Self-care trajectories of patients with coronary heart disease: a longitudinal, observational study

Abstract

Aims

To determine if distinct trajectories of coronary heart disease (CHD) self-care behaviours could be identified, linked to differences in quality of life (QoL), and predicted based on baseline characteristics.

Methods and results

A secondary analysis of a prospective, longitudinal, observational study. Patients with CHD answered questionnaires at study enrolment and six months later: Self-Care of Coronary Heart Disease Inventory (three subscales: maintenance, management, and confidence, scored 0–100, higher score = better self-care), Hospital Anxiety and Depression Scale, 12-Item Short Form Survey, 16-Item European Health Literacy Survey Questionnaire, and CHD Education Questionnaire. Latent growth mixture modelling was used to identify distinct self-care trajectories over time. On average, patients (n = 430, mean age 64.3 ± 8.9, 79% male) reported inadequate self-care (maintenance 61.6 ± 15.4, management 53.5 ± 18.5) at enrolment. Two distinct trajectories of self-care behaviours were identified: first, an ‘inadequate-and-worsening’ (IN-WORSE) trajectory (57.2%), characterized by inadequate self-care, improvement in maintenance (4.0 ± 14.5-point improvement, P < 0.001), and worsening of management over time (6.3 ± 24.4-point worsening, P = 0.005). Second, an ‘inadequate-but-maintaining’ (IN-MAIN) trajectory (42.8%), characterized by inadequate self-care, improvement in maintenance (5.0 ± 16.2-point improvement, P < 0.001), and stability in management over time (0.8 ± 21.9-point worsening, P = 0.713). In comparison, patients in the IN-WORSE trajectory had less favourable characteristics (including lower health literacy, knowledge, confidence) and significantly lower QoL. Not attending rehabilitation (OR 2.175; CI 1.020–4.637, P = 0.044) and older age (OR 0.959; CI 0.924–0.994, P = 0.024) predicted (IN-WORSE) trajectory inclusion.

Conclusion

Two self-care trajectories were identified, both suboptimal. Rehabilitation predicted membership in the more favourable trajectory and some positive characteristics were identified among patients in that group. Therefore, interventions supporting these factors may benefit patients’ self-care and QoL.

Graphical Abstract

Open in new tabDownload slide

Introduction

Cardiovascular disease, including coronary heart disease (CHD), is the leading cause of death globally.¹ Consequently, cardiovascular risk factor reduction is central to reducing morbidity and mortality. Self-care is an essential component of risk factor management to improve both cardiovascular health and quality of life (QoL).^2,3 Self-care recommendations for cardiovascular health typically include regular physical activity, a healthy diet, smoking avoidance, and maintaining normal blood pressure, body mass index, fasting plasma glucose, and cholesterol levels.⁴ However, a large majority of patients with CHD does not achieve lifestyle goals or therapeutic targets for secondary prevention.⁵ As self-care is complex and multidimensional, several factors contribute to this shortfall in risk reduction behaviours. For example, self-care of CHD and the associated change in lifestyle can be challenging, particularly for older patients with anxiety and depression.⁶ Self-care also is a dynamic process of maintaining health and managing illness with trajectories that change over time.⁷

Several investigators have explored self-care trajectories among patients with heart failure.^8–11 Trajectories of QoL and depression also have been studied in CHD;^12–14 but trajectories examining the influence of self-care on QoL have not been identified among adults with CHD so little is known about how CHD self-care influences QoL over time. Therefore, the objectives of this study were to (1) determine if distinct trajectories of CHD self-care behaviours could be identified following hospitalization, (2) link trajectories of CHD self-care behaviours to differences in physical and mental QoL over time, and (3) identify which patient characteristics are helpful in predicting trajectory membership.

Theory of self-care in chronic illness

The Theory of Self-Care in Chronic Illness informed this study.⁷ Self-care is a naturalistic decision-making process for maintaining health and managing chronic illness. Two key behaviours that are integral to the theory are self-care maintenance and self-care management. Self-care maintenance comprises the activities to maintain physical and emotional stability, such as taking medications and eating a healthy diet. Self-care management comprises the actions taken to ameliorate signs and symptoms when they occur with actions such as seeking healthcare or taking medications. Patients need skills in reflection and decision-making in their self-care as well as self-efficacy and disease-related knowledge.^7,15 Patients with lower knowledge may be less capable of performing lifestyle changes or using the healthcare system effectively.

