Is what you’re asking for really what you need? Tips for getting to the root of the problem—and finding the right solution

As a lifelong data enthusiast and healthcare leader, one of my primary objectives is to help transform a reservoir of seemingly disconnected inputs and information into meaningful insights and strategic direction. Over the years, I’ve learned that the first conclusion we find isn’t necessarily the right one, and sometimes the solution we’re asking for isn’t really what we need to solve the problem. The process of finding diamonds in the data rough takes patience and persistence—but there is nothing more rewarding (to a data scientist, at least). 

For healthcare organizations, this is an important lesson. Identifying a solution quickly can feel hopeful, but if it’s not the right answer to the real issue, you’ll be starting again from square one. In this blog post, I’ll explain how to identify the root cause of a problem, share questions you need to ask to identify the best solution, and suggest tactics you can use to help the problem-solving process go smoothly and successfully.

To begin, let’s talk about what can happen when your problem-solving process is messy.

Taking a trial-and-error approach often leads you to the wrong problem or guides you to implement a mismatched solution. These missteps can be costly—especially to organizations with limited resources. Negative outcomes include:

• Wasting time and money: At best, pouring effort into an ineffective path takes you off the project course for a few days. At worst, it breaks your budget and compromises your operations.
• Adversely impacting team/organization morale: Investing significant time into something that ultimately gets tossed out never feels good. Staff are often frustrated by the lack of progress and stressed about missing key deadlines. In extreme situations, your employees experience burnout and your teams suffer from high turnover.
• Inability to effectively execute KPIs: If you never find the real problem, you will be unable to meet the performance, access and outcomes goals demanded by your patients, members, communities and markets.

This mess is not always avoidable; sometimes trial and error is the only move you can make. In most situations, however, you can take a more strategic approach by closely examining data before making any decisions.

The strategic approach to problem-solving starts with identifying the root cause of the issue discovered.

When it seems like there is a problem, it is sometimes helpful to be skeptical at first—to ask, “Is this really a problem?” This is to elicit an explanation of why it’s a problem. (Put another way, “What discomfort is this causing for me, for my team, for my customers, etc.?”) Backing up to the “why (is this a problem)” instead of fixating on the “what (is the problem)” tends to mitigate against jumping into solution-finding before the true issue is properly identified.

Let’s explore an example:

1. You are approached by a few members of the Accounts Receivable (A/R) team, who state that there is a problem: “We need to be using predictive analytics/AI/ML.”
2. You apply the “why” lens: “Why is it a problem that you aren’t using predictive analytics/AI/ML?”
3. This question encourages the team to give the matter additional consideration from new angles. As a result, the team members re-evaluate and share: “The analytics we’re performing today do not seem to be as actionable as we expected. We are having difficulty taking tangible steps forward towards reducing unnecessary costs and addressing audit risk.”
4. With a more specific understanding of the dilemma presented, you conclude: The root of the problem isn’t that a specific technology is missing; it is that existing data insights aren’t actionable. You’re now equipped to focus less on terms and more on outcomes.

Once you gain clarity and focus on the root of the issue at hand, you can start thinking about potential solutions.

Before your fingers sprint across the keyboard to Google “how to fix XYZ problem,” take time to build on the identified root to flesh out the trunk and branches of the problem. I suggest asking these five questions to get the conversation moving:

1. Why is this a problem?
2. What happens if we don’t do anything?
3. Where does this problem exist? Is it broad or specific?
   1. For example, is it contained to one subject area (e.g., billing patterns), or is the problem prolific (e.g., missing actionable insights everywhere)? Is it only affecting one team (e.g., Accounts Receivable), or are there other stakeholders to consult?
4. How quickly do we need a solution?
   1. Can we spend some time exploring/innovating? Or are we in major crunch mode?
5. How “perfect” does the solution need to be?
   1. For example, forecasting emergency department (ED) and urgent care utilization patterns across a target population may not need to be accurate down to the number of patients; direction, trend, and magnitude are all that is needed to identify goals and implement plans for shifting sites of service. However, forecasting ED utilization for resource allocation (e.g., staffing, beds, supplies) needs to be much more accurate about exact patient numbers.

You’ll also want to consider a few broader factors that will have a significant influence on how you can move forward.

First, review the financials. Setting a reasonable budget from the beginning will keep you from wasting time on unattainable solutions.

Next, check your tech. Get a thorough understanding of what you can accomplish with what you currently have and what you could achieve with new platforms that fit in your budget. For example, many MedeAnalytics customers partner with us for our data lake, data storing and sorting capabilities that has the power to run large data science models.

Finally, rally the troops. Identify the internal and external experts in the field and choose who you want to rely on for guidance throughout the process. Internal experts can provide valuable, niche insight into the situation and serve as a champion for the chosen solution. External experts can offer a fresh, innovative viewpoint on the problem and help efficiently roll out the chosen solution.

If you’re interested in working with external experts, keep a handful of requirements in mind.

Your ideal problem-solving partner should be:

• Consultative and collaborative: Help accelerate you toward your goals, not distract from them.
• Experienced in data science: Knows the process inside out – from problem clarification and brainstorming to solution delivery and iterating.
• Tuned into your specific situation: Customizes standard solutions to meet current needs and long-term objectives (not ‘one size fits all’).

To illustrate this type of partner, let’s walk through how our MedeAnalytics team would handle the problem found above.

1. You share the challenge: Your existing A/R data insights aren’t actionable, which is a problem because cost and audit risk mitigation are strategic initiatives for your organization.
2. We help you prioritize: What areas within this domain are most hungry for data-driven decision-making?
3. You choose billing and coding analytics: This area is critical, as your organization has been struggling to efficiently understand where improvements could be made.
4. Our data science team gets to work: We identify clinicians whose spending/utilization patterns are outside the patterns of their peers for like services and surface anomalies—giving you a more actionable view of billing. With this information, you can facilitate early intervention and proactive education in coding practices to limit unnecessary spending. 

Notice that this solution did not include “predictive,” “AI” or “ML.” Like the problem statement, the solution should not be focused on the terms, either.

Start problem-solving like a data scientist

When you’re ready to ditch trial and error for a more strategic approach, MedeAnalytics is ready to help. Your problem is our challenge, and your solution is our goal. Reach out to our team of implementation, training and optimization experts today to get started.

(Written with support from Madeline Hasegawa)