first quarter of the fiscal year 1967, 387,413 inpatient hospital claims had been reported by October 15, 1966, as approved for payment by the intermediary. Almost all of these claims (99 percent) are for reimbursement for care in short-stay hospitals. As would be expected, the average number of days per short-stay hospital claim is considerably less than in the long-stay hospitals: 10.3 days compared with 23.0 days in long-stay hospitals (general and special hospitals with average stays of 30 days or over; tuberculosis, psychiatric, and chronic disease hospitals; and Christian Science sanitoria). As the program continues, average hospital stays will undoubtedly be greater because the claims presented here are only for the first quarter of the program's operation-a period not long enough to reflect many long stays.

Total charges for the 387,413 claims amounted to approximately $167 million. Almost fourfifths 78 percent-of the total hospital charges were paid by the hospital insurance program. The deductibles and noncovered items on the bill account for differences between total charges and reimbursed amounts. The actual amounts reimbursed to hospitals are based on interim per diem rates that will be adjusted in the future on the basis of actual reasonable costs. The reimbursed amount per claim averaged $334 in short-stay hospitals and almost twice that amount in long-stay hospitals because of the considerably longer stays in the latter. Total charges averaged $29 per day in long-stay hospitals, compared with $42 per day in short-stay hospitals.

Because reimbursement by the hospital insurance program was only for inpatient care beginning July 1, average stays per claim approved in the program's first month were considerably shorter than for the next 2 months. For example, the average length of stay in short-stay general hospitals was only 6.7 days in July, compared with 10.0 days in August and 11.7 days in September. Average stays in long-stay hospitals in each of the 3 months reported show a similar pattern. July claims obviously included a considerable number of stays for aged persons who were in hospitals on July 1, and August claims also included some who were admitted before the effective date of the program.

The small number of inpatient hospital claims for the month of July-8.5 percent of the 3-month

total-reflects the delays in transmittal of forms and claims at the beginning of the program rather than a small number of aged persons receiving inpatient hospital care during the month. Likewise, claims approved in September and recorded in the Social Security Administration tape record as of October 15 are 40 percent below the number for August because of the lags in reporting and recording the data. These data will be updated and revised each month and more complete information for the earlier months will be reported in future issues of the BULLETIN.

Analytical Studies

In addition to providing basic data on program operations on a recurrent basis, the statistical system has been designed to provide the basis for a variety of analytical studies to evaluate the program and measure its performance. These studies will be concerned with assessment of program operation and achievements in terms of the program goals: to protect the aged person against the catastrophic costs of hospitalization and illness and to provide quality hospital and medical care in the most efficient and economical manner.

The statistical system has been designed to make possible studies to analyze the utilization experience in relation to the demographic data available from the eligibility records, to the charges and costs of providers, and to carrier operations. Such studies will provide the knowledge necessary for appraising the program's attainment of its purposes and for determining the need for legislative changes to facilitate effective operation.

These studies can be categorized in three main groups: utilization and costs of health services, effectiveness of administration, and questions relating to specific provisions of the law. Several examples of the type of analytical studies to be undertaken are sketched below.

STUDIES OF UTILIZATION AND COSTS OF HEALTH SERVICES

The availability of a population base permits the calculation and presentation of a wide variety of utilization rates for population subgroups. In addition to the utilization data, the basic statistics

include data on total and covered charges for the various types of services. The potentialities for combining and cross-classifying utilization data by characteristics of beneficiaries and providers of services open new vistas for analysis and study of variations in patterns of use for hospital and medical services and the factors affecting such variation, including geographic and certain demographic differences.

The availability of statistical data on utilization of hospital and medical services for each individual beneficiary provides the opportunity for longitudinal studies of the patterns of covered services received by individuals over a period of time. Use of services by specific groups of individuals, beginning at age 65 (or the start of the program), can be followed and studied in terms of the characteristics of the beneficiaries and the type and extent of services received. A tie-in with the basic record system of the Social Security Administration will make possible a unique opportunity for analysis of the medical history after age 65 in relation to the person's work history in covered employment, age at retirement, and benefit status.

The considerable fund of data relating to the characteristics of the providers of service, their reimbursement rates, and the utilization of their services provides the basis for a variety of studies. Geographic differences in reimbursement rates will be analyzed in terms of the providers and the services provided. Studies will be undertaken to determine where beneficiaries in a given geographic area receive their medical services and where hospitalized persons come from.

STUDIES OF EFFECTIVENESS OF ADMINISTRATION

The central statistical system will provide the data required for a variety of studies of the program's administration. Under the hospital insurance plan each group of providers, or association of providers in behalf of their members, has nominated a national, State, or other public or private agency or organization to serve as fiscal intermediaries between themselves and the Federal Government. The intermediary determines the amount of payments due on receipt of bills from hospitals and other institutional providers and makes such payments.

Studies will be undertaken to analyze the operations of the intermediaries with respect to the

effective operation of the program. Differences among carriers in their operating costs, methods of payments, procedures for claims review, billing lags, and other administrative responsibilities will be reviewed and analyzed in detail.

