Abstract
The delivery room logbook entries of Gimbie Adventist Hospital for a period of one month were monitored and compared with the actual number of births recorded on a separate worksheet for that period. The implications of the missing data were reviewed. This data was compared to a previous audit of the same design. Eighty per cent of births were recorded during this one month period which reflected an improvement from 72% reported in a previous study. Underreporting resulted in discrepencies when calculating the number of births and signal functions such as caesarean section rates, blood transfusion, administration of parenteral anticonvulsants and removal of retained products. In turn, these discrepencies impact the calculation of process indicators of safe motherhood projects. The reliability of the delivery room logbook as the sole source of information to create health policy and to monitor and evaluate health programs is questionable.
Introduction
The delivery room logbook is the primary tool used to record all births and their outcomes. It serves several purposes. For the health-care user, it can provide proof for a birth certificate. Within the health facility itself, it can be used to track maternal and newborn health, to identify maternal and neonatal deaths and the trends and outcomes of treatment. Regionally and nationally, it can provide demographic pictures of communities, such as the incidence of maternal mortality, which in turn helps to create health, social and financial plans and influence policy decisions.
Typically, it is the birth attendant who completes the delivery room logbook. The amount of information entered varies from institution to institution, but generally includes the name, address, hospital number, date and time of delivery, method of delivery, personnel present at delivery, sex, live birth/stillbirth, Apgar score, birth weight and gestation and blood loss. In high-resource countries it has been reported that, when comparing logbook entries to the details within a patient's chart, 5–60% of entries in the logbook contained errors. 1,2 Little information exists regarding the accuracy of records in low-resource settings.
Data derived from delivery room logbooks can be useful for tracking the progress of maternal health programmes within a health-care facility as they indicate the type of services, known as signal functions, that are being delivered. During an initial needs assessment of a safe motherhood programme, the delivery room logbook is a reference point for the identification of signal functions that help to determine whether the facility provides basic or comprehensive emergency obstetric care services. The generally accepted list of signal functions in safe motherhood projects is indicated in Box 1. As part of the monitoring and evaluation of a maternal health project, these signal functions are reassessed every three to six months to ensure that the health facility is providing all six or eight of these functions.
Obstetric care signal functions
Basic emergency obstetric care
Administer parenteral antibiotics Administer parenteral oxytocic drugs Administer parenteral anticonvulsants for gestational hypertension Perform manual removal of placenta Perform removal of retained products Perform assisted vaginal delivery
Comprehensive emergency obstetric care (EmOC)
All the functions included in Basic EmOC PLUS:
(7) Perform surgery (caesarean section) (8) Perform blood transfusion
Information from delivery room logbooks may also be used to estimate process indicators. 3 In maternal health programmes, the standard process indicators used to monitor the availability and use of obstetric services is illustrated in Box 2.
Obstetric service process indicators
*The delivery room logbook may be used as a data source to calculate this indicator
Finally, the delivery room logbook is a source of case identification of maternal deaths in order to conduct a maternal death review as part of a maternal mortality surveillance system. 4 The maternal mortality surveillance cycle is designed to improve obstetric services in a health-care facility and to measure its progress.
The delivery room logbooks can provide pertinent information about events and the outcomes of pregnancies and are important tools for the monitoring and evaluation of maternal health projects and programmes. In our study, we assessed the reliability of the delivery room logbook as such a tool.
Description of the study setting
The study was conducted from 20 October to 19 November 2005 at the Gimbie Adventist Hospital (GAH) in Gimbie Town, in the West Wollega zone, about 450 km west of Addis Ababa in Ethiopia. The total population of West Wollega is 2.03 million, of which 35,000 live in Gimbie Town. The catchment area for GAH is approximately 500,000.
Maternity Worldwide (MW) has been working in Gimbie since 2003, providing a comprehensive safe motherhood programme which includes the provision of essential obstetric care services, community health education and a network of income-generating women's groups. MW's overall aim is to increase the proportion of women with complications attending emergency obstetric care facilities and to improve the quality of care provided in order to reduce unnecessary maternal deaths.
In 2003 MW undertook an initial assessment of maternal health-care services in West Wollega and compared it against UNICEF/World Health Organization (WHO)/UNFPA standards (as described in Box 1 and 2). Maternal health service provision in West Wollega was found to be woefully inadequate with only 2.1% of births taking place in emergency obstetric care (EmOC) facilities. Additionally, the caesarean section rate for women in West Wollega was approximately 0.4% of all births. These figures compare to the WHO standard of a minimum of 15% of births requiring EmOC services and a minimum of 5% of women requiring caesarean section. In 2003, the case fatality rate at GAH was 2.2% compared to the standard of <1%. 5
Method
For this particular study, a maternity worldwide midwife trainer collected data twice a day before each shift change. This involved checking the occupants of beds in the female ward, the private ward and the intensive care unit (ICU) as well as speaking with staff about admissions and events that had occurred during the preceding shift. All information was tracked on a separate worksheet and patient records were used to fill out details of the births. After one month a comparison was made between information recorded on the worksheet by the data collector with what was recorded in the delivery room logbook by hospital staff. In order to ensure an accurate picture of the birth registration process, the hospital staff was unaware that the study was being conducted.
