3 Wishes, penalties, and more

Summaries from ACP Hospitalist Weekly.


Project honoring ICU patients' dying wishes was affordable, sustainable

The 3 Wishes Project, a program that honors patients' final wishes by creating patient- and family-centered memories, was successfully implemented across multiple ICUs, a qualitative study found.

A clinician or a project team member invites the patient and family to participate in the program after the decision has been made to withdraw life-sustaining technology or if the estimated probability of death in the near future is above 95%. Participants then discuss the patient and his preferences and values at the end of life before requesting small, joyful wishes. Researchers assessed the value, transferability, affordability, and sustainability of the intervention, which was implemented at four ICUs in the U.S. and Canada. The intervention was not strictly protocolized, although implementation was guided by the original program, which launched in January 2013 in Ontario. Results were published online on Nov. 12, 2019, by Annals of Internal Medicine and appeared in the Jan. 7 issue.

Photo by Getty Images
Photo by Getty Images

A total of 730 patients (mean age, 63.1 years; 44.9% women) enrolled, and 3,325 wishes were fulfilled. Most had more than three wishes implemented, had medical diagnoses (90.0%), and were admitted to the ICU from the ED (41.5%). Overall, 23.5% of wishes were implemented by family or friends, 55.9% by clinicians, and 28.0% by the 3 Wishes Project team (wishes could be fulfilled by several people). Examples of wishes included unlimited visiting hours/number of visitors, taking the patient outside, live music, wedding or renewal of vows, snacks and beverages for the family, and pet visitation. After minimal investment for reusable materials (e.g., flameless candles, music player/speakers, coffeemaker), the average cost was $5.19 per wish, and 2,553 (76.8%) wishes had no cost. Continuing costs were mainly for supplies, and some items were donated.

Between 2017 and 2019, 167 participants participated in 80 semistructured interviews and 14 focus groups. Of 75 family members, 29 (38.6%) were spouses and 22 (29.3%) were children. Of 72 clinicians, 54 (75.0%) were nurses and nine (12.5%) were physicians. The 20 managers and administrators ranged from an ICU director to a hospital vice president. Participants said they perceived value in the way the program encourages human connection and brings the patient's identity into focus. Family members reported being comforted, and clinicians described calls from families who had participated in the program and were now requesting it for another family member on a different hospital ward. Clinicians and managers said that having enthusiastic champions, prioritizing improved end-of-life care, and having a collaborative unit culture facilitated the establishment of the program. Each site continued the program after study completion.

Limitations of the study include the inability to obtain perspectives of dying patients and the fact that all participating sites were closed ICUs in academic tertiary care centers, potentially limiting generalizability to other settings, the authors noted. “Although 3 of 4 centers are funded by a single-payer public health care system, concordance of the main results in the United States and Canada show our shared humanity and the commonality of the end-of-life experience,” they added.

The U.S. project site, a medical ICU at Ronald Reagan UCLA (University of California, Los Angeles) Medical Center, was the focus of a Success Story in the August 2019 ACP Hospitalist.

Hospital-acquired condition penalty program did not reduce conditions it targeted

The Hospital-Acquired Condition Reduction Program (HACRP), which in recent years has penalized the quartile of hospitals with the highest rates of hospital-acquired conditions, did not reduce the conditions it targeted beyond existing trends in Michigan, a recent study found.

Researchers used clinical registry data from 73 Michigan hospitals that participated in a statewide surgical quality collaborative to assess the association between implementation of the program and the incidence of risk-adjusted hospital-acquired conditions. They assessed rates of postoperative hospital-acquired conditions targeted by the program from 2010 to 2018 to estimate the impact of the policy since it was first announced by CMS in August 2013. (Although financial penalties under the program began in October 2014, hospitals had already begun to anticipate the changes.) The first year of the program targeted central line-associated bloodstream infections and catheter-associated urinary tract infections, and its second year expanded to include surgical-site infections related to colon surgery and hysterectomy. Results were published in the November 2019 Health Affairs.

Patient age, demographics, and comorbid disease burden were similar before and after the program was implemented. Patients experienced hospital-acquired conditions at a rate of 133.4 per 1,000 discharges (95% CI, 126.6 to 140.1) in the period before the program and 122.2 per 1,000 discharges (95% CI, 116.8 to 127.6) in the period after it was announced, for a pre-post difference of −11.2 per 1,000 discharges (P<0.01). Greater improvements were observed for nontargeted conditions, which decreased from 71.6 to 63.5 per 1,000 discharges (pre-post difference, −8.1 per 1,000 discharges; P<0.01), compared to targeted conditions, which decreased from 61.7 to 58.7 per 1,000 discharges (pre-post difference, −3.1 per 1,000 discharges; P<0.05). While there were no significant changes in the rates of hospital-acquired conditions during the pre-period, the adjusted annual slope changed by −6.2 per 1,000 discharges annually (95% CI, −9.2 to −3.3; P<0.01) in the post-period. The program, however, was not associated with a significant change in the adjusted annual slope of all hospital-acquired conditions (−4.3; 95% CI, −11.4 to 2.8) or for targeted conditions (0.1; 95% CI, −3.7 to 3.8).

