Induced Abortions Among Adolescent Women in Rural Maharashtra, India

Abstract

Abstract

In a study in rural Maharashtra, India, adolescents constituted 13.1% of the 1717 married women who had an induced abortion during an 18-month period in 1996–1998. The 197 adolescents who were subsequently interviewed had a lesser role in the decision-making process on abortion than women older than them. Most abortions were obtained in the private sector. Though spacing was the main reason for adolescents seeking abortion, prior contraceptive use among them was low. Additionally, they were less likely to receive post-abortion contraceptive counselling or to adopt contraception. Sex selection accounted for more than a fifth of abortions among adolescents. Additional qualitative data from 43 never-married and separated adolescents seeking abortion showed that non-consensual sex made many pregnancies unwanted, and cost, limited mobility, lack of family and partner support and the need for privacy to prevent stigma led many to go to traditional providers, even though safer options existed. Family planning programmes need to address the contraceptive needs of newly married adolescent women as well as unmarried adolescents. Informing adolescents of their legal rights, sensitising providers to adopt an empathetic attitude, and exploring innovative ways of increasing access to safe services for unmarried adolescents are all recommended.

Résumé

D'aprés une étude réalisée dans le Maharashtra (Inde rurale), les adolescentes représentaient 13, 1% des 1.717 femmes mariées ayant avorté sur une période de 18 mois en 1996–1998. Les 197 adolescentes interrogées par la suite avaient joué un rôle plus effacé dans cette décision que leurs aînées. La plupart des avortements avaient été pratiqués dans le secteur privé. L'espacement des naissances était le principal motif d'avortement chez les adolescentes, mais l'utilisation préalable de contraceptifs était faible. En outre, elles avaient moins de probabilités d'être informées sur la contraception après l'avortement ou d'adopter une méthode de contraception. Le choix du sexe motivait plus d'un cinquième des avortements chez les adolescentes. Des données qualitatives supplémentaires obtenues auprès de 43 adolescentes célibataires ou séparées souhaitant avorter ont montré que dans beaucoup de cas, ce désir était dû á des relations sexuelles non consensuelles; le coût de l'avortement, la mobilité limitée, le manque de soutien familial et du partenaire, et le souci de discrétion pour éviter la stigmatisation incitaient nombre d'entre elles á s'en remettre á des prestataires traditionnels, même en présence d'options plus sûres. Les programmes de planification familiale doivent répondre aux besoins des adolescentes - jeunes mariées ou célibataires. Il faut informer les adolescentes de leurs droits, convaincre les prestataires d'adopter une attitude compréhensive et élargir l'accés des adolescentes célibataires á des services sûrs.

Resumen

Según un estudio realizado en un área rural de Maharashtra, India, de 1717 mujeres casadas que tuvieron un aborto inducido durante un periodo de 18 meses en 1996–1998, un 13, 1% eran adolescentes. Las 197 adolescentes que fueron entrevistadas posteriormente participaron menos en la toma de decisión respecto al aborto que las mujeres mayores. La mayorı́a de los abortos se obtuvieron en el sector privado. Aunque el espaciamiento era la razón principal que aducían las adolescentes, el uso de anticonceptivos entre ellas era bajo. Era, además, menos probable que ellas recibieran consejerı́a postaborto sobre la anticoncepción o que adoptaran un método anticonceptivo. Más de un quinto de los abortos entre las adolescentes era por razones de selección de sexo. Datos cualitativos adicionales sobre 43 adolescentes nunca casadas o separadas que buscaban abortar mostraron que el tener relaciones sexuales en contra de su voluntad resultaba a menudo en embarazos no deseados. En muchos de estos casos, debido a los costos, la falta de movilidad y del apoyo familiar o de una pareja, y la necesidad de privacidad para evitar el estigma, las muchachas recurrían a proveedores tradicionales en lugar de otras opciones más seguras que existen. Los programas de planificación familiar deben atender las necesidades anticonceptivas de las mujeres adolescentes recién casadas tanto como las adolescentes solteras. Se recomienda informar a las adolescentes de sus derechos legales, sensibilizar a los proveedores para que adopten actitudes de mayor empatía, y explorar maneras innovativas de aumentar el acceso a servicios seguros para las adolescentes solteras.

