Haemophilia (2014), 20, 616–623
DOI: 10.1111/hae.12463
ORIGINAL ARTICLE Clinical haemophilia
Adherence to treatment in a Western European paediatric population with haemophilia: reliability and validity of the VERITAS-Pro scale J. LOCK,* H. RAAT,† N. DUNCAN,‡ A. SHAPIRO,‡ M. BEIJLEVELT,§ M. PETERS,§ R. Y. J. T A M M I N G A , ¶ F . W . G . L E E B E E K , * * H . A . M O L L † † and M . H . C N O S S E N * *Department of Paediatric Haematology, Erasmus MC - Sophia Children’s Hospital; †Department of Public Health, Erasmus MC, University Medical Centre Rotterdam, Rotterdam, Netherlands; ‡Department of Haematology, Indiana Haemophilia and Thrombosis Centre, Indianapolis, IN, USA; §Department of Paediatric Haematology, Academic Medical Centre Amsterdam, Amsterdam; ¶Department of Paediatric Haematology, University Medical Centre Groningen, Groningen; **Department of Haematology, Erasmus MC, University Medical Centre Rotterdam; and ††Department of General Paediatrics, Erasmus MC – Sophia Children’s Hospital, Rotterdam, Netherlands
Summary. Treatment adherence in haemophilia is strongly associated with quality of life and the cost– benefit of treatment. Therefore, it is important to quantify and monitor it. This study aimed to validate a translation of the VERITAS-Pro cross-culturally and analyse treatment adherence in a Dutch population of paediatric haemophilia patients. Children aged 1– 18 years with haemophilia were included from three Haemophilia Treatment Centres, on prophylactic clotting factor replacement therapy for more than 1 year. Parents and adolescents were analysed separately. The adherence scale for prophylactic therapy (VERITAS-Pro) was translated according to international guidelines. This instrument contains a total of six subscales (‘Time’, ‘Dose’, ‘Plan’, ‘Remember’, ‘Skip’ and ‘Communicate’) each with four items. Lower scores reflect higher adherence. Overall response rate was 85%, leading to a study population of 60 children. Mean age was 10 years (SD 4.1). Internal consistency reliability: Mean
Introduction Patients’ adherence to therapy and recommended lifestyle measures are strongly correlated with clinical outcome, quality of life and cost–benefit of treatment Correspondence: M.H. Cnossen, MD PhD, Department of Paediatric Haematology, Erasmus University Medical Centre - Sophia Children’s Hospital, P.O. Box 2040, 3000 CA Rotterdam, the Netherlands. Tel.: +31 10 7036691; fax: +31 10 7036801; e-mail: [email protected] Accepted after revision 22 April 2014 616
Cronbach’s alphas were adequate (>0.70) for total score and the subscales ‘Skip’ and ‘Communicate’. Item-own subscale correlations were stronger than most item-other subscale correlations. Convergent validity: Total scores were higher for non-adherent participants compared with adherent participants according to patient infusion logs (n = 48; P < 0.05). Test–retest correlations: Significant for all scales except ‘Dose’ (n = 58; P < 0.01). This study demonstrates applicability of VERITAS-Pro outside the United States, as total score and most subscales effectively quantified treatment adherence in a Dutch paediatric population on prophylactic therapy. Nonadherent respondents’ total scores were significantly higher, demonstrating the ability of VERITAS-Pro to identify non-adherent individuals. Keywords: adherence, children, haemophilia, reliability, Validated Haemophilia Regimen Treatment Adherence Scale-Prophylaxis (VERITAS-Pro), validity
in chronic diseases such as haemophilia [1–3]. In Europe, when indicated, haemophilia patients are treated prophylactically with clotting factor replacement therapy in the home setting (prophylaxis), which significantly decreases spontaneous bleeding and joint damage [4–7]. In case of acute bleeding, haemophilia patients are also treated on demand with clotting factor concentrate. Both strategies require patient adherence and responsibility with regard to treatment. In a meta-analysis assessing general adherence and clinical outcomes of medical treatment, DiMatteo et al. found that 26% of patients experienced a better clinical outcome if adhering to treatment [2]. In haemophilia, © 2014 John Wiley & Sons Ltd
ADHERENCE IN HAEMOPHILIA
non-adherence to prescribed clotting factor therapy is reported in up to 50% of patients on prophylaxis [8– 11], clinically manifested by both under- and overtreatment. Expectedly, regular undertreatment leads to higher risk of repetitive joint bleeding, synovial hypertrophy, cartilage- and bone damage with arthropathy and disability as a result [1,12]. On the other hand, overtreatment leads to unnecessary and avoidable costs, as well as a higher risk of complications such as development of inhibiting antibodies, adverse reactions and infectious or thrombotic complications [13,14]. The main reasons for non-adherence in haemophilia are the chronic and unpredictable course of disease, long-term dependency on treatment, obligatory lifestyle adjustments and invasiveness of intravenous administration of