Pregnancy-Related Deaths of African Women (II)
Hospital and Community Based Studies

Association for Interdisciplinary Research in Values and Social Change
Vol. 13, No. 3 March/April, 1999
Thomas W. Strahan, J.D.
Reproduced with Permission

This article is a continuation of Pregnancy-Related Death of African Women II (Volume 13, No.2, January /February, 1999) which concentrated on Background and Risk Factors. This article summarizes hospital and community based studies of pregnancy-related deaths of African women in several countries. It also includes a description of the abortion methods used, if described in the article, as well as the identity of those performing the abortion. Demographic data on the background of women who obtain induced abortions as well as who dies from pregnancy-related causes is also included. Although the studies have some important limitations, they also provide valuable insights of the various reasons why African women die from abortion and childbirth.


Seventy-four women with complications of induced abortion were studied prospectively at the Obafemi Awolowo University at IfeIfe, Nigeria between November 1988 and December, 1989. Fifty-eight percent of the women were single and never-married, 80% were Christians, 49% were students, 20% were traders, 5% were teachers with the remaining women being unemployed or casual workers.

It was reported that 32% of the abortions were performed by medical practitioners (presumably qualified), 27% were performed by non-medically qualified persona, 19% were self-administered drugs or other dangerous instruments, and 5% were performed by nurses. The most common complications from the induced abortions were sepsis (84%) and haemorrhage (51%). All women received a broad spectrum of antibiotics and 76% required evacuation of fetal parts or tissue. There were 13 reported maternal deaths from induced abortion which were part of the total of 37 maternal deaths in the hospital during that same period. Fifty-four percent of the 13 maternal deaths from the complications of induced abortion were performed by medical personnel including 5 medical practitioners, 2 nurses and 3 persons identified as non-medically qualified.

Because of these deaths the authors questioned the effectiveness of the antibiotics. Although the antibiotics may not have been effective, there could have been other reasons for their deaths. For example, the study did not state whether or not any blood tests were undertaken or whether any blood transfusions had been given. Nor was there any indication that any women had undergone laparoscopy to detect whether or not a potentially-life saving hysterectomy was needed or an ectopic pregnancy was present. Thus, it appears that certain other potenially life-saving methods were not utilized, most likely because of lack of resources or training.

Twenty(20) of the women in the study satisfactorily completed an interview schedule. Nine (45%) of the women stated they had previously aborted one or more pregnancies. All had knowledge of contraception but only 10% used any effective method of contraception. Only 20% of the women knew abortion was illegal in Nigeria. When they were told that abortion was illegal, only 10% thought abortion should be made legal. The major reasons given by the 18 women who did not want abortion legalized was that abortion was immoral and that it was against the wish of God. The authors of the study noted that this was the same argument made in the public debate by those who were opposed to the legalization of abortion in Nigeria1.

In another study of Nigerian women admitted to the gynecological ward of a hospital in Calabar, Nigeria during 1985-88, 12,117 deliveries, 1421 spontaneous abortions and 147 women with induced abortions were reported. Eleven percent (11%) of the women with induced abortion had a repeat induced abortion and 31.6% had previously carried to term. Among the women with induced abortion, 72% were under 20 years of age, 80.9% were single, and 58.1% were students. Five percent (5.1%) of the women admitted having practiced contraception. Sixty-one percent (61.8%) were reported to have voluntarily given an accurate history.

Various induced abortion techniques included oral and injectable drugs (51); enema (16); dilatation and curettage (29); suction with plastic cannula (3); locally prepared sharp metal rods introduced transvaginally into the uterine cavity (44); intravaginal herbs introduced digitally (10), penetration of the vagina with the abortionist's penis smeared with grease (2); and 15 women were unsure of the technique used. Thirty-two percent (32.3%) of the induced abortions were done by traditional healers, 23.5% by chemists, 18.4% by orthodox doctors, 8.8% were self-induced, 7.4% by quack doctors, and 9.6% were unspecified.

