Economic Factors in Hospital Planningin Urban Areas

HERBERT E. KLARMAN

THE ABSENCE of planning is not, per se,proof that chaos or anarchy prevails.

Whether planning is called for depends on thegood (commodity or service) in question and on

the circumstances surrounding its productionand consumption.Forms of Economic InterventionFor a large variety of goods and services we

tend in this country to accept the operating re¬

sults of the market. This decision is rootedpartly in faith in the beneficience of AdamSmith's discovery, the invisible hand (by whichthe individual in pursuing his own interests isalso promoting the general interest). In addi¬tion, however, it partly reflects confidence in thesuperiority of decentralized decision making,something that the socialist countries have re¬

cently come to acknowledge. It is buttressed bythe willingness of society to redistribute incomethrough various devices when the results of themarket offend its sense of fairness.

Society may intervene in economic affairsthrough additional devices (1). These devicesare listed here, without elaboration, in order toconvey their number and variety. Thus, it enacts

Dr. Klarman is professor of public health adminis¬tration and political economy, Johns Hopkins Uni¬versity, Baltimore, Md. The article is based on a

paper presented before the medical care sectionpanel at the 94th annual meeting of the AmerieanPublic Health Association, held in San Francisco,Calif., November 3, 1966.

laws to safeguard persons and property. Anti-trust laws are meant to inhibit both the powerof monopolies and their inefficiencies. Deliverysystems that are costly to duplicate, such as thetelephone or the electricity networks, are ac-

corded public utility status, to which statedprivileges and responsibilities attach. Eegula-tion and licensure of certain categories of per¬sonnel are intended to safeguard the consumer.

Subsidies (cash graiits) and tax credits or de-ductions may be employed as inducements toencourage desired courses of action. SometimesGovernment serves as the producer of servicesthat it sells (the post office) or as the purchaserof services it pays for (hospital care). Planningis another vehicle of social control.The dictionary defines planning as devising a

scheme for doing, making, or arranging. A planrefers to any detailed method, formulated be-forehand, for doing or making something. Astatement of general principles does not consti-tute planning.This paper focuses on areawide planning for

hospital care because no other concrete body ofplanning experience from the health field isavailable to us in this country. Plans for mentalhealth and mental retardation services are justcoming off the drawing boards. Currentlymoney is being allocated for drawing plans forthe regional medical programs which derivefrom the De Bakey commission's report on

heart disease, cancer, and stroke. It is knownthat these programs will encourage and facili¬tate cooperative arrangements among providers

Vol. 82, No. 8, August 1967268-233.67-5

721

of service in a region. The contents of these pro¬grams will evolve in response to local initiativeand will vary among regions, depending on

needs and opportunities and on whether pri¬mary emphasis is given to the wider delivery oi

services created by medical discoveries or to im¬proving the overall quality of medical care.

Legislation authorizing comprehensive plan¬ning of health services by health departmentshas just been enacted.My analysis of planning for hospital care will

be limited to economic factors. Such an analysisis incomplete, of course, lacking the political,social, and physical elements that also enter intoplanning.

I strongly believe, however, that the analysisof a concrete body of experience, though incom¬plete, is more valuable than any amount of dis¬cussion of generalities. The successes, opportu¬nities, and failures of planning can only beappraised in the light of experience.

Public Concern About HospitalsWhy is there public concern for the proper

development of hospital services? The reason isthat from the very beginning hospitals have ab-sorbed large masses of social capital. In thiscontext, social capital includes both philan-thropic and governmental.

It is perhaps an accident of history that thepublic has furnished the physician's workshopwithout expense to him.something it has notdone for other professions in private practice.Certain factors, however, suggest that this pol¬icy may have some rational bases. Seventy toeighty years ago capital requirements for hos¬pitals loomed large relative to operating expen¬ditures, and investment in one represented a bigchunk of capital. Free care, or care at less thancost, for the poor (who represented a majorityof hospital patients) was the accepted mode.The education and training of new physicianswas, in turn, closely associated with care of thesick poor in the hospital.The existence of public concern, however, is

not a sufficient condition for action. Anothernecessary ingredient is the possibility of doingsomething about the problem. A community or

