| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Perspective |
1 Queens University School of Medicine, Kingston, ON, K7L 3N6 Canada.
2 Departments of Diagnostic Radiology and Economics and Public Health, Yale
University School of Medicine, P.O. Box 208042, New Haven, CT
06520-8042.
Received March 20, 2005; accepted after revision June 7, 2005.
Address correspondence to H. P. Forman
(howard.forman{at}yale.edu).
OBJECTIVE. The purpose of this article is to promote insight into radiology costs through improvements in assessing patient-level cost data. Accurate patient costing is a prerequisite for establishing a proper payment system—one where the price paid for a service approximates the cost of delivering that service. In the absence of an accurate payment scheme, margins can vary significantly from one patient to the next.
CONCLUSION. The resulting financial incentives skew the radiology marketplace away from the provision of efficient and appropriate care toward the selection of patients whose costs are low relative to reimbursements.
Cost, appropriately defined, is a very useful measure. In the literature, costs are often assessed to reflect the financial performance of a department, an imaging technique, a piece of imaging equipment, or a procedure [1-3]. These measures are critical to many management and policy decisions. This article makes the case that further insight into radiology costs can be gleaned by assessing patient-level cost data. Policy aims such as access, system efficiency, and cost effectiveness rely on accurate patient costing. Thus, an improved understanding of patient-level cost data will help us better manage our specialty into the future.
Where Are We Now?
This article considers resource costs to be the resources consumed in the treatment of patients. We focus on resource costs because they are often used to assess costs in health care and also because they form the basis of the Resource-Based Relative Value Scale (RBRVS) system of reimbursements. The Current Procedural Terminology (CPT) system defines radiologic studies according to imaging technique and anatomic region imaged. The RBRVS then permits conversion into dollar units [4]. This establishes a reimbursement for a family of examinations defined by anatomy and imaging technique. The portion of radiology revenues deriving from inpatient diagnosis-related groups does consider disease state and comorbidities. However, this applies only to the technical component of inpatient charges.
Recent evidence, however, suggests that important cost differences exist among patients with similar CPT codes. Herts et al. [5] found that image acquisition times are longer for inpatients than for outpatients. This implies that inpatients generate greater image acquisition expenditures. In previous work, we showed that patient indications for CT can predict professional expenditures [6]. That is, the nature of patients' illnesses may render them more or less costly to diagnose. Blackmore et al. [7] showed that resource costs in spinal trauma radiography increase with patients' increasing probability of injury. These studies all indicate that patient characteristics bear on radiology expenditures, independent of anatomy and imaging technique.
Hsiao and colleagues [8], who performed the lion's share of the work devoted to establishing a rational Medicare fee schedule, arrive at a similar conclusion: "Because the complexity of patients' conditions can vary within a given CPT-4 code, we described (aided by physicians) typical cases for purposes of measuring resource costs." Notionally, assigning a cost to a typical patient is efficient, although it is at odds with recent cost assessments in the radiology literature. The root and resolution of this problem are founded in the fundamental methods of cost accounting, which we turn to next.
Process and Job-Order Costing
Process costing and job-order costing are two differing ways to assign a cost to patients [1, 2]. Process costing is a system whereby all costs incurred within a period are divided equally among all units produced within that period. Process costing is generally used in instances where each unit of output consumes the same resources. Job-order costing tracks separately the costs of producing separate units of output and is typically used in instances where each unit produced consumes different amounts of inputs.
The example in Table 1 helps illustrate the difference between job-order and process costing. The first two cost columns use the job-order approach, where each input is measured and assigned to the appropriate radiology patient. Patient 1 is an outpatient who is ambulatory and may be easily positioned for imaging. Patient 2 is a nonambulatory, hospitalized patient with an IV line, a chest tube, and other impediments to imaging. This patient also has findings that make the images more time consuming for the radiologists to interpret. In short, radiology care for this patient consumes more resources. The "Average" column reflects how we would view these same expenditures with process costing. Neither costing method is clearly superior; both job-order and process costing have their relative strengths and appropriate uses, as we will see later.
|
The Difference Between Revenue and Cost: Margin and Incentive
The study and design of physician payment incentives are complicated somewhat by the presence of two tiers of payment incentives. First, there is a payment structure by which funds are disbursed to imaging practices. This usually represents payments by third-party payers. Second, there is the basis on which each practice pays its physicians. Recent literature supports the claim that both tiers influence physician behavior [9]. To our knowledge, the effect of each incentive tier on the manner in which radiologists conduct their practices has not been studied. For the purposes of this article, we will adopt the perspective of the radiology practice and assume that individual radiologists' responses to incentives are identical to those of a profit maximizer.
