Unsafe abortion after legalisation in Nepal: a cross-sectional study of women presenting to hospitals


  • CH Rocca,

    Corresponding author
    • Bixby Center for Global Reproductive Health, Department of Obstetrics, Gynecology and Reproductive Sciences, School of Medicine, University of California, San Francisco, CA, USA
    Search for more papers by this author
  • M Puri,

    1. Center for Research on Environment Health & Population Activities (CREHPA), Kusunti, Lalitpur, Kathmandu, Nepal
    Search for more papers by this author
  • B Dulal,

    1. Center for Research on Environment Health & Population Activities (CREHPA), Kusunti, Lalitpur, Kathmandu, Nepal
    Search for more papers by this author
  • L Bajracharya,

    1. Capital Hospital Pvt. Ltd, Putalisadak, Kathmandu, Nepal
    2. Paropakar Maternity and Women's Hospital, Thapathali, Kathmandu, Nepal
    Search for more papers by this author
  • CC Harper,

    1. Bixby Center for Global Reproductive Health, Department of Obstetrics, Gynecology and Reproductive Sciences, School of Medicine, University of California, San Francisco, CA, USA
    Search for more papers by this author
  • M Blum,

    1. Bixby Center for Global Reproductive Health, Department of Obstetrics, Gynecology and Reproductive Sciences, School of Medicine, University of California, San Francisco, CA, USA
    Search for more papers by this author
  • JT Henderson

    1. Bixby Center for Global Reproductive Health, Department of Obstetrics, Gynecology and Reproductive Sciences, School of Medicine, University of California, San Francisco, CA, USA
    Search for more papers by this author

Correspondence: Dr C Rocca, Bixby Center for Global Reproductive Health, Department of Obstetrics, Gynecology and Reproductive Sciences, School of Medicine, University of California, San Francisco, 3333 California Street, Suite 335, Box 0744, San Francisco, CA 94143-0744, USA. Email roccac@obgyn.ucsf.edu



To investigate abortion practices of Nepali women requiring postabortion care.


Cross-sectional study.


Four tertiary-care hospitals in urban and rural Nepal.


A total of 527 women presenting with complications from induced abortion in 2010.


Women completed questionnaires on their awareness of the legal status of abortion and their abortion-seeking experiences. The method of induction and whether the abortion was obtained from an uncertified source was documented. Multivariable logistic regression was used to identify associated factors.

Main outcome measures

Induction method; uncertified abortion source.


In all, 234 (44%) women were aware that abortion was legal in Nepal. Medically induced abortion was used by 359 (68%) women and, of these, 343 (89%) took unsafe, ineffective or unknown substances. Compared with women undergoing surgical abortion, women who had medical abortion were more likely to have obtained information from pharmacists (161/359, 45% versus 11/168, 7%, adjusted odds ratio [aOR] 8.1, 95% confidence interval 4.1–16.0) and to have informed no one about the abortion (28/359, 8% versus 3/168, 2%, aOR 5.5, 95% CI 1.1–26.9). Overall, 291 (81%) medical abortions and 50 (30%) surgical abortions were obtained from uncertified sources; these women were less likely to know that abortion was legal (122/341, 36% versus 112/186, 60%, aOR 0.4, 95% CI 0.2–0.7) and more likely to choose a method because it was available nearby (209/341, 61% versus 62/186, 33%, aOR 2.5, 95% CI 1.5–4.3), compared with women accessing certified sources.


Among women presenting to hospitals in Nepal with complications following induced abortion of pregnancy, the majority had undergone medically induced abortions using unknown substances acquired from uncertified sources. Women using medications and those accessing uncertified providers were less aware that abortion is now legal in Nepal. These findings highlight the need for continued improvements in the provision and awareness of abortion services in Nepal.


Abortion was legalised in Nepal in 2002, and comprehensive services were made available in 2004. Formerly a criminal offence, abortion is now legal for any reason up to 12 weeks of gestation.[1] The contribution of illegal abortion to high maternal mortality (539 deaths/100 000 births, 1996)[2] was instrumental in the effort to legalise abortion.[3, 4] Available data indicate both a decline in maternal mortality between 1995 and 2010[2, 5, 6] and a decline in the severity of abortion-related complications at hospitals from 2001 to 2010.[7] However, more data are needed to definitively discover the health effects of abortion legalisation.