Ninety-eight percent (98.5%) of the women received broad spectrum antibiotics, 59.5% required blood transfusion, and 77.9% had uterine evacuation for retained fetal parts or tissue. During the study period, there were 41 deaths in the gynecological ward and 80 maternal deaths in the entire hospital. Induced abortion accounted for 16 deaths and spontaneous abortion accounted for 3 deaths. Induced abortion contributed 20% of the maternal deaths in the hospital during the study period2.

In another Nigerian study of 230 cases of illegally induced abortion complicated by sepsis during 1981-87 treated at the University College Hospital at Ibadan, Nigeria, it was reported that 10% of the abortions had been performed at chemists shops, 12.2% in homes, and 73.4% in private hospitals or clinics. The major symptoms of septic abortion included abdominal pain (81.3%), fever (40.4%), vaginal bleeding (35.2%), foul smelling vaginal discharge (31.3%), and abdominal distension (13.9%). Methods of treatment included suction evacuation with antibiotics (53.0%), antibiotics alone (17.0%), laparotomy with antibiotics (13.9%), and hysterectomy (1.7%). A perforated uterus was present in 36 of 230 cases. The average number of pints of blood transfused per patient was 2.23. Maternal mortality was 8.3%. Complications from sepsis included peritonitis (68.3%), pelvic abscess (26.1), jaundice (12.2%), uterine perforation (10.4%), acute renal failure (3.0%), and disseminated intravascular coagulation (1.3%). The authors believed that the rapid rise of poorly equipped hospitals that perform abortions, usually at no fee to attract patients, was mainly responsible for the septic abortions. They also observed that antimicrobial treatment regimens were not uniform, even in the same hospital3.

The incidence of disseminated intravascular coagulation (DIC) (1.3%) in this study was much higher than previous U.S. studies which reported an incidence of DIC of only 8 per 100,000 suction curettage abortions and 191 per 100,000 dilation and evacuation abortions4. DIC is defective coagulation of the blood which results in life-threatening hemorrhage. Both induced abortion and sepsis are known risk factors for DIC5.

In another study of 934 women who delivered in a rural public hospital in Ekpoma, Nigeria during 1989-91, 89.1% were reported to be spontaneous vaginal deliveries, while 10.9% were delivered by caesarean section. During this time there were 12 maternal deaths with a maternal mortality ratio of 1285 per 100,000. One-hundred forty (15%) of women who delivered had complications of which 45% were due to haemorrhage. Haemorrhage included all cases of ruptured ectopic pregnancies, threatened abortions, incomplete abortions, and ante-partum and post-partum haemorrhage. Other complications were toxemia of pregnancy (21%); prolonged labor (18.6%), sepsis (10%), and severe anaemia (4.3%).

Malfunctioning of hospital equipment, shortage of essential obstetric drugs and supplies, treatment delays, lack of blood donors, blood storage and blood replacement facilities, along with low staff morale were observed. The authors stated that the problem of high maternal mortality in sub-Saharan Africa stems not only from inadequate health facilities, but also is the end result of the interplay of numerous antecedent factors that could be social, cultural, economic and logistic; coupled with very high fertility6.

In a case-control study of risk factors for 35 maternal deaths among women who died in Obafemi Awolowo hospital in Nigeria during 1989-91, it was found that delay in arriving at the hospital was significantly higher among women who died compared with controls. Women who died were also likely to have no formal education (66% v. 11%), and were more likely to be under age 20 (34% v.11 %). Two-thirds of both cases and controls had no prenatal care. As many as 14 deaths were attributed to inadequacies in the health care delivery system. The primary complications which resulted in death included postpartum hemorrhage (29%), obstructed labor and puerperal sepsis (29%), caesarean section (29%), eclampsia (9%), hepatitis or tetanus (9%), and sickle cell disease (9%). No deaths were attributed to abortion, either spontaneous or induced,in this study7.