This study is a follow-up to a similar audit that was performed and conducted in the same manner from 14 November 2004 until 13 January 2005. After the first audit, all health providers at GAH received instructions on how to complete the delivery room logbook. The results of this study were compared to results of the initial study in order to evaluate the improvement or deterioration of the birth registration process.
Results
Summary of births conducted at Gimbie Adventist Hospital
From 20 October until 19 November 2005 the delivery room logbook reported 46 births at GAH compared to 57 births that were recorded in on the worksheet, i.e. only 80.7% of all births were recorded.
Of the 11 unrecorded births, 10 (90.9%) involved a complication requiring the assistance of a skilled attendant. According to the worksheet there were 37 (64.9%) complicated births. The delivery room logbook reported that 29 (63%) births were complicated.
During the study period, there were no maternal deaths. During the initial 2004–2005 audit, only 72% of the births were recorded. Our study reflects an improvement of 8.7%.
A description of the unrecorded births and their outcomes is summarized in Table 1.
Unrecorded births and their outcomes
SVD, spontaneous vaginal delivery; D&C, dilation and curettage
When births are not documented, information regarding the procedures (signal functions) is lost. During the period of this study 91% of the unrecorded births required an intervention: assisted vaginal delivery, caesarean section, parenteral anticonvulsants, blood transfusion and removal of retained products.
The omission of these births affects the overall picture of events that occurred in the logbook and paints a different picture of what really happened. A comparison of the delivery room logbook and the worksheet that reveals the discrepancies is presented in Table 2.
Summary of unrecorded events
Assisted vaginal deliveries
An assisted vaginal delivery is a forceps or vacuum delivery. Although the two forceps deliveries were accounted for, there were five vacuum deliveries reported when six had actually occurred. In this case it would appear that 15.2% of all deliveries required an instrumental delivery when the actual figure was only 14%.
Caesarean sections
During the study period, 15 of the 57 women admitted received a caesarean section giving a caesarean section rate of 26.3%. However, the delivery room logbook recorded that only nine women of 46 admitted women received a caesarean section, a rate of 19.6%. This resulted in a 26% underreporting of caesarean sections.
Administration of parenteral anticonvulsants
There were two cases of complicated gestational hypertension requiring parenteral anticonvulsants during the study period. However, the delivery room logbook stated that there was only one case. The reported incidence of gestational hypertension for the month was 2.1% versus the 3.5% that actually occurred. This results in a 40% underreporting of the incidence of complicated gestational hypertension.
Use of blood transfusion to treat postpartum haemorrhage
During the study period there were a total of three women who required a blood transfusion for postpartum haemorrhage: the delivery room logbook recorded only two. The incidence of the use of blood transfusions was reported as 4.1% instead of the correct figure of 5.3%. The incidence of postpartum haemorrhage requiring a blood transfusion was underreported by 23%.
Removal of retained products
The unrecorded case of a postpartum haemorrhage also required a uterine evacuation. This was the only patient who received a uterine evacuation during the study period and reflects a 100% under recording of this event. The proportion of uterine evacuation that occurred was 1.8% and yet it was reported as not having occurred at all.
Observations
A review of the unrecorded deliveries was conducted in order to attempt to explain the logbook inaccuracies. A summary of the findings is shown in Table 3.
Circumstances surrounding an unrecorded delivery
SVD, spontaneous vaginal delivery
Discussion
Accurate documentation and record keeping play a central role in health-care policy and planning. As such, the delivery room logbooks serves multiple roles—they provide baseline data, measure progress and provide a means of accountability to funders. Conversely, inaccurate or incomplete data can have negative implications on the logistical, adminstrative and financial management of health-care facilities and programmes.
This study reveals that 20% of births were unrecorded and that 90% of these births required assistance. What are the implications of the missing data?
Impact of inaccurate delivery room logbooks: monitoring and evaluating signal functions
The unreported cases revealed missed oppportunities to properly measure signal functions such as surgery, blood transfusion, removal of retained products and parenteral administration of anticonvulsants by margins between 23–40%. Of the 11 unreported cases, one created a complete misrepresentation of the hospital's capability to perform all eight of the signal functions—specifically, the removal of retained products. This was the only time that this signal function was performed during the whole month and it was not recorded in the delivery room logbook. In this case it appears that the hospital was not able to meet the standard of a comprehensive EmOC facility. Such information is vital in order to ensure that the facility receives recognition for the care that it does provide.
Process indicators
The calculation of process indicators relies on the proper identification of signal functions. The accuracy of the data recorded by the hospital staff was poor and created the impression that there were fewer births and caesarean sections. The omission of the case requiring the removal of retained protects could lead one to assume that there was a failure to meet the minimum acceptable level of an adequate number of EmOC facilities for the population when that was not the case. The case fatality rate was not affected as there were no maternal deaths to report during that period.
