Factors influencing adherence to therapeutic recommendations made after diagnostic reassessment of medically unexplained symptoms in children and adolescents

Abstract

Background:

Medically unexplained symptoms (MUS) are common among children and adolescents and may be highly impairing. Even after long diagnostic and/or therapeutic trajectories, many of these children and their parents feel dissatisfied with the advice and therapies they were given.

Objectives:

After a 2-week hospitalisation for somatic and psychiatric reassessment, children and their families were given recommendations for further treatment. This study evaluates which of these recommendations were carried out (primary outcome measure) and which factors influenced the (non-)adherence to therapeutic advice.

Methods:

Parents of 27 children aged 7–17 with impairing MUS took part in a structured telephone survey to assess adherence to and perceived effectiveness of therapeutic recommendations (cross-sectional study). Influencing factors were analysed retrospectively.

Results:

Psychotherapy was recommended to all 27 patients and their families; 19 of them (70.4%) carried out this advice. When physiotherapy was recommended, adherence proved lower (6/22 children; 27.3%). No influencing factors were found to have a statistically significant correlation with adherence. Effect sizes may be indicative of clinically relevant influential factors, but should be considered cautiously.

Conclusion:

Results suggest that more efforts need to be made to ensure adherence to therapeutic recommendations. Known risk factors for non-adherence to treatments for chronic somatic disorders may not apply for children with somatoform disorders.

Keywords

Patient compliance somatoform disorders treatment outcome

Introduction

Medically unexplained symptoms (MUS), also referred to as functional somatic symptoms, can be defined as ‘bodily complaints of unclear aetiology, which are (currently) not fully explained by well-recognized somatic pathology’ (Eikelboom, Tak, Roest, & Rosmalen, 2016). MUS are common among children and adolescents across all ages (Domènech-Llaberia et al., 2004; Rask et al., 2009; van Geelen, Rydelius, & Hagquist, 2015) and have been shown to be associated with emotional, social and/or behavioural problems, as well as absence from school (Cerutti et al., 2017; Forgeron et al., 2010; Konijnenberg et al., 2005). However, this association with functional impairment is not always present. Although MUS are usually short-lived and transient, some MUS may be highly impairing. Therefore, MUS may be considered on a continuum of severity, ranging from children with transient and relatively mild symptoms to conditions with disabling MUS (Rask et al., 2009). While most children and adolescents with MUS will recover, a small subgroup will not. In this group of patients, a somatoform disorder can often be diagnosed, especially when the MUS are impairing. In our study, we report on a group of children and adolescents with impairing MUS, who all met the criteria necessary for diagnosing a somatoform disorder. Fortunately, evidence-based treatments are available for these patients.

Interventions

Most studies regarding the treatment of MUS and/or somatoform disorders concern adult patients. A Cochrane review on pharmacological interventions for somatoform disorders in adults indicated no difference between tricyclic antidepressants and placebo for outcome severity (Kleinstäuber et al., 2014). By contrast, new-generation antidepressants and natural products were found to be effective. The combination of selective serotonin reuptake inhibitors (SSRIs) and antipsychotics proved more effective than SSRIs alone. However, adverse effects and dropout occurred relatively often. Furthermore, the overall quality of evidence was low and the risk of bias was high. No trials included follow-up assessments.

With regard to children and adolescents, research on pharmacological interventions for MUS and/or somatoform disorders is much more limited. A Cochrane review (Kaminski, Kamper, Thaler, Chapman, & Gartlehner, 2011) identified only two randomised controlled trials (RCTs) examining the effectiveness of antidepressants for the treatment of abdominal pain-related functional gastrointestinal disorders (FGIDs) in children and adolescents, both using amitriptyline. The review revealed no statistically significant differences between amitriptyline and placebo for most efficacy outcomes. The authors concluded that for the majority of antidepressant medications, no evidence supported their use for the treatment of abdominal pain-related FGIDs in children and adolescents.

A recent meta-analysis examined the effectiveness of psychotherapy for adult patients with severe somatoform disorders compared with ‘treatment as usual’ (TAU; Koelen et al., 2014). Both psychotherapy and TAU were found to be effective for psychological symptoms, but psychotherapy was found to be more effective than TAU for physical symptoms and functional impairment. These effects were maintained at follow-up. On measures of functional impairment, patients who had received psychotherapy continued to improve, whereas patients who had received TAU did not.

A Cochrane review on non-pharmacological interventions for somatoform disorders and MUS in adults found that different forms of psychotherapy resulted in less severe symptoms at end of treatment, compared to patients receiving TAU or on a waiting list (van Dessel et al., 2014). These effects – though durable after 1 year of follow-up – were considered small to medium, while the overall quality of evidence was low and the risk of bias high. Moreover, psychotherapy resulted in a 7% higher proportion of dropouts during treatment than TAU.

A recent meta-analysis (Bonvanie et al., 2017) on psychotherapeutic interventions for children with MUS reported similar conclusions as those regarding adults. The study found that psychotherapeutic interventions significantly improved symptom load, disability and school absence compared to waiting list, placebo or TAU groups. At follow-up, these treatment effects were maintained.

There is also limited evidence to support the effectiveness of physiotherapy in the management of MUS in children and adolescents. For example, a systematic review (FitzGerald, Southby, Haines, Hough, & Skinner, 2015), including 12 studies, found limited and poor-quality evidence to establish the efficacy of physiotherapy in the treatment of child and adolescent conversion disorders.