Another study involving 1077 women who were admitted to 8 hospitals in various locations in Kenya in 1988-89 for incomplete abortions, found that 169 (15.7%) had illegally induced abortions. The percentage of women with induced abortions admitted to the various hospitals as a percentage of total hospital admissions varied widely. The highest (36.4%) was at the only mission hospital which was located in a relatively affluent area of Kenya, while the lowest was in a district hospital located along the coast (0.9%) Among the women with illegal induced abortions only 20% were married, a majority (62.6%) were less than 25 years of age and 29.6% were between 10 and 19 years of age and were significantly younger than the non abortion group. There was a steady rise in the proportion of women with induced abortion with increasing level of education. Women with no formal education represented only 4.7% of the women with illegal induced abortion, while women with post-secondary college or university education represented 26.3% of the sample.

Students were 14.8% of the women with induced abortions; unemployed, including housewives were 42.6%; those formally employed were 13.0%, while bar girls and sex workers represented 11.3% of the induced abortion group. Among the induced abortion group, 68.7% said their boyfriend was responsible for their pregnancy while 20.9% said it was their husband and 10.4% said it was a stranger, others, or not known. In contrast, most of the non-abortion group reported that it was their husband (78.5%) or boyfriend (19.8%) that was responsible for their pregnancy.

Abortion-Related Deaths as Compared to All Maternal Deaths
LocationYearAbortion Deaths
Maternal Deaths
Lfe-Lfe, Nigeria (hospital)1977-8824/23210.4illegal abortion
Lfe-Lfe, Nigeria (hospital)1988-8913/3735induced abortion
Calabar, Nigeria (hospital)1985-8816/8020induced abortion
Lfe-Lfe, Nigeria (hospital)1989-910/350not mentioned
Ekpoma, Nigeria (hospital)1989-910/120not mentioned
Thika District, Kenya (institutional)1981-8816/1649.7postabortal spsis
Kwne District, Kenya (community)1989-900/350not mentioned
Harare, Zimbabwe (urban community)1989-90--23abortion complications
Masvingo, Zimbabwe (rural community)1989-90--15abortion complications
Conakry, Guinea (health facilities)1989-9013/12310.5induced abortion
Conakry, Guinea (community)1989-906/1154.5induced abortion
Ejisa, Ghana (community)1985-891/442.2septic abortion
Bagamoyo, Tanzania (community)19933/764.0induced abortion
Addis-Ababa, Ethiopia (community)1981-8312/4526.6illicit abortion
Kampala, Uganda (hospitals)1980-86117/58020.2abortion deaths
Bloemfontein, South Africa (hospital)1980-8512/8114.8postabortal sepsis

Thirty-Seven percent (37.8%) of the women with illegally induced abortions reported to the hospital within 24 hours after noticing initial symptoms, compared to 50.2% of the non-abortion women. Complications due to illegal induced abortion were reported to be sepsis (34.3%), septic shock (2.4%), haemorrhage (12.4%), anaemia (17.8%), and genital injury (16.6%) which were all substantially higher compared to women with incomplete abortions which were not illegally induced. There were two maternal deaths during the study period. One in the non-abortion group died due to hemorrhaging as there was no blood in the hospital. The other in the induced abortion group died due to severe septicaemia with septic shock, renal and hepatic failure. The case-fatality rate for women with illegally induced abortions was 6.1 per 1000 compared to 1.1 per 1000 noninduced abortion admissions8.

A community based study in Costal Kenya identified 35 maternal deaths during 1984-87. The primary direct causes of death were hemorrhage, sepsis, and obstructed labor/ruptured uterus with anemia a primary indirect cause of death. The birth attendants for the 35 maternal deaths were health worker (11%); traditional birth attendant (32%); relative (54%) and self-(4%). which was similar to the type of birth attendants for women age 15-44 giving birth in the area. In 29% of the maternal death cases women had no previously born children compared to 18% among women age 15-44 giving birth in the area In fifty-percent of the maternal deaths women had 1 to 4 children compared to 51% among women age 15-44 giving birth in the area; and in 21% of maternal deaths women had five or more previously born children compared to 31% among women age 15-44 giving birth in the area.