For the sizable privately owned radiology sector, this is probably a fair assumption. Indeed, hospitals that maximize the incomes of their physician staff members will also behave like profit maximizers [10]. Profit maximization is also a fair assumption for nonprofit entities. The not-for-profit sector will maximize profits on profitable services and use the gains to finance nonprofitable activities [11]. Thus, the profit-maximizing assumption can apply to portions of the not-for-profit sector as well.
The difference between revenue and cost is the margin. In our example, if reimbursements are equated to job-order costs, then hospital costs are covered, and each patient is charged according to the costs he or she incurred. If reimbursements are set equal to process costs, it is plainly more profitable to service the low-cost patient. The margin is also affected by revenues, more specifically by payer type. It is generally known that indemnity insurance payments (where payments typically track charges) exceed managed-care payments (which are most commonly a negotiated discounted fee-for-service). These payments, in turn, exceed Medicare payments, which typically exceed Medicaid payments, which ultimately exceed uncompensated care.
The financial incentives extended by the current payment system in radiology encourage the selection of low-cost, well-paying patients. Economic theory argues that providers have an incentive to seek out patients associated with relatively high margins and that the incentive to pursue these cases increases with the margin. Conversely, providers have an incentive to undertreat or avoid patients associated with negative margins. Such patients run the theoretic risk of facing an access problem.
Enterprising practices that actively manage their patient and payer mix through advertising, contracting, and denying treatment will service a disproportionate share of profitable patients. The remaining poorly profitable or unprofitable patients will be treated by practices that are more passive toward their patient and payer mix. These latter practices will likely find themselves in grave financial condition.
The market mechanism will best serve patients if it is harnessed to promote appropriate and efficient care instead of patient and payer selection. An improper payment system—where the price paid for radiology services does not approximate the cost of delivering those services—will tend to devolve toward selection.
Show Me the Data
Is patient selection just a theoretic problem, or is it an actual problem? The general medical literature supports the contention that incentive structures do influence physicians' practice habits [12]. Indeed, finding the best-paying, least intensive work is a fundamental human economic proclivity. The radiology literature contains evidence of this effect. It is intuitive that by managing routes of referral, practices can also manage their patient mix.
Levin et al. [13] detailed how outpatient centers may select and deselect the clinicians from whom they receive referrals. Conversely, the selection bias inherent to inpatient and trauma centers gives them a unique patient mix and a limited ability to manage that patient mix. Ginsburg [14] asserts, "Physicians' ability to change their payer mix of patients distinguishes them from hospitals." It is therefore not surprising that tertiary care centers are often incorporated into health networks, which typically would include outpatient centers. The location of these outpatient sites may be chosen to affect the patient mix, and the programs offered may, at least in part, be chosen on the basis of the expected margin.
Consider the recent bumper crop of self-referred, whole-body CT screening centers [15, 16]. The availability of this service has exploded despite a swirl of controversy over its value [17, 18]. Stolberg [18] argues, "Simply put, there is no apparent scientific basis for whole-body CT screening." One might suspect that with an uninformed consumer the supply of this service is excessive compared with its value because its revenues are excessive compared with its costs. Overtreating certain patients by providing CT screening despite a lack of evidence for its efficacy represents an inefficient use of resources.
The idea that pricing mechanisms may lead to inefficiency is not unique to radiology. Historically, the Civil Aeronautics Board (CAB) regulated airline fares. For airlines, servicing smaller towns was frequently not profitable because of low occupancy rates on those flights. The CAB would often require airlines to provide money-losing service to smaller towns in exchange for higher-than-competitive rates on routes between large cities. The assumption was that profits from one route would subsidize losses from the other. Yet the extraordinary economic profits on certain routes induced airlines to compete on those routes by offering frequent flights. This resulted in half-empty airplanes—an inefficient use of resources—subsidized by the higher than competitive fees paid by travelers. Needless to say the CAB was forced to abdicate its strategy.
