Anthroposophic therapy for children with chronic disease: a two-year prospective cohort study in routine outpatient settings

Background Many children with chronic disease use complementary therapies. Anthroposophic treatment for paediatric chronic disease is provided by physicians and differs from conventional treatment in the use of special therapies (art therapy, eurythmy movement exercises, rhythmical massage therapy) and special medications. We studied clinical outcomes in children with chronic diseases under anthroposophic treatment in routine outpatient settings. Methods In conjunction with a health benefit program, consecutive outpatients starting anthroposophic treatment for any chronic disease participated in a prospective cohort study. Main outcome was disease severity (Disease and Symptom Scores, physicians' and caregivers' assessment on numerical rating scales 0–10). Disease Score was documented after 0, 6, and 12 months, Symptom Score after 0, 3, 6, 12, 18, and 24 months. Results A total of 435 patients were included. Mean age was 8.2 years (standard deviation 3.3, range 1.0–16.9 years). Most common indications were mental disorders (46.2% of patients; primarily hyperkinetic, emotional, and developmental disorders), respiratory disorders (14.0%), and neurological disorders (5.7%). Median disease duration at baseline was 3.0 years (interquartile range 1.0–5.0 years). The anthroposophic treatment modalities used were medications (69.2% of patients), eurythmy therapy (54.7%), art therapy (11.3%), and rhythmical massage therapy (6.7%). Median number of eurythmy/art/massage therapy sessions was 12 (interquartile range 10–20), median therapy duration was 118 days (interquartile range 78–189 days). From baseline to six-month follow-up, Disease Score improved by average 3.00 points (95% confidence interval 2.76–3.24 points, p < 0.001) and Symptom Score improved by 2.41 points (95% confidence interval 2.16–2.66 points, p < 0.001). These improvements were maintained until the last follow-up. Symptom Score improved similarly in patients not using adjunctive non-anthroposophic therapies within the first six study months. Conclusion Children under anthroposophic treatment had long-term improvement of chronic disease symptoms. Although the pre-post design of the present study does not allow for conclusions about comparative effectiveness, study findings suggest that anthroposophic therapies may play a beneficial role in the long-term care of children with chronic illness.


Background
Chronic illness affects 15%-18% of children [1] and can lead to functional limitation, dependency on therapies and medication, poor school performance, and impaired quality of life [1][2][3][4]. Strategies to improve the outcome of chronic paediatric illness include enhanced healthcare provision [5] and special educational and behavioural interventions [6][7][8]. Many children with chronic disease also receive complementary therapies [9][10][11][12][13], sometimes provided by their physicians. In Germany, several physician-provided complementary therapies have been reimbursed by health insurance companies as part of special health benefit programs [14][15][16][17]. In most of these complementary therapies the physician is the active person, directly treating (e. g. giving acupuncture) or prescribing therapy (e. g. homoeopathic medications), while the patient has a predominantly passive role. Anthroposophic medicine (AM), a complementary system of medicine founded by Rudolf Steiner and Ita Wegman [18], includes active (AM art and eurythmy therapy) as well as passive therapy modalities (massage, medications).
In AM art therapy the patients engage in painting, drawing, clay modelling, music or speech exercises [19]. In addition to psychological effects (e. g. activation, emotive expression, dialogical communication with the therapist and with the artistic medium) [20,21], AM art therapy can induce physiological effects: e. g. AM speech exercises have effects on heart rate rhythmicity and cardiorespiratory synchronization which are not induced by spontaneous or controlled breathing alone [22,23].
Eurythmy therapy (Greek: eurythmy = "harmonious rhythm") is an active exercise therapy, involving cognitive, emotional, and volitional elements [24]. During eurythmy therapy sessions the patients are instructed to perform specific movements with the hands, the feet or the whole body. Eurythmy movements are related to the sounds of vowels and consonants, to music intervals or to soul gestures, e. g. sympathy-antipathy [25]. Eurythmy exercises have specific effects on heart rate variability [26]. Qualification as an AM art or eurythmy therapist requires six years' training according to an international, standardised curriculum.
Rhythmical massage therapy was developed from Swedish massage by Ita Wegman, physician and physiotherapist [27], and is practiced by physiotherapists with 1 1/2-3 years specialised training. In rhythmical massage therapy, traditional massage techniques (effleurage, petrissage, friction, tapotement, vibration) are supplemented by gentle lifting and rhythmically undulating, stroking movements, where the quality of grip and emphasis of movement are altered to promote specific effects [28].
AM medications are of mineral, botanical or zoological origin or are chemically defined substances [29].
Prior to prescription of AM medications or referral to AM therapies, AM physicians have prolonged consultations with the patients and their caregivers. These consultations are used to take an extended history, to address constitutional and psychosocial aspect of the patients' illness, to explore the patients' and caregivers' preparedness to engage in treatment, and to select optimal therapy for each patient [19,28].
AM is practiced in 67 countries [30]. Related to the AM approach is an educational philosophy implemented in more than 3,000 Waldorf Schools, kindergartens, and curative education centres worldwide [31,32]. Waldorf school attendance has been associated with a reduced risk for atopic disease [33,34], possibly mediated by effects on the intestinal microflora from restrictive use of antibiotics and antipyretics in childhood infectious disease [34] or from a vegetarian diet [35].

