Key Points
Question
Does a quality improvement intervention bundle increase use of isotonic maintenance intravenous fluids (IVF) and decrease laboratory testing?
Findings
In this stepped-wedge, cluster randomized clinical trial of 106 pediatric hospitals, the intervention bundle led to an absolute increase of approximately 5% above baseline in exclusive isotonic IVF use but did not change the mean proportion of hospital days during which a laboratory value was obtained.
Meaning
In this study, an intervention bundle significantly improved the use of isotonic maintenance IVF without a concomitant increase in electrolyte testing.
Abstract
Importance
Given that hypotonic maintenance intravenous fluids (IVF) may cause hospital-acquired harm, in November 2018, the American Academy of Pediatrics released a clinical practice guideline recommending the use of isotonic IVF for patients aged 28 days to 18 years without contraindications. No recommendations were made regarding laboratory monitoring; however, unnecessary laboratory tests may contribute to health care waste and harm patients.
Objective
To examine the effect of a quality improvement intervention bundle on (1) increasing the mean proportion of hours per hospital day with exclusive isotonic IVF use to at least 80% and (2) decreasing the mean proportion of hospital days with laboratory tests obtained.
Design, Setting, and Participants
This stepped-wedge, cluster randomized clinical trial (Standardization of Fluids in Inpatient Settings [SOFI]) was sponsored by a national quality improvement collaborative and was conducted across 106 US pediatric hospitals. The SOFI intervention period was from September 2019 to March 2020.
Interventions
Hospital sites were exposed to educational materials, a clinical algorithm and order set for IVF use, electronic medical record interventions to reduce laboratory testing, and “harms of overtesting” cards.
Main Outcomes and Measures
Primary outcomes were mean proportion of hours per hospital day receiving exclusive isotonic IVF and mean proportion of hospital days with laboratory test values obtained. Secondary measures included total IVF duration per hospital day, daily patient weight measurement while receiving IVF, serum sodium testing, and adverse events. Baseline data were collected for 2 months; intervention period data, 7 months. Outcomes were analyzed using linear mixed-effects regression models.
Results
A total of 106 hospitals were randomly assigned to 1 of 3 intervention start dates (wedges), and 100 hospitals (94%) completed the study. In total, 5215 hospitalizations were reviewed before the intervention, and 6724 hospitalizations were reviewed after the intervention. Prior to interventions, the mean (SD) proportion of hours per day with exclusive isotonic IVF use was 88.5% (31.7%). Interventions led to an absolute increase of 5.4% (95% CI, 3.9%-6.9%) above baseline in exclusive isotonic IVF use but did not change the proportion of hospital days during which a laboratory test value was obtained (estimated difference, 0.1%; 95% CI, –1.5% to 1.7%; P = .90), IVF use duration (estimated difference, –1.2%; 95% CI, –2.9% to 0.4%), serum sodium testing, or adverse events. There was an absolute increase of 4.4% (95% CI, 2.6%-6.2%) in the mean proportion of hospital days with a patient weight measurement while receiving IVF.
Conclusions and Relevance
In this stepped-wedge, cluster randomized clinical trial, an intervention bundle significantly improved the use of isotonic maintenance IVF without a concomitant increase in adverse events or electrolyte testing. Further work is required to deimplement laboratory testing.
Trial Registration
ClinicalTrials.gov Identifier: NCT03924674
This stepped-wedge, cluster randomized clinical trial assesses whether hospital exposure vs no exposure to a quality improvement intervention bundle increases use of isotonic maintenance intravenous fluids and decreases laboratory testing among hospitalized patients.
Introduction
Inappropriate use of hypotonic maintenance intravenous fluids (IVF) may cause serious hospital-acquired harm, including iatrogenic hyponatremia with neurologic sequelae.1,2,3,4,5,6,7,8 Although symptomatic, hospital-acquired hyponatremia is rare, the bulk of evidence suggests that it is a preventable harm for which hypotonic IVF are the major risk factor. In November 2018, the American Academy of Pediatrics (AAP) released a clinical practice guideline on maintenance IVF with 1 major recommendation to use isotonic maintenance IVF for medical and surgical patients 28 days to 18 years of age without preexisting adrenal, cardiac, hepatic, neurosurgical, hematologic or oncologic, and kidney disease.9
No recommendations were made in the guideline regarding laboratory monitoring (eg, electrolytes) while receiving IVF given a dearth of high-quality data. However, unnecessary laboratory testing is prevalent and contributes to health care waste, often triggering a cascade of interventions without improving quality.10,11,12,13,14,15 Given the common co-occurrence of IV placement, laboratory testing, and IVF initiation, interventions to change maintenance IVF prescribing practice present an opportunity to simultaneously reduce laboratory testing.
