Mortality by hospitalist age, NOACs in kidney disease, and more

Summaries from ACP Hospitalist Weekly.

Medicare patients had higher 30-day mortality when treated by older vs. younger hospitalists, study finds

Elderly patients treated by older hospitalists had higher mortality rates than those treated by younger hospitalists—except when the physicians cared for high volumes of patients, according to a recent observational study of U.S. hospitals.

Researchers assessed a 20% random sample of Medicare fee-for-service beneficiaries who were admitted between 2011 and 2014 and treated by hospitalists who were assigned based on scheduled work shifts. The main outcome measures were 30-day mortality, readmissions, and costs of care. Results were published online on May 16 by The BMJ.

A total of 736,537 admissions managed by 18,854 hospitalists (mean age, 42.9 years) were included in the analysis. Researchers adjusted for characteristics of patients and physicians, as well as hospital fixed effects, and found that patient mortality increased with physician age.

The overall 30-day mortality rate was 11.1%. Adjusted 30-day mortality rates were 10.8% (95% CI, 10.7% to 10.9%) for patients treated by physicians ages 39 years and younger, 11.1% (95% CI, 11.0% to 11.3%) for ages 40 to 49 years, 11.3% (95% CI, 11.1% to 11.5%) for ages 50 to 59 years, and 12.1% (95% CI, 11.6% to 12.5%) for ages 60 years and older. The association was driven primarily by physicians who treated low to medium numbers of patients, and higher patient volume seemed to provide a protective effect. For physicians with a high volume of patients (defined as more than 201 admissions per year), there was no association between their age and patient mortality.

For low-volume (fewer than 90 cases per year) and medium-volume (90 to 200 cases per year) physicians, each 10-year increase in age was associated with significantly increased 30-day patient mortality (adjusted odds ratios [ORs], 1.19 and 1.06, respectively; P<0.001 for both comparisons). However, among high-volume physicians, there was no association between physician age and patient mortality with a 10-year increase in age (adjusted OR, 1.01; P=0.29).

Readmissions did not vary with physician age, but costs of care were slightly higher in patients cared for by older physicians. Each 10-year increase in physician age was associated with a 2.4% increase in Medicare Part B spending (P<0.001). Similar patterns of results were found among general internists (both hospitalists and nonhospitalists) and in several sensitivity analyses.

The study authors noted limitations of their work, such as how results would be confounded if older physicians tend to treat patients at higher risk of death because of unmeasured factors. Also, due to the cross-sectional study design, they noted that they could not distinguish between the effects of physician age and cohort effects (i.e., changes in training) on clinical performance.

For patients treated by physicians younger than 40 years old, these findings amount to an 11% lower probability of dying, compared to those treated by physicians ages 60 years and older, the authors noted. They added that such an effect on mortality “is not only statistically significant but arguably clinically significant” and their study “suggests that continuing medical education of physicians could be important and that continual assessment of outcomes might be useful.”

An accompanying editorial noted the Institute of Medicine cites evidence that institutional resource priorities, processes, and culture create the possibility of good patient outcomes despite variation in education, training, and clinical performance of the hospital providers. As an example, the editorialist commented that there is large variation across hospitals in nursing workloads, educational qualifications, skill mix, and work environment, and that such differences are associated with mortality and other patient outcomes independent of physicians' qualifications. The editorialist added that “while this study suggests a performance advantage for younger physicians with more recent training when mortality is the outcome measured, more experienced physicians might do better on performance measures not included, particularly communication with patients and decisions informed by experience, leading perhaps to fewer invasive medical procedures at the end of life.”

Adherence to labeling for NOACs may be crucial in patients with kidney disease

When prescribed in a manner inconsistent with their labeling, non-vitamin K antagonist oral anticoagulants (NOACs) may be associated with harm rather than benefit in patients with severe kidney disease.

Researchers sought to investigate NOAC dosing and the associated risks of stroke and major bleeding in patients with atrial fibrillation in routine clinical practice. They looked at uses and outcomes of a standard NOAC dose in patients with a renal indication for dose reduction (potential overdosing), as well as the use and outcomes of a reduced dose in patients with no renal indication for dose reduction (potential underdosing).

Image by Thinkstock
Image by Thinkstock

A large U.S. administrative database was used to find 14,865 patients with atrial fibrillation who received apixaban, dabigatran, or rivaroxaban between October 2010 and September 2015. Patients were considered to have a renal indication for dose reduction if they were prescribed dabigatran and had an estimated glomerular filtration rate (eGFR) below 30 mL/min per 1.73 m2, if they were prescribed rivaroxaban and had an eGFR below 50 mL/min per 1.73 m2, or if they were prescribed apixaban and had a serum creatinine level of 1.5 mg/dL or greater.

The primary effectiveness outcome was ischemic stroke or systemic embolism, and the primary safety outcome was major bleeding. Results were published online June 5 by the Journal of the American College of Cardiology.

