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 57
OCR for page 58
This page in the original is blank.
OCR for page 59
A Presentation Summaries Presentation 1 WHAT INTERNATIONAL DATA TELL US NOW Olav Meirik, M.D. World Health Organization, Special Program of Research, Development, and Research Training in Human Reproduction Background This presentation reported provisional final results from the International Collaborative Postmarketing Surveillance led by the World Health Organization Special Program for Research, Development, and Research Training in Human Reproduction, with Family Health International and the Population Council. The purpose of the surveillance was to study, over a 5-year period, major short- to medium-term side effects of Norplant that had not been identified in clinical trials. Methodology The surveillance was based on a controlled concurrent cohort research design with a study population of women aged 18 to 40 at enrollment, and enrolled through a total of 32 family planning clinics in eight countries—Bangladesh, Chile, China, Colombia, Egypt, Indonesia, Sri Lanka, and Thailand. Index subjects were women choosing Norplant in those clinics, and controls age-matched by 5-year bands, who either chose intrauterine devices (IUDs) or sterilization. None had contraindications to Norplant or IUD use. The study population consisted of 7,977 Norplant acceptors (49.8%), 6,625 IUD acceptors (41.1%), and 1,419 sterilized women (8.9%), for a total of 16,021. The largest representation was from China: 6,114 participants, half of whom were Norplant acceptors. The initial study objective was to have no more than 15 percent loss from enrollment, that is, a follow-up level of 85 percent. In fact, overall loss to follow-up over the course of the study was a much lower 3.6 percent: 0.1 percent in China, 7.6 percent in other Asian countries, 3.2 percent in Latin America, and 3.6 percent in Egypt. The study accumulated an average of approximately 4.8 calendar years of follow-up of IUD users, and 5 years of follow-up of both Norplant users and women who had been sterilized. The result was 78,000 woman-years of
OCR for page 60
follow-up overall, 33,600 for current use of Norplant, 29,500 for IUDs, and 7,800 for sterilization. Active follow-up continued for 5 years regardless of method changes and was carried out primarily through scheduled semi-annual visits, including home visits, letters, and telephone calls. Women were encouraged to come to the clinic if they had any health problems, and data from these unscheduled visits were recorded as well. In the case of inpatient care or outpatient clinic visits, records were retrieved and their content recorded. All major health-related events, significant health problems, and contact with health care services were recorded in personal diaries. "Major health-related events" included all deaths and hospitalizations, pregnancies, and morbidity and trauma that were potentially life-threatening, that required hospitalization and/or at least 1 month convalescence, that had long-term sequelae, or that required long-term medications.* "Significant health problems" were defined as virtually anything except common colds and minor injuries, all of which were reviewed by a country coordinator for purposes of standardization. Data were managed and coordinated centrally at WHO/HRP in Geneva, where they were reviewed, entered, and checked. All diagnoses were coded according to ICD-9 categories. Oversight included regular monitoring of clinics, site visits, and extensive correspondence with each clinic and collaborators in the study countries. There were some methodological differences among clinics in assessment of side effects, so that some endpoints may reflect detection bias. For example, frequency of blood pressure checks varied by method use. Blood pressures for Norplant and IUD users were checked at least three times in about 40 percent and 31 percent of those two subpopulations respectively, in contrast to only 15 percent of the sterilized women. It may have been, however, that complaints of headaches among Norplant users led to more frequent blood pressure checks. Other possible sources of bias might be the frequency of hemoglobin measurements and overall numbers of visits. All such sources of bias are being analyzed further. Findings Age and Educational Levels Overall, the majority of women in the study fell into the groups aged 24 to 35; however, women in China and Egypt tended to be older, with age distributions among the South American and Asian countries other than China roughly similar. As for education, those at the highest education levels chose the IUD, followed by Norplant, and then sterilization. The exception was South America, where the educational levels of Norplant and IUD users were almost identical. * Results related to major health-related events were still provisional at the time of the workshop, because some analysis of individual diseases and conditions was still underway. After final review, there may be some reclassification of events.
