Authors: Xinyi Li, Xuedi Ma, Wangjun Qin, Changcheng Shi, Lihong Liu, Chen Wang
Categories: Review, COPD, pharmacist-led integrated management, randomized controlled trials, meta-analysis
Source: International Journal of Chronic Obstructive Pulmonary Disease
Doi: 10.2147/COPD.S589904
Authors: Xinyi Li, Xuedi Ma, Wangjun Qin, Changcheng Shi, Lihong Liu, Chen Wang
Chronic Obstructive Pulmonary Disease (COPD) is a leading cause of morbidity and mortality worldwide, and suboptimal medication management contributes to exacerbations and preventable healthcare utilization. Pharmacist-led integrated care has the potential to improve medication use and clinical outcomes. We conducted a systematic review and meta-analysis to evaluate the effects of pharmacist-led interventions in COPD.
This systematic review and meta-analysis was conducted and reported in accordance with PRISMA 2020. We searched PubMed, Embase, and Web of Science from inception until June 23, 2025. Randomized controlled trials (RCTs) assessing the effects of pharmaceutical care on clinical outcomes in COPD patients were included. A random-effects model was used to estimate pooled relative risks (RRs) or mean differences (MDs) with 95% confidence intervals (CIs). Risk of bias was assessed using the Cochrane Risk of Bias tool.
A total of 11 randomized controlled trials involving 2313 participants were included. Pharmacist-led interventions were associated with a lower risk of exacerbation-related hospital admissions (RR = 0.43, 95% CI: 0.33–0.55). Improvements in medication adherence and higher smoking cessation rates were also observed. Improvements in health-related quality of life were reported; however, substantial heterogeneity was present. In contrast, effects on COPD Assessment Test scores and objective disease measures, including lung function, were non-significant. Overall study quality was variable, with many trials being small and at high risk of bias.
Pharmacist-led interventions in COPD may improve selected medication-related and patient-centered outcomes; however, the available evidence is heterogeneous and limited by study quality and inconsistent effects across outcomes. These findings should be interpreted cautiously, and well-designed, adequately powered trials with standardized outcomes are needed before robust conclusions regarding clinical effectiveness can be drawn.
Chronic obstructive pulmonary disease (COPD) is a common chronic airway disease characterized by persistent airflow limitation, chronic respiratory symptoms and structural pulmonary abnormalities.1 It has become a major public health issue worldwide, with significant epidemiological impact.2 COPD affects more than 400 million people globally and is the third leading cause of death worldwide,3,4 responsible for over 3 million deaths annually.5 The economic burden is equally substantial, with estimated global costs INT$4.326 trillion in 2020–50.6 In China, the situation is particularly severe, with nearly 100 million people affected,7 placing immense pressure on the healthcare system and society. Improving treatment outcomes for these patients is therefore urgently needed.
To effectively manage symptoms, patients with COPD often require long-term adherence to inhalation therapy. However, despite considerable evolution in inhalation therapies, more than half of patients make critical errors in their use, and medication adherence remains poor.8–10 Compounding this issue, COPD patients often face polypharmacy due to frequent comorbidities, which increases the risk of adverse drug reactions, interactions, and overall treatment complexity.11 These challenges contribute to suboptimal disease control, reduced quality of life, and a higher risk of exacerbations and hospitalizations.12,13 Consequently, enhancing inhaler technique, adherence, and overall medication management is essential to improving prognosis.
Given their expertise in pharmacotherapy, clinical pharmacists are well-positioned to deliver value-added services such as medication management and adherence support for patients with chronic conditions including COPD, potentially leading to improved health outcomes. For instance, pharmacists can provide structured education on proper inhaler use, explain the purpose of treatment, dosing frequency, potential side effects, and drug interactions. They can also offer counseling on lifestyle modifications and continuous adherence support.14 Through improving effective drug delivery and persistence with maintenance therapy, such interventions may enhance real-world treatment effectiveness and reduce exacerbation-related utilization. Although randomized controlled trials (RCTs) have evaluated pharmacist-led interventions in COPD, existing evidence syntheses have limitations.
Several systematic reviews15 and meta-analyses have examined pharmacist-led interventions in patients with COPD. Earlier reviews primarily focused on specific aspects of care, such as medication adherence, inhaler technique, or health-related behaviors, and many are now outdated.16 More recent reviews have provided valuable summaries of pharmacist involvement in COPD management; however, these syntheses often included a limited range of outcomes,17 placed less emphasis on exacerbation-related healthcare utilization, or did not comprehensively integrate multiple clinically relevant endpoints. In addition, prior reviews generally provided limited critical appraisal of between-study heterogeneity arising from differences in intervention components, follow-up duration, outcome measurement tools, and healthcare settings. As a result, uncertainty remains regarding the consistency and strength of evidence supporting pharmacist-led integrated management across key clinical outcomes.
To address these gaps, we conducted an updated systematic review and meta-analysis of randomized controlled trials to comprehensively synthesize evidence across exacerbation-related healthcare utilization, quality of life, symptom burden, medication adherence, and smoking cessation, while explicitly considering heterogeneity and study quality.
The protocol for this systematic review has been registered with PROSPERO (CRD420251274607). This study adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.18 See eTable 1 for details.
Two independent investigators conducted a systematic literature search in the following electronic Embase, PubMed, and Web of Science. The search period spanned from the establishment of each database to June 23, 2025. The search strategy incorporated a combination of keywords and MeSH terms related to “pharmaceutical care” and “COPD”, along with their synonyms and variations. The search was restricted to RCTs. Furthermore, the reference lists of all retrieved articles and relevant reviews were manually screened to identify additional eligible studies. The complete search strategies for each database are provided in eTable 2.
RCTs investigating the effects of pharmaceutical care on outcomes in patients with COPD were included. The exclusion criteria were as (1) non-randomized studies; (2) studies involving non-pharmacist-led interventions; (3) duplicate publications.
