Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices (1987)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Programs and Policies CHAPTER 9 THE EFFECTS OF PROGRAMS AND POLICIE S ON ADOLESCENT PREGNANCY AND CHILDBEARING Sandra L. Hofferth This chapter will explore the effects of a variety of programs and policies on teen sexual activity, contraceptive use, pregnancy and birth. Discussion also will focus on programs and policies concerned with resolving a premarital pregnancy and with the well-being of young mothers and their infants. Although we often think in terms of developing programs and poli- cies to prevent teen pregnancy or to ameliorate its assumed conse- quences, we often tend to overlook the potential feedback effects of programs and policies on teen behavior. Programs and policies have specific objectives: programs for pregnant teens and teen mothers are designed to improve outcomes for child and mother; family planning programs are designed to improve contraceptive use. These are con- sidered to be positive or desirable effects. There may be other un- intended impacts. By reducing the negative consequences of certain behaviors policy makers may be reducing the disincentives to engage in such behaviors. For example, making family planning services available may lead some teens to initiate sexual activity at an earlier age than otherwise. Increasing the availability of welfare may increase the probability that, once pregnant, a girl will bear and keep her baby rather than have an abortion or relinquish the child for adoption. Providing special programs for pregnant teens and teen mothers may increase the probability of a repeat pregnancy. In this chapter re- search on both intended and unintended effects of policies will be reviewed. SE XUAL ACTI VI TY Several types of programs may delay the initiation of sexual activity and prevent pregnancy. Although all are educational programs, each has a slightly different focus, underlying rationale, and metho- dology. The first type to be considered is designed specifically to delay sexual involvement. Projects currently funded by the Office of Adolescent Pregnancy Programs fall into this category. These programs are primarily educational, but have a specific purpose, which is to delay sexual intercourse. One type of project focuses on helping 207

208 young people develop skills to be able to avoid sexual intercourse. Another type of project promotes parent.child communication as a means to delay teen sexual involvement. A second type of program attempts to prevent early involvement in sexual activity and early pregnancy by making young women aware of career options and life choices other than motherhood. It attempts to raise young women's self-esteem and sense of control over their lives so that such alternatives can become realistic ones. Three programs that fall into this area include Project Choice, the Association of Junior Leagues programs, and the G iris Clubs programs. A third type, the most common, falls under the heading of sex edu- cation or family life education. The major goals of sex education are to promote "rational and informed decision-making about sexuality" and to Increase a student's knowledge of reproduction" (held by 97 percent and 77 percent of school districts respectively). Fewer than half cited a desire to reduce teen pregnancy and only 25 percent cited a goal of reducing teen sexual activity (Sonenstein and Pittman, 1984~. A fourth type is a combination of traditional sex education with some new techniques which focus on training students in problem-solving and decision-making skills as well as assertiveness skills so that they can implement the decisions they make. These decisions may include avoiding sexual involvement, or selecting contraception (Schinke et al., 1981~. Such a project may reduce early sexual involvement, but that is not its primary goal. Promotion of Abstinence A major project "Postponing Sexual Involvement" underway in Atlanta, Georgia, directed by Marion Howard of Emory University, has as its goal educating adolescents concerning self-discipline and re- sponsibility in human sexuality (Howard, 1984~. The program, which is designed to help young people (under age 16) resist pressures to become sexually active before they are ready for such involvement, originally consisted of a series of four workshops for young people and an option- al ser ies of workshops for their parents. Each 90 minute session in- corporated factual information and participatory activites designed both to increase knowledge and to build a specific set of skills. Session I focussed on social pressure, Session II on peer pressure, and Session III on problem solving. Session IV, conducted three to six months later was designed to reinforce the skills learned. The critical features of this program are 1) a base in developmental theory, 2) a clear value base, 3) skill-building exercises and activi- ties as well as provision of information, 4) use of peer and parent support. The program was first field-tested in Atlanta and Cleveland, after which two curriculum guides and a companion slide-tape presenta- tion were developed. The project is currently being implemented in the Atlanta public schools as a six-week course for all eighth graders. No evaluation data are yet available.

209 The office of Adolescent Pregnancy Programs is funding a series of 21 projects which have as their goal the prevention of sexual activity among young teens through improved parent-child communication. One such project has as its goal "to enable parents to better communicate their values and attitudes regarding sexual behavior to their children and to help their adolescents develop positive self.concepts and im- proved decision-making skills to enable them to exercise greater re- sponsibility over their sexual behavior n (Montana State University, Bozeman, MT). Another project will offer a values based Life and Family National Demonstration Project to parents and their adolescents for the purpose of promoting sexual restraint among teenagers (Search Institute, Minneapolis, MN). The majority of these projects have just begun; evaluations are not yet available. However, it is apparent from a review of their evaluation plans that although most will measure parent-child communication and related attitudes and values, few will actually measure the impact of the program on the sexual activity of the teens who participate. Thus a major test of the effectiveness of such programs will not result from this research. In 1981 and 1982 the State of California tested and then imple- mented a Family Communication Program in two regions of California, first in the Fresno area and then in the San Francisco Bay Area (Salem and Associates, 19821. The primary goal of the Family Communication Program was to increase the frequency and improve the quality of parent-child communications about sexuality and thereby ultimately to reduce teen pregnancies in California. The immediate objective was for parents of children ages 10 to 17 years old to initiate a verbal communication or increase the number of verbal communications with their children about sexuality. The program made no attempt to mandate the content of those family communications; rather existing community groups were used as vehicles to facilitate and direct such communica- tion. The media campaign used radio and television advertisements, publicity, printed materials and encouragement and publicity for local organizational activities held concurrently with the publicity cam- paign. An evaluation of the program was conducted concurrently (Public Response Associates, 1982~. This included pre- and post-program public opinion surveys, conducted by telephone. The evaluation showed an increase of 14 percentage points in those parents who "use every opportunity" to teach their children about sex," and a decline of 11 percentage points in the number of parents who say that their children initiated more discussions on sexual topics than they did. There was no overall change in attitudes towards sex education, although a slightly increased proportion thought parents were responsible for their children's sex education. The impact of the program was greater in the Fresno area than in the San Francisco Bay area, probably be- cause the program reached more people in Fresno and because San Fran- cisco area parents were better communicators before the program. Finally, television was shown to be more effective than either news- papers or radio in reaching the public.

210 Life Options Approach The intent of Project Choice is to support and help young women explore future careeer options other than young motherhood (Alexander, 1984~. It focuses on making at-risk young women more aware of the variety of life choices which are or can be, available to them. The structure of Project Choice is that of an extra-curricular club with voluntary participation. Meetings are held weekly. Activities are directed toward the development of educational, career and general life options, not simply towards contraceptive behavior. Clubs are composed of students who may not yet be sexually active, as well as of students who are sexually active or who may already be parents. Since these programs are relatively new, few evaluations have been conducted. The evaluation of Project Choice showed no significant im- pact on any of its stated goals (Alexander, 19841. However, the eval- uation of the program was not very rigorously conducted: the control groups were poorly constructed, the nature of the intervention was not clearly delineated, and the outcome measures were abstract. Only a small number of young women participated in this project. Thus it can be concluded that this evaluation could not adequately evaluate the success or failure of this type of intervention. The Girls Clubs have developed an experimental program that started about March 1, 1985 and will last three years (Quinn, L985~. Eight girls clubs across the United States were selected. Four were assigned to the experimental and four to the control condition. All projects received a baseline (pre-test) and will receive a follow-up (post-test) instrument, but in only the experimental group are the programs being implemented. The program consists of 4 components, each of which will be implemented at each experimental site. Since the Girls Clubs in- clude girls of a wide range of ages, different components will be di- rected at different age groups. As the girls age, they will move from one component into another. The first two components are directed at 12 to 14 year olds. The first component consists of a parent-child program which is directed toward increasing communication about sexual issues and values. The specific mechanism will be mother-daughter workshops. The second component, consisting of specific interventions with young adolescents, is designed to postpone their sexual involve- ment. This part is modeled after the Atlanta program directed by Marion Howard. The third and fourth components are directed at older adolescents (ages 15 to 17~. The third component consists of a proj- ect "Choices" developed by the Santa Barbara Girls Clubs to help girls develop career and educational aspirations. The program approach is designed to engage participants in a set of activities that will chal- lenge them to think about their own futures, in the areas of family life and work outside the home. These activities include conducting interviews, developing family budgets, solving puzzles, writing con- clusions for hypothetical life stories, using classified ads to seek housing and employment, caring for a baby, conducting a personal skills inventory, developing a set of personal goals and objectives, and

211 developing an individualized plan for reaching these goals (Quinn, 1984~. A book entitled Choices, a workbook for young women that is based on the course's content and format, has been published. {A come parable book for males has also been developed.) An evaluation of Choices began in 1984 but no results are yet available. The fourth component of the Girls Clubs programs is the "clinic bridge" between educational and clinic services. This is modeled on the comprehensive school-based program developed by the St. Paul, Minnesota, Maternal and Infant Care Project. Educational and clinic services will be offered, perhaps on the club site, to club members. Traditional Sex Education Information on sex education in schools comes f ram two major national studies. According to a 1982 survey of 200 school districts in large US cities conducted by the Urban Institute and jointly admin- istered by the National Association of State Boards of Education, three quarters of school districts offered some sex education (not necessarily a separate course)--in their schools (Sonenstein and Pitt- man, 1984~. A 1977 survey by the National Institute of Education found that only 36 percent of public high schools offered a separate course in family life or sex education (Orr, 1982~. Most schools integrate sex education into the material in other courses. These data are consistent with reports from individual adolescents, of whom three quarters report some sex education instruction before leaving school (Zelnik and Kim, 19821. School districts are remarkably in agreement on the goals of sex education. Ninety four percent agree that a major goal is to promote rational and informed decision-making about sexuality; 77 percent agree that a goal is to increase a student's knowledge of reproduction. 25 percent say that a goal is to reduce teen sexual activity and 21 per- cent say it is to reduce teenage childbearing. Nine of ten districts include physical differences between males and females, pregnancy and childbirth and sexually transmitted diseases in their curriculum. Three quarters include information on sources of contraceptives and on personal or moral values, while only half include information on mas- turbation and homosexuality, and almost none include information on sexual techniques (AGI, 1983; Orr, 19821. Effects of Sex Education Sex education is designed primarily to enhance knowlege. That is, it is designed to improve students' knowledge and understanding about how the body functions and about human sexuality. Some courses are also designed to understand the social context of sexuality, including relationships with others and the social, moral and ethical constraints on its expression. A successful course should be expected at a minimum to increase knowledge in the areas taught. And this generally is found

212 to be the case. Numerous studies find the same thing--courses in human sexuality and family life do increase students' knowlege about the subject (Kirby, 1984; Eisen et al., 1985; Finkel and Finkel, 1984~. Sex education may also affect attitudes, but research to date has documented only limited effects. On the one hand sex education appears to make students more tolerant of others' attitudes and behavior; on the other hand it has not been found to alter the individual's personal attitudes and beliefs (Kirby, 1984~. However, it is precisely this concern that has led many sex educators to focus more attention on dis- cussion of personal and moral beliefs and values about sex and sexual- ity, to make sure that students understand the context of their person- al beliefs, even if other students have different beliefs and values. Although receiving sex education has the strongest statistical association with increased knowledge, there is substantially more in- terest in how sex education affects behavior. In particular, there is concern that sexuality education promotes early or more frequent sexual activity among teens that take such courses. There is also concern that it may not be effective enough in promoting effective and respon- sible contraceptive use among teens sexually active, in particular, effective enough to counteract the alleged incentive effect on sexual activity. Finally, there is interest in identifying, among those exemplary programs, programs that appear to be most successful, and the characteristics that are associated with these successes. Here, as before, we focus on the effects of these programs on sexual activ- ity, contraceptive use, and premarital pregnancy and pregnancy resolution. In an analysis of the effects of having had sex education in high school among a national sample of 15 to 19 year old females in 1979, Zelnik and Kim {1982) found no association between the probability of initiating sexual activity and having had sex education; however, for those who were sexually active, those who had had sex education were more likely to contracept and less likely to become pregnant than those who hadn't. This study was based on survey responses, with only a limited set of questions to measure whether the respondent ever took a sex education course, and with no measure of whether the course was taken before or after the behavior examined. However, the authors believe that the lack of information on the ordering of events pro- duces a conservative bias. That is, the results are weaker than they might have been with more appropriate data. Unfortunately, with the data available it is impossible to tell what about the program or pro- grams might have led to this result. Thus the results are useful but not definitive. There are several fairly recent studies of sex education that are of interest. The Arkansas Family Planning Council (1983) conducted a study of the impact of instituting sex education in about one-third of its schools. It found that births declined at a faster rate from 1978 to 1981 in areas where students received sex education than in areas

213 where they did not. Unfortunately, a reanalysis of this data by Doug Kirby (personal communication, Februrary 1985) for that office found that after controlling for other factors there was no residual impact of the sex education program. It is likely that initial differences between counties, rather than the sex education program, led to the initial conclusion that the program had been successful. A three-year study of Family Life Education was conducted in 1979- 81 in 12 California School Districts by ETR Associates (Cooper, 1983~. The study found that the program was highly effective in increasing students' knowledge and enhancing students' self esteem and decision making skills, in increasing parent-child communication and even in reducing pregnancy rates. Unfortunately, the study did not have a control group. As a result, it is not possible to assess whether or not the control group's knowledge, skills, etc., would have improved and their pregnancy rates declined as well. This is not a trivial concern since the Kirby study (reported below) found that changes over time in the control groups equalled changes in experimental groups. Mathtech Study The most recent evaluation of sexuality education programs was conducted by Kirby (19841. The major purpose of this research was to find, develop and evaluate promising approaches to sexuality educa- tion" (Kirby, 1984:47~. Accordingly, and guided by numerous experts, promising programs were selected and improved, before the evaluation was begun. Kirby focused on 12 exemplary programs in nine sites around the country (some sites have more than one program). Four of these programs were school-based, constituted a separate course, and were at least one semester in length. Five were short programs pre- sented either in the school, in community centers, or other locations which lasted from 5 to 10 hours, either in one day or over several weeks. Three additional programs one a peer education program, a second a parent-child program, and the last a combination education/ clinic program in a school setting--were also examined. In all of the non-clinic sites, data were collected prior to the course, imme- diate~y after the course ended and then 3.5 months later. Control groups were selected for each program. The changes across each site over time were compared with changes in the control group to determine the effects of each program. Most programs increased students' knowledge. Classes with younger students learned more than those with older students. Surprisingly, the longer courses did not appear to improve knowlege more than the shorter courses. However, there were a number of important differ- ences between short and long programs, including different lengths of follow-up periods, different control groups, and differences in par- ticipants and curricula, that might have resulted in this unexpected result.

