https://orcid.org/0000-0003-3189-7888
Abstract
This study aimed to uncover hidden patterns and predictors of symptom multi-trajectories within 30 days after discharge in patients with heart failure and assess the risk of unplanned 30-day hospital readmission in different patterns.
The study was conducted from September 2022 to September 2023 in four third-class hospitals in Tianjin, China. A total of 301 patients with heart failure were enrolled in the cohort, and 248 patients completed a 30-day follow-up after discharge. Three multi-trajectory groups were identified: mild symptom status (24.19%), moderate symptom status (57.26%), and severe symptom status (18.55%). With the mild symptom status group as a reference, physical frailty, psychological frailty, and comorbid renal dysfunction were predictors of the moderate symptom status group. Physical frailty, psychological frailty, resilience, taking diuretics, and comorbid renal dysfunction were predictors of the severe symptom status group. Compared with the mild symptom status group, the severe symptom status group was significantly associated with high unplanned 30-day hospital readmission risks.
This study identified three distinct multi-trajectory groups among patients with heart failure within 30 days after discharge. The severe symptom status group was associated with a significantly increased risk of unplanned 30-day hospital readmission. Common and different factors predicted different symptom multi-trajectories. Healthcare providers should assess the physical and psychological frailty and renal dysfunction of patients with heart failure before discharge. Inpatient care aimed at alleviating physical and psychological frailty and enhancing resilience may be important to improve patients’ symptom development post-discharge.
Three symptom multi-trajectories were identified in patients with heart failure within 30 days after discharge: mild symptom status (24.19%), moderate symptom status (57.26%), and severe symptom status (18.55%).
Patients with severe symptom status were 14.766 times more likely to have unplanned readmission within 30 days after discharge than those with mild symptom status.
Physical frailty, psychological frailty, resilience, taking diuretics, and comorbid renal dysfunction had significant effects on patients belonging to the severe symptom status group.
Introduction
Heart failure (HF) is a progressive and symptomatic clinical syndrome that affects over 64 million people worldwide.1 The importance of symptom management in patients with HF has been emphasized by HF associations in many countries.2,3 As HF progresses, patients tend to experience a significant symptom burden that can be comparable to that of patients with advanced cancer.4 Despite receiving optimal treatment, patients with HF still endure significant distress from symptoms, including persistent HF-specific physical symptoms such as dyspnoea, fatigue, and oedema, as well as psychological symptoms such as depression and anxiety.5 These symptoms have negative impacts on physical functioning, activities of daily living, and quality of life of patients with HF.6 Currently, unplanned readmission within 30 days is a widely accepted measure of HF care quality and is associated with medium-term mortality.7 It has been observed that 15.1–23.9% of patients with HF experience readmission within 30 days after discharge, with a median length of stay of ∼12 days.8,9 The risk of unplanned readmission is primarily determined by patient-specific factors rather than hospital-related factors.10 The poorly controlled HF symptoms are the most common patient-specific causes of HF hospital readmission, and dyspnoea and fatigue are frequently reported symptoms by individuals seeking emergency care for HF.11
The symptoms experienced by patients with HF are characterized by their co-occurrence, interplay, and dynamic nature. On average, patients with HF experience 7–19 symptoms, and Chinese patients report an average of 11.9 symptoms.12 Symptoms that occur in one patient often interact with each other. Dyspnoea, a classic symptom of HF, has a complex relationship with other symptoms. For example, patients with probable depression are five times more likely to experience shortness of breath than those without depression.13 Dyspnoea has also been linked to fatigue, oedema, depression, and anxiety.14 Previous studies have observed bidirectional relationships between fatigue and depression. Depression is an initial causative factor in worsening fatigue and physical activity.15 Meanwhile, fatigue can predict the onset of depression after six months.16 The illness trajectory of HF is complex and unpredictable after discharge, with different developmental trends due to inter-individual differences, such as age, comorbidities, and the severity of the condition.4 Therefore, it is crucial to fully understand the unique developmental trajectories and characteristics of symptoms co-occurred commonly in patients with HF.