Key Words:

• adolescents
• induced abortion
• abortion providers and services
• India

Abortion (or medical termination of pregnancy, MTP) has been legal in India on a broad range of medical and social grounds for nearly three decades. Over six million abortions take place each year in the country. Adolescents account for about 6% of abortion seekers [1], though some hospital-based studies have found that adolescents constitute as much as 20–30% of client load [2][3]. While changing social norms have seen an increase in pre-marital sexual activity and unwanted pregnancies among adolescents, given the near universality of early marriage in India, most adolescent abortion seekers are married [4].

Information about adolescents who seek abortion in India comes mainly from urban hospital-based studies [4][5]. A population-based study in rural Maharashtra in 1993–1996 found that while overall mortality from abortion complications was low, it was disproportionately higher among adolescents, accounting for three of 39 adolescent maternal deaths that were identified in a 35-month surveillance period. However, there were an additional four deaths among unmarried pregnant adolescents who committed suicide in order to preserve family honour and died without having attempted an abortion [4]. The present study, undertaken among both married and unmarried women who had undergone an induced abortion in the same geographic area, aimed to explore their access to abortion services, decision-making on abortion, determinants of provider choice and extent of morbidity experienced.

The study took place in 139 villages with a population of ≈325,000 in three districts in Western Maharashtra in India. Some 15% of villages in the study area lie within 5 km of a functioning government abortion facility. Services are also offered by small private hospitals and NGO clinics. The KEM Hospital has long standing links with the community and with health care providers in the study area and provides community outreach health services (including abortion) in part of that area.

Methodology and participants

The methodology evolved through group discussions and key informant interviews, including with abortion service providers [6]. We sought to identify all the women who had undergone an induced abortion in the study area in an 18-month reference period during 1996–1998. Secondary sources – community workers, women's groups and community women as well as health service providers (including those who did not themselves provide abortion services) – identified potential participants. Providers (and wherever possible, community sources) were asked to explain the study to women they identified and did not provide contact information if they felt a woman's interests would be jeopardised. While the focus of the study was currently married women, some unmarried women were also identified.

Where possible we tried to arrange for interviews to take place in the clinic setting, but in the absence of complications, post-abortion follow-up visits are not the norm so this was not always possible. Married women who could not be interviewed in a clinic were approached for interviews at their homes, provided it was within six months of the abortion (in order to minimise recall bias). Given the fact that consent is dynamic and that neither the woman nor we would have control over the presence of others at her home during interview, a second layer of protection was afforded to participants. We clustered women into geographic areas and drew up a sample of women from a list of eligible couples, among whom were included the women who had been identified as possible cases. Thus at any given time, both cases and non-cases (dummy interviewees) were being interviewed, using the same questionnaire.

Questions began with general health and pregnancy history. Women were not told that the interviewer knew they had had an abortion. If and only if, during the course of telling pregnancy history, a woman mentioned an induced abortion in the reference period was she asked permission to interview her about the abortion experience. For this we used a combination of a structured questionnaire with open-ended probes and a qualitative in-depth time-line of sequence of events. If non-cases reported an abortion in the reference period, they too became participants. By the end of the study, seven women who were initially interviewed as `dummy interviewees' became cases. Thus, the dummy interviews also served as a cross-check on our case finding.

To increase the chances of one-to-one privacy, we used a `team' of interviewers, where one conducted the actual interview while the others engaged family members in one-on-one conversations. This allowed a modicum of private time with the woman. The interviewers, who included anthropologists, physicians and medical social workers, were well trained in qualitative in-depth interviewing. They had been working in the community for several years and were also trained to divert the interview when unwanted family members or outsiders insisted on listening in.

These strategies allowed us to ascertain the woman's willingness to talk about her abortion without confronting her with information she may not have wanted to be confronted with, gave her a dignified way of refusing the interview and allowed her and us to present a non-threatening explanation of the study to family members or others if required.

Given the focus of the study and the fact that many women could face social stigma if their abortions became public knowledge, we interviewed unmarried women only if the interview could be in the hospital or other setting that afforded both privacy and anonymity. Data were collected through in-depth interviews using a broad interview guideline. Referrals for health care were provided to all cases and “dummy interviewees” who needed assistance. Additionally, 52 public and 107 private providers of abortion services in the study area and from nearby towns and cities were interviewed in depth. These included all providers who were mentioned during key informant or in-depth interviews with cases and who could be contacted and agreed to the interview.