214 Only in three short programs at one site did clarity of values in- crease. Most programs did not increase clarity of long term goals, clarity of personal sexual values and understanding of personal re- sponse to sexual situations. In general, the programs did not increase liberality or permissive- ness of student's attitudes toward premarital sex. In contrast, the attitudes of the control groups did become slightly more liberal. Programs did not affect attitudes toward the importance of birth control. Atitudes were positive and high in experimental and control groups. Scores of both groups increased. Few impacts on other attitudes, such as toward gender roles, sex- uality in life or the importance of the family were found. There was in some programs an increase in opposition to use of pressure and force in social and sexual relations. There was little impact on self esteem, little impact on skills in social decision-making, communication, sexual decision-making and com- munication about birth control. The only program to increase comfort talking about sex, birth control, sexuality with parents, and the fre- quency of reported conversations about sex, birth control with parents, friends and boyfriends was the parent.child program for young children. There was little impact upon comfort with other social and sexual activities. Finally, there was no impact of the programs on sexual behavior, contraceptive use (frequency of sex without birth control or effective- ness of birth control), or pregnancies. Student and parent assessments of the courses and their impacts were generally positive to enthusiastic. Recent Analyses of National Surveys Because of the lack of national surveys with information on sex education ever received, in the last five years questions were added to the number of national surveys which included teenagers. Recently, analyses of these new data sources have become available. Questions on whether the respondent ever had a sex education course, as well as information on sex activity and with whom ever talked with about sex were included in the 1981 wave of the National Survey of Children, a survey of children who were 12 to 16 at the time. A subset of teens 15 to 16 were analyzed (Moore et al., 19857. The authors found a sig- nificant association for white males and females and for black females, such that those who had received sex education were less likely to re- port that they had had intercourse, and this association held up when controls for family income, mother's education, the mother's age at first birth, family structure, and community size were included. The

215 only group the association did not hold for was black males. However, with no information on the timing of either the sex education course or first intercourse in these data, a causal relationship could not be tested. The National Survey of Family Growth included questions not only on whether the respondent had ever had sex education, but whether the course included instruction on methods of birth control and the age of the respondent when that instruction took place. In addition, the month and year of first intercourse was obtained. This allowed the researcher (Dawson, 1986) to establish the timing of sex education with regards to intercourse for all respondents except those who re- ported contraceptive education and first intercourse at the same year of age. As a result the researcher tested two models: the first made the extreme assumption that all women receiving formal contraceptive education at age x received that education at the start of that year of age, i.e., at exact age x. Under this assumption all women whose first intercourse occurred during the same year as sex education would have received the sex education first. The second model made the equally extreme assumption that all women receiving formal contracep- tive education at age x received that education at the end of that year of age. Under this assumption none of the women would have re- ceived sex education before first intercourse. The author found no effect of having received sex education under the latter assumption. Under the first assumption they found one significant effect: 14 year old s who had contraceptive education by exact age x were more likely to initiate sex within the next year. None of the coefficients for the other single-year-of-age groups (IS, 16, 17, 18) were significant. These results held controlling for factors such as race, parental edu- cation, and religion that are also associated with early initiation of sexual intercourse. The author concludes that no evidence for a causal effect of contraceptive education on first intercourse was found in these data. However, the author found evidence that having had contraceptive education is associated with contraceptive use at first intercourse. Under both the above assumptions those who had contraceptive education were more likely than those who had not to use a contraceptive method at first intercourse. Those who had ever had contraceptive education were more likely to have ever used contraception. The authors found no evidence for a direct effect of contraceptive education on the probabil ity of a premarital pregnancy. The third study was conducted on data from the National Longitu- dinal Survey of Youth, 1984 wave (Marsigilio and Mott, 19861. In that year questions were asked about whether the youth had had a course related to sex education and whether it included information about contraception, the female monthly cycle and so on. In addition it asked the month and year in which the first such course was taken. The survey also obtained the month and year in which the respondent first had sexual intercourse. Thus this survey provides the best

216 possible information to precisely determine the timing of first inter- course relative to a sex education course. The authors found that females who had a sex education course prior to their 15th and 16th birthdays, respectively, were more likely to have had intercourse during their 15th and 16th years. The coefficients were large and statistically significant, net of a number of other factors also found to be associated with initiation of sexual intercourse, such as race, church attendance, parental education, and birth cohort. The coeffi- cients declined in size and were no longer statistically significant at ages 17 and 18. The authors interpret these results as providing some evidence that sex education can increase the probability of en- gaging in sexual intercourse at young ages. On the other hand, there may be factors associated with instituting sex education at an early age that are also associated with early sex, such as characteristics of the school or community that are not controlled. If so, then the relationship could be spurious. This is the first rigorously conducted scientific study to suggest that sex education may be associated with increased intercourse, thus it is an important one, but more research should be conducted to test the thesis. Since the data were obtained in a survey instrument, only a very minimal amount of information about the course could be obtained. In particular, there was no in- formation about the length of the course, or any of the other factors that were considered in the Kirby (1984) study of a number of model programs. These authors (Most and Marsiglio, 1986) found some evidence that having had a sex education course was associated with a higher likeli- hood of currently contracepting among 17 and 18 year old women; how- ever, the measure of contraceptive use available in the data is very limited. Finally the authors explored but found no relationship be- tween having had a sex education course and whether had a premarital pregnancy after taking that course. The signs were negative, but the coefficients were not significant. Assertiveness/Decision-Making Approaches Schinke and Gilchrist (1984) and Schinke et al. (1981) have been utilizing what they call a Life Skills Counseling approach with adoles- cents. Its six components include information, problem-solving, self- instruction, coping, communication and support systems. Schinke and colleagues argue that pregnancy prevention requires adolescents to think analytically and rationally about their sexual behavior. To do so they need not only factual information, but also problem-solving and decision-making skills and interpersonal communication skills so as to be able to implement those decisions. Schinke and colleagues have implemented a number of programs utilizing this approach. These programs provide information, but they also train students in problem- solving and decision-making skills as well as assertiveness skills so that they can carry out their decisions. For example, if a youth de- cides to avoid involvement in sexual activity, he or she will have the

217 skills to resist pressure to participate. In contrast to the Howard project, the Schinke project takes no explicit value position on be- havior. Each individual develops his/her own goals and objectives. The Schinke project also includes contraceptive decision-making and teaches assertiveness in active sexual relationships. The skills emphasized in the course are taught through modeling, role playing and rehearsal. The Schinke project (1981) has evaluation data for ~ total of 44 subjects in the experimental condition and 49 controls. Data from the six, nine and 12 month follow-ups show that the youth who took the course had better problem.solving and communication skills, and more knowledge of reproduction and birth control than those who didn't. They also had more favorable attitudes toward family planning, more habitual contraception, greater protection at last intercourse and less reliance on inadequate birth control methods than untrained teens (con- trol group). Unfortunately, no information on sexual activity was reported for these subjects. Summary In summary, research conducted on young men and young women has shown that sex education increases knowledge about sexuality; however, the evidence for an impact on behavior is weak. Assertiveness/deci- sion-making approaches along with sex education look promising, but have not yet been adequately evaluated. Life options approaches are new and lack evaluation. Finally, research has shown that special parent-child programs can and do result in increased parent-child come munication about sexuality, at least for a short period following the program {see Kirby, 1984:317-350~. However, support for the hypothesis that communication reduces teen sexual activity is weak (see Fox, 1980; Newcomer, 1983; Kahn et al., 1984, and discussion in Chapter 1~. Un- fortunately, a major test of the effectiveness of such programs will not result from the research described above because of weaknesses in the evaluation designs. Male-female differences in responses to these programs have not been explored. CONT RACEP T T ON/ PREGNANCY P REVENT I ON The types of programs that are directed at contraceptive use and pregnancy prevention include the assertiveness/decision-making ap- proach, sex education, family planning services, school-based pro- grams, and non-school based multi-purpose youth centers. Assertiveness/decision-making approaches were covered in the previous section. Schinke et al. (1981) found a significant program impact on contraceptive use at the six month follow-up. The charac- teristics of sex education were reviewed in the previous section and won't be reviewed here. As described earlier, according to the most

218 rigorous evaluation of a small number of sex education programs across the United States, no significant impact of these programs on contra- ceptive use (e.g., frequency of sex without birth control, effective- ness of birth control used), or on pregnancies in the short term was found. However, three recent studies based on nationally representa- tive samples of young women (Zeloik and Kim, 1982; Dawson, 1986; Mott and Marsiglio, 1986) found that young women who said they had had a sex education course in school also appeared to be more effective contra- ceptors. The evidence that sex education is associated with better contraceptive use among sexually active teenagers is strong and consis- tent. Family Planning Services Family planning means the provision of information and services re- lating to birth control primarily to women. In 1981 an estimated 4.6 million women obtained family planning services from organized provi- ders; about one third or 1.5 million were teenagers. Another 1.4 mil- lion teenagers visited private physicians. These 2.9 million teens represent about 57 percent of the estimated 5 million teens under age 20 at risk of unintended pregnancy in 1981 (Torres and Forrest, 1983~. Of the 2504 different agencies providing services, the majority (56 percent) were health departments, 13 percent were hospitals, 7 percent planned parenthood agencies, and the remaining 24 percent a variety of other types of organizations. Thus family planning services include services provided by private physicians as well as family planning clinics such as those operated by Planned Parenthood and public health departments. The Relationship between Contraception and Pregnancy Much research has shown that those women who use contraceptives experience lower rates of pregnancy than those who use no contracep- tive method. Typical pregnancy rates for 100 teenage women who start out the year employing a given method range from 2 for a combined birth control pill to 54 for those using no method. Pregnancy rates are con- sistently lower for contraceptors than non-contraceptors, for consis- tent users compared with inconsistent users, and for those using medi- cally prescribed methods compared with those using non-medically pre- scribed methods (except for rhythm). Given that we know that consistent use of effective methods of contraception is associated with a lower risk of pregnancy, the ques- tions that arise include the following: 1) What contraceptive methods do teens use? 2) Where do teens obtain these methods? and 3) What pro- grams are most effective in promoting the consistent use of effective methods by teenagers?

219 Methods Teens Use What methods do teens use? A 1982 study (Forrest and Henshaw, 1983) of the methods used by women (of all marital statuses) who are at risk of unintended pregnancy shows that among those 15-19, 18 per- cent are using no method. Of those using a method, 53 percent are currently using the pill, 3 percent the IUD, 1 percent the diaphragm, 25 percent the condom, and 18 percent other methods. There were few sterilizations. Another 1982 study (Bachrach, 1984) shows that of never-married teens 15-19 who currently used a contraceptive method, 62 percent were using the pill, 0.9 percent the IUD, 6.4 percent the diaphragm, 22 per- cent the condom and 7.8 percent other methods. The results, while not identical, are similar. Where Do Teens Obtain Contraception? Where do teens obtain what kind of method? A study by Forrest and Henshaw (1983) shows that medical methods (pill, IUD, diaphragm) are obtained from physicians and clinics, while non-medical methods (con- dom, spermicides) are obtained from retail stores, with a small propor- tion obtaining them from clinics. Both physicians and clinics should be particularly effective in increasing contraceptive use and preventing pregnancy, with retail establishments ranking third. Although we have much less information on physician-obtained contraception than that obtained in clinics, the information we have suggests that individuals do not depend uniquely on one source. For example, one study (HHS, 1978) found that 12 per- cent of those clients who didn't return to the clinic in 6 months had actually changed providers, with 2 out of 3 of these going to a private physician. Differences between Physicians and Clinics There are important differences between physicians and clinics, however, in actual practice. According to the recent Orr study (1984), 92 percent of obstetrician-gynecologists, but only 66 percent of general practitioners and 32 percent of pediatricians would prescribe a contraceptive method for a sexually active unmarried teenager younger than 18 without parental consent. Of those who will prescribe contra- ceptives for adolescents, 90 percent will prescribe the pill; smaller proportions will prescribe other methods (IUD, diaphragm). Thus, in actual practice young teenagers have less access to medically pre- scribed contraception through private physicians than clinics. In addition, there are substantial cost differences between physicians and clinics. Orr (1984) estimated that the average fee

220 charged by private physicians for an initial family planning visit ranged from $34 for a GP, $35 for a pediatrician to $42 for an OB-GYN, for an average of 637. This does not include the cost of a supply of pills, which ranges from $8.75 to $15.00 per month (Atlanta, 1984) (Conversation with Bob Hatcher) (see Torres and Forrest, 1983, for an average estimate as of 1982 of $8.26 per cycle). In contrast, 1982 clinic fees averaged $9 for patients of all ages, including the 35 per- cent who received services provided free by the clinic or paid for by Medicaid. The average fee per visit among patients who paid was about $14 (Torres and Forrest, 1983~. Women who obtained their prescrip- tions through a clinic often paid much less than the full cost. Among teenagers, four of ten clinic patients are served without charge or are covered by Medicaid. For those who do pay, the average clinic fee averaged $11 in 1981. Moreover, the younger the client the less likely she is to be served at no charge--from one-third of 18-19 year olds to half of those under age 17 (Torres and Forrest, 1982~. The average full cost to a full-paying patient for an initial visit plus three months worth of pills would have averaged about $50 in a family plan- ning clinic in 1981 ($14 for first visit plus 836 for pills) compared with about $76 for a private physician ($40 for first visit plus $36 for pills). Data from Chamie et al. (1982) indicate that the primary reason adolescent clinic clients give for choosing a family planning clinic rather than a private physician for contraceptive care is that the physician is too expensive (65 percent). The second most impor- tant reason is that the physician might tell parents (26 percent). Thus it is not surprising that physicians and clinics have a dif- ferent clientele. Clinic patients are more likely to be black and to have been younger at first intercourse (Zelnik et al., 1984~. Although Zelnik et al. (1984) found that teens who obtained their contraceptive method from a clinic are more likely to become pregnant than those who obtained a method from a private physician, once controls are intro- duced for race and for age at first use of contraception, source of contraception is not a significant determinant of pregnancy. Rather the pattern of prior contraceptive use, race and SES are the major determinants of pregnancy (Zelnik et al., 1984~. Differences Between Clinic/Physician Patients and Other Women Besides the differences pointed out above between clinic patients and physician patients in race, SES and age, there is a major differ- ence between teens going to clinics or to private physicians and those who don't (who may, for example, use non-prescription methods from a drugstore). The former are at higher than average risk of unintended pregnancy, because of their greater level of sexual activity. Zabin and Clark (1981) found that 37 percent of the teens who gave a reason for making their first visit to a clinic mentioned having a closer relationship with their partner and 29 percent mentioned that they ex- pected to have sex more often. The data seem to show a higher risk of pregnancy when relationships are changing, for example, in the early

