Pulmonary rehabilitation assessment in COPD based on the ICF brief core set: a latent profile analysis

Abstract

Background

Chronic obstructive pulmonary disease (COPD) is the most burdened chronic respiratory disease in the world, resulting in a reduced quality of life and limited physical activity for patients. Pulmonary rehabilitation (PR) is an effective therapy for COPD. Effective PR relies on an accurate pulmonary rehabilitation program. An adequate pre-rehabilitation assessment helps healthcare professionals to develop an accurate pulmonary rehabilitation program. However, pre-rehabilitation assessment strategies lack specific selection criteria and an assessment of the patient’s overall functioning.

Methods

This study explored the functional characteristics of COPD patients before pulmonary rehabilitation and collected COPD patients from October 2019 to March 2022. A cross-sectional survey of 237 patients was conducted using the ICF brief core set as the study tool. Latent profile analysis identified subgroups of patients with different rehabilitation needs based on body function and activity participation.

Results

Four subgroups of functional dysfunction were identified: 5.42%, 21.03%, 29.44%, and 34.11% in the high dysfunction group, the moderate dysfunction group, the lower-middle dysfunction but high mobility impairment group, and the low dysfunction group, respectively. Patients in the high dysfunction group were older, had a higher proportion of widowed spouses, and experienced more exacerbation. Most patients in the low-dysfunction group did not use inhaled medication and had a lower participation rate in oxygen therapy. Patients with a more severe disease classification and symptom burden mostly belonged to the high dysfunction group.

Conclusions

COPD patients require an adequate assessment before implementing a pulmonary rehabilitation program to determine their rehabilitation needs. The four subgroups were heterogeneous in terms of the degree of functional impairment in body function and activity participation. Patients in the high dysfunction group can improve basic cardiorespiratory fitness; patients in the moderate dysfunction group should focus on improving cardiorespiratory endurance and muscle fitness, patients in the lower-middle-dysfunction but high mobility impairment group should focus on improving mobility and patients in the low functional disability group should focus more on preventive measures. Healthcare providers can tailor rehabilitation programs to the functional impairments of patients with different characteristics.

Trial registration

This study has been registered in the Chinese Clinical Trials Registry (ChiCTR2000040723).

Keywords:

• Chronic obstructive pulmonary disease
• Pulmonary rehabilitation
• Precision rehabilitation
• Rehabilitation needs
• ICF
• Latent profile analysis

1. Introduction

Chronic obstructive pulmonary disease (COPD) is a common chronic respiratory disease characterized by persistent lung function decline and airflow limitation. The Global Initiative for Chronic Obstructive Lung Disease (GOLD) 2022 guidelines specify that patients with COPD account for the largest proportion of patients with chronic respiratory diseases worldwide [1]. Patients with COPD suffer from dyspnea, cough, expectoration, wheezing, and other symptoms. These conditions financially burden the patient, decrease their quality of life, limit physical activity, decondition muscles, and cause poor sleep quality, anxiety, and depression [2].

Previous research demonstrated that pulmonary rehabilitation (PR) is the hallmark of treatment in patients with COPD and can improve their quality of life and exercise endurance. It can also relieve the symptoms of dyspnea and increase lung function [3]. COPD, a highly prevalent chronic airway disease with substantial personal and social impact, has clinical and biological complexity [4], and corresponding research on COPD moves towards precision medicine [5]. Individualized PR programs are widely used in the rehabilitation programs of COPD patients and reduce the disease burden for individual patients [6]. However, we found that the research on PR focused more on the intervention effect and ignored the pre-rehabilitation assessment [7].

Precision rehabilitation is a “comprehensive intervention based on a thorough patient assessment followed by patient-tailored therapies” [8]. A thorough assessment is a crucial component of a precision PR program. However, the content of pre-rehabilitation assessment in current research is different, and there is no specific standard for selecting assessment content. Therefore, the rehabilitation needs of different types of COPD cannot be obtained. For example, studies have shown that different clusters of patients with COPD can be derived with different characteristics based on their intra-pulmonary and extra-pulmonary symptoms, all of whom respond differently to pulmonary rehabilitation [9]. A review suggested that stratified medicine is a pathway to personalized medicine in COPD⁶. However, some existing approach was focused on assessing different traits, whereas integrated and holistic approaches are needed to tackle the complexity of every unique patient [10,11]. Furthermore, addressing a treatable trait requires a profound understanding of underlying and individual factors. For example, various social and personal barriers and enablers can limit or encourage participation in physical activity in COPD patients [6]. Therefore, applying a clustering approach to graded rehabilitation assessment is very important to understand the participation of patients in the PR program and facilitate personalized treatment planning. Some European researchers have applied the clustering method to the PR process, which includes the multi-dimensional assessment of the patient’s PR response and the exploration of clinical phenotypes. Spruit et al. [12] applied a clustering approach to analyze the heterogeneity of COPD patients’ responses to PR, explore the characteristics of patients with different response types to identify poor responders and population characteristics and provide individualized care interventions to those who do not respond well. In addition, some researchers have clustered patients based on their baseline characteristics to compare differences in PR response across different patients and provide information on the complexity of PR patient traits [9,13]. However, such studies emphasized the non-responder identification to tailor rehabilitation techniques to the patient’s specific needs rather than pre-PR assessment.

