Web-based decision support tool may help determine safety of outpatient PE management
An internet-based electronic clinical decision support tool may be able to increase rates of safe outpatient management of pulmonary embolism (PE), a recent study found.
Researchers at Kaiser Permanente Northern California performed a controlled pragmatic trial at 21 community EDs to examine the effect of an integrated electronic clinical decision support system used for risk stratification and decision making in patients presenting with acute PE. The study period lasted 16 months, from January 2014 to April 2015. At month 9, 10 intervention sites received a technology and education intervention while the other 11 sites acted as concurrent controls. Intervention sites also had study champions who provided in-person education about the tool to ED and hospitalist physicians and continued to promote the tool throughout the study.
The clinical decision support system was added to the patient care workflow after the diagnosis of PE was confirmed but before site-of-care decision making was completed. The system reviewed eligibility criteria and listed, defined, and automatically populated the components of the PE Severity Index (PESI), which the treating physician then edited as needed. After the PESI score was calculated, the system provided the physician with the patient's risk class and tailored estimates of 30-day all-cause mortality, based on the literature. For low-risk patients, defined as classes I and II, the system said that “outpatient management is often possible.” For higher-risk patients, defined as classes III to V, the system noted that “inpatient care is often indicated.” The PESI was supplemented with an evidence-based list of relative contraindications to PE management.
The study's primary outcome was outpatient management immediately from the ED or after a short-term stay in an ED-based outpatient observation unit. Adverse outcomes, including return visits for PE-related symptoms within five days and recurrent venous thromboembolism, major hemorrhage, and all-cause mortality within 30 days, were also evaluated. The researchers used a differences-in-differences approach to compare preintervention and postintervention changes at the intervention and control sites while adjusting for demographic and clinical characteristics. Study results were published online Nov. 13, 2018, by Annals of Internal Medicine and appeared in the Dec. 18 issue.
There were 881 eligible patients with PE at intervention sites (421 in the preintervention period, 460 in the postintervention period) and 822 at control sites (410 in the preintervention period, 412 in the postintervention period). Patients at the intervention sites were slightly younger, were more likely to be African-American, and had less chronic lung disease and heart failure and lower PESI scores. Rates of adjusted home discharge increased significantly at the intervention sites (17.4% preintervention to 28.0% postintervention) but not the control sites (15.1% preintervention and 14.5% postintervention), a difference-in-difference comparison of 11.3 percentage points (95% CI, 3.0 to 19.5 percentage points; P=0.007). Rates of PE-related return visits at five days and major adverse outcomes at 30 days did not appear to increase with implementation of the clinical decision support system.
Site allocation was not randomized, and it is difficult to know which elements of the system most contributed to outpatient management, the authors said. They also pointed out that although the patient populations at the included medical centers were diverse, the geographic area and the practice setting were limited, possibly also limiting generalizability. However, they concluded that implementation and structured promotion of a clinical decision support system can help physicians make site-of-care decisions about PE management without increasing return visits within five days or major adverse outcomes at 30 days.
“Identifying the most appropriate venue of care for patients with acute medical conditions is a key priority for transforming U.S. health care,” the authors wrote. “The use of [clinical decision support systems] to bring validated risk-stratification tools to the ED bedside could help advance this agenda and could be expanded beyond PE to improve care and resource use for other clinical conditions.”
An accompanying editorial mentioned potential challenges in implementation of the tool, including that the health system provided fast, close follow-up after discharge that may not be available at all centers and that patients treated at home may not have time for the extensive education required regarding PE. They also noted that the intervention sites had champions of the intervention who promoted use of the clinical decision support system and that uptake might not have been as high otherwise.
However, patients with PE overwhelmingly prefer to be treated at home, and the article shows that home treatment can be safe in appropriately selected patients, the editorialists said. “The investigators are to be congratulated for showing the effectiveness of this computer tool that, when intensively promoted, enabled physicians to knowledgeably select a larger proportion of patients for home treatment,” they wrote.