1717 married women were identified as having had an induced abortion during the 18-month reference period, of which 226 (13.1%) were ⩽20 years old (defined as adolescents for the purposes of the study). Of these, 16 could not be contacted, or refused to be interviewed, 11 said the abortion during the reference period was spontaneous and two did not report an abortion at all (Table 1

Table 1 Married women identified as having had one or more induced abortions in the reference period (n=1717)

	Married women ⩽20 (n=226)	Married women >20 (n=1491)
Reported an abortion and agreed to interview	197	1212
Reported a spontaneous abortion	11	91
Did not report an abortion	2	8
Could not be interviewed	16	180

). Hence, the data in this paper are from the 197 adolescent women who were fully interviewed, who reported 199 induced abortions in the reference period.

Additionally, 132 currently unmarried women were identified as having had an induced abortion in the reference period. Of these, 43 were ⩽20 years of age. Six were not interviewed because we could not ensure privacy; one had died. Basic information from the rest was collected through interviews in a hospital where they had sought an abortion or were being treated for complications. Eight of these 36 adolescents did not report having had an abortion, 12 did not want to discuss the episode in-depth and 16 agreed to an in-depth interview. These included 12 who were never-married and four who had been married but were separated from their husbands at the time of the abortion.

A descriptive profile of the married adolescent group is presented, as are comparisons with older married women who underwent an induced abortion in the same reference period. While no direct comparisons with the unmarried adolescents or older women are possible, the paper uses the qualitative interview data to highlight some of the problems specific to those who were unmarried.

Profile of married adolescent women who had an abortion

Nearly half (48.7%) of the married adolescents lived in a joint family setting, and their mothers-in-law were living with them. The mean education level was 6.9 years. While the study looked only at induced abortions during the reference period, 1.6% of this group reported having had one or more induced abortions prior to the reference period as well. A significantly higher proportion of adult women (13.7%) had one or more induced abortions prior to the reference period (Table 2

Table 2 Profile of married adolescent and adult women with induced abortion in the reference period

	Married adolescents ⩽20 years (n=197)	Married adults 20 years (n=1212)	Significance
Education (mean no. of years)	6.9 yr	6.12 yr	0.28
Mother-in-law living in same house	48.7%	34.7%	0.000
Reported one or more induced abortions prior to reference period	1.6%	13.7%	0.000
Economic independence
Has an independent source of income	11.7%	23.9%	0.000
Able to keep/spend her own money	20.0% (n=23)	36.7% (n=290)	0.13
Autonomy
Permission needed to go to market	66.1%	55.6%	0.091
Permission needed to go to health centre in same village	81.4%	67.5%	0.001
Permission needed to go to natal home	99.4%	93.4%	0.003
Major decision-making role on
Cooking, market purchases	14.6%	31.7%	0.000
Purchase of major household items	15.7%	26.4%	0.000
Seeking care for own illness	39.6%	46.1%	0.17
Discussion/communication with husband in last year on issues relating to
Child illness	62.9%	92.1%	0.000
Contraception	74.0%	84.7%	0.024
Discussion/communication with mother-in-law in last year on issues relating to
Child illness	28.5%	36.9%	0.05
Contraception	4.0%	3.4%	0.859

).

Married adolescents were less likely to have an independent source of income or to have control over the money earned as compared to the older married women. They were also significantly less likely to play a major role in household decisions like cooking or buying large household items. Mobility was more restricted for the younger women; 81.4% required their husband's permission to visit the health centre, while 67.5% of adult women required such permission. Even though the younger women were more likely to be living in families with their mothers-in-law, few of them had discussed issues relating to childcare or fertility with them. They were also less likely to have discussed childcare issues with their husbands (Table 2).

