Elective Abortion as a Primary Health Service in Rural India: Experience with Manual Vacuum Aspiration

Abstract

Abstract

India allows abortion up to 20 weeks of pregnancy but places restrictions on abortion facilities and providers. Abortion services are especially deficient in rural areas. Although vacuum aspiration is safer, sharp curettage continues to be used by providers as they lack relevant training. This paper describes the provision of first trimester abortion services using manual vacuum aspiration (MVA) in a rural clinic in the state of Rajasthan over a four-year period. Non-use of anaesthesia increased safety and allowed women to return early to a normal routine. Of 534 women, none suffered major complications; 16 required repeat evacuation and the procedure failed in two. We recommend that models based on MVA and including medical methods be piloted in rural areas of a number of states of the country, to establish the feasibility of delivering first trimester abortion as a primary health service. Ways to increase access to second trimester procedures are also required. Current requirements for certification of private facilities are excessive and the process of obtaining certification is arduous. The law in India does not extend these norms to government facilities, which remain under-equipped and lack trained providers. We recommend that certification of facilities be liberalised and applied equally to government and private institutions. There is also a need to rapidly increase training capacity across the country.

Résumé

L'Inde applique une politique libérale qui permet d'avorter jusqu'à 20 semaines de grossesse. Néanmoins, les services sont particulièrement déficients dans les zones rurales. Bien que l'aspiration manuelle soit plus sûre, le curetage continue d'être utilisé par des prestataires insuffisamment formés. Cet article décrit les services d'avortement par aspiration manuelle pendant le premier trimestre dans un dispensaire rural de l'Etat du Rajasthan sur une période de 4 ans. Le non-recours à l'anesthésie augmentait la sécurité et permettait aux patientes de reprendre rapidement leurs occupations. Sur 534 femmes, on a enregistré peu de complications et aucune complication grave; 18 avaient eu besoin d'une nouvelle évacuation, 16 pour avortement incomplet et 2 pour échec de l'avortement. Nous recommandons que les modèles basés sur l'aspiration manuelle assortis de méthodes médicales fassent l'objet d'une étude pilote dans les zones rurales de plusieurs Etats, pour décider s'il est possible de réaliser des avortements sûrs pendant le premier trimestre dans le cadre des services de santé primaires. Il faut également trouver des moyens d'accroı̂tre l'accès aux procédures pendant le deuxième trimestre. Nous avons constaté que les exigences actuelles pour l'agrément d'équipements privés sont excessives et le processus de demande est complexe. La Loi sur l'interruption médicale de grossesse n'étend pas ces normes aux institutions gouvernementales, qui demeurent sous-équipées et manquent de prestataires formés. Nous recommandons que l'agrément des installations soit libéralisé et s'applique aux institutions publiques et privées. Il faut également rendre obligatoire la formation à l'avortement par aspiration manuelle et électrique et à l'avortement médical pour tous les prestataires concernés dans les instituts de formation.

Resumen

En la India se permite el aborto por diversas razones hasta 20 semanas del embarazo. Sin embargo, los servicios de aborto son especialmente deficientes en las áreas rurales. Aunque la aspiración endouterina es más segura, los proveedores continúan practicando el curetaje por falta de otro tipo de capacitación. Este artı́culo describe la provisión de servicios de aborto en el primer trimestre usando la aspiración manual endouterina (AMEU) en una clı́nica rural en el estado de Rajastan durante un periodo de 4 años. Al eliminar la anestesia, se disminuyeron los riesgos y las mujeres pudieron volver pronto a una rutina normal. De 534 casos, se presentaron pocas complicaciones y ninguna complicación mayor; en 18 casos hubo que repetir la evacuación, 16 por aborto incompleto y 2 por falla del aborto. Se recomienda realizar modelos pilotos basados en AMEU incluyendo métodos médicos en áreas rurales de varios estados a través del paı́s, con el fin de establecer la factibilidad de proveer servicios de aborto sin riesgos en el primer trimestre como un servicio de salud primaria. Habrı́a que buscar además la forma de ampliar el acceso a intervenciones durante el segundo trimestre. Determinamos que los requisitos de certificación vigentes para los establecimientos privados son excesivos y que el trámite es arduo. El Acta MTP no extiende estas normas a las instituciones públicas, las cuales carecen de equipos y de proveedores capacitados. Recomendamos por lo tanto la liberalización de los requisitos de certificación y que estos se apliquen igualmente a instituciones públicas y privadas. Se precisa además exigir que todos los proveedores indicados reciban capacitación en la provisión de la aspiración endouterina tanto eléctrica como manual.