The introduction of legal abortion in Nepal has been characterised by a committed effort by the Ministry of Health and Population (MOHP) and private organisations to increase access to safe abortion.[8] By 2011, 500 facilities were providing aspiration abortion services, and over 1300 service providers had been trained.[8] A dedicated mifepristone/misoprostol product, Medabon® (Sun Pharmaceutical Industries, Ltd, Mumbai, India), was approved in 2008. Despite these efforts, however, Nepali women continue to experience induced abortion complications and require postabortion care.[1, 7] The few studies available suggest that significant barriers to care still exist, including lack of awareness of the availability or location of services, transportation problems and gender norms that hinder women's autonomy.[9, 10] Although the established fee for a public sector abortion is about $14, with free services for poor or marginalised women, in practice, costs vary and can be prohibitive.[8] In Nepal's cultural context, abortion may be considered a result of immoral behaviour and associated with shame and stigma.[9, 10] Furthermore, unqualified providers and referral agents can lead women to obtain unsafe or ineffective procedures or medications.[9] Under current law, medical abortion is available only from trained clinicians at certified facilities. However, recent evidence suggests that abortion medications can be obtained illegally from medical shops and pharmacies;[11, 12] no data to our knowledge are available on the medications and information being provided.

To develop effective interventions supporting access to safe abortion, existing reasons for unsafe care need to be identified. Little is known about women's unsafe abortion practices following legalisation, including the contexts in which women seek care from untrained providers and the abortion methods that bring women to postabortion care facilities. We completed clinical case reports and interviews with 527 women presenting with induced abortion complications at major hospitals. We assessed abortion law knowledge and care-seeking experiences and examined factors associated with abortion method and obtaining an abortion from an uncertified source.


Study population and procedures

This cross-sectional study was conducted in 2010, as a part of a large-scale medical chart review study examining the effect of abortion legalisation in Nepal on maternal morbidity from 2001 to 2010. The chart review study examined the number and severity of abortion complications among women treated at four major public hospitals.[7] The current study was performed among women presenting at the same hospitals with complications from induced abortion over the last year of the chart review study (December 2009–2010). Sites were selected because they are major public tertiary care centres in Nepal and serve substantial numbers of postabortion patients. Paropakar Maternity and Women's Hospital (Kathmandu) is Nepal's largest public maternity hospital and receives maternal care referrals from throughout the country. Tribhuvan University Teaching Hospital (Kathmandu), Nepal's largest academic public hospital, also receives nationwide referrals of women with serious gynaecological problems. Lumbini Zonal Hospital (Butwal) and Bharatpur District Hospital (Chitwan), located in the populous Terai region, serve as important regional referral centres for surrounding districts.

Potential participants were identified by trained clinicians involved in care of women at hospital admission or during treatment; research assistants also reviewed hospital registers and visited relevant hospital wards daily to capture all postabortion care patients. Clinicians completed a standardised clinical case report form for all women admitted for postabortion care services. The form included information on clinical symptoms, gestational age, treatment required and outcome. Women who self-reported induced abortion or had clinical signs of abortion that clinicians determined to be ‘certainly’ or ‘probably’ induced, using criteria similar to those used in previous research,[13, 14] were referred to a trained female research assistant. Women had to be at least 16 years old and have pregnancies with an estimated gestational age of ≤28 weeks to participate.

We anticipated enrolling approximately 500 women, based on estimates of the number of eligible women who would present for care at the four recruitment sites in 1 year. In total, 563 of 3826 (15%) postabortion patients who were eligible by age and gestational age were identified as cases of induced abortion. Of the 563 eligible women, 527 participated (94%); reasons for nonparticipation were that the woman was not healthy enough to participate (n = 14), the attending clinician did not refer the woman to the study (n = 9); the woman left the facility before being formally discharged (n = 8); or that the woman refused (n = 5).

Eligible women were approached and invited to participate in the study after being cleared for discharge by the attending clinician. Research assistants explained the study procedures and conducted verbal informed consent using a standardised script. Participants were offered a handout with contact information for the local study investigators and institutional review board. Because of the sensitive nature of the study topic, interviews were conducted in a private location in the hospital, with a focus on maintaining participant confidentiality. Interviewer-administered questionnaires included open and closed-ended questions, with items on sociodemographics, pregnancy and abortion history, barriers to obtaining abortion services, and details of women's abortion experiences, including methods used, location of abortion and who conducted the abortion. Responses were recorded by interviewers on paper questionnaires; data were entered into a database in the study headquarters by trained study staff using CSPro (Washington, DC, USA). Open-ended responses were translated into English by a research coordinator and the local PI (BD and MP), entered into the database, and coded based on a coding manual developed from the data.

