|Labor epidurals improve outcomes for babies of mothers at high risk for unscheduled Cesarean section
|Stuart K A, Krakauer H, Schone E, Lin M, Cheng E, Meyer G S
This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn.
The use of epidural placement for labour in women at high risk for Caesarean section.
Economic study type
The study population comprised labouring women at high risk for unscheduled Caesarean section.
The setting was secondary and tertiary care. The economic study was carried out in the USA.
Dates to which data relate
The effectiveness and resource use data were gathered mainly in 1995. The price year was 1995.
Source of effectiveness data
The effectiveness evidence was derived from a single study and experts' opinions.
Link between effectiveness and cost data
The costing was carried out prospectively on the same sample of patients as that used in the effectiveness study.
Power calculations to determine the sample size were not reported. A single group of patients was identified among eligible labouring women who were military dependent and received either peripartum and neonatal care in military treatment facilities or civilian treatment facilities with military insurance. The sample was derived from the US Department of Defense's National Quality Management Program (NQMP) in 1996. Data on 7,704 deliveries were gathered. The regression coefficients of twelve pre-admission and nine post-admission characteristics of the sample considered in the study were summed to give the threshold probability of better outcomes with unscheduled Caesarean section. Such a threshold value was 0.75. Of the 7,704 deliveries identified, 594 had a greater than 75% chance of having an unscheduled Caesarean section. However, 62 did have scheduled Caesarean section, 10 had contraindications to continue labour, and 4 had no clear indication for epidural placement. These were all excluded. Therefore, the final sample comprised 518 mother-baby pairs.
This was a historical cross-sectional study. The centres where the intervention took place were not reported. The women were not followed after the epidural placement was performed.
Analysis of effectiveness
All of the patients included in the initial study sample were considered in the effectiveness study. The outcomes used in the analysis were the probability values of unscheduled Caesarean section, vaginal delivery, and spinal or general or epidural/spinal anaesthesia. The odds ratios for the likelihood of Caesarean section considering the exposure of epidural placement were also calculated in the whole sample (7,704 mothers) and in the study sample of high-risk mothers. Regression and correlation analyses were carried out to identify potential confounding variables among 18 factors. The characteristics of the sample considered in the study were compared with both the general population of delivering mothers from the US National Natality Survey (1995), in order to assess the generalisability of the results, and the 1997 NQMP data set to validate the results.
The probability values were reported in the graphic of the decision tree. The probabilities were:
0.94 for unscheduled Caesarean section and 0.06 for vaginal delivery;
0.083 for epidural only and 0.17 for other anaesthetic procedures (0.63 for spinal anaesthesia and 0.37 for general anaesthesia) after epidural placement and unscheduled Caesarean section; and
0.93 for epidural/spinal and 0.07 for general anaesthesia after no epidural placement and unscheduled Caesarean section.
The odds ratios for the likelihood of Caesarean section considering the exposure of epidural placement were 1.45 (95% confidence interval, CI: 1.26 - 1.67) in the whole sample of 7,704 women and 0.61 (95% CI: 0.25 - 1.44) in the study sample.
The regression and correlation analyses indicated that only one variable, chorioamniotis, was correlated with both exposure and outcome for mothers and their babies. There was a positive association between epidural placement and chorioamniotis, and an unfavourable correlation between chorioamniotis and outcome.
The comparison with the national database showed that the military sample was less likely to be in an extreme age group and more likely to be married than the general US population. In addition, a lower Caesarean rate and a higher use of prenatal care were observed for military dependants. The 1997 data showed similar trends, as a significant advantage was observed for babies.
The probability values, which were obtained from the single study, were used as model inputs in the decision tree. The analysis suggested that in the general population of the 1996 NQMP data set, epidural placement was associated with an increased risk of unscheduled Caesarean section while there was no significant association in the sub-group of high-risk mothers.
A decision tree model was constructed to assess the benefits and costs of epidural placement in comparison with no epidural placement. The mothers could deliver vaginally or by unscheduled Caesarean section, regardless of whether an epidural was placed or not. General anaesthesia was required for some women. Three trees of similar structure and probability values were used for mothers, babies, and mother-baby pairs. The authors noted that the probability values could differ slightly due to the occurrence of multiple gestations.
Methods used to derive estimates of effectiveness
A modified Delphi process was used to attribute utility values to maternal and neonatal conditions occurring during the peripartum and neonatal hospitalisations. A panel of neonatal and obstetrical specialists scored all conditions on a scale of 0 (the condition had no effect on the outcome) to 12 (the condition led to the worst outcome). Averages were then computed to provide utility values for entry into the decision model. The authors also made some assumptions.
Estimates of effectiveness and key assumptions
The estimates of utility for specific conditions were generally not reported. Some exceptions were symptomatic pneumothorax in a baby (score of 4), Grade four ventricular haemorrhages (score of 8), fourth degree laceration for mothers (score of 3), eclampsia (score of 6) and pulmonary embolism (score of 8). The authors assumed that the technique for administering the epidural was relatively constant and that the monitoring of mothers and babies was uniform.