Sacubitril-valsartan reduced NT-proBNP more than enalapril in acute heart failure
Patients who began taking sacubitril-valsartan during a hospitalization for acute decompensated heart failure had greater reductions in N-terminal pro B-type natriuretic peptide (NT-proBNP) than similar patients taking enalapril, an industry-funded study found.
The trial included 881 patients with heart failure with reduced ejection fraction who were hospitalized for acute decompensated heart failure at any of 129 sites in the U.S. After they were hemodynamically stable, the patients were randomized to either a target dose of sacubitril, 97 mg, plus valsartan, 103 mg, twice daily (n=440) or enalapril, 10 mg, twice daily (n=441). The study was funded by Novartis and published by the New England Journal of Medicine on Nov. 11, 2018; it appeared in the Feb. 7, 2019, issue.
The primary efficacy outcome was the time-averaged proportional change in NT-proBNP concentration from baseline through weeks 4 and 8, and it was significantly greater in the sacubitril-valsartan group than the enalapril group (−46.7% vs. −25.3%; ratio of change between groups, 0.71; 95% CI, 0.63 to 0.81; P<0.001). The difference between groups was evident at week 1 (ratio, 0.76; 95% CI, 0.69 to 0.85). The trial's key safety outcomes were worsening renal failure, hyperkalemia, symptomatic hypotension, and angioedema, and rates of these did not differ significantly between the groups.
The study authors noted that the effect of sacubitril-valsartan on NT-proBNP was accompanied by a reduction in high-sensitivity cardiac troponin T and that an exploratory analysis showed a lower rate of readmission for heart failure at eight weeks in patients taking the combination drug. These results extend the evidence base for this medication to patient populations in whom it had not previously been extensively studied, both because patients in the trial were hospitalized and because 36% of them were black, the authors said. Limitations of the trial include that its protocols may have prolonged length of stay and that data on NT-proBNP were missing for 15% of patients.
The trial had very specific requirements for initiation of the combination drug, including a systolic blood pressure of at least 100 mm Hg for the preceding six hours, with no use of IV vasodilators and no increase in IV diuretics, as well as no use of IV inotropes in the preceding 24 hours. Despite this and other precautions, about 20% of patients in each group had discontinued treatment by eight weeks, mostly because of adverse events, which suggests that “initiation of any neurohormonal agent in this population should be performed cautiously,” the authors said.
Care coordination bundle reduced total cost of care for Medicare and Medicaid patients
At one urban academic health system, an acute care intervention bundle to improve transition planning after discharge was associated with large reductions in total cost of care, especially among Medicaid patients, and with mixed results for hospital utilization, a nonrandomized quality improvement study found.
As part of the study, which spanned from 2012 through 2016, researchers analyzed the effect of the bundle on hospital utilization, practitioner follow-up visits, and total cost of care for Medicare and Medicaid patients. The acute care intervention, which was part of the Johns Hopkins Community Health Partnership (J-CHiP) program, included six evidence-based services: 1) early screening for discharge planning, 2) daily multidisciplinary unit-based rounds, 3) patient education, 4) enhanced medication management, 5) postdischarge follow-up by phone, and 6) skilled home care, remote monitoring, and/or a skilled nurse transition guide for high-risk patients. Most of these services were set up across 35 adult inpatient units at two Johns Hopkins hospitals.
Researchers compared outcomes of intervention participants in the 90 days after their acute episode with those of a propensity score-weighted preintervention group at J-CHiP hospitals and a concurrent comparison group drawn from similar Maryland hospitals. Results were published on Nov. 2, 2018, by JAMA Network Open.
The intervention group had 26,144 beneficiary episodes for Medicare patients (mean age, 68.4 years; 52.5% female). For Medicare patients, the intervention was associated with a $29.2 million reduction in aggregate total cost of care, largely driven by relative reductions in skilled nursing facility expenses. It was associated with increases in 90-day hospitalizations and 30-day readmissions of 11 and 14 per 1,000 beneficiary episodes, respectively. The intervention was associated with a reduction in practitioner follow-up visits of 41 and 29 per 1,000 beneficiary episodes for seven-day and 30-day visits, respectively.
