Effectiveness of a standardized electronic admission order set for acute exacerbation of chronic obstructive pulmonary disease

Effectiveness of a standardized electronic admission order set for acute exacerbation of chronic... Background: Variation in hospital management of patients with acute exacerbation of chronic obstructive pulmonary disease (AECOPD) may prolong length of stay, increasing the risk of hospital-acquired complications and worsening quality of life. We sought to determine whether an evidence-based computerized AECOPD admission order set could improve quality and reduce length of stay. Methods: The order set was designed by a provincial COPD working group and implemented voluntarily among three physician groups in a Canadian tertiary-care teaching hospital. The primary outcome was length of stay for patients admitted during order set implementation period, compared to the previous 12 months. Secondary outcomes included length of stay of patients admitted with and without order set after implementation, all-cause readmissions, and emergency department visits. Results: There were 556 admissions prior to and 857 admissions after order set implementation, for which the order set was used in 47%. There was no difference in overall length of stay after implementation (median 6.37 days (95% confidence interval 5.94, 6.81) pre-implementation vs. 6.02 days (95% confidence interval 5.59, 6.46) post-implementation, p = 0.26). In the post-implementation period, order set use was associated with a 1.15-day reduction in length of stay (95% confidence interval − 0.5, − 1.81, p = 0.001) compared to patients admitted without the order set. There was no difference in readmissions. Conclusions: Use of a computerized guidelines-based admission order set for COPD exacerbations reduced hospital length of stay without increasing readmissions. Interventions to increase order set use could lead to greater improvements in length of stay and quality of care. Keywords: Length of stay, Clinical decision support, Chronic obstructive pulmonary disease, Quality improvement Background for AECOPD cost approximately USD $3.8 billion [4]and Chronic obstructive pulmonary disease (COPD) is a com- account for 51% of overall expenditures for COPD. [5]Pro- mon and progressive lung disease that is characterized by longed hospital length of stay (LOS) also have a negative shortness of breath, activity limitation, and a predisposition impact on patient function and quality of life. [6] to exacerbations. Acute exacerbations of COPD (AECOPD) Evidence-based management guidelines for AECOPD adversely affect quality of life, [1] increase the risk of disease have been developed, [7, 8] and include recommenda- progression, [2] and reduce survival. [3] Hospitalizations tions regarding pharmacotherapy and post-exacerbation care. Despite these guidelines, hospital care of patients * Correspondence: sachin.pendharkar@ucalgary.ca with AECOPD remains highly variable. [9] This variation Department of Medicine, Cumming School of Medicine, University of may contribute to prolonged LOS [10] that, in turn, in- Calgary, Calgary, AB, Canada creases the risk of hospital-acquired complications and Department of Community Health Sciences, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada adversely impacts quality of life. [6, 11] Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Pendharkar et al. BMC Pulmonary Medicine (2018) 18:93 Page 2 of 8 Order sets are grouped medical orders intended to criteria. Additional details on the methods are provided in standardize evidence-based best practice. Computerized an additional file (see Additional file 1). Physician Order Entry (CPOE) systems may improve work- flow, promote appropriate testing and treatment, reduce er- Order set development rors and improve guideline adherence, [12–17] particularly The AECOPD order set was based on published COPD when integrated into general order sets. [18]Standardized guidelines, [7] and developed by a provincial COPD admission order sets have been used in other diseases with working group comprised of physicians, nurses, and re- variable success at reducing hospital LOS. [14, 15] spiratory therapists from a variety of clinical settings, in Two observational studies have demonstrated that a series of face-to-face and teleconference meetings. order sets likely improve the quality of hospital care for The order set contained recommended testing, medi- patients with AECOPD and reduce LOS. [13, 16] How- cation (including suggested dosing and mode of deliv- ever, these studies used pre-post designs that could be ery), consultations, and a priori discharge planning influenced by secular trends in AECOPD management, interventions specific to patients with AECOPD. Some and the studies did not account for the differential interventions were pre-selected to encourage use (e.g., effects of the order set among physician groups. physiotherapy referral). The order set was built into the The objective of this study was to determine whether hospital’s existing CPOE system, Sunrise Clinical the implementation of an evidence-based computerized Manager (Allscripts Solutions, Chicago IL). Screenshots admission order set would improve the quality of in- are provided in an additional file (see Additional file 2). patient AECOPD care. A stepped wedge design was used to account for differential effects among physician Implementation groups and to minimize confounding related to the tim- The order set was implemented using a stepped wedge ing of order set implementation. Our hypothesis was design [20] among the three physician groups who admit that the implementation of a standardized order set patients with AECOPD: respirologists, general internists, would reduce hospital LOS of patients admitted for and family physician hospitalists. It was implemented se- AECOPD without increasing emergency department quentially within physician groups with each group act- (ED) or hospital readmissions. Preliminary study results ing as its own control. Study outcome data were have previously been reported in abstract form. [19] collected at baseline and at each implementation ‘step’. Implementation among respirologists, general inter- nists and hospitalists occurred in March, May and Au- Methods gust 2013, respectively. Prior to each implementation Study design step, the research team met with physicians and allied This study is an analysis of administrative health data for a health staff to introduce the order set. Order set use by quality improvement project in which an electronic stan- each individual physician was voluntary. Monthly statis- dardized admission order set for patients with AECOPD tics on order set use were posted in clinical areas. was implemented at a large, tertiary-care teaching hospital in Calgary, Alberta between March 1, 2013 and March 31, Analysis 2015. Since this was a quality improvement project, the Patient demographic, comorbidity and hospitalization University of CalgaryConjointHealthResearchEthics data were obtained from provincial administrative data Board waived the requirement for formal ethics approval. and linked to order set usage data from the CPOE sys- tem using the patient’s provincial health number. [21] Study population The primary outcome was hospital LOS for patients Patients were included if they were: older than 45 years of admitted during the implementation period compared to age; admitted to hospital between March 1, 2013 to March those admitted during the previous 12 months (pre-post 31, 2015 with an International Classification of Diseases, implementation analysis). Secondary outcomes included: Tenth Revision (ICD-10-CA) code indicative of AECOPD hospital LOS of patients admitted with and without the (J42 [unspecified chronic bronchitis], J43 [emphysema], or order set after implementation (post-implementation J44 [other chronic obstructive pulmonary disease]) in the analysis); all-cause readmissions at 7, 30 and 90 days after primary diagnosis field of the hospital discharge abstract discharge; ED visits at 7 and 30 days; and in-hospital database; and admitted to the pulmonary, general internal mortality. medicine or hospitalist clinical services. Patients were ex- Unadjusted and adjusted median regression models cluded if they were admitted to the intensive care unit or were constructed to assess the impact of the order set on any other clinical service. Historical controls from the LOS. [22–24] Covariates in adjusted models included age, 12 months prior to order set implementation in each sex, and five clinically relevant comorbidities (heart failure, group of ordering physicians were identified using similar dementia, liver disease, renal disease, and diabetes) that Pendharkar et al. BMC Pulmonary Medicine (2018) 18:93 Page 3 of 8 were strongly associated with the Charlson Comorbidity the pre-implementation period (64.5% vs. 74.3%). Patients Index (Somers’ D = 0.94). [25] Logistic regression was with co-existing heart failure and diabetes were more com- used to adjust 30-day readmission odds ratios for age, sex, monly admitted under general internists. Over 95% of pa- comorbidity and admitting physician specialty. tients were discharged home. All analyses were performed using SAS version 9.3 (Cary, NC) or R version 3.2.3; [26] p < 0.05 was consid- Order set uptake ered statistically significant. In the post-implementation period, 57% of patients admitted to the hospitalist service were admitted using Results the order set, compared to 30% of patients admitted by Of 1435 AECOPD admissions to one of the three physician general internists or respirologists. Time series analysis groups during the study period, 1413 with a LOS less than revealed that order set use increased gradually after 90 days were included in the analysis (Fig. 1). There were implementation, mostly by general internists and hospi- 857 admissions after order set implementation, of which talists (Additional file 3: Figure S1). 