JLE

Epileptic Disorders

MENU

Correlation between child and parental perceptions of health-related quality of life in epilepsy using the PedsQL.v4.0 measurement model Volume 12, numéro 4, December 2010

Auteur(s) : Zulfi Haneef1,5, Mitzie L Grant2,3,4, Ignacio Valencia3,4,5, Elizabeth F Hobdell4, Sanjeev V Kothare4,6, Agustin Legido3,4,5, Divya Khurana3,4,5

1Department of Neurology, Temple University Hospital
2Department of Psychiatry, Drexel University College of Medicine
3Department of Pediatrics, Drexel University College of Medicine
4Department of Pediatrics, St. Christopher's Hospital for Children
5Department of Neurology, Drexel University College of Medicine, Philadelphia, Pennsylvania
6Department of Neurology, Children's Hospital, Boston, Massachusetts, USA

Article reçu le 11 Ao�t 2010, accepté le 7 Octobre 2010

Epilepsy is a common disorder with about 100,000 cases identified annually in the United States (Annegers, 1994). Although traditional medical care stresses adequate seizure control with AEDs, it is increasingly recognized that a holistic approach should consider the impact on a patient's quality of life (QOL) (Jacoby, 1992; Norrby et al., 1999). Research has shown that seizure frequency and duration affect QOL (Jacoby et al., 1996; Au et al., 2002). Many non-epilepsy factors also affect QOL for treatment-resistant epilepsy, including medical and psychiatric comorbidity, social factors, and self-esteem (Schachter, 2005). Available measures to evaluate health-related quality of life (HRQOL) in children with epilepsy are limited, with some confined to a particular domain providing a limited HRQOL profile (Batzel et al., 1991; Hoare and Russell, 1995; Gilliam et al., 1997). Some are applicable only to a particular age group such as adolescents (Batzel et al., 1991; Cramer et al., 1999). Others are specific to either parent proxy-reporting (Hoare and Russell, 1995; Carpay et al., 1996; Sabaz et al., 2000; Camfield et al., 2001) or child self-report (Batzel et al., 1991; Cramer et al., 1999). Only two groups have developed epilepsy-specific instruments with both parent proxy and child self-reporting of HRQOL scores (Arunkumar et al., 2000; Ronen et al., 2003).

When the child is too young or cognitively impaired to complete an HRQOL instrument, parent proxy-reporting becomes important. Comparing the child and parental perceptions of HRQOL becomes significant in this context. Few studies have compared child and parental perceptions of HRQOL using readily available standardized instruments. The use of a generic HRQOL instrument in paediatric epilepsy comparing child self-report with simultaneous parent proxy-reporting has not been previously studied. A recent study examined the effects of single seizures and newly diagnosed untreated epilepsy on the parent proxy HRQOL scores using the PedsQL (Pediatric Quality of Life Inventory) measurement model, but child HRQOL scores were not investigated (Modi et al., 2009). The PedsQL.v4.0 (PedsQL, version 4) is a brief, standardized, generic, modular, and multidimensional instrument designed for use in healthy children as well as children with acute and chronic health conditions (Varni et al., 1999). Our study investigated both parental and child self-perceptions of QOL in children with epilepsy using the PedsQL.v4.0 model.

Methods

The study was approved by the Institutional Review Board at Drexel University College of Medicine and was performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki.

Recruitment

This study included children between 2 and 18 years of age with epilepsy diagnosed more than 3 months previously. Epilepsy was defined as the occurrence of two or more unprovoked seizures and was further categorized as refractory or well-controlled. Children with well-controlled epilepsy were those who were seizure-free for at least six months on a single anticonvulsant. Children with refractory epilepsy were those who had at least one seizure per month after three trials of AEDs. The study was performed in a children's hospital serving predominantly families from a socioeconomically disadvantaged inner-city region. In compliance with the hospital's institutional review board, after informed consent and assent were obtained, children and their families attending the neurology clinic between March 2003 and December 2004 completed a brief demographic questionnaire and the PedsQL.v4.0 questionnaire. If the primary caregiver was not the parent, the parent proxy-reporting questionnaire was given to the primary caregiver for completion. For half of the group, a study coordinator read the questionnaires to the parent and child. For the other half, parent and child completed the questionnaires during the office visit. In a few instances, families were permitted to return the questionnaire by mail when time was insufficient to complete it during the office visit.