Two reviewers independently extracted data from each included study using a pre-designed data extraction form. The following information was (1) first author, publication year, and country; (2) study design, duration, and all primary and secondary outcomes; (3) intervention and comparator group details, including number of patients and baseline characteristics (eg, mean age).
The methodological quality and risk of bias of the included studies were assessed independently by two reviewers using the Cochrane Risk of Bias Tool (RoB 1). The tool evaluates the following random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other potential sources of bias. Each domain was rated as having low, unclear, or high risk of bias.
Meta-analyses were conducted using Review Manager (RevMan) software version 5.4 when at least three studies provided sufficient and comparable outcome data. Random-effects models were applied to incorporate potential between-study heterogeneity. For dichotomous outcomes, results were pooled and expressed as risk ratios (RR) with 95% confidence intervals (95% CIs). For continuous outcomes, mean differences (MD) and 95% CIs were calculated. When continuous outcomes were reported as medians and interquartile ranges (IQRs), means and standard deviations (SD) were estimated using validated methods19,20 to enable inclusion in the meta-analysis.
Heterogeneity was assessed using the I^2^ statistic. I^2^ values were interpreted as 0%-40% indicated negligible heterogeneity, 30%-60% moderate heterogeneity, 50%-90% substantial heterogeneity, and 75%-100% considerable heterogeneity.21 A p value < 0.05 was considered statistically significant.
A total of 811 references were identified through electronic searches. From these references, 11 studies met the inclusion criteria and were selected for our study. The details of the selection process were summarized in Figure 1. Figure 1Flow diagram of studies that were assessed and included.Flowchart of study selection process from 811 records to 11 trials.
The pooled analysis included a total of 2313 patients with COPD. The mean age of participants was consistently greater than 60 years across all studies, indicating a geriatric population. In addition, the studies demonstrated a multicenter international representation, encompassing a total of 7 countries across Asia, Europe, and the Middle East. The majority of studies were conducted in Asia (n=7), with specific contributions from China (n=3), India (n=2), and Vietnam (n=2). European countries (n=3), including Belgium, Norway, and Northern Ireland, and one study from Jordan in the Middle East were also represented. The sample sizes varied widely, ranging from 40 participants to 734 participants. The duration of the interventions and follow-up also differed considerably, spanning from short-term studies to long-term trials, with the most common durations being 6 and 12 months. The characteristics of the included trials are summarized in Table 1. Table 1Characteristics of the Included StudiesStudyCountryStudy DesignDurationPrimary OutcomeSecondary OutcomeInterventions ArmsNAge (y), Mean (SD) or Median (IQR)Khdour 200922Northern IrelandRCT121) Hospital admissions for AECOPD2) ED visits for AECOPD1) QoL assessed by SGRQ2) Medication adherence assessed by MMAS-43) FEV14) Knowledge scores assessed by COPD knowledge questionnaire developed by Scherer et al5) BMI6) Number of quit smoking1) Inhaler technique training2) Disease & medication education3) Medication adherence counseling4) Respiratory self-management training5) Lifestyle counseling8665.63 (10.1)Usual care8767.3 (9.2)Jarab 201223JordanRCT6QoL assessed by SGRQ1) Hospital admissions for AECOPD2) ED visits for AECOPD3) Medication adherence assessed by MMAS-44) FEV15) Knowledge scores assessed by COPD knowledge questionnaire developed by Scherer et al6) BMI1) Disease & medication education2) Medication adherence counseling3) Respiratory self-management training4) Lifestyle counseling6661 (14) ^†^Usual care6764 (15) ^†^Tommelein 201324BelgiumRCT31) Inhalation technique assessed by checklist2) Medication Adherence assessed by MRA1) Hospital admissions for AECOPD2) QoL assessed by EQ-5D3) Dyspnoea assessed by mMRC4) COPD-specific health status assessed by CAT5) Patients with Severe exacerbations6) Number of quit smoking1) Inhaler technique training2) Disease & medication education3) Medication adherence counseling4) Lifestyle counseling37168.4 (9.6)Usual care36368.9 (9.7)Wei 201425ChinaRCT12Medication adherence assessed by pill counts plus direct interview1) Hospital admissions for AECOPD2) QoL assessed by SGRQ1) Inhaler technique training2) Disease & medication education3) Medication adherence counseling5865.2 (8.1)Usual care5963.9 (6.2)Xin 201626ChinaRCT121) Medication Adherence assessed by MRA2) QoL assessed by SGRQ1) Hospital admissions for AECOPD2) Patients with Severe exacerbations3) Number of quit smoking1) Inhaler technique training2) Disease & medication education3) Medication adherence counseling4) Lifestyle counseling5) Follow-up visit education11464.2 (14.2)Usual care11364.6 (14.5)Abdulsalim 2016; 201827,28IndiaRCT241) QoL assessed by SGRQ272) Medication adherence assessed by MMAS-428-1) Inhaler technique training2) Disease & medication education3) Medication adherence counseling4) Respiratory self-management training5) Lifestyle counseling6) Follow-up visit education13060.6 (7.9)Usual care13061.1 (8.4)Bui 202029VietnamRCT3QoL assessed by validated Vietnamese version of the CCQ-1) Inhaler technique training2) Disease & medication education3) Medication adherence counseling4) Lifestyle counseling9263.80 (9.96)Usual care9366.08 (8.67)Liu 202130ChinaRCT6DDDs of antibacterials1) Length of stay2) Costs of hospitalization3) Cases of adverse drug reactions4) Medication adherence assessed by MMAS-85) COPD-specific health status assessed by CATComprehensive MTM9675.13 (8.03)Usual care9773.25 (7.45)Vastrad 202131IndiaRCT1QoL assessed by WHOQOL-Pharmaceutical care35-Usual care35-Kebede 202232NorwayRCT12Time to readmissionCOPD-specific health status assessed by CATInhaler technique training20Female 73.1 (9.1)Male 73.4 (7.4)Usual care20Female 74.4 (9.7)Male 74.5 (6.1)Nguyen 202433VietnamRCT1Medication adherence assessed by General Medication Adherence Scale1) Dyspnoea assessed by mMRC2) COPD-specific health status assessed by CAT3) Inhalation technique assessed by checklist1) Inhaler technique training2) Disease & medication education3) Medication adherence counseling4) Respiratory self-management training5) Lifestyle counseling6) Follow-up visit education9265.2 (9.5)Usual care8966.6 (7.0)Notes: ^†^Values are presented as median (interquartile range). References27,28 and report different outcomes from the same randomized controlled trial conducted by Abdulsalim et al.Abbreviations: RCT, Randomized Controlled Trial; AECOPD, Acute Exacerbation of COPD; ED, Emergency Department; QoL, Quality of Life; SGRQ, St. George’s Respiratory Questionnaire; MMAS, Morisky Medication Adherence Scale; FEV1, Forced Expiratory Volume in 1 second; BMI, Body Mass Index; MRA, Medication Refill Adherence; mMRC, Modified Medical Research Council; CAT, COPD Assessment Test; CCQ, Clinical COPD Questionnaire; DDD, Defined Daily Dose; MTM, Medication Therapy Management; WHOQOL, World Health Organization Quality of Life.
