Data-Driven Health Economics Research Can Improve Clinical, Economic Outcomes

Lalan Wilfong, MD

With improvements in electronic medical records (EMRs), telemedicine, and other digital health solutions, clinicians have access to more data than ever to help inform decisions. New metrics such as patient volume, location, and details of treatment trajectory can help practices improve their care from a clinical standpoint, according to Lalan Wilfong, MD. But how best to leverage the economic influence of these data on clinical practice and patient outcomes remains an area of exploration.

“Practices [must] figure out how to better understand the economics of what they do, especially as the field moves to more total cost of care models and more valuebased care models, where we are more accountable for the cost we give [to the patient] and how we spend the [associated] revenue,” Wilfong said in an interview with OncologyLive^®.

Wilfong, a medical oncologist and hematologist with Texas Oncology and vice president for payer relations and practice transformation at McKesson Specialty Health, both in Dallas, Texas, said using data to retain patients from time of diagnosis through completion of therapy is a key factor in a practice’s financial stability. Data-driven health economics and outcomes research offers clinicians effective methods for improving quality of life and outcomes for patients and optimizing economic benefits for the practice.

“Quality metrics have influenced the way that we measure and perceive cancer care by focusing us more on what means most to patients, and how do we improve the outcomes that we provide vs just managing patients as we see is best,” Wilfong said. “Things [such as] metrics in the Oncology Care Model for reducing hospitalizations and reducing ED [emergency department] visits makes us think through the entire patient care continuum vs focusing on just what’s happening that day with a patient.”

Strategies for Improved Metrics

Health economics research has grown tremendously in community oncology over the past few years. Wilfong attributes that growth to the shift toward value-based care. “We had to show the benefit of the services we provided and really start thinking about the entire patient journey vs just individual data points, like we do in clinical research,” he said. “Health economic outcomes [helps us] understand risk profiling and what is the true value of the therapies we provide, not just monetarily, but also the patient’s quality of life and outcomes.”

The incorporation of machine learning and/or artificial intelligence in standard practice procedures may afford clinicians the chance to evaluate and optimize treatment economics. The SHIELD-RT study (NCT03775265) was among the first randomized prospective studies to implement machine learning as part of an interventional clinical application. The study used an EMR-based algorithm to identify patients at high risk of emergency department (ED) visits or hospitalizations during radiotherapy or chemotherapy.¹

The machine learning algorithm identified 311 patients with a greater than 10% risk of an acute care visit during radiotherapy at Duke University Hospital in Durham, North Carolina, between January 2019 and June 2019. Patients were randomized to receive a weekly evaluation during treatment (n = 157) or a twice-weekly evaluation during treatment (n = 154).¹

Overall, a reduction in acute care visits for the total population was observed from 22% to 12%. Patients evaluated twice weekly experienced fewer hospitalizations and ED visits compared with patients in the once-weekly arm: 29 vs 41 and 18 vs 33, respectively. From an economic perspective, the study authors noted that costs were reduced across revenue centers and that machine learning has shown potential to enhance quality of care in addition to keeping costs down. Specifically, the overall mean cost of acute care during radiotherapy was reduced by $2063 (95% CI, 4-4119; P = .03) when patients were evaluated twice weekly. The largest mean difference was observed in inpatient costs ($1815; 95% CI, –129 to 3760; P = .05).¹ Next steps include incorporating physician and intervention costs into the analysis.

Another potentially helpful data point clinicians may consider involves patient-reported outcomes (PROs), which can be added into the EMR and other digital health solutions to better inform treatment decisions. Data obtained at the patient level provides investigators with real-world insight that may help in understanding outcomes not observed in patients enrolled in clinical trials and assessing the unmet needs in preventing and treating adverse effects. Leveraging digital health solutions that can collect these data may improve the following:

• Communication between patients, providers, and their communities
• Patient and caregiver education
• Clinical assessment improvements for patient outcomes
• Self-monitoring education and uptake of self-management practices

Support of Real-World Analyses of Pros

In a qualitative review, Aapro et al sought to better understand the effect of PROs in oncology collected via electronic methods on routine supportive care in terms of clinical and health economic end points. In the analysis, investigators reviewed 66 studies that included data from 38 unique digital health systems used to collect PROs via remote monitoring. Of note, 21 of the systems provided patients with self-management symptom support tactics.²