Regarding pre-rehabilitation assessment, heterogeneity in patients’ structural and functional relationships must be assessed to implement an individualized PR. Healthcare providers must identify the clinical complexities of COPD through a sufficient assessment so that more accurate and effective therapies can be offered to patients [14]. Some researchers have identified subgroups of COPD with different health-related exercise patterns by assessing the physical activity and sedentary behavior patterns of COPD patients to design individualized PR plans. Other studies have analyzed the differences between patients’ respiratory physiology characteristics before and after PR [14]. Although the method mentioned above is comprehensive, it involves several medical devices and tools, is complicated, and is challenging to popularize. There is a need for a more concise and convenient assessment to explore the PR needs of COPD patients.

The International Classification of Functioning, Disability, and Health (ICF) set represents a new methodological standard for the functioning (organs and diagnosis involved), disability (symptoms and signs), social participation restriction, and environment interaction assessment [15,16]. Within the ICF, healthcare practitioners can rely for the first time on a worldwide accepted model that provides a universal language for describing and classifying functioning. Researchers have developed appropriate ICF tools in different steps of the rehabilitation cycle for different diseases, such as pre-rehabilitation assessment and evaluation of rehabilitation outcomes [17]. The ICF was applied in the assessment and documentation of rehabilitation for cerebral palsy and non-traumatic acquired brain injury, stroke, chronic low back pain, rheumatoid arthritis, and knee osteoarthritis [18,19]. Similarly, studies have demonstrated that the ICF score can describe the function of COPD patients to evaluate rehabilitation outcomes [15,20]. The ICF core set for COPD could evaluate rehabilitation effects, be integrated into the Rehab-CYCLE, and provide a valuable tool for rehabilitation management [21,22].

Researchers investigated the severity of physical function, the need for hospitalization, and COPD patients’ recovery from exacerbation after PR¹⁵. Other studies have found multiple ICF components associated with the participation and effectiveness of PR [23]. In addition, several researchers have developed a brief ICF core set for COPD that can accurately and appropriately measure daily activity in COPD [24]. In our group’s previous study, the brief ICF core set showed good reliability and validity for assessing comprehensive dysfunction in COPD patients, the results of which can be used to formulate rehabilitation strategies [25].

The ICF score has been used to assess PR in COPD patients. However, current studies are limited to patients’ suitability for PR and rehabilitation outcomes. In addition, there is a lack of adequate pre-rehabilitation evaluation to determine a specific and feasible PR program for patients with different levels of function. Therefore, this study planned to extract the characteristics of COPD patients in terms of physical function, activity, and participation based on the brief ICF core set, inferring rehabilitation needs according to the patients’ levels of functional disability. It was hypothesized that queues would be generated with different levels of disability functions, and patients with higher disease severity and more severe dyspnea would show higher functional impairment. It allowed the authors to identify the PR needs of patients and achieve the precise prevention and control goals of personalized PR programs.

2. Methodology

2.1. Study design and setting

This cross-sectional study recruited patients with COPD from October 2019 to March 2022 at five tertiary hospitals in Tianjin, China. This study used STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) checklist for observational research reporting. Ethical approval was obtained before the study’s commencement. All participants were informed of the entire contents of the study, and they submitted signed consent forms.

2.2. Participants

This study selected participants based on the following inclusion criteria. Participants should

1. meet the diagnostic criteria of the GOLD 2023 and the global strategy for the diagnosis, treatment, and prevention of COPD (updated);

2. show a post-bronchodilator FEV/FVC <0.70 in the clinical context that confirmed the presence of persistent airflow limitation and, thus, of COPD in patients with appropriate symptoms and significant exposures to noxious stimuli.

It should be noted that this study defined COPD as a heterogeneous lung condition characterized by chronic respiratory symptoms (dyspnea, cough, and sputum production). COPD occurs due to airway abnormalities (bronchitis, bronchiolitis) or alveoli (emphysema) that cause persistent, often progressive, airflow obstruction.

Exclusion criteria included participants (i) having complications with severe primary diseases leading to functional impairment, such as cardiovascular, liver, kidney, and brain diseases; (ii) being unable to respond to the research instructions.