OCR for page 61
Major Health-Related Events There were 35 deaths in the course of this study, a rate of 0.44 per thousand woman-years. Just one of those was related to a reproductive event: a Sri Lankan woman who discontinued Norplant and became pregnant a year later, had a clandestine abortion, and died in sepsis. Incidence of cardiovascular disease was in all cases below the power of the study. There were no cases of acute myocardial infarction. Three cases of stroke were recorded, two in China (one ischemic, one hemorrhagic) and one (unclassified) in Bangladesh 2 months after Norplant removal-a total of 3 per 100,000. There was one case of venous thromboembolism in a current Norplant user in Sri Lanka. Hypertension seemed somewhat more frequent in Norplant users than in the controls but, as indicated above, the extent to which this was a function of detection bias is unclear. As for neoplastic disease, there were three confirmed cases of invasive breast cancer in current Norplant users and one in a current IUD user, all in China where the age distribution of the cohort in question would predict 2.8 cases over the same time period. There was one case of a clinically diagnosed metastatic breast cancer in a woman in Bangladesh who had used Norplant and oral contraceptives, and two cases of borderline breast malignancy (one phyllodes tumor and one in situ cancer), again, both in China. There was one case of invasive cervical cancer, diagnosed in Chile in a woman in the sterilization group. There were no cases of ovarian cancer. Data were also gathered on diseases that have been associated with oral contraceptive use in general and that have also been addressed in studies of Norplant in the United States. These include gallbladder disease, found in 101 women, a rate of 1.28 per thousand woman-years, just slightly more frequent in Norplant users than in IUD users. As expected, in the 125 cases classified as anemias, incidence was highest among IUD users, lowest among women who had been sterilized. Diabetes mellitus, found in 12 subjects, was more frequent (eight cases) in Norplant users than in IUD users (three cases) or sterilization acceptors (one case), but the difference was not statistically significant. Questions have been raised about a possible relationship between Norplant and systemic lupus erythematosus and collagen diseases, including rheumatoid arthritis and polyarthropathies. The frequencies encountered in the surveillance study samples were far too low to permit any conclusions: that is, three cases of lupus in China (two IUD users) and Egypt (one Norplant user), and nine cases of varying diagnoses of arthritis-related diseases, none long-term. Significant Health-Related Problems Included in this category were mood disturbances, anxiety, and depression; migraine or other headaches; and visual disturbances. Mood disturbances were recorded more frequently among Norplant users than among IUD users, but their
OCR for page 62
incidence was similar to that generally found with other hormonal methods of contraception, mainly oral contraceptives. Incidences of these problems in IUD users and sterilized women were almost identical. Incidences of migraine and other headaches followed the same pattern. As for visual disturbances, while there seemed to be higher incidence in Norplant users (19 cases, of which 15 were in Norplant users), closer scrutiny revealed no causal relationships. Six cases proved to be disorders of refraction, requiring eyeglasses; five cases were various diagnoses including borderline glaucoma, an intraocular foreign body, thyroiditis, cestode infection, and keratitis; and 8 cases (7 in Norplant users) were 1- to 3-month complaints associated with headache or fatigue, all reversible. Continuation and Removals The surveillance study was conducted in family planning clinics chosen for their good quality; all had had experience with Norplant and were familiar with both insertion and removal procedures, which may explain why continuation rates were so high and removal problems so few. Continuation rates for both Norplant and the IUD were exceptionally high: The cumulative 5-year continuation rate for Norplant was approximately 67 percent, for the IUD, 65 percent. Out of 7,977 Norplant insertions, four were problematic (two each in two clinics). And, of the 7,827 removals that had occurred by 5 years of use, 79 had been difficult (10 per 1,000, or 1%), and 46 of those were in the same two clinics. "Difficult" was defined as cases involving broken capsules, removal requiring two sessions, or capsules that had been inserted too deeply. Since one of the clinics had been a training clinic, the question arises whether these difficult removals involved trainees. Both clinics are being evaluated. Efficacy Norplant has a very low contraceptive failure rate of 0.23 per 100 woman-years, compared to 0.15 for female sterilization. Norplant users were also found to be at very low risk of ectopic pregnancy, 0.03 per 100 woman-years, compared to a rate of 0.19 per 100 woman-years for non-contracepting women. Conclusions Prospective postmarketing surveillance studies can now be said to be feasible in developing countries. Their considerable value does not exclude, however, the need for continued follow-up, particularly in the form of operational research.
OCR for page 63
Norplant appears to have patterns of adverse effects that are very similar to those of combined oral contraceptives, with the exception of bleeding disturbances. In settings where postmarketing surveillance studies were carried out, Norplant proved to be a safe, well-tolerated, and highly effective contraceptive method. Presentation 2 DATA AND ANALYSIS FROM POPULATION COUNCIL STUDIES Irving Sivin Center for Biomedical Research Population Council Background Between 1990 and 1996, the Population Council conducted a series of studies to assess the health of women during use of either Norplant or what has been popularly referred to as "Norplant 2," now referred to as the LNG ROD or, outside the United States, Jadelle. The latter is an implant consisting of two rods that slowly release their levonorgestrel contents over an approved duration of efficacy of 3 years. The studies of this new formulation were undertaken to provide sufficient information to permit its registration, as well as to obtain additional information for revision of the efficacy labeling on Norplant. In the discussion that follows, "Norplant" refers to the slightly modified soft-tubing version approved by the Food and Drug Administration (FDA) for use in the United States. Methodology The studies involved a total population of 2,798 women in seven countries,* 43 percent of whom were from the United States. The first randomized study followed 199 women using an old version of the LNG ROD and 199 women using a new version of the LNG ROD. This study measured blood levels of drug, efficacy, and safety. The second randomized study involved 600 women using Norplant with the "soft" tubing and 600 women with the LNG ROD. It was mainly conducted outside the United States. * Chile (2 sites), Dominican Republic, Egypt, Finland, Singapore, Thailand. United States (5 sites: University of Southern California, University of California at San Francisco, Robert Wood Johnson Research Institute, New York University, and Cornell University/ New York Medical Center).
