Research in Developmental Disabilities 35 (2014) 776–783
Contents lists available at ScienceDirect
Research in Developmental Disabilities
Multimorbidity in older adults with intellectual disabilities Heidi Hermans a,b,*, Heleen M. Evenhuis a a
Intellectual Disability Medicine, Department of General Practice, Erasmus University Medical Center Rotterdam, P.O. Box 2040, 3000 CA Rotterdam, The Netherlands b Amarant Groep, P.O. Box 715, 5000 AS Tilburg, The Netherlands
A R T I C L E I N F O
A B S T R A C T
Article history: Received 7 October 2013 Received in revised form 21 January 2014 Accepted 22 January 2014 Available online 14 February 2014
Multimorbidity may be related to the supposed early aging of people with intellectual disabilities (ID). This group may suffer more often from multimorbidity, because of IDrelated physical health conditions, unhealthy lifestyle and metabolic effects of antipsychotic drug use. Multimorbidity has been defined as two or more chronic conditions. Data on chronic conditions have been collected through physical assessment, questionnaires, and medical files. Prevalence, associated factors and clusters of multimorbidity have been studied in 1047 older adults (50 years) with ID. Multimorbidity was prevalent in 79.8% and associated with age and severe/profound ID. Four or more conditions were prevalent in 46.8% and associated with age, severe/profound ID and Down syndrome. Factor analyses did not reveal a model for disease-clusters with good fit. Multimorbidity is highly prevalent in older adults with ID. Multimorbidity should receive more attention in research and clinical practice for targeted pro-active prevention and treatment. ß 2014 Elsevier Ltd. All rights reserved.
Keywords: Multimorbidity Intellectual disability Frailty Mental health
1. Introduction It has long been a common understanding, that people with intellectual disabilities (ID) ‘are old from the age of 50 years onwards’ (Perkins & Moran, 2010; Roth, Sun, Greensite, Lott, & Dietrich, 1996). Nevertheless, apart from people with Down syndrome (Roth et al., 1996), premature aging has never been scientifically confirmed for this group. Geriatric frailty occurs early in the population with ID (Evenhuis, Hermans, Hilgenkamp, Bastiaanse, & Echteld, 2012) and is considered to be a risk factor for subsequent deterioration of health and independence (Fried et al., 2001), occurs early in the population with ID. Their mean frailty index scores at age 50–59 years are comparable to those in the general population aged 70–79 years (Schoufour, Mitnitski, Rockwood, Evenhuis, & Echteld, 2013). This early occurrence of frailty might be an explanation for the perceived early aging. Frailty might be partly caused by multimorbidity, which refers to the occurrence of two or more chronic conditions. The prevalence of multimorbidity has been extensively studied in older people with normal intelligence (Glynn et al., 2011; Schram et al., 2008; Van Oostrom et al., 2011). The results of these studies imply that prevalence increases with age and is related to female gender, lower education and low social-economic status (Marengoni, Winblad, Karp, & Fratiglioni, 2008; Salisbury, Johnson, Purdy, Valderas, & Montgomery, 2011; Tucker-Seeley, Li, Sorensen, & Subramanian, 2011; Uijen & van de Lisdonk, 2008). Despite the numerous studies, treatment options are still vague. Physicians seem to treat each disease
* Corresponding author at: Erasmus University Medical Center, Department of General Practice, Intellectual Disability Medicine, P.O. Box 2040, 3000 CA Rotterdam, The Netherlands. Tel.: +31 613342840. E-mail address: [email protected] (H. Hermans). 0891-4222/$ – see front matter ß 2014 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ridd.2014.01.022
H. Hermans, H.M. Evenhuis / Research in Developmental Disabilities 35 (2014) 776–783
777
separately and show little attention for the synergy between different diseases (Bower et al., 2012), whereas lack of good treatment causes ongoing functional decline, impaired quality of life and early death (Drewes et al., 2011; Fortin et al., 2006; Hunger et al., 2011; Landi et al., 2010). People with ID seem to have an increased risk of chronic multimorbidity (McCarron et al., 2013) for several reasons. Multimorbidity may start at a young age, with conditions related to brain damage, impaired brain development, and etiologic syndromes. For example, people with cerebral palsy often have motor impairment, epilepsy and other neurologic problems (Arvio & Sillanpaa, 2003). What is more, risks to develop age-related conditions may be different because of superpositioning