Very few of the children who the mother was attempting to deliver, survived infancy Eight (8) died in utero. 10 were stillborn, 5 lived up to 6 days, 2 died after 2-4 weeks, 5 died between 1-11 months, and only 5 survived infancy9.

In another study of maternal mortality in Thika Subdistrict in Kenya from January 1981 through September, 1988 there were 86,248 live births and 164 maternal deaths in the various governmental and private institutions in the subdistrict. This was calculated to be a maternal death rate of 190 per 100,000. Only 25% of the mothers who died delivered live babies. Only 28.1% of the mothers who died had attended an antenatal clinic anywhere, and those who did attend frequently started late in pregnancy as 52.1% were already in the third trimester.

Among those who attended an antenatal clinic anywhere, only 32.6% had any blood or urine test, only 34.8% had any height or girth recording, while 95.7% had a blood pressure recording. Among the 164 maternal deaths, 46 women died from sepsis which included 15 women who died following caesarean section, 8 who died following post vaginal delivery and 16 who died from postabortal sepsis. In addition, 4 women died from postabortal hemorrhage, 2 from drug poisoning and 2 from ruptured ectopic pregnancy. Factors which contributed to maternal mortality included: unavailability of blood (22.6%); unavailability of medicines or IV fluids (18.9%); unavailability of laboratory facilities for cultures, electrolytes(17.1%); unavailable transport or impassable roads (12.2%); wrong booking (7.3%); patient factor/misjudgment or non-cooperation (15 9%)10.


A case-control study in both rural and urban Zimbabwe was undertaken in 1989-90 to determine the incidence and risk factors for maternal mortality. Information was obtained from various family members as well as from various health and vital statistics records of the women who died and compared to women who delivered without dying. The maternal mortality rate in the rural area selected was determined to be 168 per 100,000 1ive births compared to 85 per 100,000 in the urban setting. The major direct causes of deaths in the rural setting were haemorrhage (25%), abortion complications (15%), puerperal sepsis (15%), and eclampsia (5%). The major direct causes of death in the urban setting were eclampsia (26%), abortion complications (23%), puerperal sepsis (15%) and haemorrhage (10%). Compared to a woman who was the only wife of a husband, single women had an increased risk of maternal death of 4.7; divorced, separated or widowed women had an increased risk of 5.6 - 7.1; cohabiting women had an increased risk of 8.0; and being one of several wives had an increased risk of 2.0. Compared to households where the woman's husband was the head of the household, there was an increased risk of 5.8 - 6.6 if the woman herself was the head of the household. If the guardian family was someone other than the wife's or husband's family, the woman was 4 - 5 times more at risk of maternal death.

Previous or present abortion, whether spontaneous or induced, increased the risk of death by 4.0 compared to women with no abortion. 40% of the urban women who died were reported to have had one or more abortions, compared to 13% of controls. There were no significant differences between cases and controls by level of education, households approximated total income or whether domestic help was available during pregnancy, or whether or not the pregnancy was unwanted".11

One reason for lower maternal mortality among married, monogamous women is most likely due to a significantly lower incidence of sexually transmitted diseases which would reduce the risk of sepsis12. Another reason for lower maternal deaths among closer-knit families may be due to the frequent need for family members to assist the woman in obtaining hospital care by purchasing needed supplies, drugs, locating absent staff, assisting in transportation or carrying out her family responsibilities while she is ill.

A community-based investigation of maternal deaths caused by obstetric hemorrhage was undertaken in the rural Masvingo, Zimbabwe in 1989-1990. One hundred nine (109) total maternal deaths were identified of which there were 26 deaths from obstetric hemorrhage. Forty percent were due to a ruptured uterus, and 30% were due to an atonic uterus. Fifty percent of the women had no intervention whatsoever before death from obstetric hemorrhage. Among the women who died from obstetric hemorrhage, 42% were more than 35 years old, 44% had five or more children, 69% were married and most were peasant farmers. The risk of death from obstetric hemorrhage was not significantly affected by marital status, status within the household, education or religion. Many women lived a long distance from a health facility and it was concluded that provision for emergency transport would have saved 50% of the women13.