Information regarding socioeconomic status and duration of diagnosis and medications was collected from the clinic chart. The seizure type, epilepsy syndrome, and other concomitant neurological conditions were diagnosed by the treating physician. Information was abstracted from the chart concerning neurological and psychiatric comorbid conditions (NPCCs) such as attention deficit disorder, learning disabilities, mental retardation, and psychiatric conditions such as depression or aggressive behaviours. Diagnoses were made using Diagnostic and Statistical Manual-IV criteria by the treating physician or were pre-existing diagnoses made by psychiatrists whom the family had consulted or who were employed by the child's school district.

Measurements

QOL measure

Age-appropriate versions of the PedsQL were administered to parents of children 2-18 years of age and to children 5-18 years of age. The questionnaires consisted of 23 items covering four generic core scales, each with multiple items (physical: eight items, emotional: five items, social: five items, and school functioning: three items for toddlers/five items for older children). Each item was rated on a scale from 0 to 4 and scored according to the guidelines established by Varni et al. (2003). Items were subsequently reverse-scored and linearly transformed to a 0- to 100-point scale (0 = 100, 1 = 75, 2 = 50, 3 = 25, 4 = 0) with higher converted scores indicating a better quality of life.

PedsQL calculates two summary scores from the above measurements. The physical health summary score is the same as the physical functioning subscale (eight items), and the psychosocial health summary score is the mean of the emotional, social, and school functioning subscales (15 items). The total summary score was calculated from the mean of all answered items.

HRQOL scores from the current sample were compared with normative values previously reported by Varni et al. (2003) for healthy children (5,079 children, 8,713 parents) and children with chronic illness (574 children, 831 parents) including asthma, attention deficit hyperactivity disorder, depression, and diabetes. This population aged 2-16 years was sampled by mail survey whereas children aged 8-16 years self-administered the questionnaire and parents assisted with administration for younger children.

Socioeconomic status score

A socioeconomic status score was calculated based on the occupational status and educational status of the mother and father using the four-factor index of social status by Hollingshead (1975, unpublished). The occupational status of the parent was scored from 1 to 9 and the educational status was scored from 1 to 7. A retired person was scored as if he or she were still working. The social score was computed from the above variables by a formula that gives a factor weight of 5 for occupation and a factor weight of 3 for education as follows: social status score = [(occupation scale score) × 5] + [(education scale score) × 3]. If both parents were working, the social status score was the average of the individual scores. If only one partner was working, the social status score of the working spouse was taken. The score ranges from 8 to 67 and a higher score reflects a better socioeconomic status.

Data analysis

Data were analyzed using SPSS version 15.0 (Chicago, IL). Pearson's test was performed to examine correlation between HRQOL measures and parental age, age at seizure onset, length of diagnosis, duration of an average seizure, seizure frequency, and the socioeconomic status as reflected by the Hollingshead score. The Kruskal-Wallis test was used to compare the HRQOL measures with the number of NPCCs and the number of AEDs. Mann-Whitney U tests examined the association of the individual NPCCs specified earlier and the severity of epilepsy (well-controlled versus refractory) with the outcome HRQOL measures. Student t-tests were used to look for differences in HRQOL perceptions between boys and girls and to compare the means of the various individual, summary, and total HRQOL scores of our sample with the previously reported means for healthy children and for children with chronic health conditions. The HRQOL scores of children with well-controlled epilepsy with and without NPCCs and children with refractory epilepsy were compared using analysis of variance (ANOVA) and t-tests for post hoc analysis. We used GraphPad software (GraphPad Software, Inc., La Jolla, CA) for comparing our population to children with chronic health conditions and healthy children. A significant effect was shown for the method of HRQOL administration (self-report versus coordinator-assisted report completion; p < 0.03), with coordinator-assisted scores higher across all groups than self-report scores (well-controlled epilepsy with and without NPCCs and children with refractory epilepsy). Results of analysis of covariance (ANCOVA), controlling for method of administration, remained consistent with those obtained from ANOVA.

Results

A total of 106 parent proxy PedsQL.v4.0 questionnaires were administered of which six were incomplete and were excluded. Sixty-seven children completed the self-report PedsQL.v4.0. Nine were too young and 24 were otherwise unable to complete the measure. The demographic variables and seizure characteristics are listed in table 1.

