The tables below illustrate the costs for health plans and health care providers to implement the standards and the savings that will occur over time as a result of the HIPAA administrative simplification provisions. All estimates are stated in 1998 dollars -- no adjustment has been made for present value.
The tables are extracted from a report prepared by our actuaries, who analyzed the impact of the HIPAA administrative simplification provisions. Using standard actuarial principles, they utilized data from a wide range of industry sources as a base for their estimates but revised them as needed to precisely reflect the impact of the legislation. For example, the number of health care providers and percentage of EDI transactions were adjusted to reflect expected 1998 levels. Where data were not available (for example, the percentage of EDI billing for hospices), estimates were developed based on assumptions. Where data from multiple sources were in conflict, the various sources were considered in developing an independent estimate. These processes are complex and are described in detail in the actuaries’ report, both in narrative form and in footnotes to tables. The report is too voluminous to publish here, and it is not feasible to describe the processes used to arrive at each and every number. We are presenting here the data that are most critical to assessing the impact of HIPAA administrative simplification provisions and a general description of the processes used to develop those data. The full actuarial report is available for inspection at the HCFA document room and at the following web site: http://aspe.hhs.gov/admnsimp/.
The costs are based on estimates for the cost of a moderately complex set of software upgrades. The range of costs that health plans and health care providers will incur is quite large and is based on such factors as the size and complexity of the existing systems, ability to implement using existing low-cost translator software, and reliance on health care clearinghouses to create standard transactions. The cost of a moderately complex upgrade represents a reasonable midpoint in this range. In addition, we assume that health plans and health care providers with existing EDI systems will incur implementation costs related to manual operations to make those processes compatible with the EDI systems. For example, manual processes may be converted to recognize standard identifiers or to produce paper remittance advices that contain the same data elements as the EDI standard transaction. We have estimated those costs to equal 50 percent of the upgrade cost. Health care providers that do not have existing EDI systems will also incur some costs due to HIPAA, even if they choose not to implement EDI for all of the HIPAA transactions. For example, a health care provider may have to change accounting practices in order to process the revised paper remittance advice discussed above. Health plans must accept HIPAA transactions via EDI, but not all health plans will be called upon to accept all HIPAA transactions. For example, some health plans process only dental claims, while others process claims for institutional and noninstitutional services. We have assumed the average cost for non-EDI health care providers and health plans to be half that of already-automated health care providers and health plans.
Savings are based on the estimated increase in EDI attributable to the HIPAA administrative simplification provisions, multiplied by a per transaction savings for each type of transaction. Our estimates are much lower than those included in the WEDI report, primarily because we only recognize savings that would not have occurred without the legislation. While some industry estimates of gross savings (not net of costs) have been as high as $32.8 billion over five years, we believed it was important to utilize the most conservative assumptions possible. It is important to view these estimates as an attempt to furnish a realistic context rather than as precise budgetary predictions. Our estimates also do not include any benefits attributable to qualitative aspects of Administrative simplification, because of the lack of reliable data. (For example, we do not attempt to put a dollar value on improved public health practices that will result from implementation of standard identifiers.) We strongly encourage comments on how to quantitatively and qualitatively measure the efficiencies realized as a result of the HIPAA administrative simplification standards.
More detailed information regarding data sources and assumptions is provided in the explanations for the specific tables.
Table 1 below shows estimated costs and savings for health plans. The number of entities is based on the WEDI report, Department of Labor data, and various trade publications trended forward to 1998. The cost per health plan for software upgrades is based on the WEDI report, which estimated a range of costs required to implement a fully capable EDI environment. The high- end estimates ranged from two to ten times higher than the low-end estimates. We have used the lower end of the estimates in most cases because, as explained above, HIPAA does not require as extensive changes as envisioned by WEDI. The estimated percentages of health plans that accept electronic billing are based on reports in the 1997 edition of Faulkner & Gray’s Health Data Directory (5). The total cost for each type of health plan is the sum of the cost for EDI and non-EDI plans. Cost for EDI plans is computed as follows:
Total Entities x EDI % x Average Upgrade Cost x 1.5 (NOTE: As described above, the cost of changing manual processes is estimated to be half the cost of system changes.)
Cost for non-EDI plans is computed as follows:
Total entities x (1 - EDI %) x Average Upgrade Cost x .5 (NOTE: As described above, cost to non-EDI health care providers is assumed to be half the cost of systems changes.)
The $3.9 billion in savings is derived from Table 4, and represents savings to health plans for the first five years of implementation. The assumptions related to these savings are contained in the explanation to Table 4. The savings have been apportioned to each type of health plan based on the ratio of that health plan type’s cost to the cost to all health plans. For example, a plan type that incurs ten percent of the costs would be assigned ten percent of the savings. We acknowledge that this is an imprecise method for allocating savings. We have not been able to identify a reliable method for allocating savings to specific types of health plans but nonetheless believed that it was important to present costs and savings together in order to provide a sense of how the HIPAA administrative simplification provisions would affect various entities.
