Prevalence, outcomes and associated factors of labor induction among women delivered at public hospitals of MEKELLE town-(a hospital based cross sectional study)

Background Induction of labor refers to iatrogenic stimulation of uterine contractions before the onset of spontaneous labor as a therapeutic option when benefits of expeditious delivery outweigh the risks of continuing the pregnancy. This research was to study the prevalence, outcomes and associated factors of labor induction among women delivered at Ayder comprehensive specialized hospital and Mekelle general hospital in Mekelle town, Tigray, North Ethiopia. Methods A hospital based cross sectional study was conducted on 346 laboring mothers who delivered after induction of labor, from January 1st, to July 31st, 2017. Using structured questionnaire and quota sampling techniques, all eligible participants were immediately enrolled upon admission until the desired sample size was achieved. SPSS windows version 23.0 was used for analysis and both descriptive and inferential statistics were conducted; statistical significance to declare relationship between the dependent and independent variables was set at p < 0.05. Results Total of 3834 women delivered at the study area out of which 346 were induced making (9%) prevalence of induction. Out of this, 244 (70.5%) delivered vaginally, 19 (5.5%) were instrumental deliveries and 83 (24%) by Cesarean section, induction was successful in 263 (76%) while the failure rate was 25 (7.2%). All who failed induction (25) were delivered by cesarean section making a 3.3% contribution of failed induction into the overall rate of the institutions cesarean deliveries during the study period. Prolonged rupture of membranes was the commonest indication and Bishop’s score after cervical ripening significantly predicted the success of induction [AOR = 8.150, 95% CI = (1.265, 52.526)]. Conclusion Our prevalence of labor induction is very low compared to the rate of other institutions in developed countries, rate of successful inductions (76%) is slightly higher than the rate of similar institutions in Ethiopia but comparable to the regional rates while failed induction is very low in comparison to both local and regional institutions. Bishop’s score significantly predicted the success of induction.


Background
Induction of labor refers to the iatrogenic stimulation of uterine contractions, (before the spontaneous onset of labor), as a therapeutic option when the benefits of expeditious delivery outweigh the risks of continuing the pregnancy [1]. Rates of induction of labor varies from region to region with progressive increase to nearly doubling of incidence in some of the developed countries [2,3]. In the United States of America and United Kingdom about 20% of all deliveries are by induction of labor, with some institutions going to an incidence of as high as 40%, while 11.4% is reported in Latin America [2,4] and an average of 4.4% in Africa [5].
Methods for induction of labor may be divided into mechanical and pharmacological both of which have their own advantages and drawbacks [6]. Choice among this methods depends on many factors, including cervical condition, presence or absence of uterine scar, parity, available resources and obstetrician preference [6].

Statement of the problem
As compared with expectant management, induction of labor (IOL) might be associated with better maternal and perinatal outcomes as well as adverse outcome depending on many factors, despite this outcomes it is the most commonly performed obstetric procedure worldwide [7,8].
According to WHO secondary analysis study on unmet need for induction of labor, African rates of induction of labor are still very far from the expected, averaging 4.4% and 60-80.2% unmet need for labor induction [4,5].
Although there is no commonly accepted definition of "failed labor induction." the current practice acknowledges that "allowing at least 12-18 h of latent labor before diagnosing a failed induction may reduce the risk of cesarean delivery [9]", Options of management following failed induction of labor include a further attempt to induce labor after consultation with the patient or performing a caesarean section [6], Ethiopian national guideline for induction recommends Cesarean section only after declared failed induction.
Although induction of labor is a daily practice at public and private health institutions in Ethiopia, including the study area, there is a limitation in undertaking a study on prevalence, outcomes and associated factors of labor induction. Most of the studies concentrated on the success or failure rate of labor induction as a sole outcome, despite abundant literature on various outcomes [4,8,10].