A prevalent limitation across the included RCTs was the high risk of blinding due to the inherent challenges in blinding the intervention, as the nature of pharmacist-led education and counseling precludes complete masking. The result of the risk of bias assessment is shown in eFigures 1 and 2.
A review of the literature reveals that pharmaceutical interventions for COPD, despite being implemented across diverse countries, consistently employ a highly similar and comprehensive model of care. The cornerstone of this model is inhaler technique training, which typically involves pharmacist demonstration, patient practical operation, and assessment using standardized checklists, supplemented with written instructions. This core component is systematically integrated with other essential elements, including disease and medication education, medication adherence counseling, and respiratory self-management training. In addition, the interventions are further supported by lifestyle counseling on smoking cessation, physical activity, and diet, and conclude with follow-up visit education to ensure continuity of care. A detailed breakdown of these interventions is provided in Table 2. Table 2Summary of Pharmacist-Led Interventions in Included TrialsInterventionItemsInhaler technique training1) Pharmacist demonstration2) Patient practical operation3) Register patient’s inhaler technique through Standardized checklists4) Written information for inhaler techniqueDisease & medication education1) Disease knowledge encompasses the definition of COPD, the disease’s pathophysiology, the interpretation of medical tests, and the rationale behind medications2) The types, indications, doses, frequency of administration, as well as the recognition and prevention of possible side effects for each prescribed medication, along with the potential effects of drug combinationsMedication adherence counseling1) The importance of adherence2) Current problems with adherenceComprehensive MTMStandard services provided by clinical pharmacists, such as pharmacy consulting, medication monitoring, and medication education for patientsRespiratory self-management training1) Self-efficacy in managing dyspnea2) Upper and lower limb exercises and relaxation techniques3) Symptom management such as pursed-lip breathing technique4) Sputum expectoration technique such as huff cough technique5) Importance of basic exercises, symptom management and sputum expectoration techniquesLifestyle counseling1) Smoking cessation2) Exercise habit3) The importance of a well-balanced diet with sufficient intake of fresh fruits and vegetablesFollow-Up visit education1) The necessity of timely follow-up by physicians2) How to prevent covid-19 when follow-up visiting
The findings suggest potential benefits of pharmaceutical care interventions across several outcome domains, although effects varied by endpoint. The detailed results for each outcome are presented below.
Pharmaceutical care interventions were associated with reductions in severe exacerbations across contributing trials, a key clinical endpoint in COPD management. The number of patients experiencing one or more severe exacerbations was significantly lower in the intervention groups compared to usual care as evidenced by individual study results (Tommelein 24 19/371 vs 33/363, p = 0.038; Xin 26 14/114 vs 28/113, p = 0.024). In contrast, the interventions showed a limited effect on objective physiological measures. Although Forced Expiratory Volume in 1 second (FEV1) values were numerically higher in the intervention groups across studies, these differences did not reach statistical significance (Khdour 22 1.19 vs 1.05, p = 0.13; Jarab 23 1.15 vs 1.06, p = 0.55).