OCR for page 64
The third study, begun in 1990, was a non-randomized comparison study conducted mainly in the United States. This study trained clinic providers in the placement of Norplant in 600 women; the same clinics were subsequently provided with the LNG ROD for an additional 600 women. The studies included sexually active women aged 18 to 40, willing to give informed consent and to make regular visits. Women with histories of ectopic pregnancy and pelvic inflammatory disease since their last pregnancy were excluded. Women with evidence of depression, illness, and epilepsy were also excluded to ensure regular attendance. Follow-up consisted of multiple visits in Year I and semi-annual visits thereafter. Gynecological examinations were performed at each annual visit and, in many clinics, on a semi-annual basis. Findings The decision was made to analyze the three sets of studies collectively as "data on levonorgestrel-releasing implants," because both Norplant and the LNG ROD were found to have similar hormonal release rates. The amount of levonorgestrel released from Norplant was measured at approximately 70 milligrams over 5 years. Release rates for the LNG ROD were essentially identical; while, initially, there is a high daily release, at 200 days that rate decreases to about 50 micrograms per day and then slowly continues to decrease to about 25 micrograms per day. Pregnancy and Continuation Rates The two formulations are also identical in performance with regard to pregnancy and medical reasons for discontinuation: Gross pregnancy rates for both Norplant and the LNG ROD were identical at 0.4 per 100 woman-years at the end of 5 years in a randomized study overseas. For all studies taken together, for both formulations, the gross cumulative pregnancy rate at the end of 5 years was 1.0 per 100 woman-years, with a rate for the fifth year of 0.8 per 100. For the LNG ROD, the cumulative pregnancy rate at the end of 5 years was 1.2 per 100 woman-years, indistinguishable from that of Norplant. For the U.S. components of the study sample, the cumulative pregnancy rate for Norplant was 1.3 per 100 woman-years; for the LNG ROD, it was 0.7 per 100, again statistically indistinguishable and both highly effective. As for ectopic pregnancies, in 9,300 woman-years over the course of the studies, there were two ectopic pregnancies, a rate of 0.22 per 1,000 woman-years. There were no observable differences in discontinuation rates between the two formulations. The 1-year continuation rate for the studies as a group was over 90 out of 100 original adopters; at the end of 3 years, it was 70 per 100. Five-year
OCR for page 65
continuation rates, for Norplant and for the LNG ROD, of Norplant II illustrate no significant differences; both methods had a 5-year continuation rate of over 50 per 100. Reasons for Discontinuation The principal reason for discontinuation of both versions of the implant was change in menstrual patterns, most importantly prolonged or irregular menstrual flow or increased bleeding. Approximately 9 to 10 percent of women using either method had terminated by the end of 2 years as a result of one or more of these problems. Table A-1 lists, for the group of studies as a whole, the conditions reported (limited to those experienced by more than 1 percent of the sample), the associated time points, and the percentages of women discontinuing implant use for those reasons. TABLE A-1 Gross Cumulative Discontinuation Rates per 100, all 1990-1996 Population Council Norplant Studies General Reason Percentage Terminated for General Condition. 1 year Percentage Terminated for General Condition. 5 years Menstrual problems 4 20 Other medical problems 4 20 Planning pregnancy 1 20 For Individual Conditions Percentage Ever with Condition Percentage Terminated for Condition. 5 years 1 year 5 years All Studies Vaginal discharge 9 28 0 Headache 16 28 3 Pelvic pain 9 22 <1 Weight increase 5 18 3 Acne 7 12 1 Mood change 3 5 <1 Non-psychotic depression 1 3 <1 Alopecia (hair loss) 2 4 <1 U.S. Studies Vaginal discharge 10 25 0 Headache 15 25 2 Pelvic pain 10 23 <1 Weight increase 10 22 4 Acne 14 21 <1 Mood change 7 11 2 Non-psychotic depression 1 4 <1 Alopecia 5 9 <1
OCR for page 66
Removal Difficulties Of 349 Norplant removals and 388 LNG ROD removals, 2.6 percent produced a complication, and that percentage was the same for both methods. The most common complications from the use of the LNG ROD resulted from unduly deep placement and multiple or excessively long incisions, each experienced by 1.3 percent of all women using that method. With Norplant, deep placement and bruising were the most frequent difficulties. Other complications such as broken capsules or rods did not produce adverse events. Not surprisingly, the LNG ROD produces fewer removal complications than Norplant and, because it consists of two rather than six elements, requires half the removal time. Even though removal times in the United States were longer than those in the other countries studied-in part because one of the five U.S. clinic sites (University of California at San Francisco) used the more time-consuming "pop-out" method and affected the overall average-removal times for the LNG ROD were shorter than those for Norplant. Mortality and Hospitalization For purposes of control, the Population Council data were compared to a 1995 U.S. hospital discharge survey and to data on high-income women collected in the United Kingdom by Martin Vessey in 1976; the comparisons were also controlled for age since the Vessey study included only women aged 25 and over. Mortality for the age groups represented is expected to be approximately 7 per 10,000. The mortality results per 10,000 woman-years of observation for each of these studies were as follows: Population Council studies, overall (9,300 woman-years): 1.1 Population Council studies, U.S. sites (3,400 woman-years): 0.0 Population Council studies, overall, women >25 (7,400 woman-years): 1.3; Population Council studies, U.S. sites, women >25 (2,600 woman-years): 0.0; and Vessey study (women aged >25): oral contraceptives (31,076 woman-years): 4.8, diaphragm (14,730 woman-years): 4.7, and IUD (10,014 woman-years): 2.0. The single death that occurred during the 5-year course of the Population Council studies was the result of an automobile accident in Bangkok, Thailand. In other words, the mortality rate at 5 years after initiating levonorgestrel implant use was well below the expected rate.