Study design and objective
This is a prospective cohort study in a real-world medical setting. The study was part of a research project on the effectiveness, costs, and safety of AM therapies in outpatients with chronic disease (Anthroposophic Medicine Outcomes Study, AMOS) [15,46,47]. The AMOS project was initiated by a health insurance company in conjunction with a health benefit program. The present preplanned analysis concerned the subgroup of children with chronic diseases. The primary research question concerned the range of indications for AM therapy as well as the outcome of disease symptoms. Further research questions concerned quality of life, use of adjunctive non-AM therapies, therapy satisfaction, and adverse reactions.

Setting, participants, and therapy
All physicians certified by the Physicians' Association for Anthroposophical Medicine in Germany and working in an office-based practice or outpatient clinic were invited to participate in the AMOS study. Certification as an AM physician required a completed medical degree and a three-year structured postgraduate training. The participating physicians recruited consecutive patients starting AM therapy under routine clinical conditions. Patients enrolled in the period 1 January 1999 to 31 December 2005 were included in the present analysis if they fulfilled the eligibility criteria. Inclusion criteria were: 1. Outpatients aged 1-16 years.
2. AM-related consultation of at least 30 minutes followed by prescription of AM medication or referral to AM therapy (art, eurythmy or rhythmical massage) for any indication (main diagnosis).
3. Duration of main diagnosis of at least 30 days at study enrolment.
Patients were excluded if they had previously received the AM therapy in question (see inclusion criteria no. 2) for their main diagnosis. AM therapy was evaluated as a whole system [48] with subgroup analysis of evaluable therapy modality groups.

Primary outcome
Primary outcome was disease severity at six-month follow-up. Disease severity was assessed on numerical rating scales [49] from 0 ("not present") to 10 ("worst possible"): Disease Score (physicians' global assessment of severity of main diagnosis); Symptom Score (caregivers' assessment of severity of one to six most relevant symptoms present at baseline). Disease Score was documented after 0 and 6 months, Symptom Score and quality of life (see below) after 0, 3, 6, 12, 18, and 24 months. Therapy outcome rating (0-10), satisfaction with therapy (0-10), and therapy effectiveness rating ("very effective, "effective", "less effective", "ineffective" or "not evaluable") were documented by the caregivers (effectiveness rating also by the physicians) after 6 and 12 months. Adverse reactions to medications or therapies were documented by the caregivers after 6, 12, 18, and 24 months, and by the physicians after 6 months (for patients enrolled before 1 April 2001 also after 3, 9, and 12 months). The documentation included suspected cause, intensity (mild/moderate/severe = no/some/complete impairment of normal daily activities), and therapy withdrawal because of adverse reactions. Serious adverse events (death, life-threatening condition, acute in-patient hospitalization, new disease or accident causing permanent disability, congenital anomaly, new malignancy) were documented by the physicians throughout the study.

Data collection
All data were documented with questionnaires returned in sealed envelopes to the study office. The physicians documented eligibility criteria; the therapists documented AM therapy administration; all other items were documented by the caregivers or patients unless otherwise stated. The patient responses were not made available to the physicians. The physicians were compensated 40 Euro (after March 2001: 60 Euro) per included and fully documented patient, while the patients and their caregivers received no compensation.
The data were entered twice by two different persons into Microsoft ® Access 97. The two datasets were compared and discrepancies resolved by checking with the original data.

Quality assurance, adherence to regulations
The study was approved by the Ethics Committee of the Faculty of Medicine Charité, Humboldt University, Berlin, Germany, and was conducted according to the Declaration of Helsinki and largely following the ICH Guideline for Good Clinical Practice E6. Written informed consent was obtained from all patients before enrolment.