406 patients (47%) were admitted using the order set. Baseline characteristics of study participants are pre- Hospital LOS sented in Tables 1 and 2 for the pre-post and Figure 2 shows the unadjusted and adjusted differences post-implementation analyses, respectively. The hospitalist in median LOS for patients treated in the pre- and service admitted most patients with AECOPD, but admit- post-implementation periods. Median LOS was 6.37 days ted fewer in the post-implementation period compared to (95% confidence interval [CI] 5.94, 6.81; n = 556) for patients admitted before implementation and 6.02 days (95% CI 5.59, 6.46; n = 857) for patients admitted after implementation (p = 0.26). Unadjusted and adjusted comparisons of median LOS in the post-implementation analysis are presented in Fig. 3. Order set use was associated with a 1.15-day (95% CI -0.50, − 1.81) shorter median LOS, due primarily to a 1.8-day (95% CI -0.95, − 2.61) decrease for the hospitalist group (Fig. 3). Median LOS for patients admitted by general inter- nists or respirologists did not differ by order set use. Readmissions Neither order set implementation, nor order set use in the post-implementation period were associated with changes in readmissions or ED visits for all three phys- ician groups (Tables 3 and 4 and Additional file 1 Table S1). Overall in-hospital mortality did not change with order set implementation, but was lower in the hospital- ist group with order set use. Discussion This is thelargest studytoevaluatethe impact of a stan- dardized, guideline-based electronic order set for AECOPD on hospital LOS. We performed two analyses: a compari- son of LOS before and after the order set was imple- mented, and a comparison of patients admitted with and without the order set after implementation. The results re- vealed that order set implementation did not result in an overall LOS reduction, perhaps because only 47% of admit- ting physicians used it. However, the post-implementation analysis revealed that when it was used, the order set was associated with a LOS reduction of 1.15 days. Use of the order set by hospitalists, who admitted 65% of AECOPD Fig. 1 Patient flow diagram. AECOPD – acute exacerbation of patients in the post-implementation cohort, resulted in the chronic obstructive pulmonary disease; LOS – length of stay largest LOS reduction of 1.8 days. Importantly, there was Pendharkar et al. BMC Pulmonary Medicine (2018) 18:93 Page 4 of 8 Table 1 Baseline characteristics for pre-post implementation analysis Characteristics Total Pre-implementation Post-implementation p-value Number of patients 1413 556 857 Mean age, years (SD) 70 (12) 70 (12) 70 (12) 0.747 Age group, n (%) < 55 129 (9.1) 59 (10.6) 70 (8.2) 0.250 55–64 323 (22.9) 128 (23.0) 195 (22.8) 65–74 428 (30.3) 157 (28.2) 271 (31.6) 75–84 360 (25.5) 136 (24.5) 224 (26.1) 85+ 173 (12.2) 76 (13.7) 97 (11.3) Male sex, n (%) 727 (51.5) 279 (50.2) 448 (52.3) 0.441 Comorbidity, n (%) Heart failure 190 (13.5) 71 (12.8) 119 (13.9) 0.548 Dementia 45 (3.2) 24 (4.3) 21 (2.5) 0.051 Diabetes 318 (22.5) 131 (23.6) 187 (21.8) 0.444 Renal disease 36 (2.6) 13 (2.3) 23 (2.7) 0.687 Liver disease 13 (0.9) 6 (1.1) 7 (0.8) 0.614 Admitting specialty, n (%) Respirologist 148 (10.5) 51 (9.2) 97 (11.3) 0.0005 General internist 299 (21.2) 92 (16.6) 207 (24.2) Hospitalist 966 (68.4) 413 (74.3) 553 (64.5) SD = standard deviation no increase in either ED or hospital readmissions, suggest- hospitalized with AECOPD, Kitchlu et al. showed that ing that earlier discharge resulting from order set use did implementation of an order set improved the quality not occur at the expense of harm to the patient. of admission orders using pre-specified measures of Our findings extend observations from two recent guidelines-based care. [16] Using a similar design in a studies examining the impact of order sets on study of 275 patients, Brown et al. showed that phys- AECOPD care. In a pre-post design of 243 patients ician prescribing practices for AECOPD could be Table 2 Baseline characteristics for post-implementation analysis Characteristics Respirologist General internist Hospitalist No order set (n = 64) Order set (n = 33) No order set (n = 148) Order set (n = 59) No order set (n = 239) Order set (n = 314) Mean age, years (SD) 63 (11) 64 (10) 69 (11) 71 (12) 72 (12) 71 (10) Age group, n (%) < 55 8 (13) 6 (18) 12 (8) 8 (14) 15 (6) 21 (7) 55–64 27 (42) 10 (30) 37 (25) 14 (24) 39 (16) 68 (22) 65–74 20 (31) 9 (27) 50 (34) 14 (24) 72 (30) 106 (34) 75–84 9 (14) 8 (24) 36 (24) 18 (31) 65 (27) 88 (28) 85+ 0 0 13 (9) 5 (9) 48 (20) 31 (10) Male sex, n (%) 34 (53) 12 (36) 75 (51) 29 (49) 134 (56) 159 (51) Comorbidity, n (%) Heart failure 4 (6) 2 (6) 39 (26) 11 (19) 37 (15) 26 (8) Dementia 0 1 (3) 4 (3) 0 11 (5) 5 (2) Diabetes 9 (14) 8 (24) 46 (31) 16 (27) 45 (19) 63 (20) Renal disease 2 (3) 0 4 (3) 2 (3) 7 (3) 8 (3) Liver disease 0 0 1 (1) 0 2 (1) 2 (1) SD = standard deviation Pendharkar et al. BMC Pulmonary Medicine (2018) 18:93 Page 5 of 8 Fig. 2 Forest plot of implementation effects (pre-post implementation analysis). Adjusted model included age, sex, and five clinically relevant comorbidities selected from the Charlson Comorbidity Index (heart failure, dementia, mild or severe liver disease, renal disease, and diabetes. Pre – pre-implementation; Post – post-implementation; N – number of patients; Med – median length of stay; IQR – interquartile range; LOS – length of stay; CI – confidence interval improved with an electronic order set. [13]Secondary A standardized order set is appealing due to high vari- analyses in both studies revealed LOS reductions ability in inpatient AECOPD management. [9] Previous without an increase in readmissions. The current studies of order sets for AECOPD demonstrated more study extends this work by demonstrating an im- consistent use of systemic corticosteroids, appropriate provement in hospital LOS in a larger study cohort. antibiotics, and allied health providers such as physiother- Unlike the previous studies, which reported on apists, [13, 16] all of which have been shown to reduce pre-post effects of order set implementation, we hospital LOS. [27–29] The current order set similarly specifically examined the effect of actual use of the prompted admitting physicians to use these therapies, and order set, demonstrating that it could reduce LOS pre-selection of some items (e.g., bronchodilator delivery, compared to patients admitted without the order set. physiotherapy referral) provided additional clinical deci- Furthermore, the stepped wedge design is robust to sion support around guidelines-based care. secular trends in care delivery and LOS of AECOPD The LOS reduction observed after order set imple- patients, and allowed for subgroup analyses by differ- mentation was driven by improvements for patients ad- ent admitting physician groups. Both our study and mitted by hospitalists, with no differences observed in thepreviousstudies providea strong rationalefor the the other physician groups; this was an interesting find- standardization of inpatient AECOPD care using ing with many possible explanations. First, Sandhu et al. computerized order sets. demonstrated high variability in AECOPD management Fig. 3 Forest plot of the effects of the order set (post-implementation analysis). Adjusted model included age, sex, and five clinically relevant comorbidities selected from the Charlson Comorbidity Index (heart failure, dementia, mild or severe liver disease, renal disease, and diabetes. OS – order set; N – number of patients; Med – median length of stay; IQR – interquartile range; LOS – length of stay; CI – confidence interval; OS – order set Pendharkar et al. BMC Pulmonary Medicine (2018) 18:93 Page 6 of 8 Table 3 Readmissions, emergency department visits and in-hospital mortality for pre-post implementation analysis Results Pre-implementation (n = 556) Post-implementation (n = 857) p-value 7-day readmission, n (%) 33 (5.9) 60 (7.0) 0.430 30-day readmission, n, (%) 