Previous studies investigating the trajectories of HF symptoms have certain limitations. Firstly, focusing on a single symptom can lead to overlooking the interactions and interdependence among symptoms. While previous studies have provided valuable insights into specific symptom trajectories, such as fatigue,17 depression,18 and appetite,19 they may not capture the multi-trajectories of symptoms experienced mainly by HF patients. Secondly, integrating various symptoms into an overall level overlooks the inconsistency in symptom levels among patients.20,21 While existing studies have confirmed the heterogeneous trajectory of individual symptoms or overall health status in patients with HF, there is a need for research on the multi-trajectories of commonly co-occurring HF symptoms, which can provide a more comprehensive and detailed understanding of symptom experiences.
From a theoretical perspective, the theory of unpleasant symptoms (TOUS) has been developed to enhance understanding of relationships between multiple symptoms and influencing factors. The TOUS asserts that influencing factors (physiologic, psychological, and situational) are related to the symptom experiences of individuals.22 Several studies have found relationships of physiologic [e.g. age,23–25 the New York Heart Association (NYHA) functional class,23,24,26 and use of diuretics27], psychological (e.g. resilience28,29), and situational factors (e.g. social support30) to their symptom status. Moreover, non-cardiac comorbidities (e.g. anaemia, diabetes, and renal dysfunction) had a negative prognostic impact in patients with HF, especially in those who were hospitalized.31 Frailty, which overlaps with the underlying mechanisms of HF, increases the risks of functional decline and hospitalizations for HF patients.32 However, further exploration is needed to determine the impact of non-cardiac comorbidities and frailty status in symptom trajectories of patients with HF following discharge. The objectives of this study were to (1) identify the multi-trajectory groups based on common physical and psychological symptoms in patients with HF, (2) investigate the physiologic (i.e. age, NYHA class, non-cardiac comorbidities, use of diuretics, and physical frailty), psychological (i.e. resilience and psychologic frailty), and situational predictors (i.e. social support) of symptom multi-trajectory groups, and (3) determine the association between these groups and the risk for unplanned 30-day hospital readmission.
Methods
Study design and participants
This prospective observational study was conducted from September 2022 to September 2023, and the sample was conveniently recruited from the cardiology departments of four third-class hospitals in Tianjin, China. Participants were included if they (1) had HF as the primary admission diagnosis, (2) were aged 18 years or older, and (3) had the ability to read and express themselves in Mandarin without communication barriers. Participants with comorbidities that severely affected their clinical outcomes, such as cancer, and those planning to enter health care facilities for professional care after discharge were excluded. Additionally, individuals who had participated in any intervention programmes in the past three months were also excluded. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist (see Supplementary material online, Table S1).
The sample size was estimated based on a formula for logistic regression in observational studies: n = 100 + 10i (where i represents the number of independent variables).33 We assessed 13 variables in this study and accounted for a 20% sample loss rate, resulting in a minimum sample size of 288 at baseline. Participants who provided informed consent completed a baseline questionnaire regarding their information and symptom status within 24 h before discharge (T0) and provided their contact information for follow-up. Follow-up data on symptoms were collected via telephone at 2 weeks (T1) and 30 days (T2) after discharge. The duration of follow-up was established according to the Chinese guidelines for the diagnosis and treatments of heart failure 2018. A total of 301 participants completed measurements at T0 in our study.
Measurements
Symptoms
Symptom Status Questionnaire-Heart Failure (SSQ-HF)
The SSQ-HF was designed by Heo et al.34 in 2015. This questionnaire assesses the seven most common physical symptoms in patients with HF: daytime dyspnoea, dyspnoea when lying down, fatigue, chest pain, oedema, difficulty sleeping, and dizziness or loss of balance. It measures symptom presence, frequency, severity, and distress experienced by patients. Each symptom is scored on a scale of 0–12, with higher scores indicating more severe symptoms. The total score ranges from 0 to 84, and the Cronbach’s α was 0.800.34 The Cronbach’s α in the present study at baseline was 0.779.