Data analysis
The data analysis was performed on all patients fulfilling the eligibility criteria, using SPSS ® 14.0.1 (SPSS Inc., Chicago, Ill, USA) and StatXact ® 5.0.3 (Cytel Software Corporation, Cambridge, MA, USA). Diagnoses were coded according to the International Classification of Diseases, Tenth Revision (ICD-10).
For continuous data the two-tailed t-test was used. For binominal data the two-tailed McNemar test and Fisher's exact test were used. Significance criteria were p < 0.05 or 95% confidence interval (95%-CI) not including 0. Since this was a descriptive study, no adjustment for multiple comparisons was performed [53].
Pre-post effect sizes were calculated as Standardised Response Mean (= mean change score divided by the standard deviation of the change score) and classified as minimal (< 0.20), small (0.20-0.49), medium (0.50-0.79), and large (³ 0.80) [54,55]. In the main analysis, clinical outcomes were analysed in patients with evaluable data for each follow-up, without replacement of missing values.
Three pre-planned sensitivity analyses (SA1-SA3) were performed to assess the influence of patient attrition, natural recovery, and adjunctive diagnosis-relevant non-AM therapies on the 0-6-month Symptom Score outcome. SA1 concerned attrition bias: Missing values after six months were replaced with the last value carried forward. SA2 concerned natural recovery, which was assumed to be unlikely in AMOS patients with disease duration of at least one year [56]: The sample was therefore restricted to patients with disease duration of at least 12 months prior to study enrolment. SA3 concerned the effects of relevant non-AM therapies, and was performed on diagnosis groups with at least 15 evaluable patients (mental, respiratory or musculoskeletal diseases, headache syndromes, urinary incontinence). In SA3 this sample was restricted to patients not using diagnosis-related non-AM therapies during the first six study months (listed in Table 1).
Stepwise multiple linear regression analysis was performed to identify predictors of Symptom Score change from baseline after 6 and 12 months. Criterion for inclu-sion of variables in the model was p < 0.05 and for exclusion p ³ 0.10. The following independent variables were analysed: • Socio-demographics: age, gender, household size, living with father, visiting Waldorf School, health insurance coverage, year of enrolment.
• Disease status at baseline: diagnosis (five categories), disease duration, disease severity, baseline Symptom Score, number of comorbid disorders, therapies in the preceding year (number of AM therapy sessions, number of patient-months of AM medication and non-AM medications, respectively, number of sessions with physiotherapy/occupational therapy/play therapy).
• Therapy factors: physician setting (primary care or other), physician qualification (general practitioner, paediatrician), number of years since the physician's medical qualification, previous treatment by the physician, number of patients enrolled by the physician, duration of consultation with the physician at study enrolment, main AM therapy modality (eurythmy, art, rhythmical massage, medical), number of years since the AM therapist's qualification, reimbursement of costs of AM therapies, therapies in months 0-6 Missing values for independent variables were replaced by the respective mean values. Model assumptions for linear regression were checked and verified. A few outliers with studentised residuals ³ 3 standard deviations were identified and excluded from the analyses (see Results for exceptions).

Participating physicians and therapists
The patients were enrolled by 85 physicians with six different qualifications (57 general practitioners, 20 paediatricians, four internists, two otorhinolaryngologists, one gynaecologist, and one psychiatrist). Comparing these physicians to AM-certified physicians in Germany with the same six qualifications but without study patients (n = 295), no significant differences were found regarding gender (63.5% vs. 60.3% men), age (mean 46.6 ± 6.5 vs. 48 Respondents (n = 384) and non-respondents (n = 51) of the six-month patient-follow-up did not differ significantly regarding age, gender, diagnosis, disease duration, baseline Disease Score or baseline Symptom Score. Corresponding comparisons of respondents (n = 314) and nonrespondents (n = 112) of the 24-month-follow-up also showed no significant differences, except for baseline Disease Score which was 6.4 ± 1.8 points in respondents and 6.9 ± 1.7 points in non-respondents (mean difference 0.6 points, 95%-CI 0.2-0.9 points, p = 0.005). The physicians' six-month follow-up documentation was available for 89.7% (n = 390/435) of patients.  0.0-2.0) comorbid diseases. The most common comorbid diseases were R00-R99 Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified (19.7%, n = 90/458 diagnoses), F00-F99 Mental and behavioural disorders (17.5%), J00-J99 Diseases of the respiratory system (13.8%), L00-L99 Diseases of the skin and subcutaneous tissue (9.6%), and D50-D89 Diseases of the blood and blood-forming organs and certain disorders involving the immune mechanism (6.1%).

Baseline characteristics
Further baseline data are presented in Table 2.