Role in decision-making on abortion

While the great majority (75.8%) of married adolescent women felt they had been the primary decision-maker whether to have an abortion, this was significantly less compared to 88.8% of adult women. For the rest, husbands were more or less equally likely to be the main decision-maker for adolescent vs. adult women, while mothers-in-law were more likely to decide for younger than older women when living in the same house. Furthermore, more adolescent than adult women felt they had been pressured into having an abortion when they wanted to continue the pregnancy (Table 3

Table 3 Decision-making on abortion: married adolescents compared with adult women (%)

	Married adolescents ⩽20 years (n=199)^* (%)	Married adults ⩾20 years (n=1293)^* (%)
Woman played a major role in the abortion decision^***	75.8	88.8
Husband made abortion decision, woman agreed	10.7%	8.9%
Mother-in-law made abortion decision, woman agreed^**, ^***	3.4	0.2
Woman wanted to continue the pregnancy but had an abortion due to family pressure^***	4.5	1.5

* n=total abortion episodes.

** Cases where mother-in-law was residing in same house.

*** 0.005 significance.

).

“I wanted to keep it but he said, `Why are you so eager to keep it? It must be someone else's. If it is mine you will do as I say.' After that what could I do?” (19-year-old, 9th standard educated)

Reasons for abortion

Adolescent married women mentioned their previous child being too young as the most common reason for having an abortion, while the adult women cited not wanting any more children. A small number of adolescent abortions were to delay a first birth; all 14 women who aborted their pregnancies for this reason were either employed, studying in college or doing a vocational training course (Table 4

Table 4 Primary reason for abortion in married adolescent and adult women (%)

	Married adolescents ⩽20 years (n=199)^* (%)	Married adults ⩾20 years (n=1293)^* (%)
Previous child too young^***	53.3	21.2
Delay first birth^***	7.0	0.2
Family size reached, no more children wanted^***	9.5	45.1
Failure of IUD or sterilization^***	nil	3.2
Economic/work reasons	5.5	5.6
To avert birth of a girl	12.6	18.3
Fetal malformation	5.5	1.3
Maternal illness/other reasons	6.6	5.1

* n=abortion episodes.

*** 0.000 significance.

).

Only 13.1% of married adolescent girls who were trying to avoid pregnancy or postpone their next child said they had used a contraceptive at any time between marriage/last pregnancy and the pregnancy they terminated. This was a significantly lower number than the 26.3% of adult women who reported contraceptive use between the last pregnancy and this abortion. Contraceptive failure (becoming pregnant with an IUD in place, after a tubal ligation or while using the pill) was uncommon in both adolescent and adult women.

Twenty-five adolescent women had abortions to avoid the birth of a girl child. Four of them already had one living son. Four others reported having had a sex determination test in a previous pregnancy, two of them had a previous sex selective abortion as well. Other reasons for abortion included fetal malformation and maternal illness (Table 4).

Number of weeks pregnant at time of seeking abortion

The mean number of weeks of pregnancy at which the married adolescents and adult women sought an abortion was almost the same (10.9 vs. 10.8 weeks). But abortion-seeking was delayed to beyond 12 weeks of pregnancy in 52 (26.1%) of the 199 adolescent abortions. The reasons for the delay included sex selective abortions, which are mostly done after an ultrasound test at 10–12 weeks of pregnancy onwards (24 cases), fetal abnormalities detected in the second trimester (5 cases), late recognition of pregnancy conceived during the breastfeeding period (6 cases) and delays because of the need for multiple visits to the provider (8 cases) or in arranging for payment (5 cases). In four other cases, the adolescent sought an abortion in the latter part of the first trimester and was asked by the provider to return later in the pregnancy.

Choice of provider

Almost a third of adolescent women (29.6%) had abortions with providers who had not been trained, significantly more so than older women (22.1%). Untrained providers included paramedical workers, those with training in indigenous systems like ayurveda, as well as generalist doctors without specific obstetrics or gynaecology training or training in doing abortions. The use of public sector services was marginally though not significantly higher among the adolescents than the adult women (Table 5

Table 5 Abortion service providers, married adolescent and adult women (%)

	Married adolescents ⩽20 years (n=199)^* (%)	Married adults ⩾20 years (n=1293)^* (%)
Providers without formal training^**	29.6	22.1
Provider and/or place of service provision not certified	41.3	43.1
Informal/traditional provider	1.5	2.1
Public sector provider	18.6	15.9
Private sector provider	79.9	82.0

* n=abortion episodes.

** 0.025 significance.

).

