High-value care, tele-ICUs, and more

Research summaries from ACP Hospitalist Weekly.


Residents may misjudge faculty, patient attitudes toward high-value care

Residents may underestimate the importance placed on high-value, cost-conscious care (HVCCC) by faculty and administrators and overestimate its importance to patients, according to a recent study from the Netherlands.

Researchers conducted a cross-sectional survey between June 2017 and December 2018 to measure and compare attitudes about HVCCC among residents, faculty members, administrators, and patients at 66 nonacademic hospitals. Respondents completed the Maastricht HVCCC Attitude Questionnaire (MHAQ), which has subscales for high-value care, cost incorporation, and perceived drawbacks. Residents also estimated the HVCCC attitudes of the other groups and answered questions about job demands and resources. The study results were published Nov. 2, 2020, by the Journal of General Internal Medicine.

Three hundred twelve residents, 305 faculty members, 53 administrators, and 1,049 patients responded to the survey. Of 42,825 total values, 3,348 (8%) were missing. Two hundred ninety-nine residents (96%), 297 faculty members (97%), and 53 administrators (100%) filled out all of the MHAQ items, while patients answered 88% (27,554 of 31,380), with the remaining 12% (3,826 of 31,380) missing at random. Attitudes toward HVCCC were found to differ on all three subscales of the MHAQ. Faculty and administrators had more positive attitudes than residents while patients had less positive attitudes (P≤0.05 for all comparisons). Residents underestimated faculty's attitudes about physicians' duties to incorporate costs in daily practice and overestimated their attitudes toward HVCCC's drawbacks, while underestimating patients' beliefs about the latter (P<0.001 for both comparisons). Residents and faculty were more likely to have positive HVCCC attitudes with increasing age (P≤0.05), while patients were less likely to have positive HVCCC attitudes if they had lower perceived health quality (P<0.001). Residents' attitudes toward HVCCC were more likely to be positive as their autonomy increased (P≤0.05).

The researchers noted that they could not report a response rate and that they didn't include all potential stakeholders, such as nurses, in their study, among other limitations. They concluded that attitudes toward HVCCC vary in the residency learning environment and that residents may misjudge the attitudes of both faculty and patients. “Residents may benefit from educators forthrightly addressing this variation, encouraging faculty and administrators to explicitly share and model their positive views, and providing empathic patient-centered strategies for communicating benefits of HVCCC,” the authors wrote.

Tele-ICU care linked to improved outcomes for nighttime ICU admissions

Implementation of tele-ICU care may improve outcomes among patients admitted to the ICU at night, a recent study found.

Researchers performed a retrospective pre-post analysis comparing risk-adjusted ICU mortality before and after tele-ICU implementation for daytime versus nighttime admissions at Westchester Medical Center in New York State. The tele-ICU service covered a burn ICU, a cardiothoracic surgery ICU, a cardiac critical care unit, a medical ICU, a neuroscience ICU, a surgical ICU, a trauma ICU, and three mixed medical-surgical community ICUs. The study's primary outcome included a comparison of risk-adjusted ICU mortality rates stratified by admission during the daytime shift (defined as 7 a.m. to 7 p.m.) versus the nighttime shift (7 p.m. to 7 a.m.). The Acute Physiology Age and Chronic Health Evaluation (APACHE) score, Version IVa methodology, was used to calculated patient acuity and risk-adjusted ICU mortality. The preimplementation period ranged from October 2014 to October 2015, while the postimplementation period began in January 2016 and ended in December 2018. The results of the study were published Oct. 27, 2020, by CHEST.

Image by Getty Images
Image by Getty Images

Overall, 1,581 and 14,584 ICU stays were available for analysis in the preimplementation and postimplementation periods, respectively. The patients' mean age was 64.3 years versus 63.7 years in the daytime admission group before and after tele-ICU implementation (P=0.37), respectively, while the mean age in the nighttime admission group was 63.4 versus 61.9, respectively (P=0.03). The average APACHE IVa score was 46.6 versus 54.8 in the daytime preimplementation versus postimplementation groups and 47.2 versus 56.3 in the respective nighttime groups (P<0.01 for both comparisons). Risk-adjusted ICU mortality overall was 8.7% before the tele-ICU was implemented and 6.5% afterward (P<0.01). No difference was seen in risk-adjusted ICU mortality before and after in the daytime groups. In the nighttime group, risk-adjusted mortality was 10.8% before tele-ICU implementation versus 7.0% afterward (P<0.01). Standardized mortality ratio before and after tele-ICU implementation was 0.95 and 0.87 in the daytime group versus 1.30 and 0.84 in the nighttime group.