NPCCs seen in our patient population included attention deficit hyperactivity disorder (n = 14, 12.7%), learning disability (n = 16, 14.5%), global delay/mental retardation (n = 32, 29.1%), autism (n = 6, 5.5%), motor problems (n = 18, 16.4%), and other psychiatric diagnoses (n = 4, 3.6%). Forty-six children had no NPCCs, 27 had one NPCC, 25 had two, and two had three NPCCs. All children with refractory seizures (n = 20) had NPCCs. The sample with well-controlled epilepsy (n = 80) was subdivided into those with (n = 46) and those without (n = 34) NPCCs.

According to the Hollingshead score described previously, the median score for maternal education was 4, with an interquartile range (IQR) of 4-5, and median score for paternal education was 5 (IQR 4-6). The median maternal as well as paternal occupational scale score was 5 (IQR 4-6). The mean social status score for both parents combined was 35.3 (standard deviation, 12.9).

Correlations of PedsQL.v4.0 data between 67 pairs of parent and child reports were highly significant for total summary scores as well as for the physical health and psychosocial health summary scores and emotional, social, and school functioning subscales (p < 0.001), reflecting a high degree of concordance across all domains (table 4). The significant results obtained for the child self-reports and parent proxy-reports are detailed separately below.

Child self-report measure

A descriptive analysis of the PedsQL.v4.0 scores for both child and parent is given in table 2, along with comparisons to normative data for groups of healthy controls and children with chronic illness (Varni et al., 2003). Compared with scores from healthy controls, the HRQOL scores in our subjects were significantly lower (p < 0.001) across all domains. Compared with the control sample with other chronic health conditions, our sample of children with epilepsy had significantly lower school functioning (p < 0.05), physical health (p < 0.01), and total (p < 0.05) scores. Owing to NPCCs (i.e. mental retardation/global developmental delay), only four of the 20 children with refractory seizures were able to complete the PedsQL.v4.0 measure, precluding separate analysis of this group.

Among children with well-controlled epilepsy, those with NPCCs had significantly lower scores across all domains (p < 0.05, p < 0.001) (table 3). Table 4 shows the correlation of demographic variables and seizure characteristics with the HRQOL scores. Duration of epilepsy was negatively correlated with PedsQL.v4.0 total score (r = -0.35, p < 0.01). The socioeconomic status of the family (Hollingshead) and parental age correlated positively with the child's psychosocial and total summary scores. Age at seizure onset, seizure frequency, and duration of diagnosis did not correlate significantly with the child's self-reported QOL scores.
Table 1 Demographics and seizure characteristics.

Demographics (n = 100)

Child

Parent

Mother

Father

Age (y)

- Mean - SD - Range

10.1 4.2 2.1-18.8

39.2 7.5 27-60

38.6 7.1

39.8 7.9

Male gender

59 (50.9%)

Hollingshead social scores

- Mean - SD - Range

35.3 12.9 9-66

36.4 14.0 9-61

35.5 9-61

Seizure characteristics

Child

Age at onset (y)

- Mean ± SD (range)

5.2 ± 4.4 (0-16)

Duration of epilepsy (y)

- Mean ± SD (range)

4.8 ± 4.3 (0.3-18.2)

Median number of medications (range)

2 (1-11)

Number of hospitalizations in past year

- Mean ± SD (range)

1.7 ± 3.1 (0-20)

Seizure severity

- Well-controlled - Refractory

80 (80%) 20 (20%)

Seizure type

- Generalised tonic-clonic - Absence - Complex partial - Infantile spasm - Mixed/Lennox-Gastaut - Other

21 (20.0%) 13 (12.4%) 47 (44.8%) 1 (1.0%) 16 (15.2%) 7 (6.7%)


Table 4 Correlation between HRQOL measures and seizure characteristics.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