When asked the reasons and circumstances that led to the choice of provider, cost considerations were the most commonly mentioned by adolescents (29.1%). Other factors included the fact that they were referred by friends and family (25.6%), that the provider was a woman (9.9%), that the abortion procedures were completed in one visit and without the need for an overnight stay (10.5%), that the provider was known to be skilled (18.6%) or that abortion deaths had never occurred with that particular provider (9.9%). Older women mentioned similar reasons, but they mentioned cost considerations significantly less frequently than the adolescent women (12.1%). Adolescent women travelled an average of 10.1 km and adult women an average of 10.9 km to access their abortion service provider.

Cost of abortion

The cost of the abortion depended on the number of weeks of pregnancy at which it was sought and the type of provider, but did not vary by age of the woman. The median costs that women incurred for a first trimester abortion in the private sector were Rs. 490.8 (US$ 10.20), with a range of Rs. 65–2625 (US$ 1.40–54.70). Public sector abortions, though cheaper, were not free of cost; the median costs for a first trimester were Rs. 177.2 (US$3.70), ranging from nil to Rs. 550 (US$11.50). Second trimester abortion costs in the private sector averaged Rs. 1661.4 (US$34.60) and in the public sector Rs. 560.8 (US$11.70).

Experience of abortion

When asked about their experience of having the abortion, fewer adolescents than older women (29.9% vs. 35.8%) reported that the provider had explained to them the abortion procedure they were about to undergo. Providers were more likely to insist on spousal consent from adolescent women (26.1%) than from adult women (21.2%).

While the actual methods used for the abortion were difficult to determine from the women's reports, when asked if the provider had used any pain medication, sedation or anaesthesia during the procedure, 73.3% of adolescents and 70% of adults said that no form of pain management had been provided.

Fewer adolescents than adults reported receiving information on post-abortion contraceptive options (47.9% as compared to 59.3%), and the proportion of adolescent women who adopted a contraceptive method after abortion (36.9%) was lower than of adult women (57.2%). These differences were significant and persisted even after adjusting for whether the abortion took place in the private or public sector.

71.9% of adolescent and 67.6% of adult women reported one or more complaints after the abortion that were severe enough to disrupt their household routine. Excessive bleeding was the most frequently reported problem for 71 adolescents (35.7%) and 457 (35.3%) adults. Even when the bleeding was troublesome, it was usually treated with home remedies, or herbal or indigenous medicines. Only 14 adolescents (19.7%) and 97 adult women (21.2%) who complained of bleeding went to a formal provider to seek care. Other complaints mentioned included lower abdominal pain, high fever or foul smelling discharge and weakness. Incidence of life-threatening morbidity was similar in both age groups, 6.4% of adolescents and 6.7% of adult women had to be hospitalized for serious complications such as septic or incomplete abortions, or uterine perforation. Reported morbidity did not vary significantly by age but was significantly higher among women who had had a second trimester abortion.

Perceptions and actuality of the legality of abortion

Over a quarter of the 197 adolescents (26.4%) believed that abortion was illegal under all circumstances and 42.9% believed that legality depended on the marital status of the women, and that all extra-marital abortions were illegal.

Similarly, although 70.4% of adolescent abortions were done by trained providers, only 58.7% were technically legal under Indian law, in the sense that both the provider and the place where services were provided were certified to do abortions. Even those who were specialists in obstetrics and gynaecology were often performing illegal abortions (Table 5).

The situation for unmarried adolescents

While comparisons between the married and unmarried women interviewed in this study are not possible because different types of interviews were carried out, the qualitative data obtained from interviews with the unmarried adolescents suggests that the pattern of decision-making, provider choice and the abortion experience itself may have been very different among them.

All four of the adolescent women who were separated from their husbands and two of the 12 never-married adolescents interviewed said that their pregnancies occurred as a result of non-consensual sex with their employers. Four other never-married adolescents said that their pregnancies resulted from non-consensual sex with an older male relative (father, uncle). The other six said they had been in love with the man who they became pregnant with. All 16 referred to their pregnancies as being “illegal” or “stolen goods”.

