Changing patterns of emergency obstetric care at a Nigerian University hospital

Article Outline

• Abstract
• 1. Introduction
• 2. Methods
  • 2.1. Study setting
  • 2.2. Definition of terms
  • 2.3. Study design and identification of cases
  • 2.4. Data analysis
• 3. Results
• 4. Discussion
  • 4.1. Actions taken as a result of this audit for
    • 4.1.1. Administrative problems
    • 4.1.2. Medical personnel problems
    • 4.1.3. Patient-orientated problems
  • 4.2. Limitations of the study
• 5. Conclusions
• Acknowledgment
• References
• Copyright

Abstract 

Objective

To analyze the changing patterns of critical obstetric care over two consecutive 3-year periods and identify the factors responsible for the trend through combined audits of near miss and maternal mortality at a Nigerian University hospital.

Methods

Retrospective audit and comparison of “near misses” and maternal deaths recorded in 1999–2001 and 2002–2004 at a tertiary care center in southwest Nigeria. The definition of near miss morbidity was based on validated disease-specific criteria. For each near miss and maternal death, the local audit committee compared the actual management with local treatment protocols and explored avoidable factors. Case fatality rate was calculated for “critically ill obstetric patients” (CIOP-CFR) for both periods. The cause-specific case fatality rate (CFR) was used to assess the trend in standards of care for life-threatening obstetric conditions. Data were compared using the χ² or Fisher's exact test. P<0.05 was considered statistically significant.

Results

There were 175 near misses and 27 maternal deaths in 1999–2001 and 211 near misses and 44 maternal deaths in 2002–2004. The CIOP-CFRs for the two periods showed a declining (but non-significant) trend in the standard of emergency obstetric care for life-threatening conditions (13.4% to 17.3%, P=0.250). The CIOP-CFR for postpartum hemorrhage significantly increased from 3.1% to 21.1% in the 2nd period (P=0.033), reflecting a decline in the standard of care. Lack of blood for transfusion became a more significant administrative problem in the 2nd period occurring in 17.8% of all critically ill patients managed in 2002–2004. There was a notable though statistically insignificant increase in the non-adherence to treatment protocol among cases of maternal death in 2002–2004 compared with 1999–2001.

Conclusions

The standard of critical obstetric care in this center is suboptimal with no evident improvement over the 6-year period. This audit supports the feasibility of including near miss reviews in maternal death audits to provide insights into the trend in the quality of emergency services for severe maternal complications while highlighting factors associated with deficiency or improvement in care for specific maternal conditions.

Keywords: Emergency obstetric care, Near miss, Maternal death, Critical obstetric care, Case fatality rate, Quality of care

Back to Article Outline

1. Introduction 

In recent times, review of cases of severe acute maternal morbidity, also known as near miss, has been shown to be a useful complement to the traditional enquiries into maternal deaths for audit of maternal healthcare services [1], [2], [3]. Because near miss morbidities are more common than maternal deaths, clinical audits of maternal healthcare can be conducted more frequently, even in areas with low maternal mortality. Although experience is limited, audits of maternal healthcare using near miss reviews have produced meaningful results in developing countries [2], [4]. Combining near miss reviews with maternal death audits permits early detection of trends in maternal healthcare and may justify redistribution of resources at shorter intervals [2], [5], [6]. This system of audit is particularly advantageous in low resource settings, where meager government investments in obstetric services require short-term appraisals to determine whether such efforts yield the desired changes.

Since the International Conference on Population and Development (Cairo, 1994), there has been an increasing understanding of the pathways to maternal deaths and disabilities and the strategies to achieve the most favorable results. Although the provision of emergency obstetric care (EmOC) is generally accepted as the keystone of any successful approach to reduce maternal deaths and disabilities, its integration into existing health services and monitoring of its use remain a challenge to health systems in developing countries. The United Nations process indicators describe the vital elements and performance of health systems for women with obstetric complications [7]. One of these indicators, the case fatality rate (CFR), provides an insight into the quality of EmOC provided within a facility and can be used for tracking changes in the standard of emergency care over time in the same facility. While factors other than standard of care influence the CFR, it is a useful indicator of care. In facilities with a large enough number of obstetric complications, the cause-specific CFR is encouraged for assessment of the quality of care for specific maternal conditions to explore the need for specific interventions [8]. We used this indicator in a combined audit of near miss morbidity and maternal mortality to analyze the changing patterns of emergency obstetric care over two 3-year periods at a Nigerian university hospital. We also identified factors contributing to the trends in the quality of care for severe maternal complications in the institution and described the interventions initiated to tackle them.