Measure of benefits used in the economic analysis
No summary measure of benefit was derived. The model output was the utility value associated with epidural or no epidural placement in mothers, babies, and mother-baby pairs.
Discounting was not performed since the costs per patient were incurred during a short time period. The unit costs and the quantities of resources used were not presented separately. The health services included in the economic evaluation were procedures and hospital stay. The cost categories were not broken down. The cost/resource boundary of the health care system appears to have been used. Resource use was estimated using actual data coming from the sample of patients who were involved in the effectiveness study. The cost data were estimated from Medicare rates. The average costs per patient were calculated using the decision model. The price year was 1995.
Statistical analysis of costs
Statistical tests were carried out to test the statistical significance of differences in the estimated costs, but no details of the type of tests used were reported. The costs were presented as mean values with 95% CIs.
The indirect costs were not considered.
US dollars ($). The costs were presented as relative cost units (RCUs). Each RCU corresponded to approximately $33.50.
Sensitivity analyses were not performed.
Estimated benefits used in the economic analysis
No summary of benefit measure was derived. The estimated utility values obtained from the model were:
2.37 with epidural and 2.35 without epidural for mother (difference 0.02, 95% CI: -2.15 - 0.255; p=0.87);
2.42 with epidural and 3.01 without epidural for babies (difference -0.59, 95% CI: -1.175 - -0.004; p=0.048); and
4.79 with epidural and 5.37 without epidural for mother-baby pairs (difference -0.58, 95% CI: -1.227 - 0.067; p=0.079).
Therefore, only the difference in utility values for babies reached statistical significance.
The estimated costs were:
5.81 RCUs with epidural and 5.77 RCUs without epidural for mother (cost-difference 0.04, 95% CI: -0.029 - 0.089; p=0.32);
4.49 RCUs with epidural and 4.57 RCUs without epidural for babies (cost-difference 0.08, 95% CI: -0.212 - 0.052; p=0.22); and
6.09 RCUs with epidural and 6.08 RCUs without epidural for mother-baby pairs (cost-difference 0.01, 95% CI: -0.064 - 0.089; p=0.79).
Synthesis of costs and benefits
The costs and benefits were not combined. Therefore, in effect, a cost-consequences analysis was carried out.
Epidural placement was cost neutral and led to modest benefits only in relation to newborns. No impact was observed in mothers or mother-baby pairs.
CRD COMMENTARY - Selection of comparators
The rationale for the choice of the comparator was clear. No epidural placement was appropriately selected as the basic comparator because the aim of the study was to assess the active value of epidural placement. It was unclear whether it represents standard care. You should decide whether no epidural placement represents a valid comparator in your own setting.
Validity of estimate of measure of effectiveness
The analysis of effectiveness was mainly derived from a cross-sectional study, which represents a weak source of evidence. No comparison group was considered and the relationship between exposure and outcome could have been attributed to factors other than the study intervention. The authors carried out regression analyses in an attempt to reduce the impact of potential confounding factors. However, other types of bias, such as reporting bias, may have affected the study results. The utility data were derived from experts' opinion through a modified Delphi process. The use of sensitivity analyses would have been helpful to explore the issue of uncertainty in all estimates and to find critical variables.
Validity of estimate of measure of benefit
No summary benefit measure of benefit was used. Although the authors derived utility values associated with some related conditions, the costs and benefits were not combined.
Validity of estimate of costs
Although the authors stated that a societal perspective was adopted, it would appear more appropriate to state that the boundary of the health care system was adopted since Medicare rates were used to estimate the costs and the indirect costs were not included. Limited details on the cost analysis were reported. The price year was reported, but the unit costs were not presented separately from the quantities of resources used. Further, a breakdown of the cost items was not provided. It would be difficult to replicate the study in other settings. The costs were presented as RCUs after a log transformation was performed. Statistical tests of the costs were carried out but no sensitivity analyses were conducted. The cost estimates were specific to the study setting.
The authors did not compare their findings with those from other studies. However, they did state that their study was the first to address the use of epidurals for labour in women at high risk for Caesarean section. The authors noted that the data used in the study might have been based on a biased population. Therefore, some caution is required when generalising the results of the study to populations different from the military population considered in this study.
Implications of the study
The authors suggested that future studies should address a further issue, such as whether the infants could have been delivered faster if an epidural was already in place when the decision for Caesarean section was made.
Source of funding
Funded in part by a Generalist Physician Faculty Scholar Award from the Robert Wood Johnson Foundation.
Stuart K A, Krakauer H, Schone E, Lin M, Cheng E, Meyer G S. Labor epidurals improve outcomes for babies of mothers at high risk for unscheduled Cesarean section. Journal of Perinatology 2001; 21(3): 178-185
Subject indexing assigned by NLM
Adolescent; Adult; Anesthesia, Epidural /economics /statistics & Anesthesia, Obstetrical /economics /statistics & Cesarean Section /utilization; Costs and Cost Analysis; Decision Trees; Female; Hospitals, Military; Humans; Labor, Obstetric; Pregnancy; Pregnancy Outcome; Risk Factors; United States; numerical data; numerical data
Date bibliographic record published
Date abstract record published