Click for next page ( 17


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 16
LIZ Introcluction Medical care has changed dramatically in recent decades; it has become more ef- fective, more costly, and more ambitious. Changes have been impelled by new devel- opments in biology, molecular biology, pharmacology, chemistry, physics, bioen- gineering, materials science, computing, and other scientific and technical fields. Medical advances have meant new medi- cal technologies. New technologies force changes throughout the system; old procedures are discarded, new ones replace them (perhaps too soon or too late), the definition of what is accepted medical practice shifts, third- party payers pay for medical interventions that were earlier unknown and they stop paying for some superseded ones, textbooks are revised, medical school curricula change, and old equipment is replaced with new. This chapter was prepared by Lincoln E. Moses and Frederick Mosteller based on a document drafted by David Banta and Donald Young. David Banta con- tributed the material for the examples on electronic fetal monitoring, the computed tomography scanner, drug treatment for hypertension, and hysterectomy. 16 The expanding gap between the health care that can be provided by the available financial resources and the health care that could be provided if there were no finan- cial constraint makes it increasingly neces- sary that the health care technologies to be available and the conditions of availability be chosen knowledgeably. One might hope that such selection pro- cesses would be guided by an orderly, well- conceived, unified system of testing and as- sessing the new, comparing it with the old, and moving forward as warranted by valid, reliable, evaluative information. That hope is at present only partially ful- filled. Prompt and valid assessment of medical technology is important both to individuals and institutions. The use or nonuse of a new drug, device, or procedure directly concerns two individuals, the patient and the physician. The hospital, manufactur- ing firms, and insurance companies must all make and unmake various arrange- ments when new technology replaces old. Elucidation of medical technology assess- ment thus demands analysis that contem- plates both the individual and the social in- terests.

OCR for page 16
INTRODUCTION Two examples of the impact of technol- ogy assessment illustrate the contribution to be made to disease prevention and to en- hancing the quality of patient care. POLIO Paralytic poliomyelitis used to strike many children, usually during the sum- mer; swimming pool and other recreation facilities often would close because of a po- lio epidemic. The Salk vaccine was in- vented to prevent polio, though no one knew how effective it could be. The idea of a large public test of a vaccine on children was remarkable in itself, though having a President, Franklin D. Roosevelt, who had been afflicted with the disease lent sub- stantial support to the trial. To be convincing, such a trial had to be large because the annual rate of this disease was about 50 per 100,000 population but this number varied widely from year to year and from place to place. An unusual feature was that it struck well-to-do neigh- borhoods more often than lower-income groups. Less-hygienic living conditions are believed to lead to earlier childhood expo- sure to the virus while the immunity con- ferred by the mother still protected the child. It is also true that less educated and less well-to-do groups volunteered less of- ten to participate in such investigations. Originally the study design proposed was to vaccinate children in grade 2 and com- pare them with children not vaccinated in grades 1 and 3. Some state health depart- ment officials objected that such a loose de- sign might leave uncertainty no matter how the study results came out. They would not let their states participate with- out a randomized controlled study. In the end, both studies were carried out, and the results show the wisdom of having the tighter control. The placebo controlled experiment had 201,000 children in the placebo and in the vaccinated groups and 339,000 who were 1 ~ not inoculated. The resulting rates of para- lytic polio per 100,000 population (using laboratory determination) among the vac- cinated groups was 16 and among the pla- cebo group was 57. Those not inoculated had a rate of 36, which was considerably lower than the rate of 46 for the controls in grades 1 and 3 of the other study. The lat- ter were not selected by their refusal to participate and thus included all groups. The reduction from 57 to 16 was substan- tial and led to widespread use of this and other vaccines in the United States and elsewhere and to a very diminished rate of paralytic polio (Meter, 1978~. SURGICAL PROCEDURE A landmark assessment of a surgical technology was carried out on relatively few patients by Cobb et al. (19S9) and by Dimond et al. (1958) in separate but nearly simultaneous experiments. Barsamian (1977) points out that as a treatment for angina pectoris the surgical procedure of internal mammary artery ligation was rap- idly introduced by surgeons in Italy and the United States, theorizing that the oper- ation would reduce pain by shunting blood into the coronary circulation. Early re- ports from operations indicated consider- able success. Barsamian said that the oper- ation was introduced rapidly because it was safe (could be done under local anes- thesia), simple for the surgeon to learn and carry out, and much needed for a large population of patients with coronary ar- tery disease. Measuring effectiveness, though important, was not a major activ- ity for surgeons in the 1950s. An enthusias- tic report in the Reader's Digest (Ratcliff, 1957) sent many patients to surgeons ask- ing for the operation. Cobb applied the operation to 17 pa- tients; for 8 the real operation was carried out, and these patients reported a 34 per- cent subjective improvement in the first 6 months following surgery; 9 patients had a -