Table 2 illustrates the costs and savings attributable to various types of health care providers.
The number of entities (practices, not individual health care providers) is based on the 1992 Census of Services, the 1996 Statistical Abstract of the United States, and the American Medical Association survey of group practices trended forward to 1998. Estimated percentages of EDI billing are based on the 1997 edition of Faulkner & Gray’s Health Data Directory or are actuarial estimates.
The cost of software upgrades for personal computers (PCS) is based on reports on the cost of software upgrades to translate and communicate standardized claims forms. The low end is used for smaller practices and the high end for larger practices with PCS. The estimate for mainframe upgrade packages is twice the upper end for PCS. The cost per upgrade for facilities is ours after considering estimates by WEDI and estimates of the cost of new software packages in the literature. The estimates fall within the range of the WEDI estimates, but that range is quite large. For example, WEDI estimates the cost for a large hospital upgrade would be from $50,000 to $500,000. For an explanation of the method for computing Total Cost, see the explanation for Table 1.
The $3.4 billion in savings is derived from Table 4 and represents savings to health care providers for the first five years of implementation. We have included them here to provide a sense of how the HIPAA administrative simplification provisions would affect various entities. As in Table 1, the savings have been apportioned to each type of health care provider based on the ratio of that health care provider type’s cost to the cost to all health care providers.
Table 3 shows the estimates we used to determine the portion of EDI increase attributable to the HIPAA administrative simplification provisions. The proportion of claims that would be processed electronically even without HIPAA is assumed to grow at the same rate from 1998 through 2002 as it did from 1992 to 1996, except that the rate for hospitals, which is already high, is assumed to grow at one percent annually instead of the two percent that was observed from 1992-1996. The proportion of “other” provider claims is high because it includes pharmacies that generate large volumes of claims and have a high rate of electronic billing.
The increase attributable to HIPAA is highly uncertain and is critical to the savings estimate. Our actuary arrived at these estimates based on an analysis of the current EDI environment. Because the rate of growth in electronic billing is already high, there is not much room for added growth. On the other hand, much of the increase that has already occurred is attributable to Medicare and Medicaid; private insurers and third party administrators still have fairly low rates of electronic billing and may benefit significantly from standardization.
Table 4 shows the annual costs, savings, and net savings over a five-year implementation period. We assume that the costs will be incurred within the first three years, since the statute requires health plans other than small health plans to implement within 24 months and small health plans to implement within 36 months. As each health plan implements a standard, health care providers that conduct electronic transactions with that health plan would also implement the standard. We assume that no savings would accrue in the first year, because not enough health plans and health care providers would have implemented the standards. Savings would increase as more health plans and health care providers implement, exceeding costs in the fourth year. At that point, the majority of health plans and health care providers will have implemented the standards, and costs will decrease and benefits will increase as a result.
The savings per claim processed electronically instead of manually is based on the lower end of the range estimated by WEDI. We have used $1 per claim for health plans and physicians, and $.75 per claim for hospitals and other health care providers. These estimates are based on surveys of health care providers and health plans. Savings per EDI claim are computed by multiplying the per claim savings times the number of EDI claims attributed to HIPAA. The total number of EDI claims is used in computing the savings to health plans, while the savings for specific health care provider groups is computed using only the number of EDI claims generated by that group (for example, savings to physicians is computed using only physician EDI claims).
WEDI also estimated savings resulting from other HIPAA transactions. The savings per transaction was higher than the savings from electronic billing, but the number of transactions was much smaller. Our estimates for transactions other than claims were derived by assuming a number of transactions and a savings per transaction relative to those assumed for the savings for electronic billing (see table 4a). In general our assumptions are close to those used by WEDI. One major difference is that we derived the number of enrollment/disenrollment transactions from Department of Labor statistics. We used their estimate of the number of events requiring a certificate to be issued, which includes such actions as starting or leaving a firm, children “aging out” of coverage and death of policyholder. That estimate is about 45 million events. We used WEDI’s estimate that the savings per transaction is about half that of billing transactions.
We also assumed that savings could be expected from simplifications in manual claims. The basic assumption is that the savings are ten percent (per transaction) of those that are projected for conversion to electronic billing. However, it is also assumed that the standards only gradually allow health care providers and health plans to abandon old forms and identifiers because of the many relationships that have been established with other entities that will require a period of overlap.
NOTE: Figures do not total due to rounding.
Table 4a shows the savings per nonclaim transaction as a multiple of claims savings per transaction and the ratio of transactions to number of claims. These values were used to determine the savings for nonclaims transactions.