The seven-item General Anxiety Disorder (GAD-7)
The GAD-7 was developed to identify potential patients with a generalized anxiety disorder and to assess the severity of symptoms of general anxiety.35 The GAD-7 consists of seven items, with a four-point Likert scale ranging from 0 (not at all) to 3 (nearly every day). The total score ranges from 0 to 21. According to the questionnaire manual, summated scores of 5–9 indicate mild GAD symptoms, 10–14 indicate moderate GAD symptoms, and 15 or higher indicate severe GAD symptoms. In the present study, Cronbach’s α at baseline was measured at 0.934.
The depressive symptom severity scale (PHQ-9)
The PHQ-9 is a nine-item self-report measure that is used to assess depression severity and criteria for a major depressive episode.36 Items assess for symptoms of depression and response anchors range temporally from 0 (not at all) to 3 (nearly every day). The total score ranged from 0 to 27 (scores of 5–9 are classified as mild depression; 10–14 as moderate depression; 15–19 as moderately severe depression; ≥20 as severe depression). The value of Cronbach’s α in the present study at baseline was 0.875.
Potential predictors
Non-cardiac comorbidities (i.e. anaemia, diabetes mellitus, old cerebral infarction, chronic obstructive pulmonary disease, liver dysfunction, and renal dysfunction) and the NYHA class of patients were extracted from the patient’s discharge records. Information on taking diuretics after discharge was obtained from the discharge prescription.
Physical and psychological frailty was measured by the Tilburg Frailty Indicator (TFI), an effective frailty measurement tool in patients with cardiovascular diseases.37 The TFI consists of 15 items covering physical, psychological, and social domains. It adopts the two-classification scoring method, and the scoring range is from 0 to 15 points. Higher scores indicate greater levels of frailty. The internal consistency reliability was good (Cronbach’s α = 0.71) in China.38 The Cronbach’s α in the present study at baseline was 0.917.
Resilience was assessed using the 10-Item Connor–Davidson Resilience Scale (CD-RISC-10).39 It is used widely to assess resilience and, specifically, the ability to cope with adversity. It consists of 10 items that cover aspects of self-perception, adaptability, flexibility, and goal orientation. Each item is scored on a five-point Likert scale. The total scores range from 0 to 40, with higher scores indicating greater resilience. The Cronbach’s α value of the CD-RISC-10 was 0.92 for the clinical samples.40 At baseline, the Cronbach’s α in the present study was 0.965.
Social support was tested by the Social Support Rating Scale (SSRS), developed by Shuiyuan Xiao in 1994.41 It consists of 10 items divided into three dimensions: subjective support, objective support, and utilization of social support. Each item is scored on a three-point or four-point Likert scale, with total scores ranging from 12 to 66. Higher scores indicate greater perceived social support. The Cronbach’s α value was 0.81 for the full scale.41 The value of Cronbach’s α in the present study at baseline was 0.813.
Performance
Unplanned 30-day hospital readmission was selected as performance in this study. It was defined as any non-programmed hospitalization for any cause within 30 days after the current discharge.
Ethical considerations
The study protocol was approved by the Research Ethics Committee of the University [project identification code: TMuhMEC (2022021)]. The study was conducted following the tenets of the Declaration of Helsinki.42 Oral and written informed consent was provided at the beginning of the study and signed by each participant.
Statistical analysis
Data analysis was conducted using IBM SPSS Statistics Version 24.0 and STATA 17.0. Descriptive statistics were used to summarize the general characteristics of the study population, including physiologic, psychological, and situational factors, and performance. Continuous variables were presented as means and standard deviations (SD), while categorical variables were presented as frequencies and percentages. Differences in baseline characteristics among trajectory groups were compared using analysis of variance for continuous variables and χ2 tests for categorical variables.
