555819
research-article2014
CNU0010.1177/1474515114555819European Journal of Cardiovascular NursingSchoormans et al.
EUROPEAN SOCIETY OF CARDIOLOGY ®
Original Article
Clinical and psychological characteristics predict future healthcare use in adults with congenital heart disease
European Journal of Cardiovascular Nursing 2016, Vol. 15(1) 72–81 © The European Society of Cardiology 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1474515114555819 cnu.sagepub.com
Dounya Schoormans1–3, Mirjam AG Sprangers1, Joost P van Melle4, Petronella G Pieper4, Arie PJ van Dijk5, Gertjan Tj Sieswerda6, Mariët S Hulsbergen-Zwarts7, Thijs HWM Plokker8, Leo GH Brunninkhuis9, Hubert W Vliegen10 and Barbara JM Mulder2,3
Abstract Background: To deliver adequate care to patients with congenital heart disease (CHD), it is important to know which patients use what type of care. This knowledge is valuable, as modification of these factors may be used as means to regulate healthcare use. Our objective was to examine the predictive value of psychological characteristics for future healthcare use, independent of clinical characteristics. Methods: In total 845 adult CHD-patients participated in a longitudinal questionnaire study, with a two-year followup period. Linear regression analyses with negative binomial log link function were performed predicting healthcare used during the previous year. Psychological predictors were Type D personality, quality of life (QoL), depressive symptoms, trait-anxiety, happiness, optimism, and illness perceptions, independent of the number of co-morbidities, disease complexity and functional status. To control for clustering we included the variable type of centre (regional versus tertiary referral). Results: Patients who reported more healthcare use had a complex defect, a poor functional status, no Type D personality, and a poor QoL. They moreover felt their CHD had a severe impact on their life and believed their CHD could be managed by themselves or treatment. Conclusions: Healthcare use is not entirely determined by disease complexity and functional status but also by psychological patient characteristics. It can by hypothesised that reducing the negative impact experienced and informing patients about strategies to manage their CHD, will modify their future healthcare use. Additional research is necessary to examine this possibility. Keywords Congenital heart disease, healthcare use, type D personality, quality of life, illness perceptions Date received: 23 April 2014; revised: 23 September 2014; accepted: 25 September 2014
1Department
of Medical Psychology, Academic Medical Centre, the
Netherlands 2Department of Cardiology, Academic Medical Centre, the Netherlands 3Interuniversity Cardiology Institute of the Netherlands, the Netherlands 4Department of Cardiology, University Medical Centre Groningen, the Netherlands 5Department of Cardiology, Radboud University Nijmegen Medical Centre, the Netherlands
6Department
of Cardiology, University Medical Centre Utrecht, the Netherlands 7Department of Cardiology, TweeSteden Hospital, the Netherlands 8Department of Cardiology, Saint Antonius Hospital, the Netherlands 9Department of Cardiology, Orbis Medical Centre, the Netherlands 10Leiden University Medical Centre, the Netherlands Corresponding author: Dounya Schoormans, Department of Medical and Clinical Psychology, Tilburg University, P.O. Box 90153, 5000 LE Tilburg, the Netherlands. Email: [email protected]
Downloaded from cnu.sagepub.com at University of York on February 19, 2016
Schoormans et al.
73
Introduction
approval is not required for questionnaire studies under Dutch law, this study was exempted from approval of the Medical Ethics Committee of the Academic Medical Centre in Amsterdam (reference number 09.17.0207). This study was conducted in accordance with the principles of the Declaration of Helsinki.
The number of adults with congenital heart disease (CHD) has grown expansively over the last few decades. As a consequence of their residual lesions, patients with CHD experience a number of physical symptoms,1 psychological impairments (for example a poor quality of life (QoL) or increased levels of depression and anxiety),2–5 and social problems (for example reproduction, employment and financial issues).6 Therefore, it is highly recommended that care for adults with CHD is multidisciplinary.7,8 To facilitate optimal care for CHD-patients, we need to focus on healthcare utilisation provided not only by cardiologists, but also by other healthcare providers such as pulmonologists or social-workers. Studies examining the number of contacts that CHD-patients have with this wide range of healthcare providers is scarce.9,10 Research in other patient populations have shown that in addition to clinical factors, psychological patient characteristics (e.g. QoL, depression, anxiety and illness perceptions) are also predictive of healthcare use.11–15 Knowledge on such psychological factors is valuable, as modification of these factors may be used as means to regulate healthcare use. Furthermore, it is important for healthcare providers and policy makers to know which patients use what type of healthcare. This knowledge will aid in the development of future healthcare facilities and tailor care to those who need it. In the present study we therefore examined the independent predictive value of psychological characteristics on future healthcare use. Included characteristics were personality, QoL, negative and positive emotions (depressive symptoms, trait-anxiety, happiness, optimism) and illness perceptions
Methods Study population and procedure This study is part of a larger longitudinal study on healthcare use of adults with CHD.10 Adult CHD-patients from five tertiary referral and three regional centres were randomly selected from the CONCOR database, a nationwide registry for CHD-patients.16 Upon registration in the CONgenital CORvitia (CONCOR) database, patients gave informed consent to be approached for future scientific research. Inclusion criteria were: being 18 years or older and ever being diagnosed with a CHD. Exclusion criteria for this study were: a diagnosis of Marfan syndrome, being mentally impaired, or illiterate in Dutch. Eligible patients were sent an information letter explaining the nature and procedure of this study. In total, 1670 eligible patients were invited to fill out the questionnaire. After two years another questionnaire was sent to all patients who completed the baseline questionnaire. Since