With regard to adults with MUS and/or somatoform disorders, there is some more evidence in support of physiotherapy. Still, the Cochrane review on non-pharmacological interventions mentioned above identified no RCTs on the effectiveness of physiotherapy for adults with MUS (van Dessel et al., 2014). Nonetheless, the authors note that several studies have indicated that physical activity can improve mental health (e.g. mood, sleep and pain threshold) and has also proved beneficial for chronic fatigue syndrome and fibromyalgia. Another systematic review on physiotherapy for functional motor symptoms identified only one controlled intervention study (with a historical control group) and 28 case series or reports, most of which reported benefit from physical treatment, including some studies with long-term follow-up (Nielsen, Stone, & Edwards, 2013).

Adherence

To date, there are no studies reporting on adherence to therapeutic advice after diagnosis of a somatoform disorder in children. However, many studies have been published on compliance with treatment protocols for children with similar symptoms, such as children with chronic somatic illnesses. Risk factors for lower compliance may be illness-related (e.g. longer duration and lower functional impairment), treatment-related (e.g. a more complex treatment regimen), child-related (e.g. avoidant coping, high levels of emotional distress and older age), family-related (e.g. marital conflict, single parent families and lower socio-economic status) or related to the health care professional’s communication and behaviour (Fielding & Duff, 1999). Furthermore, children who perceive their ability to influence their health more positively, that is, children with a higher internal health locus of control (LOC), tend to be more compliant with therapy (Burkhart & Rayens, 2005; Nazareth et al., 2016). Similarly, a good therapeutic alliance with the health care professional, expecting a positive treatment outcome and being motivated for behaviour change are factors that correlate with the outcome of mental health treatments (De Nadai et al., 2017).

In summary, the effectiveness of psychotherapeutic interventions for the treatment of MUS and/or somatoform disorders in children is well established, and there is limited evidence for physiotherapy as a useful therapeutic intervention. However, most studies focus on how to treat these children and their families, yet fail to inform us on how to get them to engage in recommended treatments. Indeed, very little is known about (the risk factors for) non-adherence to therapeutic recommendations in this group of patients. In a systematic review (Bonvanie et al., 2017), most studies did not report on adherence to the therapeutic programme, nor did they analyse the risk factors for non-participation or dropout. Patient compliance was mentioned in only eight (out of 27) studies, which all stated (very) high compliance and/or attendance rates. Most of the interventions described were short (ranging from 3 weeks to 6 months) and not multimodal.

Objectives

To the best of our knowledge, there are no studies that focus on adherence to advice given to children and their families after diagnosis of a somatoform disorder (i.e. willingness to participate in therapeutic recommendations). This study aims to evaluate which recommendations were carried out by the children and their families and which factors were associated with (non-)adherence to therapeutic advice.

Methods

Study design, size, setting and participants

Parents of children referred to ‘KOLIK’ – a diagnostic collaboration between the Brussels University Hospital (UZ Brussel) Child and Adolescent Psychiatry and Paediatric Departments – because of impairing somatic symptoms for which no definitive diagnosis had been made were included. Impairment is defined as absence from school and/or inability to participate in everyday family or social activities, due to their somatic symptoms.

KOLIK is a multidisciplinary team, which seeks to perform a diagnostic reassessment during a 2-week hospitalisation at the UZ Brussel Paediatric ward. A somatic, psychiatric, (neuro)psychological and physiotherapeutic assessment is made. Afterwards, the child and his or her parents are provided with recommendations for further treatment. It is important to note that these children and their parents were not followed up systematically after their reassessment by KOLIK, and that KOLIK itself does not provide therapy.

The cross-sectional study was performed between 31 May 2017 and 9 September 2017. To maximise the power of this study, as many patients as possible were included, as only 41 children were assessed by KOLIK between 2012 and 2016 (see Chart 1). This study was approved by the UZ Brussel Medical Ethics Committee (B.U.N. 143201731928).

Chart 1.

Participation.

Only children with at least one Dutch-speaking parent (n = 36; 87.8%) were eligible for inclusion in this study. Parents were contacted up to three times; 27 parents (1 per child) were included. After obtaining written informed consent from the participating parent, a structured telephone survey was performed by trained examiners who were not part of the KOLIK team at the time of their child’s hospitalisation. The examiners were only aware of the recommendations that were made for further treatment, but were blind to the patients’ MUS and to the results of the diagnostic reassessment.

In the survey, parents were asked about the treatments that were started within 6 months after their child’s KOLIK reassessment. Several forms of therapy were probed individually.

Response and non-response groups did not differ regarding age (t(32.57) = 1.52; p = .14) or gender (χ² = 1.09; p = .30) of the child, educational level of mothers (U = 110.00; p = .65) or fathers (U = 115.50; p = .82) or family constellation (FE = 1.00; p = .75).

Instruments/measures

Parental satisfaction with KOLIK was assessed during the telephone survey with the Cliëntentoets or C-toets (Ramos et al., 2006), a 15-item questionnaire with a 4-point Likert-type format (score range 15–60; higher scores indicate higher satisfaction).

A number of questionnaires were filled out by the patients and their parents during the hospitalisation. Children reported depressive symptoms using the Dutch version of the Child Depression Inventory (CDI) (Braet & Timbremont, 2002; Kovacs, 1981) – a 27-item questionnaire with a 3-point Likert-type format. Health LOC was measured using the Dutch version of the Children’s Health Locus of Control (CHLC) (Parcel & Meyer, 1978) – an 18-item questionnaire with a 5-point Likert-type format containing three factors (internal, external powerful others and external chance). Only the factor ‘internal health LOC’ was used.