Limitations of the study include the inability to compare outcomes against a set of controls, since all U.S. hospitals were subjected to penalties under the program, the authors noted. They added that the findings may not be generalizable to patients or hospitals in other states.

The findings raise questions as to whether national value-based payment programs are effective in reducing hospital-acquired conditions and whether penalties under the program should be continued, the authors concluded. “Given the fundamental challenges associated with the HACRP, its costs—including hospitals' administrative efforts and distraction from quality improvement activities with a higher yield—may well exceed its benefits,” they wrote.

HIV-related admissions declined from 1996 to 2016, but readmission rates were stable

Hospital admissions of patients with HIV have shown a continued decline, but readmission rates and racial disparities are persisting, according to a recent analysis of data from 1996 to 2016.

The study included 4,323 patients in the University of North Carolina Center for AIDS Research HIV Clinical Cohort who received clinical care in 1996 to 2016 (29% women, 60% black). Their 30,007 person-years were used to estimate hospitalization, readmission, and inpatient mortality rates. Results were published Oct. 22, 2019, by Clinical Infectious Diseases.

The hospitalization rate over the entire period was 34.3 per 100 person-years (95% CI, 32.4 to 36.4) with a mean change of −3% per year (95% CI, −4% to −2%). The 30-day readmission rate was 18.9% (95% CI, 17.7% to 20.2%) and stable over the study period. Risk of hospitalization was higher in patients who were black (compared to white), were older, had HIV RNA >400 copies/mL, or had CD4 count <200 cells/μL (all P<0.05). Inpatient mortality was higher with older age and lower CD4 count (both P<0.05). Readmissions were higher among black patients, those with detectable HIV RNA, and those with lower CD4 cell counts (all P<0.05).

The findings show the importance of early diagnosis and retention in care for patients with HIV, the study authors concluded. “Despite substantial decreases over the study period in all patient groups, disparities persist in rates of hospitalization with CD4<100 cells/μL, which were more frequent among Black and Hispanic patients compared to White patients, and among patients with detectable viral loads. These findings, consistent with prior studies, may reflect late diagnosis, challenges to care engagement, and unequal health care access,” the authors said.

Limitations of the study include lack of data on discharge diagnoses or risk factors such as substance use, mental health disorders, and socioeconomic status, the study authors said. An accompanying editorial also noted this weakness of the study. “Without these data it is hard to know where to intervene in order to mitigate ongoing HIV-related hospitalizations and the racial disparities that persist,” the editorial said.

The editorial also included reflections on one of its authors' recent experience treating HIV in inpatients. “If progress is happening, it's not happening fast enough nor uniformly everywhere,” it said. “Regardless of how far we've come in the HIV epidemic, patients hospitalized with HIV-related complications are a reflection of the remaining holes we have to patch in our public health programs before we can realistically end the epidemic for all.”

Hospitalist care associated with slightly higher costs, readmissions among patients discharged to SNFs

Medicare patients who were treated by hospitalists and then discharged to skilled nursing facilities (SNFs) were slightly more likely to be readmitted within 30 days and had slightly higher costs than patients whose inpatient care was handled by nonhospitalists, a study found.

The retrospective study looked at Medicare fee-for-service beneficiaries who were over 66 years of age and discharged from a hospital to a SNF in 2012 to 2014 (n=2,839,779) to compare outcomes between hospitalist and nonhospitalist inpatient care. All of the outcomes were adjusted for patients' demographic and clinical characteristics, and hospital fixed effects were included to account for heterogeneity across facilities. Results were published Oct. 21, 2019, by the Journal of General Internal Medicine.

The study found a slightly higher 30-day rehospitalization rate among patients treated by hospitalists: 17.59% versus 17.31% in nonhospitalists' patients (adjusted difference, 0.28%; 95% CI, 0.13 to 0.44). Medicare Part A and B payments within 60 days were also slightly higher with hospitalists: $26,301 versus $25,996 (adjusted difference, $305; 95% CI, $243 to $367). On the outcome of successful discharge to the community, which the study defined as still alive and not readmitted to a hospital or SNF at 30 days and discharged to the community within 100 days of SNF admission, hospitalist care was associated with an insignificantly lower rate (adjusted difference, −0.26%; 95% CI, −0.48 to −0.04). However, hospitalists' patients also had a lower mortality rate, although the difference was not significant (adjusted difference, −0.12%; 95% CI, −0.22 to −0.02).