Measurements Demographic and clinical characteristics. At baseline, patients’ sex, age, co-morbidities, disease complexity and functional status were measured. Sex and age were assessed by self-report. The following previous or ongoing medical conditions were reported: diabetes, kidney disease, gastroenterologic problems, chronic obstructive pulmonary disease (COPD), rheumatic conditions, chronic allergies, chronic skin conditions, problems hearing or seeing, chronic back pain, restrictions in the use of an arm or leg, psychological conditions and being treated for cancer. The type of congenital heart defect was extracted from the CONCOR database and categorised into simple, moderate and complex defects according to Warnes and colleagues.17 Patients’ functional status was operationalised according to the New York Heart Association (NYHA) classification18 and assessed by a validated patient-based question.19 Psychological characteristics. The 14-item DS1420 was used to measure the distressed personality type (type D) – type D personality means having both a high tendency to experience negative affect and to be socially inhibited. The DS14 consists of two subscales; Negative affectivity (seven items) and Social inhibition (seven items). A type D personality is defined as a cut-off score of ≥10 on both subscales of the DS14, and is coded as a dichotomous variable. QoL can be defined in many ways. In general there is consensus that QoL is a multidimensional concept including at least physical, psychological and social functioning. For this study we adopt the following definition related to a person’s health: ‘QoL represents the functional effects of an illness and its consequent therapy upon a patient, as perceived by the patient’ (p.12).21 Patients filled out the Short Form Health Survey-36 (SF-36), yielding a physical component summary (PCS) and mental component summary (MCS) score.22 Additionally, disease-specific QoL was assessed with the TNO/AZL Adult Quality Of Life-CHD (TAAQOL-CHD) forming the three subscales: Symptoms (nine items), Worries (10 items), and Impact cardiac surveillance (seven items).23 Negative emotions were operationalised into measures of depression and trait-anxiety. We used the following definition: ‘a broad range of aversive mood states that form a general distress factor which can be measured either as a transient fluctuation in mood (state) or as a stable individual difference in affective level (trait)’ (p.159).24 The seven-item depression subscale of the Hospital Anxiety
Downloaded from cnu.sagepub.com at University of York on February 19, 2016
74
European Journal of Cardiovascular Nursing 15(1)
and Depression Scale (HADS) was used to measure depression.25 Trait-anxiety was assessed with the 20-item state subscale of the State Trait Anxiety Inventory (STAI).26 We used the following definition of positive emotions: ‘the feelings that reflect a level of pleasurable engagement with the environment such as happiness, joy, excitement, enthusiasm and contentment’. The fouritem Subjective Happiness scale,27 and eight-item Life Orientation Test (LOT)28 were used to assess happiness and optimism, respectively. Illness perceptions were measured by the Illness Perception Questionnaire-Revised form (IPQ-R)29 which was adapted for a CHD-population (see Schoormans et al. for a description of the adaptations made).30 The following six IPQ-R scales measuring illness representations were included in this study. Consequences (five items) describe the effects on physical, psychological and social functioning. Personal (six items) or treatment (three items) control yield questions about whether the illness or its effects can be modified by personal or treatment interference. Illness coherence (four items) measures patients’ understanding of their illness. Timeline-cycle (three items) represents the perceived changeability of the illness and its effects. Emotional representation (five items) addresses the experienced negative emotions (e.g. depressive feelings) caused by the illness. The questionnaires that were used are validated and found to have good psychometric properties.20,27–29,31–34 Please note that one study assessing the validity of the TAAQOL-CHD showed excellent levels of validity,23 whereas others found the levels to be merely adequate.35,36 For all questionnaires, but the DS14, higher scores indicate higher levels of depression, anxiety, happiness, optimism and QoL. The DS14, measuring type D personality is coded as a dichotomous variable. Healthcare use. Two years after the baseline assessment, a study-specific questionnaire (see Schoormans et al. for a detailed description)10 was administered to assess clinical and psychological healthcare use during the previous year for CHD-related reasons. Clinical healthcare yields three types; primary and cardiac outpatient healthcare, additional medical outpatient healthcare, and inpatient healthcare. 1. For primary and cardiac outpatient healthcare the number of contacts with the cardiologist (one item) and general practitioner (one item) were assessed. 2. Additional medical outpatient healthcare (nine items) was operationalised as the number of contacts with nine healthcare providers. 3. Three questions asking how often patients were hospitalised, visited the emergency room, or were operated upon in relation to their CHD, formed inpatient healthcare use.
In addition, psychological healthcare was operationalised by two items assessing the number of contacts with a psychologist and social worker.
Statistical analyses First, we calculated the frequency and percentages for all items to identify the amount of primary and cardiac outpatient, additional medical outpatient, inpatient and psychological healthcare use. Second, we examined the predictive value of psychological characteristics for different types of healthcare use at follow-up, while controlling for demographic and clinical characteristics, and type of centre. Given that data was collected from multiple centres, both regional and tertiary referral, we corrected for clustering by adding a categorical variable representing these two types of centres. Four generalised linear models with negative binominal log link function were run separately for the different types of healthcare use. Sex, age, the number of co-morbidities, disease complexity, functional status, type of centre, type D personality, the four QoL scales, depressive symptoms, trait-anxiety, happiness, optimism and the six illness perception scales were simultaneously included in the analyses. In order to include functional status as a predictor, NYHA-III and NYHA-IV were lumped together leading to enough patients in each category. Exploratory analyses showed that there was no multicollinearity among the predictors (data not shown). The Impact subscale of the TAAQOL-CHD was not included, because items overlap with the healthcare use items. Heterogeneity in our sample was corrected using robust standard errors. A significance level of 0.05 was employed. All analyses were conducted in SPSS 16.0 (SPSS, Chicago, Illinois, USA).