The researchers noted that their results could have been affected by the study's pre-post design and by unmeasured confounders, among other limitations. They concluded that in their study, risk-adjusted ICU mortality decreased with implementation of a tele-ICU, mostly due to increased physician interaction during the nighttime “off hours,” and said that benefit from tele-ICUs will likely depend on where specific institutions have room for improvement. “Identifying ICU operational deficiencies, demonstrated by observed to predicted ratios >1, may be helpful in directing Tele-ICU driven process improvements,” the authors wrote. “Future research endeavors might include identification of additional Tele-ICU success predictors related to ventilator days [and] length of stay among other important outcomes. Identifying ICUs that might benefit from Tele-ICU augmentation may be accomplished by a priori analysis of observed to predicted outcome ratios prior to implementation.”

Excess days in acute care may improve performance measurement for hospitals

Measuring hospital performance by excess days in acute care (EDAC) rather than 30-day readmission rates may provide a more comprehensive assessment, according to a recent study.

CMS currently uses 30-day readmission rates to evaluate hospital quality in the Hospital Readmissions Reduction Program (HRRP). The EDAC measure, in contrast, captures all hospital encounters, including inpatient, ED, and observation stays, within 30 days of discharge. Researchers studied hospitals that participated in the HRRP in fiscal year 2019 to determine whether using the EDAC measure instead of 30-day readmissions would change hospitals' penalty status for three targeted conditions: heart failure, acute myocardial infarction (MI), and pneumonia. The study results were published Oct. 13, 2020, by Annals of Internal Medicine.

Overall, 3,173 hospitals that participated in HRRP in fiscal year 2019 were included in the study, reflecting discharges among Medicare fee-for-service beneficiaries ages 65 years and older from 2014 to 2017. Among all hospitals, the median readmission rate was 21.6% (range, 15.9% to 29.8%) for heart failure, 16.0% (range, 12.0% to 20.7%) for acute MI, and 16.7% (range, 12.5% to 23.3%) for pneumonia, while the median EDAC per 100 discharges for these conditions was 5.1 days (range, −60.1 to 143.4 days), 4.8 days (range, −59.0 to 174.3 days), and 6.3 days (range, −57.8 to 148.9 days), respectively.

Five hundred sixty-four hospitals were ranked in the top performance group for heart failure when the readmission measure was used, and of these, 280 (49.6%) were reclassified to a lower performance group when the EDAC measure was used. Similarly, 239 of the 571 hospitals in the worst performance group (41.9%) were reclassified to a higher performance group when the EDAC measure was used. Similar reclassification patterns were also seen for acute MI and for pneumonia. It was determined that the HRRP penalty status of 769 of 2,845 hospitals (27.0%) for heart failure, 581 of 2,055 hospitals (28.3%) for acute MI, and 724 of 2,911 hospitals (24.9%) for pneumonia would change if the EDAC measure were used to evaluate performance instead of the 30-day readmission measure. The distribution of penalties by hospital characteristics would also change, with fewer small hospitals and fewer rural hospitals being penalized.

The authors noted that the study looked only at conditions initially targeted by the HRRP, that only the EDAC point estimate, not its associated margin of error, was used to determine reclassification of penalty status, and that the EDAC measure does not address limitations such as inadequate adjustment for social risk. They concluded that “CMS should consider using the EDAC measure, which provides a more comprehensive picture of hospital use within 30 days of discharge than the readmission measure, to evaluate health care system performance under federal quality, reporting, and value-based programs.”

An accompanying editorial said that while the current study is “a step in the right direction,” it had limited ability to adjust for frailty, medical complexity, and social determinants of health and did not account for which rehospitalizations represent lapses in care. “Although the EDAC measure captures the use of acute care hospital resources, both it and the readmissions metric are poorly designed to measure quality of care and patient safety. The continued use of readmission measures as a reflection of quality of care in national rankings and pay-for-performance programs is empirically suspect and an ongoing source of frustration and confusion among health care leaders and researchers,” the editorialist wrote. “We can do better.”

Team structure affected resource use but not clinical outcomes

The structure of inpatient general medicine teams did not affect clinical outcomes in a retrospective cohort study.

Researchers used administrative data to compare clinical outcomes, cost, and resource utilization among resident, advanced practice clinician, and hospitalist teams on an inpatient medicine service at an academic medical center. The study included 12,716 adult patients discharged from the service between July 2015 and July 2018. The main measures were length of stay, 30-day readmissions, inpatient mortality, normalized total direct costs, discharge time, and consultation utilization. The researchers adjusted their multivariable analysis for time of admission, interhospital transfer, and comorbidities. Results were published online on Nov. 18, 2020, by the Journal of Hospital Medicine and appeared in the December 2020 issue.