All parental questionnaires were completed by the children’s mothers. Their parenting stress was assessed with the Dutch version of the Parenting Stress Index (Abidin, 1983; de Brock, Vermulst, Gerris, & Abidin, 1992). The short version (NOSIK) consists of 25 statements with a 7-point Likert-type format. Parental health LOC was measured with the Dutch version of the Multidimensional Health Locus of Control (MHLC) scale (Halfens, 1985; Wallston, Wallston, & DeVellis, 1978) containing three factors, with six statements about health beliefs for each of these factors (internal, external powerful others and external chance). Items are rated on a 5-point Likert-type format. Only the factor ‘internal health LOC’ was used.

Variables/outcome measures and influencing factors

The primary outcome measure, adherence to therapeutic advice, is defined as starting with (or being on a waiting list for) a recommended therapy within 6 months after the KOLIK assessment.

Several influencing factors were taken into account, based on our literature review:

Illness-related: duration of symptoms;

Child-related: age, self-reported depressive symptoms (as a measure for emotional distress) and health LOC;

Family-related: family stability, parental educational level (as a measure for socio-economic status), maternal health LOC and parent-reported stress;

Health care professional-related: parental satisfaction with KOLIK.

Data sources

Information on the primary outcome measure was obtained through the telephone survey, as well as parental satisfaction with KOLIK. All other data were documented in the patients’ medical files and were retrospectively analysed.

Bias

To avoid socially desirable answers, parents were contacted by an examiner who they had never been in touch with before. They were informed that their answers would be processed anonymously and that (non-)participation in the study would not influence further care by the hospital.

To minimise memory bias and/or generalisation, the telephone survey was clearly structured and each form of recommended treatment was probed individually.

Statistical analysis

Descriptive analysis was done using frequency distributions. Differences were examined using t tests and chi-square tests (χ²). These were converted to correlation coefficients (r) for the calculation of effect sizes (ES), as recommended for correlational designs (Durlak, 2009). Note that r ⩾ .10 is usually considered ‘small’, r ⩾ .30 ‘medium’ and r ⩾ .50 ‘large’.

Results

Descriptive data

Participants

Patient and family characteristics are presented in Table 1. The group consisted of 27 children (7 boys and 20 girls) between 7 and 17 years old, at the time of their KOLIK reassessment. Children reported relatively high levels of depressive symptoms (as a measure for emotional distress). As the internal health LOC perception score can vary between 6 and 30, most children agreed moderately with the idea that they have influence on their own health. Most children grew up in a two-parent family. Most mothers were highly educated and reported moderate levels of parenting stress. Their internal health LOC scores indicated that they agreed moderately with the idea that they have influence on their own health. As satisfaction scores can vary between 15 and 60, parents were relatively satisfied with KOLIK.

Table 1.

Patient and family characteristics.

Children
Gender	20 girls (74.1%); 7 boys (25.9%)
Age (years)	Range: 7–17; mean: 13.78 (SD 2.50)
Emotional distress (CDI; percentile)	Range: 4–96; mean: 61.56 (SD 27.74)
Internal health LOC (CHLC; rough score)	Range: 6–22; mean: 16.04 (SD 4.15)
Parents
Family constitution	A total of 16 two-parent families (59.3%); 7 one-parent families (25.9%); four newly composed families (14.8%)
Educational level (mother)	18 higher education (66.7%); 9 lower education (33.3%)
Parenting stress (NOSIK; rough score)	Range: 25–112; mean: 43.26 (SD 21.25)
Internal health LOC (MHLC; rough score)	Range: 6–26; mean: 17.65 (SD 4.55)
Satisfaction with KOLIK (C-toets; rough score)	Range: 35–60; mean: 46.23 (SD 7.58)

CDI: Child Depression Inventory; LOC: locus of control; CHLC: Children’s Health Locus of Control; MHLC: Multidimensional Health Locus of Control.

Reasons for referral

Most children reported multiple somatic symptoms for which no (broadly accepted) somatic diagnosis had been made (see Table 2). The most commonly reported symptoms were sleeping problems and/or tiredness, neurological (e.g. headache, pseudo-epileptic episodes and loss of function), abdominal (e.g. vomiting, diarrhoea and constipation) and muscle or joint symptoms (e.g. aches, cramps and weakness). Thoracic (e.g. hyperventilating and chest pains) and ENT symptoms (e.g. sore throat, swollen glands and trouble swallowing) were less common. One child complained of ‘excessive drinking and peeing’ and one child reported a ‘low blood pressure’.

Table 2.

Reasons for referral.

Symptoms	Number of children
More than one symptom	24 (88.9%)
Only one symptom	3 (11.1%)
Abdominal symptoms	21 (77.8%)
ENT symptoms	4 (14.8%)
Muscle/joint symptoms	14 (51.9%)
Neurological symptoms	18 (66.7%)
Sleeping problems/tiredness	21 (77.8%)
Thoracic symptoms	10 (37.0%)
Other symptoms	2 (7.4%)

As most children reported more than one symptom, the sum is not 100%.

Pre-referral trajectory

Most children had gone through long, unsatisfying diagnostic (and therapeutic) trajectories prior to their referral to KOLIK (see Table 3). Nearly half of the children had previously had a somatic disorder that caused similar symptoms as the ones they were currently struggling with (e.g. appendicitis 2 years ago and currently abdominal pain), possibly making the diagnostic process more difficult. They had often sought medical and/or other forms of therapy, without the results they had hoped for.

Table 3.

Pre-referral trajectory.

