79%). Of the 21 children aged 12 years or older 11 could be involved in the consent process; the other 10 were under sedateion and/or mechanical ventilation. Table 1 lists the reasons for the 88 refusals (5 in line with the child’s decision), with the most frequent reason (n=59) being that the proposed research to be too burdensome for their child; all refusers spontaneously expressed the reason for refusal.

Table 1 Reasons for declining informed consent (n=88)
Full size table

The proposed prospective nutritional assessment study [14] monitored the nutritional status of critically ill children by various means from admission to 6 months after discharge, including repeated antropometry (weight, length, circumferences of head, arm, and calf, and skinfolds), knemometry (measuring lower leg length), bioelectrical impedance analysis, indirect calorimetry, blood sampling, and stable isotope studies. Blood sampling was carried out only if arterial or venous access was already available (no additional vena puncture). The stable isotope studies required oral administration of the isotope (deuterium labeled water) and urine collection.

Clinical and demographic data were obtained both from children who participated in the nutritional assessment study and from those whose parents withheld permission to the proposed study. Severity of illness was assessed by means of validated scoring systems: the Pediatric Risk of Mortality score (PRISM) [15] and the Clinical Risk Index for Babies (CRIB) [16]. Furthermore, the Therapeutic Intervention Scoring System (TISS) score [17] was used to estimate the extent of interventions for each child during the first 24 h. Children were classified into three age groups: preterm neonates (gestational age <37 weeks), term neonates (0–30 days) and older children (>30 days). Since children aged 12 years or over are involved in the consent process, we also looked at this subgroup.

Data are expressed as median and range except when indicated otherwise. Parametric data were analyzed using Student’s t test. Nonparametric data were analyzed using Pearson’s χ2 test or Fisher’s exact test, and the Mann-Whitney U test. To adjust for the effect of other factors stepwise multivariate logistic regression analysis with backward elimination (likelihood ratio, p<0.1 for entry, p<0.05 for elimination) was carried out to examine which variables affected parents’ decision at the time of request. In the preterm and term neonates we checked birth weight, postconceptional age at admission, illness severity score, TISS score, and whether undergoing surgery. In the older children additional factors were previous health status and acute/elective admission. A two-tailed p value less than 0.05 was considered to indicate statistical significance.

Results

Patient and disease characteristics and factors associated with the disease severity are shown in Table 2 for the children with and children without consent. There was a significantly higher percentage of children aged at least 12 years in the nonconsent group, and a significantly higher proportion of older children without consent had a history of underlying disease. All five children aged 12 years or over who declined to participate had a history of chronic disease and several hospital admissions. Median PRISM and TISS scores were similar in both groups, but the median CRIB score and length of stay of the preterm neonates in the consent group was significantly higher than in the nonconsent group. Among the older children the nonconsent group had a higher proportion of deaths during admission, but their PRISM scores were not significantly higher (median score 21.5 vs. 19, p=0.829).

Table 2 Patient and disease characteristics and factors associated with illness severity (PRISM Pediatric Risk of Mortality, CRIB Clinical Risk Index for Babies, TISS Therapeutic Intervention Scoring system)
Full size table

In logistic regression analyses among term neonates none of the predefined variables explained parent’s decision whether to give informed consent, whereas among the preterm neonates a significant model was found using all five predefined factors accounting for 17% of the variation (p=0.025). Birth weight and postconceptional age were significant contributors. A post-hoc analysis carried out in the preterm neonates without a CRIB score (n=41, admission later than 12 h after birth), using the variables TISS score, postconceptional age, and birth weight revealed no significant factor explaining parent’s decision whether to give informed consent.

In the older children, no significant model was found using all six predefined factors. Following backward elimination three variables remained in the model (p=0.017) accounting for 12% of the variation: previous health status, undergoing surgery during admission, and PRISM score. This model suggests that parents of children with underlying disease were 3.2 times less likely to give informed consent for participation in the study, independent of the PRISM score or whether the child had to undergo surgery.

Discussion

Our study showed, contrary to our hypothesis, that a higher degree of current illness as determined by objective scores did not negatively influence parents’ willingness to give informed consent for participation of their critically ill child in a clinical study. In the preterm neonates the illness severity scores were even found to be higher in the group of children for whom consent was obtained. Logistic regression analyses revealed some significant child-related factors that influenced parental decision. Since these factors explained only 12–17% of the variation in the decision regarding consent, we conclude that parents also base their decision on factors other than factors related to illness severity, the child, or its disease. Children aged 12 years or over should be involved in the consent process when considering participation in research [13, 18, 19]. In our study nearly half of these children (5/11) who could be involved in the consent process did not want to participate, and each had a history of chronic disease. It seems that both in parents and children this factor is important in contemplating participation in research. We found no differences between the neonates in the consent group and in the nonconsent group concerning illness severity score, other clinical factors and illness severity perceived by parents which is in accordance with a previous study of newborn infants [6].

Concern has been expressed that many parents consenting to research do not understand the information or are too intimidated to refuse. We did not investigate the integrity of the consent process [3, 6], parental educational background, social economic status or personality; however, previous studies concerning these issues have been inconclusive [4, 5, 8, 20, 21]. We believe that a short general informative document explaining the nature of research being performed at the unit would help to prepare parents for the specific research that will be proposed.

Our study can be seen as nontherapeutic interventional research with minimal risk, but moderate burden on the child [12, 22], and it is debatable whether our results can be extrapolated to other types of studies, with different risks and benefits or lower burden. In this context, Pierro and Spitz [10] observed an increasing rate of parental refusal for nontherapeutic studies from 30% for performing anthropometric measures to 70% for stable isotope intravenous infusion, gas exchange measurements and blood sampling. A blinded randomized placebo controlled study conducted in our hospital investigating the analgesic effect of routine morphine infusion in preterm ventilated neonates had an inclusion rate of 71% [23]. This rate is comparable to the hypothetical 78% inclusion rate for studies involving moderate risk but possible major benefits, as reported by Singhal et al. [24] among parents with a child admitted to a neonatal ICU.

We conclude that the severity of illness as determined by objective scores did not differ between children whose parents consented and those whose parents did not consent. This suggests that parents are not influenced by the illness severity of their child in the decision to allow their child to participate in clinical observational research.