Determining Factors of Cesarean Delivery Trends in Developing Countries : Lessons from Point G National Hospital ( Bamako – Mali )

One of the obstetric interventions introduced to address this issue is the cesarean – delivery. Cesarean delivery is defined as the birth of a fetus through incisions in the abdominal wall (laparotomy) and the uterine wall (hysterotomy) [Cunningham, 2001]. Historically, cesarean delivery was associated with a high complication rate, sometimes causing maternal death. In the era of modern medicine, however, cesarean section has become safe and is widely endorsed throughout the world as a strategy to improve pregnancy outcomes [Weil & Fernandez, 1999].

Most developing countries, however, report cesarean delivery rates well below the acceptable minimum standard of 5% outlined by the WHO.Poor healthcare access, underdeveloped healthcare infrastructure, geographical inaccessibility, cultural mistrust, poverty, and paucity of human health resources are barriers to providing cesarean deliveries to all women who need them [Dumont et al., 2001;De Brouwere et al, 2002;Ronsmans et al, 2002;Kwawukume, 2001].Large ecological studies in West Africa emphasized this gap by demonstrating increased maternal mortality in settings with a lower percentage of births supervised by a skilled attendant, fewer deliveries performed in-hospital, or a smaller proportion of deliveries performed by cesarean section.Increased access to these services correlated with lower maternal mortality rate [Ronsmans et al, 2003].
Mali is the 3 rd poorest nation in the world, with an estimated maternal mortality ratio between 464 and 830 deaths per 100,000 live births [Chou et al, 2010;Samaké et al, 2007].In 1990, the Mali Ministry of Health developed a healthcare initiative focusing on the maternal and child health.Among the key elements of the Malian healthcare system, is the clear distinction between the three levels of care provision: primary (community health centres), secondary (district referral health centres) and tertiary care (hospitals).Pregnant women are initially supposed to book at the community health centres (which is the entry point of the healthcare system) with a primary care midwife or obstetric nurse for care provision during pregnancy, birth and the puerperium.These community health centres have the pivotal role of patients selection based on risk assessment.One important innovation of this new policy was the establishment of a referral system for perinatal complications in 1994.To ensure that referral takes place in an optimal fashion, guidelines for consultation and collaboration between community health centres, district referral health centres and hospital have been formulated in the Perinatality Module and in the Standard, Options and Procedures for Reproductive Health Services Manual.In these documents, all professional groups involved in maternity care agreed on the indications for consultation and referral according to the level of care.This program augmented the healthcare system's capacity to manage obstetric emergencies by upgrading referral centres' technical trays including staff training, surgical theatre rehabilitation, creation community health centres in previously inaccessible areas, organisation of transport between the community centres and referral centres, and communities' mobilisation to own the system.The main obstetric emergency encountered was cephalopelvic disproportion and its complications.Cesarean delivery was the main obstetric procedure used to deal with these complications.Lowering of financial barriers to increase access to this major obstetric intervention was one of the strategies of the organisation of the referral system in Mali.
To date, there have not been any in-depth evaluations of cesarean delivery in Mali since the inception of this program.Poor data capture of most population health indicators have called into question the reliability of cesarean delivery reports for other developing countries [Stanton et al, 2005;Holtz and Stanton, 2007].In this context, large hospital databases of good quality provide a bird's eye view of the national health system and trends in healthcare delivery over time.
This paper aims to assess the trends of cesarean delivery at the Point G national hospital in Bamako, Mali over a period of 2 decades.We explore the impact of sociodemographic, reports, admissions records for the intensive care service, records from the internal medicine and urology services, and hospital death records.