Methods

Study design

A secondary analysis of a prospective, longitudinal, observational study was conducted. Data were collected between 2017 and 2019 with questionnaires: at discharge from a hospitalization for CHD and six months later.

Settings

The study was conducted in the two main hospitals in Iceland, a university hospital that runs the country’s only specialized cardiac units (including cardiac surgery and catheterization (PCI), and a teaching hospital located in northern Iceland.

Participants

The participants were all consecutive patients admitted acutely or electively to the hospital for CHD (e.g. for ST-segment elevation myocardial infarction, non-ST-segment elevation myocardial infarction, acute coronary syndrome, elective angiography, or coronary artery bypass graft). Patients were eligible if they were aged 18 or older, could understand Icelandic, and did not have a documented cognitive impairment or other conditions preventing them from participating in the study.

Procedure

Eligible patients who agreed to learn about the study received verbal and written information and signed a consent form if willing to participate. Participants answered a questionnaire either on paper or electronically as preferred, before hospital discharge or within 24 h of discharge. A pre-stamped envelope was included for those answering on paper. The questionnaires were repeated six months after discharge.

Measurement

Self-care was measured with the validated Self-Care of Coronary Heart Disease Inventory, version 2.2 (SC-CHDI)¹⁶ that consists of 22 items with three subscales i.e. maintenance (10 items), management (six items), and confidence (six items). Self-care maintenance scale items has five response options ranging from 1 (never/rarely) to 5 (always/daily), self-care management scale items were ranked on five response options ranging from 1 (not quickly/likely/sure) to 5 (very quickly/likely/sure), and items of self-confidence scale has four options of 1 (not sure) to 4 (very sure). For each subscale, a standardized score of 0–100 was calculated with a higher score indicating better self-care or confidence and a cut-off of ≥70 defined as adequate self-care.¹⁷ A change of eight points between time points has previously been suggested as a theoretical threshold for clinically relevant difference.¹⁸

Health-related QoL was assessed with the SF-12® Health Survey (SF-12), which has two components, mental (six items) and physical (12 items), and standardized scores that range from 0 to 100, with higher scores indicating better physical and mental health functioning.¹⁹ SF-12 has acceptable psychometric properties in the CHD population.²⁰

Symptoms of anxiety and depression were assessed with the Hospital Anxiety and Depression Scale (HADS), which consists of two independent subscales, HADS-A (symptoms of anxiety, seven items) and HADS-D (symptoms of depression, seven items). Possible scores from each subscale range from 0 to 21, with higher scores indicating more symptoms.^21,22 The Icelandic version of the HADS has demonstrated satisfactory psychometric properties among patients with CHD. Evaluated with ordinal alpha, both subscales demonstrated good internal consistency, 0.90 for anxiety and 0.88 for depression.²³

Health literacy was assessed with the Health Literacy Survey Questionnaire, HLS-EU-Q16-IS, a 16-item instrument with possible scores ranging from 0 to 16, and higher scores indicating better health literacy.^24,25 The HLS-EU-Q16 has been validated in the general population^24,25 translated into several languages, including Icelandic with satisfactory psychometric properties^24,25 and used to compare health literacy of people with and without cardiovascular disease.²⁶

Disease-related knowledge was assessed with the validated Coronary Artery Disease Education Questionnaire—Short Version (CADE-Q-SV) that consists of 20 questions, with possible scores ranging from 0 to 20, with higher scores indicating better knowledge.²⁷ The Icelandic version has demonstrated satisfactory psychometric properties among this patient cohort (Cronbach’s alpha 0.74).²⁸

Background information was collected at enrolment from medical records and the self-reported questionnaires on age, gender, education, living conditions (alone or with others, rural/urban), sufficiency of income, and hospital admission (previous hospital admission for CHD, admission diagnosis, elective/acute).