Results Response rate and demographic and clinical characteristics At baseline 1670 eligible patients were invited to participate. A total of 1109 patients filled out the first questionnaire (response rate 66.4%). Non-responders were not different from the participants who were included at baseline (n=1109) in terms of type of CHD and complexity. Two years later the follow-up questionnaire was sent to these respondents (n=1100, nine patients died during the follow-up period). Of the 1100 eligible patients, 845 patients returned the second questionnaire (response rate 76.8%). Patients included in the follow-up were not significantly different from first-time respondents regarding demographic and clinical characteristics. More than half of the 845 patients were female (Table 1). Median age was 39 years. Just over half of all patients
Downloaded from cnu.sagepub.com at University of York on February 19, 2016
75
Schoormans et al. Table 1. Demographic and clinical characteristics for the included patients (n=845). Demographic and clinical characteristics
n (%)
Sex (female) Age in median years (range) Number of co-morbidities No co-morbidity 1 co-morbidity 2 or more co-morbidities Disease complexity17 Simple Moderate Complex Functional status18 NYHA-I NYHA-II NYHA-III/IV Congenital heart defect VSD/ASD/ASD-I/PDA TOF/PS BAV/AS/SubvAS CoA TGA/ccTGA Other CHD-diagnoses
457 (54.7) 39 (18–85) 401 (47.5) 225 (26.6) 219 (25.9) 502 (59.5) 270 (32.0) 72 (8.5) 591 (71.2) 175 (21.1) 64 (7.7) 306 (36.3) 162 (19.2) 121 (14.3) 106 (12.6) 28 (3.3) 121 (14.3)
AS: aortic stenosis; ASD: atrial septum defect; ASD-I: atrial septum defect primum type I; BAV: bicuspid aortic valve; ccTGA: congenital corrected transposition of the great arteries; CHD: congenital heart defect; CoA: coarctation of the aorta; NYHA: New York Heart Association functional class; PDA: patent ductus arteriosus; PS: pulmonary stenosis; SubvAS: subvalvar aorta stenosis; TGA: transposition of the great arteries; TOF: tetralogy of Fallot; VSD: ventricular septum defect. The total sample size ranges from 830–845 due to missing values. Numbers are given in frequencies (valid %), age in median years (range).
(52.5%) reported having one or more co-morbidities. The most common co-morbidities were having chronic allergies, back pain, COPD, or visual problems (data not shown). Almost 60% of patients was diagnosed with a simple malformation, and 71.2% had a NYHA-I classification. The most common diagnoses were ventricular septal defect (VSD), atrial septal defect (ASD) and patent ductus arteriosus (PDA) (together 36.3%, Table 1) which is in concordance with the distribution of diagnoses of the CONCOR database.
Amount of healthcare use Just over two-thirds of patients (n=574) reported having had contact with their cardiologist during the previous year. Nearly 30% of patients (n=247) reported having had more than one contact (Table 2). Almost 40% of patients (n=324) reported having had contact with their general practitioner regarding their CHD. Patients reported having had less contact with additional medical healthcare providers, ranging from 1.7% for a genetic counsellor to 12.4%
for a nurse. A little over 10% reported having been hospitalised (n=87), of which 56 patients were operated upon. An additional 22 patients were operated upon, but not hospitalised. Furthermore, 10.5% (n=89) had visited the emergency room. Less than 5% reported having had contact with a psychologist and 2.7% with a social worker.