Analysis
We report trends in cesarean delivery rates at Point G National Hospital in Bamako, Mali from 1985 to 2003.Annual cesarean deliveries rates were calculated and grouped by historic time intervals to elucidate changes in cesarean utilization over time.These intervals represent 5 distinct periods in the hospital's history: 1985 to 1990 before the department of obstetrics and gynecology was established by the first Malian professor in this field ; 1991 to 1995 encompassing the introduction of the National Perinatality ProgramProgram; 1996 to 1997 when the service of obstetrics and gynecology functioned at partial capacity due to hospital renovation; 1998 to 2001 immediately after renovation; and 2002 to 2003 when the major obstetric team moved from Point G to Gabriel Touré Teaching Hospital, another teaching hospital in Bamako.
We first computed cesarean delivery rates during the five time periods according to different categories to observe general trends.Cesarean delivery rates were calculated as the percentage of pregnant women delivered after surgical opening of the abdomen.Crude and adjusted odds ratios (OR) were obtained by logistic regression and subsequently transformed into relative risks (because rates of cesarean delivery were more than 10%) by the equation: [Zhang & Yu, 1998]. (1) Characteristics considered to be of relevance for cesarean delivery were: maternal age, marital status, ethnic group, parity, hypertension or diabetes in pregnancy, gestational age, number of fetus (single vs. multiple gestations), cesarean delivery indications and referral status.
We then described the contribution of different indications to overall cesarean delivery rates following the rules of the Baltimore group on cesarean indications reporting for developing countries [Stanton et al, 2008].Interactions of these indications with maternal characteristics have been reported.The next step looked for our practice concerning specific obstetric group.This step focused on the study of cesarean delivery in ten obstetric groups.The definition of these groups appears in table 3. Based on the review of the relevant literature about this topic [Stavrou et al., 2011;Costa et al, 2010;Brennan et al, 2009;McCarthy et al, 2007;Robson, 2001], we focused on the correlation between trends of overall cesarean delivery rates and that of the cesarean delivery rates in term single cephalic nulliparas (TSCN).The term single cephalic nulliparas gathered groups 1 and 2 during the 19 year period.Pearson's correlation coefficient was used to estimate the relationship between overall CS rates and TSCN cesarean delivery rates.Independent Student t test was used to compare mean overall CS rates.The coefficient of variation (CV) was calculated as the standard deviation (SD)/mean x 100.The relevant cesarean indications characterising this specific composite group were identified.Finally, we identified individual factors influencing the cesarean delivery rates in our hospital by multilogistic regression using sequential adjustments.165 The final section of the analysis dealt with cesarean morbidity and mortality.We considered maternal as well as fetal and neonatal complications.For maternal complications, we estimated rates of intraoperative complications as well as of post-cesarean complications.We defined intraoperative complications as laceration of the uterus (uterine rupture included), cervix, bladder, vagina or bowel, intraoperative blood loss of ≥1000 ml, blood transfusion, and hysterectomy.Post-cesarean complications included post-cesarean infection, hemorrhage, deep venous thrombosis and puerperal psychosis.Regarding postcesarean infection, we specifically determined surgical infection rate as well as serious infectious morbidity rate.For surgical site infection we adopted the CDC definition as stated by Horan et al. [Horan et al, 1999].Serious infectious morbidity was defined as bacteremia, septic shock, septic thrombophlebitis, necrotizing fasciitis; peritonitis, or death attributed to infection.Risk factors for intra-operative complications and post-cesarean infection have been studied.We first computed crude odds ratios followed by adjusted odds ratios.We adjusted each factor for potential confounders in a multivariate logistic regression model.The final step in this analysis of maternal complications studied the trends of cesarean related maternal death risk and relationship between cesarean delivery and maternal mortality in a multivariate analysis of primary predictors including antenatal screening, referral status, maternal age, parity and route of delivery.In this analysis adjusted odds ratio have been produced for cesarean delivery.Regarding fetal and neonatal prognosis, we estimated trends of stillbirth rates and neonatal death rates.These indicators were studied by comparing cesarean to vaginal delivery.Stillbirth was defined as Apgar score = 0 immediately after delivery in a live-born-infant.Neonatal deaths are those occurring during the first 28 days following delivery.However, neonatal death rates presented are underestimated since our observation period was limited to the duration of hospitalization at birth; the maximum length of follow up of the neonates was 13 days.Neonates discharged healthy were assumed to have survived to 28 days.Nonetheless, the rates presented give an idea of the size of this important issue.
All calculations were performed using SPSS version 11.0 (SPSS Inc, Chicago, IL).P<0.05 was considered statistically significant.The database used for this analysis was reviewed and approved by the ethics committee of the Faculty of Medicine, Pharmacy, and Dentistry at the University of Bamako, Mali.

Characteristics of deliveries in our teaching hospital
During a nineteen year period from 1985-2003, 17,721

Cesarean delivery rates
Observed rates of cesarean delivery and relative risk are presented in table 1 above.The coefficient of variation for overall cesarean delivery rates was 27.9, and the ratio of the highest (36.9%) to the lowest (12.5%) was 2.95, indicating significant variability in overall cesarean delivery rates during the 19 years.Using year 1985 as the reference, we noted a striking increase in the cesarean delivery rate through out the study period.Since 1991, the Observed cesarean rates were relatively higher in the 35 -50 years old age group, Bambara ethnic group, grandmultiparas, women residing outside of Bamako, and those referred from other health centers.Cesarean delivery rates for unbooked pregnancies varied between 24.5% and 45.0%.Rates for women who followed antenatal screening varied between 15.6% and 31.0%.