Published interventions to increase appropriate use of isotonic IVF have been limited to single institutions.16 Given that guideline publication alone has not been consistently associated with rapid or marked changes in clinician practice,17,18,19,20 we conducted the multicenter, stepped-wedge, cluster randomized clinical trial Standardization of Fluids in Inpatient Settings (SOFI) to evaluate a quality improvement (QI) intervention bundle for isotonic IVF use, with an additional primary outcome of laboratory testing. During a 7-month period, we aimed to (1) increase the proportion of hours per hospital day with exclusive isotonic maintenance IVF use to at least 80% and (2) decrease the proportion of hospital days with laboratory values evaluated by 20% from baseline.
Methods
Context and Trial Design
The SOFI trial was sponsored by the AAP inpatient QI collaborative, the Value in Inpatient Pediatrics (VIP) Network (trial protocol in Supplement 1). All AAP members may join VIP and lead QI projects at their institution regardless of location, hospital size, or clinical structure. A stepped-wedge, cluster randomized design was used with a closed cohort of hospitals. Baseline data collection was followed by intervention implementation at 3 separate time points. Baseline data collection was intentionally selected to start 6 months before guideline publication (May 2018) and to capture the first 9 months after publication (August 2019). Random assignments occurred by hospital to 1 of 3 wedges, with intervention implementation 2 months apart for each wedge (Figure 1). Three wedges were selected to efficiently deliver the intervention bundle to all hospitals in a time frame that would minimize cross-contamination and appropriately match the operational resources of the study. All hospitals collected data through monthly electronic health record (EHR) reviews before and after intervention start times. This study followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline. The AAP institutional review board (IRB) approved SOFI, and hospital leaders handled IRB applications at their institution. The AAP IRB waived the requirement for obtaining informed consent because no protected health information was collected, and all data were presented either in aggregate or with deidentified preassigned hospital numbers. No physician received compensation for participating in this study. Maintenance of certification credits were available to participants on successful completion of the study.
Figure 1. Standardization of Fluids in Inpatient Settings (SOFI) Study Design.
All hospitals collected data for the baseline (before intervention) period from May 2018 to August 2019. At step 1 (September 2019), 37 hospitals randomly assigned to the first wedge began interventions. At step 2 (November 2019), 34 hospitals randomly assigned to the second wedge began interventions. At step 3 (January 2020), 35 hospitals randomly assigned to the third wedge began interventions. All data were collected by hospital in monthly increments throughout all periods.
Participants
An open call for participants was distributed via the AAP Section on Hospital Medicine and QI listservs. Those interested were asked to complete an application and designate an interdisciplinary leadership group with a pediatric hospital medicine physician, nursing leader, and pharmacy leader. All hospitals with a completed application were included.
Inclusion and exclusion criteria for EHR reviews aligned with recommendations in the guideline.9 Patients had to be aged 28 days to 18 years at the time of admission and to have received maintenance IVF (defined as a rate >10 mL/h) on their second hospital day at 12:01 am. We chose this time to capture a point after initial fluid resuscitation (ie, fluid bolus administration) that was reasonably early in the hospital course. Exclusion criteria included patients admitted to the intensive care unit; patients who were not discharged from the hospital; patients with active adrenal, cardiac, hepatic, neurosurgical, chronic kidney, hematologic or oncologic, or biochemical genetic or metabolic illnesses (defined by review of medical history in the EHR or an active medication related to 1 of these conditions); patients with diabetes insipidus, diabetic ketoacidosis, or severe burns; or patients with a primary behavioral health problem who required no medical treatment.
Hospitals reviewed the first 20 inpatient encounters meeting criteria per month (or, if <20, all encounters meeting criteria). The goal of 20 EHRs per month per hospital was chosen pragmatically to not oversample larger hospitals while still collecting sufficient data per hospital to detect changes in use of isotonic IVF and laboratory testing over time. Participants entered data into an online AAP data aggregator, which generated run charts in real time by hospital as well as benchmarking by aggregate performance across the study. Hospitals with more than 1 EHR reviewer conducted a quality assurance check by reviewing the same 2 EHRs and comparing entries to discuss discrepancies.