OCR for page 67
Hospitalization rates per 1,000 woman-years of observation were as follows: Population Council studies, overall: 20.7; Population Council studies, U.S. sites: 29.9; Population Council studies, overall, women >25:19.7; Population Council studies, U.S. sites, women >25:25.0; and Vessey study: oral contraceptives: 50.9, diaphragm: 54.0, and IUD: 57.7. The U.S. hospital discharge survey found a hospitalization rate for women aged 15 to 44 of 130,000 to 140,000 per 1000 woman-years depending on survey year; when pregnancy-related hospitalizations are extracted, that rate falls to approximately 62 per 1,000. The rate of 20.7 for the Population Council implant studies as a whole is still notably lower. It is also lower than all hospitalization rates found in the Vessey studies. There is no body system in which implants have elevated hospitalization rates compared with the U.S. hospital discharge survey or Vessey's studies. Conclusions Norplant and the LNG ROD provide essentially identical drug release and clinical performance through 3 to 5 years of use. The cumulative pregnancy rates through 5 years of use are approximately 1 to 1.5 per 100, comparable to those associated with sterilization or the levonorgestrel-releasing IUD. Menstrual and medical complaints associated with the use of these implants are frequent and require counseling before and during use, yet women continue to use the implants at rates higher than almost all other reversible methods of contraception. Removal is markedly faster with the LNG ROD, but training in placement and removal is still required and maintenance of skills essential. Correct insertion is the prerequisite to easy removal. Severe adverse events are uncommon among implant users. Death rates have been zero in the United States in over 3,000 women years and 1.1 per 10,000 woman-years of observation for the 1990-1996 Population Council studies overall. Hospitalization rates among users in the U.S. studies have been substantially below rates for all U.S. women aged 15 to 44. Low pregnancy rates, high continuation rates, and the safety profiles indicate that Norplant and the LNG ROD are a reasonable contraceptive choice for American and non-American women of reproductive age.
OCR for page 96
Presentation 13 IMPLANT REMOVAL AND TRAINING Paul Blumenthal, M.D. Johns Hopkins University Bayview Medical Center Background With the advent of Norplant in the United States, the Bayview Medical Center in Baltimore, Maryland, opted to adopt a structured training program where practitioners would acquire competency in the method before advancing to training in a clinical setting. The decision to do this derived from knowledge of difficulties experienced in other countries, from the desire to profit from the experience of the JHPIEGO Corporation's training program in Indonesia and later in the United Kingdom, and from understandings about the way clinical procedures are taught in the United States. Before putting Norplant on the U.S. market, its distributor, Wyeth-Ayerst, provided support for a national hands-on training program in Norplant insertion for physicians, nurse practitioners, and physicians' assistants, using master trainers in 37 hospital- and clinic-based locations. As unprecedented as it was for a pharmaceutical company to sponsor such an endeavor, the effort was affected by factors that had more to do with predominant medical culture and training traditions in the United States than anything else. First, for the most part, it had to be left to individual practitioners to present themselves for training since there was no way of requiring that they do so. Second was the indeterminate number of physicians who believe they already command enough basic skills to cope with new technical requirements that seem elementary, leaving them not motivated to find time in inevitably demanding schedules for training. Third, medical training, undergraduate or postgraduate, is often provided in circumstances where competency need not be demonstrated, documented, or required before use, an acute limitation when students are unenthusiastic about needing to be in the classroom in the first place. Fourth, the prevailing pattern is that individual practitioners, once exposed to a surgical technique, adapt it as they will, which produces an enormous amount of variability in practice. Finally, in very specific terms, at the time of its U.S. introduction the experience with Norplant, especially removals, was not as deep or extensive in the United States as it was elsewhere, notably Indonesia. Nor was there an effort to replicate the strategy used in the United Kingdom, which had chosen to prepare its master trainers on site in Indonesia, because of the experience acquired in the pre-introductory period in that country. Methodology Medical education in the United States traditionally emphasizes the transmission of information; students are evaluated according to the amount of
OCR for page 97