Statistical analysis

Characteristics of the sample at large were presented as means ± standard deviations or n (%) as appropriate. To address objective 1, latent growth mixture modelling²⁹ was used to identify distinct trajectories of CHD self-care behaviours (i.e. self-care maintenance and self-care management) over time between enrolment and the six-month follow-up. Ram criteria³⁰ were followed to identify the number of distinct trajectories. In brief, model entropy (closest to 1.0 most favourable), classification probabilities [closest to 1.0 (indicating less uncertainty about which trajectory participants belonged to) most favourable], trajectory size (no <5% of the sample), the parametric bootstrap likelihood ratio test (statistical significance), and the Lo–Mendell–Rubin adjusted likelihood ratio test (statistical significance) were used to compare models with different numbers of trajectories.²⁹ Trajectories were compared (between-group difference with t-tests) and described (within-trajectory change with paired t-tests) as well as presented visually. Our focus was on self-care behaviours (maintenance and management), but we also present data on self-care confidence as this metric of self-efficacy in the self-care behaviours is frequently presented along with data on actual self-care behaviours.⁷ At the point of trajectory identification, there was a stopping point for the full team to deliberate on key differentiating characteristics between trajectories, and by consensus determine the most appropriate nomenclature. To address objective 2, unadjusted differences in QoL (i.e. SF-12 physical functioning and mental health scores at enrolment and at 6 months) were compared between CHD self-care trajectories using between-group and within-trajectory t-tests. To address objective 3, unadjusted differences between trajectories in demographic and clinical variables at enrolment were evaluated using comparative statistics (i.e. t-tests for continuous variables, and χ² for categorical variables). Characteristics with an unadjusted significance of <0.20 were considered as candidate predictors in multivariate modelling,^31,32 because how variables function outside of a model does not necessarily predict how they function in a multivariate model, and so that potentially clinically relevant variables are not missed. Finally, multivariate logistic regression modelling was used to identify significant determinants of membership in the more favourable trajectory vs. less favourable trajectory. The results are reported in adjusted odds ratios, 95% confidence intervals, and statistical significance. Unless otherwise noted, an alpha of ≤0.05 was used to determine statistical significance. Latent-grown mixture modelling was performed using MPlus v8.7 (Los Angeles, CA); all other analysis was performed in StataMP, v.17 (College Station, TX). The results reported here represent exploratory and secondary analyses of a CHD cohort study that has been described elsewhere; hence, no a priori sample size estimation or power analysis was completed. But, multiple latent classes and trajectories have been identified using models with more indicators and smaller samples sizes^9,33 compared with the present analysis (four model indicators and a sample size of 430), and our n-to-items ratio exceeds sample size recommendations for related approaches (10–20:1).³⁴ As such, we were confident in the adequacy of the sample size to detect multiple trajectories. Missing data (14.2% including self-care measures at enrolment or follow-up) were handled using full information maximum likelihood estimation.

Ethical considerations

The study was approved by the National Bioethics Committee for Medical Research Ethics (17–159) and permission to examine patient records was granted by hospital authorities. The study conforms with the Declaration of Helsinki.³⁵ The patients received both verbal and written information about the study before signing informed consent forms.

Results

A total of 446 patients were enrolled in the study, and 430 had completed data to be analysed. Characteristics of the sample and a comparison between the two trajectories are presented in Table 1.

Trajectories of self-care behaviours

At enrolment, self-care maintenance (61.6 ± 15.4) and self-care management (53.5 ± 18.5) were inadequate on average. Two distinct trajectories of CHD self-care behaviours were identified in the best-fitting growth mixture model (entropy = 0.92, classification probabilities = 0.96 and 0.87, parametric bootstrap likelihood ratio P < 0.001, Lo–Mendell–Rubin adjusted likelihood ratio test = 9.96, P = 0.005). Models identifying three to five potential trajectories had non-significant results and entropy below 0.65, indicating excessive modelling uncertainty.

The first trajectory (n = 246) included 57.2% of the sample and was characterized by inadequate and worse self-care maintenance and management at both enrolment and the six-month follow-up (Figure 1). Participants in this trajectory had a statistically significant improvement in self-care maintenance between enrolment and the six-month follow-up (4.0 ± 14.5-point improvement, P < 0.001), and a statistically significant worsening of self-care management between enrolment and the six-month follow-up (6.3 ± 24.4-point worsening, P = 0.005). Based on these trajectory characteristics, we labelled this first trajectory of CHD self-care behaviours as ‘Inadequate and worsening’ abbreviated as ‘IN-WORSE.’

The second trajectory (n = 184) included 42.8% of the sample and was characterized by comparatively better self-care maintenance and management at both enrolment and the six-month follow-up (Figure 1). Participants in this trajectory had a modest but statistically significant improvement in self-care maintenance between enrolment and the six-month follow-up (5.0 ± 16.2-point improvement, P < 0.001), and no significant change in self-care management between enrolment and the six-month follow-up (0.8 ± 21.9-point worsening, P = 0.713). Based on these trajectory characteristics, we labelled this second trajectory of CHD self-care behaviours as ‘Inadequate but maintaining’ abbreviated as ‘IN-MAIN.’