OCR for page 16
18 sham operation (which included every- thing in the real procedure except tying off the arteries), and these patients reported 42 percent subjective improvement. Di- mond gave the real operation to 13 pa- tients, 10 of whom showed substantial im- provement, while 5 patients getting the sham operation all reported significant im- provement. Barsamian (1977) said that this test of this operation forced recognition of the possibility that surgery or medicine could have a placebo effect. From this followed the demand and acceptance of controlled studies of surgery. Much of the nation's finest scientific and technological talent conducts research in- tended ultimately to improve the health of Americans. The importance of basic re- search for these purposes is well under- stood, and both the National Institutes of Health and industry invest heavily in it. The work of development also is well rec- ognized. Less appreciated is the bridge in the road from basic research to beneficial use for human beings, namely, technology assessment. One might expect that, with all the research and development that is ac- complished, any product, device, or sys- tem produced would automatically be beneficial and that the invisible hand of marketplace economics would make a new technology cost-effective. This expectation brings frequent disappointments. For ex- ample, in the area of surgery and anesthe- sia, Gilbert et al. (1977) found that less than half of the surgical innovations brought to careful testing in a randomized controlled trial were regarded by their as- sessors as successful. In addition, Gilbert et al. (1977) and Grace et al. (1966) show that weakly controlled studies tend to fa- vor innovations more than do well- controlled studies. Gittelsohn and Wenn- berg (1977) find that small similar areas of a state have great variation in the fre- quency of performance of such surgical procedures as tonsillectomy, appendec- ASSESSING MEDICAL TECHNOLOGY tomy, and hysterectomy. Such variation raises questions about the appropriate level of use of these operations. Thus technology assessment is necessary to verify that inno- vations do or do not work in practice. Cost comparisons and social consequences also require technology assessment. It is not obvious that the step of technol- ogy assessment is required, nor are such as- sessments easy to make. Indeed, in some areas it is not known how to do them. Be- cause the necessity for this step is not widely appreciated, the nation has not thoroughly developed a system for doing it, though many groups contribute to a partial effort, as will be explained in Chap- ter 2 and in greater detail in Appendix A. Such efforts are needed, but also costly. It is the committee's belief that additional funds are required for technology assess- ment and that these incremental funds should come from the health dollar. In the matter of drugs and devices that the Food and Drug Administration regulates, indus- try pays for testing for safety and efficacy, and the public ultimately pays when prod- ucts are marketed. In other matters, such as cost-effectiveness and downstream con- sequences, the health system as a whole is involved with no natural agency or organi- zation to give support to these efforts. For example, market forces like those that sup- port assessment of drugs do not exist for surgical procedures. The nation requires a systematic ap- proach to technology assessment. A strat- egy and an organization for setting priori- ties is needed as well. Given the priorities, mechanisms are needed for actually mak- ing the assessments and implementing their findings. Finally a method is needed for paying for many of the needed assess- ments. As with any large-scale technologi- cal enterprise, it is necessary to maintain a strong body of professional personnel to carry out the assessments, and they must be encouraged to conduct work of high qual- ity and develop new techniques as re-

OCR for page 16
INTRODUCTION quired. Although in some areas parts of this overall process are in place and con- tribute well to the nation's health, the sys- tem as a whole has major gaps and defi- ciencies, which will be described. Simply knowing the outcomes for the health care system and their relation to the treatments and diagnoses employed might help in designing a more economical and effective system. Just now a strong link that connects outcome to care is not known. Partly this comes from not know- ing how to set up such a system. A vast, sprawling monitoring system cannot be what is needed. Some keen minds should think about better indicators that relate care and outcome. Part of the trouble is that much of the current health care sys- tem contributes to quality of life rather than to morbidity and mortality. By and large, accomplishments are not assessed in these softer areas, and so medicine does not get nearly the credit due for these contribu- tions to comfort and convenience. Part of the difficulty arises from the fact that the health delivery system is itself a dynamic process applied to a changing population. We would not have it otherwise, but it helps to explain how hard outcome studies for whole populations must be. Having mentioned this larger problem, we turn back now to the narrower ideas of technol- ogy assessment. The questions of who should carry out assessments, how they should be done, and who should pay for them are complicated and political and have no simple answers. Consequently the study described here was conducted. This report addresses the present state of the assessment of medical technology, gives attention to processes, problems, interested parties, successes, and failures, and finally points to some needs and opportunities for improving the present system of medical technology as- sessment. 19 EXAMPLES OF TECHNOLOGY ASSESSMENT For examples, this section begins with brief sketches of how five medical technol- ogies have recently made their entrance into medical practice. A number of com- mon themes recur, and these will help to shape a systematic treatment of the sub- ject. Because medicine moves rapidly, fur- ther work will have been done on the prob- lems treated in these examples by the time this work is published. The purpose here is not to publish the latest information but to give a feel for the varieties of technology assessment that arise. Electronic Fetal Monitoring Electronic fetal monitoring (EFM) is a technologic step beyond the stethoscope for monitoring the heart rate of the fetus dur- ing labor and delivery. EFM enables eval- uation of fetal heart rate patterns in rela- tion to uterine contractions and facilitates detection of certain types of abnormal pat- terns. Concerns about preventable perina- tal mortality and brain damage led investi- gators to seek a more reliable and valid method of following fetal status during la- bor (Banta and Thacker, 1979~. Advances in electronics during World War II made electronic fetal monitors feasible, as first demonstrated by a team at Yale Univer- sity. Such monitors became available on the market in about 1968, and their use spread rather quickly into most of the ob- stetric units in the United States. By 1980 about half of the deliveries in this country were electronically monitored (Placek et al., 1983a,b). But in the mid-1970s electronic monitor- ing already had become controversial be- cause it was suspected of being associated with inappropriate cesarean sections. This question stimulated several randomized clinical trials (RCTs) (Haverkamp et al., 1976; Renou et al., 1976; Kelso et al.,

OCR for page 16
20 1978; Haverkamp et al., 1979; Wood et al., 1981~. Although the trials consistently found no reduction in fetal or infant mor- tality from use of the electronic monitor as compared with auscultation, they did sug- gest an association with increases in cesar- ean delivery rates. The basic problem with these RCTs is that they were all too small. Any beneficial effects of EFM would be so slight as to require very large studies to cie- tect them. Another shortcoming of the RCTs is that none included low-birth- weight fetuses (including prematures), who are at greatest risk of mortality and morbidity. More recently, a trial in about 13,000 low-risk women in Dublin, Ire- landa sufficient number to yield valid results has found no benefit in terms of reduced mortality from EFM but has sug- gested a decrease in neurolo~ic damage and a dramatic decrease in numbers of in- fant convulsions with monitoring (Inge- marsson, 1981; McDonald et al., 1983~. Also the Dublin trial found no difference in cesarean rates for the monitored groups. The risks associated with electronic monitoring are of concern. These range from infection and hemorrhage to a possi- ble correlation with the incidence of cesar- ean section, which would be the greatest risk. The rate of cesarean sections in the United States was 4.5 percent of deliveries in 1965, but rose steadily to about 18 per- cent in 1980 (Placek et al., 1983a,b). EFM also is expensive. Banta and Thacker (1979) estimated an annual cost of such monitoring at $411 million, including indirect costs such as those of cesarean sec- tions and other complications. Cohen (1983), using a more thorough method of estimation, projected annual costs of from $210 million to $385 million. (These stud- ies however do not include the possible benefits of preventing necrologic damage. ~ Electronic fetal monitoring is a good exam- ple of inadequate evaluation. More than 15 years after its introduction, more work is needed to define its appropriate use. One ASSESSING MEDICAL TECHNOLOGY RCT is now being done in low-birth- weight infants and may answer some ques- tions. In the meantime, a risky and costly procedure continues in widespread use. Computed Tomography Scanning The computed tomography (CT) scan- ner is a revolutionary diagnostic device that combines x-ray equipment with a computer and a cathode ray tube display to produce images of cross sections of the human body (Office of Technology Assess- ment, 1978~. The CT scanner was the result of decades of research in such fields as mathematics, computer applications, and x-ray tomography. During the 1960s several people in the United States realized that it would be possible to develop a medi- cal diagnostic device based on this re- search, but they were unable to interest ei- ther industry or government. Late in that decade, Hounsfield, working at EMI Ltd. in England, was able to convince his com- pany to develop a prototype device (Hounsfield, 1980~. The British Depart- ment of Health and Social Services also contributed some funds to the project. The first demonstrations were held at interna- tional radiology meetings in 1972, and the device was rapidly accepted by the medi- cal community. The importance of the technological advance was recognized by the award of the Nobel Prize in medicine to some of the developers. Plaudits notwithstanding, the CT scan- ner has come to symbolize the problem of high-cost medical technology (Banta, 1980~. In part, this was because of the ra- pidity of its spread in industrialized coun- tries. In part, it was because of its expense: the typical scanner cost $300,000 or more in 1973 and by 1984 it cost almost $1,000,000. The United States now has more than 2,000 scanners, representing a capital investment of more than $1 billion. Operation of the scanners costs the United States at least another $1 billion annually.