on childhood conditions and other unfavorable factors (De Winter, Bastiaanse, Hilgenkamp, Evenhuis, & Echteld, 2012). For instance, an increased risk of cardiovascular risk factors is found both in young and older adults with ID (De Winter et al., 2012; Emerson, 2005; Haveman et al., 2011). This may not only be attributable to an unhealthy lifestyle, but also to metabolic effects of antipsychotic drug use (De Kuijper et al., 2013) and fragmented sleep–wake rhythms (Maaskant, van de Wouw, van Wijck, Evenhuis, & Echteld, 2013). Nevertheless, medical care for this group is primarily reactive, i.e. if complaints or observed symptoms are brought to the attention of the physician (Lennox, Diggens, & Ugoni, 1997). Multimorbidity and frail unhealthy life-years may be delayed by treating conditions that are to be expected during early and later adulthood, as well as anticipating healthcare, aimed at prevention and pro-active diagnosis. Consequently, healthcare costs will decrease because of less dependency caused by additional diseases. To improve healthcare for people with ID, more knowledge on multimorbidity is necessary (McCarron et al., 2013). Therefore, we studied the prevalence and associated factors of chronic multimorbidity in the broad client population, aged 50 years and over, of Dutch intellectual disability service providers. We have also studied the presence of meaningful clusters of multimorbidity, as a basis for anticipating healthcare. 2. Method 2.1. Design and study population This study was part of the cross-sectional ‘Healthy Ageing and Intellectual Disabilities’ (HA-ID) study. This study has been performed in a consort of three large formal ID service providers in the south and west of the Netherlands in both rural and urban environments. These service providers provide care (e.g. washing, nursing) and support (e.g. helping with finances, housekeeping) to a broad spectrum of clients. They cover different levels of support needs: centralized residential accommodations (mainly care), community-based homes (mainly support), day activity centers and supported independent living. The distribution of clients primarily receiving care (35%) and clients primarily receiving support (65%) is similar as in the total Dutch population using formal ID services (Hilgenkamp et al., 2011). People with ID unknown to formal ID services are not part of our study population. For the HA-ID study, all clients aged 50 years or over were invited to participate (n = 2322). The age-limit of 50 years was chosen, because it was generally accepted, though not proven, that people with ID, and not only people with Down syndrome, show signs of premature aging (Patel, Goldberg, & Moss, 1993; Perkins & Moran, 2010). Of the general Dutch population aged 50 years, 0.5% is known to formal ID services (Woittiez & Crone, 2005) of which 10% receives care or support from one of the services participating in this study. Recruitment and the informed consent-procedure have been described in detail elsewhere (Hilgenkamp et al., 2011). In short, all clients receiving care or support from one of the three participating ID service providers on the 1st of September 2008 have been invited to participate in the HA-ID study. Except for age 50 years, no exclusion criteria have been used. Of the total number invited, 49.7% of the clients or their legal representatives gave informed consent to participate of whom 98.2% (n = 1050) actually participated. This study has been approved by the Medical Ethical Testing Committee of the Erasmus University Medical Centre at Rotterdam, the Netherlands (MEC nr: 2008-234). 2.2. Definition of multimorbidity and rationale of included diseases In most studies multimorbidity is defined as two or more chronic conditions, including diseases and risk factors (e.g. hypertension). In the current study, multimorbidity has been defined as two or more chronic conditions which may negatively influence daily functioning. A selection based on included diseases in other multimorbidity studies and included diseases in the HA-ID study has resulted in a list of 20 conditions (Table 1). Conditions with subjective symptoms (e.g. migraine, lower back pain) have not been included, because these conditions are hard to diagnose in people with more severe ID due to their limited abilities to report symptoms. Dysphagia, motor impairment, epilepsy, hearing impairment, visual impairment, gastro-esophageal reflux disease, and autism have been added to the multimorbidity list, because the prevalence rates of these diseases are found to be higher in people with ID because of their association with brain damage of brain dysfunction (Bohmer et al., 1999; De Bildt, Sytema, Kraijer, & Minderaa, 2005; Evenhuis, Theunissen, Denkers, Verschuure, & Kemme, 2001; Kennedy, McCombie, Dawes, McConnell, & Dunnigan, 1997; Oppewal, Hilgenkamp, van Wijck, & Evenhuis, 2013; Ring, Zia, Lindeman, & Himlok, 2007; Splunder, Stilma, Bernsen, & Evenhuis, 2006). Depression has been added because recent research has found that major depression is five times more prevalent in older people with ID than in the general older population (Hermans, Beekman, &