Justification of the study
Since Ethiopian demographic health survey database do not include information on induction of labor, this study will stimulate further studies perhaps nationwide to address national rate of induction.
Induction when successful results in vaginal delivery but sometimes fails with potential adverse outcomes including high Cesarean delivery rate, for this reason the current international guidelines allow at least 12-18 h of latent labor before diagnosing failed induction and further attempt to induce labor after consultation with the mother, while Ethiopian national protocol for induction recommends only Cesarean section after declared failed induction, undertaking this study will help us knowing the contribution of failed induction into the overall institutional cesarean section rate as well as the other adverse outcomes. This can be used as evidence that we can build on with more well designed randomized control trials to help the policy makers revising the national guideline.
Knowledge of the outcomes of induction at the institutional level will be employed as a database to monitor the rate, common indications and outcomes for future improvement of the quality of care as well as an evidence based information for counseling of mothers for induction.

Methods
This study was conducted from January 1st, 2017 to July 31st, 2017 at Ayder comprehensive specialized Hospital (ACSH) and Mekelle General Hospital (MGH) which are public Teaching hospitals located in Mekelle town, the capital city of Tigray national regional state, north Ethiopia, and 789 Kilometers from Addis Ababa, The capital city of Ethiopia.
Both hospitals are used by Mekelle University College of health sciences as teaching institutions; they have seventy four obstetrics beds and annual deliveries of 5000 to 5500. During the study period the Department of Obstetrics and Gynecology had 8 specialists, one subspecialist, 35 residents at different residency levels (year1 -year4), 41 nurse midwifes and variable number of interns.
The study population was pregnant women who voluntarily accepted to participate in the study and admitted after 28th completed weeks for termination of pregnancy by induction of labor as per the national guideline.
The sample size will be calculated using the single population proportion formula: Where: n = sample size. Z = standard normal distribution corresponding to significance level at α = 0.05. p = proportion of labor induction success. "P", proportion of labor induction success is assumed to be 65.7% which was derived from study conducted at Jimma University Specialized Hospital [9]. d = margin of error 5%. Since both selected hospitals have the same case load and specialty care, the calculated sample size was equally divided into two giving a sample size of 173 per each hospital (Fig. 1).
After accurate revision of medical records as well as direct interview of patients, using structured questionnaire and quota sampling techniques, all eligible participants were immediately enrolled upon admission until the desired sample size was achieved.
Upon enrollment and through the period of labor and delivery, information regarding socio-demographic factors, obstetric history and physical examination findings, indication for induction, methods used for induction and outcomes, was collected within 6 hours after delivery from the medical records files, operating theatre and neonatal intensive care unit records using structured and comprehensive questionnaire designed for that. Oral interview was also used when necessary and missed information was resolved with the managing Physicians before discharge of the mother and her neonate.
One day training was conducted for data collectors (3 midwives and 2 year one residents) to outline the objectives and relevance of the study, who are the eligible participants, procedures of data collection techniques and the whole contents of the questionnaire. The questionnaire was pre-tested by taking 5% of the calculated sample size and necessary modification were made based on the nature of gaps identified. The principal investigator supervised data collectors on daily basis through revision and checking for completeness, accuracy and consistency. Research data were stored in a way that ensured security and confidentiality.
After data collection, each questionnaire was coded and entered in to prepared template of SPSS Version 23; frequency counts were conducted to check for accuracy, consistency and missed values. Age and parity were reclassified to facilitate analysis. The prevalence of labor induction as well as the rate of cesarean deliveries for failed induction was calculated out of the total deliveries during the study period. Descriptive statistics like frequency tables, mean, standard deviation, minimum and maximum, when applicable and descriptive summaries were used to describe the study variables. Associations were tested using chi square or Fisher's exact test as appropriate; the level of statistical significance was set at p < 0.05. Multivariate logistic regression analysis was employed to assess independent determinants of successful induction of labor. Odd ratios (95% confidence intervals) were used to determine the association of different factors with the success of induction.
Ethical clearance to conduct the study was obtained from the Ethics and research committee of Mekelle University, College of Health Sciences [ERC0898/2017].
The aim and purpose of the study was explained to all participants using patient information sheet. Participants were enrolled to the study only when they gave informed and written consent that they are willing to take part in the study.
Study participants were informed about their right not to give response to a question, and their right to withdraw from the study if they felt any discomfort. It was also emphasized that their withdrawal will not affect their management. The participants were followed throughout labor and postpartum without intervention apart from the standard procedures for labor induction as outlined by the national guidelines.
At Ayder Referral hospital both mechanical (balloon catheter) and pharmacological (vaginal misoprostol, oxytocin) methods are employed for induction of labor depending on favorability of the cervix and the indication for induction. With unfavorable cervices, defined by Bishop's Score less than 4 (Table 1), 25-50 μg vaginal misoprostol given in 6 hours intervals and maximum of 200 μg is the ripening agent used, sometimes, especially in cases of severe oligohydramnios, balloon catheter is used.
With the use of misoprostol some women goes into active labor and some require subsequent induction with oxytocin. Oxytocin induction chart was utilized to monitor the maternal vital signs, fetal heart rate, uterine activity, rate of oxytocin infusion and the dose of oxytocin.
Ayder referral hospital is following the national protocol [20] for induction of labor using 5 units of oxytocin in 1000 ml of IV fluids (normal saline in most of the times) for nulliparous, starting by 5mu/min with dose increment every 30 min till adequate uterine contractions are attained ( Table 2).