OCR for page 16
INTRODUCTION CT scanning is such a radical departure from conventional radiographs that it will take years of research for its proper roles to be assessed in different parts of the body and in different disorders. Although many evaluations of CT scan- r~ing have been done, few studies have ad- dresse`1 a fundamental question: For what kinds of patients is application of this diag- nostic technology worth its costs? (Wag- ner, 1980~. When dealing with diagnostic technologies, answering this question re- quires answering another question: How does a particular technology fit into an op- timal diagnostic process for a given condi- tion? In other words, which diagnostic technologies should be used in a particular patient who needs to have a diagnosis es- tablishe~l? The importance of this question may be illustrated by data indicating that millions of CT head scans were done each year for people with uncomplicated head- aches until CT was better evaluated. The United States now faces the emer- gence of nuclear magnetic resonance imag- ing, which is another example in the se- quence of technology development. With- out better evaluative studies for this and other new technologies, a great deal of money might be wasted on inappropriate diagnostic tests, and important opportuni- ties might be missed to do diagnostic tests on people who truly need them. Drug Treatment for Hypertension Hypertension, or high blood pressure, is the most common chronic disease in the United States. Estimates are that about 60 million people in this country have definite or borderline hypertension (Levy, 1982~. People with high blood pressure are more likely to have strokes, heart disease, and kidney failure than people with normal blood pressure. Hypertension can be controlled by drug treatment. In the late 1960s, the Veterans Administration supported a multi-institu- 21 tional RCT of treatment for men with the drugs hydrochlorothiazide, reserpine, and hydralazine. The control group was given placebos. The drug treatment was remark- ably effective for men with diastolic pres- sures higher than 105 mm of mercury. For example, strokes were reduced by 75 per- cent, and congestive heart failure, renal failure, and dissecting aneurysm occurred only in the control group (Veterans Ad- ministration, 1967, 1970~. The growing use of drug treatment for high blood pres- sure has been considered to be one of the factors that has led to a falling rate of death from heart disease in this country (Havlik and Feinleib, 1979~. However, although a growing percent- age of people with high blood pressure are being treated, many still are not. Recent surveys done in Connecticut, South Caro- lina, Maryland, and California have shown that 18 to 28 percent of those with definite hypertension were unaware that they had the disease, and that another 21 to 34 percent were aware but were not re- ceiving adequate therapy (National Center for Health Statistics [NCHS], 1983~. Thus, very positive results of an evaluation of drug therapy have not yet been fully im- plemented nationally and people are still dying of cardiovascular disease at an un- necessarily high rate. Mild hypertension is a different prob- lem. Clinical trials have not given clear- cut evidence as to whether people with diastolic pressures under 95 mm of mer- cury should be treated (Hypertension Detection and Follow-up Program Coop- erative Group, 1982~. Yet, surveys of phy- sicians have shown that they frequently prescribe drugs for mild hypertension (Guttmacher et al., 1981~. More recent clinical trials have not resolved this scien- tific issue for patients under 50 years of age, which is of concern because antihy- pertensive drug treatment is not benign (Joint National Committee on Detection, Evaluation, and Treatment of High Blood