Indications of cesarean delivery
In practice, the decision to perform a cesarean relies on an array of parameters.There is no general consensus universally accepted way of reporting cesarean delivery indications.Absolute numbers and specific cesarean delivery rates per indication / risk factors for cesarean delivery appeared in table 3. We report here 3 systems of reporting these indications:

Classification of cesareans by mutually exclusive clinical indications
Two independent obstetricians were asked to review our database and to point out what was the major factor leading to the decision of cesarean.They reviewed together cases where they found different factors.The results are presented in table 2 below.Of note, pelvic contraction and suspected fetal distress were the most represented and showed an increasing pattern over time.1985 -1990 N=1033 1991 -1995 N=1577 1996 -1997 N=319 1998 -2001 N=1216 2002 -2003 1985 -1990, 74.6% for 1991 -1995, 63.6% for 1996 -1997, 59.5% for 1998 -2001 and 53.8% for 2002 -2003 (p<.001).

Indications
Uterine rupture, an absolute indication for cesarean delivery occurred in 2.6% of all the 17721 deliveries and was the indication of 10.1% of the 4517 cesarean deliveries.The time trends of uterine rupture were as follow: 1.8% of all deliveries recorded in 1985 -1990, 3.9% for 1991 -1995, 2.7% for 1996 -1997, 1.5% for 1998 -2008 and 0.7% for 2002 -2003.Of all women with uterine rupture, 94.7% of cases were diagnosed at admission examination in referred patients (92.5% with patients referred emergently).
Table 3. Trends in cesarean delivery rates (total number of deliveries in each category) for sociodemographic, pregnancy and delivery characteristics by time period.

Robson's ten group classification
To further examine trends in cesarean delivery according to patient demographics, we classified our population following Robson's rules (table 4).Collectively, groups 1, 3 and 5 constituted 78.2% of deliveries.Their cesarean delivery rates are 22.6%, 13.2% and 76.4% while their contributions to total cesarean deliveries were respectively 18.14%, 26.54% and 20.86%.Although group 2 and 4 had high levels of cesarean delivery rates (91.9% and 46.3% respectively), they contributed only 8.95% of total cesarean deliveries.Trends in cesarean delivery rates for each of the ten groups appear in figure 2. There were no significant changes in abdominal delivery for Robson's group 2 (Nulliparous, single gestation, cephalic presentation, ≥ 37 weeks gestational age, induced or cesarean delivery before labor) and group 5 (Previous cesarean delivery, single gestation, cephalic presentation, ≥ 37 weeks gestational age).We observed an increasing cesarean delivery rates for groups 4, 8 and 9. Group 9 presented a two pattern aspect with rates shifting from around 75% before 1990 to around 85% thereafter.
To further understand variations in obstetric practice in our hospital, groups 1 (spontaneously laboring term nulliparas) and 2 were combined as a composite variable, the term TSCN (table 1).The annual TSCN cesarean delivery rate and contribution of TSCN to hospital deliveries are documented in Table 1.The mean cesarean delivery rate in TSCN was 25.8% (range, 10.3% -40.7%).The CV for TSCN cesarean delivery rates was 29.5%, again indicating significant variation between different years.The 19 year trends of cesarean delivery rate in TSCN follows a pattern similar to that of overall cesarean delivery rate (Table 1).Figure 1 demonstrates positive correlation between the overall and TSCN cesarean delivery rates over time (Slope = 0.876).Linear regression model suggested that 77% of the variation of the overall cesarean delivery rates can be explained by the variation observed in TSCN cesarean delivery rates (p<0.001).Our analyses suggest that the increase in overall cesarean delivery rate was not related to changes in obstetric groups since the proportion of all deliveries that were TSCN did not vary substantially.The average proportion of TSCN in this study was 21.2% (range, 18.3 -24.4%) with a coefficient of variation of only 8.4% (Table 1).