All data were kept confidential: surveys were assigned unique identification numbers and did not contain any personal identifying information. Participants were offered a wrapped soap and hand towel in remuneration for their time. Study protocols were approved by the Committee on Human Research at the University of California, San Francisco and the Nepal Health Research Council, Kathmandu.

Data collection and outcome measures

Induction complications

Interview data were paired with clinical case report data documented by clinicians overseeing the care of postabortion patients. Case reports included detailed information on morbidities at hospital presentation. We created a dichotomous variable, based on previous work on severity measurement,[7, 15, 16] indicating whether the participant presented with a serious complication. A serious complication was one characterised by signs of infection or injury, including fever >38.9°C (>102°F), pelvic infection or endometritis, peritonitis, septicaemia, septic shock, foreign body in the uterus, perforation and other injuries to the reproductive or gastrointestinal tracts.

Abortion law knowledge

Abortion law knowledge was assessed with questions about whether abortion is legal in Nepal and whether a woman needs permission from her husband to have an abortion (yes, no, don't know). These items were coded as correct or incorrect: abortion is legal (yes versus no/don't know) and a husband's permission is needed (no versus yes/don't know).

Abortion-seeking experiences

A series of survey questions assessed women's experiences seeking abortion. Participants were asked their reasons for seeking abortion, who they told that they were getting an abortion, from whom they found out where to get an abortion, sources of comfort and support around the abortion, and reason for selecting the source of abortion they did. Participants could provide multiple responses to each question. Women were also asked whether they had felt pressure to have the abortion.

Abortion induction methods

Women were asked how many times they attempted abortion before coming to the hospital and about all of the induction methods they used. The term ‘medication’ refers to any substance or drug that was taken orally or inserted vaginally, and ‘instrumentation’ refers to insertion of instruments into the vagina, including aspiration, dilatation and curettage, or objects. We created a variable for all methods used (instrumentation only, oral medication only, vaginal medication only, both oral and vaginal medication, both instrumentation and medication). For regression analyses, we used a primary induction method variable: medication or instrumentation. Women who used medications at the same time as instrumentation were coded as instrumentation, as was a participant who used heavy massage. Women reporting multiple induction attempts were coded according to their first attempt.

For women whose primary method was a medication, we asked about which medications were used. Mifepristone or misoprostol products were categorised as ‘safe and effective when used correctly’; ayurvedic or drugs for other indications were considered ‘unsafe or ineffective’; and women who could not name the medication taken were categorised as ‘unknown’. For women undergoing instrumentation, we use clinicians' assessments of type of instrumentation procedure recorded on postabortion care unit clinical case reports (aspiration, dilatation and curettage/extraction, other).

Certified and uncertified abortion sources

We created a dichotomous variable of whether sources of abortion were certified versus uncertified or likely uncertified using data from questions about where and from whom women obtained abortions. For instance, abortions provided at hospitals and clinics known to be government certified, based on an official MOHP listing of government-approved facilities, were coded as certified. Abortions provided at private clinics, nursing homes or health posts not listed on the MOHP list were considered uncertified. Abortions provided by uncertified providers, including pharmacists, relatives and traditional healers, were also considered to be uncertified. To categorise women who induced at home using medications, we used data from open-ended descriptions of how women tried to end their pregnancies, classifying as certified those who obtained medications from certified facilities. The following types of cases were considered uncertified: medications were bought directly from pharmacists or medical shops by women or other persons; unapproved medications were reportedly used; medications were obtained from non-certified health facilities; or home remedies for abortion were used. We also created variables describing whether women induced at home (yes, no) and who performed the abortion: the woman herself; a doctor or nurse; or another person such as a pharmacist, friend or traditional healer.


Participants provided information on their age (years), age at marriage (years), parity (0, 1, 2, 3+), education (≤primary education, >primary education), and whether they were currently married or employed. We assessed socio-economic status using seven items about whether participants' households had amenities such as electricity, radio or motorcycle/moped; principal components analysis was used to create a standardised scale (Cronbach's α = 0.65). We asked whether participants' current husband or partner had ever hit or beaten them to assess physical domestic violence. Gestational age at hospital presentation (weeks) was obtained from clinical case reports; we used a categorical version (<9 weeks, 9–12 weeks, >12 weeks). We included a variable for hospital site.


We described participant characteristics and clinical complications at hospital presentation. We also described knowledge of legal abortion and abortion-seeking experiences, including sources of advice and support, who women told about the abortion, and why the chosen abortion provider or source was selected. We used Student's t tests and chi-square tests to examine associations between sociodemographic variables and care-seeking experiences and knowledge of abortion legality.