Results
During the study period, a total of (3834) woman delivered at the study area, out of which (346) were induced making a prevalence of induction of (9%).Total number of cesarean deliveries in this period were 770, out of which 25 were done for failed induction of labor making 3.3% contribution of failed induction into the overall rate of Cesarean deliveries during the study period.
The mean family monthly income of the participants was 4169.4 Ethiopian Birr, a maximum income of 14,000 and a minimum income of 700 with standard deviation of 2130. (Table 3):

Methods of ripening and induction of labor
Cervical ripening was done for 225 (65%) (those with unfavorable Bishop's Score), misoprostol was the most    Table 5).

Outcomes of labor induction Birth outcomes
Out of the total number of induced mothers, 244 (70.5%) delivered spontaneously, 19 (5.5%) were assisted by instrumental delivery and 83 (24%) by Cesarean section. By excluding 58 cases that had cesarean section for reasons other than failed induction, induction was successful in 263 (76%) and the failure rate was 25 (7.2%).
The only maternal death in this study was a multiparous woman that was induced by oxytocin and delivered successfully vaginally. After delivery she developed refractory postpartum hemorrhage secondary to uterine atony and measures including bimanual compression of the uterus, B-Lynch and finally total abdominal hysterectomy were done. She was admitted and died in the ICU the second day (Table 7).

Fetal/ neonatal peripartum complications
In regard to fetal /early neonatal complications, 62 (17.9%) of the neonates had NRFHRP, 2 (0.6%) IUFD, 34 (10.9%) low Apgar score, 10 (3.2%) Early onset neonatal Sepsis and out of this 37 were admitted to the neonatal intensive care unit and half of them stayed for 4-7 days at the NICU and 4 neonates died within the first 7 days of life. Two of the four early neonatal deaths were due to Hyaline membrane disease secondary to prematurity, one due to congenital malformation and another one for perinatal asphyxia (Table 8).

Association of selected factors to the outcome of induction Cross tabulation
Using the chi-square test value, Maternal age, Residence, parity, gestational age whether the baby was alive or not before induction, Bishop score at admission, whether cervical ripening was done or not, the ripening agent used, bishop's score after ripening, Method of induction, outcome, birth weight, 5th minutes Apgar score, Neonatal resuscitation and admission to NICU were selected and analyzed for their possible association with the outcome of induction. It was found that parity (P value = 0.006), Bishop's score after cervical ripening (P value = 0.001) and Method of induction (P value = 0.000) are the only significant variable (Tables 9, 10 and 11).

