Orthopaedic physical therapy is considered safe, based on a lack of reported harms. Most of the research until now has focused on benefits. Consideration of benefits and harm involves informed consent, clinical decision making, and cost-benefit analyses. Benefits and harms are treatment and dosage specific. There is currently an insufficient number of dosage trials in orthopaedic physical therapy to identify optimal dosage for common interventions, including exercise and manual therapy. Published cases of severe adverse events following chiropractic manipulation illustrate the need for physical therapy to have high-quality data documenting the safety of orthopaedic physical therapy, including cervical manipulation. A recent systematic review identified poor reporting standards of harms within clinical research in this area. Lack of standardization of terminology has contributed to this problem. Pharmacovigilence provides a framework for terms that orthopaedic physical therapy can adapt and thereafter adopt into clinical practice and research. Adverse events are unexpected events that occur following an intervention without evidence of causality. Where temporality of an event is highly suggestive of causality, the term "adverse reaction" may be more appropriate. Future studies in orthopaedic physical therapy should adopt the CONSORT statement extension on the reporting of harms, published in 2004, to ensure better reporting. Consistent reporting of harms in both research and clinical practice requires professional consensus on terminology pertaining to harms, as well as defining what constitutes an adverse event or an adverse reaction. Widespread consultation and consensus should support optimal definitions and processes and facilitate their implementation into practice. This paper is focused on theoretical considerations and evidence in terms of harm reporting within physical therapy using cervical manual therapy as an example.