Key Words:

• induced abortion
• abortion methods
• primary health care
• India

Elective abortion was legalised in India under the Medical Termination of Pregnancy (MTP) Act, 1971 and MTP Rules, 1975 [1]. The Act permits termination of pregnancy in cases of risk of grave injury to the physical or mental health of a pregnant woman and following failure of contraception, up to 20 weeks of pregnancy, but places restrictions on the kinds of providers and facilities that can deliver the service. Abortion can be carried out either in government health facilities or government-approved private facilities, but private institutions must be certified to do so. The certification process requires installation of equipment and supplies for general anaesthesia and abdominal surgery, a guarantee that a trained provider and anaesthetist are available, site inspection and sequential recommendations by district and state health authorities. By contrast, government hospitals and health centres are considered eligible for providing abortion services provided they have a trained doctor in position – the Act and Rules do not extend facility requirements to them. Given the weak public health infrastructure and poor penetration of the private health system in rural areas, most private MTP-certified facilities have come to be located in cities or towns.

The Act allows qualified gynaecologists to perform abortions. Any other doctor wishing to perform abortions must first undergo an in-service course at a designated MTP training centre. In most states, only teaching hospitals carry out MTP training, and their courses are not accessible to doctors in private service. This has severely limited the number of legal abortion providers. Alternative providers, such as physicians trained in Indian systems of medicine and paramedical workers, are barred from carrying out abortions.

As a result of all these restrictions, most rural women have been deprived of access to legal abortion services. Clandestine providers, many of whom employ dangerous techniques, carry out the large majority of induced abortions in rural areas [2], and abortion complications constitute 17.6% of maternal deaths [3].

When the MTP Act was passed, sharp curettage (D&C) was the main technique used for terminating pregnancy in India. This technique requires cervical dilatation, provokes pain that must be countered by anaesthesia or heavy sedation, and carries a finite risk of perforation because a sharp, metallic instrument is used to scrape the wall of the uterus. Abortion by sharp curettage should ideally only be carried out by surgically experienced doctors operating in fully equipped facilities with provision for emergency abdominal surgery, should a complication arise. These limitations too have impeded access to safe abortion through India's primary health system.

Newer and simpler technology – electrical and manual vacuum aspiration – became available and widely used in the 1970s [4] and also met with a positive response in India [5]. A study of 80,437 abortion procedures in the USA from 1971–1975 revealed that total and major complication rates associated with vacuum aspiration were 5% and 0.4% respectively, compared to 10.6% and 0.9% for sharp curettage [4]. An analysis of 170,000 outpatient vacuum aspiration procedures between 1971 and 1987 found 0.9% complications, of which only 0.07% required hospitalisation, with no deaths. Vacuum aspiration was as effective and much safer than sharp curettage. Medical abortion (mifepristone with misoprostol or gemeprost for abortions up to nine weeks) has been successfully introduced since the late 1980s in a growing number of developed countries [6][7] but only in 2002 in India.

Vacuum aspiration requires less dilatation and is quicker and less painful, as compared to sharp curettage. Being independent of electricity, manual vacuum aspiration is a better choice for rural Indian settings. Sharp curettage, however, continues to be the most widely used abortion technique in rural and small urban health facilities. A study of MTP services and interviews with 479 providers in four major Indian states [8][9][10][11] revealed that sharp curettage was still the preferred method among 64% of providers, while only 29% preferred electric vacuum aspiration and a mere 7% preferred MVA. The same study revealed that only 25% of 187 graduate doctors had got the opportunity to carry out MVAs during their in-service MTP training course, probably because the faculty of teaching hospitals did not prefer it. Prominent sections of the medical and gynaecologic fraternity have voiced scepticism at conferences and policy meetings about the practicality of using MVA. Based on their experience in urban tertiary centres, they have judged MVA to be clumsy, messy and cumbersome when used for abortions above eight weeks, as compared to electrical suction. Even the Government of India “Manual for First Trimester Medical Termination of Pregnancy” relegates MVA to the category of “menstrual regulation”, to be performed only up to 14 days after a missed period [12]. Yet it has been shown that MVA is an effective technique for pregnancy termination at least up to 12 weeks [4].