“Our findings are consistent with studies that found higher utilization and costs for hospitalists' patients in the 30 days after discharge. However, we also observed a trend toward lower mortality in the hospitalist group, suggesting that there might be a trade off between utilization and short-term survival,” the study authors said. They noted, however, that the observational retrospective design limits the inferences that can be drawn regarding the causality of the study findings. Other limitations include that the latest data was from 2014, and so practices might have changed since, and that many patients are cared for by a mix of hospitalists and nonhospitalists, potentially diluting the differences.

The authors called for additional research that is prospective and looks at these outcomes in patients going to postacute care settings other than SNFs but offered some potential conclusions in the interim. “The findings suggest caution regarding expanding hospitalist services as a strategy to improve post-acute care outcomes while controlling utilization and spending,” they wrote.

Cost of heart failure, pneumonia hospitalizations varies by hospital more than by patient

Hospital costs appear to vary more by hospital than by patient, according to a recent study of Medicare patients who had been admitted to multiple hospitals with either heart failure or pneumonia.

This observational cohort study used CMS data on more than a million patients hospitalized with a principal diagnosis of heart failure or pneumonia between July 1, 2013, and June 30, 2016. The hospitals that treated them were classified into quartiles based on their Medicare payments for heart failure and pneumonia admissions. The study also focused specifically on patients who had been admitted for either heart failure (n=1,615) or pneumonia (n=708) at least twice, once to a lowest-quartile hospital and once to a highest-quartile hospital. Results were published by JAMA Network Open on Nov. 15, 2019.

In the overall analysis, 30-day mortality rates for these conditions were similar between the hospitals in the lowest quartile of payments and those in the highest. The median 30-day risk-standardized mortality rates were 8.1% (interquartile range, 7.7% to 8.5%) for heart failure and 11.3% (interquartile range, 10.7% to 12.1%) for pneumonia.

The analysis of patients hospitalized at facilities in both quartiles found that payments were significantly higher for stays at the high-payment hospitals: The difference in the 30-day episode payment was $2,118 (95% CI, $1,168 to $3,068; P<0.001) for heart failure and $2,907 (95% CI, $1,760 to $4,054; P<0.001) for pneumonia. More than half of the difference was attributed to the cost of the index hospitalization: differences of $1,425 (95% CI, $695 to $2,154; P<0.001) for heart failure and $1,659 (95% CI, $731 to $2,588; P<0.001) for pneumonia.

“By studying patients who were admitted twice for the same diagnosis at hospitals with different payment profiles, we were able to better isolate the association of the hospital and its ecosystem from other factors, such as social context,” the study authors said. The results are in accord with previous research showing significant variations in spending across hospitals, and this study “adds to the published work by demonstrating that the payment profile seemed unrelated to the patient, as the patient was held constant in this study,” they wrote.

The lower-cost hospitals could be seen as “achievable benchmarks” for higher-spending facilities, and the results should motivate investigation of new strategies to reduce costs, the authors said. Limitations of the study include that patients who are admitted to multiple hospitals are not representative of the general population with the same conditions. In addition, clinical details were not available on the patients, so the severity of their conditions may have differed during the admissions to different facilities.

Another study, published by the Journal of General Internal Medicine on Nov. 12, 2019, compared outcomes and costs between major teaching (n=339) and nonteaching hospitals (n=2,439) using matched pairs of patients admitted to each type: 43,999 pairs with acute myocardial infarction, 84,985 pairs with heart failure, and 74,947 pairs with pneumonia. It found that 30-day mortality was lower in teaching than nonteaching hospitals (10.7% vs. 12.0%, P<0.0001), as were readmissions and ICU utilization. Cost was $273 higher in the teaching hospitals' patients compared to their matches (P<0.0001).

In the quintile of highest-risk patients, there were larger differences in mortality (24.6% in teaching hospitals versus 27.6% in nonteaching) and cost ($1,289 more in teaching-hospital patients). Results for the individual conditions were similar to the combined results. The study authors concluded that, particularly for higher-risk patients, major teaching hospitals appear to provide good value for their higher costs. “The health system could use this information to promote a more optimal, efficient allocation of patients to academic hospitals, based on both their own needs and the differential capacity of facilities to address them,” they said.