Inhaled Medication Errors During Hospitalization or on Hospital Discharge in Patients Living With Chronic Obstructive Pulmonary Disease: A Literature Review

Inhaled medications, including beta-agonists, muscarinic antagonists, and corticosteroids, are the backbone of chronic obstructive pulmonary disease (COPD) treatment, and pharmacotherapy plans are frequently optimized during and following hospitalization. Clinical practice guidelines acknowledge that patients living with COPD may experience medication errors from inadequate inhaler technique or device faults, but inhaled medication errors within COPD pharmacotherapy plans remain unreported. This literature review aimed to collect and present studies describing medication errors occurring with inhaled medications in patients living with COPD during and following hospitalization. The databases searched included Ovid MEDLINE, Embase, and International Pharmaceutical Abstracts. One hundred forty-five unique studies were collected, and 10 studies were included. The rate of inhaled medication errors reported across the 10 studies ranged between 2.5% and 66% of patients living with COPD and who were hospitalized or discharged. The incidence and types of medication errors reported across the studies varied significantly. Standardization in categorizing and reporting inhaled medication errors is necessary for future studies to determine the true incidence of inhaled medication errors occurring in patients living with COPD who are hospitalized or discharged.


Introduction And Background
Patients living with chronic obstructive pulmonary disease (COPD) may take medications to palliate symptoms, improve health status, tolerate exercise, and reduce the risk and severity of exacerbations of COPD (ECOPD) [1].Inhaled bronchodilators, including short-and long-acting β agonists (SABAs and LABAs, respectively) and short-and long-acting muscarinic antagonists (SAMAs and LAMAs, respectively), represent the backbone of therapy to treat chronic COPD [1].An inhaled corticosteroid (ICS) may also be added to the pharmacotherapy plan if the patient living with COPD possesses eosinophilia or comorbid asthma [1].The 2023 Global Initiative for Obstructive Lung Disease (GOLD) Report lists 38 medications across five distinct pharmacologic classes to treat COPD [1].Fifteen of 38 medications include combinations of inhaled medications (ie, SABA/SAMA, LABA/LAMA, LABA/ICS, and LABA/LAMA/ICS).Single-agent and combination products may be formulated as dry powder, soft mist, or metered dose inhalers or administered via nebulizer [1].The availability of several devices and the combination of products are advantageous for patients and enable them to use a product that is convenient, affordable, and efficacious.However, the comprehensive formulary of inhaled medications may be overwhelming for both patients and prescribers and thus there is potential for medication errors.Hospitalization is a critical time wherein inhaled medication errors may occur and between 7% and 33% of patients living with COPD will experience at least one COPD-related hospitalization annually [2].Patients hospitalized with ECOPD often require pharmacotherapy intensification, such as adding another inhaled bronchodilator (ie, LABA plus LAMA) or inhaled corticosteroid (ie, ICS plus LABA/LAMA), to better manage symptoms and reduce the risk for subsequent ECOPD [1].For example, duplicate therapies may be erroneously prescribed (i.e., SABA via metered dose inhaler plus SABA via nebulizer).The 2023 GOLD Report acknowledges that medication errors occur, albeit as errors owing to poor inhaler technique or device delivery faults, not necessarily medication errors within the inhaled pharmacotherapy plan itself [1].Medication errors are defined as mistakes occurring throughout the stages of the medication-use process including medication selection, procurement, storage, ordering, transcribing, preparing, dispensing, administering, and monitoring [3].Thus, medication errors may encompass either patient or healthcare errors depending on the phase of the medication-use process [3].Medication errors that occur due to inappropriate prescribing or during the transition of care from hospital to non-hospital phases of care are unacknowledged within the 2023 GOLD Report.
The purpose of this review is to collect and present literature describing medication errors occurring with inhaled medications in patients living with COPD during and following hospitalization.