Back to Article Outline

2. Methods 

2.1. Study setting 

The audit was conducted at Olabisi Onabanjo University Teaching Hospital (OOUTH), Sagamu, the main referral center for other hospitals in Ogun State, southwest Nigeria. Besides the primary function of providing emergency obstetric services to women referred from other centers, the hospital also provides antenatal care and delivery services to pregnant women from Sagamu and its environs. The hospital serves an estimated population of 1.2 million; not all patients requiring tertiary care within the population present at the center. Patients must pay for the services provided though they are managed within the limits of the existing resources before all payments are made. The hospital runs a limited blood banking service and relatives of patients are expected to donate or pay for blood when transfusion is indicated, even in emergencies. The hospital has a high maternal death rate and a hospital death rate as high as 1936.1 per 100,000 live births has been reported [9].

Between 2002 and 2004, efforts of the government of Ogun State through the hospital management to upgrade obstetric services in the hospital included providing a new 100 kVA power generating plant for the operating theater (to combat irregular power supply) and hiring of a consultant obstetrician and 5 residents in addition to the existing 8 consultant obstetricians and 10 residents. More nurses/midwives were also employed to replace those who have emigrated. The costs of delivery care to the patients (excluding the operation fee) were also reduced by over 50% to improve affordability of services.

2.2. Definition of terms 

Near miss morbidity was defined as an acute direct obstetric complication that immediately threatens a woman's survival but does not result in death either by chance or because of hospital care received during pregnancy, labor or postpartum. Table 1 shows the locally adapted disease-specific criteria (5 main diagnostic groups) used by Filippi et al. [4] in similar hospital settings in West Africa to identify near miss morbidities.

Table 1. Definition of near miss morbidity

Criteria	Definition
Hemorrhage	Severe hemorrhage leading to shock, emergency hysterectomy, coagulation defects and/or blood transfusion of ≥2 L.
Hypertensive disorders	Eclampsia and severe pre-eclampsia with clinical/laboratory indications for termination of pregnancy to save the woman's life.
Dystocia	Uterine rupture and impending rupture e.g. prolonged obstructed labor with a previous cesarean section.
Infection	Hyperthermia or hypothermia or a clear source of infection and clinical signs of septic shock.
Anemia	Low hemoglobin level (<6 g/dL) or clinical signs of severe anemia in women without severe hemorrhage.

“Critically ill obstetric patients” (CIOP) refer to women who suffered life-threatening obstetric conditions that resulted in either death or near miss morbidity. For the purpose of this study, the CFR was calculated for critically ill obstetric patients (i.e. near miss plus women who died). This CIOP-CFR was determined by the number of critically ill women who died divided by the number of all critically ill obstetric patients. To appreciate the standard of care provided for each disease process, the cause-specific CIOP-CFR was calculated for each obstetric condition. This was defined as the number of maternal deaths resulting from a particular obstetric condition divided by the sum of near misses and maternal deaths occurring from such condition, expressed as a percentage. Maternal death was defined according to the 10th revision of the International Classification of Diseases (ICD-10) [10].