There are few documented experiences of providing abortion services as part of primary health care in rural settings where services are most deficient. This paper describes an intervention to provide first trimester abortion using manual vacuum aspiration in an interior rural area of Rajasthan, India's largest state [13].

The poor status of women's health in Rajasthan is reflected in a maternal mortality ratio of 670 per 100,000 live births [14], a total fertility rate of 3.78 and contraceptive prevalence of only 40%, most of which is due to female sterilisation [15]. There are 263 community health centres and 1662 primary health centres in the government sector [16] that are meant to provide abortion services, but most lack a trained doctor or equipment [17]. Although the state has certified 482 private facilities to provide abortion services, only 44,265 pregnancy terminations were reported to the government during 1997–1998 [3]. This suggests an extremely low caseload per certified institution, probably because they are concentrated in cities and towns, leaving rural areas underserved. Estimates suggest that some two to ten unreported procedures are carried out for each reported one [2][18].

Action Research and Training for Health (ARTH) is a non-profit, non-governmental organisation implementing a health programme in a rural community of 36,000 in southern Rajasthan. Scattered habitation, poor road connections and widespread poverty characterise the villages of this area. The female literacy rate (17.4%) is especially low [19], while conservative social norms restrict women's autonomy and control over household resources. A survey of 12 villages in the field area in 2000 revealed a contraceptive prevalence of only 22% [19]. Kumbhalgarh block (estimated population 135,000), in which ARTH's clinic is located, has seven government health centres. Until mid-2001, none of them provided legal abortion services as no trained physician was available.

We had learnt from focus group discussions with women in the community in 1998 that women try a variety of remedies to deal with unwanted pregnancy, including tablets, decoctions and visits to unsafe providers. Instances of women dying after clandestine abortions were well known. Most women, however, could not afford to go to the city for a safe abortion. Survey results further revealed that 79% of married women and 76% of married men were not aware that abortion was legal.

ARTH's clinic

In 1997, ARTH established a clinic in five rooms of a rented building in a road-linked village, 55 km from Udaipur city. The clinic provides a range of reproductive and child health services, including reversible contraceptive methods (condoms, pills, IUDs and injectables), pregnancy and postnatal care, delivery and obstetric first aid, and management of reproductive tract infections and other gynaecological problems. A gynaecologist visits the clinic on two fixed days each week, while pairs of nurse–midwives and clinic attendants are available round the clock. The clinic is equipped with equipment and supplies, as prescribed by the MTP Act and Rules, for carrying out legal abortions. The premises lack piped water and electric supply is irregular, however, especially during the summer months. Three local jeep-taxis have been contracted to provide quick transport to the city in case of emergency.

The gynaecologist could not commence providing abortions until the clinic was certified. This took two years starting in November 1997. First, it took a few months to arrange funds to procure statutory equipment for abdominal surgery (a shadowless overhead lamp and instruments for laparotomy) and anaesthesia (Boyle's apparatus, oxygen cylinders and resuscitation kits). It took another few months to locate an anaesthetist willing to be affiliated to the clinic. We then applied to the chief district medical and health officer and arranged for an inspection. Our application was returned twice for re-submission with additional documents which had not been listed in the application guide. These included an architectural drawing of the facility and an affidavit from the abortion provider and anaesthetist confirming their availability. At various stages our application lay dormant, till persuasive reminders at the state capital made the relevant file move from desk to desk. The technical head of the state health department then recommended certification, on the basis of which the government administrative authority gave approval and issued a certificate in December 1999. From then onwards we started providing abortion services, initially up to 10 weeks of pregnancy and once we gained experience and confidence, up to 12 weeks.

Abortion services and methods

We tested a service model wherein the visiting physician provided abortion services on a bi-weekly basis. Locally resident nurse–midwives were available to counsel and screen women with unwanted pregnancy, provide follow-up care, and stabilise and refer women with complications. They also sterilised MTP instrument kits well in advance whenever electricity was available.