H. Hermans, H.M. Evenhuis / Research in Developmental Disabilities 35 (2014) 776–783
778 Table 1 Description of data collection. Conditions
Data collection
Dysphagia
Mealtime observation performed by trained speech therapists using the Dysphagia Disorders Survey (Sheppard, 1991) Use of laxative medication (medical file, ATC classificationa: A06) Bone stiffness calcaneus measurement using an ultrasonometer (Lunar Achilles Insight) (Economos, Sacheck, Wacker, Shea, & Naumova, 2007) and medication use (medical file, ATC classification: M05) Indication from psychological file Diagnosis from medical file Diagnosis from medical file Diagnosis from medical file Ankle-arm index (systolic blood pressure ankle divided by systolic blood pressure arm) and diagnosis from medical file Diagnosis from medical file Diagnosis from medical file and medication (medical file, ATC classification: H03) Diagnosis from psychological file Diagnosis from medical file Score above cut-off (Hermans et al., 2012) on the Depression subscale of the Anxiety, Depression And Mood Scale (Esbensen et al., 2003) Score above cut-off (Hermans et al., 2012) on the General anxiety subscale of the Anxiety, Depression And Mood Scale (Esbensen et al., 2003) Diagnosis from medical file and medication (medical file, ATC classification: R03) Use of wheelchair inside and outside living accommodation recorded by professional caregivers Fasting serum glucose of 0.6 and medication use (medical file, ATC classification: A10) Diagnosis from medical file and psychological file Diagnosis from medical file Diagnosis in the past 5 years from medical file
Chronic constipation Osteoporosis
Severe challenging behavior Hearing impairment Visual impairment Epilepsy Peripheral arterial disease Gastro-esophageal reflux disease Thyroid dysfunction Autism Other cardiovascular diseases Depression Anxiety Asthma/Chronic Obstructive Pulmonary Disease Motor impairment Diabetes mellitus I and II Dementia Cerebrovascular accident Cancer a
Anatomical Therapeutic Chemical classification system.
Evenhuis, 2013). Anxiety has been included, because this is associated to depression (Hermans et al., 2013). Furthermore, McCarron et al. (2013) have recommended to include mental health issues (McCarron et al., 2013). Severe challenging behavior has been included because of its high occurrence rate, burden, and the possibility that it may be a representation of physical or psychiatric problems (De Winter, Jansen, & Evenhuis, 2011). Chronic constipation and osteoporosis have been included because the high prevalence of motor impairment probably influences the occurrence of both conditions (Bohmer, Taminiau, Klinkenberg-Knol, & Meuwissen, 2001; Wagemans, Fiolet, van der Linden, & Menheere, 1998). Osteoporosis is also more prevalent because of the high anticonvulsants use in this population (Holick, 2007). Thyroid dysfunction and dementia are included because both are more often diagnosed in people with Down syndrome (Graber, Chacko, Regelmann, Costin, & Rapaport, 2012; Roth et al., 1996). Peripheral arterial disease, cerebrovascular accident, other cardiovascular diseases, asthma/Chronic Obstructive Pulmonary Disease (COPD), diabetes mellitus type I and II, and cancer have been added because these conditions are usually included in multimorbidity studies in the general population (Salisbury et al., 2011; Van den Bussche et al., 2011; Van Oostrom et al., 2011). 