We described all induction methods used. We then used chi-square tests and multivariable logistic regression to assess differences in abortion-seeking experiences and sociodemographic characteristics of women using medication versus instrumentation as a primary induction method. Similarly, we assessed the proportion of women obtaining abortions from certified and uncertified sources and used multivariable logistic regression to assess abortion-seeking and sociodemographic factors associated with having obtained the abortion from an uncertified source. We determined a priori which abortion-seeking and sociodemographic variables to include in the two multivariable models, based on which were most likely to have influenced abortion method and source choice; all of the chosen variables were included in the models. Among women using a medication, we examined which substances were used, how they were administered, who administered the medications, and whether the medications were taken at home. For those reporting instrumentation, we described the location and provider of the procedure. Finally, we used multivariable logistic regression to assess whether having a serious abortion complication was associated with primary induction method or obtaining the abortion from an uncertified source. Results are reported at the P ≤ 0.05 level. All analyses were conducted using stata version 12 (College Station, TX, USA).


The characteristics of the 527 study participants are shown in Table 1. Most women presented to the hospital with pelvic or vaginal bleeding (90%, n = 476) and/or abdominal pain (67%, n = 351). One in ten women had a serious health complication, including sepsis/septicaemia (7%, n = 38), high fever (2%, n = 11), organ failure (1%, n = 5), pelvic infection/endometritis (<1%, n = 4), hypovolaemic shock (<1%, n = 2), pelvic inflammatory disease (<1%, n = 2) and foreign body in the uterus (<1%, n = 2).

Table 1. Participant characteristics, women presenting for postaboriton care at four hospitals in Nepal (= 527)
  1. a

    Includes only currently married women.

  2. b

    Item included in household assets scale.

  3. c

    Missing two observations.

Age (years), mean, SD 27.65.8
Currently married, n ,% 50395.5
Age of marriage (years),a mean, SD18.54.3
Parity, n ,%
>Primary education, n ,% 28554.1
Currently employed, n ,% 27952.9
Household has motorbike or car,b n,%12323.9
Household has mobile phone,b n,%46387.9
Ever experienced physical domestic violence,c n,%8015.2
Gestational age, n ,%
<9 weeks39575.0
9–12 weeks7814.8
≥12 weeks5410.3
Hospital site, n ,%
Parokpakar Maternity (Kathmandu)23344.2
Tribhuvan University (Kathmandu)7013.3
Bharatpur (Chitwan)8516.1
Lumbini Zonal (Butwal)13926.4

Table 2 describes women's abortion-seeking experiences. Abortion law knowledge was low in this sample of postabortion care women, with 44% (n = 234) of participants knowing that abortion was legal in Nepal. Compared with women who were unaware of the legality of abortion, women who knew abortion was legal were younger (mean age 26.4 versus 28.5 years), had married later (mean age 19.0 versus 18.1 years), and had higher household asset scores (mean 0.4 versus −0.3). Abortion law knowledge was also associated with having more than a primary education [64% (183/285) versus 21% (51/242) among those with a primary education or less]. Abortion-seeking experiences were generally similar between women who did and did not know that abortion was legal. However, women who did not know that abortion was legal were more likely to have sought abortion information from pharmacists than those who knew abortion was legal (38% versus 26%).

Table 2. Participant characteristics, women presenting for postaboriton care at four hospitals in Nepal: proportions and P-values from chi-square tests (= 527)
 Total Medication Instrumentation P
n = 527 (100%)= 359 (68.1%)n = 168 (31.9%)
n % n % n %
  1. n/a, P-value not available due to zero or near-zero cells; ref., reference category.

  2. a

    ≤ 0.05.

  3. b

    P ≤ 0.01.

  4. c

    P ≤ 0.001.

  5. d

    Response categories not mutually exclusive.

Know that abortion is legal 23444.414941.58550.6 a
Know that a husband's permission is not needed to have an abortion 529.93610.0169.5 
Reason for seeking abortion d
Enough children32060.720958.211166.1 
Cannot provide for another child11321.47019.54325.6 
Previous child still young6312.04913.7148.3 
Health concerns5710.8349.52313.7 
Felt pressured to have abortion 173.2143.931.8 
Who told going to get abortion d
Family member16330.99526.56840.5 c
No one315.9287.831.8 b
How found out where to get abortion d
Pharmacist17232.616144.9116.6 c
Friend13625.88323.15331.6 a
Husband11622.06718.74929.1 b
Family member7113.5318.64023.8 c
Source of comfort or support d
Husband27051.217047.410059.5 b
No one13625.811030.62615.5 c
Member8716.54813.43923.2 b
Why chose selected abortion provider d
Close by27151.420857.96337.5 c
Recommended 25348.014339.811065.5 c
Affordable9117.35314.83822.6 a
All induction methods used       n/a
Oral medication only27452.027476.300 
Instrumentation only14126.80014183.9 
Oral and vaginal medication5410.35415.000 
Instrumentation and medication407.8154.22515.5 
Vaginal medication only163.0164.500 
Heavy massage10.20010.6 
Abortion from an uncertified source 34164.729181.15029.8 c
Abortion conducted at home 24746.924668.510.6n/a
Who performed abortion
Doctor or nurse23544.67420.616195.8 
Other (pharmacist, relative)428.0359.874.2 