1. Parent PsSS

-

2. Parent PhSS

0.59**

-

3. Parent total

0.92**

0.85**

-

4. Child PsSS

0.69**

0.6**

0.71**

-

5. Child PhSS

0.57**

0.71**

0.68**

0.79**

-

6. Child total

0.68**

0.68**

0.73**

0.97**

0.91**

-

7. Aggregate total

0.87**

0.83**

0.95**

0.91**

0.86**

0.94**

-

8. Child's age

- 0.07

0.04

- 0.02

0.09

0.01

0.06

- 0.05

-

9. Mother's age

0.14

0.20

0.18

0.31*

0.36*

0.34*

0.23

0.36**

-

10. Father's age

0.41**

0.39**

0.43**

0.42*

0.32

0.41*

0.47**

0.42**

0.84**

-

11. Age at seizure onset

0.26*

0.3**

0.31**

0.22

0.15

0.20

0.3**

0.51**

0.45**

0.45**

-

12. Duration of epilepsy diagnosis

- 0.33**

- 0.24*

- 0.33**

- 0.16

- 0.16

- 0.17

- 0.35**

0.45**

- 0.19

- 0.21

- 0.54**

-

13. Length of average seizure

0.01

- 0.01

0.01

- 0.25*

- 0.3*

- 0.28*

- 0.03

- 0.15

0.01

- 0.08

- 0.01

- 0.13

-

14. Length of longest seizure

- 0.15

- 0.10

- 0.14

- 0.16

- 0.18

- 0.17

- 0.16

0.18

0.01

- 0.25

- 0.17

0.37**

0.07

-

15. SES

0.18

0.14

0.18

0.3*

0.19

0.26*

0.21*

- 0.03

0.41**

0.56**

0.22*

- 0.21*

- 0.17

- 0.05


Table 2 PedsQL scores for children with well-controlled epilepsy, chronic illnesses, and healthy controls.

Child self-report

Epilepsy (n = 67)

Healthy control (n ~5,070)

Chronic illness (n ~573)

Emotional

64.9 ± 25.7

79.2 ± 18.0

69.3 ± 21.4

Social

73.2 ± 28.5

85.0 ± 16.7

76.4 ± 21.6

School

62.8 ± 24.2

81.3 ± 16.1

68.3 ± 19.1*

Physical

73.2 ± 27.3

87.8 ± 13.1

79.5 ± 17.1**

Total

69.4 ± 23.3

83.9 ± 12.5

74.2 ± 15.4*

Parental proxy-report

Epilepsy (n = 80)

Healthy control (n ~8,690)

Chronic illness (n ~830)

Emotional

64.7 ± 24.3

81.2 ± 16.4

71.1 ± 19.8**

Social

68.9 ± 27.1

83.1 ± 19.7

75.1 ± 21.8*

School

64.7 ± 23.5

78.3 ± 19.6

65.6 ± 20.8

Physical

69.3 ± 29.2

84.1 ± 19.7

77.0 ± 20.2**

Total

67.3 ± 22.8

82.3 ± 15.6

73.1 ± 16.5**


Table 3 Comparison of PedsQL scores of children with well-controlled epilepsy with and without comorbidity and refractory epilepsy.

Parent proxy-report scores

Group 1 (n = 46)

Group 2 (n = 34)

Group 3 (n = 20)

Emotional

69.2 ± 21.8

58.6 ± 26.4*

57.9 ± 26.0

Social

77.8 ± 25.1

56.4 ± 25.0

39.5 ± 25.2

School

70.7 ± 25.6

56.6 ± 17.8**

43.3 ± 25.8

Physical

79.9 ± 24.0

54.5 ± 29.6

51.1 ± 33.4

Total aggregate

75.5 ± 21.1

56.1 ± 20.3

48.0 ± 14.7

Child self-report scores

Group 1 (n = 42)

Group 2 (n = 21)

Emotional

70.1 ± 23.8

54.5 ± 26.7*

Social

80.2 ± 25.4

59.2 ± 29.9**

School

70.0 ± 21.4

49.0 ± 23.6

Physical

79.9 ± 24.8

60.3 ± 27.8**

Total aggregate

75.9 ± 20.8

57.0 ± 23.0

Parent proxy-report

Parental PedsQL.v4.0 scores for children with epilepsy differed significantly from comparative normative data from healthy children (p < 0.001) across all individual domain scales and for total composite scores (table 2). In comparison with the PedsQL scores for children with chronic medical conditions, parents of children with well-controlled epilepsy rated their child's emotional, social, physical, and total QOL scores as significantly lower (p < 0.05, p < 0.01). No significant differences were found between the two groups on the school functioning score.

