Five-year changes in generic and disease-specific patient-reported outcome measures in stable COPD: a real-world cohort study

Abstract

Objective

Evidence comparing the long-term performance of generic and disease-specific patient-reported outcome (PRO) measures in chronic obstructive pulmonary disease (COPD) remains limited. This study compared five-year longitudinal changes and evaluative performance of commonly used generic and disease-specific PRO measures in stable COPD.

Methods

This secondary analysis used data from a single-center longitudinal cohort of stable COPD outpatients. Participants completed assessments at baseline and five-years later. Disease- specific instruments included the COPD Assessment Test (CAT) and St. George’s Respiratory Questionnaire (SGRQ). Generic instruments included the Medical Outcomes Study 36-Item Short Form (SF-36) and EQ-5D-5L. Distribution-based indices of five-year change were calculated using effect size (ES) and standardized response mean (SRM). Associations between changes in forced expiratory volume in one second (FEV₁) and PRO scores were examined using Spearman correlation coefficients.

Results

Thirty-nine participants (mean age 74 years) completed five-year follow up. Disease-specific instruments showed minimal change, with small ES and SRM values (<0.5), although the SGRQ Activity deteriorated significantly. In contrast, several SF-36 subscales, including Physical Functioning, Role Physical, Social Functioning, and Role Emotional, declined significantly. Physical and Role/Social Component Summary scores and the EQ-5D-5L utility index also deteriorated significantly. Moderate distribution-based changes (ES 0.5–0.8) were observed in selected SF-36 subscales and in the EQ-5D-5L index. Mean annual FEV₁ decline was approximately 20 mL/year and was not significantly correlated with changes in PRO scores.

Conclusion

In stable COPD, generic PRO measures showed greater long-term deterioration than disease-specific tools, suggesting value for monitoring gradual health decline in routine clinical practice.

PLAIN LANGUAGE SUMMARY

Chronic obstructive pulmonary disease (COPD) affects daily life as well as breathing. Doctors use questionnaires to understand symptoms, activity, and overall health. In this study, we followed people with clinically stable COPD for five years. “Stable” means no recent flare-up or treatment change at assessment, not absence of long-term health change. General health questionnaires showed declines in physical and social functioning, while COPD-specific questionnaires changed less overall, except for activity-related limitations. These changes may reflect aging, frailty, other health conditions, and COPD-related factors. The findings suggest that COPD-specific and general health questionnaires may provide complementary information for understanding gradual changes in everyday life.

Introduction

Chronic obstructive pulmonary disease (COPD) affects patients in ways that extend beyond measurable impairments in lung function, influencing daily activities, symptom burden, and overall well-being. Over recent decades, the clinical focus has increasingly shifted toward multidimensional assessment, incorporating patient-reported outcomes (PROs) such as health status and health-related quality of life (HRQOL). Although physiological indicators, particularly forced expiratory volume in one second (FEV₁), remain central to evaluating disease progression,^1,2 PROs offer complementary insights—especially in long-term monitoring, where deterioration is typically gradual and multifactorial.^3–7 The psychometric validity of PRO instruments is generally assessed along three axes: their discriminative ability (to distinguish between individuals), evaluative capacity (to detect longitudinal changes), and predictive utility (to anticipate future outcomes).^8–10 This study focuses on how different PRO instruments reflect long-term changes in perceived health status in clinically stable COPD, recognizing that such changes may reflect aging, frailty, comorbidity, and functional decline as well as COPD-related factors. Most prior studies on PRO change and responsiveness in COPD populations have been based on interventional designs with relatively short observation periods—typically no more than two to three years^7,11—and have focused on improvements rather than deterioration. However, in chronic conditions like COPD, long-term changes often reflect gradual decline rather than recovery. Importantly, the minimal important difference (MID) thresholds often differ depending on whether the health status is improving or worsening,¹² complicating the interpretation of longitudinal data.

Disease-specific tools are generally favored over generic instruments in interventional studies because they are often sensitive to short-term symptom changes.^13,14 However, it remains uncertain how generic and disease-specific instruments differ in long-term observational settings, where changes are gradual and may reflect aging, comorbidities, frailty, and functional decline as well as COPD-related factors. Therefore, this study aimed to describe five-year longitudinal changes and measurement patterns of commonly used generic and disease-specific PRO measures in individuals with stable COPD in routine clinical practice.