There were no differences in length of stay, readmission, mortality, or cost between teams. Resident teams admitted fewer patients at night (32.0%; P<0.001) than advanced practice clinician (49.5%) and hospitalist (48.6%) teams did. Hospitalists discharged patients 26 minutes earlier than residents did (mean hours after midnight, 14.58 [95% CI, 14.44 to 14.72] vs. 15.02 [95% CI, 14.97 to 15.08]). Adjusted consult utilization was 15% higher for advanced practice clinicians (adjusted mean consults per admission, 1.00; 95% CI, 0.96 to 1.03) and 8% higher for residents (adjusted mean consults per admission, 0.93; 95% CI, 0.90 to 0.95) than for hospitalists (adjusted mean consults per admission, 0.85; 95% CI, 0.80 to 0.90).

The study was limited by its retrospective, nonrandomized design, and the center's inpatient admission process may have placed the burden on resident teams to perform all daytime admissions, inadvertently affecting outcomes, the authors noted. “It is possible the observed benefits of a solo hospitalist team are attributable to the lack of admitting duties rather than inherent advantages of the team structure,” they wrote. “If this were the case, we would expect similar benefits among [advanced practice clinician] teams, which we did not note.”

The results may help inform medical centers experiencing workforce shortages caused by a surge of COVID-19 patients, the authors said. “Our analysis suggests clinical outcomes are not significantly affected by inpatient team structure, and the addition of general medicine inpatient [advanced practice clinician] or hospitalist teams represent safe and efficient alternatives to traditional resident teams within an academic medical center,” they wrote.

Another study, published in the December 2020 The Joint Commission Journal on Quality and Patient Safety, also looked at inpatient teams. Researchers used the Safety Attitudes Questionnaire to assess perceptions about teamwork and collaboration, with a teamwork climate score, among 380 nurses, nurse assistants, hospitalists, and resident physicians working on general medical services in four mid-sized hospitals in the U.S.

Overall, hospitalists had the highest median teamwork climate score (83.3; interquartile range [IQR], 72.3 to 91.1), and nurses had the lowest (78.6; IQR, 69.6 to 87.5), but the difference was not significant (P=0.42). However, ratings of the quality of collaboration did significantly differ based on professional category, with 50 (63.3%) of 79 hospitalists rating the quality of collaboration with nurses as high or very high and 94 (48.7%) of 193 nurses rating the quality of collaboration with hospitalists as high or very high.

“[P]rior studies suggest that the largest discrepancies in perceptions of teamwork and collaboration exist between nurses and physicians,” the study authors noted. “Although we included only four sites, the current study represents the largest study to date characterizing teamwork and collaboration in the general medical hospital setting.”

Patient photographs in EHR reduced wrong-patient order entry errors

A quality improvement initiative to display patient photographs in the banner of the electronic health record (EHR) was associated with a decreased rate of wrong-patient order entry errors in a recent cohort study.

Researchers analyzed data collected from patients who visited the ED of a large tertiary academic hospital in Boston from July 2017 through June 2019. The patient photo feature was available and enabled when the EHR was installed and, prior to this study, a small number of practitioners uploaded patient photos with their consent. The quality improvement campaign, which started in July 2018, aimed to expand the capture of patient photos in the ED by educating ED registration staff on the importance of patient photographs for identification and safety, providing equipment that enhanced the photo-taking process (e.g., switching from desktop computers with webcams to mobile devices), and inviting patient participation through posters in the waiting area. The primary outcome was the rate of wrong-patient order entry errors, measured with the wrong-patient retract-and-reorder (RAR) measure. Results were published Nov. 11, 2020, by JAMA Network Open.

The primary analysis comprised 2,558,746 orders, of which 596,346 (23.3%) were placed while the patient's photo was displayed to the practitioner. These orders were placed for 71,851 unique patients (mean age, 49.2 years; 59.4% women). The overall rate of RAR events was 186 per 100,000 orders for the no-photo group and 133 per 100,000 orders for the photo group (unadjusted odds ratio [OR], 0.72; 95% CI, 0.57 to 0.89). After adjustment for multiple variables (e.g., patient race/ethnicity, practitioner type, emergency severity index), the reduced rate of RAR events remained consistent in the photo group compared to the no-photo group (OR, 0.57; 95% CI, 0.52 to 0.61). Patients with the highest level of acuity had significantly lower odds of wrong-patient orders but were also less likely to end up in the photo group. In addition, White ethnicity was associated with a lower rate of wrong-patient errors compared to Black ethnicity (OR, 0.91; 95% CI, 0.84 to 0.98).

Limitations of the study include its single-center, quasi-experimental design and the fact that the two groups had significant differences in terms of patient acuity, the study authors noted. They added that results may be different at other institutions, particularly those using a different EHR system.

The authors noted the relatively low costs of the intervention: less than $1,600 to purchase devices and accessories for the study and a projected $1,000 per year to maintain and replace equipment. “We speculate that the expected savings from improved safety, although not measured, likely far exceed the minimal costs of this program,” they wrote.