Among the 1,473 patients with a renal indication for dose reduction, 43.0% potentially received an overdose, which was associated with a higher risk of major bleeding (hazard ratio [HR], 2.19; 95% CI, 1.07 to 4.46) but no statistically significant difference in stroke (according to a pooled analysis of the three NOACs). Among the 13,392 patients with no renal indication for dose reduction, 13.3% potentially received an underdose, which, among apixaban-treated patients was associated with a higher risk of stroke (HR, 4.87; 95% CI, 1.30 to 18.26) but no statistically significant difference in major bleeding. There were no statistically significant relationships in dabigatran- or rivaroxaban-treated patients without a renal indication.

An editorial stated that older patients in this study were particularly prone to being underdosed with NOACs, despite larger trials finding that age alone is not a reason to decrease doses.

“Although the desire to avoid bleeding is a natural inclination, this study demonstrated that reduced dosing without a formal indication does not lead to improved outcomes and may be hazardous,” the editorial said. “Although less may be more in various aspects of life, less NOAC does not provide more to patients who should receive a standard dose, and underdosing may predispose patients to higher and fully avoidable risks.”

Nurses, but not family, found that quality of life program helped dying elderly patients

Care for dying elderly patients in the hospital may be improved by implementation of a program focusing on quality of life, but nurses are more likely to see improvement than patients' family members, according to a recent study.

Researchers in Belgium performed a cluster randomized controlled trial involving acute geriatrics wards at 10 hospitals from Oct. 1, 2012, to March 31, 2015. A random number generator assigned five hospitals to the intervention group and five to the control group. Hospitals in the intervention group began using the Care Programme for the Last Days of Life (CAREFuL), which included a care guide for last days of life, training, supportive documentation, and a guide for implementation, while hospitals in the control group provided usual care. The implementation guide involved nine components to be completed as well as a two-day training package intended to promote correct, compassionate use of the care guide, which was a document designed to guide health care staff in making choices when caring for dying patients.

Patients were included in the study if they (or their representative, in case of those unable to give informed consent) consented to use of their personal medical data, if they died on the wards during the study period, and if they had been hospitalized for more than 48 hours. The study's primary outcomes were comfort around dying, as measured by the End-of-Life in Dementia—Comfort Assessment in Dying (CAD-EOLD), where scores range from 14 to 42, and symptom management, as measured by the End-of-Life in Dementia—Symptom Management (SM-EOLD), where scores range from 0 to 27. Secondary outcomes were symptoms and care needs in patients' last three days of life as assessed by nurses using the Palliative Care Outcome Scale, where scores ranged from 0 to 40; family caregivers' satisfaction with care during the last 48 hours of life as assessed by the End-of-Life in Dementia—Satisfaction with Care scale; and symptom burden as assessed by nurses. All measures were completed by patients' nurses or family caretakers after the patients' deaths. Patients and families were blinded to intervention status, but hospital staff members were not. The study results were published online May 16 by The Lancet and appear in the July 8 issue.

Overall, 118 patients in the control group and 164 patients in the intervention group had outcomes assessed in the postintervention period. In the intervention group, 132 patients (80%) were assessed by nurses and 48 (29%) were assessed by family caretakers, while in the control group, 109 patients (92%) were assessed by nurses and 23 (19%) were assessed by family caretakers. Patient comfort as assessed by nurses was significantly better in the intervention group versus the control group (baseline-adjusted mean difference in CAD-EOLD, 4.30; P<0.0001). However, no significant differences were noted in comfort as assessed by family caretakers (baseline-adjusted mean difference in CAD-EOLD, −0.62; P=0.82) or in symptom management as assessed by family caretakers (baseline-adjusted mean difference in SM-EOLD, −0.59; P=0.71) or by nurses (baseline-adjusted mean difference in SM-EOLD, −0.41; P=0.58). Nurses also noted significantly better symptoms and care needs in the intervention group, but satisfaction with care as assessed by family caregivers was significantly worse in the intervention group than in the control group.

The authors acknowledged that nurses were not blinded to study allocation and were aware that they were implementing the CAREFuL program. In addition, they noted that only 34% to 36% of patients gave consent to use their data in the trial and that not all patients were approached to participate in the study due to insufficient time, among other limitations. However, they concluded that CAREFuL led to significant improvements in nurse-assessed comfort around dying, symptoms, and care needs in patients' last days of life, although these benefits were not noted by patients' family members. While the positive results among nurses seem to support wider implementation of the program, further study of family members' and patients' perceptions are needed, they wrote.

The authors of an accompanying commentary noted concerns about the strengths of the study's findings due to the lack of blinding among the health care staff, the lack of information on the effect of the intervention on ward processes, and the low family response rate and level of satisfaction. They also pointed out that CAREFuL was based on an earlier program, the Liverpool Care Pathway for the Dying Patient (LCP), that ended in the United Kingdom in 2014 because both the public and the government were concerned about its implementation.

“The results of the CAREFuL study are welcome as an important step, but crucial questions remain unanswered about the study itself and the contextual and implementation issues that the experience with LCP revealed,” the commentary authors wrote. “We encourage future research—on both CAREFuL and other end-of-life pathways—to narrow the evidence gap we face in comforting patients dying in hospital and their families.”