The operation of a radiology department can be distilled into one simple goal: accessible, accurate information at the lowest possible cost. If information is not available, technology, interpretation, staffing, and computer systems become irrelevant. The intent of this article is to consolidate some of our findings related to radiology turnaround times from other studies, to review what we believe is a comprehensive monitoring system of services for a radiology department, and show how we approached improvements in those services.

Fairview Hospital, Cleveland, is a 453-bed general acute care hospital with a full range of services, including obstetrics; a Level III neonatal intensive care unit; an extensive cardiac program, including open heart and interventional cardiology; an active cancer center; a Level II trauma program; and residencies in medicine, surgery, and family medicine. The radiology department does more than 123,000 procedures per year and offers a full range of services with an operating budget in excess of $5 million, 87 full-time equivalents, and an eight-person radiology group.

LITERATURE REVIEW

While studies have been conducted, there does not seem to be agreement in the industry on what turnaround time for a radiology result should be or even how to measure it.

In 1994, Crabbe, Frank, and Nye (1) identified a seven-step process from examination completion to final report. They did not include the time from order to completion of the examination or the time to distribute results. They proposed to complete the cycle twice per day for a maximum turnaround of 32 hours. Across four types of examinations, improvements were implemented that brought turnaround from a baseline of 101 to 52 hours for various examinations to 59 to 44 hours for the same tests.

Hanwell and Conway (2) reported on a 1995 survey for the American Healthcare Radiology Administrators in which only 26% of the respondents monitored the time from radiologist signature to charting, and there was no monitoring of order to imaging or for receipt in the physician’s office. They found that turnaround time (image to signature) for all examinations in hospitals of 400-499 beds was between 20 and 25 hours.

Seltzer et al (3) set out to determine whether quality improvement tools could be applied to result reporting. In 1993, they reported that the mean time to sign reports dropped from 26 hours to just under 11 hours. Their subsequent work in 1997 (4) targeted “providing final reports within 48 hours.” In this study, the process is measured from examination completion to final report and they were able to drop from a baseline of 81 hours to 36 hours. They reported examination completion to dictation of 19 hours, dictation to preliminary report of 4 hours, and preliminary to final report of 13 hours. They noted that transmission of reports was not included.

The Medical Leadership Council (5) in 1996 reported a study completed by the Premier Health Alliance and The University Hospital Consortium of 80 hospital departments. The summary of findings is in Table 1.

While hospitals estimate a turnaround time of less than 24 hours, this does not seem to be the norm for most large facilities. Further, we conclude that the focus seems to be purely on image completion to radiologist finalization. This may make sense from a department perspective, as these are the components that the radiology department can control. However, if the patient does not get to the department, the process cannot begin. The studies we reviewed did not include this time frame as a component. In the minds of our ordering physicians, the process begins when they write the order, not when we finish the imaging. In our institution, transport and schedules are controlled by radiology; therefore, the time from order to image completion is a departmental responsibility. We concluded that radiology should be held accountable for this component.

In terms of report distribution, today’s information technology allows the report to be available online when the radiologist electronically verifies (signs off) the report. Therefore, finalization by the radiologist could be defined as the completion point. Unfortunately, the majority of our physicians do not have interest in or access to online reports. Again, from the perspective of our medical staff, radiology was done when they saw a report in the chart or received the report in their office. The radiology department does not control charting; however, in the minds of our customers (ordering physicians), we were being held accountable for the entire process. This forced us to develop a mechanism to monitor this aspect, as well as the beginning and middle of the process.

Where Crabbe, Frank, and Nye identified a seven-step process, we expanded it to a nine-step process as follows:

  • Order
  • Transport
  • Imaging
  • Film Assembly
  • Interpretation
  • Dictation
  • Transcription
  • Finalization
  • Distribution
  • Meltdown Imminent

In 1996, events at Fairview Hospital forced us to evaluate our process and its results. We needed to revamp our entire operation and determine a monitoring system that captured the entire process from the placing of an order to the final destination of the report.

Before 1996, Fairview’s radiology department had a tenuous grip on the requirement of providing the needed information to our medical staff. In late 1996, that grip was lost when a new information system conversion eliminated the existing homegrown system. The department (and the hospital) was sent into months of frenzied activity to meet the simple objectives of providing results, films, and access. In the end, our systems and services are better for having gone through the process, but not without a lot of anguish.

On conversion, the number of departmental complaints from patients, physicians, and nursing units grew dramatically. Report turnaround time, film availability, and telephone access were all seriously compromised. We found that due to a lack of report availability the demand for films increased. However, the staff was spending so much time trying to expedite the matching and reading of films, that the films were not being managed (filed) appropriately; they were inaccessible and therefore reported lost. In addition, the demand for films and reports clogged the telephone lines beyond our capacity (calls exceeded 280 on the day shift), patients could not schedule, and physicians could not call for films or results. One system failure fed another and the department was in a negative spiral. We were in complete meltdown.