2.3. Data collection

In the early stages, all investigators underwent uniform training on survey tools to decrease the inter-investigator bias of the questionnaire. The method of implementation is as follows: Based on the Chinese version of Assessment Standard for the lnternational Classification of Functioning, Disability and Health Rehabilitation Set [26,27], we explained each ICF question in detail to each survey participant, to ensure that they had a consistent understanding of each ICF question. Prior to the start of the study, a number of patients were selected for pretesting to collect possible patient-expressed outcomes, and these were quantified consistently according to the guidelines for the use of the ICF, with the aim of providing a consistent measure of outcomes for each investigator. We asked investigators to use uniform questions in all questioning sessions. For outcomes that could not be measured by the investigator, we asked the investigator to record them in writing and to discuss the ratings with the whole group of investigators.

Patients were briefed on the specifics of this study and interviewed face-to-face after being screened as eligible participants. This study gathered the patients’ sociodemographic and other clinical data. A portable spirometer (Micro Loop, CareFusion, Kent, UK) assessed lung function through spirometry according to the guidelines of the European Respiratory Society (ERS) [28] for further classification of the severity of the airflow limitation (GOLD 1–4).

2.3.1. Sociodemographic data

Sociodemographic data included gender, age, body mass index (BMI), marital status, medical insurance type, and smoking status (smoking habit and intention to quit). It also included the previous year’s number of exacerbations, number of comorbidities, inhalation medication use, and home oxygen therapy.

2.3.2. The brief ICF-COPD core set

The brief ICF-COPD core set assessed functioning and disability, following the ICF categories [25]. These categories were designated by specific letters: b, s, d, and e represented body functions, body structures, activities and participation, and environmental factors, respectively. ICF qualifiers were applied to rate the degree of problems in each category of the body function, structures component, and the activity participation component with a generic 5-point scale: 0, 1, 2, 3, and 4, indicating no problem, mild problem, moderate problem, severe problem, and complete problem, respectively. In addition, scores 8 and 9 indicated “unspecified” and “not applicable,” respectively, in all categories. The core set showed good reliability and validity, with a high internal consistency of 0.873 for the total scale and values of 0.750, 0.640, and 0.843 for body functions, structures, activity, and participation, respectively. The content validity index (CVI), scale-level CVI/universal agreement, and S-CVI/Ave had values of 0.80–1, 0.929, and 0.986, respectively. Among the brief ICF-COPD core set, part ‘b’ emphasized the impairment of the patient’s pulmonary and extrapulmonary functions, leading to symptoms of dyspnea and exertion, which are significant problems in pulmonary rehabilitation [14]. Part ‘d’ assessed patients’ experiences in different life domains and reflected a main goal of rehabilitation: overcoming patient participation limitations [14]. Therefore, this study selected two indicators, ‘b’ and ‘d,’ to evaluate patients’ rehabilitation needs (Table 1).

Table 1. Contents of the components b and d of the brief ICF-COPD core set.

Components	Code	Categories
body function	b440	Respiration functions.
	b450	Additional respiratory functions.
	b455	Exercise tolerance functions.
	b460	Sensations associated with cardiovascular and respiratory functions.
Activity and participation	d230	Carrying out daily routine.
	d450	Walking.
	d455	Moving around.
	d640	Doing housework.

2.3.3. COPD clinical severity grade

The severity of airflow limitation in COPD was classified based on the post-bronchodilator FEV1 value into four groups (GOLD1, GOLD2, GOLD3, and GOLD4). Mild patients (GOLD1), moderate patients (GOLD2), severe patients (GOLD3), and very severe patients (GOLD4) had predicted FEV1 ≥ 80%, 50% ≤ FEV1 < 80%, 30% ≤ FEV1 < 50%, and were FEV1 < 30%, respectively (GOLD 2022).

2.3.4. Modified british medical research council (mMRC)

The mMRC measured the severity of dyspnoea [29]. The mMRC was a 5-point scale that inquired patients to indicate how dyspnoea limited their daily activities on a 5-point scale where “0” represented no limitation, and “4” represented a very severe limitation.

2.3.5. COPD assessment test (CAT^TM)

The CAT^TM evaluated the health status of COPD patients [30]. It included eight items that assessed the levels of cough, phlegm, climbing difficulties, housework activity performance, confidence in going outside, sleep disorders, and energy loss. The evaluation was on a 6-point scale, from “0” (no effect) to “6” (maximum effects). The score ranged from 0 to 40, with “0–10,” “11–20,” “21–30,” and “31–40” representing the slight impact, medium impact, severe impact, and very severe impact of the disease on the health status, respectively.

