Robert A. Bell, PhD

The Winter Olympics have once again shown us that properly motivated and trained individuals can achieve truly amazing goals. The icon of the five intertwined Olympic rings is also a cogent metaphor for the five interdependent pillars on which high-quality diagnostic services rest. Should any one be substandard, the entire MRI service may be at risk. These key facets are:

• Patient Throughput. Examinations must be conducted in an efficient, safe, and empathetic manner.
• Examination Technical Quality. Sites must be able to produce sufficient image quality throughout an acceptable range of applications to meet the interpreter's and referrer's requirements.
• Interpretation. Readers should have sufficient skill and experience to properly diagnose from acceptable examinations and to identify those not meeting clinical needs.
• Communications with the Community. Diagnostic information must be conveyed to the referrer in an understandable and timely manner. The medical community must be made aware of new applications and techniques. Patients must have their questions addressed.
• Billing and Collections. MRI providers must get paid before bills become due.

This article focuses on the first category, enhancing patient throughput. Staff members who hold the responsibility for dealing with MRI patients, primarily technologists and receptionists, can also be gold medal competitors. But how should their performance be measured? What goals offer motivation without fear of punishment? How can one assess efficiency? Can operational bottlenecks be identified? Is it worth the effort? Appropriate benchmarks are vital to addressing these and other questions.

According to IMV International, the firm in Des Plaines, Ill, that polls thousands of hospitals and imaging centers each year, approximately 18,000,000 MRI examinations were conducted in the United States in 2001. This volume represents an 11% increase over 2000 and translates into a utilization factor of about 63 examinations per 1,000 of population per year. Such data should come as very welcome news to MRI site managers. In the present climate of limited reimbursement, volume is key to financial viability. However, the fiscal impact of volume is sometimes misunderstood and volume at any price is not necessarily beneficial.

MRI is a high fixed-asset investment. This means the majority of expenses incurred in the production of examinations does not depend on the number of examinations. Lease payments for the system, rent, utilities, maintenance costs, basic personnel expenses, management, and legal and accounting costs must be paid whether the site conducts one examination or 4,000. "Breakeven" is the number of examinations that must be completed and reimbursed to meet annual operating expenses, both fixed and variable. The breakeven number of examinations per year can be calculated by dividing the fixed costs by the payment per examination less the variable costs (see Figure 1, page 27):

If the site depicted in Figure 1 conducts and receives payment for 1,556 MRI examinations per year, it will be able just to meet the costs to produce those examinations (no profit or loss). However, if the same site conducts 1,800 examinations, it will enjoy a profit of (1,800 - 1,556) x $450, which equals $109,800. Thus, once breakeven volume is achieved, further payments fall largely to the bottom line. Each examination over breakeven contributes $450 to this hypothetical venture. As long as one's average reimbursement is higher than the actual costs to produce an examination, additional examination volume should be encouraged. However, be careful that deep discounts to some providers do not jeopardize higher paying contracts as explored in the example below.

Figure 1: MRI breakeven calculation. Costs and reimbursement are approximations based on a hypothetical MRI facility with a high-field system.

Imagining, for a moment, that the hypothetical MRI venture above is on track to complete and be paid for 1,556 examinations this year. A scan broker approaches management and offers an additional 500 examinations for payment of $300 each. This sounds like a very tempting offer since the cost for each of these additional patients appears to be only the variable component, $200. Is not this an easy profit of $50,000? A proper answer is not possible until many questions are addressed. Will this discount be demanded by other payors that now are paying more than $650 per examination? Will the extra volume decrease the quality of service offered to all patients? Will shareholders (or hospital management) expect the same level of profit next year? Will the scan broker demand even deeper discounts for the next group of patients? In general, it is prudent to accept only contracts that reimburse at levels above your average cost per scan, not variable costs alone.

Average cost per examination is volume dependent. It is calculated by dividing the total yearly fixed costs by the annual number of examinations and adding the variable cost per examination:

Cost Per Examination = (Yearly Fixed Costs/Annual Examination Volume) + Variable Cost Per Examination

In the example described in the figure, this cost is $650 for 1,556 annual examinations. However, if yearly volume is 2,000, the cost drops to $550. And if the facility can achieve a yearly volume of 3,500, the cost per examination drops to $400.

Volume Alons Is Not Enoigh

Although obtaining more volume is often the primary goal at MRI facilities, insufficient focus is usually directed toward processing the volume efficiently. When asked for their average time spent per MRI examination, many site managers answer, "We schedule every 45 minutes." This demonstrates confusion between scheduling and production. What you plan is not necessarily what you achieve. Average time per examination can be easily calculated by dividing the number of hours per week your system is available for scanning by the number of examinations you actually conducted during that period. In a survey of 60 MRI systems in 1994, the author found hospitals and freestanding centers averaged about 1.3 hours per examination. Today, best estimates indicate the average MRI facility does about 2,600 examinations per year per unit. IMV International found that in 2001 the weekday hours of MRI sites were as follows: 17% were less than 9 hours, 46% were 9 to 13 hours, and 37% were more than 13 hours. Also, they found that more than 40% of sites were open on the weekends. From this data an average estimate of 12 hours per weekday and 5 hours per weekend is not unreasonable. This total of 65 hours per week yields an average time per examination of 1 hour and 15 minutes. Thus, it appears little progress has been made in this critical production number over the past 7 years.