Among 13,921 beneficiary episodes for Medicaid patients (mean age, 52.2 years; 53.1% female), the reduction in aggregate total cost of care was $59.8 million, and practitioner follow-up visits decreased by 70 and 182 per 1,000 episodes for seven-day and 30-day visits, respectively. While the 90-day ED visit rate decreased by 133 per 1,000 episodes among Medicaid patients, hospitalizations increased by 49 per 1,000 episodes.
The study authors noted limitations, such as a lack of data from non-CMS payers and the fact that Maryland hospitals in the study started participating in health care delivery reforms in 2014. They added that the study did not include the cost of the intervention itself or a formal cost-benefit analysis.
Since the intervention used a bundle of services, it is unclear which components were most responsible for the cost reductions, which is an important consideration in resource-poor settings, an accompanying editorial noted. “It is also not clear how the costs (or savings) of care management interventions are spread across a health system or over a patient's entire continuum of care—that is, to what extent costs are not saved, but simply shifted,” the editorialists wrote.
Compared to other analyses of care management that have found more modest savings, the study suggests that comprehensive inpatient care coordination can be effective, “even, and perhaps especially, in socioeconomically disadvantaged areas like East Baltimore, Maryland,” they added.
Fewer than 1 in 4 hospitals have both low-mortality and low-cost care for AMI, HF, and pneumonia
Hospitals' mortality rates for acute myocardial infarction (AMI), heart failure (HF), and pneumonia are inversely correlated with the payments they receive for treating these conditions, a recent study found.
This national cross-sectional study investigated the association between hospital-level 30-day risk-standardized mortality rates and 30-day risk-standardized payments for AMI, HF, and pneumonia. It included more than 4,000 hospitals and hundreds of thousands of fee-for-service Medicare beneficiaries hospitalized for any of the three conditions from July 1, 2011, through June 30, 2014. Results were published by JAMA Network Open on Oct. 19, 2018.
The median 30-day mortality rates for AMI, HF, and pneumonia were 14.3% (interquartile range [IQR], 13.8% to 14.8%), 11.7% (IQR, 11.0% to 12.5%), and 11.5% (IQR, 10.6% to 12.6%), respectively. The median 30-day payments were $21,620 (IQR, $20,966 to $22,567) for AMI, $15,139 (IQR, $14,310 to $16,118) for HF, and $14,220 (IQR, $13,342 to $15,097) for pneumonia.
There were statistically significant but weak inverse correlations between the mortality rates and payments (−0.08 for AMI, −0.21 for HF, and −0.07 for pneumonia). The correlations did not differ significantly by teaching status, safety-net status, urban/rural status, or the proportion of patients with low socioeconomic status. Fewer than a quarter of hospitals had both mortality rates and payments below the median.
The study's results suggest there are “substantial opportunities to improve the efficiency and value of care,” the authors said. They noted that this inverse correlation has been found in other studies as well. “Taken together, these results suggest that in some instances efforts to reduce use may come at the expense of worse clinical outcomes,” they wrote. The study also shows that hospitals can achieve both low mortality rates and low costs, regardless of their size, setting, and teaching or safety-net status, and that payment policies could integrate spending and outcome data without systematically disadvantaging particular categories of hospitals, according to the authors.
The findings highlight the complexity of the associations between hospitals' spending and mortality rates. “Specifically, these data would suggest that some hospitals have to increase spending, while others may be able to decrease spending to maintain or improve outcomes,” the authors said, adding that future research should identify specific strategies and interventions that improve the value and efficiency of care.
