Effect of Educational Outreach Timing and Duration on Facility Performance for Infectious Disease Care in Uganda: A Trial with Pre-Post and Cluster Randomized Controlled Components: e0136966

Background Classroom-based learning is often insufficient to ensure high quality care and application of health care guidelines. Educational outreach is garnering attention as a supplemental method to enhance health care worker capacity, yet there is little information about the timing and duration required to improve facility performance. We sought to evaluate the effects of an infectious disease training program followed by either immediate or delayed on-site support (OSS), an educational outreach approach, on nine facility performance indicators for emergency triage, assessment, and treatment; malaria; and pneumonia. We also compared the effects of nine monthly OSS visits to extended OSS, with three additional visits over six months. Methods This study was conducted at 36 health facilities in Uganda, covering 1,275,960 outpatient visits over 23 months. From April 2010 to December 2010, 36 sites received infectious disease training; 18 randomly selected sites in arm A received nine monthly OSS visits (immediate OSS) and 18 sites in arm B did not. From March 2011 to September 2011, arm A sites received three additional visits every two months (extended OSS), while the arm B sites received eight monthly OSS visits (delayed OSS). We compared the combined effect of training and delayed OSS to training followed by immediate OSS to determine the effect of delaying OSS implementation by nine months. We also compared facility performance in arm A during the extended OSS to immediate OSS to examine the effect of additional, less frequent OSS. Results Delayed OSS, when combined with training, was associated with significant pre/post improvements in four indicators: outpatients triaged (44% vs. 87%, aRR = 1.54, 99% CI = 1.11, 2.15); emergency and priority patients admitted, detained, or referred (16% vs. 31%, aRR = 1.74, 99% CI = 1.10, 2.75); patients with a negative malaria test result prescribed an antimalarial (53% vs. 34%, aRR = 0.67, 99% CI = 0.55, 0.82); and pneumonia suspects assessed for pneumonia (6% vs. 27%, aRR = 2.97, 99% CI = 1.44, 6.17). Differences between the delayed OSS and immediate OSS arms were not statistically significant for any of the nine indicators (all adjusted relative RR (aRRR) between 0.76-1.44, all p>0.06). Extended OSS was associated with significant improvement in two indicators (outpatients triaged: aRR = 1.09, 99% CI = 1.01; emergency and priority patients admitted, detained, or referred: aRR = 1.22, 99% CI = 1.01, 1.38) and decli