Multivariate analysis
Finally, we performed multilogistic regression with sequential adjustment to identify explanatory factors for increased cesarean delivery rates.Unadjusted analysis revealed a 100% increase in the rates of cesarean delivery (2003 vs 1985, RR = 2).The best model identified referral status, cephalopelvic disproportion and history of previous cesarean delivery as 3 factors to account for the observed increases in overall cesarean delivery rates.However, this model explained less than half of the observed increase (Figure 4).Of note, controlling for maternal age, parity and marital status didn't affect the observed increase.Controlling for cephalopelvic disproportion alone explained 32% of the increase since we found an adjusted relative risk of 1.68.Adjusting simultaneously for cephalopelvic disproportion, referral status and previous cesarean delivery further decrease the adjusted relative risk to 1.58.We couldn't build another model better than this last one.
As expected, higher levels of abdominal delivery were observed in referred patients (table 2).Since 1986, 60 to 70% of emergency admissions during labor have resulted in cesarean delivery.The cesarean delivery rates for referred patient without emergency fluctuated from 40%-60%.Cesarean delivery rates for direct admissions were ≤ 10% before 1994 and 10-20% thereafter (figure 3).
Cephalopelvic disproportion was a common indication for cesarean delivery, with a mean rate of 39.6% of women delivering abdominally having some degree of CPD.The percentage of CPD in cesarean deliveries ranged from 30.3% in 1985 to 48.8% in 1999.Contracted pelvis constituted 87% of all CPD.Of note, 63.5% of all contracted pelvis were recorded in the referred patients, who generally came from poor rural environments.The high incidence of uterine rupture among this group may correlate with severity of pelvis contraction.
Table 5 presents an analysis of factors influencing the occurrence of intraoperative complications.Univariate analyses found four risk factors for intraoperative complications: admission during the active phase of labor (cervical dilatation ≥4cm), transverse lie, total length of labor more than 24 hours, and emergent referrals.However, in multivariate analyses, only emergent referral remained a significant risk factor with a 3.4 folds increase in the odds of intra-operative adverse events.Removing referral status from the multivariate analysis allowed two factors to be linked to intraoperative complications: ruptured membranes at admission (OR=2.1 [1.2 -3.7], p<0.01) and total length of labor (OR=1.9[1.1 -3.5], p<0.05).
Twelve of the 244 maternal deaths associated with cesarean delivery occurred before the intervention was performed (4.9%).A similar proportion occurred during cesarean.The vast majority of maternal deaths were recorded in the post-cesarean period (90.1%).
The absolute number and risk of cesarean-related maternal deaths shows a sharp decrease beginning in 1994 (Figure 5).Table 8 presents case fatality rates for direct and indirect maternal complications by route of delivery.The vast majority of maternal complications (91.8% for cesarean delivery, 83.9% for vaginal delivery) and maternal deaths (95.5% for cesarean delivery, 87.5% for vaginal delivery) were the consequences of direct maternal complications.The overall case fatality rate was 6.9% (244/3548) for cesarean delivery compared to 2.9% (104/3597) for vaginal delivery.In the cesarean delivery group among direct maternal complications, uterine rupture had the highest fatality rate (23.2%).There was a consistent decrease in the incidence and case fatality rates of uterine rupture in women delivered abdominally (figure 6).The incidence decreased from 10.5% for the period 1985 -1990 to 4.5% for the period 2002 -2003.The case fatality rates decreased from 29.9% to 6.9% in the same time periods.We examined the relationship between cesarean delivery and maternal death in the context of other known primary predictors (table 9).Cesarean delivery remained strongly associated with maternal death even after controlling for antenatal screening, referral status, maternal age, parity, abruption, placenta previa, hypertensive disorders, and malpresentation.Table 9.Odds ratios with 95% confidence interval for maternal death for primary predictors.

Stillbirth rates
Overall, the stillbirth rate for cesarean delivery was 19.3% vs. 7.3% for vaginal delivery (p<0.001).Since 2000, the gap between the two curves narrowed significantly (figure 7).Gestational age-specific stillbirth rates are shown in table 10.Preterm stillbirth rates were comparable for the two routes of delivery or higher in the vaginal route.However, there was a statistically significant difference for term stillbirth rates with higher rates observed in the cesarean delivery group.The risk of stillbirth associated with cesarean delivery was high in univariate analysis (2.9 [2.7 -3.2]).However, after adjusting for maternal age, parity, referral status, CPD, antepartum hemorrhage, hypertension in pregnancy, malpresentation and uterine rupture, the risk disappeared and cesarean delivery ws shown to be protective against stillbirth (aOR = 0.36 [0.30 -0.42]).

Neonatal mortality
Neonatal mortality rates over time are shown in Figure 8.Rates were generally higher for cesarean delivery compared to vaginal delivery.Neonatal death rates didn't vary significantly over time for either route of delivery.As expected, the younger the gestational age, the higher the neonatal death rate (Table 10).Univariate analysis revealed an increased risk of neonatal death when the delivery route was abdominal as compared to vaginal route (Table 11).However, after adjusting for maternal (age, parity, referral status), pregnancy (gestational age at delivery, booking status) and fetal / neonatal (suspected fetal distress during labor characterized by an abnormal heart beat rate and / or an abnormal amnionic