Before long, senior administration, medical leadership, and departmental management developed an ad hoc Radiology Services Committee to rectify the situation as soon as possible. Due to the comprehensive nature and complexity of the problem, a consultant was engaged.

PLAN OF ACTION DEVISED

We believed that short-term actions were needed to stabilize the system while more substantive improvements were initiated. Three additional FTEs (full-time equivalents) were approved to help with telephones and film management. At the same time, the planned completion of the radiology information system (RIS) installation was expedited to manage radiology information including film tracking. The existing dictation system was replaced with a Lanier system and interfaced to the RIS, all of which sounds simpler than it was in reality.

The Radiology Services Committee met weekly to review actions and results. With the consultant, a team of departmental staff and radiologists was developed. Two lists of actions were developed, a list of quick steps (referred to as the Just Do It list) and a set of longer-term initiatives.

To reduce turnaround time and improve film availability, the following steps were taken (again, simpler than they sound here):

1. Modem access was established to allow the radiologists to finalize reports from their homes.

2. Routine reports were developed that are bar coded into the system. These standardized reports were implemented to reduce the amount of time needed to transcribe.

3. The flow of films through the department was reorganized to limit the number of times a film was handled. Previously, a majority of films went directly to the radiologists and then to film assembly. Currently, more films go to film assembly before radiologist review.

4. Active file capacity was increased from 2 to 6 weeks.

5. Redesign of the entire front office was completed to support the flow design changes and file capacity. These changes included installing workstations for film assembly, consolidating and expanding the film retrieval window, installing additional film storage cabinets, relocating transcription and schedulers, establishing a workstation in the basement film storage area, and installing a high-density film storage system.

6. Film filing was converted from alpha to color coded patient number in which the last four digits and the year are color-coded. This makes misfiles more obvious. A daily review of the library identifies misfiled films at a glance.

7. Physicians requesting films were required to provide a medical records number. A daily roster of patient names and medical records numbers was provided at the film library for easy access and staff were trained to assist them in finding patient numbers.

8. All films, including portables and emergency department wet reads, were coded into the system prior to delivery to remote locations.

9. Formal collection rounds were introduced in the hospital and the department.

10. Subfolders were eliminated; now the entire patient record is kept together, which reduced misfiling of films and the number of times a jacket was handled.

11. Autofaxing of reports to physician offices was implemented.

We instituted a service monitoring system and some initial targets. At the time, the targets were little more than guesses, although our initial review of the literature made these guesses somewhat educated. The report reflects mean times as reported by the radiology information system. To track the entire process, we included Order to Begin. We also began monitoring charting of results at the specific nursing units.

Routine calls to physician offices were also initiated to assess the arrival of results. This practice was abandoned early as the office staff rapidly lost patience with multiple follow-up calls. The autofax seemed to meet their needs and they asked us to stop checking on reports. The practice did show the physicians the extent to which the department was willing to go to meet their needs.

RESULTS WERE FORTHCOMING

The results of the department’s efforts are depicted in the radiology performance report (Table 2). In January 1997, the Order to Begin time was 6 hours and has remained consistent across the year. Complete Examination to Final Report has dropped from 59 hours to as low as 20 hours. We have seen some regression and ended 1997 at 26 hours.

Standardized reporting was implemented in June. While helpful, the use of standardized reports has not met our initial targets. Plans to bar code standardized reports are expected to increase their utilization.

At Fairview, charting is the responsibility of the nursing units. We have found that routinely reporting chart rates at Fairview has provided an incentive to getting the results charted (Hawthorne Effect). Initial data showed a 38% chart rate within 8 hours of receipt. By the end of 1997, the average chart rate was between 64% and 93% (Table 3). We plan to move to a 4-hour standard.

We have also seen a dramatic decline in patient complaints via the phone, lost films, and reports not sent to physicians as tracked by the hospital’s Customer Relations Department.

CONCLUSIONS AND BEYOND

We have survived the meltdown and have received unsolicited compliments from some of our physicians about the improvements that have been made. We have reduced turnaround time from an average of 59 hours to 26 hours, our find rate is consistently running at 100%, charting has improved from 38% to 83%, and patient complaints have dropped to almost none. We are by no means a model department – yet. Our turnaround time is still too high. Future targets are to get the complete process from order to finalized report to less than 18 hours. With charting at 4 hours, the entire process will still take an average of 24 hours. To do this, we will be focusing on film assembly, which is the largest time component and the one over which we have the most control.

Our greatest lesson has been in establishing targets and monitoring progress. By reinforcing these processes with the staff, radiologists, medical staff, and nursing units, we have been able to move forward. Second, monitoring components has helped us to better understand the entire process and allowed us to work on areas that will get the greatest return for our efforts. Finally, it was important for us to view the entire process. If we had a 6-hour turnaround time from completion to final report, it would be meaningless if physicians and nurses did not have easy access to the results. Our objective continues to be to contribute to the overall episode of care, not just the processing of an image.

?

Victor J. Galfano is associate vice president, operations, at Fairview Hospital, Cleveland.