Journal of Obstetrics and Gynaecology (1998) Vol. 18, No. 6, 520± 523

OBSTETRICS

Does the attending obstetrician in¯ uence the modeof delivery in the `standard’ nullipara?

NERMEEN VARAWALLA and R. SETTATREESolihull Maternity Unit, Birmingham Heartlands and Solihull Hospitals NHS Trust(Teaching), UK

SummaryWe have investiga ted whethe r the attendin g obstetri cianhas an in¯ uence on the mode of operative deliver y byexamining practice s of obstetri cians working at a distric tgenera l hospita l in the United Kingdom over 3 years.Nulliparo us women (n 5 1410) with a singleton , cephali cand term fetus who had an operative deliver y were studied .The most senio r obstetri cian presen t at 89% of these deliv-eries was one of the 12 registra rs training in the unit. Theproport ion of women delivered by each mode of vagina l orabdominal operativ e deliver y by each registra r was com-pared . In nine of these registra rs the propor tion of womenthey delivered by at least one mode differed signi ® cantlyfrom the others. Caesarea n section after an unsuccessfulattempt at operative vagina l deliver y showed the largestvariatio n and caesarean section before the second stage oflabou r showed the least, with vagina l operativ e deliver ymodes and caesarean section in the second stage without atrial of vagina l deliver y intermediate between the two. Themethod can be used to provide a `pro ® le ’ of the preferencesexercised by individual operato rs.

IntroductionThere is a widely held suspicion that the attendingobstetrician has an in¯ uence on the mode of deliveryalthough there is little supporting evidence. We haveattempted to address this issue by looking at theobstetric practices of obstetricians working at an aver-age sized maternity unit in the UK over a 3-yearperiod. The concept of the `standard nullipara’ (Pater-son et al., 1991) , was adapted to identify a reasonablyhomogeneous group whose management could not bein¯ uenced by previous obstetric experience. Allparous women and nullipara with multiple pregnan-cies, non-cephalic presentations, gestation at deliveryless than 37 weeks and those selected for delivery byelective caesarean section were therefore excluded.There were no exclusions on the basis of ethnicity,pre-existing diseases, pregnancy complications, fetalsize, induction of labour, analgesia in labour, use ofsyntocinon, nor any other feature that might havein¯ uenced mode of delivery. These features could beexpected to be randomly distributed throughout theperiod of study, or to form part of the practice of theperson responsible for the delivery.

Subjects and methodsThe mode of delivery recorded for nulliparous

women with a singleton, cephalic term fetus whowere delivered at Solihull Maternity Unit (SMU)between 1 January 1992 and 31 December 1994 (a3-year period) was analysed. Women who had a`planned’ (elective) caesarean section were excluded.A `planned’ caesarean section was one in which theplan for delivery by caesarean section was madebefore the onset of labour with the intention of avoid-ing vaginal delivery. Some of these would have beenperformed in early labour because labour beganbetween the plan being made and carrying it out.Relevant records were obtained from a computerisedregister.

As is the practice in most maternity units in theUK, midwives managed normal labour and delivery.Obstetric staff became involved when any abnormal-ity was noted. The minimum obstetric cover availableat all times was a senior house of ® cer (SHO), aregistrar (or a SHO acting up as a registrar) and aconsultant. The SHO and registrar were available onsite throughout the 24 hours and the consultant wason site during of ® ce hours and later was available forattendance within 20 minutes of being called. Alldoctors present at every vaginal delivery or emer-gency caesarean section were recorded in the labourward register just after the birth. The most seniorobstetrician present at every vaginal operative deliv-ery or emergency caesarean section was noted andassumed to have been responsible for the conduct ofthe delivery. All the cases managed by locum regis-trars during the study period were analysed collec-tively as a single group. The posterior ventouse cupwas not available for use during the study period.

Apart from the locum registrars there were 12registrars who worked at SMU for varying lengths oftime during the study period. For each of them werecorded the country where they had received theirbasic medical training, whether they had achievedtheir Membership of the Royal College of Obstetri-cians and Gynaecologists (MRCOG) before or duringtheir time in post and their obstetric and gynaecologi-cal experience both in the UK and overseas beforecommencing work at SMU. The ® rst author (N.V.)was not working at SMU during the study period andapart from herself no one else was aware of theidentity attached to the codes used for the obstetri-cians reported in this study.

Correspondence to: R. Settatree , Princess of Wales Women’ s Unit , Birmingham Heartland s Hospital , Bordesle y Green East,Birmingham B9 5SS, UK.

0144-3615/98/060520-04 $9.50 ã Institute of Obstetrics and Gynaecology Trust, 1998

J O

bste

t Gyn

aeco

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

CD

L-U

C S

anta

Cru

z on

10/

30/1

4Fo

r pe

rson

al u

se o

nly.

In¯ uence of obstetrician on mode of delivery 521

For each registrar the propor tion of women deliv-ered by each of the modes of operative delivery wascompared with the 12 other registrars in the study.The probability of difference from chance, wherequoted, was based on a 2 3 2 table comparison usingchi-squared with Yate’ s correction, or Fisher’ s ExactTest if the observed value was less than 5.

ResultsDuring the study period, 1992±94, the total number ofbirths at SMU was 10 073, of which 3983 (39´5%)were to nulliparous women. The vaginal operativedelivery rate and caesarean rate for all women deliv-ered in the unit were 13´9% and 14´6%, respectivelyand those for nulliparous women were 25´1% and17´9%, respectively. After excluding caesarean sec-tions `planned’ before the onset of labour, the numberof births to `standard’ nulliparous women were 3459,constituting 86´8% of nulliparous deliveries and34´3% of all deliveries. Of the 3459 deliveries tostandard nullipara, 2049 (59´2%) were spontaneousdeliveries, 623 (18%) were straight forceps deliveries(Neville Barnes, Simpsons or Wrigleys), 109 (3´2%)were Kjellands forceps deliveries, 207 (6%) wereventouse extractions and 471 (13 6́%) were by emer-gency caesarean section. Deliveries by forceps, ven-touse or emergency caesarean section are hereafterreferred to as `operative’ deliveries.

The most senior obstetrician present at 90 (6%) ofoperative deliveries was a consultant. At 72 (5%) theonly obstetrician present was the SHO, 54 of whichwere outlet forceps and 18 were ventouse deliveries.For the remaining 1248 (89%) operative deliveriesthe registrar, or acting registrar, was the most seniorobstetrician present. Table I shows the numbers andpropor tions of women delivered by the various modesof operative delivery by each registrar. Locum regis-

trars were considered as a single group. Table II ranksthe registrars by their possession of the MRCOG andtheir total years of relevant experience both in theUnited Kingdom and abroad, and re-analyses theresults shown in Table I as the ratio of observed rateto expected rate for each mode of delivery comparedwith the group as a whole. Those ratios with aprobability of chance less than 0´05 are indicated.

The 12 registrars who were being trained at SMUduring the study period had diverse backgrounds.Their basic medical quali ® cation had been obtainedfrom one of eight different countries and three hadquali ® ed in the UK. Three registrars had achievedtheir MRCOG, one of whom had done so during thestudy period. Their individual obstetric and gynaeco-logical experience ranged from 1´5 to 7´7 years, ofwhich their experience in the UK ranged from 0´7 to3´5 years and overseas experience from 0 to 7 years(Table II).

The ratios recorded in Table II provide a `pro ® le’of the choices or preferences in mode of delivery foreach registrar. Apart from Registrars C, G, J and thelocum registrar group, all the other registrars had atleast one mode of delivery ratio which differedsigni ® cantly from the group as a whole.

With regard to caesarean section before full dilata-tion, there were two registrars more likely to deliverat this time (H and D), and one less likely (B). Innone of these three was the probability less than 0´01and none held the MRCOG, though all had more than2 years’ experience.

With regard to assisted vaginal delivery, there werethree registrars (E, B and K) who appeared to have abias in their choice of instrument, but this did notaffect their frequency of performing caesarean sectionin the second stage. Registrar E, the least experiencedof the group, clearly preferred the ventouse overstraight forceps, but still had some experience of

Table I. Mode of operative delivery by obstetric registrars (excluding planned caesarean section) of all nulliparousterm singleton cephalic presentations in Solihull Hospital over a period of 3 years. Figures in brackets arepercentages of all operative deliveries by that registrar. Results are ranked by number of deliveries, except forlocums

CS in 2ndCS in 2nd stage

stageÐ after failedno trial of trial of Total

CS before operative operative number ofIndividual 2nd stage Straight Keilland’s vaginal vaginal operativeregistrar of labour forceps forceps Ventouse delivery delivery deliveriescode n (%) n (% ) n (%) n (% ) n (% ) n (%) n (%)

A 57 (28) 130 (63) 8 (4) 9 (4) 2 (1) 0 (0) 206 (100)B 24 (20) 57 (48) 17 (14) 15 (13) 5 (4) 2 (2) 120 (100)C 34 (30) 50 (43) 13 (11) 12 (10) 1 (1) 5 (4) 115 (100)D 41 (37) 53 (48) 3 (3) 7 (6) 2 (2) 5 (5) 111 (100)E 29 (27) 14 (13) 6 (6) 52 (48) 0 (0) 7 (6) 108 (100)F 30 (30) 31 (31) 9 (9) 17 (17) 8 (8) 4 (4) 99 (100)G 20 (25) 43 (53) 5 (6) 10 (12) 3 (4) 0 (0) 81 (100)H 29 (39) 27 (36) 2 (3) 7 (9) 1 (1) 8 (11) 74 (100)I 18 (24) 28 (38) 9 (12) 7 (9) 2 (3) 10 (14) 74 (100)J 12 (26) 24 (51) 1 (2) 9 (19) 1 (2) 0 (0) 47 (100)K 15 (33) 5 (11) 14 (31) 9 (20) 1 (2) 1 (2) 45 (100)L 8 (27) 9 (30) 0 (0) 6 (20) 1 (3) 6 (20) 30 (100)Locums 34 (25) 60 (43) 7 (5) 26 (19) 3 (2) 8 (6) 138 (100)Allregistrars 351 (28) 531 (43) 94 (8) 186 (15) 30 (2) 56 (4) 1248 (100)

J O

bste

t Gyn

aeco

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

CD

L-U

C S

anta

Cru

z on

10/

30/1

4Fo

r pe

rson

al u

se o

nly.

522 N. Varawalla and R. Settatree

Table II. Experience of registrars and ratio of observed to expected rate for each registrar and each mode of deliverycompared with the group as a whole (operator pro® le). Results are ranked by achievement of MRCOG and totalyears of experience

Characteristics of registrars Mode of deliveryÐ ratio of `observed’ to `expected’CS in 2nd

stageCS in 2nd afterstageÐ no failed

Years relevant trial of trial ofexperience CS before operative operative

Non 2nd stage Straight Keilland’s vaginal vaginalRegistrar MRCOG UK UK Total of labour forceps forceps Ventouse delivery delivery

E No 1´5 0´0 1´5 1´0 0´3** 0´7 3´2** 0´0 1´4I No 1´6 0´0 1´6 0´9 0´9 1´6 0´6 1´1 3´0**F No 1´0 0´8 1´8 1´1 0´7² 1´2 1´2 3´4** 0´9B No 2´5 0´0 2´5 0´7² 1´1 1´9* 0´8 1´7 0´4L No 3´0 0´5 3´5 0´9 0´7 0´0 1´3 1´4 4´5**H No 1´0 3´5 4´5 1´4² 0´9 0´4 0´6 0´6 2´4²C No 2´0 2´7 4´7 1´1 1´0 1´5 0´7 0´4 1´0J No 1´6 4´0 5´6 0´9 1´2 0´3 1´3 0´9 0´0D No 0´7 7´0 7´7 1´3² 1´1 0´4 0´4² 0´7 1´0Locums various Ð Ð 0´9 1´0 0´7 1´3 0´9 1´3A Yes 2´0 1´0 3´0 1´0 1´5** 0´5² 0´3** 0´4 0´0**G Yes 3´5 0´0 3´5 0´9 1´2 0´8 0´8 1´5 0´0K Yes 3´0 1´0 4´0 1´2 0´3** 4´1** 1´3 0´9 0´5

² , P , 0´05.*, 0´05 , P , 0´01.**, 0´01 , P , 0´001.

Kjellands forceps. His/her frequency of resorting tocaesarean section after a failed trial of vaginal deliv-ery was above average, though not signi ® cantly so,but there were no recorded cases of delivery in thesecond stage before a trial of vaginal delivery. Regis-trar B used Kjellands forceps more than average, andperformed caesarean section less often after a failedtrial, though more often in the second stage beforeany attempt at vaginal delivery. Registrar K, withsome experience of ventouse, seemed to use Kjel-lands forceps in preference to straight forceps.

One further registrar (A) seemed to have a prefer-ence for straight forceps over both Kjelland’ s andventouse whilst apparently completely avoiding theneed for caesarean section after a failed trial offorceps. This was accompanied by low use of cae-sarean section in the second stage without a trial offorceps and average use of caesarean section beforethe second stage. Registrar A performed the largestnumber of deliveries during the study period, due tobeing the longest in post, and would appear to beexemplary with regard to avoidance of caesareansection as a mode of delivery. However such a con-clusion is unsafe without knowing the number ofspontaneous deliveries occurring during this regis-trar’ s time on call. If the propor tion of operativedeliveries, compared with spontaneous deliveries,were high this would imply early intervention whichfor a propor tion of cases would have been unneces-sary, by the standards of the other registrars, andtherefore likely to have been relatively easy techni-cally. Even so the absolute number delivered byregistrar A by caesarean section in the second stagewas remarkably small.

With regard to caesarean section in the secondstage of labour there was one registrar (F) whoappeared to resort to this much more often without atrial of vaginal delivery, despite apparently average

use of both Kjellands forceps and ventouse, thoughless than average use of straight forceps. There werethree registrars (I, L and H) who had a higher thanexpected propor tion of caesarean section after failedtrial of forceps. One of these (L) had a higher thanaverage ventouse rate, while another (I) had a higherthan average Kjellands forceps rate. The other (H)had lower than average rates for all assisted vaginaldeliveries and a higher than average ratio of cae-sarean section in the ® rst stage. This operatorappeared unwilling to deliver women vaginally!

All the four registrars (F, H, I and L) who per-formed a higher propor tion of caesarean section in thesecond stage of labour had not yet achieved theirMRCOG. Furthermore three of them (registrars F, Hand I) had less than 2 years’ obstetric experience inthe UK.

DiscussionAlthough there is considerable interest in variations inintervention rates in maternity care, particularly bythose involved in setting contracts and standards, thevariables which affect these differences have receivedlimited attention and there are no regularly usedmeasures available to correct for the effect of knownvariables. If¯ and and Gordon (1996) demonstrateddifferences between some of their obstetric registrarsin the rate of caesarean section after failure to deliverwith forceps. Our study showed a similar differencebetween registrars as well as a difference in all modesof delivery in a consecutive group of nulliparoussingleton term cephalic presentations. This lattergroup, which were close to those de® ned elsewhere as`standard nullipara’ , can be presumed to have startedlabour from a position of similar risk. If the distri-bution of additional risks of requiring interventioncan be assumed to have been approximately random

J O

bste

t Gyn

aeco

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

CD

L-U

C S

anta

Cru

z on

10/

30/1

4Fo

r pe

rson

al u

se o

nly.

In¯ uence of obstetrician on mode of delivery 523

between the registrars during the study period, then itis clear that their choices and decisions on the appro-priate mode of delivery differ substantially from oneanother. The assumption of random distribution ofrisks has not been tested, but is likely to be true.Apart from a small propor tion of private patients, andoccasionally attendance during of ® ce hours at deliv-ery of a patient in the care of their own consultant’ steam, there were no systematic features of access toregistrars, of which the authors were aware, whichwould prevent them being responsible for all `stan-dard nullipara’ during their time on-call. The pro-portion of deliveries for which the most seniorobstetrician present was a consultant appeared ran-domly distributed and in any case was small enoughnot to have much in¯ uence on the overall pattern ofrisk to which the registrars were exposed. Some ofthe deliveries ascribed to consultants would havebeen while supervising the more junior registrars, orthose new to the unit, thus reducing some of theeffects that might otherwise have been observed dueto inexperience.

Our study agrees with the observations of Johansonet al. (1993) that almost all operative vaginal deliver-ies in the UK are performed by obstetricians in thetraining grades. It is interesting to speculate howregistrars improve their judgement and skills fordeciding and executing the most appropriate mode ofoperative delivery. Some registrars have had a greatdeal of overseas experience and for the others thesenior registrars and registrars themselves contributeto the training of their junior colleagues. Howevermost registrars appear to learn, unsupervised, fromtheir own experiences. The observation that the fourregistrars who performed a higher propor tion of cae-sarean sections in the second stage of labour had notyet achieved the MRCOG and that three of them hadless than 2 years’ obstetric experience in the UKimplies that relatively junior registrars are less able orwilling to achieve instrumental vaginal delivery. Thiscould well be commendable caution with recognitionof their personal skill level and might bene® t thecases for whom they had responsibility. Equally, theirmore experienced colleagues may be achievinginstrumental vaginal delivery at the expense of otherbene® cial outcome measures.

It would have been useful to assess whether theoverall intrapartum intervention rate amongst stan-dard nulliparous women differed according to whowas the on-call obstetric registrar. However, thiswould have required ascribing spontaneous normal

deliveries to the registrars in the study as if it were anadditional mode of delivery for which they wereresponsible. We attempted to do this by assigningspontaneous deliveries in any 24-hour period, begin-ning at 09´00 hours, to the registrar on-call for theevening and night of that day. Some differences werenoted in the ratio of observed to expected, comparedwith those illustrated in Table II, although they wererelatively small. However, the analysis was con-sidered insecure as of ® ce hour weekday cover wasnot documented and was often by a registrar otherthan the one on duty after 17´00 hours. For accuratedata, a prospective study would be required whichwould assign each spontaneous delivery to the oper-ator who was involved in care before delivery orwould have been called if an abnormality, whichcould have led to operative delivery, had occurred.

It was not the purpose of this study to judgeindividual pro ® ciency in intrapartum care but simplyto see whether differences existed between differentregistrars practising in the same unit with a homoge-neous obstetric population. It is clear that the exist-ence of the individual variation that we have shownwill need to be taken account of in future studies, aswell as additional variables predictive of operativedelivery and additional outcome measures such asperineal and/or fetal damage and maternal satisfac-tion.

AcknowledgementsThe authors would like to thank Kenneth Fear, computingassistant , Tracy Gray, senio r house of® cer in obstetric s andgynaeco logy, Solihul l Hospital , and Ann Tonks of WestMidland s Perinata l Audit, for help with data collectio n.

ReferencesIf¯ and C. A. and Gordon H, (1996 ) Failed forcep s and

ventouse: an audit of registra rs in training . Journal ofObstetric s and Gynaecology, 16 , 83±85.

Johanso n R. B., Rice C., Doyle M., Arthur J., Anyanwu L.,Ibrahim J., Warwick A., Redman C. W. E. and O’ BrienP. M. S. (1993 ) A randomised prospect ive study compar-ing the new vacuum extracto r policy with forcep s deliv -ery. British Journa l of Obstetric s and Gynaecology, 100,524±530.

Paterson C. M., Chapple J. C., Beard R. W., Joffe M., SteerP. J. and Wright C. S. W. (1991) Evaluatin g the qualityof the maternity servicesÐ a discussio n paper . BritishJournal of Obstetric s and Gynaecology, 98 , 1073±1078.

J O

bste

t Gyn

aeco

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

CD

L-U

C S

anta

Cru

z on

10/

30/1

4Fo

r pe

rson

al u

se o

nly.