Methods

Study design and parent cohort participants

This study was a secondary analysis of data derived from a hospital-based longitudinal cohort of outpatients with stable COPD established at the National Center for Geriatrics and Gerontology, Japan.^5,15 The primary objective of the parent cohort was to characterize long-term trajectories of various PRO measures, and functional outcomes in clinically stable COPD under routine outpatient care. The cohort enrolled patients consecutively between 2013 and 2018 and has been described in previous publications examining longitudinal changes in respiratory symptoms and health status. The cohort study was approved by the Institutional Ethics Committee of the National Center for Geriatrics and Gerontology (No. 1138–3). The original cohort was designed as an observational study without a formal a priori sample size calculation for between-group comparisons. Instead, all eligible patients attending the outpatient clinic during the enrollment period were invited to participate in order to reflect real-world clinical practice. Therefore, the present analysis should be regarded as exploratory, focusing on descriptive patterns of long-term change rather than confirmatory comparisons between instrument types.

Participants

Eligibility criteria included: age ≥50 years; smoking history ≥10 pack-years; post-bronchodilator FEV₁/forced vital capacity (FVC) ratio <0.70; and clinical stability without exacerbation within the preceding three months. Patients with active pulmonary diseases other than COPD, physician-diagnosed asthma, or poorly controlled comorbidities requiring hospitalization were excluded. In newly diagnosed patients with COPD, health status often improves during the first six months of management.¹⁶ Therefore, to ensure that participants had reached a stable baseline following the initial treatment, the study only included those who had received continuous outpatient care for at least six months prior to enrollment. Clinical stability was defined as no acute exacerbation and no change in maintenance pharmacotherapy within four weeks prior to assessment. Throughout this manuscript, “stable COPD” refers to clinical stability of the underlying disease condition at the time of assessment, defined by the absence of recent exacerbation and recent changes in maintenance pharmacotherapy. It does not imply absence of longitudinal change in symptoms, health status, or PRO scores over the five-year follow-up. Written informed consent was obtained from all participants.

Follow-up and attrition

Participants were evaluated every six months as part of routine care, including spirometry and PRO assessments. In addition to the routine six-month assessments, comprehensive baseline data were collected between April 2017 and April 2018.¹⁷ A follow-up survey was then conducted five years later, between April 2022 and March 2023, using the same battery of questionnaires. Although routine assessments were conducted every six months, the complete PRO battery analyzed in the present study was administered only during the comprehensive baseline and five-year follow-up surveys. Therefore, intermediate time point data were not used for the present comparison because they were not consistently available for all instruments and would not allow direct comparison across generic and disease-specific measures. The number of participants with the same complete PRO battery at intermediate time points was not larger than that available for the paired baseline and five-year analysis. Of the original cohort, 39 participants completed the five-year follow-up and were included in the present analysis. Reasons for attrition included death, relocation, severe comorbidity progression, and loss to follow-up due to discontinuation of outpatient visits. Baseline characteristics of completers and non-completers were compared for assessing potential selection bias.

Pulmonary function testing

Participants were instructed to abstain from using bronchodilators for at least 12 hours before testing. Spirometry was performed more than one hour after the administration of a dry powder, long-acting bronchodilator, using a CHESTAC-8800 spirometer (Chest, Tokyo, Japan). The highest values from three acceptable maneuvers were recorded. All procedures were conducted by trained technicians in accordance with the guidelines of the American Thoracic Society and European Respiratory Society.¹⁸ Reference values for pulmonary function were derived from the Japanese Respiratory Society.¹⁹

Disease-specific measures

Disease-specific health status was assessed using validated Japanese versions of the COPD Assessment Test (CAT)^20,21 and the St. George’s Respiratory Questionnaire (SGRQ; Japanese version 2).^22,23 The CAT consists of eight items rated from 0 to 5, covering cough, sputum, dyspnea, chest tightness, exercise tolerance, activity limitation, sleep, and energy levels.²¹ Total scores range from 0 to 40, with higher scores indicating greater impairment. The SGRQ comprises 50 items grouped into three domains—Symptoms, Activity, and Impact—with each score ranging from 0 to 100; a total score is also calculated, where higher scores reflect worse health status.²³

Generic measures

Generic HRQOL was assessed using the Japanese versions of the Medical Outcomes Study 36-Item Short Form (SF-36)^24–26 and EuroQol five-dimension five-level questionnaire (EQ-5D-5L).^27,28 The SF-36 version 2 includes eight subscales: Physical Functioning (PF), Role Physical (RP), Bodily Pain (BP), General Health (GH), Vitality (VT), Social Functioning (SF), Role Emotional (RE), and Mental Health (MH). For each subscale, item responses were transformed to scores ranging from 0 to 100 according to standard SF-36 scoring procedures, with higher scores indicating better health status. These transformed subscale scores are presented in Table 2. To allow comparison with the Japanese general population, norm-based scores were also calculated for each subscale, with a mean of 50 and a standard deviation of 10 based on Japanese population norms.^24,25 These norm-based subscale scores are presented in Table 3. Additionally, three component summary scores were calculated—Physical Component Summary (PCS), Mental Component Summary (MCS), and Role/Social Component Summary (RCS)—using a Japanese three-component scoring algorithm established in 2011.²⁹

Table 2.