Duration of symptoms (months)	Range: 4–72; mean: 16.9 (SD 15.4)
	Number of children
Previous diagnoses
Somatic disorder with similar symptoms	13 (48.1%)
Chronic somatic disorder with exacerbation risk	2 (7.4%)
Chronic fatigue syndrome	1 (3.7%)
Neurobiological psychiatric disorder (e.g. ADHD)	5 (18.5%)
Depressive disorder	1 (3.7%)
Previously consulted doctors
General practitioner	10 (37.0%)
Somatic specialist doctor	15 (55.6%)
(Child and adolescent) psychiatrist	5 (18.5%)
Previous forms of therapy
Medication for symptom relief	14 (51.9%)
Psychotropic medication	7 (25.9%)
Psychotherapy	17 (63.0%)
Physiotherapy	2 (7.4%)

ADHD: attention-deficit hyperactivity disorder.

Diagnosis

After somatic and psychiatric diagnostic reassessment, a previously unknown somatic disorder was diagnosed in four children. A non-somatoform psychiatric disorder was diagnosed in more than half of the children (see Table 4).

Table 4.

Diagnoses after somatic and psychiatric reassessment.

Disorders	Number of children
Previously unknown somatic disorder	4 (14.8%)
Migraine and tension headache	1 (3.7%)
Pituitary adenoma	1 (3.7%)
Asthma	1 (3.7%)
Helicobacter pylori gastritis	1 (3.7%)
Non-somatoform psychiatric disorder	14 (51.9%)
Neurodevelopmental disorder	5 (18.5%)
Mood disorder	4 (14.8%)
Anxiety disorder	5 (18.5%)
Trauma-related disorder	1 (3.7%)
Eating disorder	1 (3.7%)

Recommendations

Therapeutic recommendations focused on psychotherapy and physiotherapy, in line with evidence-based guidelines and recent findings regarding the treatment of somatoform disorders (see Table 5). In some cases, additional treatments were recommended based on diagnostic findings (e.g. comorbid psychiatric or somatic diagnosis). Because previously unknown somatic diagnoses were a relatively rare find (4 out of 27 children), somatic treatments and/or follow-up by a somatic specialist doctor were recommended to only a few children. For all of these recommendations, referrals were carried out; with permission of the child and his or her parents, therapists were contacted by phone or received a referral letter. The recommended therapeutic approach varied for every patient, as there were often comorbid psychiatric disorders and/or family problems. The KOLIK team assessed which problems needed to be addressed first and talked this through with the patients and their parents, as well as with their future therapist(s) over the phone.

Table 5.

Recommendations after reassessment.

	Number of children
Recommendations
Psychotherapy	27 (100%)
Physiotherapy	22 (81.5%)
Other recommendations
Psychiatric treatment
Psychotropic medication	4 (14.8%)
In-patient treatment	4 (14.8%)
Somatic treatment
Follow-up by somatic specialist doctor	4 (14.8%)
Somatic medication	3 (11.1%)

Outcome data

Adherence to therapeutic recommendations

Psychotherapy was recommended to all 27 patients and their families; 19 of them (70.4%) carried out this advice within 6 months after their KOLIK reassessment. Adherence proved lower for physiotherapy, which was recommended to 22 children, of whom only 6 (27.3%) followed this advice.

None of the factors that were examined (see Tables 6 and 7) proved to have a statistically significant association with adherence to therapeutic recommendations. ES differed with regard to adherence to the recommendation of psychotherapy versus physiotherapy. Some factors had a small (r ⩾ .10) to moderate (r ⩾ .30) effect on adherence to both forms of recommended therapy, namely, shorter duration of illness, standard family constitution, lower parenting stress, lower maternal internal health LOC and higher satisfaction with KOLIK. Other factors only had an effect on psychotherapy (older age of the child and lower maternal educational level) or on physiotherapy (lower emotional distress, higher internal health LOC and higher maternal educational level).

Table 6.

Factors influencing adherence to therapeutic recommendations.

Influencing factor	Psychotherapy (recommended)
	Adherent group (n = 19)	Non-adherent group (n = 8)	χ²/t	ES (r)
Illness-related
Duration	15.0 (15.2)	21.5 (16.0)	1.0	.20
Child-related
Age	14.2 (2.1)	12.8 (3.2)	−1.4	.27
Emotional distress	60.1 (29.0)	65.1 (25.9)	0.4	.09
Internal health LOC	13.8 (3.5)	13.1 (3.4)	−0.5	.09
Family-related
Family constitution			0.4	.12
Two-parent family	n = 12	n = 4
Other family constitution	n = 7	n = 4
Educational level	3.4 (1.2)	4.0 (0.5)	1.8	.33
Parenting stress	40.9 (19.5)	48.9 (25.5)	0.9	.17
Internal health LOC	17.0 (4.6)	19.1 (4.4)	1.1	.22
Caregiver-related
Satisfaction with KOLIK	47.4 (8.0)	43.5 (6.1)	−1.2	.24

ES: effect sizes; LOC: locus of control.

Results are presented in the same way as in Table 1 (i.e. percentiles, rough scores etc.).

Table 7.

Factors influencing adherence to therapeutic recommendations, continued.