Though the unmarried adolescents came from the same geographic area and had the same physical access to abortion services as the married women, 11 of 28 unmarried women for whom information was available used a traditional/informal provider located in or near their villages. The average distance travelled by these girls in order to access the provider was 4.2 km. Thirteen went to a nearby private provider, an average distance of 5.5 km away. Four went to a public hospital in a nearby town, an average of 14 km away.

Sixteen of these adolescents said that the need to keep the pregnancy a secret from neighbours and community was of paramount importance. Only three mentioned that the concerned partner was involved in helping them to obtain the abortion. All 12 of the never-married adolescents had shared their situation with their mothers, but seven kept it hidden from others in the family, particularly their fathers.

“If he [father] comes to know, he will cut me up. He will kill me.” (16-year-old, unmarried girl).

For the separated young women, family support was more difficult and only one had taken her mother into her confidence. While secrecy was important, most of these women also mentioned that mobility was limited. Any prolonged or repeated visits away from home would have been impossible, limiting their options available for accessing care.

Private practitioners provided them the confidentiality they needed and were a reasonable distance away, but the cost of their services was high and so they were accessed only by those girls with financial support from their parents or brothers. One young girl, who was asked to pay Rs. 1200 (US$25) for a first trimester abortion, was told:

“Yes, I can do it. You can go home in two hours. No one will come to know. But for illegal cases like yours, I am taking a big risk. I have to charge accordingly.” (private provider, male)

While large public hospitals in nearby cities provided services at minimal costs, the costs of travel and the need to make more than one visit were all constraints, as were the judgemental attitudes that the women encountered.

“You village girls are all the same. Illiterate and stupid. Cannot keep yourself out of trouble. Do you not understand anything?” (Male gynaecologist, large city hospital)

Both public and private doctors often insisted on the presence of a guardian to provide consent (Indian law requires a guardian's consent for young women below 18 years of age). However, some girls reported that private practitioners were willing to forego this requirement for a fee.

“I insist on a signature. I tell them, bring your husband, bring your mother. Not that it is really needed. But otherwise everything is too easy for these girls. They will go on doing such things and then coming for abortion.” (Female gynaecologist, large city hospital)

The unmarried adolescents were a mean of 12.7 weeks pregnant when they sought abortion and morbidity reported was high in this group. All 11 girls who went to a traditional provider reported problems, of whom four had life-threatening complications and were eventually taken to large city hospitals. We identified one death in the study area during the study period, an unmarried 16-year-old, whose abortion consisted of a twig insertion done by a traditional provider.

Provider perspectives

More than half of the 159 interviewed abortion providers (55.9%) felt that the number of adolescent women seeking abortions was on the rise compared to five years ago, but most felt that the proportion of those who were unmarried had remained unchanged. Many had reservations about young women having an abortion and 86.2% said they would not do an abortion to delay a first birth.

Providers routinely referred to unmarried adolescents as “illegal” cases because their pregnancies did not occur within the context of marriage. While we did not ask them about the costs of services for unmarried clients, a few providers did mention that unmarried young women are charged three to five times the normal rate. Some 40% selectively refused services to unmarried and separated young women even though they offered them to other women. Nearly a quarter of the providers who did not provide services to unmarried women were from legally recognized MTP centres.

Most public sector providers mentioned that consent from someone other than a young woman was necessary, as per the law. Private providers, though aware of this provision, did not think it a serious barrier, and most were willing to put down the woman's age as 18 even if they suspected it was lower. However many of them did insist on a signature from the husband, as a way of protecting themselves in case of domestic disputes or subsequent questioning by the police.

“Yes, husband's signature is compulsory to confirm that the pregnancy is not illegal. I always ask for the signature of her husband or her in-laws. Especially when she is so young, we are suspicious. Also he may come tomorrow and say, `I wanted a son. Who asked you to do that?' Why should we become involved with their problems?” (Female gynaecologist, owner of private nursing home)

Discussion

Many adolescent women in this study used only abortion as a birth control measure, as other studies in India have found [7][8][9]. In order to tilt the balance towards greater use of contraception, family planning programmes in India must address the needs of newly married couples, move away from the current focus on sterilisation and provide contraceptive counselling to young women seeking abortion services.

It is unlikely that we identified every abortion in the community during the study period. Early uncomplicated abortions among unmarried women were particularly likely to have been missed, and many identified unmarried adolescents were not contacted for an interview for fear of jeopardising their social situation. Thus, generalisability for this sub-group of adolescents may be limited.