OCR for page 16
22 Pressure, 1984~. Complications associated tion with drug treatment include dizziness, im- potence, and general tiredness. Side effects can be minimized by careful medical su- pervision, but it is doubtful whether such care is usually available. The practice of treatment of mild hypertension has been criticized as seeking the technological rather than the social solution to disease, because the incidence of high blood pres- sure is often associated with stressful life situations (National Institutes of Health 1979j. The drug industry's promotion of drug treatment for mild hypertension also has been criticized. In short, drug treat- ment for hypertension is one of the most important medical advances of this cen- tury. However, questions remain that can be answered only by good assessments. Hysterectomy Surgical removal of the uterus is per- formed more often than any other major operation in the United States. The Na- tional Center for Health Statistics esti- mates that 704,800 hysterectomies were performed in the United States in 1978, compared with 678,000 in 1976. The 1978 rate is 817.3 per 100,000 women 15 years of age and older (Korenbrot et al., 1980~. More recent figures for 1980 and 1981 indi- cate rates of 563 and 573 per 100,000 women, respectively (Easterday et al., 1983~. At such a rate, more than half of American women would have their uterus removed by age 65. The high rate of hys- terectomy is not peculiar to the United States (L. J. Kozak, National Center for Health Statistics, personal communica- tion). Canada and Australia have rates ap- proximately as high as those in the United States. It is frequently alleged that many of these hysterectomies are unnecessary. Medical indications for hysterectomy are not standard, and this leads to varia- tions in the rate. Wennberg and Gittelsohn (1973) have demonstrated a strong correla- ASSESSING MEDICAL TECHNOLOGY between the numbers of surgical spe- cialists and the number of operations per- formed in different districts in Vermont. Cross-national studies comparing the United States, the United Kingdom, and Canada have demonstrated a relationship between the number of surgeons and oper- ations, including hysterectomy (Bunker, 1970; Vayda, 1973~. Some indications for hysterectomy, such as for cancer, are well accepted. The con- troversial indications are its use as a means of sterilization and its use to prevent cancer of the uterus. About 30 percent of hysterec- tomies done in the United States are done for these indications (Korenbrot et al., 1980). Hysterectomy has significant risks. Death occurs in 0.1 to 0.4 percent of cases. If 30 percent of hysterectomies are elective in the United States, a 0.1 percent mortal- ity would mean 210 deaths among this group. Much more frequent are nonfatal operative complications, including bleed- ing, infection, and complications of trans- fusion and anesthesia. In a meta-analysis of published reports carried out at Stan- ford University, 73 percent of women with nonemergency abdominal hysterectomy had some degree of morbidity, and more than 7 percent had moderate to life-threat- ening complications (Korenbrot et al., 1980~. The financial costs of hysterectomy are high. A hysterectomy was estimated in 1978 to cost from $1,700 to $2,600 in direct medical care expenses (Korenbrot et al., 1980~. This figure does not include the costs of complications, nor does it include indirect costs such as lost work or psycho- logical costs. Several cost-effectiveness studies have been done, and none found hysterectomies cost-effective for steriliza- tion or prevention of uterine cancer. The studies have found large immediate risks and costs, with some future benefits. In a recent study, Sanberg et al. (1984) found net costs to range per year from $1,200 to

OCR for page 16
INTRODUCTION $3,700 and net increases in life expectancy were on the order of 6 to 8 months. Thus, hysterectomy is frequently used for questionable indications. Many data that would be useful in assessing this kind of use are lacking, and available data do not strongly defend it. This raises questions of who will do the needed research and what will be done about the frequency of the procedure. Assuming that women are choosing the procedure voluntarily, it may be appropriate for society to decide that some benefits, such as the small, yet un- proved likelihood of cancer prevention, are not worth the large immediate costs. Medical Information System E1 C amino Hospital, a medium-sized, nonteaching, community hospital, in- stalled in 1971 a Technicon Medical Infor- mation System (TMIS), which processed a broad range of medical and administrative data. Two studies of it by Battelle Colum- bus Laboratories were funded by the Na- tional Center for Health Services Research (Coffey, 1980~. The first examined the ef- fect of TMIS on organization and adminis- tration. The second study, described here, measured the impact of TMIS on total hos- pital costs. This example illustrates a par- tial evaluation of a support system rather than of a treatment or of a diagnostic tech- nology. A more complete evaluation might also study changes in the quality of care. 23 (Although the study presented here dis- cusses costs, these may actually be some form of charges.) Costs were examined in two ways, one excluding the cost of TMIS and the other in- cluding its operational cost. The investiga- tors used three indices: cost per patient, cost per patient day, and cost per month. Cost per patient measures the social cost, expense per patient day keys into health insurance carriers' procedures, and monthly expenses give an overall picture of the hospital budget. Expenses were broken down ac- cording to nursing care, ancillary services, and support services. The study covered a 6- year period: 2 years before TMIS, 1 year of installation, and 3 years of full operation. Multiple regression methods and four control hospitals were used to adjust for variables that could not be controlled. Table 1-1 shows the overall outcome of changes in costs associated with the opera- tion of the TMIS in its third year of opera- tion. These changes exclude the cost of the TMIS itself. The reduced cost per patient for nursing was said to be due to reduced paperwork and to a reduction in the nurs- ing work force. Another important impact was that faster turnaround time on tests and execu- tion of orders led to a reduction of 4.7 per- cent in length of hospital stay. This was re- markable because E1 C amino already had a very low average length of stay. The increase of 4.5 percent in support costs per patient was, nevertheless, unex- TABLE 1-1 Partial Impact of TMIS (Excluding Its Cost) on Three Measures of Cost at E1 Camino Hospital, by Department Change in Cost Change in Cost Change in Cost Department per Patient per Patient Day per Month Nursing _ 5.0a - 2.0 5.3 Ancillary - 2.4 1.1 7.7a Support 4.5 7.5a 14.3a All departments - O.6 2.3 9.2 a Statistically significant beyond the 5 percent level.

OCR for page 16
24 pected. Some increase was due to increased medical records work and some to a deci- sion unrelated to TMIS to increase nurses' training. The investigators believed that changes in support costs should not be at- tributed to the TMIS, and so they analyzed the data in two ways. Table 1-2 shows the results analyzed with an adjustment for these support cost charges (Method 1) and a second analysis that shows exactly what happened with adjustment for support costs (Method 2~. The TMIS costs are treated separately, and the two analyses are presented in Table 1-2. The 4.5 percent difference between the monthly costs at every level of both meth- ods of calculation suggests that the TMIS costs 4.5 percent of the total budget. The investigators speculate that this means that the hospital may have to absorb 40 percent of the cost of the system,with 60 percent covered by improvements in productivity. Earlier studies had shown reduced error rates on orders and tests and improved completeness and accuracy of patient data. Thus, the medical benefits might justify the additional costs, but this issue was not part of the study. The main conclusions were that (1) nurs- ing costs per patient had been reduced by about 5 percent; (2) average length of stay was reduced by about 4.7 percent, even ASSESSING MEDICAL TECHNOLOGY though E1 C amino started with a very low rate; (3) TMIS raised overall costs per pa- tient by 1.7 or 3.9 percent depending on whether one ignores support department increases; (4) hospital costs rose 3.2 percent per patient day; and (5) adjusted overall monthly expenses rose 7.8 percent largely because of increased patient flow. Caution should be observed in transferring this ex- perience because management informa- tion systems have become more fully devel- oped since then. The outcomes might depend heavily on the patterns of work and the organizational structure of the in- stitution installing a management informa- tion system. One would want to assess such a system to see which aspects were working well and which were not; the outcome of the assessment may be unique to the insti- tution. Other studies assessing information systems in hospitals have been carried out by Rogers and Haring (1979) and Haring et al. (1982~. This analysis clearly shows that intro- ducing a new system has extensive and of- ten unexpected ramifications. Tracing the consequences is a difficult task. SOPHIE LESSONS FROM THE EXAMPLES It can be seen that medical technology is a term that embraces quite a range of ac- TABLE 1-2 Overall and Partial Impact of TMIS on Annual Expense of All Departments at E1 Camino Hospital During 1975, by Two Methods Unit of Measurement TMIS Impact Including TMIS Cost (%) TMIS Impact Excluding TMIS Cost (%) Method 1. Adjusted for support costs Per patient 1.7 - 2.8 Per patient day 3.2a -1.3 Per month 7.Sa 3.3a Method 2. Unadjusted Results Per patient 3.9 - 0.6 Per patient day 6.8 2.3 Per month 13.7 9.2 aStatistically significant beyond the 5 percent level.