Where payment is on the basis of charges for physician services and medical and other health services, the intermediaries or carriers are to take action to assure that the charge on which the reimbursement is based is reasonable and not higher than the charge used for reimbursement in behalf of the carriers' own policyholders or subscribers under comparable circumstances. In determining reasonable charges, the carriers are to consider the customary charges for services generally made by the physician furnishing the covered services, as well as prevailing charges in the locality for similar services.

Analysis will be made of the geographic variation in actual charges for physician services for comparable procedures in relation to their characteristics and those of beneficiaries. The studies will give some clues on the extent to which the carriers are effectively carrying out this important function.

Hospitals and extended-care facilities participating in the hospital insurance program must have a utilization review plan in effect, providing for review, on a sample or other basis, of the following: admissions of beneficiaries of the hospital insurance program to the institution, length of stays, and the medical necessity for services provided. Statistical studies analyzing the variations in institutional stays for comparable diagnostic categories in terms of geographic location and types of institution will assist in evaluation of the utilization review process.

STUDIES RELATING TO SPECIFIC PROVISIONS

The 1965 amendments to the Social Security Act include several special provisions embodying unique concepts in health insurance programs, the effects of which will be studied and analyzed. For example, inpatient hospital and extended-care services within specified limitations are provided under the law for each spell of illness. The term "spell of illness" is defined as beginning the first day (not in a previous spell of illness) in which an individual is furnished covered inpatient hospital or extended-care services and ending with the

last day of the first period of 60 consecutive days during which he was not an inpatient in a hospital or extended-care facility. Studies of the impact of this requirement will be made in terms of the average duration of spells of illness, number of beneficiaries who exhaust benefits during single spells of illness, average duration of time between exhaustion of benefits and beginning of a new spell of illness, and the proportion of total costs of care in hospitals not covered because of the spell-of-illness concept.

Payments to providers of service under the hospital insurance program are made on the basis of reasonable costs for the services furnished. The costs of services in hospitals and extendedcare facilities vary widely from one institution to another, reflecting differences in quality and intensity of care. Reimbursement rates and the method for determining reasonable costs will be analyzed in terms of geographic variations, type of facilities, and services provided.

One of the conditions of participation for an extended-care facility is that it must have a transfer agreement with at least one participating hospital (except under special circumstances). A transfer agreement is one that provides, in writing, for the transfer whenever such action is medically appropriate, as determined by the attending physicians. Analysis will be made of the various types of transfer agreements, the implementation of this requirement on a geographic basis, and its effect on patterns of care received under the program.

There is a lifetime limit of 190 days of covered services in psychiatric hospitals. Psychiatric care in general hospitals, however, does not count against the 190-day lifetime limit. Statistical study will be undertaken to determine the number of persons who exhaust these benefits, the number and extent of psychiatric services in general hospitals, and emerging trends in this area.

REPORTING PLANS

Many of the analytical studies described above cannot of course be carried out until the health insurance program for the aged has been in operation for some time. On the extent of services and on charges, the Current Medicare Survey is designed to yield program data on a national basis in advance of the detail to be obtained from the record. Current plans for reporting these survey data as well as basic data on program operations include publication of monthly, quarterly, and annual data in the BULLETIN and in special releases and reports by the Office of Research and Statistics as the data become available and the studies are completed.

The need for statistical data by agencies, organizations, and researchers outside the Social Security Administration will also be taken into account in our tabulation plans. In reporting all program data, the Social Security Administration's general policy relating to confidentiality will be continued. Information will not be released identifying individual beneficiaries and their specific utilization of services under the

program.

The health insurance program for the aged will have a significant impact on the entire structure of the organization and financing of health services in the country in addition to its impact on the ability of the aged individual beneficiary to meet the costs of needed hospital and medical care. The broad scope of benefits affecting this large population group and the financing of these benefits will require substantial adjustments in the entire system of health services, involving not only the aged beneficiary but the remainder of the population. In addition to the analytical studies outlined above, a broad research program will be undertaken to measure the impacts on both public and private programs, identify and define program gaps and unmet needs, and examine and evaluate the economic consequences of the program.

by SAUL WALDMAN*

When a worker suffers a temporary loss of wages because of a work-connected illness or injury, his wage loss will usually be compensated under the workmen's compensation law of his State. If the loss of income results from an illness or injury incurred off the job, which is far more common, the possibilities of wage replacement vary, depending usually on the location of his work or the protection voluntarily provided by his employer. Protection against nonoccupational sickness is provided under employers' sickness insurance and sick-leave plans and by law for workers in four States and the railroad industry. The Social Security Administration compiles data annually that measure the amount of protection provided through cash sickness-benefit plans.

same in 1965 as in 1964. About 32.6 million private wage and salary workers, or almost 61 percent of the total, had some kind of protection under public programs, private insurance plans, or formal sick-leave arrangements.

The 1965 data seem to indicate little change in recent trends in the extent of protection against income loss arising from illness. The proportion of potential income loss replaced rose continuously from 1948 (the first year these data were compiled) until 1959. Since then, the proportion has leveled off, fluctuating within a range of 29–30 percent.