2.3. Measurement procedure and instruments The measurement procedure of each disease has been described in Table 1. Detailed information on the total organization of the measurements is given in Hilgenkamp et al. (2011). Former diagnoses, etiology of ID, and medication use have been collected through the participants’ medical file using a checklist which was completed by their general practitioner or IDphysician. Level of ID, diagnosis of autism spectrum disorders, and presence of severe challenging behavior have been collected through the participants’ psychological file using a checklist which was completed by their psychologist or behavioral therapist. Level of ID was based on former data of IQ-testing. Data on dementia have been collected through the participants’ medical file and psychological file. The measurements for osteoporosis and peripheral arterial disease (see Table 1) have been performed during a physical examination. Vena puncture was performed after an overnight fast to assess serum glucose levels. Serum was transported frozen, stored at 80 8C and analyzed at the laboratory of the Erasmus Medical Center in Rotterdam. Dysphagia was studied during a mealtime observation performed by trained speech therapists using the Dysphagia Disorders Survey (DDS) (Sheppard, 1991). The DDS is a standardized meal observation developed for people with ID developed by Sheppard et al. The DDS is divided in two parts: part one (seven items) consists of related factors of swallowing and part two (eight items) is the actual mealtime observation. Inter-item reliability is good for both parts and the total DDS: ranging from a = 0.82 to a = 0.91 (Sheppard, 2002). Inter-rater reliability is considered good with a found agreement rate of 97% (Sheppard, 2002). Depression and anxiety have been assessed using the Depression subscale and General anxiety subscale of the Anxiety, Depression And Mood Scale (ADESS) (Esbensen, Rojahn, Aman, & Ruedrich, 2003). The ADESS has been developed for people with ID. The psychometric properties of the Depression subscale and the General anxiety subscale are good: internal
H. Hermans, H.M. Evenhuis / Research in Developmental Disabilities 35 (2014) 776–783
779
consistency of 0.84 and 0.88, test–retest reliability of 0.75 and 0.86, interrater reliability of 0.75 and 0.74, sensitivity of 73– 80% and 80–82%, and validity of 71–79% and 65–78% for the used cut-off scores (Hermans, Jelluma, Pas, & Evenhuis, 2012). The ADESS was completed by professional caregivers who knew the participant for at least three months. 2.4. Statistical analyses All analyses were done with the Statistical Package for the Social Sciences (SPSS) 19.0 and Statistical Analysis Software (SAS) 9.3. To be included in the analyses, participants should have data on at least two of the 20 chronic conditions. First, nonresponse analyses were done using a t-test for age and x2-tests for gender and residential setting. Occurrence rates of the separate conditions and multimorbidity (2 conditions) have been calculated using descriptive statistics. Prevalence rates were also calculated for 4 conditions and for people aged 55 years and over to enable comparison with studies in the general older population. Logistic regression analyses were used to study the association of gender, age, level of ID (borderline/mild, moderate, severe/profound), and Down syndrome (independent variables) with multimorbidity (2 and 4 conditions). All independent variables were simultaneously entered into the regression to control for their influence on other independent variables in the regression. Dummy variables were constructed for level of ID, because all independent variables should be continuous or dichotomous. The magnitude of association of the independent variables with multimorbidity were compared by calculating odds ratios. Multicollinearity refers to a high correlation between independent variables, which is not preferred in regression analysis, and was checked for all independent variables with the variance inflation factor (VIF) of linear regression analysis (Craney & Surles, 2007). The proportion of the dependent variable which is explained by the factors in the model was calculated with R2 of Hosmer and Lemeshow (1989). To determine clusters of conditions, an exploratory factor analysis using the maximum likelihood method allowing for Heywood cases (communality of 1) with a varimax rotation has been performed. The model resulting from the exploratory factor analysis has been examined using confirmatory factor analysis. Our own theoretical model, based on literature and clinical experience, has also been examined using confirmatory factor analysis. Model fit was checked with Benthler’s Comparative Fit Index (CFI), Root Mean Square Error of Approximation (RMSEA) and Standardized Root Mean square Residual (SRMR) (Brown & Cudeck, 1992). Because of the dichotomous character of the variables, a tetrachoric correlation matrix has been made which was used for the factor analysis. In factor analyses, only participants with complete data can be included, which led to loss of participants. Consequently, motor impairment had to be excluded from the list of conditions, because almost all people with motor impairment had incomplete data. The extreme high correlation of some of the data (>0.90) or low prevalence (