information absorbed. Most training in clinical procedure occurs via exposure to a technique, followed by freedom to adapt it to suit practitioner preference; heterogeneity of practice is inevitable. In contrast, the Baltimore training program, adapted from that developed by the JHPIEGO Corporation, emphasized transmission of skills and evaluation of performance. It also emphasized standardization: of those skills, the way they would be transmitted, and the eventual performance of trainees. The Baltimore group adapted a set of essential steps, to achieve technical consensus among their faculty on how insertion and removal would be performed and to standardize these agreed upon techniques so that training and skills assessment would be consistent throughout. The one requirement of practitioners recruited to the program was a clear willingness to learn procedures and to practice them repeatedly. Those who intended to become trainers were required in addition to complete a course on skill acquisition and to demonstrate proficiency before being permitted to train others. Levels of skill acquisition and proficiency were defined and a checklist developed to guide trainees and those responsible for evaluating their performance. Parameters for handling procedurally difficult removals were set and accounted for during training. One useful mechanism was development of the "VAP"-Visibility, Arrangement, Palpability-score for standardizing the assessment of the degree of difficulty of a prospective implant removal and anticipating the amount of time needed for the eventual procedure. A scale of 1 to 3 was established, with 1 indicating that all capsules or rods are visible, arrangement occurs in a fan-shaped distribution, and all implants are easily palpable with minimal pressure. Scores of 2 and 3 indicate graded complications within each of these characteristics. Findings The predominant client population in Baltimore for Norplant is young and many are teenagers. The pattern of requests for removal was a peak at 6 months, when most users have not yet experienced the settling down of menstrual irregularities that occurs in most women between 6 to 9 months. "Problem visits" to clinics generally cease after 6 months, because those who remain are generally satisfied or willing to tolerate side effects and those who are not have elected to have the implant removed. Experience with these removals proved that the VAP score did correlate with duration of procedure and was considered a good tool for predicting the time required for implant removal; it also proved useful in predicting difficult removals. The standardized procedure that had been developed for removal was found to require, on average, 15 minutes. There were two provider populations at the center: physicians and physicians' assistants. One limited assessment found that mean removal time was longer for a physician who had not attended any formal training course and was essentially self-taught, than it was for a physician's assistant who took the course.
OCR for page 98
The shorter removal time was also found to have a positive effect on patient attitudes, thought to be vulnerable to the influence of ongoing media coverage of difficult removal experiences. Patient perceptions about the level of removal difficulty also coincided more closely with the perceptions of physicians' assistants than they did with those of physicians. Finally, the speed of the removal procedure correlated strikingly with the extent to which users, even though they had had the implant removed, would still recommend the method to a friend; the perception of the very large majority of these women was that their removal experience was much easier than they had anticipated. Conclusions Another contributor to method success and improved removal experiences will be the removal technique itself. The "U" method of removal was developed and evaluated in Indonesia as a possible alternative method. Mean removal times were calculated for the standard and the U techniques in a total of 250 removals, and the number of removals required for initial and sustained competency was evaluated. Initial competency was defined as fulfilling all the required steps at least once and was assessed to determine whether a practitioner fell back to "incompetency" before his or her competency could be said to be sustained. The result was that sustained competency could be achieved an average of two patients sooner using the U technique; removal times with that technique were also significantly lower and with a smaller percentage of removal problems. Since removal time seems to be inextricably linked to overall patient perception of the method, any removal technology that could expedite removal and therefore improve client perceptions is very desirable. The Baltimore training and clinical experience demonstrated that client attitudes toward Norplant and, ultimately, use of the method in general, can be modified by provision of safe and expedient removal. For the method to be a popular contraceptive option for women, removals need to be provided by competent, well-trained personnel with experience involving a variety of levels of difficulty. Another critical variable would be standardization of implant insertion and removal procedures. At a recent symposium sponsored by Wyeth-Ayerst, participants discussed the possibility of establishing a certification requirement for the insertion and removal of Norplant but concluded that such a propostion was unlikely to gain favor. However, some degree of standardizing the procedures was considered both worthwhile and feasible.