Predictors of healthcare use at follow-up Demographic and clinical characteristics were independently predictive of healthcare use at follow-up (Table 3). More specifically, females reported more additional medical outpatient healthcare use (adjusted odds ratio (OR)=1.73, 95% confidence interval (CI)=1.19–2.51), and older patients reported less psychological healthcare use (adjusted OR=0.93, 95% CI=0.90–0.96). Individuals who reported to have one or more co-morbidities reported to have used less inpatient healthcare (adjusted OR=0.51, 95% CI=0.29–0.87; and adjusted OR=0.48, 95% CI=0.25– 0.93 respectively). Individuals who reported to have more than one co-morbidity reported less psychological healthcare use (adjusted OR=0.38, 95% CI=0.17–0.89). The complexity of the defect was independently predictive of additional medical outpatient, inpatient, and psychological healthcare use (see Table 3 for the specific contrasts). Finally, patients with a poorer functional status (NYHA-II and NYHA-III/IV respectively) reported more primaryand cardiac outpatient (adjusted OR=1.43, 95% CI=1.01– 2.03) and inpatient healthcare use (adjusted OR=4.17, 95% CI=1.41–12.29). In addition, patients who were being treated in a regional centre reported more inpatient (adjusted OR=2.16, 95% CI=1.23–3.71) and psychological healthcare use at follow-up (adjusted OR=7.48, 95% CI=2.86–19.57). Type D personality and quality of life. Patients with a type D personality reported less inpatient (adjusted OR=0.34, 95% CI=0.15–0.77) and psychological healthcare use (adjusted OR=0.17, 95% CI=0.04–0.71) than patients without a type D personality. A poor QoL was independently predictive of more future clinical healthcare use (see Table 3 for the specific significant relations). Negative and positive emotions. High-trait anxiety was predictive of increased psychological healthcare use at followup (adjusted OR=1.11, 95% CI=1.03–1.19). Depressive symptoms, happiness and optimism were not related to healthcare use in the multivariable analyses. Illness perceptions. Illness perceptions were independently predictive of healthcare use at follow-up. In more detail, patients who felt their CHD had a severe impact on their life, reported more primary and cardiac outpatient use (adjusted OR=1.06, 95% CI=1.03–1.09), additional medical outpatient use (adjusted OR=1.11 95% CI=1.06–1.17)
Downloaded from cnu.sagepub.com at University of York on February 19, 2016
76
European Journal of Cardiovascular Nursing 15(1)
Table 2. Healthcare use for the included patients (n=845). Number of contacts Clinical healthcare Primary and cardiac outpatient healthcare Cardiologist General practitioner Additional medical outpatient healthcare Internist Nurse Pulmonologist Genetic counsellor Rheumatologist Surgeon Gynecologist Pacemaker specialist Thrombosis specialist Inpatient healthcare Hospitalisations Emergency room visits Operations Psychological healthcare Psychological healthcare Psychologist Social worker
None
Once
270 (32.0) 518 (61.5)
327 (38.7) 159 (18.9)
751 (91.3) 721 (87.6) 759 (92.4) 800 (98.3) 804 (98.3) 761 (93.1) 778 (95.3) 728 (88.8) 741 (90.0)
46 (5.6) 42 (5.1) 36 (4.4) 9 (1.1) 7 (0.9) 31 (3.8) 22 (2.7) 21 (2.6) 20 (2.4)
758 (89.7) 754 (89.4) 765 (90.7)
57 (6.7) 55 (6.5) 61 (7.2)
781 (95.5) 795 (97.3)
8 (1.0) 4 (0.5)
More than once 247 (29.3) 165 (19.6) 26 (3.2) 60 (7.3) 26 (3.2) 5 (0.6) 7 (0.9) 25 (3.1) 16 (1.9) 71 (8.7) 62 (7.5) 30 (3.6) 34 (4.0) 17 (2.0) 29 (3.5) 18 (2.2)
The total sample size ranges from 814–844 due to missing values. Amount of healthcare use per item was categorised according to no contact, one contact, and more than one contact (similar for hospitalisation, emergency room visit, and operation). All numbers are given in frequencies (%). A time frame of one year was used.
and psychological healthcare use (adjusted OR=1.32, 95% CI=1.17–1.50). Patients who believed that their CHD could be managed by themselves reported more additional medical outpatient use (adjusted OR=1.09, 95% CI=1.03– 1.15) and less inpatient healthcare use (adjusted OR=0.90, 95% CI=0.84–0.97). Patients who believed that their CHD could be managed by treatment reported more inpatient healthcare use (adjusted OR=1.14, 95% CI=1.04–1.26).
Discussion To our knowledge, this is the first longitudinal study of predictors of future healthcare use in adults with CHD. Our main result is that healthcare use was not entirely determined by the clinical variable complexity of the cardiac defect or functional status, but also by psychological characteristics. Patients that reported more healthcare use were patients without a type D personality who had a poor QoL, a strong belief that their CHD had a severe impact on their life and believed their CHD could be managed by themselves or by treatment. The clinical variables complexity of the defect and patients’ functional status are among the strongest predictors of healthcare use (β=1.28, OR=3.59 and β=1.43, OR=4.17 respectively, for inpatient healthcare use). The
beta-values and ORs for the psychological variables are smaller, ranging from β= −0.02, OR=0.98 relating poor QoL (symptoms scale) to primary and cardiac outpatient care, to β= −1.79, OR=0.17 relating type D to psychological healthcare use. As expected clinical variables are stronger predictors, yet psychological characteristics impact on healthcare use as well independently of disease complexity and functional status. The negative relations between the number of comorbidities and inpatient and psychological healthcare use seem counter-intuitive. Leave-one-out analyses identified that this association was caused by the symptom-QoL subscale of the TAAQOL-CHD. This subscale addresses CHD-specific symptoms, like shortness of breath. COPD was found to be one of the most common co-morbidities in this sample, perhaps being responsible for this finding. Indeed, when performing a sensitivity analysis excluding the 152 CHD-patients with co-morbid COPD, the effect of the number of co-morbidities on inpatient and psychological healthcare use decreased. Patients with a type D personality reported less healthcare use at follow-up than patients without such a personality, irrespective of disease complexity and functional status. This is in agreement with the results of our previous cross-sectional study37 which is part of this larger research
Downloaded from cnu.sagepub.com at University of York on February 19, 2016
Downloaded from cnu.sagepub.com at University of York on February 19, 2016
Regional centree Psychological characteristics Personality Type D personalityf Quality of life PCS MCS
Demographic characteristics Sex (female) Age Clinical characteristics Number of co-morbiditiesb 1 co-morbidity 2 or more co-morbidities Disease complexityc Moderate Complex Functional statusd NYHA-II NYHA-III/IV
0.82
0.80
0.21
−0.19 (0.15) 0.83 (0.61–1.12) 0.00 (0.01) 1.00 (0.99–1.02) 0.00 (0.01) 1.00 (0.99–1.02)
0.68
0.24
research-article2014
CNU0010.1177/1474515114555819European Journal of Cardiovascular NursingSchoormans et al.
