He also assigned the experimental group weekly visits from a non-professional high school graduate who reinforced regular pill taking behavior.

[10] Haynes then shifted his research focus on documenting the problems of evidence handling by practitioners and proposing solutions.

He wrote an article explaining the constraints faced by medical practitioners in handling research and clinical evidence.

Haynes suggests the use of measurement principles and information tools to reduce the inaccuracy in the data gathering, interpretation, communication and application of the research evidence for patient care.

[11] Haynes worked on evidence retrieval from bibliographic databases and published over 50 articles on the theme of 'Developing optimal search strategies for detecting clinically sound studies in MEDLINE' as well as EMBASE, CINAHL, and PsycLIT.

He pioneered the use of machine learning to empirically determine the optimal search filters for various types of studies (e.g., prevention, treatment, diagnosis, prognosis, causation, quality improvement and cost-effectiveness).

The article concluded that there is a dire need for innovation in the methods to assist and encourage patients to follow the medical prescriptions.

[14] Haynes' research has resulted in structured abstracts of medical journal articles along with online search filters such as "Clinical Queries" which are often used by MEDLINE, EMBASE and other bibliographic databases.

He is one of the originators of Evidence-Based Medicine, led by David Sackett, along with Gordon Guyatt, Peter Tugwell and Deborah Cook) for which his work has been critically acclaimed.