Influencing factor	Physiotherapy (recommended)
	Adherent group (n = 6)	Non-adherent group (n = 16)	χ²/t	ES (r)
Illness-related
Duration	13.2 (6.9)	17.9 (18.3)	0.6	.14
Child-related
Age	14.2 (1.8)	14.3 (2.2)	0.1	.03
Emotional distress	47.0 (34.0)	64.8 (25.6)	1.3	.28
Internal health LOC	14.5 (2.1)	13.5 (3.7)	−0.6	.14
Family-related
Family constitution			2.8	.35
Two-parent family	n = 5	n = 7
Other family constitution	n = 1	n = 9
Educational level	4.0 (0.9)	3.6 (0.6)	−1.1	.24
Parenting stress	32.5 (4.5)	44.1 (21.2)	1.3	.28
Internal health LOC	15.0 (6.3)	18.7 (3.6)	1.3	.37
Caregiver-related
Satisfaction with KOLIK	49.5 (7.6)	44.5 (7.3)	−1.4	.30

ES: effect sizes; LOC: locus of control.

Results are presented in the same way as in Table 1 (i.e. percentiles, rough scores etc.)

Discussion

Key results

When evaluating whether therapeutic recommendations were carried out within 6 months after KOLIK’s reassessment, it is immediately striking that in many cases, recommendations were not carried out. Psychotherapy was recommended to all 27 patients and their families, but only 19 (70.4%) of them carried out this advice. Physiotherapy was recommended to 22 children, of whom only 6 (27.3%) followed this advice. None of the hypothetical influencing factors proved to have a significant association with adherence to therapeutic recommendations. Of course, given the small sample size, it is not entirely surprising that no statistically significant associations were found, and statistical significance is not the same as clinical relevance (Durlak, 2009). ES may be indicative of clinically relevant influential factors but should be considered cautiously. Based on ES, some factors had a small to moderate effect on adherence to recommendations. Somewhat unexpectedly, these effects differed with regard to adherence to the recommendation of psychotherapy versus physiotherapy.

Interpretation

Although many factors influence the effectiveness of therapeutic interventions, attendance at and adherence to therapy are arguably the most basic requirements for a treatment to work (Nock & Ferriter, 2005). In this study, adherence to KOLIK’s therapeutic recommendations is defined as the commencement of therapy within 6 months after KOLIK’s reassessment. It is reasonable to assume that the continuation of therapy depends less on KOLIK’s referral and more on risk factors for premature termination of therapy.

KOLIK’s therapeutic recommendations were often not carried out. While the aim for adherence should be 100%, in this cohort it proved much lower. Several explanations can be put forward to address this finding.

All children and their families had long diagnostic (and therapeutic) trajectories prior to KOLIK’s reassessment, but no somatic explanation had been found. They may well have hoped that KOLIK’s reassessment would finally reveal the ‘true problem’. However, a new somatic diagnosis was made for only four children (14.8%), never offering a full explanation for their disfunctioning. Thus, it is possible that KOLIK was not seen as ‘the end of the road’, but rather as yet another unsatisfactory step on their way to a somatic diagnosis.

This study examined whether illness-, child-, parent- or caregiver-related factors influenced adherence to therapeutic recommendations, as has been demonstrated for children with chronic somatic illnesses. Statistical analysis did not reveal any significant correlation between these factors and therapeutic recommendations. Of course, statistical analysis may be hampered by the lack of power due to the small sample, meaning interpretation of results is difficult. It is certainly possible that these factors do influence adherence, but that the sample was simply too small to be able to detect the correlation.

However, it may also be the case that these risk factors do not apply to somatoform disorders. It is reasonable to assume that in this group, patients’ (and their parents’) views on the likely cause and preferred treatment of their symptoms are very different from the caregivers’ diagnosis and advice (Goldbeck & Bundschuh, 2007), in contrast to somatic disorders. While patients with somatoform disorders tend to see their symptoms as a (purely) somatic problem, the treatments they are offered are usually not somatic in nature.

Therefore, rather than only focusing on risk factors for non-adherence to therapeutic recommendations in children with similar symptoms, it might be useful to look at children who were given similar therapeutic recommendations. A number of risk factors for premature termination of psychotherapy (which was recommended to all of our patients) have been described. Even though these risk factors may differ from factors influencing whether to commence psychotherapy, some of them may still apply. The same barriers that lead to premature termination of psychotherapy could prevent someone from taking it up in the first place. A model of treatment barriers (Kazdin, Holland, & Crowley, 1997; Kazdin, Holland, Crowley, & Breton, 1997) proposes four primary domains, which have all been shown to significantly contribute to premature termination of child psychotherapy: (1) the experience of stressors and obstacles (e.g. logistical difficulties to attending therapy), (2) a poor relationship with the therapist, (3) perceptions that treatment is not relevant and (4) perceptions that treatment is too demanding.

In this sample, the relationship with KOLIK’s team was probed using the C-toets, which included questions on communication and caregiver attitude. Parents reported higher satisfaction with KOLIK on the C-toets (which included questions on professionalism and expertise) in the adherent than in the non-adherent group for both psychotherapy and physiotherapy, although this difference was not statistically significant. ES were small to moderate.

The C-toets also included questions on professionalism and expertise. The lower scores in the non-adherent groups may indicate that parents did not trust KOLIK’s diagnostic reassessment to be definitive and/or correct (thus perceiving that the recommended treatment is not relevant). Children and parent’s beliefs are known to play a crucial role in their motivation to seek help and follow therapeutic recommendations (Morrissey-Kane & Prinz, 1999). Children and/or their parents may have remained convinced that their symptoms had a somatic cause and therefore that treatment should have been somatic in nature. Agreeing to psychotherapy may have required a shift in their illness perception.