OCR for page 16
INTRODUCTION tivities. For consistency the committee fol- lows the usage of the Office of Technology Assessment (OTA, 1982), which uses the term to refer to "techniques, drugs, equip- ment, and procedures used by health-care professionals in delivering medical care to individuals, and the systems within which such care is delivered." The five examples illustrate some of the dimensions along which medical technol- ogy varies. Two examples, electronic fetal monitoring (EFM) and computed tomog- raphy (CT), are diagnostic; two, hysterec- tomy (Hx) and drug treatment for hyper- tension (DTH), are therapeutic; one, in- formation processing, is a supporting technology; one is a drug, one is a surgical procedure, three are strongly equipment linked. At least six issues occur in three or more of the five examples: 1. Risks how probable and how severe are associated adverse effects? (EFM, DTH, Hx) 2. Appropriateness are there indica- tions for use in at least some patients? (All five) 3. Benefits what are they? How large? How sure? (All five) 4. Insufficient evidence in the extant studies. (EFM, CT, Hx, DTH) 5. Rapidity and scope of diffusion into clinical use. (The first four suggested the possibility of too rapid and extensive ac- ceptance; in addition, DTH suggested un- derutilization. ~ 6. Cost, both to patient and as a social investment. (All five) Two further issues should be adduced here, although we do not undertake to measure their intensity in the examples: 7. Assessment of costs, risks, and bene- fits can be difficult, and requires informa- tion that is hard to get or lacks conceptual clarity about subtle matters such as quality of life. 25 8. Ethical questions are inherent and include equity of access to new technology, reasonableness of allocation of scarce med- ical resources, and mistreatment of some patients because of incorrectly established indications, etc. These eight issues constitute problems that are addressed by technology assess- ment, a term that we now can treat more specifically. The term assessment of a med- ical technology is used here to denote any process of examining and reporting proper- ties of a medical technology used in health care, such as safety, efficacy; feasibility; indications for use; cost and cost-effective- ness; and social, economic, and ethical consequences, intended and unintended. Comprehensive assessment examines all of these issues. This language is chosen deliberately; it admits of an assessment that is concerned with only a portion of the full spectrum of properties. Some assessments will be only of the safety and efficacy of a technology while others may be more inclusive, add- ing information about costs and social and ethical impacts. This is intended, because much of assessment activity is partial in its scope. As will be seen, various participants in the health care system find different properties to be salient, focusing their stud- ies narrowly to address the questions that interest them. COMPREHENSIVE TECHNOLOGY ASSESSMENT Assessment of medical technology is of course a particular instance of technology assessment as practiced by industry, gov- ernment, consumers, and various agencies in other fields of applied technology such as transportation, agriculture, or housing. Technology assessment generally is an im- perfect but maturing process whose conse- quences are exemplified in the popular press by the occasional recall of automo-

OCR for page 16
26 ASSESSING MEDICAL TECHNOLOGY bites, the relatively recent redesign of tiveness of coronary care units for the bridge abutments and guard rails on high- treatment of patients with myocardial in- ways, and the constraints imposed on nu- farction is complicated by the dramatic re- duction in mortality from myocardial in- farction over the last 15 years which is partly attributable to the change in smok- ing habits of the population at risk. clear energy plants. Arnstein (1977) attributes the concept of technology assessment to Emilio Q. Dad- dario, former congressman and founding Director of the Office of Technology As- sessment. Technology assessment has more holistic implications than such usual meth- ods of technology evaluation as clinical trials, market research, cost-benefit analy- sis, or environmental impact assessment. Technology assessment ideally would be comprehensive and include evaluation not only of the immediate results of the tech- nology but also of its long-term social, eco- nomic, and ethical consequences. A comprehensive assessment of a medi- cal technology after assessment of its im- mediate effects may also include an ap- praisal of its unintended consequences, problems of personnel training and licen- sure, new capital expenditures for equip- ment and buildings, and possible conse- quences for the health insurance industry and the social security system. Technolont assessment provides a form of policy analy- sis that includes as potential components the narrower approaches to technology evaluation. Most assessments stop with a partial effort, and we include these when we speak of technology assessment. The assessment of medical technologies presents certain qualitative and quantita- tive differences from technology assess- ment in other sectors. For example, medi- cal technologies often can be assessed only on the basis of observing acutely ill people under conditions in which there is less than full control of important variables and with less than desirable characterization of individual circumstances. The assessment of medical technologies often is confounded by the occurrence of large changes in the patient or process out- come due to factors outside of the study de- sign. For example, the analysis of the effec- DIFFERENT PARTIES, DIFFERENT AIMS IN ASSESSMENTS The Office of Technology Assessment staff, in treating this subject, wrote (OTA 1982, p. 3~: Medical technology assessment is, in a nar- row sense, the evaluation or testing of a medical technology for safety and efficacy. In a broader sense, it is a process of policy research that ex- amines the short and long-term consequences of individual medical technologies and thereby becomes the source of information needed by policymakers in formulating regulations and legislation, by industry in developing products, by health professionals in treating and serving patients, and by consumers in making personal health decisions. For the purposes of this work, nearly all of health services research would be included in this definition. According to this statement, medical technology assessment involves at least four participants: a policymaker, an ad- ministrator, a health care provider, and the patient. The provider and the patient can be seen as a dyed, concerned primarily with the safety and efficacy of particular technologies under consideration for use by that practitioner upon that patient. A1- though cost, equity, profit, etc., matter a great deal to the dyed, they are not pri- mary factors in their technology assess- ment. Patients should be concerned that tech- nologies have been studied in patients simi- lar to themselves. That aspect of equity may be up to the dyed, not in each in- stance, but in the sense of a general popu- lation. Those who choose not to participate