Time is Money

For MRI sites the old adage "time is money" might be altered to "time per examination is money." Highly efficient MRI operations have demonstrated that total patient-to-patient scan times on modern high-field systems can be less than 30 minutes for the large majority of examinations. Where are the bottlenecks that keep most facilities from achieving such throughput? There is plenty of blame to spread around.

• Instrument Induced. Excessive downtime limits availability. Lower field strengths may require somewhat more time per examination to achieve acceptable signal-to-noise ratio (SNR). Older scanners may not have upgrades such as fast spin echo or phased array that can speed throughput via faster scan sequences.
• Radiologist Induced. Patients are left on the table while the technologist searches for a radiologist to review the examination. Examination protocols, selected by the radiologist, contain every sequence the vendor has ever offered and take an hour to complete. Scan protocols vary widely depending on the radiologist on duty.
• Technologist Induced. The next patient is not ready when the examination is done. Anxiety often masquerading as claustrophobia is not properly addressed. Filming holds up the patient queue.
• Management Induced. Employee motivation is poor. Needed supplies are not available in time for the examination. Staffing is inadequate.
• Scheduling/Reception Induced. Patients are not properly screened. People are late for their scheduled appointments or do not show up at all because they were not called the day before. Directions to the facility are inadequate. Patients are scheduled in longer-than-necessary increments. Complicated examinations are not provided adequate time (eg, pediatrics).
• Other Sources. Patient flow is interrupted by add-ons, emergency cases, bad weather, power failures, or other unexpected events.

Never The Patient's Fault!

Setting proper throughput goals requires input from all of the malefactors above. It is also important not to try to run a 4-minute mile when you just start jogging. If your present average examination time is 1.1 hours, perhaps a goal of 55 minutes within 2 to 3 months might be reasonable. As improvements are made, other goals can be set. Here are a few tips:

• Schedule in shorter time increments but leave more open slots. Instead of 1 hour, try 45-minute slots. This will encourage faster processing per patient but provide a cushion if the schedule falls behind. As throughput improves, fewer open slots will be needed.
• Request that the radiologists use a consistent set of modern protocols and that they review them at least twice a year to remove unnecessary sequences. Sometimes clinicians add newer or faster sequences but do not remove outdated ones for fear of lack of coverage. Provide time outside normal scan hours where the radiologists can try out new methods.
• Call every patient the day before the examination. Ask them if they have any questions, and if they know how to get to the facility and verify their appointed time. Efficient sites often schedule patient arrival 20 to 30 minutes before actual examination time. If the site is ahead of schedule, they can be pushed forward.
• Consider installing a PACS. Picture archiving and communications systems can markedly reduce or even eliminate film. This can reduce time between patients, speed report turnaround, and lower expenses.
• Coordinate patient transfer between reception and the technologist. Ensuring proper screening, the completion of paperwork, and enough time to prepare for the examination can reduce wait time between patients.
• Compliment the staff when good performance is seen. This is especially important for radiologists to observe. Often technologists and other staff are not given the consideration and respect they deserve.
• Make sure your vendor is earning their service fee. When you have a problem, repair service should be in action by phone or on site as soon as possible. Ask the service person to show you how he is assessing system performance during preventative maintenance. Consider a daily quality control program administered by your staff. The ACR MRI Accreditation Program can be very useful.

Comparisons With Other Sites

There are concrete ways to assess local performance. Below is listed a number of important benchmarks in my experience with levels of Gold (best practice), Silver (better than most), Bronze (approximate US average), and Dead Last (worst case).

Patient Throughput.

Explanation: Time available divided by number of examinations completed yields time per examination. These awards should be tempered depending on the type of equipment at the site (lower field systems may require more time per examination due to lower SNR).

Gold=25 to 30 minutes
Silver=40 to 45 minutes
Bronze=75 minutes
Dead Last=More than 90 minutes
Impact: Saving an achievable 30 minutes per scan increases capacity by 67% at no additional fixed expense in a 65-hour work week (52 to 87 patient slots). If additional volume is available, it can be processed without altering hours or personnel. If volume is limited, hours (and, hence, personnel costs) can be reduced.

Personnel Utilization.

Explanation: Personnel can be one of the largest expenses in MRI operations. Good technologists are hard to find. Proper utilization of this resource can markedly affect the bottom line. The figures below represent the number of examinations per technologist per 8-hour shift.