The effects on patient-reported symptoms were mixed. One study found a reduction in the proportion of patients experiencing significant dyspnea (Modified Medical Research Council (mMRC) ≥ 2) following the intervention (Nguyen 33 46/91 vs 65/89, p =0.002), while another found no between-group difference (Tommelein 24 130/346 vs 125/346, p =0.973). The pooled estimate suggested no consistent improvement in COPD Assessment Test (CAT) score (Mean difference = −2.61, 95% CI: −7.38 to 2.15; I^2^ = 98%; p = 0.28; 3 studies24,30,33). The extremely high heterogeneity indicates that the pooled result may not reflect a common underlying treatment effect and should therefore be interpreted with considerable caution (Figure 2). The effects of pharmaceutical care on clinical measures of COPD patients were summarized in eTable 3. Figure 2Forest plot of comparison on COPD-specific health status assessed by CAT. Mean differences (MD) with 95% confidence intervals (CI) were pooled using the inverse variance (IV) method under a random-effects model. The size of each square reflects the weight assigned to each study, and horizontal lines indicate 95% CIs. The diamond represents the pooled estimate of the overall effect. Between-study heterogeneity was high (Tau^2^ = 17.40; Chi^2^ = 130.95, df = 2, P < 0.00001; I^2^ = 98%). The overall effect was not statistically significant (Z = 1.08, P = 0.28).Table Study or Subgroup; Pharmacist-led intervention with Mean, Standard Deviation, Total; Usual care with Mean, Standard Deviation, Total; Weight; Mean Difference inverse variance, Random, 95 percent confidence interval. Rows: Nguyen2024 with Pharmacist-led intervention Mean 11, Standard Deviation 4.52, Total 91; Usual care Mean 12.18, Standard Deviation 5.65, Total 89; Weight 32.9 percent; Mean Difference minus 1.18 with 95 percent confidence interval minus 2.68 to 0.32. Tommelein2013 with 15.9, 7.8, 346; 15.9, 7.7, 346; 33.3 percent; 0.00 with 95 percent confidence interval minus 1.15 to 1.15. Liu2021 with 9.03, 1.75, 96; 15.61, 2.01, 97; 33.8 percent; minus 6.58 with 95 percent confidence interval minus 7.11 to minus 6.05. Total (95 percent confidence interval): Pharmacist-led intervention Total 533; Usual care Total 532; Weight 100.0 percent; Mean Difference minus 2.61 with 95 percent confidence interval minus 7.38 to 2.15. Graph: x-axis label Mean Difference inverse variance, Random, 95 percent confidence interval, scale from minus 100 to 100 with ticks at minus 100, minus 50, 0, 50, 100. y-axis label Study or Subgroup (no unit). A vertical reference line at 0. Squares and horizontal lines plot each study at minus 1.18, 0.00 and minus 6.58 with the listed confidence intervals. A diamond for the total spans minus 7.38 to 2.15 centered at minus 2.61. Bottom captions read Favours (Pharmacist-led intervention) on the left and Favours (Usual care) on the right. The detailed data points are as - For Nguyen2024, the pharmacist-led intervention mean was 11.00, the pharmacist-led intervention SD was 4.52, the pharmacist-led intervention total was 91, the usual care mean was 12.18, the usual care SD was 5.65, the usual care total was 89, the weight was 32.9 percent and the mean difference (IV, random, 95% CI) was minus 1.18 with 95% CI minus 2.68 to 0.32. - For Tommelein2013, the pharmacist-led intervention mean was 15.90, the pharmacist-led intervention SD was 7.80, the pharmacist-led intervention total was 346, the usual care mean was 15.90, the usual care SD was 7.70, the usual care total was 346, the weight was 33.3 percent and the mean difference (IV, random, 95% CI) was 0.00 with 95% CI minus 1.15 to 1.15. - For Liu 2021, the pharmacist-led intervention mean was 9.03, the pharmacist-led intervention SD was 1.75, the pharmacist-led intervention total was 96, the usual care mean was 15.61, the usual care SD was 2.01, the usual care total was 97, the weight was 33.8 percent and the mean difference (IV, random, 95% CI) was minus 6.58 with 95% CI minus 7.11 to minus 6.05. - For Total (95% CI), the pharmacist-led intervention total was 533, the usual care total was 532, the weight was 100.0 percent and the mean difference (IV, random, 95% CI) was minus 2.61 with 95% CI minus 7.38 to 2.15.A forest plot of COPD Assessment Test mean difference across studies, with mixed effects and pooled estimate.Abbreviations: CI, confidence interval; df, degrees of freedom; IV, inverse variance; SD, standard deviation; MD, mean difference.
Pharmaceutical care was associated with reductions in selected healthcare resource utilization outcomes. Meta-analysis showed that hospital admissions for acute exacerbations of COPD (AECOPD) were significantly reduced in the intervention groups (RR = 0.43, 95% CI: 0.33 to 0.55; I^2^ = 0%; p < 0.001; N = 5 studies22–26) with no heterogeneity observed (Figure 3). In contrast, the evidence regarding emergency department (ED) visits was less conclusive, with one study reporting significant reductions (Khdour 2009,22 p = 0.02), although another study found no significant difference (Jarab 2012,23 p = 0.79). Figure 3Forest plot of comparison on hospital admissions for AECOPD. Risk ratios (RR) with 95% confidence intervals (CI) were pooled using the Mantel–Haenszel (M–H) method under a random-effects model. The size of each square reflects the weight assigned to each study, and horizontal lines represent 95% CIs. The diamond represents the pooled estimate of the overall effect. Between-study heterogeneity was low (Tau^2^ = 0.00; Chi^2^ = 3.56, df = 4, P = 0.47; I^2^ = 0%). The pooled analysis showed a statistically significant reduction in risk in the intervention group (RR = 0.43, 95% CI 0.33–0.55; Z = 6.49, P < 0.00001).Study table Study or Subgroup; Pharmacist-led intervention (Events, Total); Usual care (Events, Total); Weight; Risk Ratio M–H, Random, 95 percent CI. Rows and Jarab2012: 3 of 66 vs 11 of 67, weight 4.3 percent, risk ratio 0.28 (0.08, 0.95). Khdour2009: 18 of 71 vs 32 of 72, weight 28.8 percent, risk ratio 0.57 (0.35, 0.92). Tommelein2013: 8 of 371 vs 24 of 363, weight 10.5 percent, risk ratio 0.33 (0.15, 0.72). Wei2014: 16 of 42 vs 38 of 45, weight 39.7 percent, risk ratio 0.45 (0.30, 0.68). Xin2016: 11 of 114 vs 35 of 113, weight 16.7 percent, risk ratio 0.31 (0.17, 0.58). Totals: Total (95 percent CI) 664 vs 660, weight 100.0 percent, pooled risk ratio 0.43 (0.33, 0.55). Total 56 vs 140. Graph: x-axis label Risk Ratio M–H, Random, 95 percent CI, unit not shown; scale ticks at 0.01, 0.1, 1, 10, 100. y-axis label not shown; y-axis lists the five studies. Each study is a square with a horizontal 95 percent confidence interval line at the listed risk ratio values. A vertical reference line is at 1. A diamond at 0.43 spans 0.33 to 0.55. Bottom annotations read Favours [Pharmacist-led intervention] on the left and Favours [Usual care] on the right. The detailed data points are as - For Jarab2012, pharmacist-led intervention events were 3 out of a total of 