Some women who came to the clinic were anxious about having missed a period, because they wished to limit or space their next birth. Others had prolonged lactational amenorrhoea or were close to menopause. A few were adolescents, widows or separated women. Before certification, women up to 14 days of a missed period were counselled and offered menstrual regulation without confirming pregnancy, while after certification, all women were offered pregnancy confirmation. Of those with a missed period or unwanted pregnancy (705), we recorded basic socio-demographic data and in the fourth year of providing services, we started recording additional data on reproductive intentions, accompanying persons and abortion techniques (236 women).

There is considerable social distance between rural women and service providers in our field area. Health personnel tend to look down upon illiterate women, who they see as bound by irrational beliefs and tradition. Women on their part lack confidence in dealing with well-educated providers. Our clinic staff made efforts to establish rapport in a warm and non-judgemental manner. The counselling process helped women to decide whether or not to have an abortion, informed them about likely discomfort during the procedure, possible side effects and complications, return of fertility and contraceptive options. Counselling was non-directive, and the decision to terminate was the woman's. However, in a few instances where the woman appeared to be undecided or was under family pressure, we questioned the initial decision to abort till she could make up her mind in privacy.

In the first two years, we insisted upon the presence of an attendant (any adult) for procedures beyond nine weeks. The logic was that the attendant could accompany the woman home after the procedure, since this generally involved walking 1–3 km after a bus journey, and in case of emergency could give consent to transferring the woman to the city. However, some women were unable to come with an attendant – they had not informed anyone else. Our records showed that 27% women had come alone, and another 21% were accompanied by a health worker rather than a relative or neighbour. Moreover, there were no emergencies. We therefore relaxed this requirement to improve access for the women.

Feedback from women revealed that the issue of confidentiality weighed heavily on their minds in relation to seeking abortion. Women employed different methods to camouflage their clinic visit – some went to their parents' home and visited the clinic from there; others attributed the visit to general ailments or brought children along for immunisation. Our clinic staff maintained complete auditory and visual privacy during counselling and clinical procedures, asked accompanying persons to wait outside the consultation room, and used camouflage terms while referring to abortion procedures in front of others. Informed consent was taken only from women themselves. Being illiterate, most could only affix a thumb impression. In virtually all cases, the procedure was done on the same day as counselling. Only when women arrived late in the afternoon or the clinic workload was exceptionally heavy did it have to be postponed till the doctor's next visit.

We opted for MVA since it is not dependent on electricity (and we did not have a generator). For pain control, we gave a mild analgesic and anxiolytic drugs (ibuprofen and diazepam tablets), plus the reassurance provided by careful pre-procedure counselling. Women remained awake throughout the procedure while the team (physician, nurse–midwife and clinic attendant) constantly engaged them in light and friendly conversation about their family, work routine, etc. While we have not systematically evaluated the level of pain experienced by women, most reported mild or no pain. From the third year onwards, we started using a prostaglandin (carboprost tromethamine, 0.125–0.25 mg intramuscularly) to prime the cervix for pregnancies over eight weeks. Carboprost was the only available prostaglandin. We observed standard infection prevention practices. Clean water from a hand pump or tube well, stored in a plastic bucket with tap, was used for hand washing. Whenever there was no electricity, a battery operated emergency lamp along with a powerful, hand-held flashlight served as light source.

After the procedure, women were instructed to take oral antibiotics and iron and advised that they could resume all routine activities immediately but to avoid heavy manual labour for 2–3 days. They were informed about danger signs and advised to return promptly in case these arose. Women left the clinic about an hour after the procedure. We scheduled routine follow-up visits for a week later and informed women that nurse–midwives were available round the clock, to handle any emergency.

To recover a fraction of the costs, we charged a subsidised amount ranging from Rs. 150 (US$3.20) for menstrual regulation to Rs. 300 ($6.40) for a 12-week abortion. These charges were all-inclusive and a quarter of those levied by private facilities in the city. Several women, mainly those belonging to the tribal community, got a further subsidy or credit. For women with complications of spontaneous or induced abortion carried out by any provider, we arranged for free treatment and transport to hospital, as required.