Anemia was a likely underlying important cause of death of these women because it contributes to maternal deaths from obstetric hemorrhage14. Sixty percent of pregnant African women have anemia and recent studies have reported even higher percentages15. Improved nutrition and iron supplement, may have prevented many of these deaths. The WHO recommends that antenatal clinics provide iron and folate for every women in areas of high anemia prevalence16.


In a study undertaken to detect all deaths of women age 15-45 years during 1989-90 in the City of Conakry, Guinea, 102 maternal deaths associated with delivery were identified. Two (2) deaths were related to ectopic pregnancy and there were 18 deaths related to abortion (unspecified) which were excluded. In addition, deaths of women who were not residents of Conakry were also excluded. The 102 maternal deaths were compared to 338 women who had given birth and survived. It was found that risk factors for maternal death associated with delivery included, anemia (2.1), low family income (2.6), hypertensive symptoms (19.8), infection (3.7), difficult birth (9.0), postpartum complications (4.0) and history of caesarean section (12.0 ).Among the women who died, 58% had anemia17.

Another study of maternal deaths of women in Conakry in 1989-1990 reported 123 registered maternal deaths at health facilities and 11 maternal deaths recorded in a community survey which included examining burial records at a Muslim cemetery. Of the 123 registered maternal deaths at health facilities, 94 (76%) were term pregnancies, 13 (10.5%) were induced abortion, 6 were spontaneous abortion, 9 were premature deliveries and 2 were ectopic pregnancies.In two-thirds of the 19 abortion deaths, women had taken abortifacient plants. In 13 of the 19 abortion death cases, women had an absence of menses for 12 weeks or more. Of the 11 maternal deaths in a community survey, 5 were term pregnancies and 6 (54,5%) were induced abortions. Among postabortion complications, sepsis (71%), perforation of the uterus (7%) and icterus (7%) were the most common18.


A 1990 community based survey in the Ejisu health district to determine all women who died in pregnancy, delivery or the puerperium during 1985-89, identified 44 women who died during that period. The main presumptive causes of death were postpartum hemorrhage (45.5%), jaundice in pregnancy (22.7%), obstructed labor (6.8%), eclampsia (6.8%), fever (2.2%), and septic abortion (2.2%). Fifty-nine percent of the women died in a hospital; 27.3% died at home; 22.7% died during pregnancy and 72.7% died during delivery. Thirty-four percent of the women who died did not attend any antenatal clinic19.

In another study, two-hundred twelve (212) women were admitted to a teaching hospital in Accara, Ghana in 1993-94 because of complications from induced abortions. Ninety-five percent of the cases were self-referral. Fifty-eight percent of these abortions had been performed outside designated health institutions despite the liberalization of the abortion law in Ghana about 10 years ago. It was stated that Ghanian society still seems to perceive abortions as illegal or unethical and hence to be procured clandestinely. The overall mortality rate was 2.4% while the mortality rate for septic abortions was 7.5% which the authors believed was comparable to other developing nations. The overall surgical intervention rate was 94% and there were laparotomies (incision into the loin) in over 10% of the cases of which 12 (5.7% overall) were for ectopic pregnancy which had been mistaken for interuterine pregnancies. The authors believed that laparotomy is essential if any reduction in mortality is to be achieved. However,it was stated that a laparoscope is not available in many hospitals in Ghana, as well as hospitals in other developing nations20.

To reduce the risk of death or serious injury from ectopic pregnancy, a diagnostic procedure has been developed to determine hCG levels along with vaginal probe ultrasound for use at abortion facilities21 or in emergency rooms of hospitals22. Failure to detect the ectopic pregnancy before the woman leaves the abortion facility, places the woman at increased risk of death23. A recent text book on abortion states that immediate laparoscopy should be offered to women with larger ectopic pregnancies, or peritoneal signs indicating rupture24. Also, there are various large medical centers in the U.S., where special units are available to especially handle ectopic pregnancy by use of drugs to effect a type of induced abortion or by operative removal of the growing embryo which is usually in the fallopian tube. Because ectopic pregnancy is potentially life-threatening, establishment of these centers is believed to have significantly reduced the incidence of death of women from ectopic pregnancy in the U.S25.


