Collaborative improvement is one approach being used in many countries to improve prevention of mother to child transmission (PMTCT) and treatment with anti-retroviral therapy (ART) services. Collaborative improvement consists of a network of teams engaging in a structured effort to learn from one another. A recent study analyzing the experiences of 27 collaboratives in 12 countries has shown collaborative improvement’s potential in achieving significant improvements in the level of the quality of care and the sustainability of such results. However, this is one of the first studies in developing countries that examines the effect of collaborative improvement in comparison to a control group.
The Ministry of Health of Cote d’Ivoire and the USAID Health Care Improvement Project (HCI) launched a collaborative improvement initiative in December 2008, in collaboration with implementing partners. The collaborative operated in two phases: the initial demonstration phase, which began in January 2009, and the extension (spread) phase, initiated in August 2010. This collaborative provided an opportunity to: 1) examine whether there is a significant difference in the level of the quality of care between sites that have participated in an improvement collaborative versus those sites that will be in the extension phase and have therefore not yet participated in the collaborative activities; and 2) identify the factors contributing to this difference (if any) in the quality of care provided in the intervention and control sites.
This study uses a modified quasi experimental design, in which the intervention group includes those sites participating in the demonstration phase of the ART/PMTCT collaborative, and the control group is composed of spread sites which had not yet been exposed to the collaborative activities but were planned to be included in the spread phase. Data were collected from 36 of the original 41demonstration (intervention) sites, and 42 spread (control) sites.
Intervention sites saw significantly more improvement in quality of care indicators than control sites for completeness of documentation for PMTCT and ART, and for testing of children born to HIV+ mothers. Complete documentation for PMTCT at intervention sites rose from 22% at baseline to 83% after the collaborative, whereas at control sites during the same period there was only an 8% increase (from 0% to 8%); Complete documentation for ART at intervention sites rose from 22% at baseline to 87% after the collaborative, control sites had a higher baseline at 46% but this indicator barely showed any improvement at the end of the year (49%). Testing of children born to HIV+ mothers also increased at intervention sites. Results related to loss to follow-up for intervention sites do show initial improvement but some of the gains were lost towards the end of the study period. However, control sites experienced significant increases in loss to follow-up over time. Data availability was significantly lower in control sites than in intervention sites.
QI competency and implementation were significantly higher in the intervention group, as were having a standardized process that would allow maintaining gains, mechanisms for orienting new staff, and systems for ensuring resource availability. Few differences in resource availability were noted. Control sites had a higher percentage of clinically trained providers. Intervention sites were likely to have generated change ideas themselves or borrowed these ideas from other participating sites rather than control sites, which, if they implemented the change, were mostly likely to have received the idea from their implementing partner.
Regression analyses, holding other independent variables fixed (resources and clinical competence), showed a strong association between being involved in the collaborative and results related to documentation and testing of children born to HIV+ mothers.
Conclusions and Recommendations
This study has shown that facilities involved in collaborative improvement are able to achieve significant improvement over their own baseline results in comparison to sites that have not participated in a collaborative. Regression analysis indicates a strong association between being involved in the collaborative and improved documentation and increased testing of children born to HIV+ mothers. Time series charts also indicate potential impact on loss to follow-up, although the results were not as well maintained over time. This study is one of the first of its kind in a developing country to demonstrate the effects of participating in collaborative improvement on results achieved in comparison to a control group.