Strengths and limitations
We report here an analysis of deliveries during a nineteen year period in a teaching hospital in Mali (West Africa).Our main findings are: (1) a striking increase in cesarean delivery rates throughout the study period; (2) women in labor referred with an emergency condition not only constituted the largest proportion of cesarean deliveries but also this mode of admission seriously jeopardized the maternal as well as the perinatal prognosis; (3) the two most common indications for cesarean delivery were cephalopelvic disproportion and previous cesarean delivery; (4) there were very low rates of planned cesarean delivery as expressed by the small contribution of Robson groups 2 and 4.
While the majority of women delivering at Point G National Hospital originate from Bamako, the substantial proportion of women referred from other regions and the sociodemographic characteristics of the pregnant women in this setting provide a perspective on hospital-based obstetric care in the whole country up to 2003.Since 2003, Mali has instituted important changes in the delivery of obstetric care.In 2004, a nationwide emergency obstetric and neonatal care program was put in place to reinforce the perinatality program.In 2006, the Malian government began to provide medically indicated Cesarean deliveries free of charge in public hospitals and referral district health centers, increasing access and leading to subsequently higher Cesarean delivery rates (Teguete et al, 2010b).Additionally, since 2002, major staffs of the department of obstetrics and gynecology have been appointed to Gabriel Toure teaching hospital, a nearby hospital of the Malian capital city with easier accessibility.These aspects are not covered by the database used here.
Trends in Cesarean delivery rates after 2003 can be examined through an improved and adapted obstetric database installed at Gabriel Touré teaching hospital after the transfer of staff.This database contains more than 400 variables related to patients' demographic, medical and obstetric factors as well as pregnancy outcomes [Teguete et al, 2008;Teguete et al, 2009].Rates of Cesarean delivery at the Gabriel Touré teaching hospital increased from 21% in 2003 to 32% in 2009 (Teguete I. et al, 2010b).After sequential adjustment for maternal demographic, obstetric, and referral characteristics as described above, 19% of the observed increase remained unexplained (figure 9), compared to 58.5% for the Point G database.CPD, referral status and previous cesarean delivery were the major determinants of cesarean delivery, as at Point G teaching hospital.These findings will be the core of the following comments.

Cesarean delivery rates in developing countries
The current situation of cesarean delivery rates in developing countries is very complex with large differences between countries, within countries, and between health centers [Fesseha et al, 2011;Cissé et al, 1998].A large ecological cross-sectional study reported that, in low income countries where cesarean deliveries rates are less than 10%, as section rates increase, neonatal and maternal mortality decrease [Althabe et al, 2006].Our database revealed a sharp increase in the rates of hospital cesarean delivery, similar to those observed in many teaching hospital maternity wards in Africa [Muganyizi & Kidanto, 2009;Geidam et al, 2009;Kwawukume, 2001].However, countrywide cesarean delivery rates remain under the minimum level of 5% advised by the WHO for optimal obstetric care (table 12).
Rural populations remain underserved in many developing countries [Leone et al, 2008, Kizonde et al, 2006].A recent regional meeting for the final evaluation of the "Initiative 2010" aimed at reducing maternal mortality ratios and neonatal mortality rates by 50% by 2010 reported that among 17 West and Central African countries evaluated; only five had national cesarean delivery rates of 5% or more [Ba, 2011].Thus, in sub-Saharan Africa where coverage in healthcare service is low, initiatives to increase cesarean rates are ongoing in order to meet the millennium development goals [El-Khoury et al, 2011].At the same time, there are calls for caution [Mbaye et al, 2011;Fesseha et al 2011] to prevent high unnecessary cesarean rates like those observed in many developed countries as well as some developing countries [Khawaja et al, 2004].