91 (16.4) 166 (19.4) 0.153 90-day readmission, n, (%) 170 (30.6) 299 (34.9) 0.093 7-day ED visits, n (%) 42 (7.6) 55 (6.4) 0.409 30-day ED visits, n (%) 124 (22.3) 196 (22.9) 0.803 In-hospital mortality, n, (%) 19 (3.4) 31 (3.6) 0.842 ED = emergency department among all specialties, with deviation from clinical guide- hospitalists, for a number of possible reasons. First, the lines occurring more often when care was provided by complexity of the patient’s presentation (e.g., respiratory physicians other than respirologists. [9] This variability failure requiring noninvasive ventilation) may have made may be due to the diversity of medical problems man- the order set less applicable at the time of admission. aged by hospitalists, which could lead to a less uniform Second, respirologists may have greater perceived approach to inpatient AECOPD management. Thus, it is self-efficacy with AECOPD management, leading them possible that the opportunity for standardization using to admit AECOPD patients without using an order set. an order set was greater for hospitalists than for other Finally, whereas the AECOPD order set was the only re- specialties. Second, heart failure and diabetes were more spiratory order set embedded in the CPOE system, sev- prevalent in patients admitted under general internists eral admission order sets for medical problems typically compared to hospitalist patients. These conditions have admitted under hospitalists (e.g., pneumonia, heart fail- both been associated with longer LOS in patients with ure) were already embedded. Thus, hospitalists may have AECOPD, [30] and are unlikely to be impacted by order more experience in order set use compared to other set use. Third, patients presenting with respiratory fail- physicians. Importantly, end users were consulted to en- ure requiring noninvasive ventilation were only admitted sure the order set was intuitive and minimally disruptive by general internists or respirologists; these indicators of to clinical workflow; these factors have been shown to more severe AECOPD were not systematically captured, increase uptake of clinical decision support systems such but could have reduced the effectiveness of the order as standardized order sets. [14, 17, 18] The increase in set. [31] Although the order set’s impact seemed to be order set use over time suggested that admitting physi- isolated to only one admitting group, the 1.8-day reduc- cians found it useful. tion in median LOS is an important finding since hospi- This study has a number of limitations. First, the talists were responsible for providing almost two thirds non-randomized study design raises the possibility that of inpatient AECOPD care in our study, and this is likely improvements in LOS were due to other differences be- to be similar in other large, tertiary-care urban hospitals tween groups admitted with and without the order set. in North America. However, when implementing complex healthcare inter- The order set was used by 47% of admitting physicians ventions such as the AECOPD order set, traditional ran- during the study period. This low uptake is a consistent domized controlled trials are impractical due to finding for voluntary order sets [13, 16, 32, 33] and is a logistical constraints and the risk of contamination known limitation of their use. [34] Respirologists and within clinical provider groups. The methodologically general internists used the order set less frequently than robust stepped wedge design minimized contamination Table 4 Readmissions, emergency department visits and in-hospital mortality for post-implementation analysis Results Respirologist General internist Hospitalist No order set Order set p value No order set Order set p value No order set Order set p value (n = 64) (n = 33) (n = 148) (n = 59) (n = 239) (n = 314) 7-day readmission, n (%) 4 (6.3) 2 (6.1) 0.971 18 (12.2) 4 (6.8) 0.257 13 (5.4) 19 (6.1) 0.760 30-day readmission, n (%) 19 (29.7) 5 (15.2) 0.116 29 (19.6) 14 (23.7) 0.508 41 (17.2) 58 (18.5) 0.689 90-day readmission, n (%) 30 (46.9) 13 (39.4) 0.482 59 (39.9) 19 (32.2) 0.305 75 (31.4) 103 (32.8) 0.723 7-day ED visits, n (%) 2 (3.1) 1 (3.0) 0.980 12 (8.1) 3 (5.1) 0.449 15 (6.3) 22 (7.0) 0.734 30-day ED visits, n (%) 18 (28.1) 5 (15.2) 0.155 28 (18.9) 14 (23.7) 0.437 54 (22.6) 77 (24.5) 0.597 In-hospital mortality, n (%) 0 1 (3.0) 0.162 8 (5.4) 5 (8.5) 0.411 12 (5) 5 (1.6) 0.021 ED = emergency department Pendharkar et al. BMC Pulmonary Medicine (2018) 18:93 Page 7 of 8 by allowing implementation and evaluation of the order Additional files set in clusters [35] while still analyzing the effect of Additional file 1: Supplement with additional details on methods and order sets within physician groups. results (DOCX 28 kb) Second, the use of administrative data prevented ana- Additional file 2: Screenshot of AECOPD order set (PDF 400 kb) lysis of patient characteristics that might have influenced Additional file 3: Figure S1. Monthly percentage of patients admitted a physician’s decision to use the order set; it is thus pos- using AECOPD order set during study period. Vertical lines represent sible that the reduced LOS in the post-implementation implementation start dates for each physician specialty. Respirologist represented by hatched line; general internist represented by black solid period is due to order set use in less complex cases. line; hospitalist represented by grey solid line. Reported probabilities are While our findings are consistent with studies per- for linear trends from time series models for each physician specialty. formed in different geographic and clinical settings, [13, These models showed no evidence of seasonality or auto-regression (JPG 112 kb) 16] we acknowledge the importance of future studies to examine whether COPD severity, presentation acuity, or Abbreviations use of specific interventions (e.g., noninvasive ventila- AECOPD: Acute exacerbation of chronic obstructive pulmonary disease; tion) impact order set use. Such an analysis would help CI: Confidence interval; COPD: Chronic obstructive pulmonary disease; to tailor strategies aimed at increasing order set uptake CPOE: Computerized physician order entry; ED: Emergency department; LOS: Length of stay by specific physician groups. The lack of clinical data on COPD severity (e.g., spirom- Acknowledgements etry) or baseline performance status also precluded the The authors acknowledge the members and staff of Alberta Health Services’ Respiratory Health Strategic Clinical Network, who helped with design and determination of differential effects of the order set implementation of the order set. between COPD subgroups; it is also possible that these factors influenced LOS or readmission rates independ- Funding This project was supported by Alberta Health Services’ Respiratory Health ently of the order set. However, this information is also Strategic Clinical Network Seed Grant. Members of the Respiratory Health often not available to clinicians at the time of admission. Strategic Clinical Network were involved in the design of the study and Thus, we chose to develop an order set that could be used revision of the manuscript. However, the seed grant used to fund this project was obtained through an independent peer review process. for all patients admitted with AECOPD, consistent with actual clinical practice. Our statistical models did account Availability of data and materials for age, sex and clinically relevant comorbidities, indicat- The datasets analysed during the current study are available from the corresponding author on reasonable request. ing that our results were robust to these covariates. Future studies could further evaluate how patient characteristics Authors’ contributions impact order set use and outcomes from the order set. All authors have met the ICMJE guidelines for responsible authorship. The authors’ substantial contributions to the study are listed below. All authors Finally, we did not analyze individual components of have critically revised previous versions of the manuscript, approve of the the AECOPD order set, and thus do not know which or- final version and agree to be accountable for all aspects of the work. Study ders were actually selected or executed (e.g., physiother- conception and design: SRP, NH, JG, PF, MB, CHM, MKS, Project implementation: SRP, NH, JG, CHM, Data acquisition and analysis: SRP, MBO, apy referral). We also cannot confirm whether there was DAS, PF, Manuscript preparation: SRP, MBO. concordance between pre-checked orders and actual or- ders selected by admitting physicians. These compo- Ethics approval and consent to participate The University of Calgary Conjoint Health Research Ethics Board waived the nents may have differential impact on LOS and could requirement for formal ethics approval. help refine the order set. An understanding of how indi- vidual components were used may also help to identify Competing interests areas for focused quality improvement. While the