After receiving certification, we began actively promoting the service through educational sessions organised by 40 women volunteers and four extension workers. We introduced an illustrated pamphlet in the local dialect, featuring a recent news clipping reporting the death of a woman in the district following clandestine abortion by an unqualified rural practitioner. The pamphlet conveys the following messages:

•	If a woman has an unwanted pregnancy, she should not attempt dangerous methods such as sticks, injections or tablets – they might disable her or even result in death.
•	Safe and legal abortion services for up to three months' pregnancy are available at the ARTH Health Centre.
•	A woman must come as early as possible after missing a period – delaying the visit might necessitate a more expensive abortion in the city.
•	ARTH's clinic offers complete privacy and confidentiality, along with contraceptive options for those who wish to delay or avoid the next pregnancy.
•	The cost of an abortion is Rs. 150–300.

The pamphlet is targeted at men and other extension workers, since women's low literacy levels would limit its accessibility to them. In the prevailing social environment, the pamphlet strategically portrays safe abortion as a health need rather than a woman's right. We felt that rights issues such as absence of spousal consent and access for women out of wedlock could be better conveyed in a discreet manner through extension workers.

Profile of women presenting with unwanted pregnancy

Between November 1997 and March 2002, 705 women came to our clinic with a missed period or unwanted pregnancy. Most were in their late 20s or 30s, 96% were married, and over two-thirds had three or more living children (Table 1

Table 1 Profile of women with missed period or unwanted pregnancy, November 1997–March 2002

Profile	% (n=705)
Age (years)
15–19	6.2
20–29	43.3
30–39	45.1
40–49	5.4
Marital status
Currently married	96.3
Never married	2.0
Separated/widowed	1.7
Living children
Nil	6.8
1–2	25.5
3–4	47.0
5+	20.7

). Most (88%) were less than 12 weeks pregnant. Those who presented after 12 weeks were more likely to be adolescents or pregnant out of wedlock, but the difference was not significant (data not shown). Data on reproductive intentions suggests that 65% did not desire any more children, 21% wished to delay the next pregnancy and 14% were either pregnant out of wedlock or were uncertain of future intentions.

Before the clinic was certified, more women (47.7%) were refused a termination because they came more than 14 days after a missed period. After MTP certification, 20.1% were refused, the commonest reason being pregnancy beyond 12 weeks (Table 2

Table 2 Proportion of women having menstrual regulation/MTP before and after certification

Period	No. of women presenting with missed period or unwanted pregnancy	No. receiving menstrual regulation or MTP		% having menstrual regulation or MTP
		Number	Weeks of pregnancy: mean (range)
Before MTP certification	107	56	5.7 (4–6)	52.3
After MTP certification	596	476	8.1 (4–12)	79.9

). A few women came when the physician was not available and did not return subsequently, while six had medical risk factors. We referred women beyond the first trimester or with risk factors to one of three designated facilities in Udaipur city for an abortion.

After July 2000, we began recording information on women's attempts to terminate the pregnancy before coming to the clinic. Of 495 successive women, 158 (32%) had tried what they thought was a simpler remedy, including tablets (65%), decoctions (33%), injections (8%), vaginal interference (1.3%) or IUD insertion (1.3%). None of them had suffered complications. Twenty other women presented with post-abortal complications. We advised referral care for eight of them; only three complied. The remaining 17 were managed at our clinic by uterine evacuation, oxytocin drip and antibiotics.

Before visiting the clinic, women often asked our volunteers about the pain likely to arise from an abortion procedure, and whether a man or woman would carry it out. Others were curious about the technique – would it be carried out by hand or by machine? Women seemed to prefer abortion “by machine”; they thought that doing it by hand was likely to hurt more. Afterwards, however, most women felt relieved and said they had not felt much pain. Others felt shy about having exposed themselves. They also said they found it to be a minor procedure; they were able to resume work the same day.

Safety and effectiveness of MVA as a first trimester abortion procedure

The mean time for the entire MVA procedure, from insertion to removal of the vaginal speculum, was 8.8 min (range 4–19 min). To render MVA safer and more effective we used certain support measures, more often for pregnancies above eight weeks. These included a prostaglandin for priming the cervix, reducing blood loss and minimising cervical and uterine injuries [20]. We collected data on these measures for 156 women (Table 3

Uterine evacuation support measures by weeks of pregnancy (n=156)

Procedure	Pregnancy <8 weeks (n=86)	Pregnancy 9–12 weeks (n=70)	Total
Cervical priming with prostaglandin	9.3%	85.7%	43.6%
Use of dilators	8.1%	74.3%	37.8%
Use of double valve syringe	5.8%	77.1%	37.8%
Use of ovum forceps	3.5%	58.6%	28.4%

). Dilators were required for only 38% of women, which reduced the need for pain control. None of the women required local or general anaesthesia. An ovum forceps was required for removing tissue during 28% of procedures. We avoided curettage to confirm that the procedure was complete, so as to minimise the need for dilatation and risk of perforation.

