This paper summarizes 10 years of evidence of the effectiveness of collaborative improvement in improving health outcomes and compliance with health care standards. The collaborative improvement approach was designed by the Institute for Healthcare Improvement (IHI) in the United States to produce rapid, significant improvements in a targeted area of health care. The paper was commissioned by USAID and analyzes the results achieved by over 1,300 teams of health care providers who participated in 27 improvement collaboratives supported by USAID during 1998-2008.   Data analyzed consisted of 135 time-series charts representing pooled data from groups of teams from 12 countries. All together, the data covered 81 distinct measures of compliance with standards and outcomes for maternal, newborn and child health, HIV/AIDS care, family planning, and malaria and tuberculosis diagnosis and treatment.

Trained clinical observers used a structured checklist at referral and district/regional hospitals in four developing countries to monitor care provided to 245 women during labor, delivery, and postpartum and their newborns during postpartum. The countries were Benin, Ecuador, Jamaica, and Rwanda. Observation periods were either continuous and lasted 72 hours or noncontinuous and lasted 12 hours over 4--6 days; all such periods included a weekend day and night. Observers marked the checklist to record the times when healthcare providers performed certain tasks and whether each had been done according to standard. Certain circumstances--such as a woman giving birth before arrival--required rigorous data cleaning. The quality of care for the different tasks is reported here by country, by hospital type, and overall. The frequency of labor monitoring was well below the rates recommended in all four countries, regardless of hospital type. Fetal heart rate (FHR) was monitored the most frequently at 1.3 times per hour, although its recommended rate in all countries is twice an hour. Other labor indicators recommended at twice per hour were checked less often: maternal pulse was taken 0.43 times per hour, contraction intervals 0.38 times per hour, and contraction duration 0.37 times per hour. The two indicators recommended at the rate of once every four hours (0.25 times per hour) were performed more frequently: maternal blood pressure at 0.63 times per hour and vaginal exam at 1.1 times per hour. On average, in 26% of the cases, no labor indicator was monitored at all. In the three study countries where partograph use is recommended, incorrect use was observed in more than half the case observations, varying substantially by country. Correct partograph use was associated with more frequent labor monitoring. (author's)

Each year, more than 500,000 women worldwide die from complications related to childbirth. With good quality obstetric care, approximately 90 percent of these deaths could be averted. The assistance of a skilled birth attendant during labor, delivery and the immediate postpartum period is one important component of quality obstetric care. An enabling environment for skilled attendance at delivery and prompt attention for women arriving at a medical facility with an obstetric complication are also key factors. However, little is known about the competence of skilled birth attendants, the elements that contribute to an enabling environment and the causes of what is commonly known as the "third delay:" the delay in receiving medical attention after a woman arrives at a healthcare facility. Through its Safe Motherhood Research Program, the Quality Assurance Project carried out three studies to explore these issues in countries with high maternal mortality ratios. The first study examined the competency of skilled birth attendants (SBAs). The second measured SBA performance and the relative contribution to performance of different enabling factors in the work environment. The last study examined causes of in-facility delays in receiving obstetric care. All three studies were carried out between September 2001 and July 2002 in Benin, Ecuador, Jamaica, and Rwanda. This report presents the results from Benin. The Benin studies included five hospitals: a tertiary care referral hospital with an active maternity department, two secondary care hospitals, and two smaller district hospitals. The competency study measured knowledge with a 72-question test covering six subject areas. We also tested skills in several key areas including ability to use a partograph, neonatal resuscitation, manual removal of placenta, bimanual uterine compression and IV insertion. Finally, we asked participants to assess their own ability to carry out common obstetric procedures. Results show that current competency levels are low. Two key skills, bimanual uterine compression and active management of third stage labor, are rarely taught and rarely practiced. We found little difference in competency by professional level. Physicians do somewhat better with complex procedures, but professional level appears inversely related to the interpersonal quality of care. There was no correlation between providers' self-assessment and their competency as measured by the knowledge and skills tests. The enabling environment study addressed the contribution of enabling factors and essential elements to health worker performance. We used an observation checklist to evaluate performance during labor, delivery and the immediate postpartum period. We reviewed medical records to evaluate performance at managing obstetric complications. We also surveyed providers in each facility about supervision, training and motivation. Finally, we inventoried the availability of essential drugs, equipment and supplies in each study hospital. Although most staff had extensive professional training, less than half said they had received any training in the last two years. Labor monitoring, including checking fetal heart rate and mother's blood pressure, was inadequate in most observed cases. Providers used a partograph only about two-thirds of the time. Few washed their hands before assisting at delivery and none suctioned the newborn after birth, but most administered oxytocin to the mother after delivery. The third delay study used direct observation to analyze patient flow in three of the five study hospitals. In addition, two physicians reviewed medical records to determine if there were delays at different points in the care of the patient. Most of the delays found in the record review occurred during diagnosis, especially for obstructed labor and severe pre-eclampsia/eclampsia. Waiting times after arrival at the hospital or the OB department varied by hospital but were unusually long at the regional referral hospital (RRH), one of the two secondary level facilities included in the study. We recommend a quality improvement initiative to improve medical records at all hospitals. For RRHs, we also recommend developing a systematic triage system to reduce waiting times for incoming patients. The Quality Assurance Project's safe motherhood research is supported with Maternal Health funds from the U.S. Agency for International Development (USAID). (author's)