221 months of experience (Zabin et al., 1979~. But frequency of sexual activity is probably still the best measure of risk, and most girls who attend clinics are currently sexually active (Zabin and Clark, 19811. The Impact of Family Planning Services on Contraceptive Use It is clear that family planning clinics change their patients' contraceptive behavior. Whereas 32 percent of teenage patients were using the pill before enrollment, 80 percent are using the pill after enrollment (according to 1976 data) (Forrest et al., 1981~. The pro- portion using the IUD more than doubles, while the proportion using condom and rhythm decline dramatically. The proportion using no method is reduced from 50 to 8 percent. Based upon these data Forrest et al. (1981) estimate that the difference in expected number of pregnancies before and after clinc attendance declines by 208-272 per 1,000 patients, frog 283-359 per 1,000 to 75-87 per 1,000. Contraceptive Continuation vs. Clinic Continuation Clinic continuation is related to, but not identical with, method continuation. Shea et al. (1984) found that 72 percent of the adoles- cents who did not return in the first three months and 71 percent who did not return in the first six months were inconsistent contraceptors. However, adolescents who made the most revisits, particularly in the first two months, were also likely to be inconsistent contraceptors. Such adolescents were probably having trouble with their method and this may have both discouraged them from using it and brought them in sooner than scheduled for a follow-up visit. The one study (Coughlin, 1978) that followed up a sample of 77 teens who had not attended a clinic for 6 months found that 23 percent (18) were not sexually ac- tive, 36 percent (28) had pregnancies, 12 percent (9) had changed pro- viders, 6 percent (5) had moved, 21 percent (16) were still sexually active but using no method or a non-prescription method, and 1 was still using the prescribed method (diaphragm) but had not returned for a checkup. Of those successfully contacted, 44 percent (32) were not at risk of a pregnancy. In summary, clinics satisfy an important need for family services, but they are not the only providers of contracep- tion. New methods are being introduced, some of which are available without a prescription (e.g., the sponge). The Growth in Family Planning Clinic Programs The growth in the number of women obtaining services from organized family planning clinic programs has been tremendous: an increase from 860,000 in 1968 to 5 million in 1983. Between 1968 and 1983 the number of provider agencies rose from about 1,400 separate agencies operating 1,800 clinics to 2,500 agencies administering 5,000 clinics (Torres and

222 Forrest, 1985~. The average number of patients served by each agency increased from 600 to more than 2,000 between 1968 and 1983. The proportion of clinic patients who were younger than 20 rose from 20 percent in 1969 to 32 percent in 1983. Patients 17 years of age or younger accounted for nearly all of the increase; the propor- tion of these young women served rose from nine percent in 1972 to 15 percent in 1979, declining slightly to 14 percent in 1983. The pro- portion of 18-19 year olds remained constant at 18 percent from 1969 to 1983. Overall, the number of teen patients utilizing family plan- ning clinics increased six-fold between 1969 and 1983, from 214,000 to 1.5 million (Torres and Forrest, 1985~. During the 1970s teen childbearing rates declined; however, preg- nancy rates continued to rise. The major reason for the increase in pregnancy rates was the substantial increase in the number of teens sexually active over the period. Pregnancy rates among those sexually active actually declined over the period. Because of increased abor- tion, birth rates for all teen women declined. Thus in the 1970s large numbers of teens were being served by family planning clinics, although coverage varied from county to county, and rates of teen childbearing were declining. Are organized family planning programs responsible for part of this decline? What impact, if any, have organized family planning programs had on teen pregnancy and childbearing? The Effect of Family Planning Programs on Pregnancy and Birth A successful family planning program should prevent pregnancies since the use of contraception reduces the odds of pregnancy among teens who have sexual intercourse. The only study to look at the impact of family planning programs on teen pregnancy was conducted by Moore and Caldwell (19771. Moore and Caldwell used data collected in 1971 by Kantner and Zelnik from a national sample of females 15 to 19. Policy and program characteristics of the respondent's state of resi- dence at or just before the time of survey, including availability of family planning and abortion, AFDC benefit levels and acceptance rates, and whether the AFDC program covered an unemployed father were appended to the individual's data record. Moore and Caldwell (1977) found that in 1971, black teens aged 16 to 18 living in areas with the most sub- sidized family planning services were significantly less likely to be- come pregnant than their peers. This was not found for other sub- groups. According to the authors, black teens are overrepresented among users of subsidized family planning services compared to whites, and thus may be more affected by the availability of such services (Moore and Caldwell, 1977~. No other study has looked at the relationship between family plan- ning and pregnancy. Rather they have looked at birthrates, lacking

223 individual or local area data on abortions and miscarriages. Since fewer than half of the unintended pregnancies prevented by the family planning program would have shown up as births, birthrates should show less impact from the program than pregnancy rates (Forrest et al., 1981~. Abortion and unintended births would be the appropriate target; given lack of information on abortions and unintended births, some researchers have focused on out-of.wedlock births, which are assumed to be unintended. In an analysis of state out-of-wedlock birth rates in 1974, Moore and Caldwell (1977) found family planning availability associated with a lower out-of-wedlock birth rate among black teens, although no sta- tistically significant effect was found for white teens or older women. In a recent analysis using 1975 state and SMSA data along with the Sur- vey of Income and Education, Moore (1980) was unable to find an impact of family planning availability on birth rates, either all teen rates or out-of-wedlock teen birth rates. The generalizability of this analysis is, however, severely limited by problems in obtaining accu- rate measures of number of births and birth rates, given the particular data used. Three other studies have found family planning to be associated with lower birth rates at the state, SMSA, and county levels. Edward Brann (1979) regressed state level indicators of teen family planning coverage, and a number of other state indicators on state birth rates in 1974 and on change in state birth rates between 1970 and 1974. He found that increased availability of family planning ser- vices was associated with a substantial drop in teen fertility between 1970 and 1974, net of race, level of income and education, percent urban, and the abortion ratio. Field (1981) regressed out-of-wedlock teen birth rates in SMSA's in 1971 on a set of characteristics of SMSA's such as race and age composition, educational attainment and income levels and a set of policy and program characteristics such as AFDC benefit levels, family planning availability, and liberality of abortion laws in 1970. Net of other factors, family planning availability was found to be asso- ciated with a lower level of out-of-wedlock teen births in SMSA's. A study which looked at characteristics of counties or groups of counties (Forrest et al., 1981) found family planning clinic enrollment to be associated with substantially lower white teen birthrates in 1976, net of other factors; and an increase in family planning clinic enrollment between 1970 and 1975 also was strongly associated with a drop in white birth rates between 1970 and 1976. When area differences in adolescent sexual activity were controlled, both white and non-white adolescent birthrates were found to have been reduced between 1970 and 1975 as a result of enrollment by teens in family planning clinics.

224 This study (Forrest et al., 1981) demonstrates how difficult it is to establish causality using regression analysis and cross-sectional data. In 1970 there was actually found a positive relationship between clinic enrollment and adolescent birthrates. That is, the greater the clinic enrollment, the higher the adolescent birthrates. However, this positive relationship weakened and turned negative over time. That is, it changed to a relationship such that greater clinic enrollments were associated with lower birthrates. The authors conclude that early on clinic sites were more likely to be located in areas that had rela- tively high birthrates. However, over time the family planning program did lead to lower birthrates (Forrest et al., 19811. The association between greater clinic enrollment and lower birthrates became stronger when statistical techniques were used to control for differences among counties in the level of sexual activity, which could not be directly measured. In conclusion, the evidence consistently shows an association be- tween family planning programs in local areas and reduced teen child- bearing in those areas. The only study that used data characterizing areas such as counties along with individual characteristics and out- comes (Moore and Caldwell, 1977) found an association between the availability of family planning services and a lower incidence of pregnancy among black teens 16-18. The Impact of Family Planning Clinic Programs on Sexual Activity It has been said that the introduction of the oral contraceptive in 1960 revolutionized society. It has been credited for everything from loosening of sexual mores to the increased employment of women outside the home. By separating the r isk of conception from sex it probably has changed the way generations that grew up with it think about sex and childbearing. It is estimated that during the twenty years after its introduction oral contraceptives have been used by an estimated 150 million women around the world (Ory et al., 1980~. How- ever, it is important to distinguish between the general availability of modern contraceptives, such as the pill, from specific places and programs that provide contraceptive services, such as family planning clinics. The pill has been available since the early 1960s. Clinics have grown up to provide needed services, but oral contraceptives have been available since the early 1960s from private physicians. Elimi- nating all family planning clinics would not eliminate pill use, al- though it would certainly make it harder to obtain. Clinics are more likely to respond to demand than to create it, although they may satisfy the needs of some groups who would ordinarily not have access to certain contraceptive methods. Critics of family planning programs have claimed that expenditures on family planning and sex education actually cause higher rates of abortions and births. Kasun (1982) presents data which suggests that California, which greatly increased levels of spending for family plan

225 ning and sex education, also had high rates of increase in teen abor- tion and illegitimacy. Her conclusion is that increases in spending on family planning increase sexual activity and, as a result, increase pregnancies, abortions and births among teenagers. However, she did not control for initial differences between California and the rest of the U.S., nor did she actually conduct a statistical analysis. The Forrest et al. (1981) study showed that there was, in fact, a positive correlation between clinic enrollment in 1970 and adolescent birthrates in the same year. However, this positive relationship weakened and turned negative over time. This trend suggests that early on clinics were more likely to be located in areas that had relatively high birth- rates, and that this was the reason for the positive association, rather than clinics causing higher birthrates. Associations do not, of course, show causation. For this reason, evidence of associated changes over time, net of initial differences, is stronger evidence for causal mechanisms. There is no such study of changes in sexual activity over time. One unintended and offsetting consequence of family planning pro- grams may be that of increasing teen sexual activity. The easy availa- bility of contraception through organized family planning programs may allow teens who would not otherwise engage in sex because of fear of pregnancy to do so or may legitimize early and non-marital sex. And, in fact, the period of greatest increase in teen sexual activity was paralleled by a tremendous growth in organized family planning activi- ties. What evidence is there on this issue? In the only study to examine the impact of the availability of family planning clinic ser- vices on the initiation of sexual activity net of other factors that also affect it (age, SES, family structure, urban/rural residence, religiosity, birth cohort), Moore and Caldwell (1977) found no asso- ciation between family planning availability and the probability of a virgin teen female initiating sexual activity. More research is needed on this issue. However, it seems most likely that the rela- tionship between the two is spurious; both respond to similar societal pressures. While there is no evidence on the basis of which to accept or re- ject the hypothesis that the availability of family planning clinics affects the sexual activity of individual women, it may still be the case that the availability of contraception in general has affected all teens. Teens know it is available so they may be less afraid to initiate sexual activity. It would be very hard to sort out the effects of the availability of contraception in general and the acces- sibility of family planning clinics from other related changes in social climate and mores over the past two decades. In addition, it seems like a rather pointless task. It is unlikely that the trend to- ward greater availiablity of contraceptives to teens will be reversed; if anything the trend is toward fewer restrictions on their availa- bility to unmarried minors (Bush, 19831.

226 Factors Affecting Clinic Use and Effectiveness in Serving Teens Previous research has shown that organized family planning programs are successful in meeting their objectives of reducing unwanted preg- nancies and births. However, there is probably substantial variation among clinics in their ability to meet these goals. In addition state laws vary, as does funding for such programs (see, for example, Bush, 1983~. A few recent studies have explored the specific aspects of clinics that are associated with greater success. There are two types of outcomes that could be considered successes: 1. Drawing teens into the clinic to begin with and 2. Keeping themr-including continuation at the clinic in terms of making and keeping appointments at regular intervals and continuing use of the contraception prescribed or obtaining a substitute if a method is discontinued. Just because teens do not continue at the same clinic does not necessarily mean they are not contracepting (they may go to a private physician or switch to non-prescription methods, for example. In addition, teens who continue at the clinic are~not necessarily contra- cepting effectively and continuously. Nor are teens who have never attended a clinic necessarily failing to contracept. The most recent study to look at what factors draw teens into clinics is the Kisker (1984) study. This study is based on AGI's sur- vey of family planning clinic directors, patients, pharmacies, and pri- vate physicians in 1981. The family planning clinic is the unit of analysis. The indicator of effectiveness in drawing teens is the mean delay among teenage clinic patients between first intercourse and first clinic visit. Planned Parenthood clinics, medium size clinics (1,000 to 2,500 clients) and non-metropolitan clinics have the lowest mean de- lay. Using multiple regression, a number of factors were identified as important determinants of mean delay. These are related to outreach and community relations, the convenience of attending the clinic, and the clinic's competition in providing contraceptive services. Clinics that offer a community education program for teens, obtain support of local church groups, develop active relationships with local youth groups, are open in the evening and on weekends, accept walk-in clients, are conveniently located, require less educator/counselor time per patient, and provide fewer services have less of a delay be- tween first intercourse and first visit. Mean delay is increased if pharmacies in the community make non-prescription contraceptives easier to obtain. The more private physicians there are who provide family planning services to teens does not appear to affect mean delay in attending clinics; however, the more they charge the shorter the delay in attending a clinic. Finally mean delay is shorter in more prosperous areas where mean levels of schooling are higher, and where levels of teen pregnancy are lower.

227 Several studies have explored factors associated with clinic con- tinuation. The Risker (1984) study used clinic directors' estimates of the number of the clinic's teen patients who were continuing patients in 1980. AGI then calculated the percentage of continuing patients in 1980 adjusted for growth or decline in the total number of teenage patients served annually and for aging into the adult patient group. Medium size and non-metropolitan clinics had a better record of clinic retention by this measure. Clinics with an active relation- ship with local youth groups, who had evening and weekend hours, who provided more services and who required less educator/counselor time retained more clients. In addition, clinics in higher socioeconomic status areas and in areas with fewer nonwhites retained more clients. Nathanson and Becker {1984) studied clients and the professional staff of 78 Maryland county health department family planning clinics. In a study using clinics as the unit of analysis, contraceptive use was measured as the proportion of time subsequent to the baseline in- terview that women at risk of pregnancy were using a medical method of contraception. To obtain clinic estimates, these individual measures were then aggregated over all clients for each clinic. The predictor - variables measured quality of interactions between clients and staff. It was assumed that other organizational and provider characteristics operate through their effects on interaction quality. There were four critical interaction dimensions, measured from the perspective of staff expectations, client expectations, and client's report of what actually happened. These four dimensions were: 1. the relative amount of control or direction exercised by clients as compared with practitioners; the scope of interaction (range of concerns about client); 3. the level of trust placed by client in practitioner; and 4. the "warmth" of the relationship. The warmth of the relationship appeared to be irrelevant to clients' contraceptive use. However, the other factors were important. "Under circumstances where clients expect, and staff employ, authorita- tive guidances in helping the clients to select a contraceptive method, mean levels of contraceptive use are substantially increased" (Nathanson and Becker, 1984:1~. Although in earlier analyses important correlates of contraceptive use were clinic size, time spent with clients, and age of clients, these variables were no longer significant once the four dimensions of client-staff interaction were included. Clinic hours also had no net impact on contraceptive use. These re- sults are consistent with studies of compliance with medical regimens which uniformly report authoritativeness to be associated with higher levels of compliance. Nathanson concludes that the approach taken by most clinics, which place a heavy emphasis on independent client decision-making, is less conducive to effective contraceptive use than a more authoritative ~medical" approach.