clotting factor concentrate [15–19]. The importance of identifying treatment non-adherence was recently established in a review by Schrijvers et al. Primary determinants were infrequent or absence of symptoms and increasing age. Motivators of adherence were experience with the disease, positive ideas regarding necessity of treatment and an optimal relationship with the health care provider [9,11,20]. Recent developments with regard to pharmacokinetic-guided dosing of replacement therapy and the development of long-acting clotting factor products, underline the importance of adherence in the haemophilia population. This, as minimal and infrequent dosing schemes in combination with non-adherence, may lead to dangerously low clotting factor trough levels with increased risk of bleeding [21]. Although recognized as important, a recent study by Chan et al. indicated that up to 18% of haemophilia professionals do not assess adherence in the clinical setting at all [22]. This is most likely due to the lack of accurate and inexpensive instruments to quantify adherence [23]. Healthcare professionals therefore rely on their personal judgment when evaluating treatment adherence, which has proven unreliable repetitively [24]. A valuable tool to quantify adherence in haemophilia is the Validated Haemophilia Regimen Treatment Adherence Scale - Prophylaxis (VERITAS-Pro), developed and validated by Duncan et al. (2010) in the United States [25]. This initiative combines qualitative research with quantitative survey techniques and has led to a time-efficient and feasible instrument to monitor adherence to prophylaxis, applicable in all age groups [26]. Until now, VERITAS-Pro has only been validated in the United States. However, to prove reproducibility and to promote broader application, it is necessary to validate it in other populations that differ with regard to language, culture and healthcare organization. In this study, we aimed to validate the Dutch-translated version of the VERITAS-Pro crossculturally and analyse treatment adherence in a Dutch paediatric haemophilia population. © 2014 John Wiley & Sons Ltd
617
Patients and methods Patients In this cross-sectional, Dutch multicentre study data were collected as part of a larger prospective study on the efficacy of home-treatment intervention by a trained haemophilia nurse (Dutch Trial Register: 2543). Between June 2010 and December 2011, we enrolled children aged 1–18 years with haemophilia A or haemophilia B on prophylactic home treatment for at least 1 year, and their parents from three Dutch Haemophilia Treatment Centres (HTC). Patients and parents with language difficulties and patients with inhibitors were excluded. One caregiver, primarily involved in daily haemophilia treatment, and adolescents aged 10–18 years, were asked to complete the web-based questionnaire. To evaluate test–retest reliability of VERITAS-Pro, questionnaires were sent 2 weeks after the first administration to consenting participants. Participants not returning the questionnaire within 2 weeks received reminders and were considered lost to follow-up after two unreturned messages. The Medical Ethical Committee granted permission to perform the study and written informed consent was obtained [MEC2010097].
Data collection Socio-demographic data were registered. For level of education, the International Standard Classification of Education (ISCED) division into low (ISCED 0–2; less than upper secondary level), medium (ISCED 3–4; upper secondary level) and high educational level was applied (ISCED 5–6; tertiary level or the achievement of two more years after upper secondary level) [27]. Haemophilia diagnosis, treatment and pharmacy dispensations were recorded. Infusion log data and bleedings were collected when available.
Adherence VERITAS-Pro. To quantify treatment adherence in children utilizing prophylaxis, we used the VERITASPro scale, which takes approximately 10 min to complete [25]. This instrument consists of six subscales, concerning a specific domain of haemophilia care. Each subscale is represented by four questions, leading to a total of 24 items (Table 1). VERITAS-Pro evaluates the necessity and dosing of clotting factor concentrate (subscales: ‘Time’, ‘Dose’) in relationship with prior prophylactic doses (subscales: ‘Remember’, ‘Skip’), and if the patient’s HTC was contacted (subscale: ‘Communicate’). Also, management of clotting factor infusion and stock is monitored (subscale: ‘Plan’). Each item is quantified on a five-point Likert scale ranging from ‘Always’ to ‘Never’. Each rating on Haemophilia (2014), 20, 616--623
618
J. LOCK et al.
Table 1. Characteristics of the 60 participants at the time of study enrolment. Characteristics Patient characteristics Age patients (years), mean (SD) Sex patients, male Diagnosis Haemophilia A Haemophilia B Disease severity Severe (75% were considered to display high adherence, based on earlier studies [9,10,15,25,29]. Bleeds were recorded in infusion logs and medical records.