Interventions
Intervention Bundle
A password-protected website was available to all participants. Two months prior to each step (Figure 1), hospitals that were randomly assigned to begin interventions received access to the intervention bundle (eAppendix in Supplement 2) to allow sufficient lead time to set up stakeholder meetings and to submit requests for EHR changes. To prevent contamination, online access to the interventions was restricted by wedge. The bundle included (1) webinars on the potential harms associated with maintenance IVF and laboratory testing; (2) a clinical algorithm and order set for maintenance IVF initiation and discontinuation; (3) recommendations to restock automated dispensing cabinets with isotonic instead of hypotonic fluids to make them more readily available; (4) a pharmacist checklist for verifying maintenance IVF orders similar to other medications (ie, checking indication, tonicity, and rate); (5) removal of automatically recurring laboratory tests from order sets and substitution of focused laboratory values instead of panels (eg, hemoglobin instead of complete blood count); (6) addition of laboratory testing plan and rationale to daily progress notes; and (7) introduction of “harms of overtesting” cards in clinical areas to encourage discussion of potential harms of laboratory testing (ie, sleep interruptions, staff time, overdiagnosis, pain, and indirect and direct costs).
Hospital leaders could implement and adapt interventions as appropriate and reported whether they had implemented specific components of the bundle on 2 separate surveys, distributed once 4 months into their intervention period and again at the end of SOFI.
Education
All SOFI participants (physicians, nurses, and pharmacists) were invited to join 5 live, recorded webinars throughout the project that provided in-depth reviews of the guideline, detailed explanations for SOFI interventions and metrics, and QI teaching. Slide decks were prepared for local leaders providing evidence behind the guideline and potential harms of excessive laboratory use.
Outcomes
The primary outcomes were (1) mean proportion of hours per hospital day with exclusive isotonic maintenance IVF use and (2) mean proportion of hospital days with laboratory values determined. The first outcome was aligned with the guideline and was calculated as hours per hospital day during which only isotonic maintenance IVF were used divided by hours per hospital day during which any maintenance IVF were used.
Exclusive isotonic IVF use was ascertained through patient EHR reviews at each hospital with categorization of maintenance IVF use as hypotonic only; hypotonic and isotonic; isotonic only; or no maintenance IVF use on hospital days 2, 3, and 4. Those hospital days were selected to capture the majority of acute medical stays after initial fluid resuscitation and laboratory workup. Laboratory testing was measured by the proportion of hospital days for which a white blood cell count was obtained. This laboratory value was intentionally selected for its lack of direct relationship with IVF use to measure the impact of our laboratory reduction interventions. Serum sodium laboratory monitoring was monitored separately as a balancing measure.
Duration of IVF use was a secondary outcome, measured as the mean proportion of hours per hospital day with any maintenance IVF use. Daily patient weight measurement was recommended by the SOFI clinical algorithm as a noninvasive marker of overall fluid status and a clinical data point that could be followed over time instead of laboratory testing. The patient weight measurement was tracked as a process measure and calculated as the mean proportion of hospital days with maintenance IVF use on which a patient weight was recorded.
Given prior concerns about the lack of power in previous studies to detect adverse effects of isotonic IVF use,9 we defined balancing measures to monitor clinically important and unintended consequences of the interventions: (1) mean proportion of hospital days with maintenance IVF use on which a serum sodium test was obtained; and (2) proportions of hospitalizations in which 1 of 4 adverse events occurred that prompted a change in clinical management (intrahospital transfer to the intensive care unit, hypertension or edema requiring a diuretic, hypertension requiring antihypertensive medication, and acute kidney injury requiring renal replacement therapy). We predicted that adverse events would be rare and examined them using logistic regression models.
Sample Size
We calculated sample size accounting for concurrent parallel comparisons and repeated assessments in the stepped-wedge design. Assuming baseline exclusive isotonic maintenance IVF use of 50% based on existing literature16 with an SD of 25%, type 1 error rate of 5%, and within-cluster correlation of 0.10, 80 sites would provide 95% power to detect 10% or higher increase, and >99% power to detect 20% or higher increase. We aimed to recruit at least 80 hospitals.
Randomization
Random assignment to 1 of 3 wedges occurred in June 2019, after baseline data for the prior 12 months were submitted. Randomization was stratified by US geographic region (Midwest, Northeast, South, or West); number of beds (≤10, 11-30 or >30); and baseline mean exclusive isotonic IVF use (0%-50%; 51%-70%; ≥71%; or no data submitted).
