番茄社区

A three-pronged, race-aware intervention dramatically closed the gap in treatment disparities between black and white patients with potentially curable non-small cell lung cancer (NSCLC), while at the same time improving outcomes for all.

Crude treatment rates before the intervention were 69% for black patients versus 78% for white patients (P<0.001), but rose to 96.5% versus 95%, respectively, after the intervention (P=0.56), Samuel Cykert, MD, of the University of North Carolina at Chapel Hill, and colleagues reported.

The adjusted odds ratio (OR) was 2.1 for the intervention's effect on treatment for black versus white patients (95% CI 0.41-10.4, P=0.39), they wrote in .

Intervention components in the study, which was conducted at five cancer centers, included a real-time warning system based on electronic health records (EHRs), race-specific feedback to clinical teams on their patients' treatment completion rates, and a nurse navigator who accessed the EHR data daily and acted as a point person engaging with patients throughout treatment.

"Application of this system-based, pragmatic approach to other cancer treatment disparities at a health system level could have positive effects on treatment completion, treatment equity, and overall outcomes," the authors wrote.

"With digital data readily available through EHRs it's very doable to build a system like this that signals disparities in care for both blacks and whites," Cykert told 番茄社区. "It's very feasible to have real-time actionable data and have one person responsible for the data without requiring extra hiring or bells and whistles."

From 2013 to 2016, the 5-year trial enrolled newly diagnosed stage I/II NSCLC patients ages 18 to 85. Left untreated, this malignancy is nearly always fatal within 4 years of diagnosis, the authors noted. No new hiring was necessary for the program since the cancer centers already used nurse navigators regularly, but navigators do need sound training in the more difficult aspects of communication, Cykert said.

"Since many doctors, especially surgeons, are not the most touchy-feely people, it's important to have navigators who can serially engage patients over time and give them information in manageable bits and bites," he said.

Nurse navigators were a pivotal component for preventing lapses in communication and care. When the warning system notified a nurse that a patient had missed an appointment or treatment milestone, the navigator then reached out to re-engage and bring the patient back in for treatment.

The most challenging part, Cykert added, is setting up the initial programming to make sure the electronic warning system is accurate. "Once you figure out how to use the EHRs, the system takes care of itself."

Nurses were encouraged to become familiar with patients and build trust in the case of a missed appointment, a miscommunication between doctor and patient, or other barriers to continuing care.

The reasons for treatment disparities go beyond socioeconomic status, age, and health status, the authors noted. A , for example, showed that clinician decision-making was a greater contributor to lower surgical rates for black patients than a patients refusal of surgery, while absolute surgical contraindications were the predominant factors that reduced surgeries for white patients. The study authors called for more research into the physician-patient encounter as a potential source of racial disparity in treatment.

In previous research, Cykert and colleagues identified multiple factors contributing to reduced treatment rates for black patients. "We found there was an implicit bias with many clinicians that made them less willing to take the same risks with patients that were different from them," Cykert said in a news release. "A black and a white patient of the same age could require the same surgery, have the same comorbidities, have the same income and insurance, yet white patients were more likely to receive the surgery and get their cancer treated."

In the current study, surgery rates increased following the intervention by 16% and 14% for black and white patients, respectively.

The trial patients were compared with included both retrospective and concurrent controls, and the primary outcome was receipt of curative treatment for early lung cancer. For the retrospective baseline, treatment difference by race was confirmed by a model adjusted for age, treatment site, cancer stage, gender, comorbid illness, and income, for an OR of 0.66 for black patients (95% CI 0.51-0.85, P=0.001).

At baseline, the mean ages in three arms were 68.8 for retrospective controls (n=2,841), 66.2 for the intervention group (n=360), and 69.5 for concurrent controls (n=597); black patients made up 15.9% of the retrospective group, 31.7% of the group consenting to intervention, and 13.2% of the concurrent control group.

In the final analysis, 144 black patients (48.3% women) and 246 white patients (42.7% women) completed the intervention. Median household income was $42,300 for black patients and $49,300 for whites.

So positive were the results that one of the study's participating institutions, Cone Health in Greensboro, North Carolina, is now working actively toward implementing this intervention as a permanent feature of its breast and lung cancer care, and Cykert is currently submitting a proposal to the National Cancer Institute to allow the intervention model to cover all types of cancer patients at treatment centers.

"Disparities in healthcare have been going on for decades and have been studied and measured for decades," Cykert told 番茄社区. "The fact is, we can now measure these disparities, design an intervention, and remeasure afterwards. That's a message that can be spread across other cancers and other chronic diseases to address disparities and promote health equity."

The main limitations of this trial were the differences in patient characteristics between the intervention and control groups. These were partially explained by the study's intentional oversampling of black patients for the intervention group. Furthermore, comorbidity scores in the retrospective group were obtained by diagnosis counts from cancer registries giving more crude estimates of disease counts and less specificity than the scores derived from direct clinical data in the other two groups.

The study did not adjust for health insurance coverage since the cohort was almost universally insured, with <5% of patients uninsured or on Medicaid. Finally, within-group analysis of the intervention group found no statistically significant racial difference in overall treatment, surgery, or stereotactic radiation -- results much different from the within-group comparisons in the retrospective control group.

Comment