Highlighting one of the pillars for improvement—communication—digital solutions that allow for self-reporting of symptoms can lead to a reduction in unexpected cost for the patient, greater use of hospital resources, and improved completion rates of treatment. For example, a single-center study cited in the analysis assessed outcomes of 766 patients who were randomized to receive email alert prompts to report symptoms experienced during their chemotherapy treatment at Memorial Sloan Kettering Cancer Center in New York, New York. Severe symptoms were reported by 63% of patients during the study and in response to email alerts of these symptoms, nurses initiated clinical action.³

The authors also cited a decrease in ED and hospital admissions when digital health solutions were used to monitor patient outcomes. Specifically, among those who received the alert (n = 441), 34% were admitted to the ED compared with 41% of patients who received standard care (n = 325; P = .02).²

Despite the observed benefits with PROs, the uptake of such solutions in practice present clinicians, patients, and supportive staff with additional hurdles. For example, integration of solutions in established systems takes time and adds an initial cost up front, not to mention that training and potential clinical trials evaluating efficacy may present delays in execution (TABLE).² “As far as improvement, I think it’s understanding all the disparate systems that we have to track, various economic metrics, and how we correlate those together into a unified platform,” Wilfong said.

Investigators of the analysis also supported that the effects of digital solutions on overall health care costs warrant further assessment. Wilfong noted that efforts are under way to assess the state of digital health systems at the practice level, but more efforts are needed to ensure that clinicians are making the most of data collected to influence financial change.

“For example, our practice management systems track new patients and revenue collection, our EMRs track the clinical activity around that, our pharmacy systems—which tend to be separate—track drug utilization, [and so on],” he said. “The biggest thing is figuring out what systems contain what data and [learning] how to piece that together and form a coherent model that drives financial outcomes.”

The data to determine the intensity of resources allocated to each patient depending on their diagnosis, treatment options, and comorbidities are not quite advanced enough yet, Wilfong said. He noted that, although this problem is complex and challenging, efforts are under way to better understand how to leverage large data sets, such as those available from Medicare, to achieve appropriate cost allocation on a per-patient basis.

COVID-19 Continues to Change Everything

Wilfong noted the profound effect the COVID-19 pandemic has had on how practices try to effectively provide care to patients, both from a quality of care and an economic standpoint. Clinicians and institutions are still working to fully quantify these changes, Wilfong said.

“[COVID-19 led to] a definite change in cancer screenings, which will translate to patients presenting with more advanced disease,” Wilfong said. “It is going to change [the] mix of staging [of our patients’ disease] that occurred because of the lack of ongoing provider visits and screenings.”

Although this presents the potential for cost increases because patients require more intensive treatment for their disease, the pandemic did provide some insight into improvements in care. “[What] the COVID-19 [pandemic] taught us was how to incorporate more digital health methods [and] resource allocation for those [is] very different from an economic standpoint,” Wilfong said.

In a survey conducted by the American Society for Radiation Oncology, investigators sought to quantify the effect of the pandemic on radiation oncology practices. Wakefield et al reported that telemedicine use was new to 89.2% of physician leaders who responded in April 2020 (n = 222). Additionally, telemedicine use for routine surveillance visits increased from 74.3% in April 2020 to 85.5% for responding physician leaders in February 2021 (n = 117). By February 2021, over half (53.8%) of responders were using telemedicine for new patient consultations.⁴

In terms of economic considerations, telemedicine has been shown to reduce cancer care costs for patients by offsetting the cost of transportation to the clinic, parking fees, lost time at work, and more.5 For oncology practices, the costs of implementing the infrastructure to support a telemedicine system (EMRs, updating computer systems, and so on) may be cumbersome initially.

However, Parikh et al used a time-driven, activity-based costing analysis to show that the transition to telemedicine saved a radiation oncology department an average of $586 per patient compared with the traditional workflow.⁶ The department saved an average of $347 for space and equipment and $239 in personnel costs.⁶ Several other cost analyses of the effect of telemedicine in oncology practice are maturing.

Community oncology practices should be aware that health care consumerism is also on the rise because of the pandemic, Wilfong said. Patients are beginning to question what they are getting out of their health care and how it can be improved, he added. Clinicians and institutions working at the community level must be prepared to adapt and continue to come up with innovative ways to ensure patients receive the best care from a clinical and consumer experience perspective in order to ensure continued economic viability, Wilfong said.