The belief that what is required in treatment exceeds what one is prepared to do or what one is able to do can be interpreted as the ‘perception that treatment is too demanding’. Children who perceive their ability to influence their health more positively tend to be more compliant with therapy (Burkhart & Rayens, 2005; Nazareth et al., 2016). In this sample, children in the adherent groups had a higher internal health LOC score than in the non-adherent groups. However, this difference was not significant and ES were trivial to small. Interestingly, an inverse trend was seen in the mothers’ health perception, perhaps indicating that their mothers perceived external factors as more influential. Again, this difference was not statistically significant and ES were small to moderate.

It seems reasonable to assume that these barriers could apply to physiotherapy as well. Comparable barriers to treatment adherence have been described regarding physiotherapy in children with somatic disorders, such as cystic fibrosis or juvenile idiopathic arthritis (Favier et al., 2018; Happ, Hoffman, DiVirgilio, Higgins, & Orenstein, 2013). In our sample, the subgroups were too small to perform further statistical analysis on possible influencing factors regarding the uptake of physiotherapy. The fact that adherence to this recommendation was lower than for psychotherapy might be due to a number of reasons. For instance, less emphasis may have been put on the need for physiotherapy in case of psychiatric comorbidity (which occurred in half of the children). The combination of psychotherapy and physiotherapy may also have been too demanding or costly.

Generalisability

MUS are common across all ages and may be highly impairing, though this is certainly not always the case. KOLIK only proposes a 2-week hospitalisation if patients have highly impairing and chronic MUS for which no definitive diagnosis has been made. Logically, this group of patients is not representative of all children with MUS or with a somatoform disorder.

KOLIK seeks to perform a diagnostic reassessment but does not offer any form of therapy to this group of patients or their parents. This way of working may be quite different from other centres, making it harder to compare the results of this study to those of other (treatment) settings. It may well be the case that adherence to therapeutic recommendations would be higher if diagnostic assessment and treatment were offered at the same site, as the caregiver relationship would only have to be established once. This is especially relevant for children with somatoform disorders, who have often gone through a long, unsatisfying diagnostic trajectory before finally getting their diagnosis. In this study, children reported having MUS for 4–72 months prior to the KOLIK reassessment. Most of them had consulted several doctors and/or had tried at least one form of therapy.

Given this often wearisome diagnostic process, it is important that different clinicians work together to reach the correct diagnosis as soon as possible, to avoid performing unnecessary additional tests and wasting time before starting with evidence-based treatments. Yet, first-line clinicians do not always work in a multidisciplinary setting, and may fear that they could miss a (rare) somatic disorder, thereby postponing the diagnosis of a somatoform disorder. In this sense, KOLIK is ideally positioned, as it is embedded in a university hospital, where the necessary expertise is at hand, both somatic and psychiatric. Furthermore, patients and their parents may hold a university hospital in higher esteem than a first-line clinician, perhaps giving KOLIK’s advice more weight. However, the threshold to seeking advice in a university hospital is much higher. KOLIK’s positioning as a collaboration between the Child and Adolescent Psychiatry and Paediatric Departments may also make patients hesitant to seek advice there, not wanting to be assessed psychiatrically, if they are convinced that there is only a somatic problem. On the other hand, patients who have been told their symptoms are probably caused by ‘stress’ may be more inclined to seek KOLIK’s advice, thanks to the somatic reassessment. These considerations should be taken into account when comparing the results of this study to those of other centres.

Limitations

An important limitation of this one-centre study is its small sample size. Over a 5-year period, 36 children met inclusion criteria, 27 of whom (75.0%) were included in this study. Several influencing factors were controlled for in the responders versus non-responders groups; no significant differences were found. A multicentre study could increase the sample size, but KOLIK’s positioning in a university hospital and its way of working (only diagnostic, not therapeutic) make this difficult. A longer inclusion period could also allow for a larger sample size.

With regard to health LOC, only the children’s mothers were asked to fill out a questionnaire at the time of their hospitalisation. Information on the fathers’ health LOC perception is therefore lacking. It would have been interesting to assess differences between the parents’ health LOC and to look into how these may have influenced adherence to therapeutic recommendations.

This study has a relatively high risk of memory bias, as some parents were contacted up to 5 years after the KOLIK reassessment. Parents may have forgotten how long it took to get a therapy in place and may have a different view on its effectiveness with the child’s current (dis)functioning in mind. Recall bias and generalisation were tried to be limited using a structured interview, in which each form of therapy was probed in detail. A prospective study design could virtually have avoided this risk.

A telephone survey does not feel like an anonymous questionnaire and may have led some parents to answer questions in a socially acceptable or non-confrontational way. Social desirability bias was tried to be limited by assuring their anonymity and by having the telephone survey performed by a researcher, the parents had never been in contact with.

Conclusion

MUS are common in children and adolescents of all ages and may be highly impairing. They often go through long and frustrating diagnostic (and therapeutic) trajectories before a somatoform disorder is finally diagnosed. Although evidence-based treatments are available, very little is known about adherence to therapeutic recommendations and the risk factors for non-adherence in this group of patients. This study aimed to address this gap in our current knowledge. Adherence to therapeutic recommendations was relatively low, especially with regard to physiotherapy. Several known risk factors for non-adherence to treatments for chronic somatic illnesses, as well as risk factors for premature termination of psychotherapy, were looked into. None of these hypothetical factors proved to correlate significantly with adherence to recommended treatments (i.e. commencement of therapy after diagnostic reassessment). The study sample may well have been too small to detect significant correlations, but it could also be the case that the same risk factors do not apply to this specific group of patients. Therefore, it remains unclear at this point which risk factors should be taken into account when assessing the risk of non-adherence. Because prognosis improves with adherence to evidence-based treatments, this remains a very important question. ES in this study may be indicative of clinically relevant influential factors, but should be considered cautiously. Furthermore, research is needed to address this issue, by means of a larger study sample and/or qualitative research to form new hypotheses.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author(s) received financial support for the research, authorship, and/or publication of this article: This study was funded by UZ Brussel (University Hospital Brussels). The UZ Brussel was not involved in the study design, in the collection, analysis and interpretation of data, in the writing of the report or in the decision to submit this article for publication.