intent Dr. Leigh reports personal fees from AstraZeneca, grants from AstraZeneca, personal fees from Boehringer Ingelheim, grants from GlaxoSmithKline, of this study was to evaluate the effectiveness of a com- grants from MedImmune, personal fees from Novartis, personal fees from prehensive bundle of orders, the improvement in LOS TEVA Canada, outside the submitted work. The other authors have no provides compelling evidence to justify a secondary ana- disclosures. lysis of individual order set components. Publisher’sNote Springer Nature remains neutral with regard to jurisdictional claims in Conclusion published maps and institutional affiliations. In conclusion, this study found that when a standardized electronic order set was used to admit patients with Author details Department of Medicine, Cumming School of Medicine, University of AECOPD, LOS was reduced without increasing readmis- Calgary, Calgary, AB, Canada. Department of Community Health Sciences, sions. Innovations such as order sets have the potential Cumming School of Medicine, University of Calgary, Calgary, AB, Canada. to lessen the burden of AECOPD hospitalizations on O’Brien Institute for Public Health, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada. Respiratory Health Strategic Clinical both patients and the healthcare system, and justify Network, Alberta Health Services, Edmonton, AB, Canada. W21C Research additional studies of clinical decision support tools for and Innovation Centre, Cumming School of Medicine, University of Calgary, AECOPD. Calgary, AB, Canada. Research Priorities and Implementation, Alberta Health Pendharkar et al. BMC Pulmonary Medicine (2018) 18:93 Page 8 of 8 Services, Calgary, AB, Canada. Division of Pulmonary Medicine, Department 21. Howe GR. Use of computerized record linkage in cohort studies. Epidemiol of Medicine, Faculty of Medicine and Dentistry, University of Alberta, Rev. 1998;20:112–21. Edmonton, AB, Canada. University of Calgary, TRW Building, Rm 3E23, 3280 22. Koenker R. Quantile Regression [computer program]. 2015. https://cran.r- Hospital Drive NW, Calgary, AB T2N 4Z6, Canada. project.org/web/packages/quantreg/index.html. Accessed 22 July 2016. 23. Koenker R, Bassett G Jr. Regression quantiles. Econometrica. J Econom Soc. Received: 24 August 2017 Accepted: 21 May 2018 1978;46:33–50. 24. Portnoy S, Koenker R. The gaussian hare and the laplacian tortoise: computability of squared-error versus absolute-error estimators. Stat Sci. 1997;12(4):279–300. References 25. Charlson ME, Pompei P, Ales KL, Mackenzie CR. A new method of classifying 1. Seemungal TA, Donaldson GC, Paul EA, Bestall JC, Jeffries DJ, Wedzicha JA. prognostic comorbidity in longitudinal studies: development and validation. Effect of exacerbation on quality of life in patients with chronic obstructive J Chronic Dis. 1987;40:373–83. pulmonary disease. Am J Respir Crit Care Med. 1998;157(5):1418–22. 26. R. A language and environment for statistical computing [computer 2. Kanner RE, Anthonisen NR, Connett JE. Lung health study research group. program]. Vienna: R Foundation for Statistical Computing; 2015. Lower respiratory illnesses promote FEV(1) decline in current smokers but 27. Niewoehner DE, Erbland ML, Deupree RH, et al. Effect of systemic not ex-smokers with mild chronic obstructive pulmonary disease: results glucocorticoids on exacerbations of chronic obstructive pulmonary disease. from the lung health study. Am J Respir Crit Care Med. 2001;164(3):358–64. Department of Veterans Affairs cooperative study group. N Engl J Med. 3. Connors AF, Dawson NV, Thomas C, et al. Outcomes following acute 1999;340:1941–7. exacerbation of severe chronic obstructive lung disease. The SUPPORT 28. Quon BS, Gan WQ, Sin DD. Contemporary management of acute exacerbations investigators (study to understand prognoses and preferences for outcomes of COPD: a systematic review and metaanalysis. Chest. 2008;133:756–66. and risks of treatments). Am J Respir Crit Care Med. 1996;154(4 Pt 1):959–67. 29. de Morton NA, Keating JL, Jeffs K. Exercise for acutely hospitalised older 4. Wier LM, Elixhauser A, Pfuntner A, Au DH. Overview of hospitalizations among medical patients. Cochrane Database Syst Rev. 2007;1:CD005955. patients with COPD. Healthcare utilization project statistical brief #106. 30. Wang Y, Stavem K, Dahl FA, Humerfelt S, Haugen T. Factors associated with a Rockville, MD: Agency for Healthcare Research and Quality; 2008. p. 2011. prolonged length of stay after acute exacerbation of chronic obstructive 5. Waye AE, Jacobs P, Ospina MB, Stickland MK, Mayers I. Economic pulmonary disease (AECOPD). Int J Chron Obstruct Pulmon Dis. 2014;9:99–105. surveillance for chronic obstructive pulmonary disease (COPD) in Alberta. 31. Matkovic Z, Huerta A, Soler N, et al. Predictors of adverse outcome in Edmonton, AB: Institute for Health Economics; 2016. patients hospitalized for exacerbation of chronic obstructive pulmonary 6. Wang Q, Bourbeau J. Outcomes and health-related quality of life following disease. Respiration. 2012;84:17–26. hospitalization for an acute exacerbation of COPD. Respirol. 2005;10(3):334–40. 32. Wright A, Feblowitz JC, Pang JE, et al. Use of order sets in inpatient 7. O’Donnell DE, Hernandez P, Kaplan A, et al. Canadian Thoracic Society computerized provider order entry systems: a comparative analysis of usage recommendations for management of chronic obstructive pulmonary disease – patterns at seven sites. Int J Med Inform. 2012;81(11):733–45. 2008 update – highlights for primary care. Can Respir J. 2008;15(Suppl A):1A–8A. 33. Lennox L, Green S, Howe C, Musgrave H, Bell D, Elkin S. Identifying the 8. Vestbo J, Hurd SS, Agustí AG, et al. Global strategy for the diagnosis, challenges and facilitators of implementing a COPD care bundle. BMJ Open management, and prevention of chronic obstructive pulmonary disease: Respir Res. 2014;1:e000035. GOLD executive summary. Am J Respir Crit Care Med. 2013;187(4):347–65. 34. Bobb AM, Payne TH, Gross PA. Viewpoint: controversies surrounding use of 9. Sandhu SK, Chu J, Yurkovich M, Harriman D, Taraboanta C, Fitzgerald JM. order sets for clinical decision support in computerized provider order Variations in the management of acute exacerbations of chronic obstructive entry. J Am Med Inform Assoc. 2007;14:41–7. pulmonary disease. Can Respir J. 2013;20:175–9. 35. Portela MC, Pronovost PJ, Woodcock T, Carter P, Dixon-Woods M. How to 10. Noon CE, Hankins CT, Cote MJ. Understanding the impact of variation in study improvement interventions: a brief overview of possible study types. the delivery of healthcare services. J Healthc Manag. 2003;48(2):82–97. BMJ Qual Saf. 2015;24:325–36. 11. Hoogerduijn JG, Schuurmans MJ, Duijnstee MS, de Rooij SE, Grypdonck MF. A systematic review of predictors and screening instruments to identify older hospitalized patients at risk for functional decline. J Clin Nurs. 2007;16(1):46–57. 12. Bates DW, Leape LL, Cullen DJ, et al. Effect of computerized physician order entry and a team intervention on prevention of serious medication errors. JAMA. 1998;280(15):1311–6. 13. Brown KE, Johnson KJ, Deronne BM, Parenti CM, Rice KL. Order set to improve the care of patients hospitalized for COPD exacerbations. Ann Am Thorac Soc. 2016;13:811–5. 14. Garg AX, Adhikari NK, McDonald H, et al. Effects of computerized clinical decision support systems on practitioner performance and patient outcomes: a systematic review. JAMA. 2005;293:1223–38. 15. Gillaizeau F, Chan E, Trinquart L, et al. Computerized advice on drug dosage to improve prescribing practice. Cochrane Database Syst Rev. 2013;11(11): CD002894. 16. Kitchlu A, Abdelshaheed T, Tullis E, Gupta S. Gaps in the inpatient management of chronic obstructive pulmonary disease exacerbation and impact of an evidence-based order set. Can Respir J. 2015;22:157–62. 17. Miller RA, Waitman LR, Chen S, Rosenbloom ST. The anatomy of decision support during inpatient care provider order entry (CPOE): empirical observations from a decade of CPOE experience at Vanderbilt. J Biomed Inform. 2005;38:469–85. 18. Munasinghe RL, Arsene C, Abraham TK, Zidan M, Siddique M. Improving the utilization of admission order sets in a computerized physician order entry system by integrating modular disease specific order subsets into a general medicine admission order set. J Am Med Inform Assoc. 2011;18(3):322–6. 19. Pendharkar S, Hirani N, Faris P, et al. Effectiveness of a standardized inpatient admission order set for acute exacerbation of COPD [abstract]. Am J Respir Crit Care Med. 2015;191:A6179. 20. Hemming K, Haines TP, Chilton PJ, Girling AJ, Lilford RJ. The stepped wedge cluster randomized trial: rationale, design, analysis, and reporting. Br Med J. 2015;350:h391. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png BMC Pulmonary Medicine Springer Journals

Effectiveness of a standardized electronic admission order set for acute exacerbation of chronic obstructive pulmonary disease

Free
8 pages

Loading next page...