Referral system and cesarean delivery
Like in many developing countries, access to healthcare for the poor and underserved remains insufficient in Mali.
After the publication of the now famous article "Where is the M in MCH?" [Rosenfield & Maine, 1985], and the introduction of the Safe Motherhood Initiative in Nairobi in 1987, maternal mortality reduction in sub-Saharan Africa garnered increased attention and commitment [UN.Report, 1994].In this context, maternal and child protection have become major targets in the implementation of the healthcare system by the Malian government.
Many developing countries paid special attention to the organisation of the referral system to improve maternal and child healh [Rudge et al, 2011].Likewise, the National Perinatality Program of Mali was conceived in 1994 and organised the referral system to improve the environment of perinatal care.Reported interventions at the community level focused on (1) educational activities to raise awareness of danger signs and encourage the use of obstetric services; (2) reducing geographical and financial barriers through emergency loan schemes / subvention and (3) improving transport and communication [Kandeh et al, 1997;Nwakoby et al 1997].This policy led to an increase in cesarean delivery rates in rural district hospitals [De Brouwere, 1997], but it was very difficult to implement in large cities like Bamako.The high incidence at Point G of uterine rupture, a preventable end stage obstetric morbidity, demonstrates the unmet needs of cesarean delivery.
Thus, like in many developing countries [Sørbye et al, 2011], access to emergency obstetric care is unsatisfactory in Mali and unequal.Despite a national obstetric referral system, many birthing women (often without complications or known risk factors) bypass referring facilities to get access directly to the higher level of obstetric care.On the other hand, many women without access to care have to travel long distances to access care during labor and delivery.Difficulties related to referral health systems are frequently reported in sub-Saharan Africa [Cissé et al;1998] and were common features in our hospital before 1994.Large population based studies emphasize the need to ensure that the women least likely to seek care are not marginalized [Jacqueline et al., 2003], requiring a functional referral system.
In Mali, access to cesarean delivery was a priority of referral system organisation from its inception.This system may be partially responsible for the decreased risk of caesareanrelated maternal death after 1994 as well as the downward trend in post-cesarean stillbirth rates.However, the risk of maternal death when caesarean delivery is needed is still high despite adjustment for other factors.The unsatisfactory initial impact of cesarean delivery on maternal and fetal / neonatal health led the Malian government to make it free of charge.Many other countries engaged in such political commitment to eliminate financial barriers.However some authors reported that, although removing user fees has the potential to improve access to health services especially for the poor, it is not appropriate in all contexts [James et al, 2006].Similarly, simulations have found that decreasing the price of Cesarean delivery has minor effects, suggesting that greater increases in access to care would come from investment in the improvement of healthcare structures and care processes [Mariko, 2003].Developing countries face serious issues in this respect, due to the lack of and inequitable distribution of human resources.For example, in 2002 in Mali, 265 midwives were posted in Bamako or in regional hospitals, while only 164 were working at the peripheral level.As a result, only 24% of deliveries were attended by a skilled professional.Similar figures have been reported from Tanzania [Olsen et al., 2005].Many basic health facilities do not even have a midwife, so, many patients have to come directly to the tertiary hospital or go nowhere at all [Gerein N et al, 2006].Many strategies have been or are being tested to solve this problem.Unfortunately, there is no one single-bullet solution [Dayrit et al, 2010].These gaps contribute to the poor performance of the health system [Lawn J E., 2009].Thus, a holistic approach has to be considered for better strengthening of the health system in order to meet the performance goals of the WHO schematic framework [WHO, 2007]

Cephalopelvic disproportion
The expression cephalopelvic disproportion (CPD) came into use prior to the 20th century to describe obstructed labor due to disparity between the dimensions of the fetal head and maternal pelvis that preclude vaginal delivery.This term, however, originated at a time when the main indication for cesarean delivery was overt pelvic contracture due to rickets (Olah & Neilson, 1994).CPD can be due to a contracted pelvis or a disproportionately large fetal head and is thus not limited to primary cesarean delivery only [Carbone B., 2000].
In a systematic review of cesarean delivery f o r m a t e rna l i n d i c a t i o n , D u m o n t A . e t a l [Dumont al, 2001] found that cephalopelvic disproportion was the commonest indication, and 1.4% to 8.5% of all deliveries resulted in cesarean birth for this indication.Similarly, a large population based study in West Africa reported that 1% of all deliveries were complicated by CPD [Ould El Jouda D et al, 2001].The proportion of all cesarean deliveries due to contracted pelvis (a sub-entity of CPD) has been reported to be between 20% in Senegal [Bouillin et al, 1994] and 37.3% in Bobo Dioulasso, Burkina Faso [Bambara et al, 2007].Comparable trends have been reported in Senegal with mean rates of 31.3% for CPD ranging from 26% to 34.9% between 1992 and 2001 [Cissé et al, 2004;N'Gom PM et al, 2004], as well as Ethiopia (34% [Fesseha et al, 2011]).Similar high incidence rates of CPD have been reported in non sub-Saharan developing countries [Festin et al, 2009].Our data pointed out the importance of contracted pelvis in CPD.Cephalopelvic disproportion was a major indication of cesarean delivery in our hospital from 1985 to 2003.A mean rate of 39.6% of women who delivered abdominally had some degree of CPD, ranging from 30.3% to 43.4% between 1985 and 2003.In addition, in our study contracted pelvis constituted the vast majority of all CPD (87%).63.5% of all contracted pelvis cases were found in patients referred most commonly from poor rural settings.The high incidence rates of uterine rupture (an end stage of obstructed labor) recorded in this group may correlate with severity of pelvis contraction and confirmed the close link between referred patients during labor / delivery and need of cesarean reported elsewhere [Amelink -Verburg et al, 2009].
The cause of high rates of contracted pelvis in rural areas may be due to several factors such as genetics, increasing recognition, or the impact of resource scarcity on the female bony pelvis [Cissé et al, 2004;Kurki, 2011].Special attention must be devoted to malnutrition in sub-Saharan Africa.189 high-risk cases and in the long term at improving nutrition.Early motherhood should be discouraged, and efforts are needed to improve nutrition during infancy, childhood, early adulthood, and pregnancy.Improving the access to and promoting the use of reproductive and contraceptive services will also help reduce the prevalence of this complication [Konje & Ladipo, 2000].