We were able to follow up just under half (49%) of the 534 women who had an abortion in our clinic. There were no instances of perforation or cervical injury. Sixteen women (3.0%) required repeat evacuation for retained products of conception; all but one of these were detected on routine follow-up and managed at the clinic itself. One woman reported at night with excess bleeding and was transported to hospital, where she underwent a repeat evacuation but did not require blood transfusion. This occurred before we started using prostaglandin. The MVA procedure failed in two women who were about five weeks pregnant; one of them underwent a repeat evacuation at our clinic, the other decided to continue the pregnancy. Women have easy access to the clinic and follow-up is free of cost. Our outreach staff did not report any complications among women in the community. Hence, although we cannot generalise from these data, those who did not come for follow-up may not have differed substantially from those who did.

Outcome of refusal of abortion at ARTH's clinic

Ongoing outreach services by our nurse–midwives include home visits to women resident in our project area who do not come for follow-up after an abortion or referral to the city. They provide contraceptive counselling during these discreet contacts, camouflaged among home visits for other purposes, and use the opportunity to find out what women refused an abortion at our clinic do about their pregnancy. Women from our project area who had visited our clinic at least three months before were visited in this manner. Only 15 of 85 had visited the city abortion facility as referred; over half had continued the pregnancy (Table 4

Outcome of referral of women refused abortion at ARTH clinic (n =85)

Outcome of referral for abortion at hospital in Udaipur	No. of women
Continued with the pregnancy	44
Underwent MTP in the city as referred	15
Underwent clandestine abortion	4
Reported spontaneous abortion	8
Denied visiting clinic/was untraceable	14

). Some who reported spontaneous abortion may in fact have resorted to a clandestine procedure. The main reason why they did not go to the city for an abortion was that they had not informed their family they were seeking an abortion. To travel to the city, a woman requires an escort and large sums of money, and must leave children and cattle unattended for an entire day. Family and neighbours would inevitably notice this.

Post-abortion contraception

Although most women were already aware of at least one contraceptive option, only 27% had actually used a contraceptive before seeking abortion. It had taken the event of an abortion for them to commence using a method. Two-thirds of the 534 women decided to use a contraceptive after the procedure – Copper-T IUD (48%), DMPA (12%), combined oral pills (6%) and condoms (1.5%).

A few women appeared ill at ease while contraception was being discussed and said they would prefer to return for another abortion if required. Some women perceived little risk because their husbands visited them only occasionally from the city. Others were of the view that since they had not conceived for 3–4 years after a previous childbirth, they would not be at risk for as many years after abortion. We had a difficult time convincing them that the truth was to the contrary.

Discussion and recommendations

Our experience suggests that it is feasible to improve access to first trimester abortion in interior rural areas by relying on MVA and that such facilities could also offer first-line treatment for complications of abortion. We improvised arrangements for light and water without compromising on asepsis or technical quality. Avoidance of heavy sedation and anaesthesia allowed for early discharge, well before the visiting physician left the facility. Meanwhile, the required anaesthesia trolley and equipment for major abdominal surgery have remained unused, occupying vital space in the small operating room.

In India, use of general anaesthesia for first trimester abortions continues, albeit for a minority of procedures. In 1999–2000, of 62,032 first trimester procedures carried out by the Marie Stopes clinics using MVA, general anaesthesia was used for 8% up to eight weeks and 15% from 9–12 weeks [21]. Our discussions with the faculty of several medical colleges confirm the impression that general anaesthesia or heavy sedation with first trimester abortions is not uncommon. Our experience, however, is that it is feasible to avoid completely use of anaesthesia or heavy sedation for first trimester abortions.

The MTP Rules and Regulations prescribe detailed measures for concealing women's identity in clinic case records and confidential reports to local health authorities. We instituted additional measures to help to ensure confidentiality in a rural setting. The very availability of low-cost services in the vicinity reduced the duration of women's absence from home, while obviating the need for an expensive visit to the city in most cases. The availability of a range of services for women and children along with discreet provider behaviour helped to camouflage women's need for an abortion. Non-use of anaesthesia or heavy sedation followed by early discharge meant that women were not groggy and could quickly resume work once they got home without families noticing anything.