66, usual care events were 11 out of a total of 67, the weight was 4.3 percent and the risk ratio (M–H, random, 95 percent CI) was 0.28 with 0.08 to 0.95 in brackets. - For Khodur2009, pharmacist-led intervention events were 18 out of a total of 71, usual care events were 32 out of a total of 72, the weight was 28.8 percent and the risk ratio (M–H, random, 95 percent CI) was 0.57 with 0.35 to 0.92 in brackets. - For Tommelein2013, pharmacist-led intervention events were 8 out of a total of 371, usual care events were 24 out of a total of 363, the weight was 10.5 percent and the risk ratio (M–H, random, 95 percent CI) was 0.33 with 0.15 to 0.72 in brackets. - For Wei2014, pharmacist-led intervention events were 16 out of a total of 42, usual care events were 38 out of a total of 45, the weight was 39.7 percent and the risk ratio (M–H, random, 95 percent CI) was 0.45 with 0.30 to 0.68 in brackets. - For Xin2016, pharmacist-led intervention events were 11 out of a total of 114, usual care events were 35 out of a total of 113, the weight was 16.7 percent and the risk ratio (M–H, random, 95 percent CI) was 0.31 with 0.17 to 0.58 in brackets. - For Total (95% CI), pharmacist-led intervention total was 664, usual care total was 660, the weight was 100.0 percent and the risk ratio (M–H, random, 95 percent CI) was 0.43 with 0.33 to 0.55 in brackets. - For Total events, pharmacist-led intervention total events were 56 and usual care total events were 140.A forest plot of hospital admissions for acute exacerbations of COPD showing lower risk with intervention.Abbreviations: CI, confidence interval; df, degrees of freedom; M–H, Mantel–Haenszel; RR, risk ratio.
Furthermore, the intervention led to more efficient and safer use of resources. The intervention group had a significantly shorter mean length of hospital stay (11.27 vs 13.46 days, p <0.0530), lower hospitalization costs (13405.45 vs 14856.51 RMB, p < 0.0530), reduced antibacterial consumption (121 vs 189 defined daily doses [DDDs], p < 0.0530), and fewer cases of adverse drug reactions (9 vs 23, p < 0.0130) (Liu 2021). There was no statistically significant effect on the time to readmission (Kebede 32 41 vs 95 days, p = 0.16). The effects of pharmaceutical care on healthcare utilization and costs of COPD patients were summarized in eTable 4.
Improvements in QoL were reported in several trials; however, heterogeneity across instruments and studies was substantial. The pooled analysis using the St. George’s Respiratory Questionnaire (SGRQ) showed improvement (Mean difference = −6.04, 95% CI: −11.10 to −0.98; I^2^ = 80%; p = 0.02; 3 studies22,25,26). However, substantial heterogeneity was observed, indicating considerable between-study variability and limiting the certainty of the pooled estimate (Figure 4). Two additional SGRQ studies not included in the meta-analysis due to insufficient statistical data showed mixed findings, with one27 reporting a significant between-group difference and the other23 showing no statistically significant improvement. Figure 4Forest plot of comparison on QoL assessed by SGRQ. Mean differences (MD) with 95% confidence intervals (CI) were pooled using the inverse variance (IV) method under a random-effects model. The size of each square reflects the weight assigned to each study in the meta-analysis, and horizontal lines indicate 95% CIs. The diamond represents the pooled estimate of the overall effect. Between-study heterogeneity was substantial (Tau^2^ = 15.59; Chi^2^ = 9.98, df = 2, P = 0.007; I^2^ = 80%). The pooled analysis showed a statistically significant improvement in SGRQ scores in the intervention group (MD = −6.04, 95% CI −11.10 to −0.98; Z = 2.34, P = 0.02).Table Study or Subgroup; Pharmacist-led intervention (Mean, Standard Deviation, Total); Usual care (Mean, Standard Deviation, Total); Weight; Mean Difference Inverse Variance, Random, 95 percent confidence interval. Rows: Khodour2009: Pharmacist-led intervention mean 61.8, standard deviation 16.0543, total 71; Usual care mean 65.3, standard deviation 18.2988, total 72; weight 28.0 percent; mean difference minus 3.50, 95 percent confidence interval minus 9.14 to 2.14. Weiland2014: Pharmacist-led intervention mean 48.86, standard deviation 12.54, total 51; Usual care mean 52.16, standard deviation 13.59, total 53; weight 30.1 percent; mean difference minus 3.30, 95 percent confidence interval minus 8.32 to 1.72. Xin2016: Pharmacist-led intervention mean 42.7, standard deviation 3.2, total 114; Usual care mean 52.4, standard deviation 5.2, total 113; weight 41.9 percent; mean difference minus 9.70, 95 percent confidence interval minus 10.82 to minus 8.58. Total (95 percent confidence interval): Pharmacist-led intervention total 236; Usual care total 238; weight 100.0 percent; pooled mean difference minus 6.04, 95 percent confidence interval minus 11.10 to minus 0.98. Graph: x-axis label Mean Difference Inverse Variance, Random, 95 percent confidence interval, unit not shown; scale from minus 100 to 100 with ticks at minus 100, minus 50, 0, 50, 100. Left caption Favours Pharmacist-led intervention; right caption Favours Usual care. Squares and horizontal lines match each study’s mean difference and 95 percent confidence interval; the pooled diamond spans minus 11.10 to minus 0.98 centered at minus 6.04. The detailed data points are as - For Khodour2009, the pharmacist-led intervention mean was 61.8, the pharmacist-led intervention SD was 16.0543, the pharmacist-led intervention total was 71, the usual care mean was 65.3, the usual care SD was 18.2988, the usual care total was 72, the weight was 28.0 percent and the mean difference (IV, random, 95% CI) was minus 3.50 with 95% CI minus 9.14 to 2.14. - For Weid2014, the pharmacist-led intervention mean was 48.86, the pharmacist-led intervention SD was 12.54, the pharmacist-led intervention total was 51, the usual care mean was 52.16, the usual care SD was 13.59, the usual care total was 53, the weight was 30.1 percent and the mean difference (IV, random, 95% CI) was minus 3.30 with 95% CI minus 8.32 to 1.72. - For Xin2016, the pharmacist-led intervention mean was 42.7, the pharmacist-led intervention SD was 3.2, the pharmacist-led intervention total was 114, the usual care mean was 52.4, the usual care SD was 5.2, the usual care total was 113, the weight was 41.9 percent and the mean difference (IV, random, 95% CI) was minus 9.70 with 95% CI minus 10.82 to minus 8.58. - For Total (95% CI), the pharmacist-led intervention total was 236, the usual care total was 238, the weight was 100.0 percent and the mean difference (IV, random, 95% CI) was minus 6.04 with 95% CI minus 11.10 to minus 0.98.A forest plot of St. George’s Respiratory Questionnaire mean difference showing an overall negative pooled effect.Abbreviations: CI, confidence interval; df, degrees of freedom; IV, inverse variance; MD, mean difference; SD, standard deviation; SGRQ, St. George’s Respiratory Questionnaire.