For example, a patient satisfaction survey from a large National Cancer Institute–designated cancer center by Natesan et al found that although overall patient satisfaction with telemedicine was high, different levels of satisfaction were observed across patient demographics. Patients born between 1981 and 1995 had higher satisfaction with telemedicine services (87%) compared with those born between 1965 and 1980 (77%) or 1946 and 1964 (74%).⁷

In another retrospective study of patient satisfaction with telemedicine during the pandemic, Swartz et al found that young adults with cancer (aged 18-39 years) had several suggestions for improving the experience. These included interpersonal communication, logistics, and addressing specific concerns with treatment, among others.⁸ Studies such as these underscore the fact that patients are demanding more from their care, especially after feeling the effects of the pandemic, and clinicians need to gather and apply as much as possible to better serve them.

“Understanding what data you have and understanding why [these] data are important can be 2 different questions,” Wilfong concluded. “Who are the stakeholders [who] need to see the data? How should those stakeholders see the data? You [must] understand what you’re able to collect and what importance [those] data may have depending on the incentives you’re trying to [provide].”

“[Access to these metrics] has led to significant changes in the way that we think and deliver cancer care. We need to think through better how we reduce the burden on reporting from practices, which is a significant issue for reporting quality metrics, while at the same time really focusing on those that [affect] outcomes the most.”

Room to Grow

In terms of the future of health economic outcomes, Wilfong said there is a long road ahead of clinicians in particularly when it comes to seamless integration into daily practice.

“Where I’d like to see us move to is better risk profiling, where instead of taking single tools or single bits of data and thinking about patient care, how do we incorporate a plethora of data points into what patients’ outcomes should be,” he said adding that this will include taking time to thoroughly consider a patient’s risk profile and address their concerns at the point of care. “[We need to take the] patients’ voice into our decision making much more robustly than we do so that we can take all of that data and really truly understand what the trajectory of that patient is.”

References

1. Natesan D, Thomas SM, Eisenstein E, et al. Impact of machine learning-directed on-treatment evaluations on cost of acute care visits: economic analysis of SHIELD-RT. J Clin Oncol. 2021;39(suppl 15):1509. doi:10.1200/JCO.2021.39.15_suppl.1509
2. Aapro M, Bossi P, Dasari A, et al. Digital health for optimal supportive care in oncology: benefits, limits, and future perspectives. Support Care Cancer. 2020;28(10):4589-4612. doi:10.1007/s00520-020-05539-1
3. Basch E, Deal AM, Kris MG, et al. Symptom monitoring with patient-reported outcomes during routine cancer treatment: a randomized controlled trial. J Clin Oncol. 2016;34(6):557-565. doi:10.1200/ JCO.2015.63.0830
4. Wakefield DV, Eichler T, Wilson E, Gardner L, Chollet-Lipscomb C, Schwartz DL. Variable effect of the COVID-19 pandemic on radiation oncology practices in the United States. Int J Radiat Oncol Biol Phys. Published online February 3, 2022. doi:10.1016/j. ijrobp.2022.01.045
5. Strowd RE, Dunbar EM, Gan HK, et al. Practical guidance for telemedicine use in neuro-oncology. Neurooncol Pract. 2022;9(2):91-104. doi:10.1093/nop/npac002
6. Parikh NR, Chang EM, Kishan AU, Kaprealian TB, Steinberg ML, Raldow AC. Time-driven activity-based costing analysis of telemedicine services in radiation oncology. Int J Radiat Oncol Biol Phys. 2020;108(2):430-434. doi:10.1016/j.ijrobp.2020.06.053
7. Natesan D, Niedzwiecki D, Oyekunle T, Emmons A, Zafar Y, Blitzblau R. Cancer patient satisfaction with telehealth: survey results from a large NCI-designated cancer institute. J Clin Oncol. 2021;39(suppl 15):1579. doi:10.1200/JCO.2021.39.15_suppl.1579
8. Swartz MC, Roth M, George G, et al. Satisfaction with and recommendations to improve telehealth visits among adolescents and young adults with cancer during the COVID-19 pandemic. J Clin Oncol. 2021;39(suppl 28):280. doi:10.1200/JCO.2020.39.28_suppl.280