Comparison of airflow limitation and scores of patient-reported outcome tools obtained at baseline and five years later in 39 participants with stable chronic obstructive pulmonary disease.

		Baseline		Five years later		Comparison
		Median	IQR	Median	IQR	p Valuea
Age (years)		74	71–78	79	76–83	—
FEV1 (Liters)		2.04	1.59–2.33	1.95	1.38–2.26	<0.001*
FEV1 (% pred)		77.2	56.5–95.4	78.9	55.5–94.3	0.581
CAT score (0–40)		7	2–11	5	2–15	0.152
SGRQ Total Score		17.8	6.7–31.6	19.6	8.1–38.0	0.058
	SGRQ symptoms	34.4	19.4–47.4	30.7	16.3–53.9	0.686
	SGRQ activity	29.5	6.0–47.7	41.1	0.0–59.5	0.036*
	SGRQ impact	8.8	0.0–15.1	6.2	1.6–20.6	0.131
SF36	PF	80.0	70.0–95.0	75.0	50.0–90.0	<0.001*
SF36	RP	87.5	68.8–100.0	68.8	37.5–100.0	0.004*
SF36	BP	100.0	74.0–100.0	84.0	62.0–100.0	0.192
SF36	GH	52.0	50.0–67.0	57.0	45.0–67.0	0.394
SF36	VT	75.0	62.5–87.5	68.8	56.3–81.3	0.054
SF36	SF	100.0	87.5–100.0	100.0	75.0–100.0	0.008*
SF36	RE	91.7	58.3–100.0	66.7	50.0–100.0	0.004*
SF36	MH	80.0	65.0–90.0	75.0	60.0–85.0	0.025*
SF36	PCS	46.5	34.3–51.0	41.9	30.7–50.1	0.027*
SF36	MCS	57.3	52.0–62.5	57.3	53.8–63.6	0.410
SF36	RCS	51.8	41.3–56.0	43.7	25.3–52.8	0.024*
EQ-5Db	utility (0–1)	1.000	0.875–1.000	0.895	0.782–1.000	0.001*

Values are presented as median (interquartile range). SGRQ scores and SF-36 subscale scores range from 0 to 100. SF-36 component summary scores were calculated using the Japanese three-component scoring algorithm. Higher scores indicate worse health status for CAT and SGRQ and better health status for SF-36 and EQ-5D-5L.

^a Wilcoxon signed-rank test.

^b N = 38.

*Statistically significant.

Abbreviations. FEV₁, forced expiratory volume in one second; CAT, COPD Assessment Test; SGRQ, St. George’s Respiratory Questionnaire; PF, Physical Functioning; RP, Role Physical; BP, Bodily Pain; GH, General Health; VT, Vitality; SF, Social Functioning; RE, Role Emotional; MH, Mental Health; PCS, Physical Component Summary; MCS, Mental Component Summary; RCS, Role/Social Component Summary.

Table 3.

Comparison of norm-based scores of SF-36 in which each subscale score has a mean of 50 and a standard deviation of 10 in the Japanese general population obtained at baseline and five years later in 39 participants with stable chronic obstructive pulmonary disease.

		Baseline		Five years later		Comparison
		Median	IQR	Median	IQR	p Valuea
SF36	PF	43.4	36.2–54.2	39.8	21.7–50.6	<0.001b
SF36	RP	49.1	39.1–55.7	39.1	22.5–55.7	0.004b
SF36	BP	61.7	50.1–61.7	54.6	44.7–61.7	0.192
SF36	GH	44.2	43.1–52.2	46.9	40.5–52.2	0.394
SF36	VT	56.3	49.8–62.7	53.0	46.6–59.5	0.054
SF36	SF	57.0	50.6–57.0	57.0	44.1–57.0	0.008b
SF36	RE	51.9	35.3–56.1	39.4	31.1–56.1	0.004b
SF36	MH	54.5	46.5–59.9	51.8	43.8–57.2	0.025b

Norm-based scores are standardized to the Japanese general population, with a mean of 50 and standard deviation of 10 for each SF-36 subscale. Values are presented as median and interquartile range.

^a Wilcoxon signed-rank test.

^b Statistically significant.

Abbreviations. PF, Physical Functioning; RP, Role Physical; BP, Bodily Pain; GH, General Health; VT, Vitality; SF, Social Functioning; RE, Role Emotional; MH, Mental Health.