Improvements in QoL were also reported in individual studies using the World Health Organization Quality of Life (WHOQOL) (Vastrad 31 64.10 vs 46.05, p < 0.001) and the Clinical COPD Questionnaire (CCQ) (Bui 29 0.81 vs 1.24, p = 0.001). In contrast, one study using the EQ-5D found no significant difference between groups (Tommelein 2013,24 p = 0.190). The effects of pharmaceutical care on health-related quality of life of COPD patients were summarized in eTable 5.
Pharmaceutical care consistently resulted in substantially improved medication adherence, as measured by a variety of tools. The proportion of patients exhibiting high adherence on the MMAS-4 was greater in the intervention groups based on the pooled analysis (RR = 1.45, 95% CI: 1.24 to 1.68; I^2^ = 14%; p < 0.001; 3 studies22,23,28) with low heterogeneity (Figure 5). This finding was corroborated by studies using the MMAS-8 (Liu 30 7.31 vs 6.05, p < 0.05), Medication Refill Adherence (MRA) (Tommelein 24 93.9 vs 85.7, p < 0.001; Xin 26 93.1 vs 83.2, p = 0.003), General Medication Adherence Scale (Nguyen 2024,33 32 vs 31, p < 0.001) and pill counts (Wei 25 66.5 vs 54.4, p = 0.039). The effects of pharmaceutical care on medication adherence of COPD patients were summarized in eTable 6. Figure 5Forest plot of comparison on medication adherence assessed by MMAS-4. Risk ratios (RR) with 95% confidence intervals (CI) were pooled using the Mantel–Haenszel (M–H) method under a random-effects model. The size of each square reflects the weight assigned to each study in the meta-analysis, and horizontal lines represent 95% CIs. The diamond represents the pooled estimate of the overall effect. Between-study heterogeneity was low (Tau^2^ = 0.00; Chi^2^ = 2.32, df = 2, P = 0.31; I^2^ = 14%). The pooled analysis showed a statistically significant increase in medication adherence in the intervention group (RR = 1.45, 95% CI 1.24–1.68; Z = 4.81, P < 0.00001).Table Study or Subgroup; Pharmacist-led intervention (Events, Total); Usual care (Events, Total); Weight; Risk Ratio Mantel–Haenszel, Random, 95 percent confidence interval. Rows: Abdulsalim2018: pharmacist-led intervention 84 events of 104 total; usual care 48 events of 98 total; weight 38.2 percent; risk ratio 1.65 with 95 percent confidence interval 1.32 to 2.06. Jarab2012: 45 of 63; 33 of 64; 25.0 percent; 1.39 with 1.04 to 1.84. Khodur2009: 55 of 71; 43 of 72; 36.9 percent; 1.30 with 1.03 to 1.63. Total (95 percent confidence interval): pharmacist-led intervention total 238; usual care total 234; weight 100.0 percent; pooled risk ratio 1.45 with 1.24 to 1.68. Total pharmacist-led intervention 184; usual care 124. Graph: x-axis label Risk Ratio Mantel–Haenszel, Random, 95 percent confidence interval (unitless), with tick labels 0.01, 0.1, 1, 10, 100. y-axis label Study or Subgroup (unitless). A vertical reference line is at 1. Each study is shown as a square with a horizontal line spanning its 95 percent confidence Abdulsalim2018 at 1.65 spanning 1.32 to 2.06; Jarab2012 at 1.39 spanning 1.04 to 1.84; Khodur2009 at 1.30 spanning 1.03 to 1.63. The pooled effect is a diamond centered at 1.45 spanning 1.24 to 1.68. Bottom annotations read Favours Pharmacist-led intervention on the left and Favours Usual care on the right. The detailed data points are as - For the study or subgroup Abdulsalim2018, pharmacist-led intervention events were 84, pharmacist-led intervention total was 104, usual care events were 48, usual care total was 98, weight was 38.2 percent and the risk ratio (M–H, random, 95 percent CI) was 1.65 with 95 percent CI 1.32 to 2.06. - For the study or subgroup Jarab2012, pharmacist-led intervention events were 45, pharmacist-led intervention total was 63, usual care events were 33, usual care total was 64, weight was 25.0 percent and the risk ratio (M–H, random, 95 percent CI) was 1.39 with 95 percent CI 1.04 to 1.84. - For the study or subgroup Khodur2009, pharmacist-led intervention events were 55, pharmacist-led intervention total was 71, usual care events were 43, usual care total was 72, weight was 36.9 percent and the risk ratio (M–H, random, 95 percent CI) was 1.30 with 95 percent CI 1.03 to 1.63. - For total (95 percent CI), pharmacist-led intervention total was 238, usual care total was 234, weight was 100.0 percent and the risk ratio (M–H, random, 95 percent CI) was 1.45 with 95 percent CI 1.24 to 1.68. - For total events, pharmacist-led intervention events were 184 and usual care events were 124.A forest plot of medication adherence risk ratios showing all studies favor pharmacist-led intervention.Abbreviations: CI, confidence interval; df, degrees of freedom; M–H, Mantel–Haenszel; RR, risk ratio.