neighborhood can, with its own resources, builda local hospital or enlarge an existing one, or re-

frain from doing either. By contrast, a commu-

nity's ability to influence its supply of physi¬cians appears to be small.From the outset, planning for hospital care

has been carried on separately from other socialplanning. City planning agencies have beeneither unwilling or unable to assume responsi¬bility for hospital planning. One can only specu-late on the reasons for their reluctance. Twofactors appear to have been especially impor¬tant. One is the complexity of hospital services.Given the difficulties of measuring the qualityof the output, the tendency is to resort to pro¬fessional.medical.judgment. The second isthe mixed nature of the hospital economy.gov¬ernmental, voluntary (nonprofit), and proprie¬tary (for profit). City planners are accustomedto plan for facilities under a single, govern¬mental form of control.

Economic Factors in PlanningAmong the economic factors that support

community planning for hospital care are thefollowing: (a) the waste of a low rate of occu-

pancy; (b) adapting to random variation inadmissions; (c) the trend toward larger hospi¬tals; (d) the indivisibility of equipment andteams; (e) the Hill-Burton program, rising unitcosts, and Roemer's law; (/) the long life of thephysical plant; (g) changes in the populationof cities and the growth of suburbs; and (h)Federal grants-in-aid.Low rate of occupancy. The high proportion

of overhead to total hospital cost was recognizedby accountants such as Charles Roswell and byadministrators long before it was measured byeconomists (2). A low rate of occupancy reducesincome much more than expenditures and can

pose a threat to the financial stability of thehospital.During the depression of the thirties, Govern¬

ment hospitals were overcrowded while volun¬tary hospitals had vacant beds. (Haven Emer-son's "Hospital Survey for New York"documents this point exhaustively.) This situa¬tion seemed particularly irrational, being con-

trary to the interests of all concerned. Theobvious remedy was to provide all patients equalaccess to all hospitals, regardless of who paidthe hospital bill. This policy also appealed on

another ground: a hospital open to all classes

722 Public Health Reports

of patients has a superior ability to serve itscommunity. (The latter point is still valid, ofcourse, and has gained in relevance with theenactment of Medicare.)With high overhead costs, a low rate of occu¬

pancy leads to a financial deficit. Therefore, itis a sufficient deterrent to overbuilding to in-form every hospital of events, plans, and prob¬able developments elsewhere which are likely toresult in overbuilding in the aggregate. Theplanning agency is in a better position to ascer¬

tain such information and to disseminate itthan any individual hospital.Random variation in admissions. One of the

chief contributions of operations research to thehealth field is its exploration of the applicationof stochastic (random) processes to hospi¬tals (3, If). A formal, systematic explanation ofthe persistence of average rates of occupancybelow 100 percent is only one consequence.In addition, various devices to stabilize hos¬

pital patient load.and to raise average occu¬

pancy.have been examined or suggested (5,6),such as postponement and improved schedulingof elective admissions, replacement of largewards with small bedrooms, designation ofswing beds between intensive and intermediatecare units in a progressive care facility, occa-

sional attempts to end the physical separationof maternity patients, and recommendations totransfer excess patients to other hospitals. Al¬though such transfers are customary from pri¬vate to governmental hospitals, they rarely takeplace in the opposite direction.

All these devices except for interhospitaltransfers can be introduced within an individ¬ual hospital at the wish of its management andprofessional staff. The transfer of patientsamong hospitals, however, encounters thetroublesome problem of staff appointments forphysicians (dealt with later).Trend toward larger hospitals. In the large

city, interest has focused much more on thedeficiencies of small hospitals than on the pos¬sible inefficiencies of large ones. A rule of thumbI have learned from several administrators isthat the best size of hospital is the current sizeof his hospital plus 100 or 200 beds, dependingon the administrator's assessment of prospectsfor financing an expansion.On theoretical grounds alone, one can argue

in favor of a U-shaped long-term cost curve

for hospitals. On the one hand, specializationand division of labor result in declining unitcost as the scale of output increases. Beyond a