Data analysis Psychometric properties of VERITAS-Pro. VERITASPro scores were described by conventional descriptive statistics. Reliability. To evaluate consistency of results across items, we determined internal consistency reliability using Cronbach’s alpha. Assuming that test items measuring the same construct are correlated [30], we chose 0.70 to be an adequate alpha coefficient [31,32]. Homogeneity of VERITAS-Pro was evaluated on the basis of inter-item correlations, which determines the correlation between each item and the subscale to which this item belongs. Inter-item correlations >0.20 are generally acceptable [30,33,34]. Average inter-item, average item-own scale and average item-other scale correlations were assessed with standardized correlation coefficients. To test reproducibility and consistency over time, we determined test–retest reliability with Spearman effect size correlation coefficients (r) for non-parametric data; it assumes that there will be no change in the construct measured. The agreement between the perceived adherence of parents and adolescents was also assessed by Spearman correlation [35]. According to established guidelines, concordance was determined as poor (0.50) [36]. To test stability of scores, we used the rank-sum test to evaluate whether scores of aggregate test–retest reliability and inter-rater agreement groups systematically changed over time or between parents and adolescents. Validity. Validity is the extent to which a test measures what it intends to measure. To determine the degree to which two constructs that are theoretically related practically relate within the test, we assessed the convergent validity [37]. It was hypothesized that patients with a high adherence according to infusion log documentation would report better adherence outcomes, thus lower VERITAS-Pro scores. In addition, that patients with the lowest quartile annual bleeding rate would report better adherence outcomes than patients with the highest quartile annual bleeding rate. Differences in median total scores were calculated and tested with Wilcoxon test, as a consequence of nonparametric data. © 2014 John Wiley & Sons Ltd
ADHERENCE IN HAEMOPHILIA
Data were analysed separately for parent-reported and adolescent-reported scales, except for the interrater agreement analysis, which compared adolescentreported with parent-reported scales. We considered P-values
DOI: 10.1111/hae.12463
ORIGINAL ARTICLE Clinical haemophilia
Adherence to treatment in a Western European paediatric population with haemophilia: reliability and validity of the VERITAS-Pro scale J. LOCK,* H. RAAT,† N. DUNCAN,‡ A. SHAPIRO,‡ M. BEIJLEVELT,§ M. PETERS,§ R. Y. J. T A M M I N G A , ¶ F . W . G . L E E B E E K , * * H . A . M O L L † † and M . H . C N O S S E N * *Department of Paediatric Haematology, Erasmus MC - Sophia Children’s Hospital; †Department of Public Health, Erasmus MC, University Medical Centre Rotterdam, Rotterdam, Netherlands; ‡Department of Haematology, Indiana Haemophilia and Thrombosis Centre, Indianapolis, IN, USA; §Department of Paediatric Haematology, Academic Medical Centre Amsterdam, Amsterdam; ¶Department of Paediatric Haematology, University Medical Centre Groningen, Groningen; **Department of Haematology, Erasmus MC, University Medical Centre Rotterdam; and ††Department of General Paediatrics, Erasmus MC – Sophia Children’s Hospital, Rotterdam, Netherlands
Summary. Treatment adherence in haemophilia is strongly associated with quality of life and the cost– benefit of treatment. Therefore, it is important to quantify and monitor it. This study aimed to validate a translation of the VERITAS-Pro cross-culturally and analyse treatment adherence in a Dutch population of paediatric haemophilia patients. Children aged 1– 18 years with haemophilia were included from three Haemophilia Treatment Centres, on prophylactic clotting factor replacement therapy for more than 1 year. Parents and adolescents were analysed separately. The adherence scale for prophylactic therapy (VERITAS-Pro) was translated according to international guidelines. This instrument contains a total of six subscales (‘Time’, ‘Dose’, ‘Plan’, ‘Remember’, ‘Skip’ and ‘Communicate’) each with four items. Lower scores reflect higher adherence. Overall response rate was 85%, leading to a study population of 60 children. Mean age was 10 years (SD 4.1). Internal consistency reliability: Mean
Introduction Patients’ adherence to therapy and recommended lifestyle measures are strongly correlated with clinical outcome, quality of life and cost–benefit of treatment Correspondence: M.H. Cnossen, MD PhD, Department of Paediatric Haematology, Erasmus University Medical Centre - Sophia Children’s Hospital, P.O. Box 2040, 3000 CA Rotterdam, the Netherlands. Tel.: +31 10 7036691; fax: +31 10 7036801; e-mail: [email protected] Accepted after revision 22 April 2014 616
