Five tips: How can radiology utilize data mining to increase quality of care

Many hospitals find themselves dealing with heavy internal pressure for increased efficiency, as well as tougher government demands on financial and quality reports. Many hospitals are also experiencing greater demand for CT and MR, but are facing stagnant budgets that are not increasing to match this heightened demand. Only by boosting operational efficiency can these challenges be addressed.


Both RIS and PACS, but especially RIS, contain large quantities of information that can give you the knowledge needed to tackle these challenge. With the right data mining tools, the data available in your departments can easily be extracted and converted into facts and figures on which actions can be based. You can identify bottlenecks and anomalies, monitor and streamline the patient flow through your department, as well as measure the effect of implemented changes. With the right data at hand, you can increase the quality of care and cope with the current financial pressure not only by allocating costs better, but by boosting revenue through increased efficiency.

In this article, we share five tips on data mining.

1. Eliminate bottlenecks to improve quality of care

One of the most visible ways to measure quality and spot bottlenecks is to look at lead times. Lead times can be used, for example, to gauge the service level given to different referring units, but this is also a powerful way to identify workflow inefficiencies for the purpose of improving the quality of care. In addition, following these values over time will enable you to quickly spot and act on anomalies. Some key performance indexes include the time from referral to when the report is read, the time from examination to when the report is read, the time before a referral is scheduled and different waiting times, such as the time from patient arrival to started examination.

2. Measure the effect of implemented changes

An easy way to measure efficiency is to look at production volumes. This will enable you to compare activities on different modalities, days of the week and times of the day, as well as to measure the effect of implemented changes. For example: Has the introduction of a web portal for patient self-scheduling in mammography had any effect on the level of scheduled time slots? Has the decision to offer exams on weeknights paid off? Relevant volumes to look at could include total number of exams, number of exams per modality, volumes on weekdays versus weekends or volumes during different time spans of the day.

3. Choose a system that enables nonstandard queries

Regardless of which business intelligence tool you invest in, you will have information needs that are not covered by the reports made available in the tool itself. Local regulations may impose demands on your clinic to report very specific data or you may wish to combine your data with data from other hospitals for an aggregate report. Answering any queries other than standard monthly reports can prove cumbersome and depend on internal IT or consultants, which is time consuming and costly. You need a tool that lets you easily access the complete database and export raw data for further analysis.

4. Don’t underestimate the value of medical imaging experience

The challenge with statistics is not usually obtaining data, but filtering out the relevant data and converting the numbers into knowledge—and actually acting on this knowledge. Find a vendor that knows the medical imaging business. There are plenty of companies out there that are familiar with business intelligence, but finding one that has experience of your type of operations will save you both time and money.

5. Give yourself room to grow or shrink

Choose a flexible solution. Many clinics and hospitals start out small with just a few simple reports. Perhaps you will never move beyond this level or perhaps you will. Make sure your tool can grow with you. Flexibility should also mean freedom of choice when it comes to selecting statistical tools, such as QlikView, Excel or something totally different.

Related products