EUROPEAN SOCIETY OF CARDIOLOGY ®
Original Article
Clinical and psychological characteristics predict future healthcare use in adults with congenital heart disease
European Journal of Cardiovascular Nursing 2016, Vol. 15(1) 72–81 © The European Society of Cardiology 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1474515114555819 cnu.sagepub.com
Dounya Schoormans1–3, Mirjam AG Sprangers1, Joost P van Melle4, Petronella G Pieper4, Arie PJ van Dijk5, Gertjan Tj Sieswerda6, Mariët S Hulsbergen-Zwarts7, Thijs HWM Plokker8, Leo GH Brunninkhuis9, Hubert W Vliegen10 and Barbara JM Mulder2,3
Abstract Background: To deliver adequate care to patients with congenital heart disease (CHD), it is important to know which patients use what type of care. This knowledge is valuable, as modification of these factors may be used as means to regulate healthcare use. Our objective was to examine the predictive value of psychological characteristics for future healthcare use, independent of clinical characteristics. Methods: In total 845 adult CHD-patients participated in a longitudinal questionnaire study, with a two-year followup period. Linear regression analyses with negative binomial log link function were performed predicting healthcare used during the previous year. Psychological predictors were Type D personality, quality of life (QoL), depressive symptoms, trait-anxiety, happiness, optimism, and illness perceptions, independent of the number of co-morbidities, disease complexity and functional status. To control for clustering we included the variable type of centre (regional versus tertiary referral). Results: Patients who reported more healthcare use had a complex defect, a poor functional status, no Type D personality, and a poor QoL. They moreover felt their CHD had a severe impact on their life and believed their CHD could be managed by themselves or treatment. Conclusions: Healthcare use is not entirely determined by disease complexity and functional status but also by psychological patient characteristics. It can by hypothesised that reducing the negative impact experienced and informing patients about strategies to manage their CHD, will modify their future healthcare use. Additional research is necessary to examine this possibility. Keywords Congenital heart disease, healthcare use, type D personality, quality of life, illness perceptions Date received: 23 April 2014; revised: 23 September 2014; accepted: 25 September 2014
1Department
of Medical Psychology, Academic Medical Centre, the
Netherlands 2Department of Cardiology, Academic Medical Centre, the Netherlands 3Interuniversity Cardiology Institute of the Netherlands, the Netherlands 4Department of Cardiology, University Medical Centre Groningen, the Netherlands 5Department of Cardiology, Radboud University Nijmegen Medical Centre, the Netherlands
6Department
of Cardiology, University Medical Centre Utrecht, the Netherlands 7Department of Cardiology, TweeSteden Hospital, the Netherlands 8Department of Cardiology, Saint Antonius Hospital, the Netherlands 9Department of Cardiology, Orbis Medical Centre, the Netherlands 10Leiden University Medical Centre, the Netherlands Corresponding author: Dounya Schoormans, Department of Medical and Clinical Psychology, Tilburg University, P.O. Box 90153, 5000 LE Tilburg, the Netherlands. Email: [email protected]
Downloaded from cnu.sagepub.com at University of York on February 19, 2016
Schoormans et al.
73
Introduction
approval is not required for questionnaire studies under Dutch law, this study was exempted from approval of the Medical Ethics Committee of the Academic Medical Centre in Amsterdam (reference number 09.17.0207). This study was conducted in accordance with the principles of the Declaration of Helsinki.
The number of adults with congenital heart disease (CHD) has grown expansively over the last few decades. As a consequence of their residual lesions, patients with CHD experience a number of physical symptoms,1 psychological impairments (for example a poor quality of life (QoL) or increased levels of depression and anxiety),2–5 and social problems (for example reproduction, employment and financial issues).6 Therefore, it is highly recommended that care for adults with CHD is multidisciplinary.7,8 To facilitate optimal care for CHD-patients, we need to focus on healthcare utilisation provided not only by cardiologists, but also by other healthcare providers such as pulmonologists or social-workers. Studies examining the number of contacts that CHD-patients have with this wide range of healthcare providers is scarce.9,10 Research in other patient populations have shown that in addition to clinical factors, psychological patient characteristics (e.g. QoL, depression, anxiety and illness perceptions) are also predictive of healthcare use.11–15 Knowledge on such psychological factors is valuable, as modification of these factors may be used as means to regulate healthcare use. Furthermore, it is important for healthcare providers and policy makers to know which patients use what type of healthcare. This knowledge will aid in the development of future healthcare facilities and tailor care to those who need it. In the present study we therefore examined the independent predictive value of psychological characteristics on future healthcare use. Included characteristics were personality, QoL, negative and positive emotions (depressive symptoms, trait-anxiety, happiness, optimism) and illness perceptions
Methods Study population and procedure This study is part of a larger longitudinal study on healthcare use of adults with CHD.10 Adult CHD-patients from five tertiary referral and three regional centres were randomly selected from the CONCOR database, a nationwide registry for CHD-patients.16 Upon registration in the CONgenital CORvitia (CONCOR) database, patients gave informed consent to be approached for future scientific research. Inclusion criteria were: being 18 years or older and ever being diagnosed with a CHD. Exclusion criteria for this study were: a diagnosis of Marfan syndrome, being mentally impaired, or illiterate in Dutch. Eligible patients were sent an information letter explaining the nature and procedure of this study. In total, 1670 eligible patients were invited to fill out the questionnaire. After two years another questionnaire was sent to all patients who completed the baseline questionnaire. Since