Cronbach’s alphas were adequate (>0.70) for total score and the subscales ‘Skip’ and ‘Communicate’. Item-own subscale correlations were stronger than most item-other subscale correlations. Convergent validity: Total scores were higher for non-adherent participants compared with adherent participants according to patient infusion logs (n = 48; P < 0.05). Test–retest correlations: Significant for all scales except ‘Dose’ (n = 58; P < 0.01). This study demonstrates applicability of VERITAS-Pro outside the United States, as total score and most subscales effectively quantified treatment adherence in a Dutch paediatric population on prophylactic therapy. Nonadherent respondents’ total scores were significantly higher, demonstrating the ability of VERITAS-Pro to identify non-adherent individuals. Keywords: adherence, children, haemophilia, reliability, Validated Haemophilia Regimen Treatment Adherence Scale-Prophylaxis (VERITAS-Pro), validity
in chronic diseases such as haemophilia [1–3]. In Europe, when indicated, haemophilia patients are treated prophylactically with clotting factor replacement therapy in the home setting (prophylaxis), which significantly decreases spontaneous bleeding and joint damage [4–7]. In case of acute bleeding, haemophilia patients are also treated on demand with clotting factor concentrate. Both strategies require patient adherence and responsibility with regard to treatment. In a meta-analysis assessing general adherence and clinical outcomes of medical treatment, DiMatteo et al. found that 26% of patients experienced a better clinical outcome if adhering to treatment [2]. In haemophilia, © 2014 John Wiley & Sons Ltd
ADHERENCE IN HAEMOPHILIA
non-adherence to prescribed clotting factor therapy is reported in up to 50% of patients on prophylaxis [8– 11], clinically manifested by both under- and overtreatment. Expectedly, regular undertreatment leads to higher risk of repetitive joint bleeding, synovial hypertrophy, cartilage- and bone damage with arthropathy and disability as a result [1,12]. On the other hand, overtreatment leads to unnecessary and avoidable costs, as well as a higher risk of complications such as development of inhibiting antibodies, adverse reactions and infectious or thrombotic complications [13,14]. The main reasons for non-adherence in haemophilia are the chronic and unpredictable course of disease, long-term dependency on treatment, obligatory lifestyle adjustments and invasiveness of intravenous administration of clotting factor concentrate [15–19]. The importance of identifying treatment non-adherence was recently established in a review by Schrijvers et al. Primary determinants were infrequent or absence of symptoms and increasing age. Motivators of adherence were experience with the disease, positive ideas regarding necessity of treatment and an optimal relationship with the health care provider [9,11,20]. Recent developments with regard to pharmacokinetic-guided dosing of replacement therapy and the development of long-acting clotting factor products, underline the importance of adherence in the haemophilia population. This, as minimal and infrequent dosing schemes in combination with non-adherence, may lead to dangerously low clotting factor trough levels with increased risk of bleeding [21]. Although recognized as important, a recent study by Chan et al. indicated that up to 18% of haemophilia professionals do not assess adherence in the clinical setting at all [22]. This is most likely due to the lack of accurate and inexpensive instruments to quantify adherence [23]. Healthcare professionals therefore rely on their personal judgment when evaluating treatment adherence, which has proven unreliable repetitively [24]. A valuable tool to quantify adherence in haemophilia is the Validated Haemophilia Regimen Treatment Adherence Scale - Prophylaxis (VERITAS-Pro), developed and validated by Duncan et al. (2010) in the United States [25]. This initiative combines qualitative research with quantitative survey techniques and has led to a time-efficient and feasible instrument to monitor adherence to prophylaxis, applicable in all age groups [26]. Until now, VERITAS-Pro has only been validated in the United States. However, to prove reproducibility and to promote broader application, it is necessary to validate it in other populations that differ with regard to language, culture and healthcare organization. In this study, we aimed to validate the Dutch-translated version of the VERITAS-Pro crossculturally and analyse treatment adherence in a Dutch paediatric haemophilia population. © 2014 John Wiley & Sons Ltd
617
Patients and methods Patients In this cross-sectional, Dutch multicentre study data were collected as part of a larger prospective study on the efficacy of home-treatment intervention by a trained haemophilia nurse (Dutch Trial Register: 2543). Between June 2010 and December 2011, we enrolled children aged 1–18 years with haemophilia A or haemophilia B on prophylactic home treatment for at least 1 year, and their parents from three Dutch Haemophilia Treatment Centres (HTC). Patients and parents with language difficulties and patients with inhibitors were excluded. One caregiver, primarily involved in daily haemophilia treatment, and adolescents aged 10–18 years, were asked to complete the web-based questionnaire. To evaluate test–retest reliability of VERITAS-Pro, questionnaires were sent 2 weeks after the first administration to consenting participants. Participants not returning the questionnaire within 2 weeks received reminders and were considered lost to follow-up after two unreturned messages. The Medical Ethical Committee granted permission to perform the study and written informed consent was obtained [MEC2010097].