Review Review methods
This was a literature review utilizing the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) protocol [4].The research question did not satisfy the technical requirements for a systematic review (to answer a specific patient-intervention-comparison-outcome (PICO) question) or a scoping review (to inform policymakers or posit gaps in practice, research, or policymaking), thus a literature review adhering to PRISMA criteria was employed [4].The research protocol was not registered because literature reviews are ineligible for registration in the International Prospective Register of Systematic Reviews (PROSPERO).
An electronic literature search was performed using Ovid MEDLINE, Embase, and International Pharmaceutical Abstracts on February 14, 2023.Search strings for each database are included in Appendix A and adhered to the PRISMA-S extension for literature search reporting [5].Study registries were not searched, print resources were not browsed, and in-article citation searching was not performed to collect records beyond the database searches.Articles published between 1946 and 2022 were eligible for inclusion.Exact keywords and filters varied depending on the database searched but consistent terms across the searches were COPD, medication error, and discharge.Studies were eligible for inclusion if they were published in English and the population studied included patients living with COPD, were taking or prescribed inhaled medication(s) to treat COPD, and were hospitalized or recently discharged from the hospital.We did not aim to capture, characterize, and assess errors in inhaler administration because inhaler technique errors are already and clearly described in the GOLD Report [1].
Results from each database search were exported as a Research Information Systems file and then imported into Covidence (Covidence, Melbourne, AU).During the initial file upload, Covidence automatically deduplicated identical records retrieved across databases.Abstracts were independently screened by two authors (KH and MG) and screening conflicts were resolved by a third author (PB).Full texts were independently screened by two authors (KH and MG) and screening conflicts were resolved by a third author (PB).Data abstraction was independently performed by two authors (KH and MG).To minimize heterogeneity across the articles and among authors, extraction was performed using a template in Covidence.Data collected included study location, study design, characteristics of the cohort (of all subjects included in the research), characteristics of patients living with COPD (i.e., a sub-group if the research reported on multiple acute or chronic diseases beyond COPD), and medication errors.There are many frameworks for reporting medication errors, medication error research, and outcomes attributable to medication errors [3].Because of this literature review design and the expected heterogeneity across included studies, we extracted medication errors as reported in each of the included studies using counts and frequencies, and utilized narrative synthesis to briefly summarize prescient outcomes in this manuscript [4].Once charting was completed, the authors compared their findings with their peers to determine agreements and conflicts, the latter of which was reconciled through discussion among all three investigators (KH, MG, and PB).Covidence calculated a Cohen's kappa (K) coefficient statistic to measure the interrater reliability of the abstracts screened and full texts reviewed [6].The primary author (PB) reviewed and adjudicated at each step of the review process and met with the research team in person for one hour each week (over 15 weeks) to provide direction and answer queries.
Figure 1 depicts the flow diagram for the search, screen, assessment, and inclusion of studies.One hundred forty-five unique articles were identified across the database searches and their abstracts were screened for eligibility.Thirty abstracts were deemed eligible and sought for full-text appraisal.Cohen's kappa coefficient statistic for interrater reliability was 0.46 and 0.62 for the abstract and full-text screening, respectively [6].

Medication-related problems
In a pilot study conducted at an academic medical center, pharmacists performing admission and discharge medication reconciliation discovered a medication-related problem (MRP) in 96% of patients during admission and discharge transitions of care, with an average of six MRPs encountered per patient [12].These MRPs included omissions, unnecessary or inappropriate medications, duplicate therapy, and orders written with the wrong dose, frequency, or directions [12].Thirty-six percent of patients in this pilot study were admitted with COPD as the reason for hospitalization but specific MRPs attributable to COPD hospitalizations were not provided [12].In a prospective study conducted in Spain at a large integrated healthcare system, 25% of all MRPs were associated with inhaled medications and were deemed as being related to need, effectiveness, and safety [15].This report did not, however, enumerate the counts and frequencies of need, effectiveness, and safety measures specific to inhaled medications [15].A North Carolina community hospital conducted a retrospective chart review on MRPs in which six MRPs (5.9%) with ICS were all determined to be potentially preventable [10].In a prospective cohort study of 25 patients at a teaching hospital, six MRPs were identified across four readmitted patients [11].Among them, one MRP (16.7% of all MRPs) was a transcribing error involving fluticasone/salmeterol, wherein the inhaler was omitted from one patient's pharmacotherapy plan, and in two other patients, prescribing errors occurred (i.e. one in each patient (33.4% of all MRPs)) when the fluticasone dose was inappropriately increased to 500 mcg twice daily [11].
One Veterans Affairs health care center conducted a pre-post study and found an absolute decrease of greater than 10% in prescribing errors upon discharge in each of the categories of SABA, LABA/LAMA, and ICS upon implementing an order set for ECOPD [8].Decreases in the incidence of inhaled medication errors post-implementation were statistically significant (p<.05) versus pre-implementation for SABAs (2.5% vs 13.9%) and LABAs/LAMAs (7.4% vs 16.5%) but not ICS (9.9% vs 18%) [8].A New Jersey community hospital conducted a retrospective cohort study and identified medication errors at discharge caused by dose or frequency errors decreased after having a pharmacist conduct a medication reconciliation at the time of discharge (14 vs 8, pre-post pharmacist intervention) [16].Omissions of SABAs, LABAs, and LAMAs on discharge were unchanged (11 vs 11, pre-post pharmacist intervention) [16].