2.4. Data analysis

Latent profile analysis (LPA) distinguished potential heterogeneity of physical function and activity participation in COPD patients. Models ranged from 1 to 5 subtypes, and specified variable variances and covariances arguments were estimated to identify the optimal number of subtypes and parameter combinations [31]. The following model fit indices determined the optimal number of profiles: likelihood-ratio (LL), Akaike’s information criteria (AIC), Bayesian information criteria (BIC), and sample-size adjusted BIC (ssaBIC). The lower values were indicative of a better fit. Bootstrap-adjusted likelihood ratio tests (BLRT) compared the K and K-1 profile models. P-values >0.05 indicated that the K-1 model was preferred. Entropy was computed to determine the accuracy of profile classification, with higher values indicative of better separation between profiles. Interpretability and parsimony were also considered in optimal model selection [32].

The mean and standard deviation described quantitative data, and the frequency percentage described qualitative data. ANOVA and Kruskal–Wallis test evaluated the differences between groups in quantitative data, and Pearson chi-square (χ²), Fisher’s exact test, or Jonckheere Terpstra nonparametric tests evaluated the qualitative data. Post-hoc Bonferroni-adjusted z-tests verified the pairwise difference between groups (0.05/6).

The IBM Statistical Package for Social Sciences (SPSS) software version 22 (IBM, Armonk, New York, USA) and R software version 4.0.2 (The R Foundation, Vienna, Austria) conducted statistical analyses with the significance level set at 0.05. Latent profile models were fit in the tidyLPA (version 1.0.8) and Mclust (version 5.4.6) R packages. IBM SPSS performed all other statistical analyses.

3. Results

3.1. Participants

In this study, 237 patients matched the inclusion criteria, of which five dropped out for personal reasons, two were unable to participate in the evaluation of lung function due to deterioration of their condition, and 16 were excluded due to a significant amount of missing data. Finally, 214 participants’ data were examined.

Patients included 156 (72.9%) men and 58 (27.1%) women, with a mean age of 68.81 ± 9.08. There were 25 (11.7%) patients with very severe airflow limitation (GOLD IV), 67 (31.3%) had severe airflow limitation (GOLD III), 98 (45.8%) had moderate airflow limitation (GOLD II), and 24 (11.2%) had mild airflow limitation (GOPD I). Next, 25 (11.7%) participants had a dyspnea mMRC grade of 0, 59 (27.6%) had a grade of 1, 42 (19.6%) had a grade of 2, 51 (23.8%) had a grade of 3, and 37 (17.3%) had a grade of 4. Regarding the physical function and activity participation module scores of the ICF brief core set identified by LPA, the best fit LPA model delineated four groups. The mean membership probability of individuals in each group ranged from 0.91 to 0.95, indicating good model adequacy (Table 2).

Table 2. Fit indices for latent profile analysis for discovery (n = 214).

Classes	LogLik	AIC	BIC	ssaBIC	Entropy	Probability_min	Probability_max	BLRT	BLRT_P	Class Probability
1	−2743.658	5519.315	5573.171	5522.471	1.000	1.00	1.00	/	/	1.00
2	−2393.439	4836.879	4921.028	4841.809	0.936	0.98	0.98	700.436	<0.010	0.34/0.66
3	−2305.604	4679.207	4793.651	4685.913	0.879	0.91	0.99	175.671	<0.010	0.36/0.22/0.42
4	−2267.065	4620.130	4764.867	4628.611	0.872	0.91	0.95	77.077	<0.010	0.21/0.15/0.29/0.34
5	−2266.876	4637.753	4812.783	4648.008	0.788	0.78	0.94	0.377	0.890	0.21/0.15/0.34/0.29/0.00

LogLik: likelihood-ratio; AIC: Akaike’s information criteria; BIC: Bayesian information criteria; ssaBIC: sample-size adjusted BIC; BLRT: Bootstrap adjusted likelihood ratio tests.

The bolded values represent the optimal value of the fitted metric for each column, which is used as a reference for selecting the number of categories for the Latent Profile Analysis.

3.2. Latent profile analysis

The likelihood ratio (LL) took the minimum value in the five-category model. AIC, BIC, and ssaBIC took the minimum value in the four-category model, and entropy took the maximum value in the two-category model. In the five-category model, the BLRT test p-value was 0.89, indicating that the four-category model was better than the five-category model. The three-category model showed a BLRT test p-value of less than 0.01, indicating that it was preferred over the two-category model. After considering the model-fitting index and clinical significance, the four-category model was chosen as the best-fitting model (Figure 1, Table 3).

Figure 1. Latent profile plot of the brief ICF-COPD core set for body function and activity participation (n = 214).

Table 3. The estimated value of each indicator variable across body function and activity participation of the brief ICF-COPD core set.