1
 
/lp/springer_journal/effectiveness-of-a-standardized-electronic-admission-order-set-for-micaEOkvWA
Publisher
Springer Journals
Copyright
Copyright © 2018 by The Author(s).
Subject
Medicine & Public Health; Pneumology/Respiratory System; Internal Medicine; Intensive / Critical Care Medicine
eISSN
1471-2466
D.O.I.
10.1186/s12890-018-0657-x
Publisher site
See Article on Publisher Site

Abstract

Background: Variation in hospital management of patients with acute exacerbation of chronic obstructive pulmonary disease (AECOPD) may prolong length of stay, increasing the risk of hospital-acquired complications and worsening quality of life. We sought to determine whether an evidence-based computerized AECOPD admission order set could improve quality and reduce length of stay. Methods: The order set was designed by a provincial COPD working group and implemented voluntarily among three physician groups in a Canadian tertiary-care teaching hospital. The primary outcome was length of stay for patients admitted during order set implementation period, compared to the previous 12 months. Secondary outcomes included length of stay of patients admitted with and without order set after implementation, all-cause readmissions, and emergency department visits. Results: There were 556 admissions prior to and 857 admissions after order set implementation, for which the order set was used in 47%. There was no difference in overall length of stay after implementation (median 6.37 days (95% confidence interval 5.94, 6.81) pre-implementation vs. 6.02 days (95% confidence interval 5.59, 6.46) post-implementation, p = 0.26). In the post-implementation period, order set use was associated with a 1.15-day reduction in length of stay (95% confidence interval − 0.5, − 1.81, p = 0.001) compared to patients admitted without the order set. There was no difference in readmissions. Conclusions: Use of a computerized guidelines-based admission order set for COPD exacerbations reduced hospital length of stay without increasing readmissions. Interventions to increase order set use could lead to greater improvements in length of stay and quality of care. Keywords: Length of stay, Clinical decision support, Chronic obstructive pulmonary disease, Quality improvement Background for AECOPD cost approximately USD $3.8 billion [4]and Chronic obstructive pulmonary disease (COPD) is a com- account for 51% of overall expenditures for COPD. [5]Pro- mon and progressive lung disease that is characterized by longed hospital length of stay (LOS) also have a negative shortness of breath, activity limitation, and a predisposition impact on patient function and quality of life. [6] to exacerbations. Acute exacerbations of COPD (AECOPD) Evidence-based management guidelines for AECOPD adversely affect quality of life, [1] increase the risk of disease have been developed, [7, 8] and include recommenda- progression, [2] and reduce survival. [3] Hospitalizations tions regarding pharmacotherapy and post-exacerbation care. Despite these guidelines, hospital care of patients * Correspondence: sachin.pendharkar@ucalgary.ca with AECOPD remains highly variable. [9] This variation Department of Medicine, Cumming School of Medicine, University of may contribute to prolonged LOS [10] that, in turn, in- Calgary, Calgary, AB, Canada creases the risk of hospital-acquired complications and Department of Community Health Sciences, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada adversely impacts quality of life. [6, 11] Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Pendharkar et al. BMC Pulmonary Medicine (2018) 18:93 Page 2 of 8 Order sets are grouped medical orders intended to criteria. Additional details on the methods are provided in standardize evidence-based best practice. Computerized an additional file (see Additional file 1). Physician Order Entry (CPOE) systems may improve work- flow, promote appropriate testing and treatment, reduce er- Order set development rors and improve guideline adherence, [12–17] particularly The AECOPD order set was based on published COPD when integrated into general order sets. [18]Standardized guidelines, [7] and developed by a provincial COPD admission order sets have been used in other diseases with working group comprised of physicians, nurses, and re- variable success at reducing hospital LOS. [14, 15] spiratory therapists from a variety of clinical settings, in Two observational studies have demonstrated that a series of face-to-face and teleconference meetings. order sets likely improve the quality of hospital care for The order set contained recommended testing, medi- patients with AECOPD and reduce LOS. [13, 16] How- cation (including suggested dosing and mode of deliv- ever, these studies used pre-post designs that could be ery), consultations, and a priori discharge planning influenced by secular trends in AECOPD management, interventions specific to patients with AECOPD. Some and the studies did not account for the differential interventions were pre-selected to encourage use (e.g., effects of the order set among physician groups. physiotherapy referral). The order set was built into the The objective of this study was to determine whether hospital’s existing CPOE system, Sunrise Clinical the implementation of an evidence-based computerized Manager (Allscripts Solutions, Chicago IL). Screenshots admission order set would improve the quality of in- are provided in an additional file (see Additional file 2). patient AECOPD care. A stepped wedge design was used to account for differential effects among physician Implementation groups and to minimize confounding related to the tim- The order set was implemented using a stepped wedge ing of order set implementation. Our hypothesis was design [20] among the three physician groups who admit that the implementation of a standardized order set patients with AECOPD: respirologists, general internists, would reduce hospital LOS of patients admitted for and family physician hospitalists. It was implemented se- AECOPD without increasing emergency department quentially within physician groups with each group act- (ED) or hospital readmissions. Preliminary study results ing as its own control. Study outcome data were have previously been reported in abstract form. [19] collected at baseline and at each implementation ‘step’. Implementation among respirologists, general inter- nists and hospitalists occurred in March, May and Au- Methods gust 2013, respectively. Prior to each implementation Study design step, the research team met with physicians and allied This study is an analysis of administrative health data for a health staff to introduce the order set. Order set use by quality improvement project in which an electronic stan- each individual physician was voluntary. Monthly statis- dardized admission order set for patients with AECOPD tics on order set use were posted in clinical areas. was implemented at a large, tertiary-care teaching hospital in Calgary, Alberta between March 1, 2013 and March 31, Analysis 2015. Since this was a quality improvement project, the Patient demographic, comorbidity and hospitalization University of CalgaryConjointHealthResearchEthics data were obtained from provincial administrative data Board waived the requirement for formal ethics approval. and linked to order set usage data from the CPOE sys- tem using the patient’s provincial health number. [21] Study population The primary outcome was hospital LOS for patients Patients were included if they were: older than 45 years of admitted during the implementation period compared to age; admitted to hospital between March 1, 2013 to March those admitted during the previous 12 months (pre-post 31, 2015 with an International Classification of Diseases, implementation analysis). Secondary outcomes included: Tenth Revision (ICD-10-CA) code indicative of AECOPD hospital LOS of patients admitted with and without the (J42 [unspecified chronic bronchitis], J43 [emphysema], or order set after implementation (post-implementation J44 [other chronic obstructive pulmonary disease]) in the analysis); all-cause readmissions at 7, 30 and 90 days after primary diagnosis field of the hospital discharge abstract discharge; ED visits at 7 and 30 days; and in-hospital database; and admitted to the pulmonary, general internal mortality. medicine or hospitalist clinical services. Patients were ex- Unadjusted and adjusted median regression models cluded if they were admitted to the intensive care unit or were constructed to assess the impact of the order set on any other clinical service. Historical controls from the LOS. [22–24] Covariates in adjusted models included age, 12 months prior to order set implementation in each sex, and five clinically relevant comorbidities (heart failure, group of ordering physicians were identified using similar dementia, liver disease, renal disease, and diabetes) that Pendharkar et al. BMC Pulmonary Medicine (2018) 18:93 Page 3 of 8 were strongly associated with the Charlson Comorbidity the pre-implementation period (64.5% vs. 74.3%). Patients Index (Somers’ D = 0.94). [25] Logistic regression was with co-existing heart failure and diabetes were more com- used to adjust 30-day readmission odds ratios for age, sex, monly admitted under general internists. Over 95% of pa- comorbidity and admitting physician specialty. tients were discharged home. All analyses were performed using SAS version 9.3 (Cary, NC) or R version 3.2.3; [26] p < 0.05 was consid- Order set uptake ered statistically significant. In the post-implementation period, 57% of patients admitted to the hospitalist service were admitted using Results the order set, compared to 30% of patients admitted by Of 1435 AECOPD admissions to one of the three physician general internists or respirologists. Time series analysis groups during the study period, 1413 with a LOS less than revealed that order set use increased gradually after 90 days were included in the analysis (Fig. 1). There were implementation, mostly by general internists and hospi- 857 admissions after order set implementation, of which talists (Additional file 3: Figure S1). 