Recognising the barriers limiting women's access to safe abortion in spite of a liberal law, India's National Population Policy 2000 gives priority to expanding the availability of abortion services through rural primary and community health centres [22]. It also recommends the adoption of MVA and medical methods, while calling for the elimination of cumbersome procedures for registration of abortion clinics. The policy has been farsighted in recommending that practical training in providing abortions should include all medical undergraduates and even mid-level providers. It further recommends establishment of additional training facilities and the involvement of private and non-governmental sectors in service delivery and training.

Our experience of obtaining a certificate to perform legal abortions as a time-consuming, energy-draining process is common to private abortion providers in India. A questionnaire on MTP registration sent out by the Federation of Obstetric and Gynaecological Societies of India (FOGSI) found that 13% of 118 centres were registered after delays of one to seven years, 44% were mired in the certification process, while12% were not even aware of the need for registration [23]. In a consultation with a group of public and private abortion providers in the state of Maharashtra, participants disclosed that they were either ill-informed or felt overwhelmed by the formalities involved in applying for MTP certification [24]. Only if the application process is simplified and made time-bound can access to legal services be increased in a timely way.

In the context of providing early abortion services in rural areas using vacuum aspiration, current certification requirements are excessive. A major hurdle in establishing a service is the requirement for anaesthesia and major surgical equipment [25]. A complete set of such equipment costs at least Rs. 150,000 ($3300), a punitive amount which adds to the cost of service delivery. In the unlikely event of a major complication occurring in a rural facility, stabilising the patient followed by prompt transport to a referral hospital would be far more viable than attempting to manage it in-house. Lastly, even with the equipment, the lack of qualified anaesthetists in interior areas would not be surmounted.

Paradoxically, the MTP Act and Rules do not mandate that government institutions meet these same stringent criteria [1]. In the absence of national guidelines on approval of government institutions, most primary and community health centres, as well as many hospitals, lack MTP providers or remain under-equipped. A facility survey sponsored by the Ministry of Health and Family Welfare in 221 districts throughout the country revealed that only 34% and 28% of community health centres had Boyle's apparatus and standard surgical kits respectively, and only 10% had an anaesthetist. Only 13% of primary health centres had a doctor trained to perform abortions, while only 16% were equipped with a vacuum aspirator [17]. Apart from simplifying facility and provider norms for certification, it is therefore essential that government health institutions be held to the same technical standards for approval as are required of private institutions.

The Government of India is planning to introduce abortion services using MVA in primary health centres for women up to eight weeks of pregnancy [26]. Our experience shows that this measure, while welcome, will fall far short of providing an adequate first trimester abortion service and will restrict access for women. Had we continued after certification to restrict our own services to women up to eight weeks, an additional 36% of those seeking abortions would have been refused. As it was, one in five women were still sent away because they were over 12 weeks of pregnancy and most were unable to access legal abortions in the city. This is a relatively invisible problem that requires resolution at local level in order to reduce unwanted births and complications from clandestine procedures. Hence PHC doctors should be equipped with skills to perform MVA till 12 weeks. We would further recommend a pilot project in a rural setting using vacuum aspiration up to 14 weeks of pregnancy, and possibly also medical abortion from 14 to 20 weeks, to try and address this problem.

There is now also evidence from western India that it is feasible to provide medical abortion through the rural primary health system [27]. We therefore also recommend that models based on MVA and medical methods be piloted in rural areas of a few states of the country, to establish the feasibility of delivering safe abortion as a primary health service, including periodic services for interior areas where doctors are not regularly available. The results can then be used to guide the development of more liberal and equitable certification norms for abortion facilities and providers. At the same time there is a need to increase training capacity across both public and private sectors, and to make the teaching of the most up-to-date abortion techniques a requirement for providers within all medical training institutions.

Acknowledgements

Interventions described in this paper were supported initially by the John D and Catherine T MacArthur Foundation, New Delhi, through a fellowship grant to the first author, and subsequently by an institutional grant from Sir Ratan Tata Trust, Mumbai.