Group-based multi-trajectory modelling (GBMTM) was utilized to examine latent clusters of patients with HF with similar symptom trajectories. The ‘traj’ plugin in Stata 17.0 was used for data analysis.43 Individuals who died or experienced more than one loss to follow-up were excluded from this analysis. Firstly, we calculated the probability of all symptoms for individuals who completed three assessments (T0, T1, T2), and those symptoms with a 30-day mean incidence > 20% were selected for inclusion in GBMTM analysis.44 Then, days after discharge were used as a timescale for the trajectories. The optimal number of classes was identified by analysing one-class through five-class models, with several polynomial types (linear, quadratic, and cubic). Typically, the model fitting starts from the lower subgroup number and the higher order. The scores of all symptom status belonged to continuous corresponding variables, so a censored Normal Model was use. Model selection was based on model fit statistics and clinical relevance: (1) the model with the lowest Bayesian information criterion (BIC) value was preferred, (2) the proportion assigned to each trajectory group (based on the maximum posterior probability rule) was >5%, (3) the average posterior probability (AvePP) of group membership was at least 0.7, and (4) the final models had sufficient clinical relevance and interpretability.45 A sensitivity analysis was conducted to assess the influence of incomplete data on the trajectory analysis. Patients who completed the survey at least two times were included, and the missing data of symptoms were handled by Linear Interpolation.
Multiple logistic regressions were used to identify factors affecting the different trajectories over time. Binomial logistic regression analysis was used to estimate associations between symptoms multi-trajectories and risk of unplanned 30-day hospital readmission. The odds ratios (ORs) and 95% confidence intervals (CIs) were calculated to assess the strength and direction of the associations. Statistical significance was based on the two-tailed P-value of <0.05.
Results
Sample characteristics
The study included 301 hospitalized patients with HF, with a cumulative follow-up of 543 times. During the 30-day follow-up period, a total of 17 patients died, and 36 patients were lost to follow-up due to refusal or changing contact information, leaving a final cohort of 248 patients who completed the study (Figure 1).
Flow diagram for the data collection procedure.
The mean (SD) age of the 248 enrolled patients was 68.17 (12.713) years, 150 (60.5%) were women. Of these patients, 27.0% had a history of drinking and 33.9% had a history of smoking. The vast majority of patients were from urban areas and had partners. Compared to those included in the current analyses, excluded patients were more likely to be older, had higher NYHA classes, and had non-cardiac comorbidities (see Supplementary material online, Table S2).
Symptom occurrence rates
As shown in Table 1, the majority of patients reported experiencing dyspnoea during day time (74.19%), fatigue (85.89%), and depression (80.65%) before discharge. Only 14.92% of patients reported chest pain, while 10.08% reported dizziness. The average incidence rates of dyspnoea when lying down, chest pain, and dizziness during the 30 days after discharge were all below 20%.
Symptom occurrence rates of patients with heart failure during 30 days after discharge
| Symptom | Occurrence rate (%) | Average occurrence rate (%) | ||
|---|---|---|---|---|
| T0 | T1 | T2 | ||
| Fatigue or lack of energy | 85.89 | 89.10 | 89.92 | 88.31 |
| Depression | 80.65 | 73.79 | 67.34 | 73.92 |
| Dyspnoea during day time | 74.19 | 52.42 | 44.76 | 57.12 |
| Anxiety | 65.73 | 54.44 | 48.39 | 56.18 |
| Difficulty sleeping at night | 52.02 | 29.03 | 29.03 | 36.69 |
| Leg or ankle swelling | 34.27 | 17.34 | 14.52 | 22.04 |
| Dyspnoea when lying down | 28.23 | 8.06 | 9.68 | 15.32 |
| Chest pain | 14.92 | 4.44 | 6.85 | 8.74 |
| Dizziness or loss of balance | 10.08 | 6.45 | 7.26 | 7.93 |
| Symptom | Occurrence rate (%) | Average occurrence rate (%) | ||
|---|---|---|---|---|
| T0 | T1 | T2 | ||
| Fatigue or lack of energy | 85.89 | 89.10 | 89.92 | 88.31 |
| Depression | 80.65 | 73.79 | 67.34 | 73.92 |
| Dyspnoea during day time | 74.19 | 52.42 | 44.76 | 57.12 |
| Anxiety | 65.73 | 54.44 | 48.39 | 56.18 |
| Difficulty sleeping at night | 52.02 | 29.03 | 29.03 | 36.69 |
| Leg or ankle swelling | 34.27 | 17.34 | 14.52 | 22.04 |
| Dyspnoea when lying down | 28.23 | 8.06 | 9.68 | 15.32 |
| Chest pain | 14.92 | 4.44 | 6.85 | 8.74 |
| Dizziness or loss of balance | 10.08 | 6.45 | 7.26 | 7.93 |
T0 represented 24 h before discharge, T1 represented 2 weeks after discharge, and T2 represented 30 days after discharge.