406 patients (47%) were admitted using the order set. Baseline characteristics of study participants are pre- Hospital LOS sented in Tables 1 and 2 for the pre-post and Figure 2 shows the unadjusted and adjusted differences post-implementation analyses, respectively. The hospitalist in median LOS for patients treated in the pre- and service admitted most patients with AECOPD, but admit- post-implementation periods. Median LOS was 6.37 days ted fewer in the post-implementation period compared to (95% confidence interval [CI] 5.94, 6.81; n = 556) for patients admitted before implementation and 6.02 days (95% CI 5.59, 6.46; n = 857) for patients admitted after implementation (p = 0.26). Unadjusted and adjusted comparisons of median LOS in the post-implementation analysis are presented in Fig. 3. Order set use was associated with a 1.15-day (95% CI -0.50, − 1.81) shorter median LOS, due primarily to a 1.8-day (95% CI -0.95, − 2.61) decrease for the hospitalist group (Fig. 3). Median LOS for patients admitted by general inter- nists or respirologists did not differ by order set use. Readmissions Neither order set implementation, nor order set use in the post-implementation period were associated with changes in readmissions or ED visits for all three phys- ician groups (Tables 3 and 4 and Additional file 1 Table S1). Overall in-hospital mortality did not change with order set implementation, but was lower in the hospital- ist group with order set use. Discussion This is thelargest studytoevaluatethe impact of a stan- dardized, guideline-based electronic order set for AECOPD on hospital LOS. We performed two analyses: a compari- son of LOS before and after the order set was imple- mented, and a comparison of patients admitted with and without the order set after implementation. The results re- vealed that order set implementation did not result in an overall LOS reduction, perhaps because only 47% of admit- ting physicians used it. However, the post-implementation analysis revealed that when it was used, the order set was associated with a LOS reduction of 1.15 days. Use of the order set by hospitalists, who admitted 65% of AECOPD Fig. 1 Patient flow diagram. AECOPD – acute exacerbation of patients in the post-implementation cohort, resulted in the chronic obstructive pulmonary disease; LOS – length of stay largest LOS reduction of 1.8 days. Importantly, there was Pendharkar et al. BMC Pulmonary Medicine (2018) 18:93 Page 4 of 8 Table 1 Baseline characteristics for pre-post implementation analysis Characteristics Total Pre-implementation Post-implementation p-value Number of patients 1413 556 857 Mean age, years (SD) 70 (12) 70 (12) 70 (12) 0.747 Age group, n (%) < 55 129 (9.1) 59 (10.6) 70 (8.2) 0.250 55–64 323 (22.9) 128 (23.0) 195 (22.8) 65–74 428 (30.3) 157 (28.2) 271 (31.6) 75–84 360 (25.5) 136 (24.5) 224 (26.1) 85+ 173 (12.2) 76 (13.7) 97 (11.3) Male sex, n (%) 727 (51.5) 279 (50.2) 448 (52.3) 0.441 Comorbidity, n (%) Heart failure 190 (13.5) 71 (12.8) 119 (13.9) 0.548 Dementia 45 (3.2) 24 (4.3) 21 (2.5) 0.051 Diabetes 318 (22.5) 131 (23.6) 187 (21.8) 0.444 Renal disease 36 (2.6) 13 (2.3) 23 (2.7) 0.687 Liver disease 13 (0.9) 6 (1.1) 7 (0.8) 0.614 Admitting specialty, n (%) Respirologist 148 (10.5) 51 (9.2) 97 (11.3) 0.0005 General internist 299 (21.2) 92 (16.6) 207 (24.2) Hospitalist 966 (68.4) 413 (74.3) 553 (64.5) SD = standard deviation no increase in either ED or hospital readmissions, suggest- hospitalized with AECOPD, Kitchlu et al. showed that ing that earlier discharge resulting from order set use did implementation of an order set improved the quality not occur at the expense of harm to the patient. of admission orders using pre-specified measures of Our findings extend observations from two recent guidelines-based care. [16] Using a similar design in a studies examining the impact of order sets on study of 275 patients, Brown et al. showed that phys- AECOPD care. In a pre-post design of 243 patients ician prescribing practices for AECOPD could be Table 2 Baseline characteristics for post-implementation analysis Characteristics Respirologist General internist Hospitalist No order set (n = 64) Order set (n = 33) No order set (n = 148) Order set (n = 59) No order set (n = 239) Order set (n = 314) Mean age, years (SD) 63 (11) 64 (10) 69 (11) 71 (12) 72 (12) 71 (10) Age group, n (%) < 55 8 (13) 6 (18) 12 (8) 8 (14) 15 (6) 21 (7) 55–64 27 (42) 10 (30) 37 (25) 14 (24) 39 (16) 68 (22) 65–74 20 (31) 9 (27) 50 (34) 14 (24) 72 (30) 106 (34) 75–84 9 (14) 8 (24) 36 (24) 18 (31) 65 (27) 88 (28) 85+ 0 0 13 (9) 5 (9) 48 (20) 31 (10) Male sex, n (%) 34 (53) 12 (36) 75 (51) 29 (49) 134 (56) 159 (51) Comorbidity, n (%) Heart failure 4 (6) 2 (6) 39 (26) 11 (19) 37 (15) 26 (8) Dementia 0 1 (3) 4 (3) 0 11 (5) 5 (2) Diabetes 9 (14) 8 (24) 46 (31) 16 (27) 45 (19) 63 (20) Renal disease 2 (3) 0 4 (3) 2 (3) 7 (3) 8 (3) Liver disease 0 0 1 (1) 0 2 (1) 2 (1) SD = standard deviation Pendharkar et al. BMC Pulmonary Medicine (2018) 18:93 Page 5 of 8 Fig. 2 Forest plot of implementation effects (pre-post implementation analysis). Adjusted model included age, sex, and five clinically relevant comorbidities selected from the Charlson Comorbidity Index (heart failure, dementia, mild or severe liver disease, renal disease, and diabetes. Pre – pre-implementation; Post – post-implementation; N – number of patients; Med – median length of stay; IQR – interquartile range; LOS – length of stay; CI – confidence interval improved with an electronic order set. [13]Secondary A standardized order set is appealing due to high vari- analyses in both studies revealed LOS reductions ability in inpatient AECOPD management. [9] Previous without an increase in readmissions. The current studies of order sets for AECOPD demonstrated more study extends this work by demonstrating an im- consistent use of systemic corticosteroids, appropriate provement in hospital LOS in a larger study cohort. antibiotics, and allied health providers such as physiother- Unlike the previous studies, which reported on apists, [13, 16] all of which have been shown to reduce pre-post effects of order set implementation, we hospital LOS. [27–29] The current order set similarly specifically examined the effect of actual use of the prompted admitting physicians to use these therapies, and order set, demonstrating that it could reduce LOS pre-selection of some items (e.g., bronchodilator delivery, compared to patients admitted without the order set. physiotherapy referral) provided additional clinical deci- Furthermore, the stepped wedge design is robust to sion support around guidelines-based care. secular trends in care delivery and LOS of AECOPD The LOS reduction observed after order set imple- patients, and allowed for subgroup analyses by differ- mentation was driven by improvements for patients ad- ent admitting physician groups. Both our study and mitted by hospitalists, with no differences observed in thepreviousstudies providea strong rationalefor the the other physician groups; this was an interesting find- standardization of inpatient AECOPD care using ing with many possible explanations. First, Sandhu et al. computerized order sets. demonstrated high variability in AECOPD management Fig. 3 Forest plot of the effects of the order set (post-implementation analysis). Adjusted model included age, sex, and five clinically relevant comorbidities selected from the Charlson Comorbidity Index (heart failure, dementia, mild or severe liver disease, renal disease, and diabetes. OS – order set; N – number of patients; Med – median length of stay; IQR – interquartile range; LOS – length of stay; CI – confidence interval; OS – order set Pendharkar et al. BMC Pulmonary Medicine (2018) 18:93 Page 6 of 8 Table 3 Readmissions, emergency department visits and in-hospital mortality for pre-post implementation