Table 4.

Summary of findings and potential implications for provider behaviour change

	Factors influencing anti-malarial prescription to patients who test negative	Positive influencers (may support adherence to test result)	Potential opportunities for influencing provider behaviour change
Clinical beliefs	Perceived treatment failure or undetectable malaria in patients who already took ACT (may be driven by limited understanding of how RDT works) If provider perceives ‘resistance’ or undetectable malaria, may fail to consider non-malaria febrile illness (i.e. miss true diagnosis)	Appreciation of importance of parasite-based diagnosis Trust in RDT accuracy and belief that a “negative RDT means the patient does not have malaria” Motivated by desire to do the right thing for the patient Motivated by reducing ACT wastage	Provide practical guidance on management of patients who have taken a partial dose of ACT (clear guidelines; may benefit from on the job training and review of context-specific cases) Clarify provider understanding of how RDT works; sharpen communication on limitations of HRP-2-based tests (including ‘prozone effect’) Improve referral guidance and appreciation of benefits (reducing unnecessary treatment, detection of treatment failures, identification of other causes of fever) Provide guidance on risk mitigation for severely ill patients, including role of RDTs
Capacity constraints (ability to make alternative diagnosis)	Limited diagnostic practices (history taking, examination); providers don’t give themselves means to identify alternative cause of fever Lack know-how, diagnostic tools, time	Satisfied to be “treating what [I] know”, despite workload Perceive RDT use as enabling; no longer rely on “guesswork” Know what to do when they get a negative result Acknowledgement of the challenge of making an alternative diagnosis	Increase awareness of the impact of missed diagnoses and overtreatment on patient outcomes (may encourage providers to invest more in the diagnostic process, maximize existing opportunities for information gathering to aid in diagnosis, and resist the urge to provide anti-malarials “just in case”) Develop better methods for assisting providers in diagnostic decision-making: make a negative result ‘actionable’ (i.e. clear algorithms in poster format on managing non-malaria fever cases)
Perceptions of patient demand	Perceive patient “refusal” as relating to test and desire for ACT		Improve interpersonal communication skills and quality of interaction with patient (may facilitate correct diagnosis and treatment) through training, support supervision or patient consultation guides/checklists