For further identification of the differences in self-care maintenance and management, the two trajectories of self-care were compared on item level. There was no difference between the groups in keeping doctor or nurse appointments, taking aspirin or other blood thinners, or taking medication as prescribed (Table 2). But patients in the IN-WORSE trajectory scored significantly lower on every other self-care maintenance and management behaviour compared with patients in the IN-MAIN trajectory.

Trajectories of coronary heart disease self-care behaviours and quality of life

Differences in physical functioning and mental health between the two trajectories of CHD self-care behaviours are presented in Figure 2. Participants in the IN-WORSE trajectory of CHD self-care behaviour had worse physical functioning and mental health at both enrolment and the six-month follow-up compared to the IN-MAIN trajectory. However, participants in both trajectories had significant improvements in physical functioning between enrolment and the six-month follow-up (IN-WORSE trajectory: average 23.3 ± 33.5-point improvement; IN-MAIN trajectory: average 18.9 ± 36.2-point improvement, both P < 0.001). Similarly, participants in both trajectories had significant improvements in mental health between enrolment and the six-month follow-up (IN-WORSE trajectory: average 4.3 ± 16.6-point improvement; IN-MAIN trajectory: average 5.7 ± 17.9-point improvement, both P < 0.001).

Trajectories of coronary heart disease self-care behaviours and self-care confidence

Confidence in self-care behaviours was significantly lower in the IN-WORSE trajectory both at baseline (−25.11 points, P < 0.001) and at follow-up (−34.93 points, P < 0.001) compared with the IN-MAIN trajectory. Importantly, confidence in self-care plummeted over time in the IN-WORSE self-care trajectory (Figure 3).

Determinants of trajectories of coronary heart disease self-care behaviours

Unadjusted differences between the trajectories are presented in Table 1. The results of the multivariate logistic regression model predicting membership in the IN-MAIN trajectory vs. the IN-WORSE trajectory are presented in Table 3. Each additional year of age was associated with a decrease in the odds of membership in the IN-MAIN trajectory of CHD self-care behaviours. In addition, participation in rehabilitation after the index hospitalization was associated with greater odds of being in the IN-MAIN trajectory of CHD self-care behaviours.

Discussion

Two trajectories of self-care were identified in this sample of 430 adults with CHD. Self-care of patients was low and inadequate in both trajectories at enrolment. Older age and not participating in rehabilitation predicted membership in the trajectory characterized by inadequate and worsening self-care over time.

Self-care maintenance improved in both groups, reaching adequate levels only for the IN-MAIN group at the six-month follow-up. Maintenance reflects the healthy behaviour expected of patients with CHD, such as risk factor management and lifestyle changes. We saw significant differences between the two trajectories in all self-care behaviours except for medication adherence and keeping appointments. These results are in concordance with previous studies, for example the EUROASPIRE studies that reported most patients with CHD have difficulties with maintaining a healthy lifestyle and managing risk factors.^6,36 The fact that self-care maintenance improved may reflect the motivation of patients to address their risk factors and change their lifestyle after a cardiac event. Frequently, patient education and rehabilitation is focused on adherence to prescribed regimens. Hospitalization can serve as a potent health warning that may motivate patients to improve their self-care maintenance. While the IN-MAIN group managed to improve their self-care maintenance to reach the adequate level of 70, the IN-WORSE group did not. Factors such as having less confidence, knowledge, health literacy and poorer mental health may explain the difference.

Self-care management did not improve over time in both groups. Self-care management involves responding to symptoms when they occur. Several factors are associated with difficulty in accurately assessing symptoms and consequently with symptom management.^2,37 For example, age, sex, and depression have been reported to interfere with a timely response to symptoms.^37–39 In this sample, symptoms of depression were significantly higher in the IN-WORSE group.

Overall, it is interesting to compare patients in the two trajectories. There was a trend that people in the IN-MAIN trajectory had more favourable characteristics for self-care that may be hypothesis-generating for future research. They were younger, more had university education, sufficient income, lived with others, and had someone to confide in. They had previous CHD admission (experience), adequate health literacy, better disease-related knowledge and self-care confidence, participated in rehabilitation, and had fewer symptoms of anxiety and depression than those belonging to the IN-WORSE trajectory. Consequently, they had better physical and mental functioning, reflecting better health-related QoL.