2.3. Study design and identification of cases 

Cases were identified through a retrospective analysis of the records of women admitted into the obstetric units of the hospital with pregnancy-related complications from 1 January 1999–31 December 2001, and 1 January 2002–31 December 2004. From the Medical Records Department of the hospital, the Near Miss Audit Committee (3 consultant obstetricians and 3 residents) retrieved the case files of women whose diagnoses met the above pre-defined criteria (as well as those with the possibility of being associated with severe acute maternal complications) for scrutiny. For each case of near miss morbidity and maternal death, data were collected on demographic characteristics, nature of obstetric complications, time between diagnosis and definitive treatment, the level of most senior person who treated the patient and the time until the senior person arrived after admission of the patient (or after diagnosis for in-patients), and any deviation from management protocol. Reasons for deviation from management protocol were also examined and classified as administrative, patient-orientated, and medical personnel problems. Administrative problems included lack of power supply, transport and communication, essential drugs, blood for transfusion, or competent staff. Patient-orientated problems included those generated by the patient or her family, such as delay in presentation, refusing intervention, or inability to pay for necessary services when due. Medical personnel problems included delay in initial assessment by at least a senior resident, incorrect diagnosis, and inappropriate management or monitoring of the critically ill-patient. Data on the total number of direct obstetric complications, number of deliveries, and maternal deaths for the two periods were also obtained from admission, labor, and delivery records. The Research and Ethics Committee of the hospital approved the study.

2.4. Data analysis 

We used Epi Info 2002 statistical software [11] for data entry and analyses. The results are presented in frequencies, percentages and descriptive statistics. The prevalence of near miss morbidity was defined as the number of near misses divided by the number of deliveries in the hospital. The CIOP-CFRs and cause-specific CIOP-CFRs for the two periods were compared to note the change in the standard of care for life-threatening obstetric conditions. The hospital maternal death rate was expressed as the number of maternal deaths per 100,000 deliveries. The overall CFR for the institution was expressed as the proportion of women who died among all women that experienced all degrees of direct obstetric complications.

Categorical variables were compared with the χ² or Fisher's exact test while continuous variables were compared with the t test. Differences were considered statistically significant when P<0.05.

Back to Article Outline

3. Results 

There were 27 maternal deaths, 175 near misses and 1076 deliveries in 1999–2001, and 44 maternal deaths, 211 near misses and 1501 deliveries in 2002–2004. In Table 2, the frequencies of near miss, maternal death and CIOP were compared for both periods. This shows that maternal deaths non-significantly increased from 27 in 1999–2001 to 44 in 2002–2004 representing a change in the hospital maternal death rate of 2509 to 2931 per 100,000 deliveries. The prevalence of CIOP decreased from 18.8 in the 1st period to 17.0% in the second period but the CIOP-CFR non-significantly increased by 3.9%. The overall CFR showed a marginal decrease of 0.6% between the two periods. The number of women with direct obstetric complications was almost 3 times as frequent in the 2nd compared with the 1st period.

Table 2. Comparison of the prevalence of near miss, ma-ternal deaths and case fatality rates for the two periods

Abbreviations: CIOP, critically ill obstetric patients (near misses+maternal deaths); CFR, case fatality rate.

Table 3 shows the trend in the standard of care provided for life-threatening obstetric conditions according to diagnosis. For 1999–2001, the cause-specific CIOP-CFR was highest for infectious complications (17.2%), followed by anemia (16.7%), hypertensive disorders (15.7%), dystocia (11.1%), and hemorrhage (9.1%). The ranking of the cause-specific CIOP-CFRs was similar for 2002–2004 viz: infection (28.6%), anemia (16.1%), hypertensive disorders (14.6%), dystocia (11.3%), and hemorrhage (11.0%). For the first period, evaluation of the cause-specific CIOP-CFR indicates that abruptio placentae (40.0%), abortion (33.3%), and eclampsia (21.6%) had the poorest level of care while postpartum hemorrhage (3.1%), placenta previa (0.0%), and severe pre-eclampsia (0.0%) had better care. For the second period, placenta previa (0.0%), abruptio placentae (0.0%), and impending rupture (0.0%) had the best form of care while infection (28.6%), uterine rupture (37.5%), postpartum hemorrhage (21.1%), and eclampsia (20.4%) had substandard care. When the two periods are compared, only ectopic pregnancy, abortion, abruptio placentae, and impending rupture showed improvement in their CIOP-CFRs, though even these differences were not statistically significant. The CIOP-CFR for postpartum hemorrhage significantly increased from 3.1% in 1999–2001 to 21.1% in 2002–2004 (P=0.033), indicating that more critically ill women died from this complication in the 2nd period. Compared with the 1st period, there was also an insignificant increase in women dying from uterine rupture and infections.