ORCID iD

Sam LB Bonduelle

Author biographies

Sam LB Bonduelle was a Child and Adolescent Psychiatry intern at the Dutch-speaking Free University of Brussels (VUB) at the time of this study. He has since completed his internship and now works as a child and adolescent psychiatrist at the Brussels University Hospital (UZ Brussel) Department for Child and Adolescent Psychiatry (PAika).

Johan Vanderfaeillie is an Educational Studies professor at the Dutch-speaking Free University of Brussels (VUB) and works at PAika as an educational specialist. His fields of interest include medically unexplained symptoms, autism spectrum disorder and foster care.

Katrien Denijs is a Clinical Psychologist at PAika. Her fields of interest include medically unexplained symptoms and autism spectrum disorder. She is the coordinator of the KOLIK team.

Annik Lampo is a Child and Adolescent Psychiatry professor at the VUB and head of the PAika Department. She has a supervising role in the KOLIK team.

Lindita Imeraj is a Child and Adolescent Psychiatry professor at the VUB and a child and adolescent psychiatrist at PAika. She has a supervising role in the KOLIK team.

References

Abidin

R. R.

(1983). Parenting stress index. Charlottesville, VA: Pediatric Psychology Press.

Bonvanie

I. J.

Kallesøe

K. H.

Janssens

K. A. M.

Schröder

Rosmalen

J. G. M.

Rask

C. U.

(2017). Psychological interventions for children with functional somatic symptoms: A systematic review and meta-analysis. The Journal of Pediatrics, 187, 272–281.e17. doi:10.1016/j.jpeds.2017.03.017

Braet

Timbremont

(2002). Children’s Depression Inventory (Nederlandstalige versie). Amsterdam, The Netherlands: Harcourt Test Publishers.

Burkhart

P. V.

Rayens

M. K.

(2005). Self-concept and health locus of control: Factors related to children’s adherence to recommended asthma regimen. Pediatric Nursing, 31, 404–409. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16295156

Cerutti

Spensieri

Valastro

Presaghi

Canitano

Guidetti

(2017). A comprehensive approach to understand somatic symptoms and their impact on emotional and psychosocial functioning in children. PLoS ONE, 12(2), e0171867. doi:10.1371/journal.pone.0171867

de Brock

A. J. L. L.

Vermulst

A. A.

Gerris

J. R. M.

Abidin

R. R

. (1992). Nijmeegse Ouderlijke Stress Index (Handleiding experimentele versie) [Nijmegen Parental Stress Index (Manual for experimental version)]. Lisse, The Netherlands: Swets & Zeitlinger.

De Nadai

A. S.

Karver

M. S.

Murphy

T. K.

Cavitt

M. A.

Alvaro

J. L.

Bengtson

. . . Storch

E. A

. (2017). Common factors in pediatric psychiatry: A review of essential and adjunctive mechanisms of treatment outcome. Journal of Child and Adolescent Psychopharmacology, 27, 10–18. doi:10.1089/cap.2015.0263

Domènech-Llaberia

Jané

Canals

Ballespí

Esparó

Garralda

(2004). Parental reports of somatic symptoms in preschool children: Prevalence and associations in a Spanish sample. Journal of American Academy of Child and Adolescent Psychiatry, 43, 598–604. doi:10.1097/00004583-200405000-00013

Durlak

J. A.

(2009). How to select, calculate, and interpret effect sizes. Journal of Pediatric Psychology, 34, 917–928. doi:10.1093/jpepsy/jsp004

10.

Eikelboom

E. M.

Tak

L. M.

Roest

A. M.

Rosmalen

J. G. M.

(2016). A systematic review and meta-analysis of the percentage of revised diagnoses in functional somatic symptoms. Journal of Psychosomatic Research, 88, 60–67. doi:10.1016/j.jpsychores.2016.07.001

11.

Favier

L. A.

Taylor

Loiselle Rich

Jones

K. B.

Vora

S. S.

Harris

J. G.

. . .Modi

A. C.

(2018). Barriers to adherence in Juvenile idiopathic arthritis: A multicenter collaborative experience and preliminary results. The Journal of Rheumatology, 45, 690–696. https://doi.org/10.3899/jrheum.171087

12.

Fielding

Duff

(1999). Compliance with treatment protocols: Interventions for children with chronic illness. Archives of Disease in Childhood, 80, 196–200. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10325743

13.

FitzGerald

T. L.

Southby

A. K.

Haines

T. P.

Hough

J. P.

Skinner

E. H.

(2015). Is physiotherapy effective in the management of child and adolescent conversion disorder? A systematic review. Journal of Paediatrics and Child Health, 51, 159–167. doi:10.1111/jpc.12630

14.

Forgeron

P. A.

King

Stinson

J. N.

McGrath

P. J.

MacDonald

A. J.

Chambers

C. T.

(2010). Social functioning and peer relationships in children and adolescents with chronic pain: A systematic review. Pain Research & Management, 15, 27–41. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20195556

15.

Goldbeck

Bundschuh

(2007). Illness perception in pediatric somatization and asthma: Complaints and health locus of control beliefs. Child and Adolescent Psychiatry and Mental Health, 1, 5. doi:10.1186/1753-2000-1-5

16.