The EQ-5D-5L, developed by the EuroQol Group,²⁷ comprises a five-dimension descriptive system (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression), each rated on five severity levels, along with a visual analogue scale (EQ VAS). Utility scores (EQ-5D index) were derived using the Japanese time trade-off tariff.²⁸ Frailty was also assessed using the Kihon Checklist, a widely used Japanese tool for screening frailty.^30,31 Its total score ranges from 0 (no frailty) to 25 (severe frailty), with participants categorized as robust (0–3), pre-frail (4–7), or frail (8–25) based on their total score.

All PRO instruments were administered in validated Japanese versions using paper booklets, completed by participants under staff supervision. One of the authors (KN) reviewed all responses to ensure completeness and identify any missing items.

Statistical analysis

Results were expressed as mean ± standard deviation (SD) or median and interquartile range (IQR), as appropriate. The magnitude of longitudinal change was summarized using distribution-based indices, including effect size (ES) and standardized response mean (SRM), which quantify the magnitude of change relative to score variability and are commonly used in studies evaluating longitudinal performance of PRO instruments.^32–35 These indices were used descriptively and were not intended to establish responsiveness to a predefined clinical anchor or known clinical change. ES was calculated as the mean change in score divided by the SD of baseline scores; SRM was calculated as the mean change divided by the SD of the change scores. According to Cohen, the magnitude of ES and SRM was interpreted using absolute values, with values from 0.2 to <0.5, 0.5 to <0.8, and ≥0.8 classified as small, moderate, and large, respectively.³³ These thresholds were also applied to interpret SRM values.³² Between-group comparisons were made using the Mann–Whitney U test, and within-group changes from baseline were assessed using the Wilcoxon signed-rank test. Associations among outcome measures were analyzed using Spearman’s rank correlation. Exploratory associations among changes in PRO measures were also examined using Spearman’s rank correlation coefficients. Changes were calculated as five-year follow-up minus baseline. Because of the small sample size and multiple comparisons, these analyses were interpreted descriptively. All statistical analyses were performed using SPSS Statistics, version 28.0 (IBM Corp.), with a two-sided p-value <0.05 considered statistically significant.

Results

Of the original cohort, 39 participants completed the five-year follow-up with the same PRO assessment battery and were included in the present analysis. Baseline characteristics of the study population are summarized in Table 1. The mean age was 74 years, and the study population consisted predominantly of men. Nearly half of the participants were classified as GOLD 1,³⁶ indicating that this cohort mainly represented older, clinically stable COPD outpatients with relatively preserved lung function. Although 47 patients who did not complete the five-year follow-up were excluded from the analysis, a comparison between completers and dropouts showed no statistically significant differences in baseline scores, except for one item (Supplementary Table S1).

Table 1.

Characteristics at baseline of 39 patients with chronic obstructive pulmonary disease who completed a five-year study.

		Mean	SD	Median	IQR	Max	Min
Age	Years	74.3	5.5	74	71–78	84	62
Cumulative smoking	Pack-years	56.0	29.0	50	39–60	138	16
FVC	% pred.	99.3	21.5	98.4	80.7–116.4	140.7	53.6
FEV1	Liters	1.90	0.63	2.04	1.59–2.33	3.16	0.65
FEV1	% pred.	74.3	25.3	77.2	56.5–95.4	115.1	24.9
FEV1/FVC	%	58.1	11.9	63.1	53.3–67.8	68.9	27.9
Sex (male/female)	Number (%)	37 (95%)/2 (5%)
GOLD stage (GOLD 1/GOLD 2/GOLD 3 + GOLD 4)	Number (%)	18 (46%)/14 (36%)/7 (18%)
Robust/Pre-frail/Frail	Number (%)	23 (59%)/12 (31%)/4 (10%)

Abbreviations. SD, standard deviation; IQR, interquartile range; FVC, forced vital capacity; FEV₁, forced expiratory volume in one second; GOLD, Global Initiative for Chronic Obstructive Lung Disease.