Table 3. Comparison of cause-specific case fatality rates according to diagnosis distribution

Abbreviations: NM, near miss; MD, maternal death; CIOP-CFR, critically ill obstetric patient case fatality rate.

Table 4 shows the identified areas of substandard care for the near misses and maternal deaths for 1999–2001 and 2002–2004. The definitive treatment was started after 1 h more frequently in 2002–2004 for maternal deaths, near misses and both combined. The level of most senior personnel below the level of senior resident showed a marginal reduction in its contribution to substandard care in the 2nd period. There was a notable though statistically insignificant increase in the non-adherence to management protocol for cases of maternal death in the later period compared with the earlier one. Overall, the time interval between diagnosis and definitive treatment was longer than 1 h in almost half of all deaths recorded during the six years, the most senior personnel was below the level of senior resident in 12.7% of cases and deviation from management protocol was noted in 57.7% of cases. When the two 3-year periods were compared, none of the differences in any area of substandard care for either maternal death, near miss or both combined were significant (P>0.05).

Table 4. Identified areas of substandard care for critically ill women^a

Factor	1999–2001			2002–2004			1999–2004
	MD	NM	CIOP	MD	NM	CIOP	MD	NM	CIOP
	n=27	n=175	n=202	n=44	n=211	n=255	n=71	n=386	n=457
Diagnosis-intervention interval >1 h	12(44.4)	45(25.7)	57(28.2)	23(52.3)	61(28.9)	84(32.9)	35(49.3)	106(27.5)	141(30.9)
Most senior attendant below Senior Registrar	3(11.1)	16(9.1)	19(9.4)	6(13.6)	12(5.7)	18(7.1)	9(12.7)	28(7.3)	37(8.1)
Diagnosis-treatment interval by most senior person >1 h	4(14.8)	23(13.1)	27(13.4)	3(6.8)	29(13.7)	32(12.5)	7(9.9)	52(13.5)	59(12.9)
Deviation from protocol	12(44.4)	72(41.1)	84(41.6)	29(65.9)	89(42.1)	118(46.3)	41(57.7)	161(41.7)	202(44.2)

^aValues given as number (percentage).

Abbreviations: MD, maternal death; NM, near miss; CIOP, critically ill obstetric patients.

Avoidable factors responsible for the deviation from management protocol are highlighted in Table 5. Medical personnel problems were the most frequent avoidable factor in the 1st period while patient-orientated problem was the most frequent in the 2nd period. Administrative and medical personnel problems were less frequent reasons for substandard care in the 2nd compared with the 1st period although only administrative problems were significantly so (P=0.010). However, lack of blood for transfusion became a more significant administrative problem in the 2nd period occurring in 17.8% of all critically ill patients managed in 2002–2004. The delay in assessment by senior personnel was the main avoidable medical personnel problem identified during both periods.

Table 5. Identified avoidable factors responsible for deviation from protocol in the management of critically ill obstetric patients

	1999–2001	2002–2004
	n=84	n=118
Administrative problemsa	34 (40.5)	28 (23.7)
Lack of power supply	24	7
Lack of blood for transfusion	3	21
Lack of personnel	1	0
Delays in the theatre	2	0
Transport/Communication	4	1
Patient orientated problemsa	28 (33.3)	54 (45.7)
Personal	9	23
Family	19	31
Medical Personnela	39 (46.4)	47 (39.8)
Delay in initial assessment	32	38
Problem recognition	4	5
Management plan	3	1
Monitoring	0	3

Note that some patients had more than one avoidable factor.