An accompanying commentary noted that the study raises several important questions, including why there isn't a stronger correlation between hospital spending and good outcomes and what can be done to incentivize hospitals “to become more efficient producers of health outcomes without either undergoing impossible structural changes (eg, moving away from a high-cost location) or causing unforeseen adverse social consequences (eg, systematically avoiding patients potentially likely to have high costs and/or worse outcomes).”
Review offers a framework for hospitalists on emergency preparedness
Hospitalists should take a lead role in their facilities' emergency preparedness planning, according to a recent paper that offered a framework for disaster preparedness in hospitals.
The review article, published in the October 2018 Journal of Hospital Medicine, provides an overview of disaster management principles for hospitalists, focusing particularly on mass casualty incidents that cause a rapid surge of patients with acute needs. It includes a framework for hospitalists' involvement in the preparation, response, and coordination of a hospital's response to such a disaster.
The paper includes seven targeted goals for hospitalists participating in a response, which were adapted from the Hospital Incident Command System (HICS):
- 1. Facilitate rapid intake of pending emergency admissions to medicine services
- 2. Facilitate rapid intake of incident patients requiring admission to medicine services
- 3. Offload noncritically ill patients from acute care services
- 4. Provide consultative medical management for incident patients on acute care services
- 5. Coordinate rapid discharge of patients from general and specialty medicine services
- 6. Assist with minor trauma overflow
- 7. Ensure that current census of patients continues to receive medical care
Hospitalist leaders should also prepare for emergency operations by creating Job Action Sheets (JASs), which codify the roles of everyone involved in an emergency response, the review said. It also recommended regular drills of emergency plans, including frequent small-scale drills.
“Hospitalists need to assume a leadership role in emergency preparedness to integrate seamlessly into hospital incident command structures and to shape the interdepartmental relationships vital to success—skills at which hospitalists excel. Although no plan can address all possible disasters, familiarity with HICS and well-prepared and well-written JASs should help groups respond and succeed in almost all hazards,” the paper said.
Estimates of ICU benefit differ among critical care physicians
Critical care physicians vary widely in their estimates of ICU benefit, according to a recent study.
Researchers performed a randomized study of critical care physicians in the United States to examine agreement about which types of patients benefit from ICU care and whether factors unrelated to illness severity affect estimates of benefit. Critical care physicians recruited for the study were asked to evaluate eight clinical vignettes of hypothetical patients and estimate whether the patients would benefit from ICU care. For each vignette, a single patient or hospital factor was randomized across physicians, with four factors related to severity of illness and four unrelated. The study's primary outcome was estimate of ICU benefit as assessed by a 4-point Likert-type scale. Results were published online Oct. 9, 2018, and in print in January 2019 by Critical Care Medicine.
Overall, 8,792 physicians were contacted about the study, and 1,223 agreed to participate. Of this group, 75% completed all eight clinical vignettes. It took physicians a median of 7 minutes to complete the vignettes (interquartile range, 5 to 11 minutes). Agreement about ICU benefit was poor, with a mean intra-class correlation coefficient of 0.06 (range, 0 to 0.18) for each vignette. Vignettes where illness severity was higher yielded a higher estimate of the benefit of ICU care.
Physicians were more likely to think ICU care was beneficial when one ICU bed was available versus when bed availability was not mentioned (absolute increase in ICU benefit, 7.3%) and were less likely to think ICU care was beneficial when family were said to be present versus when family presence was not mentioned (absolute decrease in benefit when a patient's wife was said to be present, 9.9%). Estimates of ICU benefit also appeared to be affected by patient age, with benefit assumed to be greater for older patients, but not by race and ethnicity.
The researchers noted that their recruitment rate of 14% was relatively low and that the clinical vignettes did not include an option for physicians to respond neutrally or recommend intermediate care, among other limitations. However, they concluded that physicians' clinical estimates of ICU benefit vary widely and are related to factors other than illness severity, which could lead to inappropriate ICU use. Guidelines recommending that decisions about ICU admission be based primarily on potential benefit are imprecise, the authors said.