Symptom occurrence rates of patients with heart failure during 30 days after discharge
| Symptom | Occurrence rate (%) | Average occurrence rate (%) | ||
|---|---|---|---|---|
| T0 | T1 | T2 | ||
| Fatigue or lack of energy | 85.89 | 89.10 | 89.92 | 88.31 |
| Depression | 80.65 | 73.79 | 67.34 | 73.92 |
| Dyspnoea during day time | 74.19 | 52.42 | 44.76 | 57.12 |
| Anxiety | 65.73 | 54.44 | 48.39 | 56.18 |
| Difficulty sleeping at night | 52.02 | 29.03 | 29.03 | 36.69 |
| Leg or ankle swelling | 34.27 | 17.34 | 14.52 | 22.04 |
| Dyspnoea when lying down | 28.23 | 8.06 | 9.68 | 15.32 |
| Chest pain | 14.92 | 4.44 | 6.85 | 8.74 |
| Dizziness or loss of balance | 10.08 | 6.45 | 7.26 | 7.93 |
| Symptom | Occurrence rate (%) | Average occurrence rate (%) | ||
|---|---|---|---|---|
| T0 | T1 | T2 | ||
| Fatigue or lack of energy | 85.89 | 89.10 | 89.92 | 88.31 |
| Depression | 80.65 | 73.79 | 67.34 | 73.92 |
| Dyspnoea during day time | 74.19 | 52.42 | 44.76 | 57.12 |
| Anxiety | 65.73 | 54.44 | 48.39 | 56.18 |
| Difficulty sleeping at night | 52.02 | 29.03 | 29.03 | 36.69 |
| Leg or ankle swelling | 34.27 | 17.34 | 14.52 | 22.04 |
| Dyspnoea when lying down | 28.23 | 8.06 | 9.68 | 15.32 |
| Chest pain | 14.92 | 4.44 | 6.85 | 8.74 |
| Dizziness or loss of balance | 10.08 | 6.45 | 7.26 | 7.93 |
T0 represented 24 h before discharge, T1 represented 2 weeks after discharge, and T2 represented 30 days after discharge.
Group-based multi-trajectory model of symptoms
Group-based multi-trajectory modelling showed that the model with three subgroups fit best. The model fit statistics were presented in Supplementary material online, Table S3. Despite the progressive reduction in BIC values as the number of model classes increased and all AvePP values surpassed 0.7, a category with a proportion of 5–6% emerged when the model class was four or five. To mitigate model overfitting and consider the clinical relevance of the model, a three-class model with lower order was identified as the optimal choice (Figure 2). After including individuals with missing follow-up data, we observed a similar pattern for the three trajectory groups (see Supplementary material online, Figure S1), with comparable model adequacy (see Supplementary material online, Table S4).
Symptoms multi-trajectory model of patients with HF. Solid lines represented the average estimated dyspnoea during day time, fatigue or lack of energy, leg or ankle swelling, difficulty sleeping at night, anxiety, and depression over time. Dashed lines represented the 95% confidence interval.
A total of 24.19% of patients belonged to the mild symptom status group (group 1). In this group, fatigue was the most prominent symptom, although it was relieved after discharge. Dyspnoea during day time had a rapid remission after discharge and persisted at a low level. Oedema and difficulty sleeping at night were not significant burdens, and there was no reported anxiety or depression. More than half (57.26%) of patients belonged to the moderate symptom status group (group 2). Fatigue did not show significant relief after discharge and remained at a moderately high level. There was a slight remission in dyspnoea. Other symptoms in group 2 had higher severity than those in group 1, although they followed a similar trend. Group 3 included 19.1% of patients with severe symptom status. Unlike the other two groups, patients in this group experienced worsening fatigue symptoms after discharge and remained at a high level. Additionally, these patients exhibited moderate to high levels of anxiety and depression, with higher psychological distress compared to the other two groups.