Back to Article Outline

4. Discussion 

This audit provides insight into the changing patterns of EmOC over a 6-year period in the same institution to indirectly assess whether public investments in the system have brought improvement. Our review shows that although the number of deliveries was relatively low, many of the women managed in this center presented with life-threatening obstetric complications, indicating the extent to which obstetricians in this center are confronted with life-saving situations despite the prevailing limitations in their provision of EmOC. The changes in the overall CIOP-CFRs (13.4% to 17.3%) indicate that the care for severely ill pregnant women in this center not only did not improve in the 2nd period but in fact deteriorated in spite of efforts to improve it. This highlights the significance of evidence-based allocation of resources, as the efforts of the hospital management to improve care did not achieve the desired changes because they were not well targeted. The review shows that while a few administrative problems improved during the 2nd period, some deteriorated considerably with significant effect on the level of maternal mortality. An example of this is postpartum hemorrhage and uterine rupture that were not well managed in the 2nd period as a result of deficiency in blood transfusion services.

The overall CFRs (for all degrees of severity) for both periods were much greater than the recommended <1%, suggesting that a lot needed to be done to improve the standard of care in this center. Of interest is the fact that the CIOP-CFR increased while the overall CFR showed a slight improvement in the 2nd period. This indicates that the standard of care for life-threatening conditions in the 2nd period was poorer even though there was a slight improvement for the care provided for all women with obstetric complications. This disparity may partly be explained by increased burden (rather than prevalence) of critically ill women and increased frequency of direct obstetric complications that were not life threatening that deflated the overall CFR in the 2nd period. This observation supports the need to exercise caution in interpreting institutional overall CFR when reported in isolation as advised by Paxton et al. [8]. While the overall CFR may suffice for primary and secondary care level hospitals, there may be a need to standardize definitions and discriminate between the severity of obstetric complications that are used in assessing the standard of care in referral centers to obtain a clearer picture of their true emergency capabilities. It can be argued that while some deficiencies in care for women with mild to moderately severe obstetric complications may not lead to catastrophic outcome, the outcome for critically ill women is very sensitive to any deficiency in the quality of care.

The level of the CIOP-CFRs was not surprising in view of the proportion of critically ill women in whom management protocols were not followed. It should be realized, however, that while adherence to management protocol is largely dependent on the clinicians in settings where services are covered by health insurance, adherence to management protocol is also influenced by other factors beyond the control of medical personnel in settings where patients pay before receiving care. The management protocol was not followed in more than half of the reviewed cases and was worse in the 2nd period as a result of decreasing patients' ability to pay.

As shown by Vandecruys et al. [5], one of the advantages of using combined audits of near miss and maternal deaths is that the analysis can be disease specific. Postpartum hemorrhage, which was one of the well-managed disease processes in the 1st period, became one of the poorly managed conditions in the 2nd period. This was related to the increasing scarcity of donated blood in the hospital, as many potential donors are reluctant to come forward out of fear of pre-donation HIV screening. This is probably the effect of HIV awareness that has increased over time and the unfavorable consequences of testing positive within the community. Part of the solution to this is increased compliance with active management of third stage of labor for prevention of postpartum hemorrhage to generally reduce the need for blood transfusion in women who deliver primarily at the hospital.

The decline in the standard of care for severe complications due to infection and uterine rupture is a cause for concern. The ranking of the standard of care according to diagnosis for both periods shows an inverse relationship between the frequency of severe maternal condition and the standard of care. It appears that less frequent conditions such as infection and anemia did not receive enough attention compared with more frequent ones, and this may account for the trend shown by infectious complications in the 2nd period.

Patients’ inability to pay for ever more expensive antibiotics when needed is an important factor as the economic situation becomes increasingly difficult. The decline in the standard of care for uterine rupture is also a reflection of the limitations of the blood transfusion services in the hospital. Patient-orientated factors that contributed to poorer blood transfusion services included unwillingness of relatives to donate, inability to pay for blood and rarely, refusal of blood transfusion by patient. Contributory administrative problems included non-availability of blood in stock (even when patients are willing to pay) and delay in screening for HIV and hepatitis when blood is sourced from private laboratories. Failure of health personnel to recognize the problem also contributed to the poor standard of care for uterine rupture.

4.1. Actions taken as a result of this audit for 

4.2. Limitations of the study 

The main limitation of this study is its retrospective nature and its reliance on records by health workers working under emergency conditions. Therefore, the possibility of introducing detection bias by personnel involved in data collection and under-reporting of circumstances surrounding the management of maternal complications should be considered. We attempted to overcome this potential source of error by ensuring that cases were selected and avoidable factors were identified by at least two senior members of the audit committee. It is unlikely that this would affect the results of our comparisons as the same personnel conducted the review for both periods.