The parents of children with refractory epilepsy rated the overall QOL of their children as significantly below (p < 0.01) the level reported for children with well-controlled epilepsy, except for emotional functioning (table 3). These two groups did not differ in gender (p = 0.33), age (p = 0.34), or seizure frequency (p = 0.30) but differed in age at seizure onset and duration of epilepsy (p < 0.001).

Parental QOL scores also differed significantly (p < 0.05, p < 0.001) between children with well-controlled epilepsy with and without NPCCs (table 3); however, no significant differences were found in total scores between children with well-controlled epilepsy with NPCCs and children with refractory epilepsy.

Table 4 lists correlations for demographics, seizure characteristics, and HRQOL measures. Total parental scores showed a positive correlation with the age of the father (R = 0.43, p < 0.01). Neither the duration of the average seizure nor seizure frequency correlated with parental total score.

Internal consistency

The Cronbach α for our sample was calculated for reliability of the measure and was 0.81 for the parents and 0.90 for the children for the various subscales of the HRQOL, exceeding the generally accepted criterion of 0.70 for adequate reliability in group comparisons.

Construct validity

Pretest hypotheses were analyzed to check construct validity of the instrument in paediatric epilepsy. As would be expected from previous evidence, we found a significant correlation between the HRQOL as measured by PedsQL.v4.0 and the duration of diagnosis of epilepsy (Hoare and Russell, 1995; Sabaz et al., 2000; Hoare et al., 2000) (table 4); the severity of epilepsy (Hoare and Russell, 1995; Sabaz et al., 2000; Camfield et al., 2001; Austin et al., 1996) (table 3); the number of medications (Sabaz et al., 2000; Sabaz et al., 2001) (data not shown); and the existence of NPCCs (Hoare and Russell, 1995; Camfield et al., 2001) (table 3).

Discussion

Our study supports the findings of other investigators that epilepsy has an adverse impact on children's quality of life. HRQOL scores were lower in our population compared with children with other chronic health conditions, corroborating earlier studies showing QOL to be more impaired in individuals with epilepsy compared with other medical conditions such as asthma (Modi et al., 2009; Austin et al., 1996). Areas of greatest concern for parents and children as reflected by lowest domain scores were school and emotional functioning. This finding is similar to previous studies of PedsQL that have reported emotional factors (Felder-Puig et al., 2004) or school and emotional factors (Talarska, 2003) to be of most concern in epilepsy. Other workers also recently demonstrated that global delay occurs across all domains of the parent proxy PedsQL scores in children who have had a single seizure or who have newly diagnosed, untreated epilepsy (Modi et al., 2009).

Our study using the PedsQL.v4.0 shows a positive and very significant correlation between children with epilepsy and their parents for both the overall HRQOL scores and the physical, emotional, social, and school performance subscores. This suggests that the QOL as perceived by the parent closely approximates the self-perceived QOL by the child. These findings are comparable to those of a study using a German version of PedsQL.v4.0, to examine the impact of cancer and epilepsy, which also described good parent-child correlation, especially for epilepsy (Felder-Puig et al., 2004). This agreement between parent and child perceptions of HRQOL is important in paediatric epilepsy, a condition in which young age or NPCCs can impair completion of HRQOL instruments.

Although we found a good correlation between the severity of epilepsy and HRQOL measures for parents, our study also underscores the importance of underlying NPCCs. In children with well-controlled epilepsy, parent and child perceptions of QOL were significantly better for children with epilepsy alone compared with those with comorbid conditions. The presence of such NPCCs appears to significantly reduce the perceived QOL compared with epilepsy in isolation. This finding is consistent with a similar comparison of parental HRQOL scores by Modi et al. (2009) among patients with a single seizure or newly diagnosed, untreated epilepsy.

An interesting finding in our study was the positive correlation between parental age and the child's psychosocial scores and overall QOL. Possible reasons for this finding include better economic situation or better coping mechanisms in older parents. The combined socioeconomic status of the parents, as measured by the Hollingshead scale, correlated with the emotional and school, but not the physical and social, subscores of the HRQOL. Neither the individual scores of the mother or father for education or employment, nor the combined index, correlated with any of the measures of HRQOL. This would seem to suggest that parental education or employment status does not impact the HRQOL. However, contrary to our findings, Devinsky et al. (1995) have shown positive HRQOL correlations with the educational and employment status of the parents, and Sabaz et al. (2003) have shown correlation with the estimated family income but not the educational status of the parents. Our study also corroborates work by previous investigators showing that age at seizure onset, duration of diagnosis, number of AEDs, and number of NPCCs all affect QOL in children with epilepsy (Hoare and Russell, 1995; Hoare et al., 2000; Sabaz et al., 2000; Camfield et al., 2001; Sabaz et al., 2001; Modi et al., 2009).