certain point, however, complexities of manage¬ment intrude and coordination of efforts be¬comes more difficult, so that unit cost rises.Application of the theoretical model to realdata is complicated, unfortunately, by differ¬ences among hospitals in range, complexity, andquality of services and by differences in salarylevels and educational programs. Various at¬tempts have been made to deal with these prob¬lems in order to determine the relationshipbetween output and cost (7, 8), and progressis being made. It is only fair to say that a final,definitive answer is not yet at hand.Economic analysis apart, small hospitals are

unable to meet two other criteria for a satis-factory modern hospital. They cannot concen-

trate enough patients for teaching, and theycannot be truly general in the patients theyserve and the services they render. These con¬

siderations, rather than economy, may havebeen decisive in fostering the movement in citiesagainst small hospitals of say 100 beds or so.

While the average size of hospitals has in¬creased, no hospital, however small, has beendebarred from caring for any category of pa¬tients. Moreover, a small hospital has fre¬quently ceased to be one by expanding. Hos¬pitals of larger size permit a concentration ofpatients for the convenience of physicians. Insum, this policy poses no disadvantage to pro¬viders of service, except possibly to hospitalsthat are unable to expand.Two sets of objections can be advanced

against the trend toward larger hospitals. Pa¬tients and prospective visitors may prefershorter travel time to one of the more numerous

smaller hospitals over longer travel to fewerand larger hospitals. In the production ofgoods, the lowest cost for a specified quality isan unexceptionable objective. In the productionof a service, the consumer must travel to theplace where it is produced (or less often, theprovider of services visits the customer). Thecost of production is only part of the real costinvolved, travel time and inconvenience beingothers (9).The second objection is that the optimum size

Vol. 82, No. 8, August 1967 723

of hospital for inpatient services may differfrom that for outpatient services. When thepatient has a family physician, there is lessneed for all medical services to be integrated ata single facility than when the patient dependson that facility exclusively. The original basisfor promoting integration of medical care serv¬

ices was to assure continuity of care and toavoid fragmentation and the poor quality ofcare associated with it. More recently, integra¬tion of services is also intended to help certainpeople who are regarded as incapable of mak¬ing good choices in buying health services.

Indivisibility of equipment and teams. Thehospital today has much more expensive equip¬ment than formerly and employs large special¬ized teams to perform certain diagnostic andtherapeutic procedures. Good, almost ubiqui-tous, examples of facilities that come in fairlylarge units are cobalt bombs for radiationtherapy, teams for open heart surgery, and.just emerging.renal dialysis units for chron-ically ill patients. The costs of larger pieces ofequipment are given in an earlier article ofmine (10); the cost of chronic renal dialysisis estimated at $15,000 a year.To serve but a few patients a facility must be

established that could serve 10, 20, or even 100patients at relatively little additional cost.

When many such facilities are set up in a com¬

munity, the average workload for each is small(11) and the unit cost high. Moreover, theskills of the personnel may deteriorate throughdisuse.An obvious remedy is to restrict the number

of facilities in an area. Some planners expectthat knowledge of the facts would lead hospi¬tals to cooperate in meeting the community'sneeds. Failure to cooperate is regarded as a fail¬ure to understand or as the unfortunate by-product of institutional vanity.This view of the situation may be too simple,

in my opinion, for at least two reasons. When a

hospital establishes a specialized service facil¬ity, the physician associated with it who is pro-fessionally qualified to use the facility benefits.A decision not to establish the facility in thephysician's hospital deprives him of income andof the continuing learning experience on whichhis specialized skills depend. Moreover, if hos¬pital A establishes such a facility, not only does

it and its staff gain while hospital B and itsstaff lose, but the community may incur an ad¬ditional loss through the deterioration of theskills of hospital B's staff and the obsolescenceof their knowledge. These losses can be averted,however, by a policy of selective duplication ofhospital staff appointments for physicians. (Un¬der this policy, not all physicians but only phy¬sicians who require access to the specialfacilities.which are to be located in a smallnumber of hospitals.would have appointmentsto staffs of hospitals other than their own.)The presence of a facility or program has

spillover effects for the other parts of an insti¬tution. Renal dialysis is intimately connectedwith advances in kidney transplantation, forexample. Radiation therapy is only one of themodalities applied in treating cancer.