Back to Article Outline

5. Conclusions 

The standard of emergency care in this tertiary hospital is suboptimal with no evident improvement over a 6-year period despite the governmental and hospital management efforts to improve it. This paper highlights what stands to be gained when investment in obstetric services is based on findings from a clinical audit. Strategies to improve maternity services in this hospital should cut across all severe maternal complications, as there were no significant positive changes in the standard of care for any category of life-threatening condition. In view of the large number of women with severe complications, efforts should be directed at managing critically ill women so as to improve on the present maternal health profile in this environment. Present efforts are, however, expected to make a difference in the quality of EmOC at this institution. Our audit supports the utility of including near miss morbidities in maternal death reviews to understand trends in the quality of EmOC and the factors responsible for improvement (or depreciation as in this case) in care. Other teaching hospitals are encouraged to conduct similar audits to allow critical comparison between institutions and learning from one another.

Back to Article Outline

Acknowledgment 

The authors acknowledge the immense contribution of the members of the Near-Miss Audit Committee, Dr. J.O. Sotunsa, Registrar, Department of Obstetrics and Gynecology and staff of the Medical Records Department of Olabisi Onabanjo University Teaching Hospital, Sagamu, Nigeria.

Back to Article Outline

References 

1. Ronsmans C, Filippi V. Reviewing severe maternal morbidity: learning from survivors from life-threatening complications. In: World Health Organization. Beyond the Numbers: Reviewing Deaths and Complications to Make Pregnancy Safer. vol. 7:Geneva: WHO; 2004;p. 103–124
   • View In Article
2. Cochet L, Pattinson RC, MacDonald AP. Severe acute maternal morbidity and maternal death audit — a rapid diagnostic tool for evaluating maternal care. S Afr Med J. 2003;93:700–702
   • View In Article
   • MEDLINE
3. Pattinson RC, Buchmann E, Mantel G, Schoon M, Rees H. Can enquiries into severe acute maternal morbidity act as a surrogate for maternal death enquiries?. BJOG. 2003;110:889–893
   • View In Article
   • MEDLINE
   • CrossRef
4. Filippi V, Ronsmans C, Gohou V, Goufodji S, Lardi M, Sahel A, et al. Maternity wards or emergency obstetric rooms? Incidence of near-miss events in African hospitals. Acta Obstet Gynecol Scand. 2005;84:11–16
   • View In Article
   • MEDLINE
   • CrossRef
5. Vandecruys HIB, Pattinson RC, Macdonald AP, Mantel GD. Severe acute maternal morbidity and mortality in the Pretoria Academic Complex: changing patterns over 4 years. Eur J Obstet Gynecol Reprod Biol. 2002;102:6–10
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (247 KB)
   • CrossRef
6. Say L, Pattinson RC, Gülmezoglu AM. WHO systematic review of maternal morbidity and mortality: the prevalence of severe acute maternal morbidity (near miss). Reprod Health. 2004;1:3
   • View In Article
7. UNICEF , WHO , UNFPA . Guidelines for monitoring the availability and use of obstetric services. New York: UNICEF; 1997;
   • View In Article
8. Paxton A, Bailey P, Lobis S. The United Nations process indicators for emergency obstetric care: reflections based on a decade of experience. Int J Gynecol Obstet. 2006;95:192–208
   • View In Article
9. Sule-Odu AO. Maternal deaths in Sagamu, Nigeria. Int J Gynecol Obstet. 2000;69:47–49
   • View In Article
10. World Health Organization: ICD-10. In: International statistical classification of diseases and health-related problems. Tenth Revision. vol. 2:Geneva: World Health Organization; 1993;
    • View In Article
11. Centers for Disease Control and Prevention (CDC) and World Health Organization. Epi Info 2002 — Database and statistics software for public health professionals. Atlanta, Georgia, USA: Centers for Disease Control and Prevention; 2002;
    • View In Article