Table 2 presents the comparison of PRO scores at baseline and at the five-year follow-up. Over the five-year follow-up, disease-specific instruments showed relatively small changes, whereas several domains of generic instruments demonstrated significant deterioration. Among the disease-specific measures, no statistically significant changes were observed except for the SGRQ Activity component. In contrast, statistically significant declines were observed in several generic measures, including five SF-36 subscales (PF, RP, SF, RE, and MH), two summary scores (PCS and RCS), and the EQ-5D-5L utility index. These findings suggest measurable deterioration in these domains over the five-year period. No significant changes were detected in the remaining three SF-36 subscales (BP, GH, and VT) or in the MCS score. Since the SF-36 allows comparison with national norms using norm-based scoring, a parallel analysis was conducted. This analysis similarly identified statistically significant changes in PF, RP, SF, RE, and MH subscales (Table 3). Because SF-36 norm-based scores are standardized to the Japanese general population with a mean of 50 and standard deviation of 10, these scores provide a reference for comparison with population norms. At baseline, median norm-based scores were below the population mean for PF (43.4) and GH (44.2), close to the population mean for RP (49.1) and RE (51.9), and above the population mean for BP (61.7), VT (56.3), SF (57.0), and MH (54.5). At five years, median scores declined below or further below the population mean in PF (39.8), RP (39.1), and RE (39.4), indicating deterioration particularly in physical and role-related subscales. Because several SF-36 subscales showed high baseline scores, potential ceiling effects were examined. At baseline, ceiling scores were observed in 72% of participants for SF, 51% for BP, 49% for RE, and 36% for RP, indicating substantial ceiling effects in selected SF-36 subscales.

Distribution-based indices of change over five years were calculated using ES and SRM (Table 4). In disease-specific instruments, all ES and SRM values were below 0.5, suggesting a small effect according to Cohen’s criteria. In contrast, several generic measures showed moderate distribution-based change: PF, RE, RCS, and the EQ-5D-5L utility index showed both ES and SRM values between 0.5 and 0.8. RP and SF subscales showed moderate change by ES, although SRM values remained within the small effect range.

Table 4.

The effect size and standardized response mean for patient-reported outcome tools over five years.

		Effect size	Standardized response mean
CAT score		0.246	0.274
SGRQ total score		0.262	0.324
	SGRQ symptoms	−0.024	−0.026
	SGRQ activity	0.291	0.327
	SGRQ impact	0.280	0.263
SF36	PF	−0.511	−0.613
SF36	RP	−0.621	−0.492
SF36	BP	−0.228	−0.204
SF36	GH	−0.097	−0.111
SF36	VT	−0.292	−0.370
SF36	SF	−0.628	−0.454
SF36	RE	−0.530	−0.506
SF36	MH	−0.352	−0.390
SF36	PCS	−0.290	−0.358
SF36	MCS	0.097	0.105
SF36	RCS	−0.710	−0.526
EQ-5D-5L	Utility	−0.710	−0.521

n = 39, except for EQ-5D-5L utility (n = 38). Positive values indicate deterioration for CAT and SGRQ, whereas negative values indicate deterioration for SF-36 and EQ-5D-5L. Abbreviations. CAT, COPD Assessment Test; SGRQ, St. George’s Respiratory Questionnaire; PF, Physical Functioning; RP, Role Physical; BP, Bodily Pain; GH, General Health; VT, Vitality; SF, Social Functioning; RE, Role Emotional; MH, Mental Health; PCS, Physical Component Summary; MCS, Mental Component Summary; RCS, Role/Social Component Summary.

To explore factors associated with changes in PRO scores, we investigated the relationship between changes in FEV₁ and changes in PRO measures (Supplementary Table S2). Over the five-year period, FEV₁ declined by an average of 20 mL per year. Changes in FEV₁ were not significantly correlated with changes in any PRO scores (Spearman r = 0.020 to 0.174 in absolute values, all p > 0.05). Exploratory correlations among changes in PRO measures are shown in Supplementary Table S3. Because of the small sample size and multiple comparisons, these correlations were interpreted descriptively. Change in the SGRQ Activity component was significantly associated with change in EQ-5D-5L utility (rs = −0.420, p = 0.009) and showed a borderline association with change in SF-36 Physical Functioning (rs = −0.315, p = 0.051). Because higher scores indicate worse health status for SGRQ, whereas lower scores indicate worse health status for SF-36 and EQ-5D-5L, these negative correlations indicate concordant deterioration.

Frailty was concurrently assessed using the Kihon Checklist. At baseline, 23 participants (59%) were categorized as robust, 12 (31%) as pre-frail, and 4 (10%) as frail. At follow-up, the distribution had shifted: 14 (36%) were robust, 15 (38%) pre-frail, and 10 (26%) frail. Over the five-year period, 8 participants progressed from robust to pre-frail, 4 from pre-frail to frail, and 2 from robust to frail. Participants were divided into two groups based on whether frailty classification worsened (n = 14) or did not worsen (n = 25). A comparison of changes in PRO scores between these two groups is presented in Supplementary Table S4. While the deterioration in EQ-5D-5L utility scores was more pronounced in the group with worsening frailty, no statistically significant differences were observed for other PRO measures.