228 The effect of scope of interaction differs depending on whether the preference is that of nurse or client. Clients' belief that the scope should be broad, including learning about clients' personal problems, is associated with greater contraceptive use in the clinic; however, contraceptive use is lower in clinics in which nurses want to talk about the clients' personal problems. The authors interpret this as suggesting that medical personnel undermine their authority by becoming too friendly. Finally, both indicators of client trust--in confidentiality of the visit and in the advice given--were found to be associated with lower mean clinic contraceptive use. The authors suggest that, consistent with findings from other studies, healthy skepticism may be related to better compliance with a medical regime. A third study (Shea et al., 1984) explored the relationship between clinic use and contraceptive use among adolescents. Adolescents making their first visits to one of nine federally funded family planning clinics in the Philadelphia area were interviewed during their initial clinic visit and followed-up by telephone six and 15 months later. Data on contraceptive behavior and clinic attendance were obtained from 359 sexually active adolescents age 13 tolls. Consistent contracep- tors were those who had used a reliable method of contraception (i.e., pill, IUD, diaphragm, foam, and condoms) during all periods of sexual activity, with the exception of the month in which the clinic visit occurred. Inconsistent contraceptors included those adolescents who used a reliable method most of the time, but not continously, as well as those adolescents who primarily relied on the least effective methods or no method at all. There is a close relationship between contraceptive behavior and clinic use. Almost three quarters of the adolescents who did not return in the first three and six months were inconsistent contraceptive users. However, adolescents who made the most revisits in the first two months were also likely to be inconsis- tent contraceptive users. Almost three-quarters of teens who made 3 or more visits to the clinic in the first three months were inconsis- tent users. It appears that those who return most are probably having problems with the method, and that is associated with inconsistent use. A revisit within the second or third month is important since most adolescents received an initial 3 month supply of oral contracep- tives, and, therefore, would need to renew the prescription or pick up another three month supply. The researchers also looked at factors related to clinic use, as measured by 1) the probability of making a revisit, 2) the total number of clinic visits, 3) the timing of the first months. Of all these variables, the only consistently significant factor was the adolescent's satisfaction with her contraceptive method. Organized family planning programs have been in existence for quite a while and a considerable amount of research and evaluation has been conducted on them. Several specialized programs specifically designed

229 to serve teens have developed recently as offshoots of hospital or clinic based family planning programs. These are programs based either in schools or in multi-purpose youth centers. They differ from clinics in that they offer more services and are directed at youth populations. Family Planning Clinic Based Programs for Males Fewer than half of one percent of all family planning clinic patients in the United States are male (Dryfoos, 1985~. Reasons for the low rate of male utilizaton include lack of funding, negative staff attitudes toward males, lack of male staff or difficulty integrating male staff into the program, and the general perception of clinics as "woman oriented," which result in difficulty recruiting males directly (Dryfoos, 1985~. However, males often attend clinics with their part- ners, attend counseling sessions and receive educational publications. The fact that males who do so attend are probably very different from those who do not make evaluation of the impact of the programs directed twoard males in such settings difficult. Programs that reach males in locations such as community centers, recreation areas, schools and even the street have used films, workshops and the media to attempt to imp prove their awareness of and use of contraception. No evaluation of the success of such efforts has been made (Dryfoos, 1985~. Condom Distribution Program for Males While the distribution of condoms may be part of family planning services offered in clinics, several innovative programs have aimed specifically at distributing condoms and encouraging their use among teenage males (Dryfoos, 1985~. In one early program (Arnold, 1973), small neighborhood stores and shops distributed free condoms to adoles- cent males. An increase in condom use over the 13 week period of the study and in the six months following the study was reported (Arnold 1972~. Contact one year after the study found that 69 to 81 percent of the respondents had used a condom at last intercourse. Other pro- grams have developed similar condom distribution activities; however, few have been rigorously evaluated. Dryfoos (1985) recommends that further review and evaluation of such programs be undertaken in light of early indications of success of such programs, the clear advantages of condom use for reducing pregnancy and preventing the spread of disease, and the low cost and risk of the condom. Parenthood Programs for Males Dryfoos (1985) reviewed several small scale intervention studies aimed at increasing the involvement of older fathers with their chil- dren. Parke and Neville (in this volume) suggests both that many adolescent unwed fathers are interested in their infants and children and that involvement by the father can yield positive effects for them.

230 Recently interest has grown in developing programs targeted to adoles- cent fathers (see, for example, Dryfoos, 1985~. There are at least three concerns that need to be addressed before major commitments to this approach are made. First, rigorous evaluation of such programs must be undertaken. Unfortunately, evaluating such programs is made very difficult by the fact that fathers who participate are a select groups of fathers; they have stayed in touch with the mother and child. The programs are likely to be successful for that reason alone. A second and serious concern is that involvement of the father may not always be to the best interest of the mother and the child. This would be the case if the father were abusive and non-supportive of the mother. In addition, contact with the natural father may prevent the mother from making contacts with another male who might serve as a more supportive father to her children. While it is true that parenting education would be useful for all males to have, maintaining ties (other than economic ones) with their natural children and their mothers may not always be helpful for all parties involved. The model of the importance of the father has been developed from the married couple family. Research suggests that other supportive family members such as the grandmother of the child or other relatives may serve equally well as substitutes for an absent father (Kellam et al., 1979) and contribute to the well-being of the child. Third, and finally, what the content of such programs should be has not been established (see the section on approaches, below). Where there is a clear commitment to actively participate in the rearing of a child of an adolescent mother, this participation could be made more effective (Dryfoos, 1985) by making sure that fathers learn and practice basic caretaking skills and by establishing regular supervised contact with their infants. Various types of interventions during the postpartum period are reviewed in Dryfoos, 1985, and Parke, in this volume). Most of these interventions were conducted on older fathers, not adolescents. So the extent to which they would apply to younger men is not known. A major program which attempts to intervene with adolescent fathers is the Teen Father Collaboration Project, dis- cussed in the section on economic approaches below. School-Based Programs for Young Men and Women Although teenagers will attend adult clinics, several studies have found common features of clinics that attract and keep teen clients. In particular, availability, confidentiality, affordability and loca- tion apear to most influence attendance (URSA, 1976; Coughlin, 1978; Zabin and Clark, 1983; Kisker, 1984~. Location on site in a school would appear to satisfy many of these criteria. Clinic and contracep- tive continuation are facilitated by ease of return visit. Follow.up is facilitated by provider accessibility to school schedules. Con- fidentiality is increased by providing a number of non-family planning services such as athletic, job and college physicals, immunizations and a weight control program. Males can be attracted and served as well as

231 females. Finally, the services are generally provided free of charge to registered students. The oldest and most successful ongoing project is the St. Paul Maternal and Infant Care (TIC) Project, which first opened a comprehen- sive heath care clinic in a local junior/senior high school in 1973. Since then it has expanded to four senior high schools and as of 1983- 84 served some 70 percent of the student population. About one-third of the students served use the clinic for family planning. Services provided include educational counseling and family planning services, prenatal and post-partum care, nutrition education, day care, and parenting, family life and sexuality education. Data from the St. Paul project show that the fertility rate in the schools with clinics dropped substantially over the period: from 59 per 1,000 in the 1976-77 school year to a low of 21 per 1,000 in 1979- 80. The rates increased in 1980 to 39 per 1,000, due to an increase in the refugee pouplation in St. Paul. Rates were down again, to 26 per 1,000, in 1983-84. These figures compare favorably with national birth statistics, which showed a birth rate of 45 per 1,000 for whites in 1977 and in 1982. No information is available from the St. Paul Project on the trend in pregnancies and abortions, so we don't know how much of the decline is due to a decline in pregnancies and how much to an increase in abortions. Doug Kirby (1984) is just beginning a project which will evaluate the effectiveness of comprehensive school based programs in preventing adolescent pregnancy in 9 to 11 sites around the country. Potential project sites, listed in Table 1, meet the following criteria for being included in the study: they are multi-service, provide family plan- ning, are located on the school campus, have or will have a good working program, have a willingness and enthusiasm to participate, have the ability to collect good data, and reflect variation in fea- tures. Three are just opening so in these there will be an opportunity to have a true pretest-posttest design. The evaluation designs vary from project to project, but the major strategy is to administer ques- tionnaires in both program and matched non-program schools at two points in time. The study will not be longitudinal (except for a very small case-study sample). Rather it will be based on a series of cross sectional surveys of all students in the study and comparison schools. Since no names will have to be recorded for follow-up, confidentiality can be assured. Comprehensive Non-School Based Prevention There are several programs that provide family planning and other services to male and female teenagers, but that are not located in school settings. These are primarily located in youth centers. How- ever, they can be included as pregnancy prevention programs because of a strong family planning component. Two familiar names include The

232 Door (New York City) and The Bridge (Boston). A third, recent addi- tion, is the "Self.Center, n located in Baltimore, Maryland. The latter falls somewhat between school and non-school based programs. The cen- ter is actually located across the street from the school; however, clinic personnel do work in the school and refer individuals to the freestanding clinic. A fourth, the West Dallas Youth Clinic, is also located adjacent to a high school. The Door is a comprehensive multi-service center for youth 12 to 21 located in Manhattan. It provides free medical and gynecological services, family planning and sex counseling services, nutrition coun- seling, psychiatric counseling and therapy and social services. It also provides crisis intervention services, education, vocational and legal counseling services. It has a learning center and a gymnasium with locker rooms and showers. It offers creative workshops in arts, crafts, poetry, music, dance and theater (from Philliber et al., 1983~. No systematic evaluation has been conducted of the family planning ser- vices component of the Door, although one is planned (Nowlan, private communication, 1985; see also Fink, Kosecoff and Roth, 1983~. The Bridge is a multi-service center for youth located in Boston, in the heart of the Boston tenderloin district. Services provided in- clude counseling, medical care, family planning, STD, employment coun- seling and placement and, for teen mothers, child care and parenting training. Particular attention is paid to assisting youngsters to attain high school equivalency degrees (Dryfoos, October 1983~. In conjunction with the Johns Hopkins University Adolescent Preg- nancy prevention program, a cooperative arrangement was made among a junior high school, a senior high school and a nearby free-standing clinic (the Self-Center) (Zabin et al., 19861. Nurse practitioners and social workers were stationed in the schools during the day to provide sex and family life education and counseling and referral for the students. They also worked at the clinic to which the students were referred for birth control. An evaluation of this program conducted by Zabin et al. (1986) demonstrated substantial program impacts. Over the course of the 2-1/2 years that the program was in operation, substantial increases in sexual and contraceptive knowledge were shown among teenagers in program schools, compared to teenagers in non-program schools. The authors attribute this to the increased accessibility of the staff and clinic. Some delay in the initiation of first intercourse occurred in program compared with non-program schools, delay substantial enough to lay to rest fears that access to contraceptive services in schools would increase levels of sexual activity, and to suggest that such pro- grams may, in fact, delay first intercourse. One of the most important findings was that students in program schools attended clinics sooner after initiating sexual activity than prior to the program and in come parison with non-program schools. Junior high school students used the clinics at levels comparable to those of older teenagers, and, sur

233 prisingly, junior high teenage males used the clinics as frequently as girls of the same age. In contrast, senior high boys were much less frequent clinic users than seior high girls. Thus suggests that junior high boys are more recepitve than seenior high boys to such programs; getting them early may improve later male contraceptive practice. There was also evidence of improved contraceptive practice among both males and females in program compared to non-program schools. Finally, each measure used showed a reduction in pregnancy rates among older teenagers and a delay in the rapid increase in pregnancy rates or a decline in pregnancy rates among younger teenagers in program compared to non-program schools. The study suggests greater effects for younger than older teenagers, and suggests that the program works by encour- aging younger teenagers to develop patterns of knowledge and behavior usually associated with older adolescents, coupled with delaying first intercourse by about 7 months. The West Dallas Youth Clinic (WDYC) of the Children and Youth proj- ect of the University of Texas Health Service Center at Dallas is located in a building adjacent to the area's only high school. Family planning and other medical services are provided to male and female adolescents 5 days a week from 9 a.m. to 4 p.m. as part of a compre- hensive adolescent health care service. An evaluation was recently conducted to determine the effect of the program on the teen birth rate (Ralph and Edgington, 1983~. An area of Dallas similar in characteris- tics to West Dallas was selected and birth rates were compared for the period 1971-74, when the program began, and for the period 1975-78. A second comparison was made between birthrates of WDYC registrants and non-WDYC registrants in 1977. Results show the birth rate to have de- creased faster between 1975 and 1978 in West Dallas than in the matched area. In addition, the birth rates in 1977 were lower among WDYC par- ticipants than among non-participants. The authors concluded that the clinic program did reduce teen births in West Dallas. Parental Notification In February 1982 the Department of Health and Human Services pro- posed to require family planning projects funded by Title X to notify both parents or the legal guardian of patients under the age of 18 within 10 days after the adolescent receives prescription drugs or devices from a clinic (Kenney et al., 1982~. Drugs for the treatment of veneral disease were specifically exempted. The department also proposed to eliminate current rules that required eligibility for ser- vices to be determined based on the minor's own income. In that year ( 1982 ~ there was only one state that had a law already on the books requiring parental notif ication before contraceptives could be prod vided to anyone under age 18. Thus the passage of such a regulation would have indeed dramatically changed the way family planning services are prov ided in the Uni ted State s. Even thoug h non-T itle X f unded ser- vices could still be provided conf identially it was expected that other programs would probably also follow suit. The question that was hotly

234 debated was the impact of parental notification on adolescents and their families and on birthrates. After a lengthy debate, the regula- tion went to the courts, where it was stopped and never implemented. However, the debate is not yet settled. Four issues will be discussed here. First, what are current practices among clinics and other pro- viders regarding the provision of contraception to teens under 18. Second, what proportion of parents already know about their teens' attendance at clinics, and what is the relationship between notifica- tion and family involvement? Third, what effect does communication with parents about contraception or sexual activity have on children's behavior, and under what circumstances? Fourth, what effect does fear of parental knowledge about their sexual behavior have on children's use of contraception and attendance at clinics? What effect would a parental notification rule have on attendance at family planning clinics and on contraceptive use? Provider Policies In 1978 AGI conducted a survey of family planning agencies, hos- pitals and freestanding clinics to find out about current clinic policies and practices regarding parental consent or notification for the provision of abortion and medical contraceptive services to teen- agers (Torres et al., 1980~. Twenty percent of family planning agen- cies do not provide medical contraceptive services to patients aged 15 or younger without parental consent or notification; 10 percent have such requirements for all patients under age 18. About 25 percent of hospitals, but only 3 percent of Planned Parenthood affiliates report such restrictions for those under 15; the comparable figures are 19 and 1 percent respectively for those under 18. Of those with restrictive policies, about half require patients to bring a parent or parent's written permission to the clinic with them. The majority will waive these requirements under certain circumstances other than legal emanci- pation or court order. AGI also conducted a survey in 37 counties of all physicians in practice of general, obstetric-gynecologic and pediatric medicine of their policies regarding age, parental consent and prescription of contraceptive methods to adolescents in 1981 (Orr, 1984~. They found that although 86 percent are willing to prescribe contraceptive methods to adolescent women, only 59 percent are willing to serve unmarried minors without parental consent. Of the types of physicians, ob. gyns are more likely and pediatricians least likely to prescribe contracep- tives. Their policies are related to state laws; as expected, physi- cians are more liberal in states with liberal policies. Thus it can be said that parental notif ication or consent is curently required by a small minority of institutions before contraception can be provided to unmarried minors, but by 2 out of 5 physicians.