Prescribing errors and omissions
One retrospective cohort study of older adults hospitalized in Palestine leveraged the Screening Tool of Older Persons' Prescriptions and Screening Tool to Alert to Right Treatment (STOPP/START) to determine potentially inappropriate prescribing via inappropriate medications and omissions [7].The authors identified 221 discrete instances of potentially inappropriate prescribing, of which seven were in patients living with COPD who were prescribed an oral corticosteroid instead of an ICS [7].Out of 247 patients in the sample, approximately 5% of all potential prescribing omissions occurred in patients with respiratory diseases such as COPD [7].
In one multisite retrospective cohort study conducted in Australia, 93.2% of hospitalized COPD patients received nebulized ipratropium, with 48% of these patients deemed as having received inappropriate nebulized ipratropium because they were ordered (duplicative) inhaled ipratropium concomitantly [14].At discharge, 45% of patients were inappropriately ordered nebulized ipratropium when inhaled ipratropium (i.e.hydrofluoroalkane (HFA) device) should have been prescribed [14].More than 30% of patients ordered ipratropium on discharge were prescribed a supply for longer than clinically necessary [14].In one retrospective chart review of inpatients in Canada, 17.5% of eligible patients were not prescribed oral or inhaled steroid therapy upon discharge [9].Prescribing omissions related to ICS, specifically, were not provided in this report and it is plausible that fewer than 17.5% of the eligible patients (not prescribed oral and inhaled steroids) had an ICS omitted from their treatment plan [9].This study also reported that as many as 78% of patients admitted with ECOPD were prescribed an inhaler but may not have been taught or assessed on appropriate inhaler techniques [9].One project capturing information from the Ontario Drug Benefit database conducted a population-based retrospective cohort study on medication discontinuation in patients living with COPD and found hospitalization was associated with a 50-62% increased risk of inhaled medications, specifically LAMA monotherapy and ICS/LABA combination therapy, being unintentionally discontinued compared to patients who were not hospitalized [13].

