Denig 1990.
| Methods | Study design: RCT Unit of allocation: physicians Stratification by: village or town Type of comparison: PEM only vs. nothing
|
|
| Participants | Physicians Clinical speciality: general practice/family medicine Level of training: fully trained Setting/country: general practice/The Netherlands |
|
| Interventions | The PEM consisted of a 'bulletin' that looked like a regular issue of the monthly Geneesmiddelenbulletin distributed by the Dutch government to all physicians and pharmacists. The bulletin used for the evaluation concerned the use of antispasmodic drugs for 2 kinds of spasms commonly seen in general practice, IBS and renal colic. The bulletin advised against (a) fixed combinations of antispasmodics with chlordiazepoxide, (b) PO/rectal butylscopolamine, and (c) fixed combinations of antispasmodics with metamizole. Recommended for renal colic were (d) diclofenac preparations. The objective was to evaluate the effects of a direct mailed drug bulletin on drug choice and prescribing practice in physicians | |
| Outcomes | 2 process outcomes:
|
|
| Notes | ‐ | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Quote from correspondence with author: "The allocation was conducted by using envelopes drawn by a person who was not involved in the research project |
| Allocation concealment (selection bias) | Unclear risk | No information is provided |
| Baseline characteristics similar (selection bias) | Low risk | Quote, pg. 6: "the physicians participating in this study were similar to their colleagues in The Netherlands with regard to years in practice, size of practice, percentage of elderly patients, and sex distribution of patients (table 5.1). Moreover, there were no significant differences in these characteristics between the control and intervention groups of the study (t‐tests; P > 0.05)" |
| Baseline outcome measurements similar (selection bias) | Low risk | Quote, pg. 7: "...before the intervention, the study groups did not differ significantly in terms of knowledge, perceived drug utility, or stated prescription. Nor did a significant difference occur in actual prescribing between the intervention and control groups (Tables 5.2‐5.5). The physicians in both study groups who were interviewed, however, prescribed fewer antispasmodics in general as well as fewer undesirable antispasmodics than the physicians who did not agree to be interviewed but permitted the use of their prescribing data" |
| Incomplete outcome data (attrition bias) All outcomes | Unclear risk | Reasons are provided for the 25 withdrawal/ineligible participants who agreed to join, but did not form part of the group analysed, but no indication of the distribution between control and intervention |
| Blinding of outcome assessment (detection bias) All outcomes | Low risk | The outcome was objective |
| Contamination protection (contamination bias) | Unclear risk | Quote, pg. 3: "physicians living in the same village or town were stratified into the control or intervention groups." From this quote it is UNCLEAR if a clustered approach was used to randomise participants |
| Selective reporting (reporting bias) | Low risk | All relevant outcomes in the methods section were reported in the results section |
| Other bias | Low risk | There was no evidence of other risks of bias |