Parameter	High dysfunction group	Moderate dysfunction group	Lower- middle- dysfunction but high mobility impairment group	Low dysfunction group
b440	2.701	1.878	1.668	0.950
b450	2.345	1.266	0.850	0.228
b455	3.086	2.287	1.617	0.854
b460	2.692	1.658	1.596	0.955
d230	3.199	1.817	1.256	0.472
d450	3.415	1.989	1.118	0.400
d455	3.943	2.807	2.834	0.629
d640	3.652	2.716	0.767	0.296

Patients in Class 1 displayed poor body function and activity participation abilities. Therefore, this study designated them to “the high dysfunction group.” The patients in the Class 2 group had moderate body function and activity participation levels, so they were assigned to “the moderate dysfunction group.” Class 3 patients had upper-middle-level body function and activity participation ability, but the scores in d455 (moving around in different places) were higher than other indicators. It indicated that patients had low exercise ability except in walking. Therefore, this group was named “the lower-middle-dysfunction but high mobility impairment group.” Patients in Class 4 who scored lower on body function and activity participation tests were designated to “the low functional disability group.”

3.3. Group differences

Most patients in the high dysfunction group were elderly, with an average age of 74.43 ± 9.20. There were no statistically significant differences among the four groups of patients in terms of gender, body mass index (BMI), education level, and smoking status. Compared to the other groups, the high dysfunction group had a significantly higher percentage of widowed patients (24.2%). Patients in the low dysfunction group experienced fewer exacerbations in the preceding year. In contrast, those in the high-dysfunction group experienced more exacerbations. Patients (19.2%) who did not use inhaled medication were more in the low dysfunction group than in the other groups, with a statistically significant difference. Patients’ frequency in the low dysfunction group participated in oxygen therapy less than patients in the other groups, whereas patients in the high dysfunction group participated in oxygen therapy more. As predicted, patients in the high dysfunction group showed more severe disease grade, degree of dyspnea, and symptom burden. Patients in the high dysfunction group had more dyspnea and were more severely afflicted by the illness, with more than half (54.5%) having a Grade 4 mMRC and a CAT score of 27.18 ± 4.49 (Table 4).

Table 4. Sociodemographic data and respiratory-related indicators of the four functional disability classifications.

Variables	Entire sample n = 214	High dysfunction group (n = 33, 15.42%)	Moderate dysfunction group (n = 45, 21.03%)	Lower- middle- dysfunction but high mobility impairment group (n = 63, 29.44%)	Low dysfunction group (n = 73, 34.11%)	Chi-square or ANOVA
Age	68.81 (9.08)	74.43 (9.20)^a	70.49 (6.66)	66.11 (9.27)^b	67.58 (9.08)^b	7.676***
Sex						5.302
Male	156 (72.9%)	19 (57.6%)	33 (73.3%)	50 (79.4%)	54 (74.0%)
Female	58 (27.1%)	14 (42.4%)	12 (26.7%)	13 (20.6%)	19 (26.0%)
BMI	23.41 (4.44)	23.19 (5.07)	23.21 (4.40)	24.03 (4.82)	23.09 (3.83)	0.587
Marital status						14.903***
Married	200 (93.5%)	25 (75.8%)	44 (97.8%)	62 (98.4%)	69 (94.5%)
Widowed	14 (6.5%)	8 (24.2%)^b	1 (2.2%)^a	1 (1.6%)^a	4 (5.5%)^a
Education						8.554
Illiteracy	8 (3.7%)	2 (6.1%)	1 (2.2%)	2 (3.2%)	3 (4.1%)
Primary school	50 (23.4%)	8 (24.2%)	11 (24.4%)	12 (19.0%)	19 (26.0%)
Junior high school	84 (39.3%)	12 (36.4%)	17 (37.8%)	31 (49.2%)	24 (32.9%)
Senior middle school	59 (27.6%)	11 (39.3%)	13 (28.9%)	13 (20.6%)	22 (30.1%)
Bachelor’s degree and above	13 (6.1%)	0 (0.0%)	3 (6.7%)	5 (7.9%%)	5 (6.8%)
Smoking status						2.251
Not smoked	54 (25.2%)	11 (33.3%)	10 (22.2%)	13 (20.6%)	20 (27.4%)
Smoked	160 (74.8%)	22 (66.7%)	35 (77.8%)	50 (79.4%)	53 (72.6%)
Have quit smoking	59/160 (36.9%)	3/22 (13.6%)	11/35 (31.4%)	22/50 (44.0%)	23/53 (43.4%)
Number of exacerbations in the previous year		1.21 (0.82)^b	1.07 (0.96)^b,c	0.70 (0.89)^a,c	0.48 (0.77)^a	24.430*
No exacerbations	103 (48.1%)	7 (21.2%)	16 (35.6%)	33 (52.4%)	47 (64.4%)
One time	66 (30.8%)	13 (39.4%)	13 (28.9%)	20 (31.7%)	20 (27.4%)
Two times	34 (15.9%)	12 (36.4%)	13 (28.9%)	6 (9.5%)	3 (4.1%)
More than two times	11 (5.1%)	1 (3.0%)	3 (6.7%)	4 (6.3%)	3 (4.1%)
Number of comorbidities	2.99 (3.40)	6.12 (4.34)^b	3.27 (3.00)^a	2.52 (3.48)^a	1.81 (2.03)^a	32.565***
Inhalation medication use						7.518*
Yes	190 (88.8%)	31 (93.9%)	40 (88.9%)	60 (95.2%)	59 (80.8%)^a
No	24 (11.2%)	2 (6.1%)^a	5 (11.1%)^a	3 (4.8%)^a	14 (19.2%)^b
Home oxygen therapy						13.774***
Yes	80 (37.4%)	20 (60.6%)^a	21 (46.7%)	19 (30.2%)	20 (27.2%)^a
No	134 (62.6%)	13 (39.4%)^b	24 (53.3%)^a	44 (69.8%)^a	53 (72.6%)^b
GOLD		2.88 (0.86%)^a	2.49 (0.84%)	2.40 (0.75%)^b	2.23 (0.84%)^b	10.550*
I	24 (11.2%)	0 (0.0%)^a	3 (6.7%)	5 (7.9%)	16 (21.9%)^a
II	98 (45.8%)	14 (42.4%)^a	24 (53.3%)	33 (52.4%)	27 (37.0%)^a
III	67 (31.3%)	9 (27.3%)^a	11 (24.4%)	20 (31.7%)	27 (37.0%)
IV	25 (11.7%)	10 (30.3%)^b	7 (15.6%)	5 (7.9%)	3 (4.1%)^a
mMRC		4.24 (1.00)^a	3.58 (1.16)^a	2.73 (1.14%)^b	2.53 (1.18%)^b	50.400***
Grade 0	25 (11.7%)	0 (0.0%)	2 (4.4%)	8 (12.7%)	15 (20.5%)
Grade 1	59 (27.6%)	3 (9.1%)	8 (17.8%)	23 (36.5%)	25 (11.7%)
Grade 2	42 (19.6%)	4 (12.1%)	7 (15.6%)	14 (22.2%)	17 (23.3%)
Grade 3	51 (23.8%)	8 (24.2%)	18 (40.0%)	14 (22.2%)	11 (15.1%)
Grade 4	37 (17.3%)	18 (54.5%)	10 (22.2%)	4 (6.3%)	5 (6.8%)
CAT	18.05 (7.90)	27.18 (4.49)^a	20.95 (7.09)^a	18.33 (7.05)^b	15.03 (7.72)^b	10.462***