235 Parental Knowledge about Teen Sexual/Contraceptive Behavior AGI also surveyed over 1,200 family planning patients unmarried and under age 18 in 1978 (Torres et al., 1980~. Of those, 54 percent say that their parents know about clinic attendance; in the majority of cases (30 percent) they told their parents voluntarily. In 21 percent the parents suggested the visit. In only 3 percent did the parents find out from others, including from the clinic. The younger the patient, the more likely the parents are to know. What is the relationship between parental notification and family or parental involvement in family planning programs? Just because clinics do not notify parents does not mean that they discourage or prevent family involvement. A study of Title XX grantees conducted by the Family Planning Council of Southeast Pennsylvania in 1981 (Fursten- berg et al., 1982) found that 85 percent of these programs involve par- ents directly through counseling, advisory groups, discussion groups, and training workshops for parents. The analysts found that agencies that encourage parental notification (e.g., Planned Parenthood pro- grams) are significantly more likely to operate multiple services for parents (57 percent) than either those agencies that require notifi- cation (e.g., hospitals) or those that take no position (e.g., health departments) (44 percent). The authors conclude that agencies that encourage notification involve parents more than those that require it or those neither requiring nor encouraging it; agencies that mandate it may consider it a substitute for other activities toward that goal or it may simply be a bureaucratic requirement with no relationship to the goal of improving family communication (Furstenberg et al., 19821. Family Communication and Teen Sex and Contraception The last issue is that of the relationship between family communi- cation and teen sex and contraceptive use. There is no consistent research evidence that daughters' communication with parents leads to postponing sexual initiation, net of other factors. It has been hypo- thesized that that increased communication about sexual issues often follows debut. Nor is there much evidence that daughters' communica- tion with parents leads to better contraceptive use; again, contracep- tive use may lead to increased communication about sexual and contra- ceptive issues (Fox, 1980; Furstenberg et al., 1984; Newcomer, 1983; Newcomer and Udry, 1983) (see also Chapters 1 and 2~. Presumably it is not communication per se that matters, but the closeness of the rela- tionship and the content of the communication (see, for example, Fox, 1980; Newcomer, 1983~. A recent study {Kahn et al., 1984) found no association between several measures of parent-child communication and sexual activity of daughters, but an association for sons. For sons, greater communication with mothers was associated with less sexual activity. In contrast, greater communication with fathers was associ- ated with greater sexual activity. Mothers and fathers may communicate different messages to their sons. Daughters rarely communicated with their fathers on sexual matters (Kahn et al., 1984~.

236 Impact of Notification on Teen Behavior In one study (as reported by Kenney et al., 1980), respondents were asked a hypothetical question about their behavior if parental notifi- cation were mandated. Of the 41 percent whose parents didn't know, 56 percent said they would not attend if parental notification were re- quired by the clinic. Of these, 65 percent would use a non-prescrip- tion contraceptive, 17 percent would use no method, 9 percent would not have sex, and 9 percent didn't know what they would do. AGI estimated in 1982 that an additional 33,000 adolescents age 17 or younger would become pregnant if the proposed regulation were adopted (Kenney et al., 1980~. Of course, what the actual behavior of adolescents would be is unknown. In conclusion, parental involvement in and knowledge about chil- dren's contraceptive behavior may be a worthy goal; however, at the present time the evidence that it will reduce the sexual activity of their adolescent children or improve their contraceptive behavior is weak or non-existent. Certainly there is no evidence that communica- tion would be furthered through requiring parental notification for family planning services. In fact, there is some evidence that organi- zations that are not required to notify parents, but that encourage parental involvement are more successful in involving parents. The relationship between parental involvement in family planning agency activities and clinic use by their own children is unknown. The evi- dence suggests that although most parents already know about their teens' attendance at clinics, that there is a group of teens who would not attend clinics if parents had to be notified and that will continue to have sex but will use less effective or no method of contraception. Only a small proportion of sexually active teens (two percent) say they would not have sex if parents were notified of clinic attendance. On the other hand there is some evidence that by the time they have reached adolescence, parent-child interaction patterns are already well-established. It may be too late to change them radically. Re- search in sex education, in parent-child communication programs and other areas suggests that improving communication prior to adolescence may be a more successful strategy (see earlier discussion). Summary Family planning programs have been evaluated and shown to be very successful in improving contraceptive use, and therefore preventing pregnancies and births to teen women. Several recent types of pro- grams, schoos based and non-school based clinic programs that are di- rected toward male and female teenagers, have shown remarkable succes- ses. Evaluations are still being conducted, but such efforts appear to be consistently successful in reducing births to young women and keeping them in school. Their success among young men has not general- ly been evaluated.

237 A large proportion of family planning programs do attempt to in- volve parents in their activities. The effect of a mandated program to notify parents or require their consent for minors to receive family planning services is unknown, although a number of studies suggest that it would reduce contraceptive use while not changing patterns of sexual activity among teens. PREGNANCY AND PREGNANCY RESOLUTION Several types of programs are designed specifically for pregnant teenage women. These include nutritional programs such as WIC, Maternal/Child Health programs with emphasis on prenatal care, and public school programs for pregnant adolescents. Finally, there are several programs that are directed at resolving a pregnancy: abortion and adoption. The programs discussed in this section end at birth or shortly thereafter. There is overlap with programs discussed in the following section, but in general I have tried to distinguish those that end at birth or shortly thereafter from those that continue for some time following the birth. Abortion Services The U.S. abortion rate remained essentially stable in 1981-84, after rising each year between 1973 and 1980, and the same was true for the rate for teenagers. In spite of the apparent widespread use of abortion, especially by teenagers, about 40 percent of all teenage pregnancies ended in abortion, compared to 26 percent of pregnancies for women of all reproductive ages, there are apparently still wide gaps in the geographic availability of abortion services. The 2,900 providers in 1982 were located in only 22 percent of all U.S. counties (Henshaw et al., 1984~. Thus 78 percent of all U.S. counties had no identified provider of abortion services in 1982. Abortion providers are concentrated in urban areas. Only 2 percent of abortions were per- formed in nonmetropolitan counties although 26 percent of women of re- productive age live in such counties. Abortion services are most available in states on the East and West coasts. In 1982, 82 percent of abortions were performed in nonhospital facilities: 56 percent in clinics which specialize in abortion services, 21 percent in other clinics and 5 percent in physicians' offices. In 1983, women paid an average of S200 for a nonhospital abortion. Other factors affect access to abortion services: for example, only a few states and the District of Columbia provide public funding through Medicaid for eligible women, and only 78 percent of health maintenance organizations cover abortion. Although the great majority of women (91 percent) in 1980 who terminate their pregnancies do so within 12 weeks, teenagers are more likely to delay the decision to abort. Only 32 percent of facilities provide abortion after 12 weeks gestation. Finally, some hospitals limit the circumstances of abortion to medical reasons.

238 Effects of Abortion Availability and Use The evidence consistently shows that in the U.S. higher abortion rates in an area, whether a state or county, are associated with lower fertility (Forrest et al., 1980; Field, 1981; Brann, 1979; Moore and Caldwell, 1977~. This is the case whether abortion availability is measured by liberality of state abortion laws, by abortion rate, or by abortion ratio. Abortion is also associated with total as well as out- of-wedlock fertility. The only study that does not find a negative association (Moore, 1980) finds no association. The relationship be- tween Medicaid policies on abortion in the state and teen fertility was also found to be only weak or non-existent. This is not surprising, since it appears that the majority of women denied publicly funded abortions are able to obtain them anyway (Trussell et al., 1980~. Abortion didn't start in 1973, when it first became legal nation- wide in the United States. Although greater availability of abortion services does not cause abortion, which has been around for a long time; it has resulted in its increased use. Between 1973, when legal- ized in the United States, and 1982, the most recent year for which data are available, the estimated proportion of teen pregnancies ter- minated by induced abortion has increased by almost two-thirds. Recently there has been a leveling off of the availability of abortion and greater stability in rates and ratios as well. Abortion rates reflect underlying pregnancy rates. The higher the pregnancy rates, the higher the abortion rates. Thus, for example, comparing blacks and whites, the abortion rate for blacks is higher than that for whites, reflecting the higher pregnancy rate among blacks than whites. In the international study conducted by AGI (Jones et al., 1985), abortion data were available in 11 to 13 countries. In these countries a higher abortion rate was associated with a higher birth rate, reflecting, presumably, a higher pregnancy rate, and a somewhat stable ratio of abortions to pregnancies. In the U.S., the rate of abortions is higher for blacks than for whites, reflecting a greater pregnancy rate, but the ratio of abortions to pregnancies is somewhat lower among black teens than among white teens, while it is higher among black older women than among white older women. This re- flects the differential use of abortion among blacks and whites. Black women use it more to terminate childbearing, while whites use it to postpone childbearing. It is possible to have low fertility rates and low abortion rates as well, as in most of the Western European nations (Jones et al., 19851. It has been argued that the availability of abortion may make women more careless contraceptors, since they know they can always resort to an abortion. However, research by Moore and Caldwell (1977) using the Kantner-Zeloik data from 1971 with added state level vari- ables, found no evidence for an impact of abortion availability on the probability of pregnancy. Nor did they find any evidence of abortion availability on the transition to sexual activity. They did, however,

239 find that abortion availability increases the chance of a premaritally pregnant teen having an abortion, and decreases her chances of bearing an out-of-wedlock child. Thus this analysis suggests that the availa- bility of abortion affects the probability that an unmarried woman will chose to have an abortion rather than having an out-of-wedlock birth or marrying once a woman is pregnant. It does not appear to affect sexual activity or the probability of becoming pregnant. In fact, a recent study suggests that in 1979 the probability that a young women who aborted a previous first pregnancy will become premaritally pregnant again within 24 months was substantially lower than that of a compar- able woman who carried th.e first pregnancy to term (Koenig and Zelnik, 1982~. These data suggest better contraceptive practice among those who had a prior abortion than those who didn't. There is a paucity of data and analysis of this important question, however. Parental Involvement in Teenage Decision Making Re Unwanted Pregnancy Resolution In spite of the fact that legalization of abortion in 1973 meant that a pregnant minor could get an abortion without telling her par- ents, a study (Rosen, 1980) based on data collected in 1974-75 sug- gested that teenagers do involve parents in pregnancy resolution decision-making. Few adolescents consulted their parents when they first thought they might be pregnant, but more than half involved their parents in decision-making to resolve the pregnancy. The mother's influence was strongest for those whites who aborted and those who gave birth but gave the baby up for adoption and was strong for all black women. Maternal influence was least for white girls who kept the baby. Many of these young women were not living at home, and some lived with a male partner. Adoption Services There is very little information available on providers of adop- tion services, either the nature of the institutions or their develop- ment. Apparently the origin of adoption services is attributed to religious groups of the late 19th century, which were supplanted by by social welfare organizations in the early 20th century (Muraskin, 1983~. Most care was apparently provided by voluntary non-profit organizations that both cared for pregnant young women and screened/ evaluated prospective adoptive parents. After the mid 1950s, however, the relative importance of adoption oriented residential and non-resi- dential facilities declined and care alternatives for young women pregnant out of wedlock and who will keep and raise their children have increased (Muraskin, 1983~. One example of a private adoption agency is The Children's Home Society of Minnesota which operates a residential program for approxi

240 mately 300 pregnant adolescents per year. Of these about 10 percent release their babies for adoption. According to the data presented by Jane Bose, there were 6,107 births to teens in 1982 in Minnesota and only 45 newborn adoptions in Minnesota in 1983 and 60 in 1984. It is clear from these data and the stories of researchers attempting to study adoption as an alternative resolution to an unplanned teen preg- nancy that it is infrequent. In recent years, some reversal of emphasis has occurred and the Office of Adolescent Pregnancy Programs (OAPP) has been charged with finding ways of making adoption a more attractive alternative for adolescents pregnant out of wedlock. According to Muraskin (1983) this is both a new role for a federal agency and a difficult one, coming at a time when few adolescent mothers choose to terminate parental rights. In line with its new mandate, the Office of Adolescent Pregnancy Programs has funded several programs whose purpose is to increase the proportion of pregnant teens who opt for adoption over abortion or childbearing. OAPP is also funding two ongoing studies (Kallen, 1984; Resnick, 1984) which will, in the future, provide some information on the factors affecting the adoption decision. The results of previous studies of this process are discussed in Chapter 4. Title IX of the Education Amendments of 1972 One regulation which has had a substantial impact on teens is Title IX of the Education Amendments of 1972. This law prohibits dis- crimination in education against teens because of their pregnancy/ childbearing/marital status. Teens cannot be expelled from school or barred from any program, course or extracurricular activity because of pregnancy, parenthood, or marriage. Schools can institute special programs, but they must be voluntary and comparable to regular pro- grams (Zellman, 1982~. Of course, this rule only applies to schools receiving federal funds; private schools are not covered and may still bar pregnant teens from classes. Although no one has attempted to test an association, recent data do show that in the late 1970s and early 1980s teens who bore chil- dren were much less likely to leave school than they were in the late 1960s and early 1970s (Most and Maxwell, 1981~. This suggests that more liberal policies regarding school attendance (and the establish- ment of special school programs) may have made some difference. On the other hand, liberal policies reflect the more liberal climate regarding out-of-wedlock childbearing in general. However, the Education Amend- ments of 1972 were an early and a landmark step in equalizing oppor- tunity between the two sexes, a step whose consequences are only now beginning to be recognized.

241 Public School Programs for Pregnant Teenage Women In her review of 12 school sponsored programs in 11 school dis- tricts in 7 states, Zellman identified three types of programs schools have developed in response to the problem of teen pregnancy. The first type is all-inclusive and consists of an intensive curriculum. Preg- nant teens are separated from regular classes and offered an educa- tional curriculum supplemented by classes in parenting and child devel- opment. Other services such as health monitoring and child care may be offered. However, the common features are that students enrolled in such a program do not attend regular classes, are physically sepa- rated from regular school, and must return to regular classes within a few weeks after birth. The second type consists of a supplementary curriculum. In this type students remain in regular classes most of the day but can take relevant for-credit courses in parenting and child development as well as obtain special services such as child care and counseling. Young mothers can remain in such programs or use their services after the immediate postnatal period. The third type consists of noncurricular programs in which students may get counseling, medical care and referral but obtain no credit for participating. All three types of programs are administratively separate from the regular school. Zellman found advantages and disadvantages to each type of program. The advantage of the intensive curriculum is its separate supportive environment and attention to parenting skills. Its disadvantage is a relatively weak academic curriculum and limitation to the period of pregnancy. Most teens need support as much or more after as before the birth; they may have trouble adapting to regular classes. The advan- tage of the noninclusive programs is that services continue after birth; students may remain in regular school classes. The disadvantage for some students may be the embarrassment of staying in regular classes. All school programs have problems in detecting potential en- rollees since there is no mechanisms for detecting pregnancy and coun- seling a girl on her options. A girl must announce her pregnancy, after which she will be referred to a program. Thus many fall through the cracks. Schools are even less helpful to teen parents. Provision of services during pregnancy is the limit to which most schools are willing to go. Because of the passive attitudes of school officials, and their expectations that the programs will be very expensive, the initiation of a special program often depends on the persistence of a single individual. Finally, Zellman concluded that special programs are uneven in quality and that each type meets some needs of some teens but not all those of all teens. WIC The Special Supplemental Food Program for Women, Infants and Chil- dren (WIC) provides supplemental foods and nutrition education to high risk pregnant and lactating women, infants, and children up to 5 years

242 of age. Participants receive vouchers for food such as milk, cheese, fruit juices, eggs, dry beans, peanut butter and iron fortified cereals and infant formula, a value of about 630 per month (Kotelchuck et al., 19841. An evaluation of the prenatal part of the program was conducted in 1978 (Kotelchuck, 1984~. Experimental and control groups were matched on age, race, parity, years of education, and marital status. Results show improved pregnancy outcomes, including decrease in low birthweight incidence, an increase in gestational age, and a reduction in inadequate prenatal care. The impacts of WIC on teens and on un- married women were stronger than those on other subgroups. Improved Pregnancy Outcome (IPO) Projects An evaluation of the effect of the North Carolina Improved Preg- nancy Outcome (IPO) Project on the use of prenatal care and the inci- dence of low birthweight on black registrants was conducted by the University of North Carolina. The project site was an underserved poor rural two.county area. ~ subsample of 297 black teens was selected for special attention. The intervention consisted of introducing nurse- midwives to provide prenatal and post-partum care, the expansion of health department services to include nutrition counseling, social services and health education, coordination with a statewide perinatal care program, and outreach and transportation. Data were taken from records of vital events maintained by the North Carolina State Center for Health Statistics, matched with IPO registrant data. Black women in two geographically proximal counties served as controls. The proj- ect appeared to greatly increase the proportion of teens who received adequate prenatal care. There was no effect, however, on the birth weight of the infants of teen mothers. Three explanations for the lack of effect include 1) preexisting differences between experimental and control groups, 2) lack of information on other health conditions, and 3) low intensity of the intervention. Summary Both abortion and adoption prevent adolescent females from be- coming parents. Adoption is used relatively infrequently. The effectiveness of programs to increase its use is as yet unknown. Abortion is widely used, particularly by teenagers. Improved contra- ceptive use would reduce the reliance of many on abortion to avoid childbearing. A nutritional program (WIC) for pregnant women was very successful in improving pregnancy outcomes for teens. !