The primary method of induction for 359 (68%) of the postabortion care women in our study was medication, whereas 168 (32%) used instrumentation. However, some women used a combination of medical and surgical methods. Furthermore, although 489 (93%) women presented to hospital after a single abortion attempt, 38 (7%) reported two or three independent attempts. Sociodemographic characteristics and gestational age of the pregnancy did not differ between women using medication compared with instrumentation, although knowledge that abortion was legal was lower among those using medication. On bivariable analysis, women who used medications were more likely to have told no one about the abortion, to have received no source of comfort or support, and to have found out from a pharmacist where to get an abortion, compared with women undergoing instrumental methods of abortion. Those women undergoing instrumental abortion were more likely to have found out where to get an abortion from their husbands or other family members than women who underwent instrumentation (Table 2).

In the multivariable model, women who used medical abortion remained more likely to have obtained abortion information from pharmacists [adjusted odds ratio (aOR) 8.1, 95% confidence interval (95% CI) 4.1–16.0] and to have told no one about the abortion (aOR 5.5, 95% CI 1.1–26.9), than women who used instrumentation (Table 3). They were also more likely to have selected the abortion method because it was available nearby (aOR 1.6, 95% CI 1.0–2.6) and less likely to have selected the method because it was recommended to them (aOR 0.4, 95% CI 0.3–0.7). Women presenting at the two rural hospitals were more likely to have used medical abortion than women presenting at urban sites.

Table 3. Factors associated with use of medical abortion as primary induction method, adjusted odds ratios, P-values and 95% confidence intervals from multivariable logistic regression (= 527)
 Use of medication (reference: Instrumentation)
aOR95% CI
  1. a

    P ≤ 0.05.

  2. b

    P ≤ 0.01.

  3. c

    ≤ 0.001.

Age 1.00(0.95–1.05)
Parity 0.77(0.56–1.05)
>Primary education 0.83(0.46–1.47)
Household assets 1.07(0.83–1.38)
Gestational age >12 weeks
<9 weeksref 
9–12 weeks0.93(0.52–1.65)
≥12 weeks1.34(0.62–2.88)
Know that abortion is legal 0.71(0.43–1.18)
Found out from pharmacist where to get abortion 8.05c(4.05–16.02)
Told husband going to get abortion 1.53(0.58–4.06)
Told no one going to get abortion 5.46a(1.11–26.90)
Chose method because close by 1.60a(1.01–2.55)
Chose method because recommended 0.41c(0.25–0.68)
Chose method because affordable 0.83(0.47–1.48)
Hospital site
Parokpakar Maternity (Kathmandu)ref 
Tribhuvan University (Kathmandu)1.52(0.73–3.18)
Bharatpur (Chitwan)3.68c(1.96–6.90)
Lumbini Zonal (Butwal)3.49c(1.90–6.43)

Among women using a medication, 324 (84%) could not name the substance used. Only 42 (11%) named drugs that can be safe and efficacious when used correctly, including mifepristone/misoprostol or Medabon®. Nineteen participants (5%) reported use of ayurvedic or other potentially unsafe or ineffective substances, including contraceptive pills, pain relievers or herbs. Among women obtaining instrumentation abortion, clinician assessment at the hospital identified 138 (82%) women as having had aspiration procedures; 22 (13%) dilatation and curettage/extraction; and 8 (5%) as having used other or unidentifiable forms of abortion.