Contents lists available at ScienceDirect
Research in Developmental Disabilities
Multimorbidity in older adults with intellectual disabilities Heidi Hermans a,b,*, Heleen M. Evenhuis a a
Intellectual Disability Medicine, Department of General Practice, Erasmus University Medical Center Rotterdam, P.O. Box 2040, 3000 CA Rotterdam, The Netherlands b Amarant Groep, P.O. Box 715, 5000 AS Tilburg, The Netherlands
A R T I C L E I N F O
A B S T R A C T
Article history: Received 7 October 2013 Received in revised form 21 January 2014 Accepted 22 January 2014 Available online 14 February 2014
Multimorbidity may be related to the supposed early aging of people with intellectual disabilities (ID). This group may suffer more often from multimorbidity, because of IDrelated physical health conditions, unhealthy lifestyle and metabolic effects of antipsychotic drug use. Multimorbidity has been defined as two or more chronic conditions. Data on chronic conditions have been collected through physical assessment, questionnaires, and medical files. Prevalence, associated factors and clusters of multimorbidity have been studied in 1047 older adults (50 years) with ID. Multimorbidity was prevalent in 79.8% and associated with age and severe/profound ID. Four or more conditions were prevalent in 46.8% and associated with age, severe/profound ID and Down syndrome. Factor analyses did not reveal a model for disease-clusters with good fit. Multimorbidity is highly prevalent in older adults with ID. Multimorbidity should receive more attention in research and clinical practice for targeted pro-active prevention and treatment. ß 2014 Elsevier Ltd. All rights reserved.
Keywords: Multimorbidity Intellectual disability Frailty Mental health
1. Introduction It has long been a common understanding, that people with intellectual disabilities (ID) ‘are old from the age of 50 years onwards’ (Perkins & Moran, 2010; Roth, Sun, Greensite, Lott, & Dietrich, 1996). Nevertheless, apart from people with Down syndrome (Roth et al., 1996), premature aging has never been scientifically confirmed for this group. Geriatric frailty occurs early in the population with ID (Evenhuis, Hermans, Hilgenkamp, Bastiaanse, & Echteld, 2012) and is considered to be a risk factor for subsequent deterioration of health and independence (Fried et al., 2001), occurs early in the population with ID. Their mean frailty index scores at age 50–59 years are comparable to those in the general population aged 70–79 years (Schoufour, Mitnitski, Rockwood, Evenhuis, & Echteld, 2013). This early occurrence of frailty might be an explanation for the perceived early aging. Frailty might be partly caused by multimorbidity, which refers to the occurrence of two or more chronic conditions. The prevalence of multimorbidity has been extensively studied in older people with normal intelligence (Glynn et al., 2011; Schram et al., 2008; Van Oostrom et al., 2011). The results of these studies imply that prevalence increases with age and is related to female gender, lower education and low social-economic status (Marengoni, Winblad, Karp, & Fratiglioni, 2008; Salisbury, Johnson, Purdy, Valderas, & Montgomery, 2011; Tucker-Seeley, Li, Sorensen, & Subramanian, 2011; Uijen & van de Lisdonk, 2008). Despite the numerous studies, treatment options are still vague. Physicians seem to treat each disease
* Corresponding author at: Erasmus University Medical Center, Department of General Practice, Intellectual Disability Medicine, P.O. Box 2040, 3000 CA Rotterdam, The Netherlands. Tel.: +31 613342840. E-mail address: [email protected] (H. Hermans). 0891-4222/$ – see front matter ß 2014 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ridd.2014.01.022
H. Hermans, H.M. Evenhuis / Research in Developmental Disabilities 35 (2014) 776–783
777
separately and show little attention for the synergy between different diseases (Bower et al., 2012), whereas lack of good treatment causes ongoing functional decline, impaired quality of life and early death (Drewes et al., 2011; Fortin et al., 2006; Hunger et al., 2011; Landi et al., 2010). People with ID seem to have an increased risk of chronic multimorbidity (McCarron et al., 2013) for several reasons. Multimorbidity may start at a young age, with conditions related to brain damage, impaired brain development, and etiologic syndromes. For example, people with cerebral palsy often have motor impairment, epilepsy and other neurologic problems (Arvio & Sillanpaa, 2003). What is more, risks to develop age-related conditions may be different because of superpositioning on childhood conditions and