Statistical Analysis
Descriptive statistics were used to summarize hospital characteristics. Baseline data collection was intentionally selected to include periods before and after the guideline publication. For the stepped-wedge analysis, we defined baseline as the 2 months immediately prior to SOFI launch (July and August 2019, after guideline publication), and compared this with 7 months of SOFI intervention period data (from September 2019 to March 2020). Linear mixed-effects regression models were constructed to examine the effect of the intervention on outcomes by comparing differences between SOFI-exposed and non–SOFI-exposed periods, within and across hospitals. We examined secular trends and potential effect modification with time by including a time period variable in the regression model. All mixed-effects models had a hospital-specific random intercept to account for clustering of clinician practices within hospitals. We adjusted P values to account for multiple comparisons such that the familywise error rate was controlled at 5% for all outcomes.
We observed that the guideline had an unexpectedly large and immediate effect on exclusive isotonic IVF use. We therefore performed an interrupted time series analysis with segmented regression21to show the guideline impact as distinct from the SOFI intervention impact and to disaggregate exclusive isotonic IVF use during the SOFI study period based on hospital exposure to the interventions. Analyses were performed using R Software, version 4.0.3 (R Project for Statistical Computing).22
Results
Of 115 hospitals that applied to participate in SOFI, 106 completed the onboarding requirements and were randomly assigned to 1 of 3 wedges. A total of 37 hospitals received the intervention in September 2019, 34 in November 2019, and 35 in January 2020. The SOFI trial was originally slated to end May 2020 to enable 9, 7, and 5 months of data collection after the intervention was implemented for wedges 1, 2, and 3, respectively. However, the COVID-19 pandemic sharply decreased inpatient volumes and disrupted hospital workflows such that we decided to shorten the follow-up period by 2 months (March 2020). After randomization, 6 hospitals withdrew, bringing the total number of hospitals included in analysis to 100 (94%). Across the entire study period, data from 11 939 hospitalizations were reviewed: 5215 before intervention and 6724 after intervention (Figure 2). Of 100 included hospitals, most (89) were medium size (11-30 beds) or large size (>30 beds) and had trainees providing direct patient care (75 hospitals). All hospitals had EHRs with computerized order entry (Table 1). The distribution of EHRs reviewed from hospitals of varying sizes remained consistent across the study periods; of 11 939 hospitalizations, 1039 (9%) were from hospitals with 10 or fewer acute care pediatric beds; 3952 (33%) were from hospitals with 11 to 30 beds; and 6948 (58%) were from hospitals with more than 30 beds.
Figure 2. Standardization of Fluids in Inpatient Settings (SOFI) Study Flow Diagram.
IVF represents intravenous fluids.
Table 1. Hospital Site and Wedge Characteristics in SOFI.
| Characteristic | Hospital sites, No. (% of total) | No. (%) of hospital sites | ||
|---|---|---|---|---|
| Wedge 1 | Wedge 2 | Wedge 3 | ||
| Total number of sites | 100 (100) | 34 | 32 | 34 |
| US Census region | ||||
| Midwest | 28 (28) | 9 (26) | 11 (34) | 8 (24) |
| Northeast | 15 (15) | 4 (12) | 5 (16) | 6 (18) |
| South | 39 (39) | 14 (41) | 10 (31) | 15 (44) |
| West | 18 (18) | 7 (21) | 6 (19) | 5 (15) |
| No. of acute care pediatric bedsa | ||||
| ≤10 | 11 (11) | 6 (18) | 2 (6) | 3 (9) |
| 11-30 | 33 (33) | 8 (24) | 11 (34) | 14 (41) |
| >30 | 56 (56) | 20 (59) | 19 (59) | 17 (50) |
| Mean exclusive isotonic IVF use, baseline period, %b | ||||
| No data | 38 (38) | 9 (26) | 15 (47) | 14 (41) |
| 0-50 | 25 (25) | 8 (24) | 8 (25) | 9 (26) |
| 51-70 | 13 (13) | 7 (21) | 4 (13) | 2 (6) |
| 71-100 | 24 (24) | 10 (29) | 5 (16) | 9 (26) |
| Hospital type | ||||
| Freestanding children’s hospital | 27 (27) | 10 (29) | 8 (25) | 9 (26) |
| Children’s hospital within a larger hospital | 39 (39) | 12 (35) | 16 (50) | 11 (32) |
| General hospital | 34 (34) | 12 (35) | 8 (25) | 14 (41) |
| Location | ||||
| Rural | 3 (3) | 1 (3) | 1 (3) | 1 (3) |
| Suburban | 31 (31) | 10 (29) | 9 (26) | 12 (35) |
| Urban | 66 (66) | 23 (68) | 22 (69) | 21 (62) |
| Majority of direct patient care provided by traineesc | 75 (75) | 25 (74) | 26 (81) | 24 (71) |
| Electronic health records present | 100 (100) | 34 (100) | 32 (100) | 34 (100) |
| Pediatric intensive care unit present | 58 (58) | 20 (59) | 23 (72) | 15 (44) |
Abbreviations: IVF, intravenous fluids; SOFI, Standardization of Fluids in Inpatient Settings.