Let us consider a more common facility, theobstetrical service. In many hospitals its rateof occupancy is low. Yet the presence of such a

facility affects the strength of the pediatrics de¬partment, the gynecology service, and internand nurse training. A service that is too costlyin terms of unit cost may make sense in termsof the overall mission of a hospital, once it isdetermined that this hospital should continuein operation. A decision by a hospital to roundout its services tends to be both self-confirma¬tory and cumulatively reinforcing.For the first time in this analysis one en-

counters possible conflicts of interest betweenthe individual hospital and the larger com¬

munity, the individual hospital being con¬cerned with overall institutional strength andthe community seeking to minimize the totalcost of a particular service. The hospital mayexaggerate the adverse spillover effects of fail¬ure to establish a certain facility. In addition,the hospital tends to assume little responsibilityfor the quality of medical care in the commu¬

nity outside its walls. Decisions on its staff ap¬pointments of physicans are made withoutregard for services supplied to ambulatorypatients.In its present dimensions, the problem of hos¬

pital appointments for the visiting staff hasemerged only within the past generation. Thepresence of a resident staff, and more recentlyof a full-time clinical staff, reduces the value tothe hospital of the voluntary attending staff.

724 Public Health Reports

At the same time the staff appointment no longerserves as the vehicle for training toward spe¬cialty practice, so that the practicing physicianis not so willing as formerly to give the hospitalhis time and energies (12).The regional medical programs for heart dis¬

ease, cancer, and stroke may substantially affectthis situation. On the one hand, in the hospitalselected to house a unique facility, the need foran equitable distribution of staff privileges tophysicians on other hospital staffs who need touse the facility will be made explicit. If publicfunds are employed, such a distribution of staffprivileges may become imperative. On the otherhand, there may ensue an increasing concentra¬tion of specialists in hospitals who will spendfull time on clinical services, rather than inresearch.Hill-Burton, unit cost, and Roemer's law.

Three sets of events have led to increased recog¬nition of the advantages offered by coordinatedcommunity action.The Hill-Burton program for assisting in the

construction of nonprofit (voluntary or govern¬mental.mostly the former) hospitals seems tohave accomplished its mission of bringing hos¬pital services to the rural population. The majorproblems are now in the cities, where moderni¬zation, improvement, and coordination are seen

as the imperative goals, rather than expansion.The unit cost of hospital care continues to rise

at a high rate. The explanation that hospitalsare catching up with other industries in wagesand working conditions fails to explain remain¬ing inequities, which require correction fromtime to time. Medical progress accounts for onlypart of the cost increase. Many economists be¬lieve that the most important factor is the con¬

tinuing lag of the hospital industry in achievinggains in productivity comparable to thoseachieved in the economy at large (13, 14).Somers, however, dissents (15). If this explana¬tion is correct, then, in the absence of substantialopportunities for automating many functionsof the hospital, the high rate of increase in thehospital's unit cost is likely to continue. Indeed,the more progressive the economy as a whole,as measured by increases in productivity, thegreater the increase in hospital unit cost.Another emerging factor is the increasingtendency to reimburse hospitals at actual cur¬

rent cost. Incentives to operate efficiently are

lacking.To keep expenditures for hospital care under

control, it would be necessary to curtail the use

of hospitals.Perhaps the major impetus for hospital plan¬

ning recently has come from still another source,namely, recognition that hospital use may notbe a good thing, per se, that relatively low use

need not reflect deprivation (16,17), and indeedthat the basis for determining the proper levelof use is constantly shifting, with the availablesupply of beds possibly exerting a strong influ¬ence on demand (18,19). One can no longer as-

sume that need, as medically determined, andfinancial ability to pay combine to create a

uniquely determined, appropriate criterion forplanning hospital use. Controversy still sur-

rounds the so-called Roemer's law.that underfinancing through prepayment newly built hos¬pital beds do not go empty (20-22). Acceptanceof the law, however, directly points to the de-sirability of limiting the total number of bedsin an area. If under third-party financing andvariable standards of hospital use, the threat ofvacant beds in the individual hospital has lostits potency, recourse to direct control or veto ofhospital building plans by an outside agencymay be necessary.The obvious desirability of avoiding recur-