other unfavorable factors (De Winter, Bastiaanse, Hilgenkamp, Evenhuis, & Echteld, 2012). For instance, an increased risk of cardiovascular risk factors is found both in young and older adults with ID (De Winter et al., 2012; Emerson, 2005; Haveman et al., 2011). This may not only be attributable to an unhealthy lifestyle, but also to metabolic effects of antipsychotic drug use (De Kuijper et al., 2013) and fragmented sleep–wake rhythms (Maaskant, van de Wouw, van Wijck, Evenhuis, & Echteld, 2013). Nevertheless, medical care for this group is primarily reactive, i.e. if complaints or observed symptoms are brought to the attention of the physician (Lennox, Diggens, & Ugoni, 1997). Multimorbidity and frail unhealthy life-years may be delayed by treating conditions that are to be expected during early and later adulthood, as well as anticipating healthcare, aimed at prevention and pro-active diagnosis. Consequently, healthcare costs will decrease because of less dependency caused by additional diseases. To improve healthcare for people with ID, more knowledge on multimorbidity is necessary (McCarron et al., 2013). Therefore, we studied the prevalence and associated factors of chronic multimorbidity in the broad client population, aged 50 years and over, of Dutch intellectual disability service providers. We have also studied the presence of meaningful clusters of multimorbidity, as a basis for anticipating healthcare. 2. Method 2.1. Design and study population This study was part of the cross-sectional ‘Healthy Ageing and Intellectual Disabilities’ (HA-ID) study. This study has been performed in a consort of three large formal ID service providers in the south and west of the Netherlands in both rural and urban environments. These service providers provide care (e.g. washing, nursing) and support (e.g. helping with finances, housekeeping) to a broad spectrum of clients. They cover different levels of support needs: centralized residential accommodations (mainly care), community-based homes (mainly support), day activity centers and supported independent living. The distribution of clients primarily receiving care (35%) and clients primarily receiving support (65%) is similar as in the total Dutch population using formal ID services (Hilgenkamp et al., 2011). People with ID unknown to formal ID services are not part of our study population. For the HA-ID study, all clients aged 50 years or over were invited to participate (n = 2322). The age-limit of 50 years was chosen, because it was generally accepted, though not proven, that people with ID, and not only people with Down syndrome, show signs of premature aging (Patel, Goldberg, & Moss, 1993; Perkins & Moran, 2010). Of the general Dutch population aged 50 years, 0.5% is known to formal ID services (Woittiez & Crone, 2005) of which 10% receives care or support from one of the services participating in this study. Recruitment and the informed consent-procedure have been described in detail elsewhere (Hilgenkamp et al., 2011). In short, all clients receiving care or support from one of the three participating ID service providers on the 1st of September 2008 have been invited to participate in the HA-ID study. Except for age 50 years, no exclusion criteria have been used. Of the total number invited, 49.7% of the clients or their legal representatives gave informed consent to participate of whom 98.2% (n = 1050) actually participated. This study has been approved by the Medical Ethical Testing Committee of the Erasmus University Medical Centre at Rotterdam, the Netherlands (MEC nr: 2008-234). 2.2. Definition of multimorbidity and rationale of included diseases In most studies multimorbidity is defined as two or more chronic conditions, including diseases and risk factors (e.g. hypertension). In the current study, multimorbidity has been defined as two or more chronic conditions which may negatively influence daily functioning. A selection based on included diseases in other multimorbidity studies and included diseases in the HA-ID study has resulted in a list of 20 conditions (Table 1). Conditions with subjective symptoms (e.g. migraine, lower back pain) have not been included, because these conditions are hard to diagnose in people with more severe ID due to their limited abilities to report symptoms. Dysphagia, motor impairment, epilepsy, hearing impairment, visual impairment, gastro-esophageal reflux disease, and autism have been added to the multimorbidity list, because the prevalence rates of these diseases are found to be higher in people with ID because of their association with brain damage of brain dysfunction (Bohmer et al., 1999; De Bildt, Sytema, Kraijer, & Minderaa, 2005; Evenhuis, Theunissen, Denkers, Verschuure, & Kemme, 2001; Kennedy, McCombie, Dawes, McConnell, & Dunnigan, 1997; Oppewal, Hilgenkamp, van Wijck, & Evenhuis, 2013; Ring, Zia, Lindeman, & Himlok, 2007; Splunder, Stilma, Bernsen, & Evenhuis, 2006). Depression has been added because recent research has found that major depression is five times more prevalent in older people with ID than in the general older population (Hermans, Beekman, &