Defined as medical and surgical pediatric beds, including observation beds and excluding intensive care unit and well-infant nursery.
Stratified at the time of randomization (June 2019).
Trainees defined as medical students, residents, or fellows.
Primary Outcomes
Before exposure to SOFI, the mean (SD) proportion of hours per day with exclusive isotonic maintenance IVF use was 88.5% (31.7%), exceeding the project goal of at least 80%. Exposure to SOFI led to an absolute increase of 5.4% above that in exclusive isotonic maintenance IVF use (95% CI, 3.9%-6.9%; P < .001). However, exposure to SOFI did not change the mean (SD) proportion of hospital days (before intervention, 12.5% [27.3%]; after intervention 11.4% [26.1%]) during which a laboratory value was obtained (estimated difference, 0.1%; 95% CI, –1.5% to 1.7%; P = .90) (Table 2).
Table 2. Primary and Secondary Outcomes Based on Exposure of Study Site to SOFI interventions.
| Outcome | Mean (SD) | Adjusted mean difference, β (95% CI)a | |
|---|---|---|---|
| Sites not exposed to SOFI intervention | Sites exposed to SOFI intervention | ||
| Total number of electronic health records | 5215 | 6724 | NA |
| Primary outcomes | |||
| Exclusive isotonic IVF use, mean % of hours per day | 88.5 (31.7) | 95.5 (20.5) | 5.4 (3.9 to 6.9) |
| Serum WBC count, mean % of hospital days | 12.5 (27.3) | 11.4 (26.1) | 0.10 (–1.5 to 1.7) |
| Secondary outcomes | |||
| Mean % of hours receiving any maintenance IVF per hospital day | 66.3 (28.1) | 63.9 (28.3) | –1.2 (–2.9 to 0.4) |
| Mean % of hospital days receiving maintenance IVF with a weight measurement | 16.2 (31.3) | 19.8 (33.2) | 4.4 (2.6 to 6.2) |
| Mean % of hospital days receiving maintenance IVF with a serum sodium laboratory value determined | 15.0 (30.0) | 14.2 (28.9) | –0.7 (–2.5 to 1.1) |
Abbreviations: IVF, intravenous fluids; NA, not applicable; SOFI, Standardization of Fluids in Inpatient Settings; WBC, white blood cell.
Calculated using linear mixed-effects regression models with SOFI exposure status and a time period variable; hospital-specific random intercept was included to account for within-hospital clustering.
Secondary Outcomes and Balancing Measures
The duration of maintenance IVF use did not change (adjusted mean difference, –1.2%; 95% CI, –2.9% to 0.4%). There was an absolute increase of 4.4% in the mean proportion of hospital days with a patient weight measurement while receiving maintenance IVF (95% CI, 2.6%-6.2%; Table 2). Exposure to SOFI did not lead to an increase in sodium monitoring (adjusted mean difference, –0.7%; 95% CI, –2.5% to 1.1%) while receiving maintenance IVF nor did it have a significant effect on adverse events (eTable 1 in Supplement 2).
Guideline Publication Effect
In the month after guideline publication in late November 2018, there was a 24.0% increase in the proportion of hours per hospital day with exclusive isotonic maintenance IVF use (95% CI, 22.3%-26.2%; P < .001). The increasing trend in exclusive isotonic maintenance IVF use was sustained throughout the period after guideline publication, including after SOFI implementation (Figure 3A). During the SOFI study period, SOFI-exposed sites had higher rates of exclusive isotonic maintenance IVF use and decreased practice variation compared with unexposed sites (Figure 3B).
Figures 3. Interrupted Time Series Analyses of Exclusive Use of Isotonic Maintenance Intravenous Fluids (IVF).
A, Exclusive isotonic maintenance IVF use before and after publication of the American Academy of Pediatrics (AAP) guideline. The first dotted vertical line indicates AAP guideline publication in late November 2018; the second dotted vertical line, beginning of the Standardization of Fluids in Inpatient Settings (SOFI) intervention period (September 2019). B, Exclusive isotonic maintenance IVF use for the SOFI study period disaggregated between hospitals exposed and unexposed to the intervention. Vertical lines indicate sequential launch dates by wedge.
