Prescription Opioids: Risk Management and Strategies for Safe Use _ _________________________________
REPORTING BIAS Reporting bias includes a diverse range of bias, mis- representation, distortion, omission, exaggeration, or dismissal of data reported by the authors of a study or of data from other publications [208]. The effect, if not the intent, of reporting bias is to influ- ence reader perception through a persuasive argu- ment that favors the agenda, paradigm, or interest of the author, agency, or institution, or to diminish or discount a competing or opposing perspective. Reporting bias is just as widespread in pain research as it is in other areas of medicine, often appearing as concluding statements of safety or efficacy that are not supported by the actual evidence. A medical issue or problem is considered “hot” when it becomes the focal point of publicity and intense investigation. Reports of research findings are less likely to be true in hotter areas of research. Prejudice can dominate a hot medical field to further undermine the validity of research findings. Highly prejudiced stakeholders can also create obstacles and obstruct efforts to publish information with opposing results [209]. Pressures of vested interests can lead to disappoint- ing research outcomes being “spun” to present the findings in a more favorable light by creative use of data, statistics, and linguistics. Examples of linguistic spin include [210]: • “Treatment X is expected to be a very impor- tant approach in the management of Disorder Y” • “Treatment X effect size approached conven- tional statistical significance.” The use of “spin”—claiming treatment benefit without any supporting evidence from the data—is common, and safety claims with spin without sup- porting data also occur [211; 212; 213]. BIAS IN CLINICAL PRACTICE GUIDELINES Concerns have sometimes been raised regarding bias in the development of clinical practice guidelines, involving the reviewed research, misrepresentation of the data, or failure to assess the quality of the evi- dence supporting the recommendations. Inadequate or weak evidence may lead to conclusions based
APPENDIX: BIAS AND VALIDITY IN PAIN RESEARCH
In addition to training, experience, and clinical judgment, safe and effective treatment of pain is guided by developments in the area of pain medi- cine research. Clinician awareness of refinements, advances, and breakthroughs in the diagnosis and treatment of pain is most directly acquired from reading the published research. Conducting well- designed clinical research is challenging and com- plex. Obtaining accurate and relevant information to apply to patient care is often made more difficult by inadvertent bias and lack of reliable validity in the reporting of research findings. Outright data fraud is rare, but false claims and biased interpretation of results (often unintentional) are commonplace in publications of medical research in general and pain research specifically. In the area of pain treatment with opioid analgesics, major stakeholder influence over the reporting of dangers, risks, benefits, and effectiveness is pervasive [2; 97; 198; 199; 200]. Clinicians trying to make the most of their limited time by reading study abstracts may also be misin- formed. A random selection of studies with abstracts from six widely read and influential medical journals ( JAMA , BMJ , Lancet , NEJM , Annals of Internal Medi- cine , and the Canadian Medical Association Journal ) found that 18% to 68% of abstracts reported infor- mation that was inconsistent with or absent from the body of the paper [201]. PUBLICATION BIAS Publication bias occurs when trials showing statisti- cally significant and positive results are dispropor- tionately published, relative to trials with negative or inconclusive findings. This type of bias is common in published pharmacological research. Pharmaceu- tical industry research sponsorship is associated with significantly higher rates of pro-industry conclusions, publication constraints, and propensity to ignore the publication of negative findings [202; 203; 204; 205; 206; 207].
52
MDMS1526
Powered by FlippingBook