H. Hermans, H.M. Evenhuis / Research in Developmental Disabilities 35 (2014) 776–783
778 Table 1 Description of data collection. Conditions
Data collection
Dysphagia
Mealtime observation performed by trained speech therapists using the Dysphagia Disorders Survey (Sheppard, 1991) Use of laxative medication (medical file, ATC classificationa: A06) Bone stiffness calcaneus measurement using an ultrasonometer (Lunar Achilles Insight) (Economos, Sacheck, Wacker, Shea, & Naumova, 2007) and medication use (medical file, ATC classification: M05) Indication from psychological file Diagnosis from medical file Diagnosis from medical file Diagnosis from medical file Ankle-arm index (systolic blood pressure ankle divided by systolic blood pressure arm) and diagnosis from medical file Diagnosis from medical file Diagnosis from medical file and medication (medical file, ATC classification: H03) Diagnosis from psychological file Diagnosis from medical file Score above cut-off (Hermans et al., 2012) on the Depression subscale of the Anxiety, Depression And Mood Scale (Esbensen et al., 2003) Score above cut-off (Hermans et al., 2012) on the General anxiety subscale of the Anxiety, Depression And Mood Scale (Esbensen et al., 2003) Diagnosis from medical file and medication (medical file, ATC classification: R03) Use of wheelchair inside and outside living accommodation recorded by professional caregivers Fasting serum glucose of 0.6 and medication use (medical file, ATC classification: A10) Diagnosis from medical file and psychological file Diagnosis from medical file Diagnosis in the past 5 years from medical file
Chronic constipation Osteoporosis
Severe challenging behavior Hearing impairment Visual impairment Epilepsy Peripheral arterial disease Gastro-esophageal reflux disease Thyroid dysfunction Autism Other cardiovascular diseases Depression Anxiety Asthma/Chronic Obstructive Pulmonary Disease Motor impairment Diabetes mellitus I and II Dementia Cerebrovascular accident Cancer a
Anatomical Therapeutic Chemical classification system.
Evenhuis, 2013). Anxiety has been included, because this is associated to depression (Hermans et al., 2013). Furthermore, McCarron et al. (2013) have recommended to include mental health issues (McCarron et al., 2013). Severe challenging behavior has been included because of its high occurrence rate, burden, and the possibility that it may be a representation of physical or psychiatric problems (De Winter, Jansen, & Evenhuis, 2011). Chronic constipation and osteoporosis have been included because the high prevalence of motor impairment probably influences the occurrence of both conditions (Bohmer, Taminiau, Klinkenberg-Knol, & Meuwissen, 2001; Wagemans, Fiolet, van der Linden, & Menheere, 1998). Osteoporosis is also more prevalent because of the high anticonvulsants use in this population (Holick, 2007). Thyroid dysfunction and dementia are included because both are more often diagnosed in people with Down syndrome (Graber, Chacko, Regelmann, Costin, & Rapaport, 2012; Roth et al., 1996). Peripheral arterial disease, cerebrovascular accident, other cardiovascular diseases, asthma/Chronic Obstructive Pulmonary Disease (COPD), diabetes mellitus type I and II, and cancer have been added because these conditions are usually included in multimorbidity studies in the general population (Salisbury et al., 2011; Van den Bussche et al., 2011; Van Oostrom et al., 2011). 2.3. Measurement procedure and instruments The measurement procedure of each disease has been described in Table 1. Detailed information on the total organization of the measurements is given in Hilgenkamp et al. (2011). Former diagnoses, etiology of ID, and medication use have been collected through the participants’ medical file using a checklist which was completed by their general practitioner or IDphysician. Level of ID, diagnosis of autism spectrum disorders, and presence of severe challenging behavior have been collected through the participants’ psychological file using a checklist which was completed by their psychologist or behavioral therapist. Level of ID was based on former data of IQ-testing. Data