Intervention Implementation
Surveys were completed by 95 hospitals during month 4 of the intervention period and 93 hospitals at the end of SOFI. Six hospitals completed only 1 survey. In aggregate, hospitals reported higher uptake of interventions related to isotonic IVF (range, 66%-74%) compared with interventions to reduce laboratory testing (range, 15%-40%; eTable 2 in Supplement 2).
Discussion
The interventions in this stepped-wedge, cluster randomized clinical trial led to a significant increase in the use of exclusive isotonic maintenance IVF at 100 pediatric hospitals of varying size, type, and organizational structure across the United States. We did not find increases in serious adverse events or sodium monitoring, whereas measurement of patient weight as a noninvasive marker of fluid status increased. These results support the routine use of isotonic fluids in medical and surgical inpatients as delineated by the AAP clinical practice guideline and address an important gap in the literature on maintenance IVF use before and after guideline publication.
The absolute increase of 5.4% in exclusive isotonic IVF use is on par with effect sizes found in prior systematic reviews of implementation science19,23 and is clinically important given how common IVF use is in the inpatient setting. Using the denominator of 5215 patient EHRs from hospitals that were unexposed to SOFI and a mean of 33 hours of total isotonic maintenance IVF use across hospital days 2 to 4, a 5.4% increase would translate to an increase of more than 9000 patient-hours of exclusive isotonic IVF use above baseline. The hospitals exposed to SOFI showed improved exclusive isotonic maintenance IVF use and less practice variation compared with hospitals that were unexposed to SOFI. The observed effects may have been limited by a ceiling effect.
We did not expect to find as large and sustained an association between guideline publication and increased exclusive isotonic maintenance IVF use. This finding contrasts with numerous studies describing long delays in uptake of evidence-based practices and the consensus that guideline publication is unlikely to be an effective strategy for changing clinician behavior.18,20 For example, in 1999, Cabana et al17 investigated barriers to physician guideline adherence and concluded that 2 of the most reported barriers were absence of awareness and familiarity with guidelines. Since then, there have been several randomized clinical trials and systematic reviews assessing the efficacy of numerous types of interventions on guideline uptake by physicians that generally reinforce the limitations of guideline publication on clinician behavior.19,23,24,25,26,27 It is possible that the simplicity and focus of the guideline, which had 1 strong recommendation to use isotonic fluids, impacted the acceptability and speed with which behavior change occurred. It is also possible that our cohort of hospitals was better prepared for prompt guideline uptake.
Our second major aim to decrease laboratory testing by 20% was not achieved. Baseline rates of laboratory testing in our sample were low (<15%), and the guideline made no recommendation regarding laboratory testing, which may have impacted the perceived importance and relevance of this aim. The SOFI participants reported lower uptake of laboratory reduction interventions than isotonic IVF interventions. Our findings reinforce the literature reporting the difficulties of deimplementation.28,29 Prior QI interventions, mostly at single institutions,30,31 have shown success in reducing laboratory testing; however, these studies incorporate frequent education, reminders, structured rounds of discussions, auditing, and feedback, which were not feasible in SOFI.
Limitations
This study had limitations. First, participating hospitals were part of a national QI collaborative, with many sites reporting prior participation in other multisite QI initiatives, introducing potential selection bias. Thus, generalizability may be limited to hospitals with particular interest in implementing evidence-based guidelines, institutional QI experience, or support and resources to implement interventions. In future investigations, we plan to compare the performance of the group of hospitals used in the present study to a control group of hospitals that did not participate in SOFI. Second is the issue of contamination, in which sites may have implemented interventions prior to their SOFI intervention period. By allowing staggered online access to the intervention bundle by wedge, we attempted to limit this effect. Third, stepped-wedge, cluster randomized clinical trials are subject to confounding by secular trends and, in this case, by the guideline itself, which was published 9 months prior to SOFI and appeared to have a much larger effect on isotonic IVF use than anticipated. This was an unexpected limitation in our study for which we tried to account in post hoc analyses. Fourth, although the study team conducted webinars and disseminated resources to ensure accurate data collection, it is possible that EHR reviews were subject to inaccuracies and measurement error. The reviewers of EHRs at each hospital were asked to complete a quality assurance check for this reason to mitigate inconsistencies and inaccurate data entry. Fifth, the short duration of the trial did not allow for assessment of sustained changes beyond March 2020 or of the impact of the pandemic.