The interventions successfully enhanced patients’ understanding of their disease and treatment. Knowledge scores, assessed by a standardized COPD questionnaire, were significantly higher in the intervention groups compared to the controls (Khdour 22 75.0 vs 59.3, p = 0.001; Jarab 23 60.7 vs 43.6, p = 0.007). The effects of pharmaceutical care on COPD-related knowledge of COPD patients were summarized in eTable 7.
A direct and profound effect of the interventions was observed on the practical skill of inhaler use. The proportion of patients demonstrating correct inhalation technique was higher in the intervention group in one study (Tommelein 24 237/346 vs 114/346, p < 0.001), a finding that was sustained in another (Nguyen 33 90/91 vs 77/89, p = 0.001). The effects of pharmaceutical care on correct inhalation technique of COPD patients were summarized in eTable 8.
The pooled analysis showed a higher quit rate in the intervention group (RR = 1.38, 95% CI: 1.12 to 1.69; I^2^ = 0%; p = 0.002; 3 studies22,24,26), although this estimate was based on a limited number of trials (Figure 6). Figure 6Forest plot of comparison on smoking cessation. Risk ratios (RR) with 95% confidence intervals (CI) were pooled using the Mantel–Haenszel (M–H) method under a random-effects model. The size of each square reflects the weight assigned to each study in the meta-analysis, and horizontal lines represent 95% CIs. The diamond represents the pooled estimate of the overall effect. Between-study heterogeneity was low (Tau^2^ = 0.00; Chi^2^ = 0.32, df = 2, P = 0.85; I^2^ = 0%). The pooled analysis showed a statistically significant improvement in the intervention group (RR = 1.38, 95% CI 1.12–1.69; Z = 3.08, P = 0.002).Table Study or Subgroup; Pharmacist-led intervention with Events and Total; Usual care with Events and Total; Weight; Risk Ratio with Mantel–Haenszel, Random, 95 percent confidence interval. Rows: Khdour2009: pharmacist-led intervention 4 events of 18 total; usual care 2 of 19; weight 1.7 percent; risk ratio 2.11 with 95 percent confidence interval 0.44 to 10.15. Tommelein2013: 15 of 170; 9 of 147; weight 6.5 percent; risk ratio 1.44 with 95 percent confidence interval 0.65 to 3.20. Xin2016: 81 of 114; 59 of 113; weight 91.8 percent; risk ratio 1.36 with 95 percent confidence interval 1.10 to 1.68. Total (95 percent confidence interval): pharmacist-led intervention total 302; usual care total 279; weight 100.0 percent; pooled risk ratio 1.38 with 95 percent confidence interval 1.12 to 1.69. Total pharmacist-led intervention 100; usual care 70. Graph: x-axis label Risk Ratio with Mantel–Haenszel, Random, 95 percent confidence interval, unitless, ranging 0.01, 0.1, 1, 10, 100. y-axis label Study or Subgroup, unitless, listing Khdour2009, Tommelein2013, Xin2016 and Total (95 percent confidence interval). A vertical reference line at 1. Individual study markers align at 2.11, 1.44 and 1.36 with horizontal confidence intervals spanning 0.44 to 10.15, 0.65 to 3.20 and 1.10 to 1.68. The pooled diamond spans 1.12 to 1.69 centered at 1.38. Bottom annotations read Favours [Pharmacist-led intervention] on the left and Favours [Usual care] on the right. The detailed data points are as - For Khdour2009, pharmacist-led intervention events were 4, pharmacist-led intervention total was 18, usual care events were 2, usual care total was 19, weight was 1.7 percent and the risk ratio (M–H, random, 95 percent CI) was 2.11 with 0.44 to 10.15. - For Tommelein2013, pharmacist-led intervention events were 15, pharmacist-led intervention total was 170, usual care events were 9, usual care total was 147, weight was 6.5 percent and the risk ratio (M–H, random, 95 percent CI) was 1.44 with 0.65 to 3.20. - For Xin2016, pharmacist-led intervention events were 81, pharmacist-led intervention total was 114, usual care events were 59, usual care total was 113, weight was 91.8 percent and the risk ratio (M–H, random, 95 percent CI) was 1.36 with 1.10 to 1.68. - For Total (95 percent CI), pharmacist-led intervention total was 302, usual care total was 279, weight was 100.0 percent and the risk ratio (M–H, random, 95 percent CI) was 1.38 with 1.12 to 1.69. - For Total events, pharmacist-led intervention total events were 100 and usual care total events were 70.A forest plot of smoking cessation risk ratios showing an overall effect favoring pharmacist-led intervention.Abbreviations: CI, confidence interval; df, degrees of freedom; M–H, Mantel–Haenszel; RR, risk ratio.
The interventions had no significant effect on Body Mass Index (BMI) at follow-up (Khdour 2009,22 p =0.09; Jarab 2012,23 p =0.61). The effects of pharmaceutical care on smoking cessation and BMI of COPD patients were summarized in eTable 9.
In this updated systematic review and meta-analysis of 11 RCTs including 2313 patients with COPD, pharmacist-led integrated interventions were associated with improvements in selected domains of COPD management, although effects varied across outcomes. Among the evaluated outcomes, reductions in exacerbation-related hospital admissions were observed across several trials with low statistical heterogeneity; however, these findings should be interpreted in light of variability in study design and healthcare context. Improvements were observed in medication adherence and quality of life in several trials, and pooled analyses suggested higher smoking cessation rates; however, these effects varied across outcomes and studies.