Information about the demographic, clinical, and symptoms-related factors of the three latent groups was presented in Supplementary material online, Table S5. Whereas gender, age, educational level, and economic status were largely similar among the three groups, group differences in regular exercise were found. The proportion of patients with regular exercise was higher in group 1 than in the other two groups. The proportion of patients with NYHA class Ⅱ was the largest in group 1, while the prevalence of NYHA class Ⅲ was the largest in groups 2 and 3.
Predictors of symptoms multi-trajectory groups
Multiple logistic regression was used to identify the predictors of trajectory class group (Figure 3, Supplementary material online, Table S6). Compared to the mild symptom status group, the moderate symptom status group had a higher physical frailty (OR = 1.476, 95% CI: 1.060–2.054), a higher psychological frailty (OR = 1.703, 95% CI: 1.229–2.361), and comorbid renal insufficiency (OR = 5.075, 95% CI: 1.182–21.786). The severe symptom status group and the mild symptom status group differed significantly at baseline in terms of physical frailty (OR = 2.206, 95% CI: 1.452–3.350), psychological frailty (OR = 1.755 95% CI: 1.068–2.885), resilience (OR = 0.925, 95% CI: 0.858–0.997), taking diuretics (OR = 4.342, 95% CI: 1.396–13.507), and comorbid renal insufficiency (OR = 7.539, 95% CI: 1.350–42.099).
Baseline predictors of multi-trajectory group membership by multiple logistic regression. Figure (A) was the result of the multiple logistic regression for group 2 when group 1 was used as the reference category, and figure (B) was the result of the multiple logistic regression for group 3 when group 1 was used as the reference category. NYHA class, New York Heart Association (NYHA) functional class; COPD, chronic obstructive pulmonary disease; OR, odds ratio; CI, confidence interval. Bold values are statistically significant values.
Association between multi-trajectory groups and performance
The unplanned 30-day hospital readmission rate for patients with HF was 11.69% in this study. Group 1 had the lowest rehospitalization rate of 1.67%, while groups 2 and 3 had higher rates of 13.38% and 19.57%, respectively. Compared with patients in the mild symptom status group, patients in the severe symptom status group reported a higher risk of unplanned 30-day hospital readmission after adjusting for all covariates (OR: 14.766, 95% CI: 1.534–142.163) (Table 2).
The risk of unplanned 30-day hospital readmission by symptom multi-trajectories
| Trajectories | Model 1a | Model 2b | ||
|---|---|---|---|---|
| P-value | OR (95% CI) | P-value | OR (95% CI) | |
| Group 1 | ref | ref | ||
| Group 2 | 0.033 | 9.114 (1.191, 69.720) | 0.050 | 8.087 (0.998, 65.525) |
| Group 3 | 0.012 | 14.750 (1.793, 121.361) | 0.020 | 14.766 (1.534, 142.163) |
| Trajectories | Model 1a | Model 2b | ||
|---|---|---|---|---|
| P-value | OR (95% CI) | P-value | OR (95% CI) | |
| Group 1 | ref | ref | ||
| Group 2 | 0.033 | 9.114 (1.191, 69.720) | 0.050 | 8.087 (0.998, 65.525) |
| Group 3 | 0.012 | 14.750 (1.793, 121.361) | 0.020 | 14.766 (1.534, 142.163) |
OR, odds ratio; CI, confidence interval.
aModel 1 was not adjusted by any covariates.
bModel 2 is adjusted by physical frailty, psychological frailty, renal insufficiency, resilience, and use of diuretics.