Discussion

This exploratory study described five-year changes in generic and disease-specific PRO measures among older, clinically stable COPD outpatients. Generic instruments showed clearer deterioration in several physical and role/social domains, whereas disease-specific measures showed smaller overall changes except for the SGRQ Activity component. These findings should be interpreted as descriptive patterns of long-term PRO change rather than definitive evidence of comparative instrument performance.

The larger changes observed in generic HRQOL measures should not be interpreted as evidence that generic instruments better track the course of COPD. Generic instruments are designed to capture broad deterioration in health status regardless of cause, including changes related to aging, frailty progression, comorbid disease burden, and functional decline. In contrast, disease-specific instruments may remain more focused on respiratory symptoms and COPD-related activity limitations. Thus, the present findings suggest complementary measurement patterns rather than superiority of one instrument type over another.

The term “stable COPD” should be interpreted as clinical stability of the underlying disease condition at the time of assessment, rather than absence of longitudinal change in symptoms, health status, or PRO scores. Thus, although the same participants were followed over five years, they were considered clinically stable because they had no recent exacerbation or recent change in maintenance pharmacotherapy before assessment. This distinction is important for interpreting the present findings: even among clinically stable COPD outpatients, gradual deterioration in physical function, frailty status, comorbid burden, and perceived health may occur over time. This may partly explain why disease-specific symptom-oriented measures showed relatively small changes, whereas generic instruments captured broader deterioration in physical and role/social domains.

The present study differs from most previous research, which has primarily focused on short-term improvements over periods of up to two or three years, by examining gradual deterioration in HRQOL over an extended follow-up. The clinical context of long-term decline differs fundamentally from that of short-term interventional response; therefore, different patterns of score change between instrument types are not unexpected. In this setting, generic measures showed larger absolute and distribution-based changes over the five-year interval, but these changes should not be interpreted as COPD-specific progression. These findings underscore the importance of considering the time horizon and broader health context when interpreting changes in PRO scores.

Over the five-year follow-up, changes in SF-36 subscale and summary scores were not uniform. Notably, PF, RE, and the RCS showed statistically significant declines, indicating clear deterioration in subscales related to physical capacity and social participation. Previous cross-sectional studies have shown that exercise tolerance and dyspnea are major contributors to HRQOL in COPD,^22,37 and these factors may increasingly influence social and role functioning over time. In contrast, BP and MCS showed little or no statistically significant change. Although pain has been reported in some individuals with COPD,³⁸ our prior survey suggested that it was not sufficiently prevalent to indicate systematic under-recognition.³⁹ Thus, the relative stability of BP observed in the present study is consistent with earlier findings. The norm-based analysis further indicated that physical and role-related SF-36 subscales were below or declined below Japanese population norms, supporting the interpretation that deterioration was most evident in functional domains rather than uniformly across all aspects of generic HRQOL.

Despite the frequent coexistence of COPD with psychosocial conditions such as depression and anxiety, MCS scores remained stable over five years. This observation aligns with prior longitudinal research by Oga et al., who reported statistically significant but modest annual increases in Hospital Anxiety and Depression Scale (HADS) scores for anxiety (+0.16/year) and depression (+0.17/year).⁶ Given our relatively small sample size and the exclusion of dropouts, the study may have lacked sufficient statistical power to detect such subtle changes.

There is relatively limited literature examining long-term changes in generic HRQOL measures in COPD. Choate et al. reported that, over six years, individuals with alpha-1 antitrypsin deficiency-related COPD showed decline mainly in physical SF-36 subscales, whereas mental subscales remained relatively stable.⁴⁰ Similarly, Oishi et al. found that several SF-36 subscales declined over five years in community-dwelling middle-aged and older adults, while MCS and Mental Health scores showed little change.⁴¹ These findings are consistent with the present results, in which deterioration was more evident in physical and role/social domains than in mental domains. The EQ-5D-5L utility index also showed statistically significant deterioration and moderate distribution-based change in our cohort. This contrasts with short-term intervention studies, in which disease-specific tools often appear more responsive,⁴² but is broadly consistent with long-term observational data from the German COSYCONET cohort showing gradual decline in EQ-5D scores over six years.⁴³ Together, these findings suggest that generic instruments may capture broader deterioration in overall health status during long-term follow-up, although the causes of such deterioration cannot be determined from the present observational design.