The reduction in exacerbation-related hospitalization is particularly important because AECOPD is a major driver of disease progression, impaired functional status, and healthcare costs in COPD.34 Pharmacist-led interventions are likely to reduce exacerbations through several complementary mechanisms. First, inhaler technique training improves effective drug delivery, which is essential for realizing the benefits of inhaled maintenance therapy.35,36 Second, adherence counseling and reinforcement may increase persistence with long-term pharmacotherapy and reduce treatment gaps.37 Third, pharmacists can identify and address medication-related problems in the context of polypharmacy, potentially reducing adverse drug reactions and optimizing regimens.38,39 Finally, lifestyle counseling and follow-up support may facilitate smoking cessation and self-management behaviors, further mitigating exacerbation risk.40 Together, these pathways may help explain the observed associations with utilization and selected patient-centered outcomes. A recent scoping review of pharmacist-physician collaborative models in COPD similarly highlighted substantial variability in intervention components, practice settings, and reported outcomes, underscoring the structural heterogeneity within this field and the need for standardized, trial-based evaluation of clinical effectiveness.41
In contrast, we did not observe a statistically significant improvement in COPD-specific health status measured by CAT, and heterogeneity was considerable. This finding should be interpreted cautiously given the limited number of CAT-contributing trials and the substantial differences among them. In addition, CAT is influenced by baseline disease severity, comorbidities, and concurrent clinical management, which likely differed across settings.42 These considerations may explain the inconsistent CAT findings and underscore the need for harmonized outcome assessment and standardized follow-up time points in future trials.
Overall, the evidence regarding pharmacist-led interventions in COPD demonstrates both converging and diverging patterns across outcomes. Consistent benefits were observed for medication-related and behavioral outcomes, including medication adherence, inhaler technique, and smoking cessation, which were directionally favorable across most included trials. In contrast, effects on symptom burden and objective disease measures were inconsistent. Meta-analysis of CAT scores showed substantial heterogeneity, and lung function outcomes were reported in too few studies to permit reliable quantitative synthesis.
Importantly, the observed heterogeneity was high for key patient-reported outcomes, reflecting marked differences in intervention components, intensity, follow-up duration, outcome measurement tools, and baseline patient characteristics across studies. These factors substantially limit the interpretability and generalizability of pooled estimates and suggest that pharmacist-led interventions may not uniformly translate into improvements in clinical symptoms or physiological measures. Taken together, these findings indicate that while pharmacist-led care may improve selected process-related and patient-centered outcomes, its impact on core disease outcomes remains uncertain. The present review therefore does not establish robust clinical effectiveness, but rather highlights variability in effects and identifies areas where evidence is inconsistent or limited.
Our findings are consistent with prior reviews that highlighted benefits of pharmacist involvement in COPD, particularly for medication-related behaviors and inhaler skills.15–17,43,44 However, earlier syntheses were either outdated16 or focused on selected outcomes.17 By integrating more recent RCT evidence and evaluating multiple clinically meaningful endpoints, this study provides updated quantitative synthesis of available RCT evidence for pharmacist-led integrated care.
Pharmacists contribute to COPD management through multi-dimensional, medication-focused services that are relatively consistent across settings, including inhaler technique training, adherence support, and medication review. The reproducibility of these core components suggests potential feasibility of integrating pharmacists into routine COPD care, although contextual adaptation would be required. This perspective also aligns with the growing recognition of pharmacists’ expanded role in chronic disease management.45
Several limitations warrant consideration. Many included trials were single-center and sample sizes varied substantially, including some small studies, which may limit generalizability. Blinding was frequently not feasible due to the behavioral nature of the interventions, introducing potential performance and detection bias, particularly for patient-reported outcomes. Intervention components, intensity, and follow-up durations varied widely, which likely contributed to heterogeneity for some endpoints. In addition, for several outcomes the number of available studies was limited, reducing precision and precluding robust exploration of effect modifiers and publication bias for specific endpoints. Accordingly, limitations related to small sample sizes, high risk of bias, lack of blinding, and substantial between-study heterogeneity should be considered integral to the interpretation of the findings, rather than ancillary concerns. Interpretation of pooled estimates for outcomes with very small numbers of studies warrants particular caution. Meta-analyses based on as few as three trials provide limited precision, and estimates of between-study heterogeneity may be unstable under such conditions. In such cases, pooled results should be interpreted with caution and should not be considered confirmatory evidence of treatment effect.
Future research should prioritize adequately powered, multi-center RCTs with standardized intervention frameworks and core outcome sets, including exacerbation-related utilization, QoL, and validated measures of adherence and inhaler technique assessed at harmonized time points. Technology-enabled approaches, such as smart sensors and digital monitoring of inhaler use, may provide objective and time-stamped measures of medication-taking and technique, enabling personalized feedback and reducing reliance on self-report.46,47 Trials should also identify subgroups most likely to benefit,48,49 such as patients with frequent exacerbations, poor adherence, high inhaler error rates, or complex comorbidity, and incorporate implementation and economic evaluations to inform scalable integration of pharmacists into COPD care pathways.
Pharmacist-led interventions in COPD were associated with reductions in exacerbation-related hospital admissions and improvements in medication adherence, with pooled analyses suggesting higher smoking cessation rates. However, these findings are derived from a limited number of studies, and the overall evidence base remains heterogeneous with variable methodological quality. Effects on symptom burden and objective disease measures were inconsistent or non-significant. Therefore, the results should be interpreted cautiously and do not establish robust or generalizable clinical effectiveness. Moreover, as most included trials were conducted in Asia and other low- to middle-income healthcare settings, differences in pharmacy practice scope and care pathways may further limit transferability to high-income systems.