Overall, 65% (n = 341) of the sample obtained an abortion from a provider or site deemed to be uncertified. Almost one-third of women undergoing instrumentation obtained the abortion from an uncertified source (30%, n = 50), as did 81% (n = 291) of those using medications. The majority of women using medication conducted the abortion at home. Most medications were administered by the women themselves (Table 2). In multivariable analyses, women who obtained abortions from an uncertified source were less likely to know that abortion is legal (aOR 0.4, 95% CI 0.2–0.7), and more likely to have obtained abortion information from pharmacists (aOR 8.7, 95% CI 3.9–19.6), than women going to a certified source (Table 4). Choosing an abortion method because it was available nearby was associated with going to an uncertified source (aOR 2.5, 95% CI 1.5–4.3), whereas choosing a method because it was affordable was negatively associated with uncertified source (aOR 0.3, 95% CI 0.2–0.6). Odds of suffering from a serious complication did not differ by induction method or by whether the abortion was obtained at an uncertified site.

Table 4. Factors associated with obtaining the abortion from an uncertified source, adjusted odds ratios, P-values and 95% confidence intervals from multivariable logistic regression (n = 527)
 Uncertified Source/Site
  1. a

    P ≤ 0.05.

  2. b

    P ≤ 0.01.

  3. c

    ≤ 0.001.

Age 0.94a(0.89–1.00)
Parity 1.08(0.76–1.54)
>Primary education 0.57(0.30–1.10)
Household assets 1.14(0.85–1.53)
Gestational age > 12 weeks
<9 weeksref 
9–12 weeks1.49(0.76–2.94)
≥12 weeks0.88(0.39–1.99)
Gestational age >12 weeks 0.81(0.36–1.83)
Know that abortion is legal 0.41b(0.24–0.72)
Found out from pharmacist where to get abortion 8.73c(3.89–19.61)
Told husband going to get abortion 0.82(0.27–2.55)
Told no one going to get abortion 3.56(0.61–20.80)
Chose method because close by 2.54c(1.50–4.30)
Chose method because recommended 0.91(0.52–1.60)
Chose method because affordable 0.31c(0.16–0.61)
Hospital site
Parokpakar Maternity (Kathmandu)ref 
Tribhuvan University (Kathmandu)0.87(0.37–2.06)
Bharatpur (Chitwan)0.33b(0.16–0.67)
Lumbini Zonal (Butwal)1.43(0.72–2.83)
Medication as primary abortion method7.18c(4.22–12.22)


Main findings

In this study of women receiving care at major public hospitals in Nepal after induced abortions, fewer than half of women knew that abortion was legal, and most erroneously thought a husband's permission was required. Two-thirds of participants had used a medication to induce abortion, one-third had undergone an aspiration or other instrumentation procedure. Women who used medications were more likely to have obtained information from pharmacists, to have chosen the method because it was available close to home, and less likely to have told others about the abortion, compared with those who had an instrumentation procedure. Medical abortion was also more common among women presenting to the rural hospitals. The majority of those who used medications obtained them from an uncertified source, administered the medications at home, and could not identify the medications taken. A quarter of the sample had had no source of comfort or support, and 6% told no one about the abortion.

Strengths and weaknesses

The limitations of our study are important. Although the sample was probably reasonably representative of women who present at public hospitals with abortion complications, it was not representative of all women who undergo abortion, nor all women who experience complications, because we did not capture women who do not seek care for complications. Similarly, we could not evaluate the experiences of women who may have obtained uncomplicated abortions from uncertified providers. Moreover, research has indicated that low-severity induced abortion problems often present as spontaneous abortions and can be missed by hospital clinicians.[16] The select nature of the sample limits the generalisability of our results. In addition, despite our attempts to gather in-depth data, we were largely unable to determine the drugs and dosages that women took because the women themselves did not know. We were therefore unable to characterise inappropriate medications being used. Furthermore, data on women's abortion experiences were self-reported and vulnerable to social desirability biases, particularly because some women may not have known whether their behaviours were legal and because abortion can be a stigmatised topic in Nepal.[9, 10] However, interviewers had extensive experience conducting abortion research and expertise in rapport-building, which probably improved women's willingness to disclose their experiences.

Our results are useful for identifying areas in which improvements in the abortion care delivery system are needed. The few studies that have examined abortion care-seeking and decision-making in Nepal after legalisation have focused on women seeking care at certified facilities[17-19] or have been qualitative investigations of perceived barriers to care.[9, 10] Research has indicated that complication rates at certified facilities are relatively low.[17] No study to our knowledge has collected in-depth information from women who had abortions requiring hospital care, capturing the experiences of women seeking care outside the formal medical system. Results contribute to our understanding of abortion-seeking factors that result in harm to women.