In order to determine the number of maternal deaths in rural Tanzania during 1993, a study was undertaken by doctors in the area to examine records, by networking with persons having possible information in order to establish a verbal autopsy determining the cause of death of Muslim women. The district studied had 31 dispensaries, 4 health centers and one district hospital. A total of 76 maternal deaths were found which represented a maternal mortality ratio of 961 per 100,000 live births. The leading causes of death were postpartum hemorrhage with retained placenta (13.2%), anaemia (13.2%), postpartum hemorrhage without retained placenta (11.9%). HIV/ARC syndrome (11.2%), obstructed labor (9.2%), puerperal sepsis (9.2%), malaria (5.3%), induced abortion (4.0%) and spontaneous abortion (1.3%).

Maternal deaths occurred irrespective of factors such as access to a main road, the presence of antenatal risk factors, or contact with health care personnel or a nearby facility before death. Antenatal care had been received by 97.2% of the mothers before death but risk factors such as high parity, anaemia, twins, or transverse lie were not taken into account and could have been prevented. The referral rate, even with the presence of a known risk factor, was only 34.6%, and patient compliance to the referral was only 44.4%. Many women returned to their family of origin instead of entering the health care system26.


In a study of maternal mortality at five hospitals in Kampala, Uganda during 1980-86, it was found that the nonabortion mortality rate was 2.65 per 1000 deliveries while the abortion related mortality rate was 3.58 per 1000 abortions. Of all maternal deaths, 80% were nonabortion and 20% were abortion related. Immediate causes of death were sepsis, haemorrhage, ruptured uterus, anesthesia and anemia. The most common patient management factors which contributed to deaths were lack of blood for transfusion, lack of drugs and intravenous fluids, theatre problems and doctor-related factors. The authors noted an increase in maternal mortality during the study period which they thought may be due to worsening economic conditions. The authors believed that maternal mortality could be reduced if health workers were imaginative with respect to each patient, tried not to operate as though they were working in a developed country, and created relevant solutions for local problems27.


In a study of 1153 adolescent gynecological in-patients at Queen Elizabeth Hospital in Blantyre, Malawi in 1994-95, 81% were identified as having complications of abortion, 9% were identified as septic abortion and 3% as criminal abortion. There were no recorded deaths. Ectopic pregnancy (1.%), molar pregnancy (0.2%), pelvic inflammatory disease (7.6%) sexually transmitted disease (2.5%) were among the listed complications28. It was unclear as to how the gynecological complications were considered to be abortion complications, since there was no data presented as to whether or not the adolescents had been pregnant, or whether or not they were now pregnant or admitted to having been pregnant in the past.


A two-stage probability study among women age 13-49 was conducted in Addis-Ababa, Ethiopia in 1981-83. Of the 9315 women estimated to be pregnant during that time, 45 died from the complications of pregnancy, delivery and the puerperium. Mortality was highest for nulliparous women, unmarried women and women employed as maids/janitresses, and students. There were 24 direct obstetric deaths and 21 indirect obstetric deaths. Thirteen (13) women died without delivering, and 9 of these were indirect obstetric deaths. Of the 24 direct obstetric deaths, abortion accounted for 13 cases of which 12 were identified from various sources as illicit methods (26.6% of maternal deaths). There were 38 women whose pregnancy desirability was known, and of these 38, 26 reportedly had an unwanted pregnancy. Among the 45 women who died, 16 did so before 20 weeks gestation, 13 died between 20-26 weeks gestation, and 16 women died after 36 weeks gestation. Twenty-four women died in a hospital and 21 died either at home or on the way to the hospital29. In this study nearly two-thirds of the women who died were at 20 weeks gestation or more. According to U.S. studies where supposedly the safest techniques are used, childbirth has been established as safer if the induced abortion takes place at 16 weeks gestation or longer based upon the relative incidence of maternal mortality30.