OCR for page 16
INTRODUCTION in studies need not expect the findings to apply to them. The medical scientists who test and try out new technologies may concentrate ef- forts mainly on one feature of the technol- ogy. The epidemiologist may study whether a particular adverse outcome is systematically related to the use of a cer- tain treatment (e. g., thromboembolism from using high-estrogen contraceptive pills, or vaginal carcinoma in daughters of mothers who used diethylstilbestrol). The diagnostic imaging specialist may want to assess the relative or absolute information obtained with a test (e.g., echocardiog- raphy, radionuclide studies, ultrasound, etc.) because it relates to the efficacy of those technologies. Another medical scien- tist, evaluating diagnostic or therapeutic algorithms, may focus on a narrow ques- tion, such as "Can the algorithm be im- proved by introducing test Y at some stage?" The makers of devices, drugs, or other medical equipment may have yet other in- 27 of safety and efficacy. Professional societies often establish standing or ad hoc panels to study questions relating to the appropriate use of technologies. It sometimes happens that different organizations reach differing conclusions on the same issue. Third-party payers contribute substan- tially to medical technology assessment. Requests for payment for a new procedure are likely to trigger a review of the technol- ogy. Usually, this review is not deliberately directed at safety and efficacy, but rather it is to help in deciding whether the new procedure conforms to accepted practice. A negative decision about reimbursing a new technology will tend to inhibit its dif- fusion. But sometimes such a decision can stimulate further assessment studies, un- dertaken by proponents of the technology. Thus, it is reported (Cutler et al., 1973) that some of the studies concerning cost- effectiveness of routine health checkups originated in efforts to justify the provision of reimbursement for such checkups. ~ ~ Large employers often are major pur- terests in assessment of their technologies. ~ ~ ~ The manufacturer of a drug, for instance, will seek to develop information to satisfy the requirements of the Food and Drug Administration; both the data and the manner in which they were acquired are subject to regulatory review. Manufactur- ers more generally will be concerned with the size of the possible market for their product, its costs as seen by the buyer, and its strengths and weaknesses in comparison with competitive products. Various institutional components of the health profession also are involved in tech- nology assessment, and their concerns re- late to their roles. Editors of journals influ- ence what becomes known to readers who rely upon their journals; the influence is exerted both through editorials and through decisions about which articles to publish. Medical school teachers and text- book authors must form, and then propa- gate, opinions that amount to assessments chasers of medical coverage for their em- ployees. Extension of coverage benefits, perhaps as an item of union contract nego- tiation, raises the question of which exten- sions would be most worthwhile. This may quickly lead to comparison of the costs, risks, likely frequency of use, and effective- ness of several medical technologies that are alternative candidates for new cover- age. Hospitals and health maintenance orga- nizations (HMOs) often look at new medi- cal technologies as possible major invest- ments, including both capital costs and costs of operation. The prospective pur- chaser must assess this technology from many points of view besides its costs. Its probable revenues are of equal concern. Questions of feasibility may be dominant: What are the space requirements? What training is required for personnel to oper- ate it? Will re-education of medical profes- sionals be called for? How about computer

OCR for page 16
28 support? Costs from the point of view of the patient may receive little attention if reimbursement is assured, or much atten- tion if, as in the HMO, recovery must ulti- mately be sought through overall increased membership fees, reserves, or other sources. Some associations of hospitals and other institutions offer to their members assess- ment-like advice about new technologies. Several of these are described in Chapter 2. For example, the American Hospital Asso- ciation (AMA) works to assist hospital ad- ministrators facing management and in- vestment decisions about new and existing technologies. The AHA program evaluates diagnostic systems, therapeutic systems, computer technologies, and the like, but evaluates medical procedures only as they relate to equipment purchase or nonmedi- cal hospital personnel. The evaluations fo- cus on: cost and organizational implications; installation costs; staffing and training requirements; probable number of patients affected; effects on other hospital resources, such as the extent to which a technology will enable the replacement of existing re- sources, or the extent to which it will neces- sitate the addition of new resources; clinical effectiveness: not patient out- comes as such, but process outcomes such as inpatient versus outpatient application, average length of stay, etc. Government agencies are active in med- ical technology assessment, often through decisions about reimbursement or about hospital investment in large equipment. Such assessments are concerned with eco- nomic efficiency, as well as with safety and efficacy, but U.S. government agencies have generally not emphasized assessments for economic efficiency and cost-effective- ness. The guidelines prepared by the fed- eral government of Canada to assist prov- inces who request guidance in their ASSESSING MEDICAL TECHNOLOGY hospital investments typically have 10 components (OTA, 1980~. These compo- nents form a checklist of considerations to be entertained before establishing a new unit: patient load; bed requirements; recommended distribution; administrative policy, procedures and control; . . services; staff establishment and coverage; staff training and qualifications; specific supporting departments and space allocation, utilization, and spe- cific design features; equipment; relationship with other departments and services. Concern with feasibility and coordina- tion loom large in these governmental guidelines for assessment. The contrast between the Canadian guidelines and the U.S. health policy state- ments in the 1975 health planning law and its 1979 amendments is worth noting. Title XV of the Public Health Services Act (P.L. 93-641) Section 1502 sets forth a number of National Health Priorities intended to guide national planning and investment, especially governmental, in capital facili- ties for health care, including expensive equipment. Among these priorities are provision of primary care services for medically underserved populations, with emphasis on rural and economically de- pressed areas; coordination and consolidation of in- stitutional health services by developing multi-institutional systems (specialty ser- vices such as radiation therapy, intensive and coronary care, and emergency trauma care are singled out for special attention); development of multi-institutional systems for sharing support services; development of health services institu-