Data collection Socio-demographic data were registered. For level of education, the International Standard Classification of Education (ISCED) division into low (ISCED 0–2; less than upper secondary level), medium (ISCED 3–4; upper secondary level) and high educational level was applied (ISCED 5–6; tertiary level or the achievement of two more years after upper secondary level) [27]. Haemophilia diagnosis, treatment and pharmacy dispensations were recorded. Infusion log data and bleedings were collected when available.
Adherence VERITAS-Pro. To quantify treatment adherence in children utilizing prophylaxis, we used the VERITASPro scale, which takes approximately 10 min to complete [25]. This instrument consists of six subscales, concerning a specific domain of haemophilia care. Each subscale is represented by four questions, leading to a total of 24 items (Table 1). VERITAS-Pro evaluates the necessity and dosing of clotting factor concentrate (subscales: ‘Time’, ‘Dose’) in relationship with prior prophylactic doses (subscales: ‘Remember’, ‘Skip’), and if the patient’s HTC was contacted (subscale: ‘Communicate’). Also, management of clotting factor infusion and stock is monitored (subscale: ‘Plan’). Each item is quantified on a five-point Likert scale ranging from ‘Always’ to ‘Never’. Each rating on Haemophilia (2014), 20, 616--623
618
J. LOCK et al.
Table 1. Characteristics of the 60 participants at the time of study enrolment. Characteristics Patient characteristics Age patients (years), mean (SD) Sex patients, male Diagnosis Haemophilia A Haemophilia B Disease severity Severe (75% were considered to display high adherence, based on earlier studies [9,10,15,25,29]. Bleeds were recorded in infusion logs and medical records.
Data analysis Psychometric properties of VERITAS-Pro. VERITASPro scores were described by conventional descriptive statistics. Reliability. To evaluate consistency of results across items, we determined internal consistency reliability using Cronbach’s alpha. Assuming that test items measuring the same construct are correlated [30], we chose 0.70 to be an adequate alpha coefficient [31,32]. Homogeneity of VERITAS-Pro was evaluated on the basis of inter-item correlations, which determines the correlation between each item and the subscale to which this item belongs. Inter-item correlations >0.20 are generally acceptable [30,33,34]. Average inter-item, average item-own scale and average item-other scale correlations were assessed with standardized correlation coefficients. To test reproducibility and consistency over time, we determined test–retest reliability with Spearman effect size correlation coefficients (r) for non-parametric data; it assumes that there will be no change in the construct measured. The agreement between the perceived adherence of parents and adolescents was also assessed by Spearman correlation [35]. According to established guidelines, concordance was determined as poor (0.50) [36]. To test stability of scores, we used the rank-sum test to evaluate whether scores of aggregate test–retest reliability and inter-rater agreement groups systematically changed over time or between parents and adolescents. Validity. Validity is the extent to which a test measures what it intends to measure. To determine the degree to which two constructs that are theoretically related practically relate within the test, we assessed the convergent validity [37]. It was hypothesized that patients with a high adherence according to infusion log documentation would report better adherence outcomes, thus lower VERITAS-Pro scores. In addition, that patients with the lowest quartile annual bleeding rate would report better adherence outcomes than patients with the highest quartile annual bleeding rate. Differences in median total scores were calculated and tested with Wilcoxon test, as a consequence of nonparametric data. © 2014 John Wiley & Sons Ltd
ADHERENCE IN HAEMOPHILIA
Data were analysed separately for parent-reported and adolescent-reported scales, except for the interrater agreement analysis, which compared adolescentreported with parent-reported scales. We considered P-values