Gold=7 to 10
Silver=6
Bronze=4
Dead Last=3 or less
Impact: Improvements in this category usually mirror those in time per examination. Increasing from four to six per shift is a 50% increase in technologist productivity. By removing some of the obstacles to smooth operation (eg, consistent and concise protocols, lower filming requirements, efficient pipelining of patients), technologists can concentrate on what they do best: patient care. This category must be treated with some care since many sites employ only technologists while others have a mixture of technologists and aides. I suggest counting aides as 0.5 FTE technologists if the latter is the case.

Film Utilization:

Explanation: Film is a surprisingly large expense in MRI. Primarily, this is not because of the inherent cost of the medium but rather due to the associated costs of long-term storage, retrieval, technologist labor to produce, manipulation (couriers, hanging films), and chemistry (environmental costs). The actual cost of film is estimated to be $3 to $5 per sheet. The average site uses about seven sheets per examination and duplicates about 30% of cases (more than nine sheets per examination total). The numbers below represent total number of film sheets per case including any duplication.

Gold = 3 sheets or less (usually incorporates PACS)
Silver = 6 to 7
Bronze = 9 or more
Dead Last = 12 or more
Impact: At an average variable cost of $35 to $40 per examination due to film, most MRI facilities can pay for a miniPACS in less than 3 years based on film reduction of about 70%. Lowering film use also saves technologist time and can increase radiologist productivity.

Contrast Utilization.

Explanation: This parameter can differ among sites for clinical reasons due to variations in examination distribution by anatomic region (eg, head examinations generally use more contrast than spine examinations), age ranges, and referrer preferences.

Gold=Less than 20% of examinations (typical for high-field systems)
Silver=25%
Bronze=35%
Dead Last=More than 80%
Impact: The choice to employ contrast is a medical decision, but it is also the radiologist's responsibility to counsel referrers who request contrast for the wrong reasons. Reducing unnecessary contrast usage can speed throughput and lower liability.

Claustrophobics.

Explanation: Claustrophobics are patients who occupy scanner time but do not generate revenue, since an acceptable examination is not produced for a variety of reasons. Included in and perhaps dominating this category are those who may be nervous about their MRI and express their anxiety as a fear of the system but are not true claustrophobics. It is my experience that this number does not show a significant dependence on type of equipment (open-sided vs tubular MRI) at well-managed MRI facilities.

Gold=Less than 2% of patient volume
Silver=3%
Bronze=4%
Dead Last=More than 5%
Impact: Improvements in this category offer only minor capacity increases, but usually indicate significant advances in patient feedback.

Lost Patients.

Explanation: Every site has some patients who do not appear on their schedule day. Although many no-shows are rescheduled and scanned at a later date, the slot on their initial visit often goes unfilled. Calling the day before to verify the appointment helps to impress patients with the need to be on time.

Gold=Less than 2%
Silver=3%
Bronze=4% to 6%
Dead Last=More than 20%
Impact: As with claustrophobics, usually only minor capacity benefits are possible, but it is a good measure of contact with patients.

Late Patients.

Explanation: Those who arrive more than 15 minutes late for their scheduled appointment can disrupt patient flow.

Gold=Less than 2%
Silver=3%
Bronze=4% to 6%
Dead Last=More than 20%
Impact: Again, usually only minor additional scan slots are possible, but a large number can indicate a need to alter schedule policies. Consider asking all patients to come in an additional 15 minutes before the time you intend them to be on the table.

Start Time and Last Patient.

Explanation: The time in minutes between the scanner's availability for the first patient of the day and the actual time the patient gets on the table is deemed "start time." "Last patient" is the time between the final patient leaving the table and the close of imaging hours.

Start Time:

Gold=Less than 10 minutes
Silver=15 to 20 minutes
Bronze=30 to 40 minutes
Dead Last=More than 60 minutes

Last Patient:

Gold=less than 20 minutes
Silver=30 to 40 minutes
Bronze=60 to 90 minutes
Dead Last=More than 150 minutes
Impact: If start time is too long, scan time is wasted and examinations are backed up. If last patient times are long, again time is wasted and perhaps scan hours should be reduced.

Average Annual Cost Per Examination.

Explanation: Highly efficient sites promote high volume, excellent throughput, and low expense while maintaining superb quality. Cost per examination is a measure of progress attaining the first three of these.

Gold=Less than $350 including reading fee
Silver=$425
Bronze=$500
Dead Last=More than $600
Impact: The combination of efficient throughput, scan hours appropriate to local volume and expense control optimizes the opportunity for profit.

SUMMARY

No two MRI facilities are identical so comparisons should be made with some caution. However, opportunities for improvement exist everywhere. MRI growth is strong and new applications arise each month. Preparing to handle more volume at lower cost per examination without jeopardizing quality is a difficult task but one that offers Olympic gold for those who succeed.

Robert A. Bell, PhD, is president of a health care consulting film in Encinitas, Calif