a, b, c: Means in a column identified with up script letter a, b, and c are significantly different among different letters; *: p<0.05; **: p<0.01,***: p<.001.

CAT: COPD Assessment Test.; mMRC: modified Medical Research Council dyspnoea questionnaire.

4. Discussion

This study aimed to develop theoretical advice for creating precise rehabilitation interventions for COPD patients by measuring the functioning and health of patients using the ICF brief core set and analyzing the underlying characteristics of patients. The latent profile analysis identified four characteristics of disability in COPD patients: the high dysfunction group (15.42%), the moderate dysfunction group (21.03%), the lower-middle dysfunction but high mobility impairment group (29.44%), and the low dysfunction group (34.11%).

The high and moderate dysfunction groups shared the same traits, including lower activity and participation compared to physical function and lower exercise tolerance. The lower-middle-dysfunction but high mobility impairment group had better daily living skills and respiratory function, but their mobility was significantly impaired. Patients in the low-dysfunction group had better physical functioning and activity and participation levels but worse body functioning than other variables. This study found that functioning and health in all groups were related to age, marital status, number of exacerbations in the previous year, comorbidities, inhaled medication use, oxygen therapy, GOLD, and mMRC score.

The review has shown that a precisely prescribed PR program requires a certain level of cardio-respiratory endurance and good physiological adaptations in patients [33]. Patients in the high-dysfunction group showed lower activity and participation than the body function. However, poor respiratory function and exercise tolerance were observed in high-dysfunction group, challenging patients in basic daily activities. In addition, the patients’ average age in this group was 74.43 ± 9.20, thus, they presented poorer physical function and health status. Ageing leads to a decline in lung function and structural dysfunction, reducing patients’ respiratory function and physical activity [34]. Therefore, most patients in this group use home oxygen therapy (60.6%), indicating that they might suffer from hypoxia during daily activities. They also had reduced mobility due to the anaerobic metabolism in the muscles developed by gas exchange restrictions [35]. The high-dysfunction group was more likely to have a greater number of comorbidities than other groups. Multimorbidity is a major risk factor for basic daily living disabilities [36]. Pain from multiple conditions may limit physical activity and motor function [37]. In addition, comorbidities lead to more adverse signs and symptoms in patients, exacerbating dyspnoea and reducing activity tolerance in COPD patients [38], contributing to impairment in activities of daily living and motor function [39]. These factors trigger poor physical function in such patients, making it challenging for them to meet the basic needs of daily living activities and improve their basic physical function in PR, which is consistent with the findings of Raskin et al. [39] PR programs for such patients should aim to restore the patient’s respiratory function, cardiopulmonary endurance, respiratory muscle strength, and oxygen-carrying capacity [40].