However, we can safely say that adolescent sexual activity certainly occurs outside the marital context, and this must be accepted and addressed programmatically by providing contraceptive counselling and services to this group in a sensitive manner. This needs to be coupled with efforts to increase adolescent reproductive and contraceptive knowledge in general.

While small families have become the norm in India, son preference remains strong. Despite the illegality of sex detection tests, this study has shown that sex selective abortions, even among young women early in their reproductive careers, is not uncommon. Further, these abortions are mainly second trimester procedures and as our data show, they may put women at greater risk for morbidity.

Husbands and mothers-in-law played a major role in some 25% of abortion decisions among married adolescents, to the extent of forcing some women to have an abortion against their wishes. For unmarried adolescents, isolation and lack of family support and involvement of chosen sexual partners were negative factors and influenced access to appropriate services as well. Women who are separated or widowed are at even greater risk of facing the abortion process without family support [10]. The role that sexual abuse and coercion play in making unmarried and separated women pregnant, in determining whether a pregnancy is wanted or not and in influencing the way women are able to deal with unwanted pregnancies have all been noted in India [11][12], but need to be explored more fully.

Although the study area, as the rest of Maharashtra state, has better geographic access to both private and public sector abortion services than many other parts of India [13][14][15], inadequacies in the certification process ensure that many technically qualified providers continue to exist in an illegal environment. Streamlining the certification process is needed and must be accompanied by efforts to ensure that certified and legal providers maintain required technical standards and fulfill reporting requirements. The levels of self-reported morbidity among the women in this study were high. While self-reported morbidity may not correlate well with clinical morbidity, the fact that women perceived these problems as severe enough to disrupt their normal functioning, and several required hospitalization, highlights the need for monitoring of sevice delivery for safety and the ability to deal with complications.

The findings of this study also suggest that the existence of services alone is not enough to ensure access to appropriate care. Even married adolescent women, impeded by a lack of autonomy, mobility and control over resources within the household, were more likely to use the services of untrained, less costly providers. At the same time, legally recognised public sector services were not always acceptable to adolescent women, either because of barriers created by provider attitudes or lack of confidentiality [16][17][18].

Issues relating to consent have been mentioned as barrier in other studies as well [10][16]. This is a complex issue, especially for adolescents. Spousal consent is not a legal requirement for married women in India. If a woman is above the age of 18, her consent alone is adequate. However, Indian law requires parental consent for women under age 18. The issue is further complicated by the fact that many such young women are married, although legally the age of marriage is also 18 [19]. There is a compelling need to increase legal awareness among adolescents and providers, to ensure that barriers with no basis in law or policy are removed. However, sexual abuse and coercion and the social stigma of illegitimate pregnancies are still the biggest barriers to the ability to respond to unmarried women's needs.

Given all these barriers, it is not surprising that even in areas where access to safer options exists, as in this study area, unmarried adolescents often turn to informal, untrained providers as the only way to meet their needs for confidentiality and respectful treatment within their limited means and mobility. If the situation is to change, there is clearly a need to find innovative ways to make abortion services more anonymous, and at the same time increase access. One approach may be mainstreaming these services within other medical services.

Service-related reforms alone are unlikely to make a substantial difference, however. An attitudinal shift towards accepting a woman's right to make her own reproductive decisions is also needed.

Acknowledgements

At the time the study was done, both authors were senior research Scientists with KEM Hospital Research Centre. The study was funded by a grant from the Ford Foundation to the KEM Hospital Research Centre, Pune. The authors are grateful to Dr. Banoo Coyaji, Chairman, Dr VN Rao, Director Research and the study team of S Ambike, L Garda, D Ghanwat, R Kale, N Suryawanshi and S Walawalkar from the KEM Hospital Research Centre. Fellowship grants from the MacArthur Foundation for Leadership Development, India and the Gates Institute of Reproductive Health, Department of Population and Family Health, Johns Hopkins School of Public Health, Baltimore, USA, provided the opportunity to write the paper. This paper is based in part on a presentation at the International Conference on Adolescent Reproductive Health, organized by WHO, Mumbai, November 2000.