Measuring outpatient safety at scale infection prevention and control practices in Kenya.


Primum non nocere--first, do no harm. This most basic tenet of medical care is routinely violated in clinics and hospitals around the world today ( But the absence of routine data means that when it comes to improving patient safety, we are often in the dark. Available studies frequently rely on self-reported data from healthcare workers, focus on a single domain such as hand hygiene and are based on small samples: one review ( found that only 10 of 41 studies on hand hygiene interventions were collected in more than one hospital!

Beyond the problems of generalizing from self-reported data in small samples, the focus on single domains can impede much needed epidemiological modelling given multiple pathways of nosocomial (i.e., healthcare-associated) infections and a range of violations to safety practices that can occur in these settings. To address these problems, in 2015 we conducted the largest patient safety survey ( across low- and middle-income countries, covering 1,035 facilities in three Kenyan counties and a total of 1,680 healthcare workers and 14,328 patients.

These data--collected as part of the larger Kenya Patient Safety Impact Evaluation or KePSIE trial ( on substantial investments by WHO in developing and validating [dagger] ( he tools ( required to study violations across multiple domains such as safe injection practices and hand hygiene. As a start, in 2013, we visited several facilities to understand the 'average' patient experience. Apart from glaring violations (see picture), we noticed that consultations in Kenya typically involve visits to different parts of a facility. For instance, a patient may start off in the consultation room, then be advised to get his or her blood tested, then return to the consultation room, and finally, be sent to the injection room for treatment.

Clearly, restricting ourselves to a single domain would not present the full picture--patients could face violations at any point, and each of these safety violations could contribute to nosocomial infections. The new tool that we implemented specifically addressed this point.

What did we do?

We identified three key procedures for...

To continue reading