Recent articles [19, 20] have documented the supply and distribution of professional radiologist resources. The findings indicate a relative shift in radiologists away from academic practice (inpatient oriented) toward private outpatient centers. This is, at least in part, likely a reflection on reimbursements relative to work effort [21]. The distribution of radiologist resources is consistent with the argument we set forth concerning cost accounting, patient mix, payer mix, and response to incentives. However, future studies correlating payment incentives and the distribution of work could yield further insights into risks and rewards in radiology.
Where To Now?
In radiology, hospital-based practices are more likely to service patients with serious disease. Indeed, hospitalized patients have been associated with greater radiology resource utilization [5]. At the same time, it is well known that hospitals are called on to deliver a disproportionate share of underreimbursed and unreimbursed care. This is undoubtedly adding to the recent crisis in academic medicine [22, 23]. The rationale for rebalancing payment rates in radiology is not complicated. Radiology remains a one-on-one, service-intensive specialty.
Practitioners should be encouraged to take on challenging patients, not rewarded for skimming low-cost, high-reward cases. Under the current system, providers are not necessarily paid more for doing more. Rebalancing rates would seek to attenuate the incentives to systematically avoid some patients while catering to others. We have discussed the theoretic risks to patient access and practice sustainability and shown empiric evidence in the context of whole-body CT screening, managing routes of referral, and the distribution of radiology resources.
It is relatively simple to argue in favor of more accurate patient costing, and hence more accurate reimbursements. It is far more difficult to successfully implement an enhanced costing scheme. Unfortunately, it would be expensive and tedious to continually detail the materials and services received by each patient, as required by a job-order approach.
We will have more success if we group patients according to patient type and empirically establish how patient type influences the cost of an imaging examination. The hope is to introduce patient-based refinements into the CPT system. This approach offers enhanced patient costing, potentially without the hassle of an overly complex costing system.
What do we mean by patient type? It is reasonable that patient characteristics such as indication for imaging, diagnosis, inpatient versus outpatient, age, and sex may affect radiology costs. The bottom line is that there is not enough evidence at the moment to determine how well any of these predict radiology costs. The science of predicting health care costs based on patient characteristics is addressed in a field of study known as risk selection and risk adjustment [24]. Unfortunately for us, researchers in risk adjustment have focused on patients' overall health care costs as seen from the perspective of insurers.
Briefly, of all the patient characteristics that may be used for radiology rate rebalancing, each has its unique advantages and disadvantages. Age and sex are the prototypical adjustors. However, the general medical literature suggests that age and sex are actually poor predictors of individual expense. We surmise that this is also true for radiology expense. Inpatient versus outpatient status may reasonably be used as an adjustor. This has the advantage that the information is easily coded and may help ease the shortage of academic radiologists.
Route of referral could be used. Under this scheme, referrals from an oncologist would yield a different rate of pay than referrals from, say, a urologist. Route of referral would be an easily coded, potentially powerful predictor of radiology expense. Diagnostic information is another potentially powerful adjustor, though data may be difficult to code or may be unreliable. Indication for imaging could also be a good predictor of expense, particularly since radiologists are already responsible for providing this information in their reports and bills.
A distinct yet related challenge is to determine how many patient types to define. As the number of defined patient types increases, costing precision will also increase but so will system complexity. It is unlikely that patient costing will ever capture true resource consumption. There will always be varying margins associated with differing patients. The goal of rate rebalancing is to temper this variability and to mitigate the incentives for patient selection.
In the end, the notion of patient-based payment rates in radiology requires more study before any scheme can be implemented. Many of the recently published discussions on departmental finance have focused on the role of clinical revenues in subsidizing academic pursuits. The tendered approach has been to increase patient throughput to increase clinical profits. Yet, the best means to understand the profitability of the clinical enterprise is to appreciate the financial effect of different patients on the practice. Academic radiology departments, which in all likelihood service patients with the lowest margins, should be keen to elucidate what drives costs in patient treatment. In this way, cost accounting can lead to rebalanced payment rates based on case complexity and can help ensure access for patients and sustainability for radiology practices.
References
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