ring, periodic requests for increases in the pre-miums of Blue Cross hospitalization planspoints in the same direction. State commis-sioners of insurance who review these applica¬tions recognize the advantages of financing theoperations of a smaller supply of beds.A^ain, a potential conflict of interests arises

between the individual hospital and the commu¬nity. It may make sense to exhort the public notto abuse health insurance benefits and not to askfor expensive amenities in the hospital; but it ispointless, if Roemer's law is valid, to exhorthospitals not to build. A firm No is required, as

in New York State, where areawide planning isnow compulsory instead of voluntary.

Life of physical plant. In depreciation tables,hospitals are shown with a life of at least 40years. Hospital facilities, therefore, must beplanned for a long time ahead. Since nobodyowns a clear crystal ball and the years betweenthe decennial censuses do not provide firm base

Vol. 82, No. 8, August 1967 725

lines, planning agencies usually compromiseand project bed requirements 10 to 15 yearsahead.Planning for hospital care always entails

planning for small geographic areas. Popula¬tion projection is difficult from a technicalstandpoint and always subject to outside forcesthat are neither well understood nor readilycontrolled. Allowing a margin for error is a safeprecaution. The demographic and socioeconomiccomposition of the population is even more un-

certain than its total number, and the implica-tions for hospital use of differences or changesin a population's composition are by no means

clear (23). The effects of future technologicalchange are certainly not known, other than thesteadily increasing ratio of square feet per hos¬pital bed. It is no exaggeration to say that a

large proportion of a given total of forecasts ofrequired hospital use are bound to be in error.

I infer that sound judgment as to directionwill probably be more helpful than precisearithmetic calculations. The most reliable de¬vice for minimizing the consequences of error isnot more careful long-range forecasting but pro-vision for as flexible use of facilities as possi¬ble (21^, 25). It should be recognized that a

plant built today will not be ideally suited -forthe conditions foreseen for a decade hence; nor

will the plant be precisely adapted to today'sconditions or volume of output. The extra costof flexibility represents a built-in diseconomyof operation (26).A major task of planning agencies, I con-

clude, is to search for, develop, and test devicesthat will promote the flexible adaptation and use

of hospital facilities over time.Population changes and shifts. Certain

changes in society at large affect planning forhospital care. The close tie between medicaleducation and the provision of free hospitalcare has kept the ratio of beds to population inthe central cities higher than it would otherwisebe. The institutions supplying hospital care

have also supplied care to indigent ambulatorypatients on an organized basis.With the advent of Medicare, and if liberal

Medicaid plans are adopted by the States, pa¬tients who receive free care will furnish a stead¬ily declining fraction of all teaching material.If private patients are used for teaching, then

a hospital with a large teaching program willbe freer than formerly to move from one site toanother. The advantage in quality of care thataccrued to an inner-city location will diminish.One alternative to removal will be an intensifiedconcern on the part of the hospital for renewalof the area in which it is located. Acting aloneto carry out renewal, a hospital can accomplishlittle. Acting in concert with other agencies andgroups, it may contribute to the conservationof its community.An independent hospital is likely to feel freer

to move than one that is a member of a religiousor ethnic network. In the case of the hospitalthat is a member of a network more of the fac¬tors that reflect the community's diverse needscan be brought to bear on its decisions, while tothe independent hospital some of these factorsappear to be beyond its ability to control.One of the important contributions of a plan¬

ning agency is to make relevant to the decisionsof an individual institution certain factors thatnormally do not concern it. By enlarging thearea of planning, benefits or costs accruing else¬where are converted into factors that may betaken into account explicitly.Federal grants-in-aid. Rufus Rorem has said,

"Cash is the prince of coordinators." At the timehe was referring to the leverage that could beexercised through construction grants. Federalmatching grants to areawide planning agencieswere still in the future.Matching grants have proved to be very in¬

fluential. Of 63 hospital planning councils nowin existence, 55 have been organized since 1962,when Federal monies for this purpose began toflow. Before 1962 the hospital planning move¬

ment was making slow headway. One agencywas founded in the 1930's, two in the 1940's,two in the 1950's, and three in the early 1960's.In 1962, 13 councils were organized, followedby 13, 5,11, and 13 in each of the next 4 years.(These data are from the Division of Hospitalsand Medical Facilities, Public Health Service.)