Measurements Demographic and clinical characteristics. At baseline, patients’ sex, age, co-morbidities, disease complexity and functional status were measured. Sex and age were assessed by self-report. The following previous or ongoing medical conditions were reported: diabetes, kidney disease, gastroenterologic problems, chronic obstructive pulmonary disease (COPD), rheumatic conditions, chronic allergies, chronic skin conditions, problems hearing or seeing, chronic back pain, restrictions in the use of an arm or leg, psychological conditions and being treated for cancer. The type of congenital heart defect was extracted from the CONCOR database and categorised into simple, moderate and complex defects according to Warnes and colleagues.17 Patients’ functional status was operationalised according to the New York Heart Association (NYHA) classification18 and assessed by a validated patient-based question.19 Psychological characteristics. The 14-item DS1420 was used to measure the distressed personality type (type D) – type D personality means having both a high tendency to experience negative affect and to be socially inhibited. The DS14 consists of two subscales; Negative affectivity (seven items) and Social inhibition (seven items). A type D personality is defined as a cut-off score of ≥10 on both subscales of the DS14, and is coded as a dichotomous variable. QoL can be defined in many ways. In general there is consensus that QoL is a multidimensional concept including at least physical, psychological and social functioning. For this study we adopt the following definition related to a person’s health: ‘QoL represents the functional effects of an illness and its consequent therapy upon a patient, as perceived by the patient’ (p.12).21 Patients filled out the Short Form Health Survey-36 (SF-36), yielding a physical component summary (PCS) and mental component summary (MCS) score.22 Additionally, disease-specific QoL was assessed with the TNO/AZL Adult Quality Of Life-CHD (TAAQOL-CHD) forming the three subscales: Symptoms (nine items), Worries (10 items), and Impact cardiac surveillance (seven items).23 Negative emotions were operationalised into measures of depression and trait-anxiety. We used the following definition: ‘a broad range of aversive mood states that form a general distress factor which can be measured either as a transient fluctuation in mood (state) or as a stable individual difference in affective level (trait)’ (p.159).24 The seven-item depression subscale of the Hospital Anxiety
Downloaded from cnu.sagepub.com at University of York on February 19, 2016
74
European Journal of Cardiovascular Nursing 15(1)
and Depression Scale (HADS) was used to measure depression.25 Trait-anxiety was assessed with the 20-item state subscale of the State Trait Anxiety Inventory (STAI).26 We used the following definition of positive emotions: ‘the feelings that reflect a level of pleasurable engagement with the environment such as happiness, joy, excitement, enthusiasm and contentment’. The fouritem Subjective Happiness scale,27 and eight-item Life Orientation Test (LOT)28 were used to assess happiness and optimism, respectively. Illness perceptions were measured by the Illness Perception Questionnaire-Revised form (IPQ-R)29 which was adapted for a CHD-population (see Schoormans et al. for a description of the adaptations made).30 The following six IPQ-R scales measuring illness representations were included in this study. Consequences (five items) describe the effects on physical, psychological and social functioning. Personal (six items) or treatment (three items) control yield questions about whether the illness or its effects can be modified by personal or treatment interference. Illness coherence (four items) measures patients’ understanding of their illness. Timeline-cycle (three items) represents the perceived changeability of the illness and its effects. Emotional representation (five items) addresses the experienced negative emotions (e.g. depressive feelings) caused by the illness. The questionnaires that were used are validated and found to have good psychometric properties.20,27–29,31–34 Please note that one study assessing the validity of the TAAQOL-CHD showed excellent levels of validity,23 whereas others found the levels to be merely adequate.35,36 For all questionnaires, but the DS14, higher scores indicate higher levels of depression, anxiety, happiness, optimism and QoL. The DS14, measuring type D personality is coded as a dichotomous variable. Healthcare use. Two years after the baseline assessment, a study-specific questionnaire (see Schoormans et al. for a detailed description)10 was administered to assess clinical and psychological healthcare use during the previous year for CHD-related reasons. Clinical healthcare yields three types; primary and cardiac outpatient healthcare, additional medical outpatient healthcare, and inpatient healthcare. 1. For primary and cardiac outpatient healthcare the number of contacts with the cardiologist (one item) and general practitioner (one item) were assessed. 2. Additional medical outpatient healthcare (nine items) was operationalised as the number of contacts with nine healthcare providers. 3. Three questions asking how often patients were hospitalised, visited the emergency room, or were operated upon in relation to their CHD, formed inpatient healthcare use.
In addition, psychological healthcare was operationalised by two items assessing the number of contacts with a psychologist and social worker.
Statistical analyses First, we calculated the frequency and percentages for all items to identify the amount of primary and cardiac outpatient, additional medical outpatient, inpatient and psychological healthcare use. Second, we examined the predictive value of psychological characteristics for different types of healthcare use at follow-up, while controlling for demographic and clinical characteristics, and type of centre. Given that data was collected from multiple centres, both regional and tertiary referral, we corrected for clustering by adding a categorical variable representing these two types of centres. Four generalised linear models with negative binominal log link function were run separately for the different types of healthcare use. Sex, age, the number of co-morbidities, disease complexity, functional status, type of centre, type D personality, the four QoL scales, depressive symptoms, trait-anxiety, happiness, optimism and the six illness perception scales were simultaneously included in the analyses. In order to include functional status as a predictor, NYHA-III and NYHA-IV were lumped together leading to enough patients in each category. Exploratory analyses showed that there was no multicollinearity among the predictors (data not shown). The Impact subscale of the TAAQOL-CHD was not included, because items overlap with the healthcare use items. Heterogeneity in our sample was corrected using robust standard errors. A significance level of 0.05 was employed. All analyses were conducted in SPSS 16.0 (SPSS, Chicago, Illinois, USA).
