This review of 26 English-language reports of behavioral intervention research on gambling published between 2000 and 2011 suggests a need for improved reporting transparency. The average report received positive responses to just over 61.5% of 59 ATQ questions, and the average ATQ question received a positive response from about 64% of studies. Further, about 30% of the particularly important questions were answered positively in just 25% of the reviewed studies.
The gambling behavioral intervention literature’s need for improvement is most evident in its methodological deficits. For instance, fewer than 25% of the reviewed reports described their sampling methods and how they minimized potential bias due to non-randomization or randomization. Without sufficient information about the methods that are used to assign participants to groups, readers cannot fairly evaluate the likelihood of bias in group assignment [26
]. In keeping with current standards, reports should describe who generated the allocation sequence, who enrolled participants, and who assigned participants to interventions.
Users of effectiveness research must be able to evaluate whether an intervention is likely to be effective in their patients. This may not be possible for consumers of much of the gambling intervention literature because fewer than 25% of study reports described the baseline characteristics of study participants lost to follow-up and retained overall or by study condition. Further, the reports did not compare whether the study participants were likely to be similar in important characteristics (such as age, gender, severity of gambling disorder) from those in the target population.
Gambling behavioral intervention reports need to be more transparent in their descriptions of who participated in the research, who dropped out, and if the drop out rates differed between experimental and control groups. TREND advocates using a flow diagram that can assist investigators in reporting the study patients’ progress throughout the course of the research. The diagram enables readers to more accurately evaluate the potential impact of the intervention on their patients. The gambling literature also needs to be explicit about unanticipated consequences of participation and provide an explanation for the interventions effects in patients who have the selected characteristics.
Reporting quality did not change over time unlike other fields in which reporting quality has improved in recent years with the introduction of guidelines [15
]. It is important to note, however, that although reporting in other fields has improved, uniform standards for comparing reporting quality across fields are unavailable [16
], and the gambling literature’s place on the quality spectrum cannot be known.
Adherence to reporting quality is essential if the gambling field is to progress, especially within the context of the growth of evidence-based health care [27
] and its reliance on meta-analyses. A meta-analysis is the use of statistical techniques in a systematic literature review to integrate the results of included studies. Its application requires that investigators report their sampling strategies and statistical methods in great detail. Unfortunately, many guidelines for statistical transparency, such as describing the sampling methods, are neglected in the gambling literature, inhibiting its usefulness for meta-analysis.
Reporting guidelines are important to a field’s growth because they can have a beneficial effect on the way research is conducted. For instance, as the American Psychological Association has noted [28
], the TREND guideline suggesting that dropout rates be reported may encourage researchers to consider what levels of attrition are acceptable, and thus they may come to employ more effective procedures for maximizing the number of participants who complete their study.
Reporting guidelines are not without their critics. Some researchers have raised questions about the criteria for assessing reporting quality [16
]. They point out that reviewers tend to use heterogeneous criteria and do not rely on uniform definitions, thereby possibly limiting the relevance of their reviews [16
]. Another concern raised by researchers is that strict adherence to guidelines may lead to excessive standardization. For example, compliance with reporting standards may fill articles with details of methods and results that are inconsequential to interpretation, resulting in a loss of critical facts to an excess of minutiae [28
]. TREND, however, is a minimum set of standards, so its use does not preclude providing other information that investigators believe will maximize the reader’s understanding of their study’s objectives, methods, and findings.
Despite some concerns about reporting guidelines, almost all major biomedical and public health journals have accepted and even promoted their use. The CONSORT Statement, for example, has been adopted by over 450 journals, and TREND is amassing an increasing number of supporters including Addiction
, the American Journal of Public Health
and the Journal of Alcohol and Drug Addiction
Behavioral gambling intervention reports can improve their transparency by focusing on particularly important guidelines currently neglected in the literature. These include describing the baseline characteristics of study participants who are lost to follow-up and those who remain throughout the study, comparing participants at baseline with the target population; describing the methods used to minimize potential bias due to group assignment, and reporting adverse events or unintended effects that occur in the experimental and control groups.
This review examined published reports in English. It is possible that there are unpublished studies in other languages whose reporting quality may have elevated the gambling literature’s average transparency. However, we had outside experts review our list of eligible study reports, and they found it to be complete and representative.
We reviewed 26 reports. This number represents only a segment of the gambling literature. Other reviewers may have selected different search terms, scoring criteria, or reporting standards. We did not review case or cross sectional studies primarily because checklists like TREND are not currently available for these research designs, which are not typically used to evaluate the effectiveness of interventions.
The number of reports included in this study, however, is consistent with reviews in other fields. For example, a review of the effectiveness of brief alcohol interventions in primary care included 21 studies [30
]. The Cochrane Collaboration’s review of psychosocial smoking cessation interventions to help people with coronary heart disease stop smoking included 16 studies [31
], and its review of combined-pharmacotherapy and psychological therapies for-posttraumatic stress disorder relies on four studies [32
Another potential limitation of this review is that we may have penalized studies for neglecting to report negative events and deviations from protocol as required by the ATQ. It is conceivable that the reports did not contain this information because the investigators did not observe any adverse events and implemented their studies as planned. Without explanatory information from the authors, however, we were unable to distinguish between intentional reporting omissions and lack of transparency. Additionally, standardized transparency scores are not available, and our criteria for evaluating reporting quality may be considered arbitrary or set too high. Finally, individual reports may have had unique characteristics that the TREND statement is not designed to uncover, and if so, our review necessarily failed to identify and discuss them.