Surgical site infection and associated factors among women underwent cesarean delivery in Debretabor General Hospital, Northwest Ethiopia: hospital based cross sectional study

Hospital based cross-sectional study was conducted from May 20 to December 15 /2017.

This study was conducted in DGH located in Debretabor town of Amhara Regional State. Debretabor town is about 98 km away to the East of Bahir-Dar (the capital of Amhara regional state) and about 667 km away to North of Addis Ababa (the capital of Ethiopia). DGH is the only general hospital in the south Gondar Zone having over 102 beds. Obstetrics and gynecology service is organized to a ward with a capacity of 22 beds and an operation room for emergency cesarean deliveries, maternal and child health unit, gynecology referral clinic and safe abortion clinic. The service is provided by two obstetrician and gynecologists, five integrated emergency surgical officers and over 22 professional and diploma midwives. About 3600 deliveries are conducted every year of which about 700 of them are by CS.

Women who delivered by CS from May 20 to December 15/2017 at DGH were included in this study. Women who didn’t have tracing mechanisms (if either the woman and/or her attendant/family didn’t have phone number) after discharge were excluded.

Census approach was used to include women delivered by CS from May 20 to December 15/2017. There were a total of 359 CDs of which 12 of them were excluded due to absence of tracing mechanism after discharge. The remaining 347 study participants were included in the study but 334 of them were traced for 30 days (Fig. 1).

For the sake of clarity, some of the operational definitions used in this research include: SSI was defined as per the criteria of the Centers for Disease Control and Prevention (CDC) guideline [1]. ASA physical status was defined as assessment of the woman’s preoperative physical condition using the ASA Classification of Physical status [29].

Wound class was defined as assessment of the degree of contamination of a surgical wound at the time of the operation by a person involved in the surgical procedure [29]. Steroid therapy was defined as preoperative maternal treatment with either betamethasone or dexamethasone for fetal lung maturity. Duration of membrane rupture was determined by the time lapsed in hour from either mother’s complain of leakage of liquor or artificial rupture of membrane to skin incision. Duration of labor was determined by the time lapsed in hour from mother’s complain of symptoms of true labor to skin incision. Duration of operation was determined by the time lapsed in hour from skin incision to skin closure. EBL was determined by the number of packs and gauze used during time of operation multiplied by estimated absorbable amount of blood by a single pack (1 soaked pack = 100 ml) and gauze (1 soaked gauze = 20 ml of blood) respectively. Post-operative hemoglobin represented hemoglobin level in gram/deciliter after 8 hours of post-operative period. Obesity represented maternal preoperative BMI > 25 kg/m².

Ethical approval to conduct this study was obtained from department of midwifery Ethical Review Committee under the delegation of Institutional Review Board (IRB) of University of Gondar. Six diploma midwives (two per shift) from obstetrics and gynecology ward were used to collect both the risk factor and outcome data. Two other Bachelor of Science (BSc) midwives from Debretabor University were assigned to supervise the data collection process. Before data collection process the data collectors informed each study participant about the purpose and anticipated benefits of the research project. The study participants had also been informed on their full right to refuse, withdraw or completely reject part or all of their part in the study. They had also been informed that their address and phone number would be used for the sole purpose of follow up and all data obtained from them would be kept confidential. Written consent was obtained from study participants prior to data collection. Study participants who were ≥ 18 years of age gave their written consent. Study participants who were < 18 years of age gave their written consent via their family members. Family members used to provide consent on the behalf of participants under the age of 18 were participants’ husbands and either of the participants’ parents. Priority was given for husbands and parents were considered in cases of either husbands’ absence or husbands’ minority age (under 18 years of age). We can confirm that the ethics committee had approved the use of attendants (husband or either of the parents) on the behalf of participants under the age of 18.

Data were collected by face to face interview of study participants, direct observation of the CS procedure and by reviewing of study participants’ charts. A semi structured and pre-tested questionnaire/data extraction tool prepared in English, translated to Amharic (local language) and then translated back to English was used to collect data. Relevant preoperative and intra operative events were observed, asked and recorded. Postoperatively, surgical wounds were examined immediately before discharge and their status recorded. Discharging study participants were counseled about symptoms of SSI and the need to return to the hospital immediately for any infected wound complaints. Their relevant addresses (phone number, place of residence and the name of the nearest health care institution) were well documented. Regular telephone calls were made at 15th and 30th post operation day [14] with an option to communicate at any time if suspicion of wound infection arose. SSI was identified before discharge and/or by readmission and evaluation of those participants with concerns of wound infection during phone calls or came directly with complain of symptoms of wound infection after discharge based on the criteria of CDC [29].

