Quality of cares, management of obstetrical risk and mode of delivery in Quebec (QUARISMA)- English version

RATIONAL: The caesarean section rate continues to rise in industrialized countries. In Quebec, the rate increased from 18.5% in 2000 to over 22.6% in 2004.  The World Health Organization (WHO) recommends that not more than 15% of births should be delivered by caesarean section. The rising caesarean rate and the growing perception of caesarean births as routine raise important public health issue. Morbidity and the long-term complications resulting from caesarean births are not well understood, in particular among low-risk patients of obstetric complications. Since the mid-1990s a number of studies have raised concerns about the risk of materno-foetal morbidity associated with caesarean births compared to vaginal delivery. This study has been developed to provide an adapted solution to this problematic, especially when the elective primary caesarean section rate without medical justification increases. QUARISMA is a multifaceted evidence medicine program based on audit and feedback, local opinion leader and educative activities, according to WHO guidelines for the evaluation of quality of care in obstetrics. This program focuses on the active participation of local health professionals in the intervention process.

HYPOTHESIS:  Primary - The QUARISMA program will result in a 20% reduction in the rate of caesarean section among the hospitals following the intervention compared to control hospitals. Secondary - This program will result in: 1) a reduction in materno-foetal morbidity, including a reduction in severe morbidity among low risk patients; and 2) no augmentation in materno-foetal morbidity among high risk patients.

STUDY DESIGN: This is a cluster randomized controlled trial in Quebec, multi-centered, stratified and non-blinded, to evaluate the effectiveness of the QUARISMA program between an intervention group and a control group. Randomisation and intervention unit is the reference hospital, and the unit of analysis is the patient.

POPULATION: Hospitals will be recruited in Quebec. Public hospitals with functional surgical rooms, more than 300 deliveries per year, a caesarean rate ≥ 17%, and written agreement to participate in the study from the directors of maternity services and professional services, will be included in the trial. In order to participate in the study patients must meet the following criteria: Inclusion - Patients carrying a viable foetus during the course of the study. Exclusion - Patients that give birth or abort before 24 weeks of gestation; patients that die prior to giving birth; patients that give birth to a baby weighing less than 500 grams.

METHODS: 26 hospitals will be stratified according to the level of care offered with 13 randomized to the QUARISMA program and 13 to the control group. A minimum of 7,800 patients per year will be recruited during the 3.5 years of the program, with a mean expected recruitment of 25,000 patients per year. A one year pre-intervention period, during which time baseline data will be collected, will be followed by a year and a half intervention phase, and a subsequent one year post-intervention phase. Four types of data will be analysed during the audit sessions: 1) a situational analysis; 2) the medical files of low-risk patients delivered by cesarean section; 3) statistics of the department; and 4) Satisfaction investigations for the patients included in the audit sessions. The trial data will be obtained from hospital registries and patient medical files before, during and after the intervention period, using standardized procedures. The caesarean section rate during the post-intervention period will be the primary outcome for judging the efficacy of intervention, and will be estimated by the Odds Ratio given by the GEE method and tested by the unilateral Wald's test. This study will also examine the health economic questions. The study will be managed by a multi-disciplinary committee with expertise in clinical research. An independent committee will follow the progress of the study and monitor the quality of data and the analysis.

RESULTS:  This study will provide strong evidence to promote an approach in which the health professionals are actively implicated in the analysis and the modification of health care processes to improve the quality of the obstetric care services. The results of this study will help health care professionals to improve the quality of the obstetric services, institute standards of quality control, meet the information needs of patients with regard to childbirth options, and significantly reducing the rates of caesareans section rate.

 

QUESTIONS 1: What is your opinion about the design of the study and the description of the intervention ?

QUESTION 2: In your opinion, is this program is a transdisciplinary program ? and why ?

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Dear Dr. Chaillet, Thank you

Dear Dr. Chaillet,

Thank you for posting a description of your proposed cluster randomized trial evaluating an intervention to reduce cesarean delivery rates (while also monitoring maternal, fetal and infant outcomes). This is a very topical issue and your study will likely generate a lot of interest. Best wishes for success in the CIHR grants competition.