THE VALUE of time lost from work because of short-term sickness in 1965 is estimated at $11.2 billion. Of this potential loss, about $3.3 billion or 29.5 percent was replaced under various public and private insurance and sick-leave plans.

The incidence of illness among workers in 1965 was somewhat higher than the relatively low rate in 1964, according to estimates derived from the National Health Survey. Because of the higher sickness rate and the substantial rise in income subject to loss that reflected the relatively high level of economic activity during the year, the potential income loss in 1965 was nearly $1 billion more than the amount for 1964. Since the potential income loss increased at a somewhat greater rate than aggregate sickness benefits did, the proportion of income loss replaced declined slightly to 29.5 percent. Among the various types of sickness protection, voluntary group insurance and sick leave showed substantial increases, but individual insurance and government temporary disability insurance payments changed only slightly.

The proportion of workers in private industry covered under cash sickness plans was about the

'Office of Research and Statistics. Earlier articles in this series have appeared in the January issue of the Bulletin.

MEASURING INCOME LOSS

Estimating Income Loss From Short-Term Sickness

The estimates of income loss used in this series are designed to cover the loss of current earnings during the first 6 months of nonoccupational illness or injury, including loss during the first 6 months of a long-term disability. This concept of short-term income loss is based on traditional usage developed in connection with accident and sickness insurance practices and later adapted by government disability insurance programs. In designing various types of insurance policies and programs, the 6-month period was considered a useful administrative device for distinguishing between short-term and long-term disability. Disability that has already lasted such a substantial period of time is customarily dealt with under plans designed for long-continued or permanent disability. The first 6 months of any illness is thus included in the short-term category regardless of the eventual span of illness. From the viewpoint of the actual nature of disability, these distinctions are of course arbitrary. There is little evidence to indicate that 6 months represents any significant point in the distribution of the duration of disabilities.

The term income loss, as used in this article, refers to the value of potential as well as actual loss. It includes, for example, income that would have been lost if not replaced under a sick-leave plan that continues wages and salaries during periods of illness or under another type of arrangement. Sick leave is counted among the types of benefits that offset the potential wage loss.

Estimates of the number of days of work lost in the year are computed separately for the various components of the labor force-wage and salary workers in private employment, Federal civilian employees, State and local government workers, and self-employed persons. The standard disability rates for each group are modified to reflect year-to-year variations in sickness rates, based on an index derived from the National Health Survey data on days of disability. The sickness index uses 1958 as the base year, with an index number of 100. The estimates of the amount of income loss are obtained by applying the modified sickness rates for each employment group to income data for that group.

The rate of sickness among workers, as measured by the sickness index, rose from 101 in 1964 to 103 in 1965-a moderate degree of sickness. The index has ranged between 97 and 107 since 1958.

The total income loss from nonoccupational short-term sickness in 1965 was $11.2 billion, a rise of $990 million or 9.7 percent. A major factor in the increase, in addition to higher sickness rates, was the growth in the amount of earnings subject to loss, which rose 7.5 percent, reflecting the relatively high level of employment and earning during the year. Each of the major laborforce groups-private industry employment, public employment, and self-employment-contributed to the rise in income loss, with each group showing an increase of 9-11 percent.

Total Economic Cost of Illness

Estimates of the loss of earnings resulting from short-term illness, similar to those included in this article, are presented in a recently issued study of the costs of illness. These data for short

'Dorothy P. Rice, Estimating the Cost of Illness, Public Health Service, Health Economics Series Number 6, May 1966.

1 Short-term or temporary non-work-connected disability (lasti more than 6 months) and the first 6 months of long-term disability. 2 Beginning 1960, data include Alaska and Hawaii.

Annual payrolls of wage and salary workers in private emplo from Survey of Current Business, Department of Commerce, multipl 7 (estimated average workdays lost per year due to short-term sic and divided by 255 (estimated workdays in year).

Total annual payrolls of wage and salary workers in industries c by temporary disability insurance laws in Rhode Island, California Jersey and New York and in the railroad industry, multiplied by divided by 255.

5 Difference between total loss for all wage workers in private emplo and for those covered by temporary disability insurance laws.

Federal civilian payroll in United States from U.S. Civil Service mission, multiplied by 8 (estimated average workdays lost per year short-term sickness) and divided by 260 (scheduled workdays in yea 7 Annual wage and salary payrolls of State and local government emp from Department of Commerce data (see footnote 3), multiplied (estimated average workdays lost per year due to short-term sicknes divided by 255 (estimated workdays in year).

Annual farm and nonfarm proprietors' income from Departm Commerce sources cited in footnote 3, multiplied by 7 (estimated i loss days per year due to short-term sickness) and divided by 300 (esti workdays in year).

Computed as for earlier years, then adjusted to reflect changes in si experience (average number of disability days) in 1959-65, as reported National Health Survey.

term illness are one component of the estim presented in the study, which represents a prehensive systematic approach to estimating total economic costs resulting from illness, bility, and premature death. The report deve methodology for estimating the direct annual of morbidity and death, such as the prevent detection, and treatment of illness, and the indi costs including the loss of manhours and earni Also included are estimates of the present v of future earnings lost because of mortality 1963. Most of the estimates are distributed age, sex, and diagnosis.