Therapy
At study enrolment, the duration of the consultation with the AM physician was < 30 min in 53. Use of diagnosis-related non-AM therapies within the first six study months was analysed in patients with a main diagnosis of mental, respiratory or musculoskeletal diseases, headache syndromes (ICD-10 G43-G44, R51) or urinary incontinence, altogether n = 335 patients. Patients were classified as users if they had used at least one of the listed therapies (Table 1) for at least one day per month. Out of 296 evaluable patients, 82.1% (n = 243) had no diagnosis-related non-AM therapy.

Primary outcome
Disease and Symptom Scores (Figure 1) improved significantly and progressively between baseline and all subsequent follow-ups. After six months, an improvement of ³ 50% of baseline scores was observed in 46.7% (n = 169/ 362) and 41.7% (n = 159/381) of evaluable patients for Disease and Symptom Scores, respectively. Standardised Response Mean effect sizes for the 0-6 month comparison were large for both scores (1.30 and 0.97, respectively). Symptom Score improved significantly in all four therapy modality groups (medical, eurythmy, art, rhythmical massage) ( Table 3).
We performed three sensitivity analyses of the 0-6 month outcome of Symptom Score ( Predictors of Symptom Score improvement from baseline after 6 and 12 months were identified by stepwise multiple linear regression analysis (see Methods for details).
The models for 0-6 month and 0-12 month improvement explained 23% and 26% of the variance, respectively (Table 5). Three variables were significant predictors in both models: • Symptom Score at baseline: For each 1.00 point increase in baseline Symptom Score (increase means worse symptoms), the Symptom Score improvement will increase by average 0.67 and 0.71 points after 6 and 12 months, respectively. Baseline Symptom Score was the strongest predictor, explaining 17% and 20% of the variance after 6 and 12 months, respectively.
• Disease Score at baseline: For each 1.00 point increase in baseline Disease Score (increase means more severe disease), the Symptom Score improvement will decrease by average 0.18 points and 0.21 points after 6 and 12 months, respectively.
• Disease duration: For each year of disease duration prior to study enrolment, the improvement will decrease by 0.10 and 0.14 points after 6 and 12 months, respectively.
In addition, the 0-6 month improvement was positively predicted by the variables 'main AM therapy modality medical' and 'previous treatment by physician' (these two predictions disappeared when one identified outlier was included in the analysis, see Methods for details) and negatively predicted by the number of sessions with physiotherapy, occupational therapy or play therapy in the previous year, while the 0-12 month improvement was positively predicted by enrolment by a general practitioner and negatively predicted by a main diagnosis of F00-F99 Mental and behavioural disorders.

Secondary outcomes
The quality of life scores in all age groups improved significantly between baseline and nearly all subsequent follow-ups (23 significant and 2 non-significant improvements in 25 pre-post comparisons). Effect sizes for the 0-6 month comparison were medium for KITA Psychosoma (0.52) and small for the remaining four scores (range 0.29-0.41, Table 3).
At six-month follow-up, the caregivers' average therapy outcome rating (numeric scale from 0 "no help at all" to 10 "helped very well") was 7.12 ± 2.56 and the caregivers' satisfaction with therapy (from 0 "very dissatisfied" to 10 "very satisfied") was 7.87 ± 2.27.
Serious adverse events occurred in five patients. Four patients were acutely hospitalised and one patient had permanent disability from a whiplash injury. None of these events were causally related to any therapy or medication.