Interpretation of long-term changes in generic HRQOL measures in elderly COPD patients requires caution. Frendl and Ware reported that, in randomized, double-blind, placebo-controlled drug trials using the SF-36, primary clinical outcomes and SF-36 outcomes were generally concordant, but clinically efficacious therapies for COPD did not achieve the MID threshold for SF-36 summary scores.⁴⁴ This suggests that even in controlled trials, generic SF-36 PCS and MCS outcomes may not fully reflect COPD-specific clinical effects. In the present observational cohort, without a control group or predefined clinical anchor, changes in generic HRQOL measures should be interpreted as broader health deterioration that may include aging, frailty, comorbid burden, and functional decline, rather than COPD-specific progression alone.

The significant declines observed in SF-36 physical, role, and social subscales and in the EQ-5D-5L utility index should be interpreted in the context of aging. Such changes may be expected over five years in older adults and may reflect broader HRQOL deterioration regardless of cause, including aging, frailty progression, comorbid burden, and functional decline. This broad sensitivity is a strength of generic HRQOL instruments, but without a comparison group, the observed changes cannot be attributed specifically to COPD progression.

In addition to aging, multimorbidity may contribute to declines in generic HRQOL measures.^45,46 Because comorbidities were not comprehensively quantified in the present analysis, we could not distinguish age-related decline from the effects of multiple chronic conditions. Thus, deterioration in generic measures should be interpreted as broader health deterioration rather than COPD-specific progression alone.

Another notable finding was the lack of a significant correlation between changes in FEV₁ and changes in PRO scores. This observation should be interpreted cautiously because the association between lung function and generic HRQOL is often weak, and substantial variability in HRQOL exists among patients with similar levels of airflow limitation. With only 39 participants, the present study was underpowered to evaluate small shared variance between changes in FEV₁ and changes in PRO scores. Therefore, the absence of significant correlations should not be interpreted as evidence of no relationship. Rather, it reinforces the need to interpret PRO changes as multidimensional health changes that may reflect aging, frailty progression, comorbid burden, and functional decline as well as COPD-related factors.

The distribution-based indices used in this study should be interpreted cautiously. ES and SRM quantify the magnitude of score change relative to variability, but they do not establish responsiveness to a predefined clinical anchor or known clinical change. In the present exploratory analysis, these indices were used to summarize the magnitude of five-year change in PRO scores. Therefore, the observed differences between instruments should be interpreted as descriptive measurement patterns rather than definitive evidence that one instrument type is more responsive than another. Because these indices were calculated from two comprehensive assessment points, they should be interpreted as summaries of paired five-year differences rather than as estimates of continuous longitudinal trajectories.

The mean increase in the SGRQ Total score over five years was approximately four points (median 17.8 to 19.6; mean ± SD 20.8 ± 15.2 to 24.8 ± 20.6), which is comparable to the MID for this instrument, suggesting modest but potentially clinically meaningful deterioration in overall health status. Notably, deterioration was more pronounced in the SGRQ Activity component, where the median score increased from 29.5 to 41.1, indicating substantial worsening of activity-related health status. This pattern was consistent with the decline observed in the physical functioning subscales of the SF-36, suggesting that long-term changes in COPD may be reflected more strongly in physical functioning than in respiratory symptoms alone. Although the median CAT score decreased from 7 to 5, the paired comparison did not show a statistically significant change (p = 0.152). Because the CAT primarily reflects current symptom burden, its score may have been influenced by treatment optimization and day-to-day symptom variability. The apparent divergence between CAT and the SGRQ Activity and SF-36 PF subscales likely reflects differences in the constructs assessed, with activity and functional limitation deteriorating gradually even when respiratory symptoms remain relatively stable.

Although the SGRQ is a disease-specific instrument, its activity component assesses limitations in physical activity and daily functioning rather than respiratory symptoms alone. Therefore, it overlaps conceptually with generic domains related to physical functioning, role/social functioning, mobility, usual activities, and overall health status. The exploratory correlation analysis among PRO changes further supported this interpretation: changes in SGRQ Activity were significantly associated with changes in EQ-5D-5L utility and showed a borderline association with SF-36 Physical Functioning. These associations should be interpreted cautiously because of the small sample size, but they help explain why deterioration in SGRQ Activity was directionally consistent with declines in selected generic measures.