Results Response rate and demographic and clinical characteristics At baseline 1670 eligible patients were invited to participate. A total of 1109 patients filled out the first questionnaire (response rate 66.4%). Non-responders were not different from the participants who were included at baseline (n=1109) in terms of type of CHD and complexity. Two years later the follow-up questionnaire was sent to these respondents (n=1100, nine patients died during the follow-up period). Of the 1100 eligible patients, 845 patients returned the second questionnaire (response rate 76.8%). Patients included in the follow-up were not significantly different from first-time respondents regarding demographic and clinical characteristics. More than half of the 845 patients were female (Table 1). Median age was 39 years. Just over half of all patients
Downloaded from cnu.sagepub.com at University of York on February 19, 2016
75
Schoormans et al. Table 1. Demographic and clinical characteristics for the included patients (n=845). Demographic and clinical characteristics
n (%)
Sex (female) Age in median years (range) Number of co-morbidities No co-morbidity 1 co-morbidity 2 or more co-morbidities Disease complexity17 Simple Moderate Complex Functional status18 NYHA-I NYHA-II NYHA-III/IV Congenital heart defect VSD/ASD/ASD-I/PDA TOF/PS BAV/AS/SubvAS CoA TGA/ccTGA Other CHD-diagnoses
457 (54.7) 39 (18–85) 401 (47.5) 225 (26.6) 219 (25.9) 502 (59.5) 270 (32.0) 72 (8.5) 591 (71.2) 175 (21.1) 64 (7.7) 306 (36.3) 162 (19.2) 121 (14.3) 106 (12.6) 28 (3.3) 121 (14.3)
AS: aortic stenosis; ASD: atrial septum defect; ASD-I: atrial septum defect primum type I; BAV: bicuspid aortic valve; ccTGA: congenital corrected transposition of the great arteries; CHD: congenital heart defect; CoA: coarctation of the aorta; NYHA: New York Heart Association functional class; PDA: patent ductus arteriosus; PS: pulmonary stenosis; SubvAS: subvalvar aorta stenosis; TGA: transposition of the great arteries; TOF: tetralogy of Fallot; VSD: ventricular septum defect. The total sample size ranges from 830–845 due to missing values. Numbers are given in frequencies (valid %), age in median years (range).
(52.5%) reported having one or more co-morbidities. The most common co-morbidities were having chronic allergies, back pain, COPD, or visual problems (data not shown). Almost 60% of patients was diagnosed with a simple malformation, and 71.2% had a NYHA-I classification. The most common diagnoses were ventricular septal defect (VSD), atrial septal defect (ASD) and patent ductus arteriosus (PDA) (together 36.3%, Table 1) which is in concordance with the distribution of diagnoses of the CONCOR database.
Amount of healthcare use Just over two-thirds of patients (n=574) reported having had contact with their cardiologist during the previous year. Nearly 30% of patients (n=247) reported having had more than one contact (Table 2). Almost 40% of patients (n=324) reported having had contact with their general practitioner regarding their CHD. Patients reported having had less contact with additional medical healthcare providers, ranging from 1.7% for a genetic counsellor to 12.4%
for a nurse. A little over 10% reported having been hospitalised (n=87), of which 56 patients were operated upon. An additional 22 patients were operated upon, but not hospitalised. Furthermore, 10.5% (n=89) had visited the emergency room. Less than 5% reported having had contact with a psychologist and 2.7% with a social worker.