The risk of unplanned 30-day hospital readmission by symptom multi-trajectories
| Trajectories | Model 1a | Model 2b | ||
|---|---|---|---|---|
| P-value | OR (95% CI) | P-value | OR (95% CI) | |
| Group 1 | ref | ref | ||
| Group 2 | 0.033 | 9.114 (1.191, 69.720) | 0.050 | 8.087 (0.998, 65.525) |
| Group 3 | 0.012 | 14.750 (1.793, 121.361) | 0.020 | 14.766 (1.534, 142.163) |
| Trajectories | Model 1a | Model 2b | ||
|---|---|---|---|---|
| P-value | OR (95% CI) | P-value | OR (95% CI) | |
| Group 1 | ref | ref | ||
| Group 2 | 0.033 | 9.114 (1.191, 69.720) | 0.050 | 8.087 (0.998, 65.525) |
| Group 3 | 0.012 | 14.750 (1.793, 121.361) | 0.020 | 14.766 (1.534, 142.163) |
OR, odds ratio; CI, confidence interval.
aModel 1 was not adjusted by any covariates.
bModel 2 is adjusted by physical frailty, psychological frailty, renal insufficiency, resilience, and use of diuretics.
Discussion
To the best of our knowledge, this is the first study to investigate symptom multi-trajectories of patients with HF within 30 days after discharge. While there is a growing body of research on symptom trajectories in HF patients, most studies have focused on individual symptoms. In our study, we leveraged the advantages of GBMTM to examine the diverse developmental multi-trajectories of the six most common symptoms after discharge. We identified three distinct symptom multi-trajectories during the 30 days following discharge: the mild symptom status, the moderate symptom status, and the severe symptom status. The common risk factors for patients in the moderate symptom status group and severe symptom status group were comorbid renal dysfunction and worse physical and psychological frailty within 24 h before discharge. Continuing diuretics use after discharge and low resilience at baseline were only associated with the severe symptom status group, which was linked to a higher risk of unplanned 30-day hospital readmission.
The multi-trajectories estimated in this study provided a clear picture of different trends in symptom development within 30 days after discharge. A total of 24.19% of HF patients experienced mild symptom status, more than half experienced moderate symptom status, and <20% experienced severe symptom status. Overall, the symptom status of the three groups aligned with their respective classifications. However, each group exhibited distinct temporal patterns. Patients in the mild symptom status group demonstrated favourable disease control following discharge, characterized by swift resolution of daytime dyspnoea and gradual alleviation of fatigue. Patients in the moderate symptom status group suffered persistent moderate fatigue, which was slightly relieved over time, along with self-reported recovery of daytime dyspnoea. Patients in the severe symptom status group showed a worsening trend in fatigue. Notably, compared to the two other groups, they reported more pronounced sleep difficulty, although this phenomenon was reduced within two weeks post-discharge. Despite the inconsistent fatigue levels among the three groups, it was the most serious and challenging symptom of all physical symptoms. In our study, combined renal dysfunction increased the risk of moderate and severe symptom status multi-trajectories. Renal dysfunction is a well-established risk factor for poor clinical outcomes in patients with HF, as comorbid HF and renal dysfunction comprise a vicious cycle.46 In addition, we found that patients using diuretics according to the discharge prescription were more likely to experience severe symptom status among the three symptom multi-trajectories. Healthcare providers prescribed diuretics for HF patients who had severe symptoms. Use of diuretics may not always lead to no symptoms or mild symptoms if those patients had some other pathophysiological reasons for more severe symptoms.47 The coexistence of multiple physical symptoms, such as fatigue, sleep disorders, and oedema, has significantly promoted the occurrence and progression of depressive and anxiety symptoms.14,48 Patients in the severe symptom status group may require more intensive psychological support and management due to experiencing moderate to high levels of anxiety and depression within 30 days after discharge. In conclusion, early support and intervention for psychological and physical symptoms should be provided for patients in the moderate and severe symptom status groups to alleviate their symptom multi-trajectories after discharge.