Additionally, the high-dysfunction group had a significantly higher percentage of widowed patients than the other groups, which might be associated with the role played by partners in COPD development. Evidence suggests that partners play a significant role as unofficial caregivers for COPD patients [41]. In a cross-sectional survey, three-quarters of all the informal caregivers were partners [42]. The patient’s spouse assisted them in activities of daily living, provided medical support, and offered emotional support and comfort [43,44]. Meanwhile, the interdependence and partner effect between COPD patients and caregivers have also been established. When the spouse is the primary caregiver, the patient has a higher capacity for self-care, and a stronger bond benefits the patient’s prognosis and positive outcomes [45]. The spouses’ assistance and company result in better functional levels and health status for the patients, whereas widowed patients are more likely to show higher functional impairment.

The moderately-dysfunction group’s trajectory characteristics were consistent with the high-dysfunction group’s. However, there was lesser impairment in physical function and activity participation, and cardiopulmonary and muscle group dysfunctions existed. Patients in this group had better respiratory function and exercise tolerance and could tolerate some levels of exercise training. Regarding clinical characteristics, acute exacerbations occurred more frequently in the moderate-dysfunction group than in the low-dysfunction group. It might be related to the respiratory muscle dysfunction accompanying COPD exacerbation and cardiac dysfunction. Studies showed that inspiratory and expiratory muscle strength forces were reduced when patients were hospitalized with COPD exacerbation. In addition, the prevalence of diaphragmatic dysfunction increased [46,47]. These changes result in respiratory and respiratory assistance dysfunction. Such patients have cardiopulmonary and muscle group dysfunctions but have some capacity for exercise training. PR programs can aim to improve their cardiopulmonary and physical activity. According to the American Association of Cardiovascular and Pulmonary Rehabilitation (AACVPR) guidelines, aerobic, strength, balance, and flexibility training addresses the patient’s cardiopulmonary and muscle group dysfunctions [48]. The aerobic training includes respiratory muscle and upper and lower limb endurance training for the patient. In addition, research has shown that resistance training in pulmonary rehabilitation can lead to greater muscular endurance and improved peripheral muscle strength [49]. The improvement in peripheral muscle strength is associated with improved walking ability, which supports resistance training to improve functional outcomes of patient activity and participation [50].

The lower-middle-dysfunction but high mobility impairment group exhibited lower grades of functional impairment in body function and participation in activities but prominent impairment in mobility. It may be related to patients’ peripheral skeletal muscle dysfunction; skeletal muscle weakness and fat-free mass deficit are common manifestations in COPD patients, resulting in reduced functionality and motor abilities [51]. Greater muscle strength might enhance the ability to perform general motor skills such as jumping, sprinting, and changing direction [52]. Therefore, increasing muscular strength and endurance should be the main rehabilitation goals for such patients. The American Thoracic Society/European Respiratory Society indicated that resistance (or strength) training has additional benefits in terms of muscle strength, in which local muscle groups are trained by repetitively lifting relatively heavy loads. In patients with chronic respiratory disease, combining constant load/interval training and strength training improves outcomes (exercise capacity and muscle strength) significantly than individual strategy without unduly increasing training time ⁸ and is appropriate as a major component of PR programs. In addition, patients’ average age in this group was 66.11 ± 9.27, and there was a smaller likelihood of skeletal muscle dysfunction due to aging. There was a prominent mobility impairment due to the patients’ lack of willingness to move. The patients’ lack of awareness of the benefits of exercise rehabilitation and inadequate knowledge of their muscle strength affects whole-body functional performance [53]. Healthcare professionals should identify the underlying causes of the patient’s mobility impairment and intervene when developing rehabilitation programs for such patients.