Predictors of healthcare use at follow-up Demographic and clinical characteristics were independently predictive of healthcare use at follow-up (Table 3). More specifically, females reported more additional medical outpatient healthcare use (adjusted odds ratio (OR)=1.73, 95% confidence interval (CI)=1.19–2.51), and older patients reported less psychological healthcare use (adjusted OR=0.93, 95% CI=0.90–0.96). Individuals who reported to have one or more co-morbidities reported to have used less inpatient healthcare (adjusted OR=0.51, 95% CI=0.29–0.87; and adjusted OR=0.48, 95% CI=0.25– 0.93 respectively). Individuals who reported to have more than one co-morbidity reported less psychological healthcare use (adjusted OR=0.38, 95% CI=0.17–0.89). The complexity of the defect was independently predictive of additional medical outpatient, inpatient, and psychological healthcare use (see Table 3 for the specific contrasts). Finally, patients with a poorer functional status (NYHA-II and NYHA-III/IV respectively) reported more primaryand cardiac outpatient (adjusted OR=1.43, 95% CI=1.01– 2.03) and inpatient healthcare use (adjusted OR=4.17, 95% CI=1.41–12.29). In addition, patients who were being treated in a regional centre reported more inpatient (adjusted OR=2.16, 95% CI=1.23–3.71) and psychological healthcare use at follow-up (adjusted OR=7.48, 95% CI=2.86–19.57). Type D personality and quality of life. Patients with a type D personality reported less inpatient (adjusted OR=0.34, 95% CI=0.15–0.77) and psychological healthcare use (adjusted OR=0.17, 95% CI=0.04–0.71) than patients without a type D personality. A poor QoL was independently predictive of more future clinical healthcare use (see Table 3 for the specific significant relations). Negative and positive emotions. High-trait anxiety was predictive of increased psychological healthcare use at followup (adjusted OR=1.11, 95% CI=1.03–1.19). Depressive symptoms, happiness and optimism were not related to healthcare use in the multivariable analyses. Illness perceptions. Illness perceptions were independently predictive of healthcare use at follow-up. In more detail, patients who felt their CHD had a severe impact on their life, reported more primary and cardiac outpatient use (adjusted OR=1.06, 95% CI=1.03–1.09), additional medical outpatient use (adjusted OR=1.11 95% CI=1.06–1.17)
Downloaded from cnu.sagepub.com at University of York on February 19, 2016
76
European Journal of Cardiovascular Nursing 15(1)
Table 2. Healthcare use for the included patients (n=845). Number of contacts Clinical healthcare Primary and cardiac outpatient healthcare Cardiologist General practitioner Additional medical outpatient healthcare Internist Nurse Pulmonologist Genetic counsellor Rheumatologist Surgeon Gynecologist Pacemaker specialist Thrombosis specialist Inpatient healthcare Hospitalisations Emergency room visits Operations Psychological healthcare Psychological healthcare Psychologist Social worker
None
Once
270 (32.0) 518 (61.5)
327 (38.7) 159 (18.9)
751 (91.3) 721 (87.6) 759 (92.4) 800 (98.3) 804 (98.3) 761 (93.1) 778 (95.3) 728 (88.8) 741 (90.0)
46 (5.6) 42 (5.1) 36 (4.4) 9 (1.1) 7 (0.9) 31 (3.8) 22 (2.7) 21 (2.6) 20 (2.4)
758 (89.7) 754 (89.4) 765 (90.7)
57 (6.7) 55 (6.5) 61 (7.2)
781 (95.5) 795 (97.3)
8 (1.0) 4 (0.5)
More than once 247 (29.3) 165 (19.6) 26 (3.2) 60 (7.3) 26 (3.2) 5 (0.6) 7 (0.9) 25 (3.1) 16 (1.9) 71 (8.7) 62 (7.5) 30 (3.6) 34 (4.0) 17 (2.0) 29 (3.5) 18 (2.2)
The total sample size ranges from 814–844 due to missing values. Amount of healthcare use per item was categorised according to no contact, one contact, and more than one contact (similar for hospitalisation, emergency room visit, and operation). All numbers are given in frequencies (%). A time frame of one year was used.
and psychological healthcare use (adjusted OR=1.32, 95% CI=1.17–1.50). Patients who believed that their CHD could be managed by themselves reported more additional medical outpatient use (adjusted OR=1.09, 95% CI=1.03– 1.15) and less inpatient healthcare use (adjusted OR=0.90, 95% CI=0.84–0.97). Patients who believed that their CHD could be managed by treatment reported more inpatient healthcare use (adjusted OR=1.14, 95% CI=1.04–1.26).
Discussion To our knowledge, this is the first longitudinal study of predictors of future healthcare use in adults with CHD. Our main result is that healthcare use was not entirely determined by the clinical variable complexity of the cardiac defect or functional status, but also by psychological characteristics. Patients that reported more healthcare use were patients without a type D personality who had a poor QoL, a strong belief that their CHD had a severe impact on their life and believed their CHD could be managed by themselves or by treatment. The clinical variables complexity of the defect and patients’ functional status are among the strongest predictors of healthcare use (β=1.28, OR=3.59 and β=1.43, OR=4.17 respectively, for inpatient healthcare use). The
beta-values and ORs for the psychological variables are smaller, ranging from β= −0.02, OR=0.98 relating poor QoL (symptoms scale) to primary and cardiac outpatient care, to β= −1.79, OR=0.17 relating type D to psychological healthcare use. As expected clinical variables are stronger predictors, yet psychological characteristics impact on healthcare use as well independently of disease complexity and functional status. The negative relations between the number of comorbidities and inpatient and psychological healthcare use seem counter-intuitive. Leave-one-out analyses identified that this association was caused by the symptom-QoL subscale of the TAAQOL-CHD. This subscale addresses CHD-specific symptoms, like shortness of breath. COPD was found to be one of the most common co-morbidities in this sample, perhaps being responsible for this finding. Indeed, when performing a sensitivity analysis excluding the 152 CHD-patients with co-morbid COPD, the effect of the number of co-morbidities on inpatient and psychological healthcare use decreased. Patients with a type D personality reported less healthcare use at follow-up than patients without such a personality, irrespective of disease complexity and functional status. This is in agreement with the results of our previous cross-sectional study37 which is part of this larger research
Downloaded from cnu.sagepub.com at University of York on February 19, 2016
Downloaded from cnu.sagepub.com at University of York on February 19, 2016
Regional centree Psychological characteristics Personality Type D personalityf Quality of life PCS MCS
Demographic characteristics Sex (female) Age Clinical characteristics Number of co-morbiditiesb 1 co-morbidity 2 or more co-morbidities Disease complexityc Moderate Complex Functional statusd NYHA-II NYHA-III/IV
0.82
0.80
0.21
−0.19 (0.15) 0.83 (0.61–1.12) 0.00 (0.01) 1.00 (0.99–1.02) 0.00 (0.01) 1.00 (0.99–1.02)
0.68
0.24
