Measuring Interdepartmental Outcomes In Healthcare: Challenges And Solutions

ScienceSoft offers three techniques for interdepartmental healthcare data analytics of outcomes. Implement them to understand your performance better.

Although all caregivers can benefit from healthcare data analytics, the challenge of comparing various departments against each other frequently makes outcomes analysis troublesome. Each department has its own specific outcomes and their averages. Even diseases within one unit have diverse outcome levels. For example, two patients in pulmonology — the one with COPD, the other with bronchitis — will naturally have different lengths of hospital stay, times to return to work, and readmission instances. At ScienceSoft, we came up with the following three solutions to this challenge.

1. Using Normalized KPIs To Analyze Different Departments
This is one of the effective ways to measure different departments that share common values, and we suggest adopting the following values:

• admission/readmission
• mortality rate
• AHRQ patient safety indicators
• infection rates
• complications
• patient experience
• patient complaints/responses

To accurately compare different departments, we suggest implementing the normalization factor, similar to risk adjustment in CMS. Let’s have a look at the example on how this factor can be calculated.

Imagine you have 23 hospitals in your health system. We take the rate of complications as a KPI, and we want to determine the most or least productive gastroenterological department among all facilities. Each department has its own percent of complications, which is enough to calculate the average rate of gastroenterological complications for the entire health system. Looking at this rate, we can determine ‘good’ and ‘bad’ performers. All departments with the indicator above average are considered underperforming in this case as we target complications.

Now, along with the gastroenterological batch of departments we have ophthalmology, for which we can also define the average rate of complications. Then we divide the average rate of complications in gastroenterology by the average rate of complications in ophthalmology. The result is our normalization factor for ophthalmology.

Say, the average for gastroenterology was 0.12 percent and 0.04 percent for ophthalmology. In this case the factor equals 3, and using it we are able to compare the performance of both departments. Every other department (neurology, pulmonology, etc.) will need its own normalization factor.

2. Using Target HIT To Compare Different Departments
This technique suggests using unique outcomes of your departments to compare them head to head and highlight issues. For example, let’s take three different departments — cardiology, neurology, and endocrinology. Each one has a whole set of its own outcomes. We suggest using the following algorithm to build the comparison:

1. Define key performance indicators for each department. For example, neurology has percentage of non-urgent electroencephalography (EEG) carried out within (≤) 8 weeks of request, ophthalmology needs to assess the rate of infectious endophthalmitis following an intraocular surgery, and so on.
2. Estimate the desired outcomes for each KPI — these are your expected values.
3. Calculate your outcomes according to your predefined KPIs — these are your actual values.
4. Compare the actual values with the expected ones and analyze which outcomes of which departments are within your targets.

With this algorithm, you can assess the performance of all departments, compare them with each other, detect positive and negative trends, and identify areas for improvement.

3. Using Using National And State Benchmarks For Interdepartmental Analytics
It is always helpful to look at how your fellow healthcare organizations perform across the country. To gain these insights for benchmarking, you can download data files from the official Medicare website that offers the Hospital Compare datasets to patients and caregivers.

As these sets contain a limited scope of measures, we recommend thoroughly sorting out the downloaded information and picking the right measures for you instead of integrating all the data into your analytics as it is.

For example, you can use the ambulatory surgical measures dataset that includes information about each facility-specific surgical area that could be matched with your departments. The list of areas includes:

• gastrointestinal
• eye
• musculoskeletal
• skin
• genitourinary
• respiratory
• nervous system

The readmission and mortality rate dataset has certain limitations: patients are 65 and older, quality measures include only heart and lungs conditions (heart attack, heart failure, stroke, COPD, pneumonia, CABG surgery) and also a hip / knee replacement measure for readmission evaluation.

Same occurs when comparing your hospital’s complications and healthcare associated infections with the national and state measures. It all requires a plenty of pre-analytic work to sort out the data useful for your organization.

Data Sources
You will need to get the required data from the following sources:

• EHR/EMR
• AHRQ quality reporting system
• Medicare national data storage

Afterword
Due to unique needs and specifics of every caregiver, implementing any of these three techniques of interdepartmental healthcare data analytics can also pose a challenge. Even if an organization already has an analytic solution, it may require a major update to add missing dimensions, measures, and KPIs.

However, it is a necessity to compare the productivity of different departments and understand their performance rates against the national and state benchmarks. Otherwise, no health organization can ensure data-driven decision making, which in practice means working in a blindfold.

About The Author
Natallia Babrovich is a Business Analyst at ScienceSoft, a software development and consulting company headquartered in McKinney, Texas. Since 2013, her advanced focus is IT solutions for healthcare and financial sector, with such projects in her portfolio as reporting for CMS and a care coordination portal.