One of the main determinants of self-care trajectory was the patients’ rehabilitation status. Although comprehensive cardiac rehabilitation (CR) historically has been recommended as secondary prevention in CHD, there are several issues ongoing globally concerning effectiveness, access, and attendance of current programmes. Cardiac rehabilitation is only available in half of 203 countries recently investigated, and capacity is grossly insufficient where programmes are offered, reducing benefits associated with participation.⁴⁰ Although CR seems to be cost-effective, especially with exercise as a component, the most cost-effective design of CR has yet to be determined.⁴¹ There is insufficient and even contradictory evidence on the long-term benefits of contemporary exercise-based CR on QoL^42,43 and psychosocial and diet recommendations in clinical guidelines on CR are contradictory or inconsistent.⁴⁴ Several enrolment obstacles for patients to participate in conventional centre-based rehabilitation also have been identified, such as schedule flexibility, time commitment, travel distance, cost, patient preference, perceived need, clinical status, and social support.^45,46 These factors may become even more significant among older populations. Quality-assured home-based and technology-based models of CR have been recommended as an adjunct or alternative to traditional programmes to improve access.⁴⁷ Technology-assisted CR (e-Health, telehealth/telemedicine, and mHealth) has comparable results to conventional centre-based rehabilitation and is a potential alternative to remove obstacles to attendance in the current programmes.⁴⁸ Finally, patients are open to the use of technology to partially replace the traditional face-to-face, centre-based rehabilitation approach in CR.⁴⁶

The results of this study also indicate the importance of confidence, disease-related knowledge and health literacy for the CHD self-care of patients. When health literacy is limited, it can prevent patients and their families from developing the knowledge, skills, and confidence needed for their engagement and participation in their own care.⁴⁹ Knowledge is an important prerequisite for self-care² and patient education focused on improving knowledge, skills, and attitudes can help with behaviour change.⁵⁰ Patients in this cohort reported limited disease-related knowledge and several unfulfilled educational needs that may have affected their self-care.²⁸ Therefore, interventions to improve health literacy should be a priority of health providers. As there is growing evidence that patient education delivered with technology can help people with CHD modify their risk factors,^51,52 the use of e-health might facilitate more effective implementation of educational activities.

Last, but not least, this study showed that significant differences in QoL were linked to the two self-care trajectories. Although mental and physical QoL improved in both groups over time, participants in the IN-WORSE group had comparatively worse physical and mental QoL at both time points. Evidence of the association between self-care and QoL is slowly emerging but is inconclusive. As regards heart failure, the evidence is still conflicting^53,54 and the literature is limited within the CHD patient population although QoL has been found to be better among patients who exercise and consume an appropriate diet after elective PCI.⁵⁵ More research is therefore needed in the future on this association and related factors.

Clinical implications

There is considerable room for improvement in self-care among patients with CHD. Better understanding of how self-care changes over time, what factors are associated with different trajectories, and how these factors may affect outcomes can help clinicians tailor their care to meet patients’ needs better.

Limitations

Several limitations should be considered when interpreting these results. First, these data are derived from a relatively small and homogenous population that may be a threat to external validity. Second, the vast majority of data was collected by self-report. Although that is the standard for measuring self-care behaviours broadly, there may be unresolved threats to internal validity. Third, this analysis was secondary to the primary aims of the study; as such, there is a risk of false discovery even with a relatively large sample size. Collectively, these data will need to be confirmed to understand the ways in which our findings are globally generalizable.

Conclusions

Two self-care trajectories of patients with CHD were identified and found to be suboptimal. While self-care maintenance improved slightly over time, self-care management either was maintained at a suboptimal level or decreased. Participation in rehabilitation predicted membership in the more favourable trajectory and some positive characteristics were identified among patients in that group. Therefore, interventions supporting these factors may benefit patients’ self-care and QoL after a cardiac event.

Acknowledgements

We thank all the patients who participated in the study and the nurses who assisted with patient recruitment and data collection.

Funding

The study was funded by the Icelandic Regional Development Institute; the Landspitali University Hospital Research Fund; the Akureyri Hospital Research Fund; the University of Akureyri Research Fund; the Icelandic Nursing Association Research Fund; the KEA Research fund; and the Akureyri Heart Association.

Data availability

The data underlying this article cannot be shared for ethical/privacy reasons.

References

Author notes