243 PARENTHOOD In this section are discussed comprehensive services to pregnant and parenting teenage females and other services not directed specifi- ca~ly at teens but which affect them, such as AFDC. A small number of programs that focus on teenage males are also discussed. Comprehensive Services to Pregnant and Parenting Teens Family planning programs are directed at the prevention of concep- tion, at facilitating the planning of births. Although data on preg- nancy, the most appropriate outcome var table to measure are not avail- able because of problems with abort ion statistics, still it appears as though they are relatively effective in achieving their goals. Special programs for adolescents, in contrast, usually focus on preg- nant adolescents or teen parents. Thus their goals are somewhat dif- ferent. In general, their goals are to improve outcomes for both mother and child. In particular, by providing prenatal care they aim to improve health of the neonate and the mother around delivery, to improve the young mother's socioeconomic circumstances by facilitating the completion of her schooling, to increase her independence by pre- paring her for employment, and, finally, to reduce the probability of rapid repeat childbearing. "Comprehensive programs for the already pregnant schoolgirl comprise short-term intervention with short-term goals: a healthy mother and baby, continued education, and the solu- tion of immediate social problems that may complicate the life of the mother" (Klerman and Jekel, 1973:10~. In the following pages I describe the results of eight evaluations, which cover some 51 different project sites across the United States. I will highlight important aspects of the programs, summarize the re- sults, and critique the evaluations. Klerman and Jekel's Evaluation of the Young Mother's Program The goals of the Young Mothers Program (YMP), evaluated by Klerman and Jekel (1973) in 1967-69 were the following: 1. short term: a healthy pregnancy, an uncomplicated delivery, a healthy infant and return to school postpartum. 2. Long range (two years postpartum): completion of high school or its equivalent, deferral of subsequent pregnancy, evidence of em- ployability, and progress toward economic independence (Klerman and Jekel, 1971:31~. The major focus of this evaluation was the Young Mothers Program (YMP), a program for never marrried, pregnant girls under 18 in New Haven, Connecticut. A second program was initially used as a control

244 group, but wound up being a second exper imental g roup . This was an Interagency Services Program (IAS) for unmarried pregnant girls enrol- led in grades 9.12 of a Hartford public school. The control group was a g roup of young women age 17 or under, unmar r led, res ident of New Haven who delivered a baby at the New Haven Hospital in 1963-65 after having been pregnant at least 20 weeks. YMP participants and the comparison group of mothers differed only in the incidence of toxemia, which was higher for the latter. There were, however, substantial differences in infant health. Those born to mothers in the comprehensive program were significantly healthier dur- ing the perinatal period than infants of similar mothers in the com- parison group. However, subsequent infants born to mothers who par- ticipated in the medically oriented comprehensive program (YMP) had very high risk of poor outcome. There was no significant difference in repeat pregnancy. Although subsequent pregnancy was not quite as rapid among experimental as con- trols in the first 15 months, by 36 months there were no significant differences in repeat pregnancy. Contraceptive use was a poor pre- dictor of subsequent pregnancy--being in school at 3 months post- partum was a better predictor of delaying subsequent pregnancies than was acceptance of contraceptives. Other program effects. Participants in the YMP program were more likely to stay in school during pregnancy, to return to school after delivery and to graduate from high school than controls. There was no difference among experimental and control groups in employment; how- ever, the follow-up period differed for experimental and control groups 2 years versus 6 years after delivery. There were few differences between the two program groups in "success"--still in school or completed school and no pregnancy--by 26 months postpartum; about 35 percent of each group fell in this cate- gory. About 31 percent of the YMP group and 24 percent of the IAS group both dropped out and had a pregnancy. Conclusions. The experimental programs had important short term impacts in terms of mothers' and infant health. Repeat pregnancies were also delayed temporarily. The major long term impact was an in- crease in the proportion who completed school, an important conse- quence. Program effects on childbearing were apparently temporary. The subsequent childbearing of the experimental group caught up to that of controls after 30 months; in addition, subsequent pregnancies to the experimentals were at very high risk of poor outcomes. No control was available on which to compare employment experience.

245 AIR Evaluation of Project Redirection Project Redirection, a demonstration program of services for low- income teenage mothers and pregnant teenagers began enrolling partici- pants in mid 1980. The purpose of the demonstration was to Assess the feasibility and impacts of a comprehensive service program that attempts to 'redirect' the lives of young women from low-income back- grounds, those most at risk of welfare dependency because of their early parenthood" (Polit et al., 1983:2~. The specific goals of Project Redirection were the following: continued schooling, the development of marketable skills, acceptance and use of needed health care and social services, and planning for eventual employment and self-sufficiency. Specific objectives included completion of a school or GED program, delay of subsequent pregnancy, attainment of job skills, and improved maternal and infant health. Program eligibility was restricted to teenagers 1) under age 18, 2) pregnant or a mother, 3) without a high school diploma or GED, and 4) receiving welfare or living in a welfare-dependent family. The four sites included in the evaluation were Boston, Harlem, Phoenix, and Riverside, California. Four hundred teens are included, about 200 in experimental and 200 in control groups. A second sample of 175 each was collected in 1982 but not included in the analysis so far. Matched comparison sites were Hartford, Bedford-Stuy~esant, San Antonio, and Fresno, California. The similarities across sites were 1) the comprehensive mix of services including educational counseling, employability and employ- ment counseling, personal counseling and referral; referral to health care services; and parenting and life management education, 2) A com- munity woman assigned to each teen to act as support and assistance as well as mediator between program and girl, 3) An Individual Partici- pation Plan (IPP) which specifies individual goals to work toward along the route to self-sufficiency, and 4) Linkage to WIN. A second characteristic of this project was that an ethnographic study (Levy and Grinker, 1983) was conducted simultaneously. Inter- views were collected from 18 participants, who were followed exten- sively over the project period. This provides an interesting sup- plementary source of information and confirmation of the main findings from intensive study. The major conclusion from Project Redirection was that the program had a small impact on participants within the first year, but that most of the effects had disappeared by the second year follow-up. The proj- ect had its strongest impact on the most disadvantaged of the partici- pants: those who were not in school at baseline, those with no work experience at baseline, younger teenagers and Puerto Rican teenagers, and teenagers with a subsequent pregnancy. The major impacts occurred in the area of schooling and employment, which were, not surprisingly,

246 the areas of focus for the program. Among all participants, those in Redirection were more likely to have enrolled in school after delivery, and they spent a greater number of semesters enrolled. Length of par- ticipation in the program was associated with a greater number of semesters enrolled. Among those who were not in school at baseline, Project Redirection participants were more likely than non-participants to be in school or to have completed school at 24 months post-baseline. Project Redirection enrollment was also associated with a higher rate of school enrollment and completion for those with a subsequent preg nancy. Project Redirection participants were less likely than controls to be neither in school/completed school nor in the labor force 24 months post-baseline. They held a larger number of jobs, and a larger propor- tion had ever been employed at 24 months. Among teenagers in AFDC households, Project Redirection teens were more likely than control group teens to be currently employed. These differences were much stronger for disadvantaged teens high school dropouts, those with no work experience, younger teens and Puerto Rican teens and for those with a subsequent pregnancy. The outcomes were also stronger the longer the length of time in Project Redirection. The impact of Project Redirection on teen contraceptive use was small at 12 months; there was no impact at all at 24 months post-base- line. In contrast, the birth control knowledge scores of participants consistently exceeded those of non-participants. The rate of repeat pregnancy was slightly lower for Project Redirection Participants than non-participants at 12 months; by 24 months the difference between the groups had disappeared. There were no differences between Project participants and comparison teens in prenatal care, length of hospital stay or birth weight. Although the study was rigorously designed, with carefully matched control groups, it turned out that a substantial proportion of the control teens also obtained services from a variety of programs during pregnancy and postpartum. Thus the true comparison turned out to be between Project Redirection and a variety of other types of programs that provide services to teens. This is probably the reason for the larger impact on participants in areas unique to Project Redirection (education and employment) and the smaller impact on infant outcomes. Unfortunately, however, since this was not anticipated, the evaluation did not gather information to evaluate which particular aspects of the program or which services were most successful. In conclusion, this is a useful evaluation of the program in general, but it doesn't provide enough information on exactly which parts of the program or which services were most successful. One good example is that of child care services. The evaluators conclude that it was not very important to these girls; however, child care assis- tance was available only in 2 sites--Harlem and Phoenix (Polit et al., I983:45-48~--and, in fact, the only thing done appears to be the

247 "brokering of child care arrangements for those who need it" (Polit et al., 1983:45~. What was actually done is not clear since "Many of the young mothers brought their children with them to on.site program activities. The sites usually provided child care, often with the help of community women or other volunteers" (Branch et al., 1984:62~. Although over 90 percent indicated that their current arrangements did meet their needs, one-fifth expressed a desire to change their arrange- ments. It is not clear which young women received which services, making an evaluation of the program impact difficult. Urban Institute Evaluation of OAPP Programs The purpose of this study (Burt et al., 1984) was to evaluate how OAPP grantees implemented their programs and what impact program par- ticipation had on the lives of program clients. The goals of OAPP funded projects are to prevent unwanted initial and repeat pregnancies, to assist adolescents to obtain proper pre- natal care, and to assist pregnant adolescents and parents to become productive independent contributors to family and community life, with primary emphasis on services to adolescents who are 17 years of age and under and who are pregnant or parents. Short term objectives are to reduce the incidence of low birth weight, baby's complications, and mother's complications, and to increase school enrollment. Long term objectives are to reduce the incidence of repeat pregnancy, increase educational attainment, increase the number who obtain training and employment and reduce welfare dependency. In OAPP projects eligibility criteria varied from project to proj- etc. The majority of participants (64 percent) entered pregnant; a small proportion (14 percent) were mothers and a small proportion (11 percent) had never been pregnant. Analyses were conducted only on females, although males made up 10 percent of the clients. In keeping with the overt goal of delivering services to young teens, 9 percent were 14 or younger, 64 percent were 15 to 17, and 26 percent were 18 or older. About one third were receiving welfare or medicaid. This evaluation differs from others in that the programs themselves collected the data, which Urban Institute researchers subsequently analyzed. Urban Institute researchers developed the protocols so that they would be standardized across programs and trained and then subse- quently provided assistance to ensure the procedure was adequately understood and followed. Unfortunately, cooperation with the data gathering part was voluntary on the part of the programs; as a result, not all cooperated. This report is based on data from 20 of 26 gran- tees, encompassing 30 individual projects. The advantage of the evalu- ation is that data were collected in roughly uniform manner across projects, and therefore are comparable. The disadvantage is that the forms were very simple and a bare minimum of entry characteristic in- formation was collected. All projects collected aggregate data; only

248 a subset of 23 sites also provided individual client data. Thus in- dividual cases could only be followed and analyzed on a subset of projects. Two of the most important findings were the importance of child care and the length of time in the program. First, child care was a very important service. Child care reduced the probability of a repeat pregnancy by 12 months after delivery for participants pregnant at entry. For women who were mothers at entry into the program, it was also associated with greater schooling completed and being employed 12 months after birth. Second, length of time in the program was strongly associated with outcomes. The longer the number of months in the proj- ect the greater the educational attainment 12 months postpartum (among teens, women pregnant at entry), and the lower the probability of a repeat pregnancy 12 months postpartum (teens already mothers at entry). There are several other findings of note. Receipt of family planning services was associated with a lower probability of repeat pregnancy 12 months postpartum for women who were mothers at entry into the pro- gram. For many of the counseling variables cause and effect links are tenuous; many appear to be outcomes rather than causes, because women at risk of poor outcomes were often identified and receive greater attention as a result. For example, family planning and counseling services were associated with higher incidence of hospitalization of the infant 12 months postpartum. Women receiving welfare appeared to be more likely to receive a variety of types of counseling. Finally, transportation assistance was associated with school enrollment at delivery. A third important contribution of the OAPP study was the distinc- tion between women pregnant at entry and those who already have a child. The services needed are obviously different, and the impacts of services also differ somewhat. They should be considered separate groups. Finally, whether services were provided on or off site did not seem to make a difference to program success. The case management approach which consisted of one individual devoted to management did seem to be associated with program success compared to no management or to divi- sion of responsibility among several individuals. The lack of a control group greatly restricted the ability of the Urban Institute team to compare young women in the program with those who did not experience a program. The evaluators used national data from NCHS and other national studies to compare the outcomes of program participants. In general, the results were favorable on most preg- nancy outcomes. On repeat pregnancies and on proportion in school/ graduated, the OAPP mothers compare favorably with those in other programs and differ little from national data. They do not compare so well on employment and on welfare dependency; the mothers in the OAPP programs are doing worse than those in other programs and the national samples.