The PedsQL.v4.0 is concise and convenient with only 23 items (18 for toddlers) yet is broad enough in assessing four domains of physical and social functioning. The instrument can be completed within five minutes (Modi et al., 2009), which compares favourably with comprehensive instruments such as the QOLIE (Quality of Life in Epilepsy; 28 minutes) (Devinsky et al., 1995) and the Impact of Childhood Illness Scale (15 minutes) (Hoare and Russell, 1995) and is well within the range of an average waiting room time in a busy outpatient clinic. Reviewing the results of the PedsQL.v4.0 would provide added depth to the clinical encounter without draining the clinician's time. Finally, because it provides information about patient concerns, a quality-of-life assessment as a brief, office-based psychological test may be reimbursable (Cramer et al., 1996).

It is perhaps questionable whether generic HRQOL instruments are comparable to the in-depth and tailored assessment of the various components of HRQOL which epilepsy-specific instruments provide. A study comparing a generic measure of HRQOL (the Impact of Childhood Illness scale [ICI]) to epilepsy-specific measures (the Impact of Child Neurologic Handicap Scale [ICNH] and the Hague Restrictions in Epilepsy Scale [HARCES]) found that all three were adequate measures to assess HRQOL in intractable childhood epilepsy (Sherman et al., 2002). Generic HRQOL measures such as the PedsQL.v4.0 remain valuable as they are portable across different diseases such as cancer, diabetes, arthritis, and epilepsy, many of which may be comorbid in the same individual. Another advantage of generic scales over epilepsy-specific scales is that they have been well validated (Carpay and Arts, 1996), as has the PedsQL.v4.0 (Varni et al., 2001). A generic measure would also be better understood and deployed by a non-specialist physician or nurse.

Internal consistency and construct validity were demonstrated for PedsQL.v4.0 by measuring the HRQOL of children with epilepsy. The corroboration of the pretest hypotheses based on existing literature proved the construct validity of the instrument. However, we did not have test–retest data to validate this measure. The strong psychometric properties of the various HRQOL subscales, as measured by PedsQL.v4.0, make this a good instrument for future trials and clinical work.

The limitations of our study were as follows. As only three children had new-onset seizures (which were subsequently excluded), a comparison of HRQOL indices between new-onset and chronic epilepsy could not be done. However, this comparison has been done elsewhere using PedsQL (Modi et al., 2009). A study coordinator supervised the completion of a proportion of the questionnaires and the rest were self-administered, limiting standardization of the testing situation. Data were not available regarding the proportion of patients who accepted the invitation to participate in the study versus those who declined.

A significant proportion of the subjects had NPCCs, which can affect HRQOL. However, NPCCs are expected to accompany epilepsy in real life, especially refractory epilepsy. Comparing the results to a previous report of normal children has an inherent weakness in variability of test administration settings. However, our comparison was based on seminal work proving the validity of PedsQL, which is potentially the best external benchmark available for this purpose.

Our research shows that the PedsQL.v4.0 assesses the various aspects of HRQOL in agreement with previously published literature. We have demonstrated good correlation between child self-report and parent proxy-report, thus establishing the PedsQL.v4.0 as a useful tool in situations in which child self-reporting is unreliable or unrealistic. A head-to-head comparison is needed for the validity of simpler instruments with more comprehensive measures which have been established as reliable in larger populations with epilepsy.

In summary, we demonstrate significant impairment in HRQOL among children with epilepsy compared with the normal population and children with other chronic health conditions. Neuropsychiatric comorbidities have a significant impact on HRQOL in this population. We demonstrate good concordance between parental and child perceptions of HRQOL using the PedsQL.v4.0 model. The ease of use and reliability of this instrument may make it valuable for routine use in the clinic.Disclosure

None of the authors has any conflict of interest or financial support to disclose.