This study highlights the need for improvements in medical abortion provision in Nepal, as well as the opportunity it presents for expanding access. The proportion of postabortion patients in our sample who used a medication was high (68%) compared with women nationally (29%, 2006–11).[1] Although factors related to sample selection could be responsible, the higher representation of medical abortion users in our sample may indicate that, now that aspiration has become safer and more accessible,[8, 17] women using medications represent higher proportions of those suffering complications. Alternatively, we may be capturing a general trend of increased medical abortion use since 2006, with Medabon® becoming legally available in 2008, 2 years before this study. The use of harmful or partially effective regimens by women seeking medications from uncertified providers would also contribute to their greater representation in the postabortion wards of hospitals. In any case, that most women using medications had obtained them from uncertified sources and could not identify the medications indicates that potentially dangerous substances or incorrect dosages are bringing women to hospitals.

At the same time, medical abortion was probably an important option for some women. Medical abortion provided advantages in terms of privacy and proximity: choosing a method because it was available close to home, telling nobody about the abortion, and presenting to rural hospitals were more common factors among those using medical abortion. Medical abortion—including home self-induction with medications without direct medical assistance[20]—has the potential to expand women's access to safe abortion in Nepal, as in other low-resource settings.[21]

Our findings also point to the important role that private pharmacists play in abortion care in Nepal, including as sources of abortion information.[11, 22] Not only are they known in communities, but they offer other benefits over clinics and hospitals, including accessibility, anonymity and less costly services.[23, 24] However, pharmacists have varied levels of training and professional qualifications: whereas some hold a bachelor degree in pharmacy, anyone with 12 years of education who has received 2 weeks of orientation from the Department of Drug Administration can open a pharmacy shop. Mifepristone/misoprostol products are legally available in Nepal only through government-certified facilities and providers. Provision of abortion medications by pharmacists is not currently legal, but pharmacists do dispense drugs for abortion, with and without prescriptions, including a range of potentially unsafe or ineffective medications.[11, 12, 25] Even pharmacists providing mifepristone/misoprostol may not provide appropriate dosage and usage information.[11] Interventions with pharmacists might be an effective way to disseminate information about safe abortion services.

Finally, misperceptions and lack of knowledge of the legal status of abortion is prevalent among women seeking postabortion care, a decade after legalisation. The proportion of women who knew that abortion is legal in our study (44%) was similar to the proportion of all Nepali women aware of the law (38%), based on 2006–11 estimates.[1] Also, women who sought care from uncertified providers in our study had lower knowledge of the law than women who obtained care from certified sources. The effectiveness of the abortion law may be limited without additional strategies that raise awareness of women's abortion rights.


There has been a rapid expansion of facilities and clinicians providing legal abortion services and care to women in Nepal since legalisation in 2002.[8] Nevertheless, women continue to obtain clandestine or unsafe abortions requiring hospital care. Ensuring that women receive safe care remains an important challenge. Campaigns to improve women's knowledge of the abortion law may help to support their willingness to seek information and safe care. As medical abortion availability expands, harm reduction strategies for stemming the provision of potentially unapproved and unsafe medication regimens are needed. Efforts to educate women and providers about which drugs are legally indicated and proven safe will be critical. Evidence-based research on the types of medications women are using to self-induce, and what information and dosages they are being given, will be needed. Specifically, research is needed to understand the counselling and prescribing practices of pharmacists and to determine how the pharmacist community might best be mobilised to promote safe use of medication abortion pills.

Disclosure of interests

We have no competing interests to disclose.

Contribution to authorship

JTH, CCH and MP conceptualised, designed and served as the investigators of the study. MP, BD, MB and LB carried out study activities, including obtaining insitutional review board approval, designing study tools and data collection. MP and BD transcribed and cleaned the data. CHR designed and conducted the analysis and wrote the manuscript. CHR, MP, CCH, MP and JTH interpreted results. All authors read and revised manuscript drafts and reviewed the final manuscript.

Details of ethics approval

Human subjects approval for the study protocol was obtained from the University of California, San Francisco, Committee on Human Research (original approval date: 2 December 2008; approval #: H47920-33296-01). The study was also reviewed and approved by the Nepal Health Research Council (original approval date: 21 December 2008; reference #: 626). Due to the low literacy of the participants in the study, we received approval from the institutional review boards to document participants' verbal informed consent.


This research was funded by grants from the Society of Family Planning, the Lucille and David Packard Foundation, the Richard and Rhonda Goldman Fund, and an anonymous donor. The views and opinions expressed are those of the authors and do not necessarily represent the views and opinions of these funding agencies.


We are grateful to Philip Darney for making this research possible, to Prabhat Lamichhane and CREHPA staff for collecting data, to recruitment hospital staff for supporting this research, and to the Nepal Health Research Council for reviewing this study.