Discussion
This two-year prospective cohort study is the first large study of AM therapy for children with chronic disease performed in office-based settings. We aimed to obtain information on AM therapy under routine conditions in Germany and studied the disease spectrum and clinical outcomes in children aged 1-16 years starting AM treatment for chronic diseases. Most frequent indications were mental/behavioural disorders (primarily hyperkinetic, emotional, and developmental disorders) and respiratory diseases. Following AM therapy, statistically significant improvements of disease symptoms and quality of life  were observed. The symptom improvement was similar in patients not using diagnosis-relevant adjunctive non-AM therapies. A strong positive predictor of improvement after 6 and 12 months was higher baseline symptom intensity (caregivers' rating); weaker and negative predictors were higher disease severity (physicians' rating) and longer disease duration. Adverse reactions to AM therapies and medications were infrequent and not serious.
Strengths of this study include a detailed assessment of the therapy setting and therapy-related factors, a long follow-up period, and a high representativeness due to the participation of 14% of eligible AM physicians and AM therapists in Germany. The participating physicians and therapists resembled eligible but not participating AM physicians and AM therapists with respect to demographic characteristics, and the included patients resembled not included patients regarding baseline characteristics (except for a possible over-representation of mental disorders). These features suggest that the study to a high degree mirrors contemporary AM use in outpatient settings.
A limitation of the study is the absence of a comparison group receiving conventional treatment or no therapy. Accordingly, for the observed improvements one has to consider several other causes apart from the AM treatment. We therefore conducted a sensitivity analysis of Symptom Score, estimating the influence of attrition bias, adjunctive non-AM therapies, and natural recovery. These three factors together explained only 5% of the improvement. According to a previous analysis from this research program [56], regression to the mean due to symptom fluctuation with preferential self-selection to therapy and study inclusion at symptom peaks explained up till 0.43 points (14%) of the improvement of Disease Score, which would correspond to approximately 18% of the Symptom Score improvement in the present analysis. Other possible confounders are psychological factors and non-specific effects. However, since AM therapy was evaluated as a whole system [48], the question of specific therapy effects vs. non-specific effects (placebo effects, context effects,  Since patients were treated by AM physicians who could possibly have an interest in AM therapy having favourable outcomes, study data were largely collected by the patients and not the physicians. Any bias affecting the physicians' documentation would not affect Symptom Score or quality of life, since these clinical outcomes were documented by the patients or caregivers. This analysis assessed AM as a whole system [48], with subgroup analyses of major therapy modality groups. In the bivariate analyses, symptom improvement was more outspoken among patients receiving AM medical therapy and less outspoken in the rhythmical massage therapy group (Table 3), but this was not reliably confirmed in the multivariate predictor analyses. Other measures of variability of AM treatment, such as the duration of the initial consultation or the number of AM therapy sessions, did not predict symptom improvement.
The strongest predictor of symptom improvement was baseline symptom severity as rated by caregivers. This finding can have several causes, such as more room for improvement and more regression to the mean with higher score values, the hello-goodbye effect, and a higher patient motivation with therapies working better at higher symptom levels [55,56]. Baseline disease severity, rated by the physicians, predicted future improvement in the opposite direction than baseline symptoms (i. e. lower disease severity predicted more improvement while lower  symptom intensity predicted less improvement). A possible explanation for this seeming paradox is that physicians incorporate medical knowledge about the patients' prognosis into their severity rating, while caregivers focus more strongly on the symptoms. -A limitation of the predictor analysis is that parental education and occupational levels as well as household income could not be assessed, since these items were not documented for children. In a corresponding predictor analysis in adult patients from the AMOS project, these factors were not associated with symptom improvement (Hamre et al, submitted for publication).
Previous studies have evaluated AM therapy for children with chronic disorders including anorexia nervosa [36,38], atopic diseases [37,43], hyperkinetic disorders [24,45], epilepsy [44], hepatitis B [39], and immune suppression with recurrent respiratory infections following radioactive exposure after the Chernobyl nuclear accident [40][41][42]. All these studies had some favourable outcomes; the three largest studies (range 79-125 AM patients) found high anorexia cure rates [36], reduced infection rates and normalisation of immune parameters in Chernobyl children [42], and successful epilepsy therapy without conventional anticonvulsive drugs [44]. Except for a pilot study with five patients [24], these studies were performed in inpatient hospitals [36][37][38][39] or outpatient clinics [40][41][42][43][44][45]. In accordance with these studies from secondary care, our study from a predominantly primary care setting showed significant improvements of mental/behavioural, respiratory, and other chronic diseases in children following AM treatment. The largest improvements (large effect sizes, half of patients improved by at least 50% of their baseline scores) were observed for the items which directly measure the conditions treated with AM, i. e. Disease and Symptom Scores.
This study also underlines the role of naturalistic outcome studies in the evaluation of complementary therapies and other complex therapy systems [57,58]. For such outcome studies, several features are called for [48,[57][58][59][60], which are found in the present study: primary assessment of the whole therapy complex and secondary assessment of therapy components; recruitment of patients in the setting where they are usually treated (here: predominantly primary care); documentation of routine therapy practice while minimising distortion from experimental study conditions; widely used outcomes (here: numerical rating scales); long-term follow-up; and, especially in the case of single-arm studies, a systematic approach to assess and minimise bias.

Conclusion
In this study, children under AM treatment for mental/ behavioural, respiratory, and other chronic diseases had long-term reduction of disease severity and improvement of quality of life. Improvements were similar in patients not using adjunctive non-AM therapies. Although the prepost design of the present study does not allow for conclusions about comparative effectiveness, study findings suggest that AM therapies may play a beneficial role in the long-term care of children with chronic illness.