Table 2 Scenarios used to facilitate discussions

Scenario 1: Use for routine audit

Clinical hospital data (with people’s names taken off) being used by the County Health Team to assess and report on patterns of different diseases at different times, such as the number of people admitted to hospital with malaria in a given time period.

Scenario 2: Use for evaluations

A public health manager in the County Health Team uses clinical and laboratory data from individual patients who have been treated for malaria in hospital (with names taken off) to evaluate whether new guidelines that have been introduced for the in-patient treatment of malaria are improving clinical outcomes overall and over time.

Scenario 3: CER: Non-randomized Pragmatic Clinical Trials

It is common in medical practice that there are several different treatments available to treat a given condition, without clear evidence that one treatment works better than the other. For example, many different antibiotics are recommended to treat particular infections, like boils, ear infections or lung infections. In this situation, doctors tend to choose a treatment based on their own or their patients’ personal experiences/preferences. If there was more evidence about which treatments work best and in which situations, both patients and doctors would benefit. One way for researchers to do this is to compare routine clinical data on patient outcomes (e.g. how quickly or completely patients got better after being treated by one drug compared to another). In this kind of research, the researchers DON’T introduce anything different to the normal practice. They only analyze clinical data from patients who were treated to compare the effectiveness of different antibiotics used.

Scenario 4: CER: Randomized Pragmatic Clinical Trials

There are many different antibiotics currently approved and used routinely for treating pneumonia. For some of these antibiotics, it’s not known if they work better than others available. For example, let’s think about two such treatments, and call them antibiotic X and antibiotic Y. Both are already approved drugs and are in use at the moment. They are given in similar ways and have similar types and risks of any side effects or more serious reactions. (Serious reactions are very rare). It is therefore unlikely that patients or physicians would have a personal preference for one drug over the other. To find out if there are any differences between these treatments, researchers can ask physicians to prescribe one of these drugs based on a system of chance, and observe over time how well patients respond to the treatments. Over time, the outcomes of patients being treated with one of these two antibiotics can be compared to learn which treatment works best. Once this is known, all the patients can be given the option to change to that treatment.