Logistic regression
On

Discussion
In this study the prevalence of labor induction was found to be 9%, while successful induction constituted 76%, failed induction represented 7.2% with 3.3% contribution of failed induction into the overall institutional rate of cesarean deliveries during the study period. The most common indication for induction was prolonged rupture of membranes; Bishop's score after cervical ripening was a significant predictor of successful induction.
The rate of labor induction (9%) was found to be slightly higher than other studies done at two teaching hospitals in Addis Ababa (4%) [11]. Although it is not a nationwide study, it is also higher than the average rate of induction in Africa (4.4%) ranging from (1.4%) in Niger to [6.8%] in Algeria and lower than the average rates in Asia (12.1%), ranging from (2.5%) in Cambodia to 35.5% in Sri Lanka [5]. This rate was found to be comparable with study done at Nigerian catholic maternity Hospital [12].
Induction of labor is directly relevant to the health related millennium development and sustainable goals (MDGs &MSGs), Given the increasing attention to    reducing perinatal and maternal morbidity and mortality, rates of induction of labor have continued to rise over the past few decades specially in the developed countries, the issue that has contributed a lot in the reduction of their maternal and perinatal mortality and morbidity. Induction rate of 9% reported in this study is comparable to the rate of some institutions in developed countries about thirty years back, confirming that our rate is still very low. Reasons for this low rate might be due to variation in practice since our guidelines don't allow for elective induction, induction after cesarean section or high unmet need for induction.
Cesarean delivery is one of adverse maternal outcomes of labor induction specially when it fails, the contribution of failed induction into the overall rate of cesarean deliveries in this study (3.3%) is relatively small and this can be explained by the low failure rate of labor induction (7.2%) compared to other studies done in similar institutions in Ethiopia. In Hawassa Public Hospital (17.3%) [13], Jimma University specialized hospital the failure rate was (21.4%) [14], Addis Ababa Teaching Hospital (28.4%) [11] and Addis Ababa Army Hospital (40.3%) [15].
Successful inductions were 76% of all inductions and goes to 91.3% if we exclude the other cesarean deliveries done for indications other than failed induction; this rate is slightly higher than the studies of Addis Ababa Army Hospital (59.7%) [15], Jimma University Hospital (65.7%) [14], Hawassa (61.6%) [13]. And it is comparable (75.9%) to a study done at Nigerian Catholic Maternity Hospital [12]. These variation can be due to the use of Oxytocin infusion as the most common Method of induction in these local institutions, while titrated oral misoprostol was used as one of the methods of induction in our study area as well as the Nigerian study, explaining the comparability to this studies [12]. Similar to many other local and regional institutions [5,[13][14][15][16], prolonged rupture of membranes, 143(41.1%) was the most common indication for labor induction, while elective induction was the most common indication in Asia [5]. This can be explained by the commonly shared risk factors for prolonged rupture of membranes both locally and regionally, and the difference in practice guidelines that allow elective induction as in Sri Lanka.
Bishops score after cervical ripening was found to be the only predictor of successful induction in this study hence induction is eight times more likely to succeed if the Bishop's score is favorable, [AOR =8.150, 95% CI = (1.265,52.526)], unlike other studies that found age, parity and Bishop's score at admission [14,15]. This might be due to the wide confidence interval having more successful inductions and few who had failed.
As a limitation of this study, cross sectional studies are of low level of evidence in comparison to other study designs and they cannot be taken as strong evidence to change practice. Another limitation was that Assessment of Bishop's score was done by more than one resident throughout the study period making it difficult to assess its precision.

Conclusion
The study revealed the prevalence of induction of labor to be 9%, which is very low in comparison to the rate of other institutions in developed countries. The rate of successful inductions (76%) is slightly higher than the rate of similar institutions in Ethiopia but comparable to the regional rates, while the rate of failed induction (7.2%) is very low in comparison to both local and regional settings. Like many local and regional Hospitals, prolonged rupture of membranes was the most common indication for induction. Bishop's score after cervical ripening significantly predicted success of induction and induction was eight times more likely to succeed if the Bishop's score was