on dementia have been collected through the participants’ medical file and psychological file. The measurements for osteoporosis and peripheral arterial disease (see Table 1) have been performed during a physical examination. Vena puncture was performed after an overnight fast to assess serum glucose levels. Serum was transported frozen, stored at 80 8C and analyzed at the laboratory of the Erasmus Medical Center in Rotterdam. Dysphagia was studied during a mealtime observation performed by trained speech therapists using the Dysphagia Disorders Survey (DDS) (Sheppard, 1991). The DDS is a standardized meal observation developed for people with ID developed by Sheppard et al. The DDS is divided in two parts: part one (seven items) consists of related factors of swallowing and part two (eight items) is the actual mealtime observation. Inter-item reliability is good for both parts and the total DDS: ranging from a = 0.82 to a = 0.91 (Sheppard, 2002). Inter-rater reliability is considered good with a found agreement rate of 97% (Sheppard, 2002). Depression and anxiety have been assessed using the Depression subscale and General anxiety subscale of the Anxiety, Depression And Mood Scale (ADESS) (Esbensen, Rojahn, Aman, & Ruedrich, 2003). The ADESS has been developed for people with ID. The psychometric properties of the Depression subscale and the General anxiety subscale are good: internal
H. Hermans, H.M. Evenhuis / Research in Developmental Disabilities 35 (2014) 776–783
779
consistency of 0.84 and 0.88, test–retest reliability of 0.75 and 0.86, interrater reliability of 0.75 and 0.74, sensitivity of 73– 80% and 80–82%, and validity of 71–79% and 65–78% for the used cut-off scores (Hermans, Jelluma, Pas, & Evenhuis, 2012). The ADESS was completed by professional caregivers who knew the participant for at least three months. 2.4. Statistical analyses All analyses were done with the Statistical Package for the Social Sciences (SPSS) 19.0 and Statistical Analysis Software (SAS) 9.3. To be included in the analyses, participants should have data on at least two of the 20 chronic conditions. First, nonresponse analyses were done using a t-test for age and x2-tests for gender and residential setting. Occurrence rates of the separate conditions and multimorbidity (2 conditions) have been calculated using descriptive statistics. Prevalence rates were also calculated for 4 conditions and for people aged 55 years and over to enable comparison with studies in the general older population. Logistic regression analyses were used to study the association of gender, age, level of ID (borderline/mild, moderate, severe/profound), and Down syndrome (independent variables) with multimorbidity (2 and 4 conditions). All independent variables were simultaneously entered into the regression to control for their influence on other independent variables in the regression. Dummy variables were constructed for level of ID, because all independent variables should be continuous or dichotomous. The magnitude of association of the independent variables with multimorbidity were compared by calculating odds ratios. Multicollinearity refers to a high correlation between independent variables, which is not preferred in regression analysis, and was checked for all independent variables with the variance inflation factor (VIF) of linear regression analysis (Craney & Surles, 2007). The proportion of the dependent variable which is explained by the factors in the model was calculated with R2 of Hosmer and Lemeshow (1989). To determine clusters of conditions, an exploratory factor analysis using the maximum likelihood method allowing for Heywood cases (communality of 1) with a varimax rotation has been performed. The model resulting from the exploratory factor analysis has been examined using confirmatory factor analysis. Our own theoretical model, based on literature and clinical experience, has also been examined using confirmatory factor analysis. Model fit was checked with Benthler’s Comparative Fit Index (CFI), Root Mean Square Error of Approximation (RMSEA) and Standardized Root Mean square Residual (SRMR) (Brown & Cudeck, 1992). Because of the dichotomous character of the variables, a tetrachoric correlation matrix has been made which was used for the factor analysis. In factor analyses, only participants with complete data can be included, which led to loss of participants. Consequently, motor impairment had to be excluded from the list of conditions, because almost all people with motor impairment had incomplete data. The extreme high correlation of some of the data (>0.90) or low prevalence (