Could I request you to discuss the design issues a little further? Specifically, I was interested in the outcome being analysed and the specific contrast being made. I am unclear as to which of the options listed below was being proposed

a) Compare the cesarean delivery rate in the post-intervention period in the intervention arm with the same rate in the non-intervention arm.

b) Compare the change in cesarean delivery rate from the baseline period to the post-intervention period in the intervention arm vs the same change in the non-intervention arm.

I wonder if the answer to the above question depends on how you would check to see if randomization produced comparable/balanced groups. Would you look at the baseline cesarean delivery rate in the 2 groups? What if they are different? What if the 2 arms have similar cesarean delivery rates at baseline but the populations have different profiles in terms of risk factors (e.g., older maternal age, obesity, etc).

I'm not very familiar with cluster randomized trials and your trial appears to have some unique features. I apologize if my thoughts on this are a little fuzzy. Looking forward to your comments on these issues.

Best wishes.

K.S. Joseph

Dalhousie University and the IWK Health Centre, Halifax

Dear, Thank you for your

Dear,

Thank you for your very interesting comment

Baseline imbalance represent a large problematic in Cluster Randomized Controlled Trial and can be adressed by several way.

1) By the determination of the Odds Ratios of the cesarean section rate between intervention and control group during the follow-up, adjusted by the cesarean section rate between intervention and control group during the baseline (Ukoumunne, 2001, Statistics in Medicine, 20, 417-433)

2) In this trial, we choosed another similar method, the Generalized Estimating Equations (GEE). This method consider the change in variation between baseline groups and follow-up groups (inter -intra-cluster variation) and allow adjustement procedures if potential imbalance are detected between groups. Effect size is given by an adjusted Odd Ratio and tested by the unilateral Wald test.

Thank you again for your comment

I ll be really happy to continue this very interesting topics

Sincerely

Nils Chaillet

Thanks for your reply Nils,

Thanks for your reply Nils, which I found to be helpful.

This topic seems to hold a lot of potential for discussion. I have a few more questions. Suppose the baseline cesarean delivery rates in the intervention arm and the control arm of the trial are identical (which would be great!). Do you see any potential for post-baseline confounding by subject characterisitcs? What factors are you planning to control for in your GEE model besides baseline cesarean delivery rates? Would you control for baseline subject characteristics (e.g., older maternal age) or postintervention subject characteriscs or both?

Looking forward to your reply.

Thanks and best wishes.

KS

Dear KS and Nils, Thank you

Dear KS and Nils,

Thank you for starting an interesting discussion. Decisions about how to deal with baseline variables in cluster randomized trials are tricky, and imbalances are more probable in cluster trials than in individual-randomized trials.

Donner and Klar recommend something along the lines of what Nils is doing - analysis using GEE or some other suitable approach for correlated binary outcomes adjusting for relevant imbalances. I prefer to not adjust unless it is necessary, and if it is, to present both analyses.

Robert Platt

Dear Yes I partage the point

Dear

Yes I partage the point of view of Robert Platt

But Imbalance between control and intervention group is very probable in C-RCT, in this case adjustement would take into account the cluster level (here the hospital) instead the individual level (better adapted for an RCT)

However, to present the both results seem to be a very good compromize

Tks for this information

Nils

Hi Nils, Good to see that

Hi Nils,

Good to see that you have submitted your project and I think it looks very interesting. Would love to see how the intervention actually works and I think it is great that someone actually works on actively getting the rate of caesarean sections down!

However, I do have a couple of questions regarding your analysis of complications.

1. As far as I can see, you are planning to assess the complication rate from the medical chart. Would it be possible to assess this prospectively, e.g. by a form that physicians fill in at delivery, caesarean section or at discharge? Or even have a research nurse keep the records on a daily basis? I think that would improve the quality of your data significantly since doctors are not always consistent in keeping the medical records up to date and trying to retrieve the data from patient charts is a major hassle.

2. Complication rates are much higher in elective, planned caesarean sections and emergency caesarean sections. This may potentially confound your results if in your intervention group the decrease elective CS rate results in an increase in emergency caesarean sections. Are you planning to assess the type of CS (elective/emergency) and indication for CS as well?

Just a couple of my thoughts... Again, my compliments on your study!

Barbra