Must meet at least one of the following criteria:

Proper designing and pre-testing of the questionnaires was made on 18 study participants at Debark district Hospital. Data collectors and supervisors had 1 day training before the actual data collection. Every day after data collection, questionnaires/data extraction tools were reviewed and checked for consistency and completeness by the supervisors and the necessary feedback was offered to data collectors. Any missed or incorrectly filled questionnaire was sent back to the respective data collector for correction. Epi Info version 7 was used to control Data entry errors. Data clean up and cross-checking was done before analysis.

All the questionnaires were checked for completeness manually and coded, then entered into Epi info 7 and exported to SPSS version 20 for further analysis. Descriptive analysis results were presented in the form of tables and text using frequencies and summary statistics such as mean, standard deviation and percentage. First Bivariate binary logistic regression analysis was done to determine the crude association of each independent variable with the outcome variable via enter method. Those crudely associated independent variables with the dependent variable using odds ratio with 95% confidence interval were entered via backward likelihood ratio for multivariate binary logistic regression analysis to adjust the influence of various independent variables (confounding effect) on the outcome variable. The Hosmer and Lemeshow (HL) test was used to check model fitness in our study. According to HL, if the significance value is more than 0.05, it indicates the model fits data well [31]. The HL test indicated that our final model was fit (p = 0.605). Odds ratio with 95% confidence interval were used to estimate the strength of the association between independent variables and dependent variable.

A total of 334 women were included in the study, making 96.25% response rate (Fig. 1). The age of the study participants were between 17 and 43 years with mean (±SD) age 26.38 ± 5.47 years. All of them were married and from Amhara ethnicity. About 96% of them were orthodox Christians and about 59% of them have attended secondary and above education. More than two-thirds (69%) of them were urban dwellers.

The health status of about 70% of study participants was ASA class I, about 8% of them were chronic hypertensive cases and 4.5% of them had preexisting anemia (Table 1).

About 43 and 50% of participants were primigravidas and primiparas respectively with respective mean gravity and mean parity status of 2.28 times and 1.59 times. The maximum number of gravidity and parity was nine pregnancies and eight deliveries respectively. About 78% of them had term pregnancy and bout 86% of them had ANC follow up. At least one form of PIH (either gestational hypertension, pre eclampsia or eclampsia) were diagnosed in 27.1% of study participants.

On labor status of study participants, about 75% of them had labor before the operation and about 43% of them had ruptured membrane before the operation with the mean duration of 14 h and 10.51 h respectively. Eleven percent of participants had either preoperative or intraoperative chorioaminitis based on clinical findings (Table 2).

Of the total operations about 86% of them were emergency CDs. About 31% of indications for CD was NRFHRP (Table 3), Abdomen was entered via pfannensteil incision for about 87% of operations and lower uterine segment transverse cesarean section was done for all cases. Almost all wounds were clean contaminated wounds and skin was closed via sub cuticular suture for about 93% of case. About 57% of operations were done by integrated emergency surgical officers and about 17% of participants had post operation haemoglobin value of less than 11 g/decilitre (g/dl) (Table 4).

Of the total 334 study participants, 27 of them had developed SSI, making proportion of 8.1% (95%Cl: 5.4, 11.6). Three of them were identified before discharge and the remaining 24 were traced through follow up and readmission. All SSI cases were superficial incisional wound infection developed from emergency CDs.

In bivariate logistic regression analysis, SSI was significantly associated with ASA class III health status classification, presence of both chronic hypertension and PIH, operation after onset of labor, operation after ROM, presence of chorioaminitis, midline skin incision, interrupted skin closure, and EBL of greater than 500 ml and post-operative hemoglobin of less than 11 g/dl. Among variables found to be significantly associated with development of surgical site infection using bivariate logistic regression analysis, PIH (AOR = 4.75, 95%CI: 1.62, 13.92), chorioaminitis (AOR = 4.37, 95%CI: 1.53, 12.50), midline skin incision (AOR = 5.19, 95% CI: 1.87, 14.37 and post-operative hemoglobin less than 11 g/dl (AOR = 5.28, 95%CI: 1.97, 14.18) were also found to be significant in multivariate logistic regression analysis (Table 5).