Conclusions
In conclusion, an intervention bundle consisting of clinical algorithms, order sets, and changes to IVF ordering and verification was associated with significant improvements in guideline-concordant use of maintenance IVF without a concomitant increase in adverse events or electrolyte testing. Next steps include further work on deimplementation strategies to reduce laboratory testing and sustain changes in clinician behavior.
Trial Protocol
eTable 1. Odds of Adverse Events Following SOFI Interventions
eTable 2. Hospital-Reported Implementation of Specific SOFI Interventions
eAppendix. SOFI Intervention Bundle
Data Sharing Statement
References
- 1.McNab S, Ware RS, Neville KA, et al. Isotonic versus hypotonic solutions for maintenance intravenous fluid administration in children. Cochrane Database Syst Rev. 2014;2014(12):CD009457. doi: 10.1002/14651858.CD009457.pub2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Moritz ML, Ayus JC. Maintenance intravenous fluids in acutely ill patients. N Engl J Med. 2015;373(14):1350-1360. doi: 10.1056/NEJMra1412877 [DOI] [PubMed] [Google Scholar]
- 3.Foster BA, Tom D, Hill V. Hypotonic versus isotonic fluids in hospitalized children: a systematic review and meta-analysis. J Pediatr. 2014;165(1):163-169.e2. doi: 10.1016/j.jpeds.2014.01.040 [DOI] [PubMed] [Google Scholar]
- 4.Choong K, Kho ME, Menon K, Bohn D. Hypotonic versus isotonic saline in hospitalised children: a systematic review. Arch Dis Child. 2006;91(10):828-835. doi: 10.1136/adc.2005.088690 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Friedman JN, Beck CE, DeGroot J, Geary DF, Sklansky DJ, Freedman SB. Comparison of isotonic and hypotonic intravenous maintenance fluids: a randomized clinical trial. JAMA Pediatr. 2015;169(5):445-451. doi: 10.1001/jamapediatrics.2014.3809 [DOI] [PubMed] [Google Scholar]
- 6.McNab S, Duke T, South M, et al. 140 mmol/L of Sodium versus 77 mmol/L of sodium in maintenance intravenous fluid therapy for children in hospital (PIMS): a randomised controlled double-blind trial. Lancet. 2015;385(9974):1190-1197. doi: 10.1016/S0140-6736(14)61459-8 [DOI] [PubMed] [Google Scholar]
- 7.Wang J, Xu E, Xiao Y. Isotonic versus hypotonic maintenance IV fluids in hospitalized children: a meta-analysis. Pediatrics. 2014;133(1):105-113. doi: 10.1542/peds.2013-2041 [DOI] [PubMed] [Google Scholar]
- 8.Neville KA, Verge CF, Rosenberg AR, O’Meara MW, Walker JL. Isotonic is better than hypotonic saline for intravenous rehydration of children with gastroenteritis: a prospective randomised study. Arch Dis Child. 2006;91(3):226-232. doi: 10.1136/adc.2005.084103 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Feld LG, Neuspiel DR, Foster BA, et al. ; Subcommittee on Fluid and Electrolyte Therapy . Clinical practice guideline: maintenance intravenous fluids in children. Pediatrics. 2018;142(6):e20183083. doi: 10.1542/peds.2018-3083 [DOI] [PubMed] [Google Scholar]
- 10.Shrank WH, Rogstad TL, Parekh N. Waste in the US health care system: estimated costs and potential for savings. JAMA. 2019;322(15):1501-1509. doi: 10.1001/jama.2019.13978 [DOI] [PubMed] [Google Scholar]
- 11.Zhi M, Ding EL, Theisen-Toupal J, Whelan J, Arnaout R. The landscape of inappropriate laboratory testing: a 15-year meta-analysis. PLoS One. 2013;8(11):e78962. doi: 10.1371/journal.pone.0078962 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Eaton KP, Levy K, Soong C, et al. Evidence-based guidelines to eliminate repetitive laboratory testing. JAMA Intern Med. 2017;177(12):1833-1839. doi: 10.1001/jamainternmed.2017.5152 [DOI] [PubMed] [Google Scholar]
- 13.Korenstein D, Husain S, Gennarelli RL, White C, Masciale JN, Roman BR. Impact of clinical specialty on attitudes regarding overuse of inpatient laboratory testing. J Hosp Med. 2018;13(12):844-847. doi: 10.12788/jhm.2978 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Koch CG, Li L, Sun Z, et al. Hospital-acquired anemia: prevalence, outcomes, and healthcare implications. J Hosp Med. 2013;8(9):506-512. doi: 10.1002/jhm.2061 [DOI] [PubMed] [Google Scholar]
