The Case for Compassionate Analytics: 4 Steps for Using Data to Humanize the Delivery of Care

Data analytics has improved the quality of American healthcare by arming providers with the tools and information they need to be more proactive and targeted in their patient outreach. Data can preemptively notify care providers about patients’ health trajectories and lists of members with care gaps or that are at high risk of adverse events.

Data can also help identify social determinants of health (SDOH) like safe housing or access to food that can impact a person’s health. Only recently have SDOH been considered for tracking in healthcare settings. By tracking SDOH we’re able to measure the impact they can have on an individual patient level and the steps needed to mitigate those factors.

But for all the improved outcomes we can achieve using analytics, data can also make healthcare faceless and dehumanize patients. Healthcare data commonly tracks a patient’s condition, but rarely tells the complete story needed to optimally treat or influence the patient’s decisions outside the office. For example, a patient diagnosed with diabetes understands the need to manage their blood sugar. But without access to healthy food options or the money to afford their medication, managing their chronic disease ranks low on their hierarchy of needs. Medical algorithms track data about the patient, but often fail to describe the context around the patient that can make the output impactful or actionable. No one deserves to be just a box on someone else’s checklist.

The Unique Needs of Dual-Eligible Patients

This is particularly true for dual-eligible patients—those who are eligible for both Medicare and Medicaid. Dual-eligibles are typically the most complex patients because they often lack the resources to manage their care, whether financial or otherwise. They also have a deeper distrust of the healthcare system based on prior healthcare experiences. This is compounded by the fact that 78% of duals have two or more chronic conditions and 80% are more likely to have depression or a mental health diagnosis than a general Medicare patient.

The lack of compassion in the analytically driven healthcare machine can be disheartening, causing disengagement from the healthcare system and poorer results for those that need the most help. Rule-based decision-making and increasingly complex healthcare models remove even more empathy from care delivery, making it harder for care teams to explain to patients why or how a decision was reached.

It’s time to make healthcare data work for people rather than making people work for the data. That’s where compassionate analytics comes in—helping providers relate solutions to the individual and humanize the healthcare decision making process.

There are four steps organizations throughout the healthcare system should consider in creating a more compassionate analytically driven experience.

  1. Build or use more holistic member models

To understand the needs of the patient we need to build more holistic member models that use a variety of data sources, not just claims or EHR data. Understanding only the medical or pharmacological side of the patient story will miss the social indicators of the patient’s health. This is why it’s important to build in SDOH data, as well as administrative data from care managers and provider practices to show a more holistic picture of a patient. Once we have that full picture, we can allow models to tailor more specific outputs and recommended actions.

A great example of a holistic approach is found in Mr. Z’s story. Mr. Z is a dual-eligible patient with diabetes and hypertension and was occasionally a no-show at his doctor’s appointments. When he did get sick, he had a history of emergency room use but wasn’t what is described as a “frequent flyer,” and his medical risk score was below average for someone who is dual-eligible.

Mr. Z’s care management team performed an SDOH assessment and learned that Mr. Z prioritizes his time, energy, and money caring for his wife. With this information in hand, the risk algorithm added increased importance to the missed appointments and helped to predict that Mr. Z’s health was likely getting worse given his home situation. Based on this analysis, Mr. Z was moved to the top of the care team’s outreach list.

Upon outreach, the care team confirmed that Mr. Z’s health was in fact getting worse. A nurse care manager visited Mr. Z’s home to do a needs assessment and connected him with a community-based organization to provide healthy food delivery options that met the nutritional requirements and tastes of Mr. Z and his wife. The care manager also got him enrolled in a caregiver’s support group for increased topical social interaction. A pharmacist then worked to set up mail order medications and provide Mr. Z with education on the importance of medication adherence.

Since being assigned a care management team in 2022, Mr. Z has had no ER visits or admissions and his medical adherence has been nearly perfect. The data-driven approach, with layered SDOH triggers, helped provide Mr. Z with more stability in the management of his life, his health, and most importantly for him, his wife’s health.

Gaps in engagement are often the inflection points where patients become disenfranchised with the healthcare system. In traditional risk or predictable event modelling this member would not have been flagged for outreach until after and adverse event like the multiple avoidable ER visits and hospitalizations Mr. Z had in 2021. A holistic patient view allowed us to practice more proactive care with Mr. Z and better relate the importance of his own care to his real priority, taking care of his wife.

  1. Compassionate Analytics – Put faces to the models

As analysts and data consumers it’s imperative to look beyond personas because personas are based on averages—and patients are anything but average. When building or training new models at Belong Health, our analysts are asked if they would be willing to have their final product care for their own families. This puts a face behind every model we have. We often find that a data analyst who has a family member with a chronic illness can better understand the challenges of managing such an illness and identify ways to improve the care process for patients with similar conditions. However, it is common for healthcare analysts to not have a family member who is dual-eligible.

So, how do we develop a deeper understanding of a member beyond the data We share and tell their stories. Just like Mr. Z, there is more to the story than what the data shows us. His resources and energy were going into supporting his wife. And because of that, his health was declining. In order to help him become healthier, we needed to support both Mr. Z and his wife. And that is something a typical analyst can understand.

This compassionate analytics approach helps to demonstrate the humanity behind the model and encourages analysts to view patients as people with unique needs and challenges.

  1. Break down workflow silos

Healthcare analysts may collect and analyse data from many different sources, but they aren’t working alone and need to understand the business problems they are trying to solve. Analysts and the supporting engineering teams need to be highly connected with their clinical and operational counterparts.

During the initial stages of building more compassionate data and analytics programs it’s helpful to build deep personal relationships across teams to establish a mutual understanding of how they define success. Stakeholders need to be willing to invite people to observe their jobs, and analysts need to leave the comfortable distance a computer screen puts between them and their customers.

  1. Incorporate behavioural design into workflows

Incorporating behavioural design—the science of how people think, feel, and behave—into workflows is beneficial to understanding psychological drivers and the biases that can surface when interpreting results. Embracing behavioural design to make analytics more compassionate empowers the creation of consistent tools, while encouraging individual connections with patients. Additionally, the scientific method generates hypotheses that challenge our own assumptions with reliable data. In short, it can steer us clear of the dreaded sunk cost fallacy.

The goal of healthcare is to enable people to live their lives outside of healthcare, and compassionate healthcare is a journey that patients, care managers, and providers embark on together. Healthcare providers need to provide meaningful and effective care that keeps patients engaged in the process. Compassionate analytics play a critical role in establishing trust between patients and healthcare providers by humanizing the process and helping providers understand the value of patient-centric care. Following the steps outlined in this article, healthcare stakeholders can create a more compassionate experience that uses analytics to prioritize the needs of patients, improves health outcomes, and builds trust with patients.

By Mac Davis is Vice President of Digital Product & Data at Belong Health.

The post The Case for Compassionate Analytics: 4 Steps for Using Data to Humanize the Delivery of Care appeared first on .

0 replies

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

Your email address will not be published. Required fields are marked *