It is evident that few communities were will¬ing to spend their own money on hospital plan¬ning activities. In one city, for example, whenoutside funds were withdrawn, operations were

curtailed substantially.The history of areawide planning agencies

726 Public Health Reports

once more demonstrates the magic discoveredby the Rockefeller Foundation, namely, themultiplicative power of the outside dollar thatis to be matched locally. It is not possible togauge what would happen if Federal funds werewithdrawn or what will happen when grantsare no longer earmarked for hospital planning.It seems prudent to begin thinking, however,about evaluating the programs for planning andjustifying them.How to evaluate? We cannot conduct con¬

trolled experiments comparing what is withwhat would otherwise have been. One possibledevice is to set targets and to measure howclosely they are approached.How is one to justify? This effort is best

undertaken in the light of available and prob¬able alternatives. Why is the course recom¬

mended by the planning agency believed to bethe superior one? Its recommendations usuallyreflect a balancing of competing objectives.What are they, and what scale of importance isattached to each? The spelling out of objectivesand of their respective weights, along with a

presentation and evaluation of alternative waysto achieve the objectives, will enable the publicto judge the desirability of recommendations.

ImplicationsIn relation to total expenditures for hospital

care, the costs of maintaining a hospital plan¬ning agency are modest. Both the modal andmedian annual budgets for such an agency to¬day are less than $80,000 (according to Divisionof Hospitals and Medical Facilities, PublicHealth Service). The potential benefits.posi¬tive or negative.are large. If a planningagency is effective, it reduces the risk of a multi-tude of small or moderate mistakes but it raisesthe risk of a few large ones.

We must try to develop planning agencies forhealth care that will make sensible analyses ofthe important facets of a problem and advancerecommendations which are geared to flexibil¬ity. Such an agency must play several partssimultaneously. It needs to know almost everything concerning the community and its healthservices; it should also be aware of what it doesnot know about them. Such an agency shouldkeep abreast of the significant issues of healthpolicy, study some of these in depth, and make

recommendations on those for which a solutionis known or for which a solution is impera¬tive.whatever the current state of knowledge.A knowledgeable and sensitive planning agencywill be able to anticipate some of the problemsthat will emerge in the next few years, beforethey become acutely pressing. An effectiveagency will divorce itself from current fads andescape the awesome authority of arithmetic,relying instead on the skillful analyses of itsstaff and the mature judgments of its board.

SummaryEconomic intervention by Government can

take many forms. Planning is one of them. Inrecent years the Federal Government has sup¬ported the large-scale expansion of areawidehospital planning agencies in this country.The original basis for areawide hospital plan¬

ning in the 1930's was recognition that overheadcost contitutes a high proportion of total hos¬pital cost. It follows that a low rate of bed oc¬

cupancy reduces income much more than ex¬

penditures do and that large numbers ofvacant beds threaten the financial stability ofhospitals.Avoidance of duplication among hospitals of

expensive facilities and services requiresrecognition of the importance of selectiveduplication of staff appointments for physi¬cians. (Through selective duplication of ap¬pointments, facilities located in only a smallnumber of hospitals can be made available tophysicians on staffs of other hospitals who needto use them.)In a number of instances, possible conflicts of

interest are noted between the individual hos¬pital and the community. Under these circum-stances, voluntary cooperation may not be forth-coming. Perhaps the outstanding example ofsuch conflict is the possibility that additionalhospital beds will tend to be used wheneverthird-party financing of hospital care is pre¬dominant.

If the increase in hospital unit cost is largelyattributable to productivity gains in the hos¬pital lagging behind the rest of the economy,primary reliance in controlling hospital care ex¬

penditures must be placed on the control of hos¬pital use. A firm No to hospital building plansmay be required.