The low-dysfunction group of patients had the mildest grade of functional impairment, with a better ability to perform daily activities and exercise function. Patients in this group showed poorer respiratory function and cardio-respiratory endurance compared to their activity participation ability, suggesting that patients’ body dysfunction had less impact on activity and participation. In terms of clinical characteristics, this group had a younger mean age (67.58 ± 9.08) and less severe disease. More patients in the low dysfunction group did not use inhalation medication compared to those in the other groups (19.2%), and most patients did not participate in oxygen therapy (72.6%), which might be related to the lower symptom burden in this group. Previous studies have noted that PR programs improved exercise tolerance and reduced exacerbations in patients with mild COPD, including warm-up training, endurance training, psychosocial support, and education [54]. However, high dropout rates are a common challenge in conducting PR programs for patients with mild COPD, and patient compliance with PR programs is poor [45]. It might be associated with patients’ less seriousness towards their disease symptoms and management. Studies showed that low adherence in patients to PR programs was associated with current smoking. Most patients with mild COPD have problems with quitting smoking, affecting patient adherence to PR programs and reducing PR efficacy [55]. To address these issues, the PR program for such patients needs to pay particular attention to adherence behaviours based on maintaining cardiopulmonary endurance.

Patients with more severe COPD and dyspnea had higher levels of dysfunction, which was in line with the prior hypothesis. Patients with low lung function had worsened dyspnea symptoms, reduced exercise tolerance, and impaired respiratory function [56]. Regarding physical activity, lung function impacts physical activity because exercise limitation results from respiratory conditions [57]. According to this study, the severity of dyspnea was correlated with functional impairment, indicating that dyspnea became more severe with increased impaired physical function and activity participation. Interestingly, some patients with better pulmonary function, like those in GOLD II, might still have moderate and high levels of dysfunction. It might be related to the course of COPD being variable, with some patients having a high degree of obstruction and minimal symptoms. In contrast, others with better lung function have a greater symptom burden [58]. Evidence suggested that symptoms should be assessed at each visit because FEV1% does not strongly correlate with symptoms at the individual level [59]. Similarly, before the patient participates in the PR program, FEV1% should not be the leading indicator for medical staff to judge the rehabilitation prescription. Lung function has not been found to infer PR responsiveness or change with this intervention.

This study has some limitations. First, the study’s sample size was insufficient, and the number of patients in each group after classification was small. The findings need to be validated with a large sample size. Second, this study was conducted in five hospitals in one Chinese city, and the findings do not represent all COPD patients. Thirdly, the sample population we included had more male than female patients and there may be some potential bias in the selection of the sample. It makes them potentially unrepresentative of the COPD population as a whole. Fourthly, psychological variables such as anxiety and depression were not included in the ICF core set and thus we did not include psychological factors in COPD patients in our analysis. It may be important to include in future stratified assessments of pulmonary rehabilitation in COPD patients. In addition, the evaluation results might deviate from the actual situation since there is no uniform quantitative standard for each item in the ICF-COPD core set. Therefore, they might be inter-investigator bias in the results obtained by different investigators, which requires further review.

5. Conclusion

This study assessed patients’ PR needs by analyzing the level of functional impairment in COPD patients based on the ICF brief core set. A potential profile analysis classified COPD patients’ functional impairment into four clusters: the high dysfunction group, the medium dysfunction group, the lower-middle- dysfunction but high mobility impairment group, and the low dysfunction group. There was heterogeneity in the functional impairment characteristics of patients in different clusters, which caused the patients’ needs for PR to differ. PR programs for patients in the high-dysfunction group should aim to improve basic cardiopulmonary function. Optimizing cardiopulmonary endurance and muscular function in the medium dysfunction group is recommended. In the lower-middle dysfunction but high mobility impairment group, patients with mobility dysfunction should be offered a rehabilitation program that includes improving their motor awareness and exercise capacity. Healthcare providers should take a more preventive perspective for patients in the low-dysfunction group to control the factors that worsen the disease and reduce the patient’s risk of exacerbation. This study provided new ideas for assessing the implementation of precision PR in COPD patients, beginning from the functional impairment stage, to improve the rehabilitation outcome of patients.

Author contributions

Xinyuwang was responsible for data measurement and writing the main content of the article.

Xiaoxuanmeng, Yongmeizhang, Zhenjieyu were involved in data collection.

Yaneli, Xiyu, Jingchunhe and Jinglingzhang were responsible for the diagnosis and follow-up of the patients.

Lanwang is the corresponding author of this article and is responsible for the research content and data.

Ethical Approval

This research involving human participants, human material, or human data, have been performed in accordance with the Declaration of Helsinki and have been approved by an the Medical Ethics Committees of Tianjin Medical University (TMUhMEC2020028).

Consent Form

All patients participating in this study have been informed and agreed to participate in all elements of the study.

Acknowledgements

The authors would like to thank the study participants in Tianjin Medical University General Hospital, Tianjin First Central Hospital, Tianjin Chest Hospital, Tianjin Fourth Central Hospital, for their participation. The authors also thank the investigators who assisted in the collection of patient information in these hospitals. In addition, the author would like to thank freescience for their help with the language of this article.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

This research was supported by National Natural Science Foundation of China (71804125).