Halfens

R. J. G.

(1985). Locus of control: de beheersingsorientatie in relatie tot ziekte- en gezondheidsgedrag [Locus of control: control orientation in relation to illness and health behavior]. Maastricht, The Netherlands: Rijksuniversiteit Limburg.

17.

Happ

M. B.

Hoffman

L. A.

DiVirgilio

Higgins

L. W.

Orenstein

D. M.

(2013). Parent and child perceptions of a self-regulated, home-based exercise program for children with cystic fibrosis. Nursing Research, 62, 305–314. doi:10.1097/NNR.0b013e3182a03503

18.

Kaminski

Kamper

Thaler

Chapman

Gartlehner

(2011). Antidepressants for the treatment of abdominal pain-related functional gastrointestinal disorders in children and adolescents. Cochrane Database of Systematic Reviews, 7, CD008013. doi:10.1002/14651858.CD008013.pub2

19.

Kazdin

A. E.

Holland

Crowley

(1997). Family experience of barriers to treatment and premature termination from child therapy. Journal of Consulting and Clinical Psychology, 65, 453–463. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9170769

20.

Kazdin

A. E.

Holland

Crowley

Breton

(1997). Barriers to treatment participation scale: Evaluation and validation in the context of child outpatient treatment. Journal of Child Psychology and Psychiatry, and Allied Disciplines, 38, 1051–1062. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9413802

21.

Kleinstäuber

Witthöft

Steffanowski

van Marwijk

Hiller

Lambert

M. J.

(2014). Pharmacological interventions for somatoform disorders in adults. Cochrane Database of Systematic Reviews, 11, CD010628. doi:10.1002/14651858.CD010628.pub2

22.

Koelen

J. A.

Houtveen

J. H.

Abbass

Luyten

Eurelings-Bontekoe

E. H. M.

Van Broeckhuysen-Kloth

S. A. M.

Geenen

(2014). Effectiveness of psychotherapy for severe somatoform disorder: Meta-analysis. British Journal of Psychiatry, 204, 12–19. doi:10.1192/bjp.bp.112.121830

23.

Konijnenberg

A. Y.

Uiterwaal

C. S. P. M.

Kimpen

J. L. L.

Van Der Hoeven

Buitelaar

J. K.

De Graeff-Meeder

E. R.

(2005). Children with unexplained chronic pain: Substantial impairment in everyday life. Archives of Disease in Childhood, 90, 680–686. https://doi.org/10.1136/adc.2004.056820

24.

Kovacs

(1981). Rating scales to assess depression in school-aged children. Acta Paedopsychiatrica, 46, 305–315. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/7025571

25.

Morrissey-Kane

Prinz

R. J.

(1999). Engagement in child and adolescent treatment: The role of parental cognitions and attributions. Clinical Child and Family Psychology Review, 2, 183–198. doi:10.1023/A:1021807106455

26.

Nazareth

Richards

Javalkar

Haberman

Zhong

Rak

. . . van Tilburg

M. A. L.

(2016). Relating health locus of control to health care use, adherence, and transition readiness among youths with chronic conditions, North Carolina, 2015. Preventing Chronic Disease, 13, E93. doi:10.5888/pcd13.160046

27.

Nielsen

Stone

Edwards

M. J.

(2013). Physiotherapy for functional (psychogenic) motor symptoms: A systematic review. Journal of Psychosomatic Research, 75, 93–102. doi:10.1016/j.jpsychores.2013.05.006

28.

Nock

M. K.

Ferriter

(2005). Parent management of attendance and adherence in child and adolescent therapy: A conceptual and empirical review. Clinical Child and Family Psychology Review, 8, 149–166. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15984084

29.

Parcel

G. S.

Meyer

M. P.

(1978). Development of an instrument to measure children’s health locus of control. Health Education Monographs, 6, 149–159. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/689889

30.

Ramos

Stams

G. J.

Stoel

Faas

van Yperen

T. A.

Dekovic

(2006). Het meten van cliënttevredenheid in de jeugdzorg [Measuring client satisfaction in youth care]. Kind en Adolescent, 27, 99–105. doi:10.1007/BF03060989

31.

Rask

C. U.

Olsen

E. M.

Elberling

Christensen

M. F.

Ørnbøl

Fink

Skovgaard

A. M.

(2009). Functional somatic symptoms and associated impairment in 5–7-year-old children: The Copenhagen child cohort 2000. European Journal of Epidemiology, 24, 625–634. doi:10.1007/s10654-009-9366-3

32.

van Dessel

den Boeft

van der Wouden

J. C.

Kleinstäuber

Leone

S. S.

Terluin

van Marwijk

. (2014). Non-pharmacological interventions for somatoform disorders and medically unexplained physical symptoms (MUPS) in adults. Cochrane Database of Systematic Reviews. Retrieved from https://www.cochranelibrary.com/cdsr/doi/10.1002/14651858.CD011142.pub2/full

33.

van Geelen

S. M.

Rydelius

P.-A.

Hagquist

. (2015). Somatic symptoms and psychological concerns in a general adolescent population: Exploring the relevance of DSM-5 somatic symptom disorder. Journal of Psychosomatic Research, 79, 251–258. https://doi.org/10.1016/j.jpsychores.2015.07.012

34.

Wallston

K. A.

Wallston

B. S.

DeVellis

(1978). Development of the Multidimensional Health Locus of Control (MHLC) scales. Health Education Monographs, 6, 160–170. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/689890