OCR for page 16
INTRODUCTION tions "of the capacity to provide various levels of care. . . on a geographically inte- grated basis." The public policy goals also emphasize alternative health care systems to hospi- tals, encouraging the development of medical group practice; the training and utilization of allied health professionals such as nurse clini- cians and physician assistants; the promotion of activities for the pre- vention of disease, including studies of nu- tritional and environmental factors affect- ing health and the provision of preventive health services. Some additional priorities are consumer education so that the general public might use "proper personal health care" includ- ing prevention; cost containment and the "adoption of uniform cost accounting and simplified reimbursement"; and activities to "achieve needed improvement in the quality of health services." Federal Medicare reimbursement for the capital portions of the hospital bill are de- nied to health care institutions that expand beds or certain programs without a certifi- cate of need. The 1975 law approached cost containment primarily through insti- tutional coordination, regionalization, the sharing of services, and nonhospital alter- natives. Technology assessment as a means of achieving these goals is not given specific mention. However, the 1979 amendments emphasized the importance of "the identi- fication and discontinuance of duplicative or unneeded services and facilities" tP.L. 96-79, Section 102(a) (1) i. To the section of the previous law promoting uniform cost accounting and improved management procedures for institutions offering health services was added the words "and the de- velopment and use of cost-saving technol- ogy" [Section 102(a)~2~. 29 Finally, legislative bodies can be deeply involved in assessment of medical technol- ogy. Congress established the Office of Technology Assessment in 1972 as an advi- sory arm. OTA uses these words in describ- ing its mission: "The assessment of technol- ogy calls for exploration of the physical, biological, economic, social, and political impacts which can result from applications of scientific knowledge." For medical tech- nology this broad construction of the task reaches far beyond safety and efficacy. The multiplicity of organizations carry- ing out assessments, the variety of kinds and purposes of assessments, and the amount spent on various kinds of assess- ments are described in Chapter 2. That in- ventory calls attention to the fact that no agency has the task of attending to the needed research for the nation, such as noting which medical technology assess- ments need to be carried out and assigning priorities and financing their execution. Appendix A supplements Chapter 2 by de- scribing in more systematically gathered detail the work of a set of the agencies car- rying out assessments of medical technolo- gies. To lay a foundation of methods used in medical technology assessment, Chapter 3 describes and illustrates the major meth- ods, explains their strengths and limita- tions, and outlines new research in each method whose results might strengthen its use. To help us understand how medical technologies come to be adopted, dropped, or ignored, Chapter 4 examines the role of assessment, education, publications, and other stimuli to diffusion. The role of reimbursement in encourag- ing and paying for the assessment of medi- cal technologies has been much debated, and Chapter 5 discusses some of the history of these developments, both state and na- tional. The international scene, at least in principle, is a two-way street for technol- ogy assessment, and Chapter 6 explores the current position of the United States with

OCR for page 16
30 respect to use of international assessments. The question of what information can be usefully exchanged and what long-term policies should be instituted go beyond the purview of this report. Chapter 7 summarizes the state of the nation's assessment of medical technologies and makes general recommendations for creating a pluralistic national system. It does not summarize the narrower recom- mendations scattered through the report, highlighted in the chapters, but focuses on the gaps that now prevent the United States from having a system and proposes a set of gradual steps for creating one. REFERENCES Arnstein, S. R. 1977. Technology assessment: Op- portunities and obstacles. IEEE Trans. Systems, Man, Cybernetics, Vol. SMC-7~8~:571-582. Banta, H. D. 1980. The diffusion of the computed tomography (CT) scanner in the United States. Int. J. Health Services 10:251. Banta, H. D., and S. B. Thacker. 1979. Assessing the costs and benefits of electronic fetal monitoring. Obstet. Gynecol. 34:627. Barsamian, E. M. 1977. The rise and fall of internal mammary artery ligation in the treatment of angina pectoris and the lessons learned. Chapter 13, pp. 212- 220 in J. P. Bunker, B. Barnes, and F. Mosteller, eds., Costs, Risks and Benefits of Surgery. New York: Ox- ford University Press. Bunker, J. P. 1970. Surgical manpower: A compar- ison of operations and surgeons in the United States and in England and Wales. N. Engl. J. Med. 282:135. Cobb, L. A., G. I. Thomas, D. H. Dillard, et al. 1959. An evaluation of internal mammary-artery li- gation by a double-blind technic. N. Engl. J. Med. 260:1115. Coffey, R. M. 1980. How a medical information system affects hospital costs: The E1 Camino Hospital experience. NCHSR Research Summary Series, DREW Pub. No. (PHS) 80-3265. Washington, D.C.: Department of Health, Education, and Welfare. Cohen A. 1983. Decision and Policy Analysis for Electronic Fetal Monitoring. Ph.D. dissertation. Har- vard School of Public Health, Cambridge, Mass. Cutler, J. L., S. Ramcharon, R. Feldman, et al. 1973. Multiphasic check-up evaluation study. Prev. Med. 2:197. Dimond, E. G., C. F. Kittle, and J. E. Crockett. ASSESSING MEDICAL TECHNOLOGY 1958. Evaluation of internal mammary artery liga- tion and sham procedure in angina pectoris. Circula- tion 18:712. Easterday, C. L., D. A. Grimes, and J. A. Riggs. 1983. Hysterectomy in the United States. Ob. Gyn. Obstet. Gynecol. 62:203-212. Francis, K., Jr., et al. 1955. An evaluation of the 1954 poliomyelitis vaccine trial-summary report. Am. J. Public Health 45~5~:1-63. Gilbert, J. P., B. McPeek, and F. Mosteller. 1977. Program in surgery and anesthesia. Pp. 124-169 in J. P. Bunker, B. A. Barnes, and F. Mosteller, eds., Costs, Risks and Benefits of Surgery. New York: Ox- ford University Press. Gittelsohn, A. M., and J. A. Wennberg. 1977. On the incidence of tonsillectomy and other common sur- gical procedures. Pp. 91-106 in J. P. Bunker, B. A. Barnes, and F. Mosteller, eds., Costs, Risks and Bene- fits of Surgery. New York: Oxford University Press. Grace, N. D., H. Muench, and T. C. Chalmers. 1966. The present status of shunts for portal hyperten- sion in cirrhosis. Gastroenterology 50:684. Guttmacher, S., M. Teitelman, G. Chapin, G. Garbowski, and P. Schnall. 1981. Ethics and preven- tive medicine: The case of borderline hypertension. Hastings Center Report 11:12. H~rin~ ~ M. P. M. Wortman R. A. Watson. 7 ~ and N. P. Goetz. 1982. Automating the medical record: An assessment of impact on process and out- come of care in hypertension, obesity, and renal dis- ease. Med. Care 20:63-74. Haverkamp, A. D., M. Orleans, and S. Lange- doerfer, et al. 1979. A controlled trial of the differen- tial effects of intrapartum fetal monitoring. Am. J. Obstet. Gynecol. 134:399. Haverkamp, A. D., H. E. Thompson, J. G. McFee, et al. 1976. The evaluation of continuous fetal heart rate monitoring in high-risk pregnancy. Am. J. Ob- stet. Gynecol. 125:310. Havlik, R. J., and M. Feinleib, eds. 1979. Proceed- ings of the Conference on the Decline in Coronary Heart Disease Mortality. Pub. No. 79-1610. Bethesda, Maryland: National Institutes of Health. Hounsfield, G. N. 1980. Computed medical imag- ing. Science 210:22. Hypertension Detection and Follow-up Program Cooperative Group. 1982. The effect of treatment on mortality in "mild" hypertension: Results of the Hy- pertension Detection and Follow-up Program. N. Engl. J. Med. 307:976. Ingemarsson, E., I. Ingemarsson, and N. W. Sven- ningsen. 1981. Impact of routine fetal monitoring during labor on fetal outcome with long-term follow- up. Am. J. Obstet. Gynecol. 141:29. Joint National Committee on Detection, Evalua- tion, and Treatment of High Blood Pressure. 1984. The 1984 report of the Joint National Committee on

