Thinklytics

Community Health Network · Healthcare · Indianapolis, IN · 18 weeks

Self-service analytics deployed to 340 clinical staff

Community Health Network’s analytics team handled 120 ad-hoc report requests weekly from clinical staff, which used up four of every five working days. We created a self-service analytics layer in Tableau for 340 clinical users, allowing them to generate their own reports. This reduced ad-hoc requests to 14 per week.

Challenge

The 8-person analytics team spent four of every five working days handling ad-hoc report requests from clinical and operational staff. With 120 requests weekly and a six-day average turnaround, clinical managers relied on outdated data to make staffing and scheduling decisions.

Approach

We interviewed 40 clinical managers to find their 20 most common questions. Then, we created a Tableau self-service platform with 20 verified data sources and 35 starter dashboards focused on staffing, patient flow, quality metrics, and supply use. Over four weeks, we led eight training sessions for 340 clinical staff to ensure proper adoption.

Outcome

Ad-hoc report requests fell from 120 to 14 per week within six weeks. This gave the analytics team back four of every five working days for strategic projects. Annual labor costs for these reports dropped from $890K to $124K.

We grabbed the main players and hashed out clear, doable requirements before jumping into development.

We started by sitting down with 40 clinical managers over three weeks. Our mission was simple: figure out the 20 questions they ask the most. Guess what? Those 20 questions accounted for 1,700 out of 2,000 random requests every month. So instead of guessing, we built the tool around what really mattered to them.

We ditched the messy raw tables and brought in clean, verified data. That way, everything’s way more accurate and trustworthy.

Here’s how we tackled it: we built a self-service layer that connects 20 certified data sources, all tightly governed and secure. The clinical team jumps in there to pull trusted data, no more sifting through messy raw tables. This way, their reports always line up perfectly with what our analytics folks call the official numbers.

How We Figured Out the Exact Skills We Needed to Sharpen

We ran eight training sessions, around 90 minutes each, over four weeks. The group? Clinical managers, department heads, and power users. But we didn’t just chat theory, we dove into real data the whole time. Out of 340 folks who joined, 309 were up and running solo within a month. Pretty solid turnout!

Results

  • $890K Annual ad-hoc labor saved
  • 120 to 14 Weekly ad-hoc requests
  • 340 Clinical staff self-sufficient
  • 18 wks From kickoff to full deployment

We used to get hit with around 120 ad-hoc requests every week, which meant no time for strategic work. Now, it’s dropped to 14, and the clinical staff are answering their own questions.

Director of Analytics, Community Health Network

In the client's terms

The question

What decisions is this reporting supposed to support?

The situation

An eight-person analytics team spent four of every five working days on ad hoc requests. At 120 requests a week with a six-day turnaround, clinical managers were making staffing decisions on data that had aged out before it arrived.

What we found

The queue was the symptom. Forty clinical managers turned out to be asking twenty questions between them, over and over, because there was no way to answer one without going through an analyst.

What it produced

Weekly ad hoc requests fell from 120 to 14 within six weeks. The analytics team recovered four of every five working days. Annual labour on ad hoc reporting fell from $890K to $124K.

The roadmap

Interviews with forty clinical managers established the twenty questions that actually recurred. A Tableau self-service environment was built on twenty certified sources with thirty-five starter dashboards covering staffing, patient flow, quality and supply use. Eight training sessions over four weeks took 340 clinical staff through it, because a platform nobody is taught to use is a platform nobody uses.

The work behind this result

Related case studies

Frequently asked questions

How do you reduce ad hoc reporting requests to an analytics team?

At Community Health Network, weekly requests fell from 120 to 14 within six weeks of launching self-service. The step that made it work was interviewing forty clinical managers first to establish the twenty questions that actually recurred, which accounted for 1,700 of roughly 2,000 monthly requests.

What does an ad hoc reporting queue cost?

Community Health Network was spending $890K a year in analyst labour on ad hoc reporting, with an eight-person team giving four of every five working days to it. After self-service that fell to $124K a year.

Does self-service analytics compromise data accuracy?

Not when it runs on certified sources. The Community Health Network platform exposes twenty governed data sources rather than raw tables, so what clinical staff produce reconciles with what the analytics team publishes. Of 340 staff trained, 309 were working unaided within a month.