Previous cesarean delivery
One third of our patients with a history of previous cesarean delivery were allowed a trial of labor (TOL), and the probability of successful vaginal delivery in this group was 50.1%.In our guidelines for trial of labor after cesarean birth (TOLAC), neither labor induction nor labor augmentation were permitted.No TOLAC was attempted when the number of previous cesarean delivery was >1 or in the case of a previous history of uterine rupture.These strict criteria explained our relative low rate of TOLAC.
In a meta-analysis of 963 papers, the range for TOLAC and VBAC rates was large (28-82 percent and 49-87 percent, respectively).Predictors of women having a TOL were having a prior vaginal delivery and settings of higher-level care, namely tertiary care centers [Guise et al, 2010].Similar findings have been reported in sub-Saharan [Boulvain et al, 1997]; the percentage of TOL ranged from 37% to 97% across reports, with probability of successful vaginal delivery of 69% (95% CI 63-75%).Maternal mortality among all women with a previous cesarean section was 1.9/1000 (95% CI 0-4.3).Uterine rupture and scar dehiscence occurred in 2.1% (95% CI 1.0-3.2).With our restrictions on VBAC, we recorded fewer vaginal deliveries, but also less uterine rupture / dehiscence, as was found in rural Zimbabwe [Spaans et al, 1997].In settings where such cautions were not applied, higher morbidity levels were observed [Olagbuji et al., 2010, Adanu & McCarthy, 2007;Olusanya & Solanke, 2009;Sepou et al, 2003;Nwokoro et al 2003;Oboro et al, 2010;Wanyonyi & Karuga, 2010].A large multicenter propective study in a western country with a uniform and well organised delivery care system emphasized the greater perinatal risk associated with a trial of labor [Landon et al, 2004].Although these findings can be a subject of debate [Greene, 2004]; they must be considered and women deserve to be well informed of the risks and benefits of TOL and VBAC [Kraemer et al, 2004].A recent systematic review suggests that VBAC is a reasonable choice for the majority of women and found that adverse outcomes were rare for both elective repeat cesarean delivery and trial of labor [Guise JM et al., 2010a, 2010b].The need for studies identifying patients at greatest risk is of primary importance in sub-Saharan Africa where high levels of morbidity are often reported.
Overall, in sub-Saharan Africa a selective policy of trial of labor after a previous cesarean delivery has a success rate comparable to that observed in developed countries.Vaginal birth after cesarean appears to be relatively safe and applicable in this context and contributes significantly to the global cesarean delivery rate.

Low rates of elective delivery
There are many reports emphasising on the high levels of emergency delivery in sub-Saharan hospitals and health centres [Fesseha et al, 2011;Shah et al, 2009;Wylie & Mirza, 2008;Dumont et al, 2001].In contrast to wealthier countries, planned delivery remains an underused option in Sub-Saharan Africa [Stavrou E. P. et al, 2011].For example, during the two decades at Point G teaching hospital, only 212 pregnant women underwent labor induction.In many sub-Saharan African countries, labor induction is not common as necessary medications are not readily available.Before the year 2000, oxytocine was the only medication available in Mali for labor induction and was only used for very favorable cases with a Bishop cervical score ≥ 7. Despite the lower rates of labor induction, we observe a mean rate of post-induction cesarean delivery of approximately 90% (Robson group 2).High levels of cesarean delivery following labor induction in nulliparas have been qualified as universal [Brennan et al, 2009 ;Main et al., 2006;McCarthy et al, 2007;Robson, 2001;Costa et al 2010;Yeast et al, 1999] Unpublished data from the WHO Global Survey on Maternal and Perinatal Health, which included 373 health-care facilities in 24 countries and nearly 300 000 deliveries, showed that 9.6% of the deliveries involved labor induction.Overall, the survey found that facilities in African countries tended to have lower rates of induction of labor (lowest: Niger, 1.4%) compared with Asian and Latin American countries (highest: Sri Lanka, 35.5%) [WHO, 2010].
One point is that many indications of labor induction are associated with preterm delivery.The lack of neonatal resuscitation facilities [Hofmeyr et al, 2009] and the poor outcomes of preterm neonates lead many sub-saharan obstetric teams to avoid preterm labor induction or preterm elective cesarean delivery.Even in hospitals, staffs are frequently not trained in resuscitation and equipment is not available.A national service provision assessments in 6 African countries demonstrated that only 2%-12% of personnel conducting births had been trained in neonatal resuscitation and only 8%-22% of facilities had equipment for newborn respiratory support [Wall et al, 2009].This important gap certainly impacts clinical decision making in Sub-Saharan obstetric units.
Therefore, it is a challenge for healthcare workers and policymakers dealing with pregnancy management in developing countries to examine critically ways to increase percentage of planned delivery.This challenge can be met firstly with preventive measures at a public health level (e.g.counselling and education), at the pregnant women's level (e.g.improved utilisation of the antenatal care services), and at the caregiver's level (e.g.better selection of cases in order improve the percentage of pregnant referred without emergency, an overt contracted pelvis mustn't begin labor at the level of primary care where obstetric surgery is not available).

