Differences in nulliparous caesarean section rates across models of care: a decomposition analysis

Brick, Aoife; Layte, Richard; Nolan, Anne; Turner, Michael J.

doi:10.1186/s12913-016-1494-3

BMC Health Services Research

Table 3 Decomposition of the Differential in Elective and Emergency Caesarean Section Rates between Public and Private Mothers

From: Differences in nulliparous caesarean section rates across models of care: a decomposition analysis

	Elective				Emergency^a
	% pt diff	% of \( {\overset{-}{Y}}^{pub}-{\overset{-}{Y}}^{pri} \)	z-stat		% pt diff	% of \( {\overset{-}{Y}}^{pub}-{\overset{-}{Y}}^{pri} \)	z-stat
\( {\overset{-}{Y}}^{pub}-{\overset{-}{Y}}^{pri} \)	−0.061				−0.060
Explained^b	−0.032	53.3			−0.047	79.7
Unexplained^c	−0.028	46.7			−0.012	20.3
Mother Age	−0.012	18.9	−7.8	***	−0.048	80.6	−14.9	***
Social Class
Professional/managerial (REF)
Clerical	0.000	0.3	−1.1		0.000	0.4	−1.3
Skilled/semi-skilled	0.000	−0.7	1.5		0.000	−0.7	0.9
Unskilled	0.001	−1.4	1.7	*	0.002	−4.0	2.3	**
Unemployed	0.000	−0.8	1.4		0.001	−1.0	0.9
Home duties	0.002	−3.4	2.7	***	0.004	−6.5	2.9	***
Other^d	0.000	−0.8	1.6	*	0.000	0.5	−0.6
Marital Status
Married (REF)
Not married^e	−0.001	1.0	−0.6		0.005	−7.7	1.9	*
Country of Birth
Ireland (REF)
UK	0.000	−0.1	0.6		0.000	0.3	−1.2
EU-15 (excl. IRE & UK)	0.000	0.0	0.0		0.000	−0.1	0.5
EU-27 Accession States^f	0.000	0.7	−0.7		−0.008	13.7	−6.3	***
Africa	0.000	−0.8	1.5		0.002	−4.1	4.9	***
Asia	0.000	0.6	−2.0	**	0.001	−2.2	1.9	*
Other^g	0.000	0.2	−1.5		0.000	0.4	−1.6
Obstetric History
Previous miscarriage	0.000	0.4	−1.7	*	−0.001	1.3	−2.0	**
Baby Characteristics
Birthweight	−0.001	1.5	−2.9	***	−0.004	6.4	−8.7	***
Clinical Risk Factors^h
Breech	−0.009	14.8	−36.6	***	0.000	−0.2	1.6
Diabetes (pre-existing)	0.000	−0.1	1.0		0.001	−0.9	5.5	***
Eclampsia or pre-eclampsia	0.000	−0.6	1.9	*	−0.001	1.6	−8.8	***
Gestational diabetes	0.001	−0.8	3.0	***	0.001	−1.8	4.9	***
Hypertensive disorders	0.000	0.0	−0.1		0.000	0.5	−3.7	***
Malpresentation	−0.013	21.5	−19.1	***	−0.001	1.0	−5.0	***
Placenta praevia or placental abruption	−0.002	3.7	−9.0	***	0.000	−0.2	1.2
Poor fetal growth	0.000	0.0	0.1		0.000	−0.2	1.5
Induction					−0.007	12.0	−12.0	***
Hospital Characteristics
Academic teaching hospital	0.000	−0.8	1.1		0.006	−9.6	6.8	***

Results are presented as marginal effects, with robust standard errors in parentheses
Using the public (pub) coefficients as the reference. Using the private (pri) coefficients as the reference, the explained components were 84.4 % for elective and 82.2 % for emergency, while using the pooled coefficients results in an explained component of 58.6 % for elective and 78.3 % for emergency (full results available on request from the authors)
*** p < 0.01, ** p < 0.05, * p < 0.1
^a Robustness checks were carried out on the EMCS model by excluding clinical characteristics and the proportion explained by the model increased (87.3 % explained). In particular the proportion explained by age and EU-27. The EMCS model was also run excluding age and the proportion explained by the model halves (44.7 % explained). This highlights that age explains a large proportion of the difference between public and private EMCS rates and when removed from the model it becomes much less powerful. This is not to say that age is not picking up some omitted variables such as BMI. Results available on request
^b The differential that is estimated based on differences in observed characteristics (i.e., the first term in Equation (2))
^c The differential that is estimated based on differences in omitted variables e.g. BMI, smoking, etc. (i.e., the second term in Equation (2))
^d Includes farmers and farm managers, other agricultural occupations and fisheries workers, and not classifiable
^e Includes never married, divorced, separated, and widowed
^f Accession States
^g Includes the Rest of Europe, the Americas, Australia, New Zealand (incl. Oceania), multi-nationality, non-Irish, and no nationality
^h Clinical risk factors imply that the relevant diagnosis code(s) appeared as one of the up to 20 diagnosis codes (one principal and up to 19 additional) recorded in HIPE, a woman may have more than one clinical risk factor. The presence of a particular risk factor does not imply that this was the ‘cause’ of the CS
Source: HIPE/NPRS 2009 data

Back to article page

ISSN: 1472-6963

Contact us

General enquiries: journalsubmissions@springernature.com