OCR for page 99
Presentation 14 INTRODUCTION OF NORPLANT INTO INDONESIA Ruth Simmons, Ph.D. University of Michigan School of Public Health Background In 1989, when Norplant moved from the research stage into full-scale introduction in Indonesia, many difficulties ensued. To prepare for introduction, the Population Council that same year performed a small study in three provinces of Indonesia focusing on quality of service delivery, in particular on three critical factors: (1) choice, that is, whether women accepting Norplant were given a choice among a range of methods and could make their eventual choice freely; (2) whether removal on demand was available; and (3) the extent to which the Indonesia program was capable of ensuring 5-year removal tracking. Findings Choice Evaluation of the Indonesian program found that choice was not consistently realized. Across the entire program, women's access to information was limited, both with respect to the implant technology itself and alternative method options. Beyond the general tendencies of providers to be authoritative in their guidance and brief with their time, the belief prevailed that women did not need much information, a belief reinforced by a national policy of emphasizing long-acting methods and by the fact that sterilization is negatively sanctioned for religious reasons. Not only was that policy well known but field staff and providers were rewarded for the number of women they recruited to all those methods, including Norplant. These factors were further reinforced by the community-based campaign style of the method's introduction. Much effort went into mobilizing communities to accept the premise that this was the method that women should adopt, and community buildings and schools became the sites of mass efforts to provide insertions, typically under considerable time pressure. Removal on Demand The second critical element identified in the Population Council study was removal on demand. Removal on demand had been assured during field trials when, quite rightly, program managers welcomed the need for training in removal techniques. However, as the method went to full-scale introduction through the
OCR for page 100
national program, women were often given the message that they were making a 5-year commitment and women who later sought to have the implant removed earlier than that were likely to encounter provider reluctance and resistance. Tracking Capacity Even though program managers recognized that not enough providers had been appropriately trained in removal skills, the assumption was that the program basically had 5 years in which to catch up in this regard. When the 5 years were up, concerns about the national program's capacity for tracking adopters proved justified. That capacity was at best uneven and particularly challenged in dense urban areas. Furthermore, the national program could not count on women showing up for removal of their own accord at the end of Norplant's 5-year period of approved efficacy, since it was not at all clear that most women had received enough information at the time of insertion to appreciate the importance of removal 5 years later. In addition, the study identified several problems related to the technical quality of care. The reasons were many: the sheer volume of activity, the pressures of time, the fact that many providers were inadequately prepared in insertion and removal techniques, equipment and supplies were inadequate, and maintenance of an aseptic environment was not a priority. Program evaluation discovered serious problems that emerged when Norplant went to scale. The difficulties were philosophical and strategic, perhaps the natural result of pushing women toward a particular method rather than helping them make choices among methods. The difficulties were also administrative and technical. Launching a brand new, complex contraceptive method for use by large numbers of women in a short period of time would compromise the quality of care and provider-client interaction in most delivery systems, even more so given the cultural constraints and limited resources in the Indonesian program. Conclusions The Indonesian experience, together with lessons from IUD introduction in India in the early years of the method, as well as more recent lessons from Cyclofem introduction in Indonesia, prompted the World Health Organization (WHO) to develop a new program designed to avoid repetition of errors. The foundation of the approach is a broad assessment of key factors before a decision is made to introduce a method: this has already affected the ways in which some countries are making decisions about introducing new methods and overall method mix. Vietnam, for instance, reversed itself on a decision to introduce Norplant, and instead decided to focus more systematically on improving quality of care through the introduction of injectables and to hold off on Norplant introduction until quality of care can be better assured. It is reasonable to expect that, as a general
OCR for page 101
matter, optimal system preparation should head off many of the problems that seem to attend the introduction of contraceptive methods, whatever they are. Presentation 15 LOOKING TO THE FUTURE A FEDERAL STANDARDS DEFENSE: WHAT DIFFERENCE MIGHT IT MAKE?1 Michael D. Green, J.D. College of Law, University of Iowa Background In considering whether a federal standards defense might have made a difference in the case of Norplant, two areas are especially germane: (1) efforts already made to address product liability at both the state and federal levels, and (2) the role and capacity of the Food and Drug Administration (FDA). A number of states have enacted some form of regulatory approval defense that is specific to pharmaceuticals. Of those, Michigan law is the most protective, providing that there is no liability for any drug that has been approved by the FDA, as long as no fraud or bribery has been involved in obtaining premarketing2 approval from that agency. At the federal level, a regulatory approval defense was included in the House version of the Common Sense Products Liability and Legal Reform Act of 1995 but was removed in conference as part of a compromise strategy for avoiding a presidential veto; the veto occurred and the attempt at an override failed.3 The experience with this piece of legislation serves to highlight a major consideration, that is, the extent to which the entire topic of products liability is politicized and the breadth of legal, regulatory, industrial, and consumer interests at play in this particular arena. Discussion The Role of the FDA Much of the strength of a regulatory approval defense rests on the processes of oversight by the FDA that are intended to assure compliance by manufacturers with the agency's regulations, both in the premarketing and postmarketing periods. During the premarketing phases, manufacturers must comply with FDA regulations for adequate and well-controlled studies, accurate reporting of results, and truthful responses to inquiries from the agency associated with New Drug Application (NDA) review. The FDA, in turn, depends for its decisions about product approvals on extensive preclinical and clinical trials that are carried out either by the manufacturer or a contract research organization hired by the manufacturer to assure product safety and efficacy. Oversight of data collection in