249 The most important result of the evaluation was the identification of specific services offered and the evaluation of their effects on clients who received the service relative to those who didn't receive the service. This is the only evaluation seen to date that does have information on timing of events and service delivery. In particular, these include date of birth, date of program entry, date services were provided, date left program, and date of fol~ow-up. So the researcher can disentangle cause and effect to some extent. Unfortunately, the evaluators were not able to analyze these data in their report. As a result, some of the results are illogical. For example, the research showed a positive association between adoption counseling and repeat pregnancy. This was probably because women who had a repeat pregnancy were more likely to have adoption counseling rather than because adop- tion counseling caused a repeat pregnancy. The report also showed that women with poor outcomes for their babies were likely to have had more medical visits; again, medical visits are probably the result of anticipated poor outcomes or pregnancy problems, rather than the cause of poor outcomes. Further research is needed on data such as these to disentangle the complicated causal connections among services and out- comes. St. Paul Maternal and Infant Care {MIC) Program Outcomes from the MIC school clinics are favorable relative to national data and to data from the population of all MIC patients (in- cluding the hospital clinic--Edwards et al., 1977, 1980~. Seven percent of the young women delivering during the 1978-79 school year had low birth weight children; 11 percent were premature. There were complications in about 15 percent of the cases, more than twice the level in older MIC patients, but comparable to national data on teen- agers. The proportion who started prenatal care in the first trimester is high--94 percent--compared with slightly over half among a national sample of white teen mothers in 1982. A fol~ow-up study of 150 prenatal patients who delivered between 1974 and 1980 indicated that 80 percent completed high school (13 per- cent dropped out, 7 percent are unknown or still in school--Edwards, 1984. Of those who stayed in school, the repeat birth rate was 1.3 percent. These figures are favorable relative to national statistics which show about 58 percent of young women in school or completed school one year after the birth and a repeat pregnancy rate of 20 per- cent (Most and Maxwell, 1981; Koenig and Zelnik, 1982~. Of course, the repeat pregnancy rate provided by the MIC program applied only to those who remained in school. Since drop-outs are more likely to have births, this inflates the estimates of the success of the program.

250 Too-Early Childbearing Network The Too-Early Childbearng Network is a network of primary preven- tion, care, and prevention/care programs receiving direct support from the Charles Stewart Mott Foundation (Mitchell and Walker, 1984~. The Foundation also funds an Impact Evaluation Project that provides tech- nical assistance to individual programs and encourages networking among programs. The evaluation team assists the programs with imple- mentation of an impact evaluation model designed to produce credible evidence of program effectiveness (Mitchell and Walker, 1984~. The results so far from the projects are somewhat encouraging. Four programs showed the incidence of low birth weight babies below their local comparison data and three showed the incidence to be below national comparison figures. All projects showed that a very high per- centage of children had regular health care, and this equaled or bet- - tered local and national statistics. Three projects showed the inci- dence of repeat pregnancy to be lower than comparable local women or comparable national incidence of repeat pregnancy to be lower than comparable statistics. Five projects showed the proportion of partici- pants in school or who have completed a high school program to be greater than among local and national comparison data. Three projects showed that participants were more likely than local comparison groups to be economically independent (not on Welfare) at follow-up. In two projects the proportion independent exceeded national figures. The problem that the impact evaluation team has had is in finding adequate locate comparison groups. In general, the team tried to come pare figures to local data, but these were not always available with the appropriate age and race breakdowns. As a backup, they used national statistics from NCHS and other sources. However, in the case of repeat pregnancy the appropriate national statistics were not avail- able, and those used produced a much more favorable outcome picture than warranted from the data. JRB Associates Projects In their report on a national study of teen pregnancy, JRB Asso- ciates (1981) evaluated the effectiveness of five model teen pregnancy programs. All of the programs showed a lower proportion of low birth- weight infants than either controls or national figures. The propor- tion entering prenatal care in the first trimester was greater than for controls in two of the projects; however, only two projects showed any improvement over national figures. In four projects repeat preg- nancies were lower among program participants than among controls or national data. School reenrollment was higher than either controls or national data in the four relevant projects. Employment was higher than national figures in three of the projects. Finally, the proportion

251 welfare free was higher than in controls or national data. Complica- tion rates were comparable to national data. Not enough information was included on the evaluation to be able to critique it. McAnarney Project RAMP The Rochester Adolescent Maternity Project (RAMP) was begun in 1969 as part of the Adolescent Program at the University of Rochester School of Medicine in Rochester, New York. An evaluation was undertaken to compare the outcomes for adolescents in RAMP to those for adolescents who received care in either a community health center (CHC) or in a hospital obstetrics clinic (HOC). RAMP delivered more services than the other programs, including a complete psychosocial evaluation, regu- lar nurse visits, prenatal classes, social worker visits, and home visits. Results show that RAMP girls had more pre and post-natal visits, as expected. RAMP girls showed greater use of contraception pre-preg- nancy and post-partum. They had fewer repeat pregnancies, live births and abortions at the two year follow-up. There were no differences however, in average gestational age, birthweight, 5 minute Apgar score, live birth status, or fetal distress. There was no difference in school enrollment or graduation two years later, nor was there any dif- ference between the groups in receipt of public assistance. Unfor- tunately, the number of participants in the evaluation was very small: 25 RAMP patients, 37 in the HOC, 20 in the CHC. Thus the results are only suggestive, not definitive. McAnarney Project START START, which is under the auspices of the Rochester Adolescent Maternity Project (RAMP) at Strong Memorial Hospital, stands for Ser- vices, Training, Analysis of effectiveness, and Regional Training. The purpose of the program is to provide services to pregnant 10 to 14 year old adolescents and their families. (RAMP serves adolescents 18 years of age and younger). Besides the traditional prenatal services, the services of nonprofessional family counselors and certified nurse midwives are utilized. The family counselor visits the adolescents' homes and facilitates their interactions with the clinic. Community health nurses also visit the the adolescent and her infant. The number of patients served is very small--30 patients were served in 1980-81. Data from the evaluation show that only one baby was low in birthweight (an incidence rate of 4 percent). The incidence of repeat pregnancy appears to be rather high: 16 percent were pregnant again within one year (approximately). The educational component did not appear to have been successful, and McAnarney recommends that it be dropped in the future. Only half the teens accepted educational counseling, and fewer

252 than half of these are still attending school, for a continuation rate of under 20 percent. All these adolescents had serious academic prob- lems prior to the pregnancy. fully. This is an interesting program, but too small to evaluate success Johns Hopkins Adolescent Pregnancy Program The Johns Hopkins Adolescent Pregnancy Program started out in 1974 as a special obstetric clinic to improve pregnancy outcomes for young teenagers delivering their babies in high risk areas around the Johns Hopkins hospital. It was soon observed that a sizable number of the adolescents had a second pregnancy soon after f irst delivery, and that many babies had health problems 4-6 weeks postpartum. This led to the initiation in 1976 of a follow-up component which provided preventive health services for both mother and baby at the same visit to the follow-up center. The pattern of comprehensive services was extended through community linkages. Preventive care was extended to primary health care for adolscents and their babies for about 50 percent of those delivered, the highest risk patients. Two evaluation studies of the JHAPP programs were conducted tHardy, 1983~. The first study compared the repeat pregnancy experience of those girls who were referred after the postpartum visit to community agencies for their continued care (control) with those enrolled in the follow-up program of the Hopkins Center (TAC clinic). Girls were matched on race, year of age, and date of delivery. The repeat preg- nancy experience of the experimental group was substantially better than that of the control group; the repeat pregnancy rates for the latter were similar to those of national samples of teens. The second study compared young women enrolled in the Hopkins Com- prehensive Care Clinic with those in the Teenage Clinic (TAC clinic referred to above). The former provided continuing medical care and family planning services for the young mothers, but not the intensive health and parenting education nor the psychosocial support that was available in the latter. Results showed that 36 percent of those en- rolled in the Teenage Clinic and 34 percent of those enrolled in the Comprehensive Care Clinic experienced repeat pregnancies within two years of delivery. Thus there was no difference between the programs, and very little improvement over national figures for repeat pregnancy within two year (37 percent) (Hardy, 1983~. This evaluation is specifically focused on assessing repeat preg- nancy rates. Unfortunately, the methodology used to assess such rates are not clear, and from the written report it is difficult to extract repeat pregnancy rates at 2 and 3 years. Data are not given for 1 year repeat pregnancy rates. A life table methodology would have been pre- ferable. The second problem with the evaluation is that the comparison

253 groups are simply other adolescent pregnancy programs. Yet it is not entirely clear what services the adolescent received in each of the separate programs mentioned and how they might differ. In any case, using other programs as comparisons would seem to reduce the likeli- hood of finding significant program effects. Prenatal/Early Infancy Project This study evaluated a comprehensive nurse-home visitation and transportation service designed to improve pregnancy outcomes and child health and development in a group of families at risk for preg- nancy and childrearing dysfunctions. The nurse home-visitation pro- gram began during pregnancy and followed the families through the second year of the child's life. Mothers of all ages were included in the study, but analyses were conducted and data tabulated separately for high risk groups (e.g., adolescent mothers). The study consisted of a true experimental design, with families randomly assigned to one of four treatment groups. Families in the first condition received no services during pregnancy (other than the excellent quality standard prenatal care provided locally). Screening was provided at 12 and 24 months. Families in the second condition received free transportation to regular prenatal and well-child visits as well as the screening offered the first group. Families in the third condition received nurse home visitation during pregnancy, in addition to transportation and screening. Families in the fourth condition received nurse home visitation during the children's first two years of life, in addition to the prenatal home visitation, transportation, and screening offered the third group. The major differences in treatments were between the first-second and third-fourth groups. Treatment effects were strongest for the highest risk groups--adolescents, smokers and unmarried women. The nurse-visited young (14-16) adolescents gave birth to newborns who were nearly 400 grams heavier than those born to adolescents in the comparison group. 89 percent of the poor unmarried, young mothers (under 19) who were visited by a nurse had either completed or returned to school, compared to 52 percent of those comparable mothers who had no nurse. There was essentially no reduction in the incidence of re- peat pregnancy among adolescents. This is a very well-done, interesting study of the impact of an innovative program. It takes a number of factors placing women at risk for poor maternal and child outcomes and considers their effects both jointly and separately. It is especially good because smoking is included as a risk factor. The results show clearly that the program has some positive benefits for adolescents; the benefits for older women, particularly married women are less clear. The random assign- ment feature and the use of different types of treatments is very good. Unfortunately, since most women can obtain excellent care through exis- ting services, the effect of the treatment is really only marginal.

254 Thus the fact that some effects are significant among important sub- groups such as adolescents is important. In addition, since the pro- gram included women of all ages, the the sample of unmarried women under 18 consisted of only 112 cases, which were divided into 4 dif- ferent treatment groups. However, the fact that significant effects were found is encouraging. Comparison of These Evaluations One of the important contributions of Klerman and Jekel's evalua- tion was methodological. It constituted a very careful analysis of the effects of programs and served as a model for later analyses. Substan- tively, it showed clearly the decay of short term gains that resulted from a program, and it pointed out the importance of distinguishing be- tween short.term and long term effects. Finally, it pointed out the importance of looking at repeat pregnancies as an important outcome. Women having repeat pregnancies were especially at risk of undesirable outcomes The Project Redirection staff also discovered the importance of focusing on delaying repeat pregnancy. They did not focus much impor- tance on it at first, but, soon alarmed at the number of repeat preg- nancies, began putting more pressure on participants to contracept effectively. Project staff monitored contraceptive use and the com- munity woman checked up on participants assigned to them. In one site (Phoenix), teens who became pregnant again were terminated (see Branch et al., 1984:49~. The issue of repeat pregnancy is a tough one for programs. Klerman and Jekel noted how hard it is for a program/its personnel to say 1) you can postpone the next baby if you want to but 2) if you become pregnant again come see us again early. They hypothesized that perhaps young mothers felt guilty about returning pregnant to a program which had stressed contraception, and that is why subsequent pregnancies had less adequate prenatal care (Klerman and Jekel, 1973:68~. (On the other hand, it may simply reflect a selection problem--the most care- less girls become pregnant again, although Klerman and Jekel didn't seem to think this was so, Klerman and Jekel, 1973:68~. The contribution of the American Institutes for Research evalua- tion of Project Redirection was also methodological it represented a very careful evaluation of overall program impacts over a relatively short term (24 months). However, there was no possibility of deter- mining which aspects of the program produced which results. Later analyses to sort out different program aspects and their impacts would be helpful, especially since there were a number of innovations. The effect of the community woman would be particularly helpful to know. The contribution of the OAPP evaluation was precisely the attempt to look at the contribution of specific services provided teens. Un

255 fortunately, this evaluation was not capable of determining overall program impact because of the lack of a control group. The Prenatal/Early Infancy Project represents the trend toward more rigorous program evaluations. Subjects were randomly assigned to treatment/comparison groups; several types of interventions were tested simultaneously; and a number of risk factors were considered simulta- neously in looking at outcomes of the project. In addition, the proj- ect was located in one area, treatment conditions were uniform across all subjects in each condition, and the project was relatively modest in size, but not too small for effective evaluation. This project appears to provide a valuable model for other evaluations. Probably the most important contribution was to show that separate aspects of the program could be evaluated at the same time by developing dif- ferent treatment conditions. The Too Early Childbearing Network has serious problems in develop- ing effective evaluation because of lack of appropriate local control groups. The national data are useful, but are hard to make comparable enough (in terms of race, age, SES) to be of use. The Mcknarney Project was simply too small for effective evalua- tion. The evaluation of the Johns Hopkins program mixes evaluations of different types of programs and does not clarify what types of ser- vices are being provided in which program. As a result, the conclu- s ions are weak and unclear. Finally, not enough information is in- cluded in the desc ription of the JRB Associates program evaluation to adequately evaluate it. Economic Approaches One major barrier to the participation of males in teenage father programs is that services often fail to meet their real life needs. "Commonly, young men will not enter a program unless it provides them with job-related skills and training; they will not stay in a program over time unless it helps them deal with their more personal, relation- ship-oriented problems" (Klinman et al., 1985:14~. This suggests that a substantially different type of prog ram is needed to attract and keep males than programs tradit iona lly oriented to teenage mothers. Job training should constitute a major part of such programs. An appropriate set of services such as legislatively mandated for teenage mothers has not been established for teenage fathers. Delaying family formation or preventing repeat childbearing has not been a primary goal of job training programs. In fact, most pro- grams appeared to ignore the family responsibilities of enrollees, in spite of the fact that a large proportion of male and female partici- pants have children of their own. As a result, there is not much evi- dence as to the impact of such programs on teen childbearing. One evaluation of the Job Corps conducted by Mathematica Policy Research