Several limitations should be considered when interpreting these findings. First, the number of participants completing the five-year follow-up was modest, reflecting the challenges of maintaining long-term observational cohorts in COPD.⁴⁷ This limited statistical precision and precluded definitive conclusions regarding the comparative performance of generic and disease-specific PRO instruments. Second, changes were estimated from only two comprehensive assessment points over five years. Therefore, the observed differences may have been influenced by random within-person variability, including day-to-day symptom fluctuation, seasonal variation, and regression to the mean. The two-point design also precluded modeling of individual trajectories or nonlinear patterns of change. Nevertheless, the paired five-year observations using the same complete PRO battery provide exploratory information on long-term change patterns across instruments. The small sample size also limited our ability to evaluate associations between changes in FEV₁ and changes in PRO scores, particularly because lung function and generic HRQOL measures may share only a small proportion of variance. Third, this was a single-center study, and the study population consisted predominantly of older men, with nearly half of participants classified as GOLD 1. This demographic and clinical composition limits the generalizability of the findings to broader COPD populations, including women, younger patients, and those with more advanced airflow limitation. Fourth, frailty was assessed using the Kihon Checklist, a screening tool developed in Japan, which may restrict direct international comparability. Fifth, exacerbation frequency and comorbid conditions were not comprehensively incorporated into the present analysis and may have influenced longitudinal changes in HRQOL. In particular, we could not distinguish age-related decline from the effects of multimorbidity or other chronic conditions, which are known to affect generic HRQOL measures. Although SF-36 norm-based scores allowed comparison with Japanese population norms, no age- and sex-matched non-COPD comparison group was included; therefore, we could not distinguish COPD-related changes from aging-related decline or other causes of HRQOL deterioration, and these comparisons should be interpreted descriptively. Additionally, several SF-36 subscales showed substantial ceiling effects at baseline, particularly SF, BP, RE and RP. These ceiling effects may have limited the ability of these subscales to detect further improvement and may have influenced the magnitude of observed longitudinal change. Although the present analysis focused primarily on deterioration over five years, the SF-36 findings should be interpreted with this potential measurement limitation in mind. Larger longitudinal studies including more diverse populations and more comprehensive clinical characterization are therefore needed to confirm these findings. Nevertheless, the extended five-year follow-up provides valuable real-world insight into the long-term changes in PRO measures in subjects with stable COPD.

From a practical perspective, the present findings may help clinicians select PRO instruments according to the purpose of long-term monitoring in routine care. Disease-specific instruments remain useful for assessing respiratory symptoms and COPD-related activity limitation, whereas generic instruments may provide complementary information on broader health deterioration, including physical, role/social, and utility-related domains. In real-world COPD follow-up, where aging, frailty, comorbidities, and functional decline often coexist, using both disease-specific and generic PRO measures may provide a more comprehensive view of patients’ long-term health status than either instrument type alone.

Conclusions

This study provides exploratory longitudinal evidence regarding changes in commonly used generic and disease-specific PRO instruments in stable COPD. Over a five-year observational period, generic instruments captured clearer deterioration in several domains of HRQOL, particularly those reflecting physical and social functioning. These findings do not suggest that generic instruments are inherently superior to disease-specific tools, but rather that different instruments may capture different aspects of health status depending on the clinical context and time horizon. In long-term observational settings characterized by gradual health decline, generic instruments may provide complementary information that is not fully reflected by disease-specific measures alone. These results highlight the importance of selecting PRO instruments according to the purpose of measurement, particularly when monitoring long-term health changes in patients with stable COPD. Further studies in larger and more diverse COPD populations are needed to determine the generalizability of these findings.

Funding Statement

This study was partly supported by Research Funding for Longevity Sciences (grant number 25-23) from the National Center for Geriatrics and Gerontology (NCGG), Japan. The funding body had no role in the design of the study; data collection, analysis, or interpretation; or in the writing of the manuscript.

Transparency

Declaration of financial/other relationships

The authors declare that they have no financial or other relationships that could be perceived as potential conflicts of interest.

Author contributions

The authors meet the authorship criteria recommended by the International Committee of Medical Journal Editors. All authors made important contributions to the work, including the conception, study design, execution, data acquisition, analysis, and interpretation, or a combination of these areas. KN (Koichi Nishimura) took the lead in designing the research, collecting and analyzing data, and drafting the manuscript. MK and AS contributed to data acquisition and critical revision of the manuscript. KN (Kazuhito Nakayasu) was primarily responsible for the statistical analysis. All authors critically revised the manuscript, approved the final version, and agree to be accountable for all aspects of the work.

Data availability statement

The datasets generated and/or analyzed during the current study can be found in the article and its supplementary information files. Anonymized participant data will also be made available upon reasonable request to the corresponding author.

Complying with ethics of experimentation

The study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Ethics Committee of the National Center for Geriatrics and Gerontology (No. 1138–3).

Consent

Written informed consent was obtained from all participants prior to enrollment.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this manuscript, the authors used generative AI-based language tools (ChatGPT by OpenAI) to assist in revising sentence structure, improving clarity, and checking grammar. All intellectual content, scientific conclusions, and final decisions were made entirely by the authors, who take full responsibility for the integrity and accuracy of the work.