“In this face of this uncertainty, critically ill patients may be harmed by inconsistent ICU admission decision-making,” they wrote. “There is a critical need for an empirical base of evidence identifying patients who benefit from ICU admission.”
A separate study, published Oct. 16, 2018, by Annals of Internal Medicine, looked at population trends in ICU admissions in the U.S. among Medicare beneficiaries from 2006 to 2015. The researchers used the Medicare Provider Analysis and Review file to evaluate all hospitalizations involving acute and ICU care and found that in claims from 88,402,008 hospitalizations, 14,787,690 (16.7%) were associated with ICU care during 289,391,446 person-years of coverage.
The ICU admission rate was 6,117 per 100,000 person-years in 2006 and 4,247 per 100,000 person-years in 2015 (P<0.001). Admission rates differed at the state level by almost threefold in 2015, ranging from 2,117 per 100,000 person-years in Hawaii to 6,312 per 100,000 person-years in Mississippi. National ICU bed count increased 11.4% between 2006 to 2015, but state-level changes ranged widely, from −38.1% in Rhode Island to 54.4% in Washington. The authors did not observe an association between percentage change in ICU beds and admissions across the states.
The data used for the study represented only those enrolled in Medicare, and the researchers did not adjust for patient-level factors. However, they concluded that ICU admission in the U.S. is not solely driven by bed availability and that this factor varies among states. “Institution of federal policies governing critical care workforce training, reimbursements for critical care services, and state-level approvals of certificates of need will thus require more local and granular data,” the authors wrote.
Simplified hand hygiene technique resulted in higher adherence than conventional method
Compared to the six-step technique for applying alcohol-based hand rub outlined in World Health Organization guidelines, a simplified three-step method was associated with better hand hygiene adherence and similar microbiological effectiveness, a recent study found.
From October through November 2015, researchers performed a cluster-randomized trial at a tertiary academic care center in Switzerland. They randomly assigned 12 wards to either the six-step method or a three-step method. The three steps are 1) covering all surfaces of the hands with an alcohol-based rub, 2) rotational rubbing of fingertips in the palm of the alternate hand, and 3) rotational rubbing of both thumbs. Results were published online on Nov. 3, 2018, by Clinical Infectious Diseases.
Overall, 2,923 hand hygiene indications were observed, and adherence to indications was 70.7%. On wards assigned to the three-step technique, adherence to indications was 75.9%, compared to 65.0% on wards assigned to the six-step technique (P<0.001). The difference remained significant after adjustment for imbalances in hand hygiene indications and for health care worker characteristics (e.g., wearing of finger rings), with an odds ratio for performing a hand hygiene action when indicated of 1.39 (95% CI, 1.00 to 1.93; P=0.049).
Adherence to hand hygiene technique was 51.7% (595/1,151) on wards assigned to the three-step technique, compared to 12.7% (116/915) on wards assigned to the six-step technique (P<0.001). The difference remained significant after adjustment for differences in hand hygiene indications and health care worker characteristics (odds ratio for fulfilling all steps of the assigned technique, 6.27; 95% CI, 3.52 to 11.17; P<0.001).
To assess effectiveness, researchers conducted microbiological assessments of hands (n=84 in each group) during the following indications for hand hygiene: after touching patient surroundings (60.7%), after touching a patient (38.1%), and after risk of exposure to body fluid (1.2%). In an analysis of 113 assessments that fulfilled requirements (56 from three-step wards and 57 from six-step wards), the reduction factor of bacterial colony counts did not differ between groups (P=0.629).
“The strengths of the proposed 3-step technique include its simplicity, thus enhancing its perceptibility and memorability, as well as its specific focus on those areas of the hands most involved in potential transmission of pathogens, namely thumbs and fingertips constituting ‘high-touch’ surfaces during direct patient-contact,” the study authors wrote.
They noted limitations, such as the study's single-center design and the hospital's longstanding history of hand-hygiene promotion, which may impede generalizability to other settings. Further trials in other clinical settings are needed to externally validate the results, they said.