OCR for page 102
clinical settings may be carried out by monitors hired by the manufacturer to assure the quality and integrity of the data and the process of its collection. The data required are typically complex and amounts may be massive. For example, the New Drug Application (NDA) for Norplant consisted of 53 volumes of data and analysis. In addition, there are some tensions among the various incentives for industry compliance with FDA requirements that have to do with the costs of testing, in time and money; pressures for approval so as to get a product to market, especially first to market; and the goal that all product risks be fully identified. FDA's capabilities for fulfilling its responsibilities are substantial and constitute the strongest case for a regulatory standards defense. The agency customarily, and increasingly, does an accurate and reliable job during this period, surpassing what can be accomplished on a regular basis through the present tort system because of the technical expertise resident on the agency's staff, as well as available through its advisory committees. Furthermore, the flexible FDA decision-making processes are quite different from the adversarial model of the tort system, as are its processes for assessing relative risks and benefits. The postmarketing period is more problematic and raises issues that are critical for conceptualizing a government standards defense.4 Many adverse reactions simply cannot be identified within the time frames and samples customary for premarketing trials and will emerge only after market introduction as the experienced population increases in size and heterogeneity. A 1990 General Accounting Office study found that of the 198 drugs approved between 1976 and 1985 for which data were available, 102 (51.5 percent) had serious postapproval risks, as evidenced by labeling changes on 96 drugs and removal from the market of 6 drugs. The large majority of the drugs were deemed by the FDA to have benefits outweighing their risks, so that resulting label changes either limited the intended target population for the drug or required addition of major precautionary warnings regarding its use.5 When clinical trials have been well designed and executed and the resulting data have been fully and openly reported to the FDA, later discoveries of adverse reactions are "no one's fault." However, reasonable prospects for such discoveries mean that well-designed, well-executed postmarketing surveillance becomes crucial. In this connection, it is also important for manufacturers to gather and report on adverse events so that risk can be identified and communicated and necessary modifications to labeling can be effected expeditiously. The role of labeling in protecting companies from liability is substantial. The Potential Role of a Federal Standards Defense There is little disagreement about these realities but there is a range of opinions about their implications for developing the sort of defense proposed in the 1990 and 1996 IOM committee reports,6 and pros and cons abound. For example, one interpretation is that a government standards defense might dilute
OCR for page 103
manufacturer interest in reporting adverse events in the postmarketing period because such a defense would have been predicated on data from the premarketing period. An alternative view is that incorporating a postmarketing surveillance requirement into the defense would be an appropriate hedge against such an eventuality and that, in any case, pharmaceutical companies already tend to see such surveillance as prudent and desirable. Furthermore, the FDA already requires reports of adverse events as a matter of course. Looking more broadly, another thought is that a government standards defense might motivate attorneys to probe more deeply into the thoroughness and integrity of company compliance with FDA requirements, rather than to take FDA approval at face value. The effect might be to increase incentives to push such determinations into the courtroom more frequently, an eventuality some might find unappealing. The question was posed as to whether some kind of federal standards defense might have constituted a deterrent to the kind of litigation seen in connection with Norplant. The claims related to Norplant have been for relatively modest injuries, for example, difficulties related to removal, alleged norgestrel-related effects, and silastic-related claims. Thus, something like Michigan's product liability legislation might have had a substantial deterrent effect. Even though such provisions might not provide complete immunity from suit, claims for modest injuries would probably be filtered out, although claims for more serious injuries, such as birth defects, would still be likely to generate suits. This interpretation is, of course, hypothetical, suggesting that an empirical study of state-level product liability laws and their effects would be useful, perhaps essential, if the concept of a federal standards defense is to be addressed further. Another related, bottom-line and, in effect, hypothetical question is whether such legislation would, in fact, encourage companies to re-enter or remain in the contraceptives market. The prevalent wisdom in the field continues to be that the threats of litigation and damaged corporate image are what have chilled industrial interest in contraceptive research and development. However, a number of studies have concluded that the incidence of punitive damage awards against pharmaceutical companies that have become final after all appeal are much lower than is generally perceived.7 Still, the sheer perception of the risk of such damages, with their unpredictably high costs, both in dollars and goodwill lost, persists as a significant factor in pharmaceutical R&D decision-making nevertheless and, although this might be seen as overreaction to what have been found by some analysts to be relatively rare events, the perception in itself appears to be powerful and durable. Whether a workable government standards defense could remove that chill yet maintain incentives for compliance with FDA standards, especially for postapproval risks, and permit reasonable compensation in those instances when a manufacturer's non-compliance creates real risks, is difficult to know. The proof would necessarily have to be in the testing, that is, the actual adoption of such a reform. With regard to the general hypothetical question, as well as the particulars of such a reform, the issues are more complicated than many observers and commentators may have appreciated. Finally, there is the issue of the scope of the defense in question and which approach might be most effectively applied to contraceptives, or at least subsume
OCR for page 104