Vol. 82, No. 8, August 1967 727

The prospects for accurate forecasting of hos-pital use in a given local area are not bright.Planning should therefore concentrate on de-veloping devices that will permit flexible use offacilities.

REFERENCES

(1) Tobin, J.: National economic policy. Yale Uni-versity Press, New Haven, Conn., 1966, pp. 5-14.

(2) Feldstein, P. J.: An empirical investigation of themarginal cost of hospital services. GraduateProgram in Hospital Administration, Univer-sity of Chicago, Chicago, 1961, p. 49.

(3) Blumberg, M. S.: Distinctive patient facilitiesconcept helps predict bed needs. Mod Hosp 97:75, December 1961.

(4) Thompson, J. B., Avant, 0. W., and Spiker, E. D.:How queuing theory works for the hospital.Mod Hosp 94: 75, March 1960.

(5) Garrett, R. Y.: Seven-day work week improvesservices. Mod Hosp 103: 5, November 1964.

(6) Long, M. F.: Efficient use of hospitals. In Theeconomics of health and medical care, edited byS. J. Mushkin. Bureau of Public Health Eco-nomics, University of Michigan, Ann Arbor,1964, pp. 211-226.

(7) Lave, J. R.: A review of the methods used tostudy hospital costs. Inquiry 3: 57, May 1966.

(8) Yett, D. E., and Mann. J. K.: The costs of provid-ing long-term inpatient care: an econometricstudy. University of Southern California, LosAngeles, 1967. Mimeographed.

(9) Long, M. F., and Feldstein, P. J.: Economics ofhospital systems: peak loads and regional co-ordination. Paper delivered before AmericanEconomic Association, San Francisco, Decem-ber 1966.

(10) Klarman, H. E.: On the hospital. New Repub 149:9, Nov. 9, 1963.

(11) Crocetti, A. F.: Cardiac diagnostic and surgicalfacilities. Public Health Rep 80: 1035-1053,December 1965.

(12) Klarman, H. E.: Hospital care in New York City.Columbia University Press, New York, 1963,pp. 155-157.

(13) Brown, R. E.: The nature of hospital costs.Hospitals 30: 46, Apr. 1, 1956.

(14) Kilarman, H. E.: The increased cost of hospitalcare. In The economics of health and medicalcare, edited by S. J. Mushkin. Bureau of PublicHealth Economics, University of Michigan, AnnArbor, 1964, pp. 227-254.

(15) Somers, A. R.: The continuing cost crisis. Hos-pitals 40: 44, June 16, 1966.

(16) Densen, P. M., Balamuth, E., and Shapiro, S.:Prepaid medical care and hospital utilization.American Hospital Association, Chicago, 1958.

(17) Falk, I. S., and Senturia, J.: Medical care pro-gram for steelworkers and their families.United States Steelworkers of America, Pitts-burg, Pa., 1960.

(18) Shain, M., and Roemer, M. I.: Hospital costsrelated to the supply of beds. Mod Hosp 92: 71,April 1959.

(19) Roemer, M. I.: Bed supply and hospital utiliza-tion: A natural experiment. Hospitals 35: 36,Nov. 1, 1961.

(20) Airth, D., and Newell, D. J.: The demand forhospital beds. University of Durham, Neweastle-upon-Tyne, England, 1962.

(21) Rosenthal, G. D.: Hospital utilization in theUnited States. American Hospital Association,Chicago, 1964, pp. 55-62.

(22) Somers, H. M., and Somers, A. R.: Doctors,patients, and health insurance. BrookingsInstitution, Washington, D.C., 1961.

(23) Feldstein, P. J., and German, J. J.: Predictinghospital ultilization: an evaluation of threeapproaches. Inquiry 2: 13, June 1965.

(24) Burgun, J. A.: Flexibility-the key to holdingoff obsolescence. Hospitals 38: 35, Oct. 1, 1964.

(25) Llewellyn-Davies, R.: Facilities and equipmentfor health services. Milbank Mem Fund Quart49: 249, July 1966.

(26) Stigler, C. J.: The theory of price. The MacmillanCompany, New York, revised 1952, p. 118.

728 Public Health Reports