A five year retrospective community based study was conducted in 1990 to determine the maternal mortality rate in Jimma town in southwestern Ethiopia. The maternal mortality rate was determined to be 4.02 per 1000 1ive births. The major causes of death were sepsis, eclampsia and abortion. More than 50% of the deaths occurred after delivery of a child. The study indicated a poor trend in following antenatal care and family planning among maternal mortality cases31.

South Africa

In a recent South Africa study, 803 women presented with incomplete abortion at 56 hospitals in 9 regional centers in a two-week period in September-October, 1994. Sixty-six (66.8%) of the overall sample were single women,15.3% were teenagers. and 38.7% were in the second trimester. Using South Africa's old race categories, 84% of the women were black, 11% were coloured, 4% were Asian and I% were white. Overall, 21.9% of the women had signs of infection, 3.2% had organ failure, and 18.4% had offensive products on evacuation, 3.9% had mechanical injury to genitalia and 1.2% had a foreign body. The condition of 15% of the women was considered to be severe according to definitions set forth in the study. Seventeen (17%) of the women received blood transfusions and 95.9% were in the hospital one day or more. Five percent (5.5%) of the women had a hemoglobin level (g/dl) of 6.5 or less, 9.5% were between 6.6-8.5, 32.4% were between 8.6-10.5, and 52.6% were greater than 10.5. Antibiotics were given to 49.5% of the women but 15% of women with serious infections did not receive any antibiotics. Only sixty women (7.5%) with incomplete abortions were determined to have undergone an induced abortion. Seventy-eight percent of the women with incomplete abortion had no signs of infection and only 23.5% had offensive products, mechanical injury to genitalia or a foreign body present. This indicates that the actual incidence of attempted induced abortion or attempted abortion could have been somewhat higher than the 7.5% which was reported.

Three maternal deaths were reported during the study period. The women who died were age 15, 21 and 27 years of age; all were black and one had two children. All died of complications of illegal abortion (5% of identified induced abortions). One woman reportedly used a mixture of Dettol, soap and lotion to induce the abortion.

The relatively low percentage (7.5%) of women with incomplete abortion who were identified as having had an induced abortion was believed by the authors to be too low. They also commented that the definition of "unsafe abortion" developed by the World Health Organization did not fit the results of their study32. Less than one-fourth of the women had any sign of infection and less than one quarter had offensive products, mechanical injury to genitalia or a foreign body present which might indicate an incomplete abortion or one performed by an untrained or unskilled person. Other African studies have also reported a low incidence of septic abortion or admitted induced abortion among women with incomplete abortion33.

Another South African study of 2450 women admitted to King Edward III Hospital in Durban during a 15 month period in 1983-84 with a diagnosis of 'abortion', found that 647 (26.4%) satisfied the criteria for septic abortion and were presumed to be cases of illegal abortion. Of the 647 patients with septic abortion, 42 (6.5%)underwent laparotomy. It was determined that a hysterectomy was necessary in 35 patients. Eighteen women died (1.8%) out of the 647 patients including 7 which had undergone a hysterectomy34.

In another South African study of 81 maternal deaths at Pelonomi Hospital, Bloemfontein from 1980-85, the primary direct causes of death were identified as puerperal sepsis (30.9%), postabortal sepsis (14.8%), hypertensive diseases (21.0%), and haemorrhage (13.6%). Preventable factors were identified in 80% of the cases. Factors attributable to the patient included late presentation (22 deaths), refusal of treatment (3 deaths), and the unbooked state (14 deaths). Hospital failures included failure to diagnose, failure to institute appropriate treatment, and delay in transfer (23 deaths). The mortality rate was much higher among unbooked patients (11.13 per 1000) compared to booked patients (0.32 per 1000). Among the 81 maternal deaths, 44.4% of the women had no prior born children, 11.I% had one prior child, 6.2% had two prior children, 4.9% had three prior children, 2.5% had four prior children, and 14.8% had five or more prior children35.

Next Page: References:
1, 2