them. There are several options for consideration: a defense for all consumer products that must meet government standards; a defense limited to FDA-approved pharmaceuticals; or a narrow defense targeted to a specific product group for which there is a widely and urgently perceived need that is likely to rally constituency for protective legislation. Two examples of the last category are the General Aviation Revitalization Act of 1994 and the National Children's Vaccine Act of 1986 and the associated National Vaccine Injury Compensation Program (VICP) implemented in 1988. The Revitalization Act was enacted to reduce the liability faced by aviation manufacturers, perceived as having substantially reduced R&D investment, with consequent negative effects on R&D advances, product quality, export potential, and employment. The VICP is a federal no-fault system designed to provide compensation to those injured by childhood vaccines, whether administered in the private or public sector. It came into being because of a decrease in the number of vaccine-producers owing to liability claims concerning adverse events, and because of apprehension about the effects of that diminution on supplies of existing vaccines and on new vaccine R&D. The program is generally believed to have had salutary effects in these regards. The question of the scope of a statutory corrective that would encompass contraceptives remains unresolved. ENDNOTES 1. Two Institute of Medicine committees studying contraceptive research and development in 1990 and 1996 recommended that the U.S. Congress enact a federal product liability statute that would make FDA approval of contraceptive drugs and devices available to contraceptive manufacturers as a defense to punitive damages. assuming proper compliance with FDA regulatory requirements. Both committees contended that for controversial products that contribute importantly to the public health yet produce only modest profit margins, limits on liability could act as an incentive for research and development or at least could reduce the amount of disincentive. The 1990 committee argued that pharmaceuticals and medical devices are unique among products in the United States in the degree to which quality is regulated before they are released in the market, so that the need for liability as a quality control mechanism is greatly reduced (National Research Council and Institute of Medicine. Developing New Contraceptives: Obstacles and Opportunities. L Mastroianni Jr, PJ Donaldson, TT Kane, eds, "Washington. DC: National Academy Press, 1990; Institute of Medicine. Contraceptive Research and Development: Looking to the Future. PF Harrison, and A Rosenfield, eds. Washington. DC: National Academy Press, 1996). As conceptualized by that committee, with such a statute—variously referred to as a federal, government, or regulatory standards defense; regulatory approval or compliance defense; or simply as an "FDA defense”—companies would not be held liable for punitive damages in a lawsuit under the following assumptions: if the drug or medical device involved had received approval from the FDA and if that company had fully complied with all of the agency's requirements for premarketing testing and postmarketing surveillance. The defense would not, however, bar plaintiffs from obtaining compensatory damages. Nor
OCR for page 105
would it be available to a manufacturer found to have withheld from the FDA either information gathered for purposes of premarketing approval, or information developed after approval for review so as to determine whether the product in question. its marketing, or its labeling should be modified. Some have expressed the view that any violation of the comprehensive regulatory scheme overseen by the FDA that might be causally related to pharmaceutical injury would fall outside the scope of an FDA compliance defense. In other words, a consumer injured by a pharmaceutical or medical device would be free to recover compensatory and punitive damages if the injury would not have occurred if the manufacturer had complied fully with all FDA regulations. 2. Arizona, Colorado, New Jersey, Ohio, Oregon, and Utah have passed legislation that allows the manufacturer of an FDA-approved product to assert a government standards defense in response to claims for punitive damages. In addition, Illinois and North Dakota have adopted a defense to punitive or exemplary damages for products that have been approved by a state or federal regulatory agency with authority to approve the product in question (Institute of Medicine, op. cit., 1996). 3. The regulatory approval defense included in the House version (H.R. 956) of the 1995 bill barred punitive damages in cases where a medical device or drug had won premarketing approval from the FDA. H.R. 956 also capped punitive damages. at either $250,000 or at three times any economic losses, and was written to be applicable to any civil litigation, not just product disputes. The Senate version (S. 565) did not include a regulatory approval defense; also capped punitive damages at $250,000 or three times economic losses, but defined the latter as the greater of lost wages or medical expenses; and applied only to product liability cases (Institute of Medicine, op. cit., 1996). However, the compromise legislation that emerged from conference was vetoed and failed to muster enough votes to override. 4. The 1990 IOM committee, in recommending enactment of a federal product liability statute, spoke frankly on the inadequacy of existing postmarketing surveillance systems for contraceptive products and on the ethical, practical, and economic obstacles to successful postmarketing surveillance. That committee recommended establishment of a comprehensive postmarketing surveillance system to provide systematic and timely feedback about positive and negative health effects of contraceptive products. In addition, both the 1990 and 1996 committees noted that because a regulatory standards defense would necessarily interact with postmarketing surveillance efforts, any recommendation for such a statute would be more compelling were formal postmarketing surveillance studies to be an integral and general requirement. 5. U.S. General Accounting Office. FDA Drug Review: Postapproval Risks 1976-1985 (Report of the Chairman, Subcommittee on Human Resources and Intergovernmental Relations, Committee on Government Operations, House of Representatives). Washington, D.C.: General Accounting Office, Program Evaluation and Methodology Division, 1990. 6. Institute of Medicine, op. cit., 1990 and 1996. 7. MD Green. Statutory compliance and tort liability: Examining the strongest case. University of Michigan Journal of Law Reform 30(2&3), Winter-Spring, 1997). See also: S Daniels, and J Martin. Myth and reality in punitive damages. 75 Minnesota Law Review 1:28-43, 1990; MJ Saks. Do we really know anything about the behavior of the tort litigation system-And why not? 140 University of Pennsylvania Law Review 1147:1254-1262, 1992; S Garber. Product Liability and the Economics of Pharmaceuticals and Medical Devices. Santa Monica, CA: The Rand Corporation, 1993.
OCR for page 106
This page in the original is blank.
Representative terms from entire chapter: