Rationalizing the use of caesarean section in Egypt through implementing 10 group Classification system

- The Robson Classification Report Table

In order to make the most of the information provided by the Robson Classification in local settings and to allow comparisons between settings, the data is best reported in a standardized way (the “Robson Classification Report Table”)

Table 6 : The Robson Classification Report Table

Setting name: Hospital ABC				period:
Colum n 1	Column 2	Column 3	Column 4	Column 5	Column 6	Column 7
Group	Numbe rof CSin group	Number of women in group	Grou p Size1 (%)	Grou p CS rate2 (%)	Absolute group contributio n to overall CS rate3 (%)	Relative contributio n of group to overall CS rate4 (%)
1
2
3
4
5
6
7
8
9
10
Total*	Total	Total	100%	Overall	Overall CS	100%
	number CS	number		CS rate	rate
		women
		delivere
		d

1.Group size (%) = n of women in the group / total N women delivered in the hospital x 100

2. Group CS rate (%) = n of CS in the group / total N of women in the group x 100

3. Absolute contribution (%) = n of CS in the group/ total N of womendelivered in the hospital x 100

4. Relative contribution (%) = n of CS in the group / total N of CS in the hospital x 100

Interpretation of the Robson Classification Report Table

Benefits of interrelation

Useful insights into the quality of data collection

- the type of population served by the hospital

- the CS rates of each group

- how each of the individual 10 groups contributes to the overall rate of CS in your setting

- The overall philosophy of care of in a maternity unit.

- Can guide the necessary training and advocacy at each units as well as country level for decreasing cs rate.

- can be used to analyze changes over time

- compare differences between hospitals and to help modify clinical practice to optimize CS rates in specificgroupswhileensuringgood maternalandperinataloutcomes.

GENERAL PRINCIPLES and tips of Interpretation of Robson Report Tables

- In settings with low volumes of delivery, the interpretation should take into account the effect of small changes in numbers on the percentages.

- It is important to assess the sizes of all the 10 groups to ascertain the balance and makeup of the whole obstetric population so it should be clear that No individual group should be interpreted unless the whole 10 groups are analysed first.

- The main three reasons for differences in sizes of groups or events and outcomes within groups are the following:

1- Poor data quality (incorrect information in the patient records or errors in retrieving information from the records)

2- Differences in significant epidemiological characteristics of the populations (age, BMI, etc...)

3- Differences in clinical practice. Only consider differences in practice after you have evaluated quality and epidemiological variables.

Steps of intepretatin the Robson Classification Report Table Three major steps

1) Assessment of data quality

2) Assessment of type of obstetrics population

3) Assessment of CS rate

Table 7 : Steps to assess quality of data using the Robson Classification Report Table

Step	Interpretation by Robson	Example: MCS population**	Further Interpretation
Lookatthe total numbersof CS andof women deliveredin your hospital	These numbers should be identicaltothetotal numberof CSandof womendeliveredin your hospital.	NA	Ifthesenumbersdonot match,thendatais missing or incorrect. Some women may not have been classified in the Robson groups because of missing variables or were incorrectly classified as to type of delivery.
2. Look at the size of Group 9 (Column 4)	It should be less than 1%.	0.4%	If this is > 1%, it is probable that women with breech (or other) presentations have been misclassified as transverse /oblique lie and allocated to this group. As the classification includes all women who have delivered, if any one group is smaller or bigger, look to the other groups which sometimeswillshowwhere the manifestation is.
3.Lookat theCSrate ofGroup9 (Column 5):	Itshouldbe 100% by convention.	88.6%	By convention,if the woman gives birth vaginally by internal version,itshouldbeclassifyas either cephalicorbreech. CS rateinGroup9should be 100%

NB: The examples in interpretation are based on two sources; one was developed by Michael Robson based on his international experience applying the classification since 1990 (28, 32,33) andthe secondsourceisthe WHO Multicounty Survey on Maternaland Newborn Health (WHO MCS) (34, 35) Itshould beemphasizedthat neither of these sources has been formally validated and the CS rates by group presentedinthistablehavenotbeen linked to improved outcomes. In particular, please note that the rates ofCSineach ofthe Robsongroupsin the WHOMCS refertoanaverage obtained from over 60 health facilitiesinlow-andmiddle-income countriesandtherefore may be used as a guide only but cannotand should not be taken as a recommendation to be followed by everyone around theworld

Table 8 : Steps to assess type of population using the Robson Classification Report Table

Step	Robson guideline	Example: MCS population	Further Interpretation
. Look at thesizeof Groups1+ Group 2 (Column 4)-	Thisusually represents35-42% of obstetric population of most hospitals.	38.1%	Insettingswithhighproportionof womenwhohave onlyonechildrather thanmorethanonechild,the groupof nulliparouswomeni.e.Groups1and2 tendsto belarger.Insettingswherethe oppositeistrue,thesize ofGroups1+ Group2willbesmallersincemostofthe populationwillberepresentedby multiparouswomen
Look at the size of Groups3+ 4(Column 4)-	This usually represents about 30% of women	46.5%	Insettingswithhighproportionofwomen withmore thanonechildratherthanonly onechild,thesizeof Groups 3 + Group 4 will be higher than 30% (provided theyhavedeliveredvaginally).Another reasonforalowsizeofGroups3and4 couldbethat thesizeofGroup5isvery highwhichwouldbe accompanied byaveryhighoverallCSrate.
Look at the size of Group 5 (Column 4)	Itisrelatedtothe overallCSrate. The size of Group 5 is roughly usually abouthalfofthetotal CS rate.Insettings withlowoverallCS rates,itisusually under10%.	7.2%	ThesizeofGroup5isusuallyrelatedto theoverallCS rate.Ifthesizeofthis groupislarger,itmeansthat therehas beenahighCSrateinthepastyearsin that hospitalandmainlyinGroups1and 2.Inplaceswith highCSrates,thesizeof thisgroupcouldbe>15%.
Look at thesizeof Groups6+7(Column4)	It should be 3- 4%	2.7%	Ifthetotal ismuch over 4%, the most commonreason isusuallya highrateof pretermdeliveriesorahigher proportion of nulliparous women. Therefore look at sizeofGroup10(Column4).Ifthatisover 4-5%,this hypothesis could be true.
. Look at the size of Groups 8 (Column 4)-	It should be 1.5 - 2%	0.9%	Ifitishigher,thehospitalisprobably tertiary(high risk,referral)orrunsa fertilizationprogram.If lower, probably a lot of the twins are referred

			out especially if the remaining twins have a low caesarean section rate
Look at the size of Groups 10 (Column 4)-	Itshouldbeless than5%in most normalrisk settings.	4.2%	Ifitishigher,thehospitalisprobably tertiary(high risk,referral)orthereisa highriskofpreterm birthsinthe populationthatthehospitalserves.If, inaddition, the CSrate is low inthis group, it could represent a preponderance of spontaneous pretermlabour. If the CS rate inthis group is high, itcouldsuggestmore providerinitiatedpre-labour CSfor fetalgrowthrestrictionorpre- eclampsia andotherpregnancyor medicalcomplications.
Lookat the Ratio of the size of Group1 versusGroup2 (Divide the size of Group 1 bythesize of Group 2, Column 4)	It is usually 2:1 or higher	Ratio 3.3	Ifitislower,suspectpoordata quality:nulliparous women who received oxytocin for augmentation (acceleration) oflabour(andshould be in Group 1) mayhavebeen misclassifiedas“induction”(and incorrectly classified as Group 2). Ifdatacollectionis correct, alower ratio may indicatethatyou havea highinduction/prelabour CS issue which may indicate a high risk populationinnulliparouswomen andarelikely therefore to have a high CS rate. Additional informationonpre-labourstillbirths wouldbethe next question to ask. Onthecontrary,iftheratioisvery high,youmay wanttolook atyourpre- labour stillbirth ratein this population whichmayindicatethatyouarenot inducing enough. Or alternatively you may have a very low risk population
Lookat the Ratio of the size of Group3 versus Group 4.	It is always higher than the ratio of	Ratio 6.3	Ifitislower,suspectpoordata quality:multiparous women who

(Divide thesizeof Group3 bythesizeof Group 4, Column 4):

Group 1/Group 2 in the same

institution, i.e, largerthan2:1.

This is very reliable finding in confirming data quality and culture of the organization.

received oxytocin for “augmentation” oflabour(and shouldbeinGroup3)mayhave been misclassified as “induction” (and incorrectly classified as Group 4).

Alowratio(duetolargeGroup4b) maysuggesta poor previous maternal experience in vaginal delivery and a request for pre- labour CS in multiparouswomen. Another explanation may be pre- labour CS done to perform tubal ligation (common in settings where family planning is not easily

available).

.Lookat the Ratio ofthesize of Group6 versus

Group 7. (Divide thesizeofGroup6 bythesize of Group 7, Column

It is usually a 2:1 because breeches are more frequent in nulliparous women than in multiparous women.

Ratio 0.8

Iftheratioisdifferent,suspecteither unusual nullipara/multipararatio or inaccurate data collection

Table 9: Steps to assess caesarean section rates using the Robson Report Table

Step	Robson guideline	MCS reference population	Further Interpretation
Look at the CS rate for Group 1 (Column 5)	Ratesunder10%are achievable	9.8%	This rate can only be interpreted accurately when you have considered the ratio of the sizes ofGroups1and2.Inprinciple,thehigherthe ratio of size of Groups 1:2, the higher the likelihoodofboththeCSrateinGroup1and2 being individually higher. However, the overall CSrateinGroups1and2combinedmaystillbe low or the same.
Lookatthe CS rate forGroup2 (Column 5):	Consistentlyaround 20-35%	39.9%	CSratesinGroup2reflectthesizeandratesin 2aand2b.IfsizeofGroup2bislarge,theoverall CSratesinGroup2isalsogoingtobelarge.If Group2bisrelativelysmall,thenhighratesof CSinGroup2mayindicatepoorsuccessrates for induction or poor choice of women to induce andconsequentlyahighrateofCSinGroup2a. Remember the general principle of not interpreting one single subgroup on its own without knowing what is left out. The interpretationof group 2arequires knowingthe relative sizes of Groups 1 and 2b.
Lookatthe CS rate for Group3 (Column 5)	Normally, no higher than 3.0%.	3%	InunitswithhigherCSratesinthisgroup,this may be due to poor data collection. It is possible thatwomenwithpreviousscars(Group5) were incorrectly classified as Group 3. Other possible reasons for high rates could be for exampleto do tubal ligation in settings with poor access to contraception, or maternal request.
Look at the CS ratefor Group 4 (Column 5)	Itrarelyshouldbehigher than 15%	23.7%	CS rates in Group 4 reflect the size and rates in 4aand4b.Ifsizeof Group4bis large,theoverall CSrates inGroup 4is alsogoingto behigh.If Group 4b is relatively small, then high rates of CS in Group 4 may indicate poor success rates for induction or poor choice of women to induce and consequently a high rate of CSin Group 4a. Poor data collection could also be a reason for high CSrates inGroup 4; forexample dueto inclusion of women with previous scars in this group (when they

			should bein Group 5). Lastly, ahigh CS rate in Group 4 may reflect ahigh maternal request for CS even if these women havedeliveredtheir first pregnancy vaginally. This may be because of a previously traumatic or prolongedlabour ortodotuballigationin settings withpooraccesstocontraception.
Lookatthe CS rate forGroup5 (Column 5)	Rates of 50-60% are considered appropriate provided you have good maternal and perinatal outcome.	74.4%	Ifratesarehigher,thisispossiblyduetoalarge Group5.2(womenwith2ormoreprevious CS). Thiscouldalsobeduetoapolicyof scheduling pre-labour CS for all women with 1 previous scar without attempting a trial of labour.
Lookat the CS rate for Group8 (Column 5)	It is usually around 60%.	57.7%	Variations will depend on the type of twin pregnancy and the ratio of nulliparous/multiparouswithorwithouta previous scar.
Lookatthe CS rateinGroup10 (Column 5):	Inmost populations it isusually around 30%	25.1%	Ifhigherthan30%,itisusuallyduetomany casesof high risk pregnancies (e.g.fetalgrowth restriction,preeclampsia)thatwillneedpreterm pre-labourCS.Iflowerthan30%,itsuggestsa relatively higher rate of preterm spontaneous labour andhencealoweroverallCSrate.
Look at the relative contributionof Groups1,2and5 totheoverallCS rate(addthe contribution of each of these groups in Column 7)	These three groups combined normally contribute to 2/3 (66%)ofallCS performedin most hospitals.	These three groups combined contributed to 63.7%ofallCS	Thesethreegroupsshouldbethefocusof attentionifthehospitalistryingtolowerthe overall CS rate. The higher the overall CS rate, thegreaterthefocusshouldbeinGroup1.
Look at the absolute contributionof Group5tothe overall CSrate (Column 7)		Thisgroupwas responsible for 28.9% ofall CS	If it is very high, this may indicate that in previous years,CSratesinGroups1and2have beenhigh and it is worth exploring further.

Table 10 Challenges in implementing Robson Classification system and how to overcome

Challenge	Way to overcome
Lack of awareness by this classification system among obstercian and its importance	Training and awareness campaign
Lack of training of physician	Training workshop as this one (TOT and ordinary workshops)
Need of dedicated staff for collecting data	designate a person if possible (clinician, nurse, clerk, manager or other) to be in charge of organizing data collection and producing the Robson Report Tables at weekly or monthly intervals
Lack of definition or consensus on the core variables used in the classification	We therefore recommend that each hospital creates a clear written definition (a glossary) of the variables that may vary in different settings will provide here the international standard accepted in Egypt .
Quality of the data	Ensuring good quality of the data should not be taken for granted and it can be challenging even in high- resource settings.
Misclassification of women in wrong groups	Data review Electronic format ( available here also)
Cases that cannot be classified due to missing data:	The size of “Unclassifiable” categoryisanimportantindicatorofthe quality of the data in the individual patientrecords.
The lack of validation of the interpretation rules	encourageusersoftheclassification to collect theirown data on maternal aswellasperinatalmorbidityand mortality per Robson group and analyze these data regularly