Commentary on ‘Unsafe abortion after legalization in Nepal: a cross-sectional study of women presenting to hospitals’

Worldwide, an estimated 43.8 million abortions take place every year. Almost half of these are unsafe abortions with an overwhelming majority occurring in developing countries (Sedgh et al. Lancet 2012;379:625–32). Whatever moral, religious or cultural views individuals or governments take of abortion, there is wide consensus that preventing unwanted pregnancy is preferable. However, unsafe abortion is still responsible for 13% of all maternal deaths worldwide (Sedgh et al. Lancet 2012;379:625–32), whereas mortality due to abortion is rare where safe services are readily available. For example, only 0.6 deaths per 100 000 abortions occur in the USA (Raymond and Grimes. Obstet Gynecol 2012;119:215–19). Restrictive abortion laws are not associated with low abortion rates and countries with restrictive laws have some of the world's highest abortion rates as well as unsafe abortion rates. Therefore, the case for access to safe legal abortion is compelling. Maternal deaths due to unsafe abortion can be avoided by the use of effective contraception, and provision of safe abortion and postabortion services.

Seeking to address its own high maternal mortality ratio of 539 per 100 000 births in the period between 1989 and 1995, Nepal legalised abortion in 2002 and made more comprehensive services available from 2004. Although improved access to safe abortion is unlikely to be the only determinant of the declining maternal mortality ratio observed during 1996–2006, it has certainly played an important role (Bhandari et al. BJOG 2011;118 suppl. 2:26–30). However, Rocca et al. present new data showing that legalisation has not, at a stroke, removed the practice of unsafe abortion in Nepal, and point to some important residual barriers. In a cross-sectional, interview-based study of women presenting between 2009 and 2010 to four major public hospitals with abortion-related complications, they found that over 68% of this admittedly highly selected population had used unsafe, ineffective or unknown substances to effect abortion. Compared with surgical methods, medical abortion was more strongly associated with use of either ineffective methods or use of unauthorised illegal sources.

These associations are in themselves consistent with what would be expected. It is unsurprising that women's choice of method was affected by knowledge, availability, accessibility and convenience. In a recent study, the odds of choosing a medical instead of surgical abortion was three times higher among Nepalese women who knew about both methods compared with those who did not (Tamang et al. Int J Gynaecol Obstet 2012;118 suppl. 1:S52–6). With increasing use and availability of medical abortion, more women, particularly in rural remote areas, are likely to obtain such procedures and many abortions are likely to take place at uncertified private facilities and homes without trained medical supervision. Although such abortions would be classified as unsafe by current definition, they might be safer compared with some traditional methods practiced to induce abortions. There is emerging evidence that both surgical and medical abortion services provided by trained mid-level providers are as effective and safe as those provided by the physicians (Renner et al. BJOG 2013:120:23–31; Ngo et al. Int J Women's Health 2013;5:9–17). The extent and complexity of training needed to provide safe medical abortion may be less compared with that needed to provide safe surgical abortion. Therefore, the possibility of extending training in safe medical abortion to include nurses and pharmacists should be considered to improve accessibility and safety.

One of the most striking findings was this population's relatively poor knowledge of abortion law and the reasons given for seeking abortion. In this group of predominantly married, parous women commonly choosing abortion because they felt unable to provide for another child, only 44% actually knew that abortion was legal, while as many as 61% believed that they required their husband's permission. Add to this the finding that 15% of women interviewed reported physical domestic violence; almost certainly an under-representation of the true prevalence of all forms of domestic violence, which might impact on women's freedom to make sexual health choices. The investigators did well to gather information on women's knowledge about abortion as well as reasons for seeking abortion, and also their degree of social support. They did not however, gather data on other potential psychological influences, such as shame, ambivalence or fear of stigmatisation, on abortion-seeking behaviour.

The authors acknowledge that these data, gathered from a highly selected group of women presenting with complications, cannot give a representative picture of all women seeking abortion in Nepal. Similarly, the data cannot be read as a comment on the relative safety of medical versus surgical abortion. However, they provide some useful quantitative data that highlight the organisational, cultural, educational and psychological hindrances that remain, post legalisation. They underline the importance of accessibility to safe abortion services and the need for better education as well as qualitative research into the hidden effects of cultural attitudes on women's feelings and behaviours.

Disclosure of interests

Neither author has anything to disclose.

  • G Acharya & S Goldbeck-Wood

  • Department of Clinical Medicine, Faculty of Health Sciences, University of Tromsø, Tromsø, Norway