analysis Results Pre-implementation (n = 556) Post-implementation (n = 857) p-value 7-day readmission, n (%) 33 (5.9) 60 (7.0) 0.430 30-day readmission, n, (%) 91 (16.4) 166 (19.4) 0.153 90-day readmission, n, (%) 170 (30.6) 299 (34.9) 0.093 7-day ED visits, n (%) 42 (7.6) 55 (6.4) 0.409 30-day ED visits, n (%) 124 (22.3) 196 (22.9) 0.803 In-hospital mortality, n, (%) 19 (3.4) 31 (3.6) 0.842 ED = emergency department among all specialties, with deviation from clinical guide- hospitalists, for a number of possible reasons. First, the lines occurring more often when care was provided by complexity of the patient’s presentation (e.g., respiratory physicians other than respirologists. [9] This variability failure requiring noninvasive ventilation) may have made may be due to the diversity of medical problems man- the order set less applicable at the time of admission. aged by hospitalists, which could lead to a less uniform Second, respirologists may have greater perceived approach to inpatient AECOPD management. Thus, it is self-efficacy with AECOPD management, leading them possible that the opportunity for standardization using to admit AECOPD patients without using an order set. an order set was greater for hospitalists than for other Finally, whereas the AECOPD order set was the only re- specialties. Second, heart failure and diabetes were more spiratory order set embedded in the CPOE system, sev- prevalent in patients admitted under general internists eral admission order sets for medical problems typically compared to hospitalist patients. These conditions have admitted under hospitalists (e.g., pneumonia, heart fail- both been associated with longer LOS in patients with ure) were already embedded. Thus, hospitalists may have AECOPD, [30] and are unlikely to be impacted by order more experience in order set use compared to other set use. Third, patients presenting with respiratory fail- physicians. Importantly, end users were consulted to en- ure requiring noninvasive ventilation were only admitted sure the order set was intuitive and minimally disruptive by general internists or respirologists; these indicators of to clinical workflow; these factors have been shown to more severe AECOPD were not systematically captured, increase uptake of clinical decision support systems such but could have reduced the effectiveness of the order as standardized order sets. [14, 17, 18] The increase in set. [31] Although the order set’s impact seemed to be order set use over time suggested that admitting physi- isolated to only one admitting group, the 1.8-day reduc- cians found it useful. tion in median LOS is an important finding since hospi- This study has a number of limitations. First, the talists were responsible for providing almost two thirds non-randomized study design raises the possibility that of inpatient AECOPD care in our study, and this is likely improvements in LOS were due to other differences be- to be similar in other large, tertiary-care urban hospitals tween groups admitted with and without the order set. in North America. However, when implementing complex healthcare inter- The order set was used by 47% of admitting physicians ventions such as the AECOPD order set, traditional ran- during the study period. This low uptake is a consistent domized controlled trials are impractical due to finding for voluntary order sets [13, 16, 32, 33] and is a logistical constraints and the risk of contamination known limitation of their use. [34] Respirologists and within clinical provider groups. The methodologically general internists used the order set less frequently than robust stepped wedge design minimized contamination Table 4 Readmissions, emergency department visits and in-hospital mortality for post-implementation analysis Results Respirologist General internist Hospitalist No order set Order set p value No order set Order set p value No order set Order set p value (n = 64) (n = 33) (n = 148) (n = 59) (n = 239) (n = 314) 7-day readmission, n (%) 4 (6.3) 2 (6.1) 0.971 18 (12.2) 4 (6.8) 0.257 13 (5.4) 19 (6.1) 0.760 30-day readmission, n (%) 19 (29.7) 5 (15.2) 0.116 29 (19.6) 14 (23.7) 0.508 41 (17.2) 58 (18.5) 0.689 90-day readmission, n (%) 30 (46.9) 13 (39.4) 0.482 59 (39.9) 19 (32.2) 0.305 75 (31.4) 103 (32.8) 0.723 7-day ED visits, n (%) 2 (3.1) 1 (3.0) 0.980 12 (8.1) 3 (5.1) 0.449 15 (6.3) 22 (7.0) 0.734 30-day ED visits, n (%) 18 (28.1) 5 (15.2) 0.155 28 (18.9) 14 (23.7) 0.437 54 (22.6) 77 (24.5) 0.597 In-hospital mortality, n (%) 0 1 (3.0) 0.162 8 (5.4) 5 (8.5) 0.411 12 (5) 5 (1.6) 0.021 ED = emergency department Pendharkar et al. BMC Pulmonary Medicine (2018) 18:93 Page 7 of 8 by allowing implementation and evaluation of the order Additional files set in clusters [35] while still analyzing the effect of Additional file 1: Supplement with additional details on methods and order sets within physician groups. results (DOCX 28 kb) Second, the use of administrative data prevented ana- Additional file 2: Screenshot of AECOPD order set (PDF 400 kb) lysis of patient characteristics that might have influenced Additional file 3: Figure S1. Monthly percentage of patients admitted a physician’s decision to use the order set; it is thus pos- using AECOPD order set during study period. Vertical lines represent sible that the reduced LOS in the post-implementation implementation start dates for each physician specialty. Respirologist represented by hatched line; general internist represented by black solid period is due to order set use in less complex cases. line; hospitalist represented by grey solid line. Reported probabilities are While our findings are consistent with studies per- for linear trends from time series models for each physician specialty. formed in different geographic and clinical settings, [13, These models showed no evidence of seasonality or auto-regression (JPG 112 kb) 16] we acknowledge the importance of future studies to examine whether COPD severity, presentation acuity, or Abbreviations use of specific interventions (e.g., noninvasive ventila- AECOPD: Acute exacerbation of chronic obstructive pulmonary disease; tion) impact order set use. Such an analysis would help CI: Confidence interval; COPD: Chronic obstructive pulmonary disease; to tailor strategies aimed at increasing order set uptake CPOE: Computerized physician order entry; ED: Emergency department; LOS: Length of stay by specific physician groups. The lack of clinical data on COPD severity (e.g., spirom- Acknowledgements etry) or baseline performance status also precluded the The authors acknowledge the members and staff of Alberta Health Services’ Respiratory Health Strategic Clinical Network, who helped with design and determination of differential effects of the order set implementation of the order set. between COPD subgroups; it is also possible that these factors influenced LOS or readmission rates independ- Funding This project was supported by Alberta Health Services’ Respiratory Health ently of the order set. However, this information is also Strategic Clinical Network Seed Grant. Members of the Respiratory Health often not available to clinicians at the time of admission. Strategic Clinical Network were involved in the design of the study and Thus, we chose to develop an order set that could be used revision of the manuscript. However, the seed grant used to fund this project was obtained through an independent peer review process. for all patients admitted with AECOPD, consistent with actual clinical practice. Our statistical models did account Availability of data and materials for age, sex and clinically relevant comorbidities, indicat- The datasets analysed during the current study are available from the corresponding author on reasonable request. ing that our results were robust to these covariates. Future studies could further evaluate how patient characteristics Authors’ contributions impact order set use and outcomes from the order set. All authors have met the ICMJE guidelines for responsible authorship. The authors’ substantial contributions to the study are listed below. All authors Finally, we did not analyze individual components of have critically revised previous versions of the manuscript, approve of the the AECOPD order set, and thus do not know which or- final version and agree to be accountable for all aspects of the work. Study ders were actually selected or executed (e.g., physiother- conception and design: SRP, NH, JG, PF, MB, CHM, MKS, Project implementation: SRP, NH, JG, CHM, Data acquisition and analysis: SRP, MBO, apy referral). We also cannot confirm whether there was DAS, PF, Manuscript preparation: SRP, MBO. concordance between pre-checked orders and actual or- ders selected by admitting physicians. These compo- Ethics approval and consent to participate The University of Calgary Conjoint Health Research Ethics Board waived the nents may have differential impact on LOS and could requirement for formal ethics approval. help refine the order set. An understanding of how indi- vidual components were used may also help to identify Competing interests areas for focused quality improvement. While the intent Dr. Leigh reports personal fees from AstraZeneca, grants from AstraZeneca, personal fees from Boehringer Ingelheim, grants from GlaxoSmithKline, of this study was to evaluate the effectiveness of a com- grants from MedImmune, personal fees from Novartis, personal fees from prehensive bundle of orders, the improvement in LOS TEVA Canada, outside the submitted work. The other authors have no provides compelling evidence to justify a secondary ana- disclosures. lysis of individual order set components. Publisher’sNote Springer Nature remains neutral with regard to jurisdictional claims in Conclusion published maps and institutional affiliations. In conclusion, this study found that when a standardized electronic order set was used to admit patients with Author details Department of Medicine, Cumming School of Medicine, University of AECOPD, LOS was reduced without increasing readmis- Calgary, Calgary, AB, Canada. Department of Community Health Sciences, sions. Innovations such as order sets have the potential Cumming School of Medicine, University of Calgary, Calgary, AB, Canada. to lessen the burden of AECOPD hospitalizations on O’Brien Institute for Public Health, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada. Respiratory Health Strategic Clinical both patients and the healthcare system, and justify Network, Alberta Health Services, Edmonton, AB, Canada. W21C Research additional studies of clinical decision support tools for and Innovation Centre, Cumming School of Medicine, University of Calgary, AECOPD. Calgary, AB, Canada. Research Priorities and Implementation, Alberta Health Pendharkar et al. BMC Pulmonary Medicine (2018) 18:93 Page 8 of 8 Services, Calgary, AB, Canada. Division of Pulmonary Medicine, Department 21. Howe GR. Use of computerized record linkage in cohort studies. Epidemiol of Medicine, Faculty of Medicine and Dentistry, University of Alberta, Rev. 1998;20:112–21. Edmonton, AB, Canada. University of Calgary, TRW Building, Rm 3E23, 3280 22. Koenker R. Quantile Regression [computer program]. 2015. https://cran.r- Hospital Drive NW, Calgary, AB T2N 4Z6, Canada. project.org/web/packages/quantreg/index.html. Accessed 22 July 2016. 23. Koenker R, Bassett G Jr. Regression quantiles. Econometrica. J Econom Soc. Received: 24 August 2017 Accepted: 21 May 2018 1978;46:33–50. 24. Portnoy S, Koenker R. The gaussian hare and the laplacian tortoise: computability of squared-error versus absolute-error estimators. Stat Sci. 1997;12(4):279–300. References 25. Charlson ME, Pompei P, Ales KL, Mackenzie CR. A new method of classifying 1. Seemungal TA, Donaldson GC, Paul EA, Bestall JC, Jeffries DJ, Wedzicha JA. prognostic comorbidity in longitudinal studies: development and validation. Effect of exacerbation on quality of life in patients with chronic obstructive J Chronic Dis. 1987;40:373–83. pulmonary disease. Am J Respir Crit Care Med. 1998;157(5):1418–22. 26. R. A language and environment for statistical computing [computer 2. Kanner RE, Anthonisen NR, Connett JE. Lung health study research group. program]. Vienna: R Foundation for Statistical Computing; 2015. Lower respiratory illnesses promote FEV(1) decline in current smokers but 27. Niewoehner DE, Erbland ML, Deupree RH, et al. Effect of systemic not ex-smokers with mild chronic obstructive pulmonary disease: results glucocorticoids on exacerbations of chronic obstructive pulmonary disease. from the lung health study. Am J Respir Crit Care Med. 2001;164(3):358–64. Department of Veterans Affairs cooperative study group. N Engl J Med. 3. Connors AF, Dawson NV, Thomas C, et al. Outcomes following acute 1999;340:1941–7. exacerbation of severe chronic obstructive lung disease. The SUPPORT 28. Quon BS, Gan WQ, Sin DD. Contemporary management of acute exacerbations investigators (study to understand prognoses and preferences for outcomes of COPD: a systematic review and metaanalysis. Chest. 2008;133:756–66. and risks of treatments). Am J Respir Crit Care Med. 1996;154(4 Pt 1):959–67. 29. de Morton NA, Keating JL, Jeffs K. Exercise for acutely hospitalised older 4. Wier LM, Elixhauser A, Pfuntner A, Au DH. Overview of hospitalizations among medical patients. Cochrane Database Syst Rev. 2007;1:CD005955. patients with COPD. Healthcare utilization project statistical brief #106. 30. Wang Y, Stavem K, Dahl FA, Humerfelt S, Haugen T. Factors associated with a Rockville, MD: Agency for Healthcare Research and Quality; 2008. p. 2011. prolonged length of stay after acute exacerbation of chronic obstructive 5. Waye AE, Jacobs P, Ospina MB, Stickland MK, Mayers I. Economic pulmonary disease (AECOPD). Int J Chron Obstruct Pulmon Dis. 2014;9:99–105. surveillance for chronic obstructive pulmonary disease (COPD) in Alberta. 31. Matkovic Z, Huerta A, Soler N, et al. Predictors of adverse outcome in Edmonton, AB: Institute for Health Economics; 2016. patients hospitalized for exacerbation of chronic obstructive pulmonary 6. Wang Q, Bourbeau J. Outcomes and health-related quality of life following disease. Respiration. 2012;84:17–26. hospitalization for an acute exacerbation of COPD. Respirol. 2005;10(3):334–40. 32. Wright A, Feblowitz JC, Pang JE, et al. Use of order sets in inpatient 7. O’Donnell DE, Hernandez P, Kaplan A, et al. Canadian Thoracic Society computerized provider order entry systems: a comparative analysis of usage recommendations for management of chronic obstructive pulmonary disease – patterns at seven sites. Int J Med Inform. 2012;81(11):733–45. 2008 update – highlights for primary care. Can Respir J. 2008;15(Suppl A):1A–8A. 33. Lennox L, Green S, Howe C, Musgrave H, Bell D, Elkin S. Identifying the 8. Vestbo J, Hurd SS, Agustí AG, et al. Global strategy for the diagnosis, challenges and facilitators of implementing a COPD care bundle. BMJ Open management, and prevention of chronic obstructive pulmonary disease: Respir Res. 2014;1:e000035. GOLD executive summary. Am J Respir Crit Care Med. 2013;187(4):347–65. 34. Bobb AM, Payne TH, Gross PA. Viewpoint: controversies surrounding use of 9. Sandhu SK, Chu J, Yurkovich M, Harriman D, Taraboanta C, Fitzgerald JM. order sets for clinical decision support in computerized provider order Variations in the management of acute exacerbations of chronic obstructive entry. J Am Med Inform Assoc. 2007;14:41–7. pulmonary disease. Can Respir J. 2013;20:175–9. 35. Portela MC, Pronovost PJ, Woodcock T, Carter P, Dixon-Woods M. How to 10. Noon CE, Hankins CT, Cote MJ. Understanding the impact of variation in study improvement interventions: a brief overview of possible study types. the delivery of healthcare services. J Healthc Manag. 2003;48(2):82–97. BMJ Qual Saf. 2015;24:325–36. 11. Hoogerduijn JG, Schuurmans MJ, Duijnstee MS, de Rooij SE, Grypdonck MF. A systematic review of predictors and screening instruments to identify older hospitalized patients at risk for functional decline. J Clin Nurs. 2007;16(1):46–57. 12. Bates DW, Leape LL, Cullen DJ, et al. Effect of computerized physician order entry and a team intervention on prevention of serious medication errors. JAMA. 1998;280(15):1311–6. 13. Brown KE, Johnson KJ, Deronne BM, Parenti CM, Rice KL. Order set to improve the care of patients hospitalized for COPD exacerbations. Ann Am Thorac Soc. 2016;13:811–5. 14. Garg AX, Adhikari NK, McDonald H, et al. Effects of computerized clinical decision support systems on practitioner performance and patient outcomes: a systematic review. JAMA. 2005;293:1223–38. 15. Gillaizeau F, Chan E, Trinquart L, et al. Computerized advice on drug dosage to improve prescribing practice. Cochrane Database Syst Rev. 2013;11(11): CD002894. 16. Kitchlu A, Abdelshaheed T, Tullis E, Gupta S. Gaps in the inpatient management of chronic obstructive pulmonary disease exacerbation and impact of an evidence-based order set. Can Respir J. 2015;22:157–62. 17. Miller RA, Waitman LR, Chen S, Rosenbloom ST. The anatomy of decision support during inpatient care provider order entry (CPOE): empirical observations from a decade of CPOE experience at Vanderbilt. J Biomed Inform. 2005;38:469–85. 18. Munasinghe RL, Arsene C, Abraham TK, Zidan M, Siddique M. Improving the utilization of admission order sets in a computerized physician order entry system by integrating modular disease specific order subsets into a general medicine admission order set. J Am Med Inform Assoc. 2011;18(3):322–6. 19. Pendharkar S, Hirani N, Faris P, et al. Effectiveness of a standardized inpatient admission order set for acute exacerbation of COPD [abstract]. Am J Respir Crit Care Med. 2015;191:A6179. 20. Hemming K, Haines TP, Chilton PJ, Girling AJ, Lilford RJ. The stepped wedge cluster randomized trial: rationale, design, analysis, and reporting. Br Med J. 2015;350:h391.

Journal

BMC Pulmonary MedicineSpringer Journals

Published: May 30, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off