Case identification
The delivery room logbook is a source of identfication of maternal deaths in order to perform a review of the care provided. Although there were no maternal deaths reported during this study, there were two neonatal deaths that had occurred which were not reported. As these cases are not recorded in the delivery room logbook, they would not have been identified in a review in order to evaluate and improve the standard of the neonatal care provided.
Discrepencies
Underreporting created the largest discrepency when calculating the number of births, caesarean section rates, incidence of gestational hypertension and postpartum haemorrhage. The underreporting distorted the reality of the services of the hospital. This in turn could have a negative effect on the programme and financial planning for the obstetric ward within the hopital. If the administration only sees a partial picture of what is really going on, insufficient money and staff might well be allocated.
Of the unrecorded deliveries, 82% involved situations where both midwives and obstetricians provided care (such as in the case of a transfer of care). These were complex cases and involved complex pathways of care. This same problem was identified after the first evaluation of the logbook and, in response, pathways of care were changed and roles and responsibilities for each staff member were defined. Although this reflected an improvement from 72% to 80.7% of recorded deliveries, other strategies may also need to be put into place to ensure that these deliveries and their outcomes are included. Reinforcement and retraining of the roles and responsibilities of the staff should be part of regular staff meetings. The training should include an explanation of how data from delivery room logbooks is used and how it provides the information required to forecast future health-care spending.
In this present study, insufficient staffing may have also contributed to the underreporting as 63.6% (seven) of the unrecorded cases occurred on a busy day. Increasing the number of personnel may allow staff the time to document accurately. The difficulty of increasing the number of personnel may be that there are an insufficient number of people in a rural area able to work as health-care providers.
Illegible information
Quite often, the information recorded in the delivery room logbook was untidy. Difficulty in reading the logbook could lead to a misinterpretation of the events. The delivery room logbook at GAH is essentially a locally purchased large, blank note book of lined paper. The staff draw in the column lines and labels. At times, the lines became so crooked that there was no space in which to write the information. A pre-printed standardized book might help improve the tidiness and avoid this problem.
Erroneous information
On several occasions even when a delivery was recorded, some of the information was simply incorrect. There was confusion among staff about the difference between a stillbirth and a neonatal death, as well as understanding the terms used in the obstetric history of a woman (e.g. gravida, para, abortion). Quite often staff would record information from the woman's old hospital card, reporting that the woman came from one village when in fact the woman had moved or the name of the village was incorrect. As MW provides community outreach to certain villages, this information has become vital in order to monitor the impact of these activities.
Missing information
Although the main details of a delivery may have been recorded, there were often some items that were left blank, such as the ‘hospital card number’ and the ‘comments’ column. The hospital card number was frequently omitted because the baby was delivered quickly while her spouse was elsewhere in the hospital arranging for her admittance. Also the card room is only open during office hours and some women have to wait a couple of days for a card. When there is a delay in obtaining a card number, health-care staff may forget to follow-up and complete information entered into the birth registration book. If a card number is not identified in the delivery room logbook it may become extremely difficult to retrieve the woman's chart when she has been discharged and thus an essential source of information for performing a case review may have been lost.
It is unfortunate that the ‘comment’ column was frequently left blank. There were often interesting details of the birth that could have been included, such as use of parenteral oxytocin, antibiotics and anticonvulsants. Quite frequently these interventions are not indicated in the logbook making it difficult to identify which signal functions were being performed.
Limitations of the study
One limitation of this study was the small number of cases and as such it may have distorted the overall effects of the underreporting. In addition, the small number of cases meant that we were unable to perform a statistical analysis.
Despite the short time span and small number of cases, this prospective study does shed some light on the strengths and weaknesses of the birth registration process within GAH.
Conclusions
Since MW has been present in GAH the rate of reporting events in the delivery room logbook has improved. This improvement is due to greater supervision of MW staff on the floor as well as providing training for the obstetric staff about their roles and responsibilities.
Despite the successes at MW, it is recognized that approximately 20% of births are still not recorded. Retraining, creating realizable protocols and increasing staff may help to rectify this. In addition, a printed, standardized delivery room logbook that ensures there are columns for all the signal functions may, over time, facilitate the monitoring and evaluation of the maternal health services offered at GAH.
The delivery room logbook could become a valuable tool for assessments, monitoring and evaluating safe motherhood programmes and identifying cases for peer review. Our experience at GAH demonstrates that, although the completeness and accuracy of the delivery room logbook had improved with training and supervision, there was still a large amount of erroneous, inconsistent, illegible or missing information and caution is required when drawing conclusions from these data sets. The reliability of the delivery room logbook as the sole source of information in creating health policy and managing safe motherhood programme is questionable.
This small study illustrates that, while facility-based data can provide important data to evaluate a safe motherhood programme, its use is hampered by inaccurate data. More prospective studies over a greater period of time, involving a larger number of births are needed to provide more rigorous conclusions.