- 15.Tchou MJ, Hall M, Shah SS, et al. ; Pediatric Research in Inpatient Settings (PRIS) Network . Patterns of electrolyte testing at children’s hospitals for common inpatient diagnoses. Pediatrics. 2019;144(1):e20181644. doi: 10.1542/peds.2018-1644 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Rooholamini SN, Clifton H, Haaland W, et al. Outcomes of a clinical pathway to standardize use of maintenance intravenous fluids. Hosp Pediatr. 2017;7(12):703-709. doi: 10.1542/hpeds.2017-0099 [DOI] [PubMed] [Google Scholar]
- 17.Cabana MD, Rand CS, Powe NR, et al. Why don’t physicians follow clinical practice guidelines? a framework for improvement. JAMA. 1999;282(15):1458-1465. doi: 10.1001/jama.282.15.1458 [DOI] [PubMed] [Google Scholar]
- 18.Grimshaw JM, Eccles MP, Walker AE, Thomas RE. Changing physicians’ behavior: what works and thoughts on getting more things to work. J Contin Educ Health Prof. 2002;22(4):237-243. doi: 10.1002/chp.1340220408 [DOI] [PubMed] [Google Scholar]
- 19.Grimshaw JM, Eccles MP, Lavis JN, Hill SJ, Squires JE. Knowledge translation of research findings. Implement Sci. 2012;7(1):50. doi: 10.1186/1748-5908-7-50 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Balas EA, Boren SA. Managing clinical knowledge for health care improvement. Yearb Med Inform. 2000;09(1):65-70. doi: 10.1055/s-0038-1637943 [DOI] [PubMed] [Google Scholar]
- 21.Bernal JL, Cummins S, Gasparrini A. Interrupted time series regression for the evaluation of public health interventions: a tutorial. Int J Epidemiol. 2017;46(1):348-355. doi: 10.1093/ije/dyw098 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.R Core Team . R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing; 2020. [Google Scholar]
- 23.Ivers N, Jamtvedt G, Flottorp S, et al. Audit and feedback: effects on professional practice and healthcare outcomes. Cochrane Database Syst Rev. 2012;2012(6):CD000259. doi: 10.1002/14651858.CD000259.pub3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Flodgren G, O’Brien MA, Parmelli E, Grimshaw JM. Local opinion leaders: effects on professional practice and healthcare outcomes. Cochrane Database Syst Rev. 2019;6(6):CD000125. doi: 10.1002/14651858.CD000125.pub5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Shojania KG, Jennings A, Mayhew A, Ramsay C, Eccles M, Grimshaw J. Effect of point-of-care computer reminders on physician behaviour: a systematic review. CMAJ. 2010;182(5):E216-E225. doi: 10.1503/cmaj.090578 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Squires JE, Sullivan K, Eccles MP, Worswick J, Grimshaw JM. Are multifaceted interventions more effective than single-component interventions in changing health-care professionals’ behaviours? an overview of systematic reviews. Implement Sci. 2014;9(1):152. doi: 10.1186/s13012-014-0152-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Moja L, Polo Friz H, Capobussi M, et al. Effectiveness of a hospital-based computerized decision support system on clinician recommendations and patient outcomes: a randomized clinical trial. JAMA Netw Open. 2019;2(12):e1917094. doi: 10.1001/jamanetworkopen.2019.17094 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Coe M, Gruhler H, Schefft M, et al. Learning from each other: a multisite collaborative to reduce electrolyte testing. Pediatr Qual Saf. 2020;5(6):e351. doi: 10.1097/pq9.0000000000000351 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Niven DJ, Mrklas KJ, Holodinsky JK, et al. Towards understanding the de-adoption of low-value clinical practices: a scoping review. BMC Med. 2015;13(1):255. doi: 10.1186/s12916-015-0488-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Tchou MJ, Tang Girdwood S, Wormser B, et al. Reducing electrolyte testing in hospitalized children by using quality improvement methods. Pediatrics. 2018;141(5):e20173187. doi: 10.1542/peds.2017-3187 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Johnson DP, Lind C, Parker SES, et al. Toward high-value care: a quality improvement initiative to reduce unnecessary repeat complete blood counts and basic metabolic panels on a pediatric hospitalist service. Hosp Pediatr. 2016;6(1):1-8. doi: 10.1542/hpeds.2015-0099 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Trial Protocol
eTable 1. Odds of Adverse Events Following SOFI Interventions
eTable 2. Hospital-Reported Implementation of Specific SOFI Interventions
eAppendix. SOFI Intervention Bundle
Data Sharing Statement