OCR for page 16
INTRODUCTION Detection, Evaluation, and Treatment of High Blood Pressure. Arch. Intern. 144:1045-1057 Kelso, I. M., R. J. Parsons, G. F. Lawrence, et al. 1978. An assessment of continuous fetal heart rate monitoring in labor. Am. J. Obstet. Gynecol. 131:526. Korenbrot, C., A. B. Flood, M. Higgens, et al. 1980. Elective Hysterectomy, Costs, Risks, and Bene- fits. Prepared for the Office of Technology Assess- ment, Congress of the United States, Washington, D.C. Levy, R. I. February 1982. The National Heart, Lung, and Blood Institute Overview 1980: The direc- tor's report to the NHLBI Advisory Council. Circula- tion 65:217. McDonald, D., A. Grant, M. Pereira, P. Boylan, and I. Chalmers. 1983. The Dublin Randomized Controlled Trial of Intrapartum Electronic Fetal Heart Rate Monitoring. Presented at the 23rd British Congress of Obstetrics and Gynecology, Birmingham, July 14. Meter, P. 1978. The biggest public health experi- ment ever: the 1954 field trial of the Salk poliomyelitis vaccine. Pp. 3-15 in M. Tanur et al., eds., Statistics: A Guide to the Unknown, 2nd Ed. San Francisco: Holden Day. National Center for Health Statistics, Department of Health and Human Services. 1983. Prevention Pro- file: Health United States. Washington, D.C.: U.S. Government Printing Office. National Institutes of Health. 1979. Diagnosis and Management of Hypertension, A Nationwide Survey of Physicians' Knowledge, Attitudes, and Reported Behavior. DHEW Pub. No. (NIH) 79-1056. Bethesda, Maryland. Office of Technology Assessment. 1978. Policy Im- plications of the Computed Tomography (CT) Scan- ner. Pub. No. OTA-H-56. Washington, D.C.: U.S. Government Printing Office. Office of Technology Assessment. 1982. Strategies of Medical Technology Assessment, p. 3. Washington, D.C.: U.S. Government Printing Office. Office of Technology Assessment. October 1980. The Implications of Cost-Effectiveness Analysis of 31 Medical Technology; Background Paper #4: The Management of Health Care Technology in Ten Countries. Washington, D. C.: U. S. Government Printing Office. Placek, P., K. Keppel, S. Taffel, and T. Liss. 1983a. Electronic Fetal Monitoring in Relation to Ce- sarean Section Delivery for Live Births and Stillbirths in the United States, 1980. Presented at the American Public Health Association Meetings, Dallas, Novem- ber 16. Placek, D., S. Taffel, and K. Keppel. 1983b. Ma- ternal and infant characteristics associated with cesar- ean section delivery. Health, United States: 1983. Hy- attsville, Md.: National Center for Health Statistics. Ratcliff, J. D. 1957. New surgery for ailing hearts. Reader's Digest 71:70. Renou, P., A. Chang, I. Anderson, and C. Wood. 1976. Controlled trial of fetal intensive care. Am. J. Obstet. Gynecol. 126:470. Rogers, J. L., and O. M. Haring. 1979. The impact of a computerized medical summary system on inci- dence and length of hospital stay. Med. Care 7:618- 630. Sanberg, S. I., B. A. Barnes, M. C. Weinstein, and P. Braun. 1984. Elective Hysterectomy: Benefits, Risks, and Costs. Institute for Health Research (sub- mitted for publication). Vayda, E. 1973. A comparison of surgical rates in Canada and in England and Wales. N. Engl. J. Med. 289:1224. Veterans Administration Cooperative Study Group on Anti-Hypertensive Agents. 1967 and 1970. J. Am. Med. Assoc. 202: 1028 and 213: 1143. Wagner, J. L. 1980. The Feasibility of Economic Evaluation of Oiagnostic Procedures: The Case of CT Scanning. Prepared for the Office of Technology As- sessment, Congress of the United States, Washington, D.C. Wennberg, J., and A. Gittelsohn. 1973. Small area variations in health care delivery. Science 182:1102. Wood, C., P. Renou, J. Oats, E. Farrell, N. Beis- cher, and I. Anderson. 1981. A controlled trial of fetal heart rate monitoring in a low-risk obstetric popula- tion. Am. J. Obstet. Gynecol. 141:527.