256 found that, among young women, participation in the Job Corps appeared to delay family formation and reduce the incidence of extramarital childbearing, and that the impact on employment, earnings, education and welfare receipt was larger for women without children than for those with children (Mallar et al., 1978~. A second study which promises to provide more information on the impact of youth programs on family formation is one currently under way by Olsen and Farkas. The program whose data they will evaluate is the Youth Incentive Entitlement Pilot Projects (YIEPP). This program guaranteed jobs to 16 to 19 year olds enrolled in school if they stayed in or returned to school and met specified attendance and performance standards (from Simms, 1984~. Although early results suggested no imp pact of the program on the rate of childbearing (which was high) among females, a three year study cited above will analyze these data in greater detail, looking for impacts among certain groups of partici- pants (Olsen, 19841. A third project, the Teen Fathers Project, was conducted between April 1983 and March of 1985 by Bank Street College in 8 sites around the country. This project obtained teen fathers through their female partner, outreach workers, the school, YMCA and word of mouth. Some 400 teen fathers received job training, job skills, job referral ser- vice, educational counseling, and instruction on site in parenting skills. The goal of the program was a change in educational status-- obtaining a GED or returning to school, obtaining employment or job training, parenting skills, and increased knowlege of and involvement with the child. Since this was a demonstration project, the developers were not sure how successful they would be at recruiting fathers. Pre- test and posttest data were collected; no control group was obtained. The evaluation will simply compare before and after measures on the fathers. Since the program has been successful in recruiting teen fathers, the project will continue with a more rigorous evaluation. The final report on the first phase of the project is due in the fall of 1985. A fourth study, currently being conducted by Public/Private Ven- tures is is its second year. This project provides summer employment and remedial education to 14-15 year olds during two summers in a rwo, with a support component during the school year. The goal of the proj- ect in the short term are 1) to produce learning gains instead of de- cline in the summer months, and 2) to improve knowledge of birth con- trol and outcomes of teen pregnancy. Interim (medium length) goals are 1) improved school performance. 2) improved labor market perfor- mance, and 3) improved contraceptive use. Long term goals include 1) high school graduation, 2) improved labor market performance, and 3) lowered teen parenting. Those eligible for the program are those eli ~ ~ ~ ~ ~ ~ in school, poor, and educa- is being conducted along treatment or control groups, telephone follow-ups, infor gible for job training programs--enrolled tionally deficient. the evaluation which with the program randomly assigns youth to and obtains information through a pretest,

257 mation from school records each semester, and a final interview 6 months after putative graduation date. The control group receives full-time summer work, but no other intervention. The first phase of the program was implemented in 1985-86 with 1,600 youth in 5 U.S. cities. Results from the first summer of the program show that the program had some positive impacts on enrollees. In particular, the program was successful in stemming the learning losses that would have occurred in the absence of the treatment. While treatment youth scored higher than control group youth at post-test in both reading and math, both groups experienced losses. The losses experienced by the control group were significantly larger than those experienced by the treatment group. Females and Hispanics appeared to benefit the most from the program. Participants' knowledge of birth control was increased. Finally, more treatment boys reported abstaining from sex during the program than control boys. Welfare and Medicaid AFDC is the major federally funded and state administered welfare program in the United States. It provides cash assistance to economi- cally needy individuals who are eligible by virtue of being a female family head with children under 18, and meeting certain income require- ments. Female subfamily heads are also eligible for themselves and their children or for their children only. Some states have instituted a program which makes eligible families in which the father is present but unemployed and who also fall below a certain minimum income level. Medicaid is a health insurance program available to all AFDC recip- ients. There has been a considerable attention paid to the hypothesis that generous transfer payments create an economic incentive to early childbearing. It is not possible to test empirically whether the existence of transfer payments under the AFDC program affect early childbearing, since all states provide transfers under the program. However, there is substantial variation in level of payments and in eligibility requirements, which are reflected in acceptance rates, and these can be related to fertility levels. This debate has a long history, going back at least to the late 1960s. A study by Placek and Hendershot (1974) that tested a number of propositions drawn from this belief for welfare recipients of all ages found no support for what has been called the "Brood Sow" myth. In fact, they found welfare mothers more likely to use contraception when on than when not on welfare, less likely to define pregnancies as wanted, and less likely to have a subsequent pregnancy when on than when not on welfare. Of course, this paper did not specifically ad- dress the issue of teen childbearing. In an early study of out-of- wedlock birth rates in 58 SMSA's, Janowitz (1976) found that a higher

258 level of state welfare benefits was associated with higher out-of- wedlock birth rates among teenagers, net of other factors such as level of schooling, earnings and unemployment. However, whether the state also had a unemployed fathers program in AFDC(AFDC-U) was also found to be associated with higher out-of-wedlock birth rates for teenagers, a counter-intuitive finding. Thus the AFDC variable may be a proxy for variables which are associated with it, but which are not controlled in the model. This is especially a problem for this analysis since AFDC payments are only obtained at the state level, while birth rates were measured for each SMSA as a whole. Several recent studies have tested the-association between welfare and teen childbearing. None found any impact of either level of AFDC benefits or acceptance rates on out-of- wedlock birth rates among teens in 1971, 1974 or 1975 (Field, 1981; Moore and Caldwell, 1977; Moore, 1980~. Moore and Caldwell (1977) also explored the impact of AFDC benefit level and acceptance rates on initiation of sexual activity, pregnancy and pregnancy resolution among teens. They found weak and inconsis- tent effects which led them to conclude that there was no association between welfare generosity or acceptance rates and the probability of initiating sexual activity or becoming pregnant. However, they did find effects of relatively generous AFDC benefit levels on the proba- bility of abortion. Teens in states with relatively generous benefit levels were less likely to have an abortion. On the other hand, teens in states with low AFDC acceptance rates were found to have a higher probability of bearing an out-of-wedlock child. The authors concluded that there was no statistically significant evidence linking welfare availability with the probability of carrying an out-of-wedlock preg- nancy to an out-of-wedlock birth (Moore and Caldwell, 1977:166-167~. The authors also looked at the impact of having an unemployed father program on pregnancy resolution. They found that women in states with such a program were, in fact, no more likely to marry, but were less likely to bear an out-of-wedlock child. Recent analyses of California data collected in 1972-74 (Leibowitz et al., 1980; Eisen et al., 1983) found that premaritally pregnant girls receiving state financial assistance (welfare) were less likely to marry and more likely to bear an out-of-wedlock child than compa- rable girls not currently receiving such assistance. This study is flawed. The major problem is that eligibility for welfare receipt is measured by current welfare receipt. In fact, many of those young women not currently on welfare would be eligible if they gave birth to a child out of wedlock. This measure of welfare availability is prob- ably contaminated by attitudes toward welfare recipiency, by awareness of welfare availability, and by prior intention to carry out or abort the pregnancy. Those intending to bear a child out-of-wedlock have a greater incentive to establish welfare eligibility than those not plan- ning to bear the child. Thus the causal direction of the association is unclear (Moore and Burt, 1982~.

259 A recent study by Ellwood and Bane (1984) used data f rom several large nationally representative data sets, the Survey of Income and Program Participation, and the 1960 and 1970 Censuses to explore the impact of welfare on AFDC on divorce and separation, births to non- married women, single parenthood, female headship and the living arrangement of single mothers. The study differs from the others de- scribed earlier in that it controlled statistically for a number of unmeasured differences between states. In addition, it used three dif- ferent types of methodologies to answer the question of the impact of AFDC: a comparison of likely versus unlikely recipients, of eligibles versus non-eligibles, and of APDC benefit levels and changes in family and household structure and fertility over time. The authors concluded that they could isolate no impact of the maximum AFDC benefit for a family of four on births to non-married women. However they did iden- tify several effects on family and household structure. In particular, divorce and separation rose slightly with a $100 increase in the AEDC maximum benefit, as did single parenthood and female headship. The largest increase was in the probability that a young single mother would live independently, an increase of 50 to 100 percent in one analysis. This analysis supports previous work that shows no impact on fertility. However, it also showed that AFDC does affect the living arrangements of young women; it permits them to live independently of their families. The conclusion is that level of welfare benefits and acceptance rates and other indicators of availability do not appear to be asso- ciated with sexual activity or pregnancy; however, they may be asso- ciated with whether or not a girl who is pregnant decides to abort, marry or bear an out-of-wedlock child, and with whom she chooses to live. The evidence is not very strong; more research is needed. The availability of Medicaid and other health care benefits might also affect teen childbearing. In particular, by improving the health of the mother and baby during pregnancy it could reduce the number of miscarriages and increase the number of live births. The availability of Medicaid to cover the cost of an abortion might increase the number of teens who would opt for abortion over a live birth. However, this would apply to teens who either already had one child, or those living in families receiving welfare already. The only research on this to date has been conducted on the issue of Medicaid coverage of abortions and teen fertility (Moore, 1980~(see earlier discussion). Results were inconclusive. Given the lack of effects of AFDC found so far, it is unlikely that Medicaid benefits would be found to have an impact either. Recent legislation extends Medicaid eligibility to pregnancy if a woman would be eligible at birth. The fact that medical costs are covered may, in fact, encourage a young woman to bear and keep the baby. However, reported time delays in getting on Medcaid may reduce the likelihood that such coverage will affect childbearing.

260 SUMMARY AND CONCLUSIONS It is important to caution the reader, first, that it is difficult to show causal relationships between policies or programs and individ- ual behavior. There are so many other factors, many of which are dis- cussed in earlier chapters, that do affect individual behavior, factors more proximate or immediate in terms of the consequences to the in- dividual. Programs and policies should not be expected to have large impacts. That some have been found to have impacts and that others have not is, therefore, of great interest and importance. Programs to Prevent Sexual Activity Sex education programs appear to be consistently associated with increased knowledge about sex and sexual behavior among participants; however, there is little evidence for a relationship with sexual ac- tivity. In spite of common beliefs, those taking a sex education course do not appear to change their own beliefs and values, although they do become more tolerant of the beliefs and values of others. Parental involvement is more a moral issue than one of conse- quences. That is, even if parental knowledge or communication about the sexual activity of their children were not found to have any rela- tionship at all to their sexual and contraceptive activity, many would still consider it an important issue. There is some evidence that parent-child programs do increase communication about sex and birth control between parents and children. The evidence for an impact of communication on their children's behavior is weak. Programs to Prevent Pregnancy There is growing evidence that sex education programs are asso- ciated with improved contraceptive use among those young women who are sexually active. More needs to be known about what aspects of sex edu- cation programs are associated with improved contraceptive use among teenagers. Family planning programs have been frequently and effectively evaluated. As a result, we know a lot about such programs. There is no evidence that the availability of family planning services increases sexual activity among female teenagers; however, it does appear to im- prove contraceptive use and reduce their chances of having an unplanned pregnancy and out-of-wedlock birth. A number of clinic characteristics appear to be associated with attracting teens early in their sexual careers. These are related to outreach and community relations efforts, the convenience of attending the clinic, and the availability of contraceptive services from physi- cians and pharmacies. Most of the same factors also promoted clinic

261 continuation among those who initiated clinic attendance. Another study found that mean levels of contraceptive use were highest in clinics in which clients expected and staff employed authoritative guidance in helping the clients to select a contraceptive method. Finally, a third study found client satisfaction with her method to be the best predictor of contraceptive continuation. The only relatively rigorous evaluation of a school-based program is that of the St. Paul MIC project. The evaluation data suggest sub- stantial effects on birth rates in the school (which declined) and sub- stantial effects on post-delivery enrollment among mothers (which in- creased). Neonatal and maternal outcomes do not differ from national statistics. One available evaluation of a non.school based prevention program (West Dallas Youth Clinic) suggested that the program had been success- ful in reducing teen birthrates. Pregnancy and Pregnancy Resolution For a young woman who is premaritally pregnant, the availability of abortion is associated with a greater probability that she will have an abortion and a lower probability that she will bear a child out-of-wedlock. There are no available evaluations of the recently funded programs directed at increasing adoption as a resolution to teen pregnancy. Given the small incidence of termination of parental rights, such a program is not likely to have a major impact on teen childbearing (Bachrach, 1985; Muraskin, 1984), although more research is needed on this issue. Public school programs for pregnant teens are limited in availa- bility and scope. A complete review and evaluation of these programs was not attempted in this chapter. WIC, a nutritional program for pregnant and parenting women, has been shown to successfully reduce the inside we of low birth weight babies, an impact which is especially strong for teenagers. Parenthood Teenage pregnancy programs are directed primarily at pregnant teens and/or teen mothers. Such programs appear to have relatively short term impacts on specific targetted goals. For example, medical pro- grams improve infant health; education programs improve educational outcomes; programs emphasizing employment improve employment outcomes. Several types of programs do appear to reduce subsequent childbearing; however, the effects are somewhat weak and may be relatively short-term

262 on impact. Substantial long-term effects on schooling appear to be robust. Programs for teenage fathers are relatively new and have not yet been evaluated. An area of great interest is that of the impact of welfare availa- bility and benefit levels on sexual activity and out-of-wedlock child- bearing. There is little theoretical rationale for expecting an asso- ciation between welfare benefits and initiating sexual activity, and no evidence exists for an empirical relationship either. However, there is a stronger argument that for some pregnant teens, welfare may be an attractive option. It may not be a salient option for a middle income girl; however, it may be one for a girl from a low income family. In fact, of the two studies to ask this question, one shows that pregnant girls who are receiving financial assistance in the form of welfare are more likely to bear a child out-of-wedlock and less likely to marry or abort. However, since the other study to explore this issue fails to find such a relationship, this association does not appear to be definitive. A third study found, rather, that AFDC does affect living arrangements, increasing the likelihood that a teen mother will live independently. The one study that looked at contra- ceptive use by welfare mothers found that welfare mothers were more likely to use contraception, less likely to define pregnancies as wanted, and less likely to have a subsequent pregnancy when on than when not on welfare. Unfortunately, these recipients included women of all ages, not just teenagers. Conclusions A variety of types of programs have been reviewed in this chapter. They have been grouped according to their primary focus: prevention of pregnancy versus amelioration of the unfavorable potential outcomes of "early" childbearing. Unfortunately, approaches to prevent preg- nancy other than providing contraception (or information about it) are still in the developmental stage. There is no evidence so far that such approaches prevent sexual activity, encourage contraceptive use or prevent pregnancy. In contrast, all the evidence provided over- whelmingly supports the effectiveness of provision of contraception in preventing pregnancy. Whether this be conducted by family planning clinics or private physicians depends in part on characteristics of the individual seeking such services. Clinics are especially impor- tant sources for black and poor teenagers. Recent work has focused on placing the clinics where they can serve teenagers even better--in the schools. The little evidence there is (one project) does suggest some success in reducing birth rates and keeping teens in school. There is no evidence as to whether pregnancies are actually prevented by such school clinics, however. Once pregnant, a number of factors influence the decision as to how to resolve an unplanned pregnancy. The availability of abortion may result in a young woman selecting abortion over either adoption or

263 bearing an out-of-wedlock child. At the present time the number of young women who terminate their parental rights to a child in the United States scans to be ire ry small. There is little or no research that looks at the long term nonhealth effects of either abortion on the young woman or on the long term nonhealth effects of giving up a child for adoption on the young woman and her child. Nutrition programs (WIC, in particular) are effective in improving infant outcomes to teen mothers. Finally, there are several models of programs to serve teen mothers that have been implemented and evaluated. These evaluations suggest short-term effects in the areas of focus long-term effects for all teenagers have of the program. Substantial _ _ yet to be demonstrated; some research folds; Polit et al.) suggests that the most disadvantaged teens may benefit most from such programs. If so, this is somewhat en- couraninq news. On the other hand. the McAnarnev program illustrates , _ _ _ . _ , _ , _ _ _ Am, the u'ttlCUlty of reaching the youngest teens. The evaluations re- viewed do point out the importance of evaluating not just the program as a whole against other programs, but of evaluating specific program components. The research so far provides less than adequate informa- tion to evaluate what aspects of the programs currently in operation produce what results for what types of young women {and men). As a result, it may be too early to promote particular exemplary program models. Note 1. Authoritative guidance refers to the nurse, an authority figure either telling the client what birth control method to use or per- suading her to use a particular method.