The predictors of latent groups identified in this study summarized baseline (before discharge) characteristics that can differentiate the distinct developmental trends of commonly co-occurring symptoms and provide potential targets for personalized interventions. In this study, we found that both physical frailty and psychological frailty were risk factors for the development of moderate and severe symptom multi-trajectories in patients with HF. Frailty reduces the resistance of patients with HF to myocardial ischaemia and pressure, volume overload, and increases the risk of arrhythmias, causing decompensation and rapid functional deterioration.49 Frailty is a multidimensional concept, and the coexistence of physical and psychological frailty significantly increases the risks of functional decline and depression.50 Therefore, the combined assessment of physical and psychological frailty in clinical practice allows for more detailed risk stratification of patients with HF. Besides, frailty is potentially preventable, reversible, and slowable, offering hope that it can be targeted for the management of HF symptoms.32 In addition, this study observed an encouraging finding that patients with lower resilience prior to discharge were more likely to experience a trajectory of severe symptom multi-trajectory, which suggested that resilience is another important target for relieving symptom status in trajectories in the severe symptom status group. A growing body of research has identified resilience as a way to help people mitigate the impacts of disease and related symptom distress. Longcoy et al.51 found a relationship between resilience and symptom distress among patients with breast cancer. A recent systematic review and meta-analysis concluded that resilience was positively associated with self-care in people with chronic conditions, and resilience had the potential to buffer the adversities of daily self-care to maintain physical and emotional well-being.52 Patients with HF who experienced moderate or severe symptom multi-trajectories were more likely to be influenced by physiologic and psychological factors rather than social factors. Based on the findings of this study, it is recommended that healthcare professionals initiate multidimensional frailty and resilience assessments as soon as the patient’s condition stabilizes. Personalized comprehensive disease management plans should be initiated early during hospitalization while assisting patients in preparing for discharge to improve their symptom management capabilities and reduce the burden of symptoms after discharge.
Importantly, our study revealed that a longitudinal multi-trajectory pattern of severe symptom status predicted a higher risk of unplanned 30-day hospital readmission. This finding was consistent with a previous study, suggesting that more severe symptoms were associated with poorer clinical outcomes.53 A high 30-day hospital readmission rate not only indicates suboptimal patient outcomes but also reflects challenges in managing symptoms. It is important to acknowledge that patients with HF often struggle to manage their symptoms alone and may have difficulty coping with changes in symptoms.54 Therefore, optimizing support and interventions for these patients is crucial. We should focus on the characteristics of symptom multi-trajectories, develop personalized intervention plans for each muti-trajectory, and attach importance to the predictors, broaden multidisciplinary interventions that begin during hospitalization and bridge the transition to home.
Limitations
There were several limitations to consider in this study. Firstly, the presence of survival bias may have resulted in the exclusion of individuals with poorer health who had a higher likelihood of mortality during the follow-up period. However, focusing on understanding and addressing the symptom multi-trajectories and clinical outcomes in survivors is more beneficial from a policy perspective. Secondly, this study was conducted in hospitals in China, which may introduce certain biases in population characteristics. Although a multi-centre study design was employed to enhance the credibility and generalizability of our research findings, it is important to conduct similar studies in other countries to comprehend the pattern of co-occurring symptoms in HF patients after discharge. Thirdly, patients were categorized into three symptom multi-trajectory groups with varying proportions. However, there is limited evidence to suggest the extent to which the same trajectory groups occur in different samples. Additionally, this study was subject to inherent limitations related to residual confounding from unmeasured covariates, such as genetic background. While the TOUS was utilized to guide the measurement of potential factors, it is impossible to guarantee the absence of overlooked influencing factors.
Conclusion
Three distinct symptom multi-trajectories among patients with HF were identified. Patients in the severe symptom status group had higher risks of unplanned 30-day hospital readmission compared to those in the mild symptom status group. Combined renal dysfunction, physical frailty, and psychological frailty were significant predictors of the moderate and severe symptom status groups. The severe symptom status group was also characterized by lower resilience and the use of diuretics. Our study provides valuable insights into the symptom multi-trajectories and their association with unplanned 30-day hospital readmission of HF patients after discharge. These findings could be further utilized to develop personalized intervention plans and reduce the symptom burden of HF patients.
Supplementary material
Supplementary material is available at European Journal of Cardiovascular Nursing online.
Funding
This work was supported by Humanities and Social Science Fund of Ministry of Education of China (23YJAZH189); the National Natural Science Foundation of China (72304206; 71974142; 72274134).
Data availability
The datasets used and/or analysed during the current study are available from the corresponding authors on reasonable request.
References
Author notes
Q.L. and Xiaon.Z. made equal contributions to this manuscript.
Conflict of interest: none declared.
Comments