Discussion
The rate of inhaled medication errors was heterogeneous across the 10 studies in this review [7][8][9][10][11][12][13][14][15][16].Brown and colleagues reduced the rate of SABA errors to as low as 2.5% after implementing an order set at discharge following ECOPD [8].Peterson and colleagues determined that 45% of patients erroneously received nebulized ipratropium on hospital discharge [14] and between 40% and 66% of patients experienced errors via medication reconciliation in the study by Singh and colleagues [16].Based on the studies included in this review, the rate of inhaled medication errors was heterogeneous and could lie anywhere between approximately 2% and 70% [7][8][9][10][11][12][13][14][15][16].Future research, including population health analytics and case-control analyses, is needed to determine more homogenous estimates of medication errors.
Studies conducted using administrative datasets, such as the project by Gershon and colleagues [13], may posit finite, accurate, and reasonable estimates of inhaled medication error rates, and similar international investigations should be undertaken.Five studies in this review highlighted errors occurring with an ICS [7][8][9]11,13].Not all patients living with COPD are eligible for an ICS [1].Patients who are most likely to benefit include those with comorbid asthma (or a history of asthma), frequent ECOPD, or eosinophilia [1].Patients living with COPD who are inappropriately prescribed an ICS are at increased risk for adverse drug reactions such as oropharyngeal candidiasis and carry increased risks for pneumonia [1].The 2023 GOLD Report provides ample guidance via narrative explanations, key summaries, and flowcharts to help prescribers choose the correct inhaled medications to treat COPD [1].A plausible reason for inappropriate ICS prescribing could be prescriber familiarity with and the commercial availability of ICS/LABA combination therapies, which are recommended for all steps of asthma treatment [17,18].The ICS/LABA combination inhalers were commercially available for at least five years before LAMAs and LABA/LAMAs entered the market [1].Compared to North America, short-acting inhalers and oral bronchodilators are prescribed more often internationally than long-acting inhalers because short-acting and oral bronchodilators are substantially cheaper than long-acting inhalers [1].Mitigation strategies to prevent ICS errors in patients living with COPD, such as electronic order sets, interdisciplinary rounds, and pharmacist-led medication reconciliation, have both nominally and statistically reduced ICS errors as reported by four studies in this review [8,11,12,16].Three studies in this review reported inhaled medication errors as they pertained to MRPs [10,12,15].The Pharmacy Quality Alliance (PQA) MRP Framework is an evidence-based tool and is familiar to many clinicians [19].Only the study by Cooper and colleagues used a validated framework (ie, PQA) to categorize MRPs [10].To improve medication error and intervention reporting consistency and transparency, future studies should strive to use validated and recognized frameworks.Beyond the PQA MRP, an additional instrument includes the National Coordinating Council for Medication Error Reporting and Prevention (NCC MERP) index [20].The NCC MERP index is an easy-to-use tool that categorizes errors and identifies the degree of harm to the patient.By adhering to PQA MRP or NCC MERP, future studies will be standardized, and their results may be more accurately compared or aggregated in future analyses.
Treatment for COPD has drastically evolved over the last 60 years [21].Inhaled bronchodilators and corticosteroids first gained attention as therapeutic options in the 1990s and these inhaled medications replaced existing systemic treatments including antibiotics, sympathomimetics, and methylxanthines [21].
The GOLD group and initiative was formally established in 2001, and their recommendations supporting or refuting inhaled corticosteroids have changed often over the last 20 years as our knowledge of COPD pathophysiology has improved [1,21].This may explain, to some extent, the heterogeneity of inhaled medications used to treat COPD across the 10 studies included in our review [7][8][9][10][11][12][13][14][15][16].To improve transparency, further research should reference the GOLD Report or another clinical practice guideline with current recommendations (at the time the study was conducted) to provide context on inhaled medication appropriateness during the study period.
This literature review is not without limitations.Although the authors utilized a systematic approach to searching, assessing, and presenting the included studies, this literature review did not meet the technical standards of either a systematic or scoping review.Future systematic reviews on the themes and topics raised within this literature review are necessary to answer patient-specific questions.The authors did not perform a bias assessment of each article and doing so would have qualified the methodologic rigor of the included studies; although six studies were retrospective designs [7,9,10,13,14,16] and none were randomized controlled trials, it may be reasonable to infer that the included studies were of low-tomoderate methodologic quality [7][8][9][10][11][12][13][14][15][16].Spirometry is required to accurately diagnose COPD [1]; none of the 10 studies in this review confirmed COPD using spirometry and thus there may be bias toward diagnostic momentum or misclassification [22,23].All 10 articles included were published in or after 2003 [7][8][9][10][11][12][13][14][15][16], despite the search parameters permitting articles published as early as 1946.This is possible because (the term) COPD was coined in the early 2000s and it replaced previous terms of this respiratory disorder: chronic bronchitis and/or emphysema [21].Although the latter two terms are currently included as sub-categories under the COPD medical subject heading, it is possible that the databases searched may index and misattribute studies between 1946 and 2000 using retired terms.A subsequent search of the literature using all possible keywords connoting COPD is necessary.The Cohen's kappa coefficients calculated during the abstract (K=0.46) and full-text screening (K=0.62) are indicative of weak and moderate interrater reliability, respectively [6,23,24].However, the authors ultimately resolved disagreements in screening, inclusion, and data abstraction via discussion and consensus-building in order to ensure the correctness of the data presented.

Conclusions
As few as 2.5% or as many as 66% of patients living with COPD who are hospitalized may experience an error in their inhaled medication pharmacotherapy plan.Pharmacologic classes with reported errors include short-and long-acting beta-agonists, muscarinic antagonists, and inhaled corticosteroids.Standardization in categorizing and reporting inhaled medication errors is necessary in subsequent studies; the Pharmacy Quality Alliance Frameworks or Medication Error Reporting and Prevention Index are validated and widely used instruments that may be considered in future studies.Future research should aim to clearly compare and contrast findings across clinical settings and investigations and determine the true incidence of inhaled medication errors.