Conclusion
From 1985 to 2003, cesarean delivery rates at Point G National Hospital increased substantially.Most of the increase in cesarean delivery rates is explained by higher proportions of outside referrals, cephalopelvic disproportion, and history of previous cesarean delivery.The increased cesarean delivery rate cannot, however, be fully explained by these factors or other characteristics collected by this study, and is likely the multifactorial impact of psychosocial determinants of healthcare utilization and systemic problems of healthcare delivery.These variables are beyond the scope of this study.Future ecological studies addressing clinical, financial, and geographical considerations as well as cultural acceptability of cesarean delivery are needed.Since emergency referrals for caesarean during delivery significantly worsen the maternal and fetal prognosis, more Cesar ean del i ver y r at es f ur t her adj ust ed f o r l abo r i nd uct i on Cesar ean del i ver y r at es f ur t her adj ust ed f o r r ef er r al st at us Fig. 9. Observed rates of cesarean delivery and adjusted rates after sequentially adjusting for age, parity, maternal conditions, obstetric practice and referral status (Teguete I. et al, 2010b).

Acknowledgments
To Evan and Lauren Oreinstein, for their valuable contribution to this work.
www.intechopen.comDetermining Factors of Cesarean Delivery Trends in Developing Countries: Lessons from Point G National Hospital (Bamako -Mali) 171 Fig. 1.Correlation between overall cesarean delivery rates and cesarean delivery rates in TSCN.

Fig. 5 .
Fig. 5. Trends of annual number and risk of cesarean related maternal deaths.

.2 Classification of Baltimore group on cesarean indications
Determining Factors of Cesarean Delivery Trends in Developing Countries: Lessons from Point G National Hospital (Bamako -Mali) 169 previous cesarean delivery, however, increased over time.In 1985 -1990, 2.90% of cesarean deliveries were performed primarily because of breech presentation; this rate reached 5.64% during 2002 -2003.These rates were 1.06% and 4.30% respectively for eclampsia and 7.84% and 16.66% for previous cesarean delivery. www.intechopen.com

Table 4 .
Prevalence of Robson ten groups ; cesarean delivery rate by group and contribution of each group to cesarean delivery.

Table 5 .
Risk factors for intraoperative surgical complications.

Table 6 .
Post-Partum Infectious Complications by Delivery Route.

Table 7 .
Risk factors for post-partum infection.

Table 8 .
Absolute numbers of cases, number of deaths and case fatality rates of direct and indirect maternal complications for cesarean delivery and vaginal delivery.
Fig.6.Time trends of uterine rupture incidence rate and case fatality rates among 4517 cesarean deliveries at Point national hospital,Bamako, Mali, 1985 -2003.

Table 10 .
Total numbers of births, stillbirth, and stillbirth rates and neonatal death for cesarean and vaginal delivery,Point G teaching hospital, 1985Point G teaching hospital,   -2003.   .
Determining Factors of Cesarean Delivery Trends in Developing Countries: Lessons from Point G National Hospital (Bamako -Mali) 191 widespread access to facilities capable of performing caesarean sections as well as earlier referrals of high risk pregnancies from primary health centers before the onset of complications would likely lead to substantial improvements in maternal and neonatal outcomes.A holistic need assessment will govern improved healthcare delivery strategies and aid progress towards meeting the millennium development goals in developing countries.Observed cesarean delivery rat esCesar ean del i ver y r at es adj ust ed f o r age and p ar i t y Cesar ean del i ver y r at es adj ust ed f o r al l mat er nal char act er i st i cs Cesar ean del i ver y r at es f ur t her adj ust ed f o r mat er nal medi cal co nd i t i ons www.intechopen.com M.K. & McCarthy M.Y. (2007); Vaginal birth after cesarean delivery in the West African setting.International Journal of Gynecology and Obstetrics Vol .98,No3,(September 2007), pp.(227-231).PMID: 17603060.Althabe F, Belizán J, Villar J, Alexander S, Bergel E, Ramos S, Romero M, Donner A, Lindmark G, Langer A, Farnot U, Cecatti JG, Carroli G. & Kestler E.(2004).Latin American Cesarean Section Study Group: Mandatory second opinion to reduce