Table 1. Quality assessment standard for a cross-over study.
| Item | Description | Scoring | |
|---|---|---|---|
| 1. Appropriate cross-over design | Three points are considered: (1) the condition of the patients should be chronic and stable; (2) the intervention should not provide permanent change, but rather temporary relief; (3) the effect of the first intervention should not last into the second treatment period. | Low: all the three points are absolutely correct; | |
| Unclear: it hard to judge because some information was missing or ambiguous; | |||
| High: one or more points are incorrect. | |||
| 2. Randomized treatment order | The order of receiving treatments should be randomized adequately. | Low: the method is appropriate and clearly described; | |
| Unclear: it is described as “randomized”, but it is hard to judge whether the implementation was adequate because some information (method, etc.) was not provided; | |||
| High: the method is inappropriate, or no randomization is applied. | |||
| 3. Carry-over effect | The authors should evaluate the carry-over effect and provide relevant information clearly. | Low: carry-over effect was evaluated and the results showed no carry-over effect; | |
| Unclear: carry-over effect was not evaluated, and it is hard for evaluators to judge; | |||
| High: carry-over effect was evaluated and the results showed apparent carry-over effect, or indicated evidently from some other provided information. | |||
| 4. Unbiased data | That only first-period data are available is considered a risk of bias. | Low: data for every period are provided; | |
| Unclear: data are unavailable for part of outcomes, or only analytical results are provided and it is hard to judge whether the results are analyzed based only on data from the first-period or every period. | |||
| High: only first-period data are available. | |||
| 5. Allocation concealment | The study should apply appropriate approaches to ensure the allocation sequence is concealed. | Low: allocation sequence was concealed adequately by appropriate methods; | |
| Unclear: concealment approaches were not described, or relevant information was ambiguous; | |||
| High: no approaches to allocation concealment were used, or concealed inadequately. | |||
| 6. Blinding | The study should apply a proper blinding method to prevent performance and detection bias. Those involved in blinding (participants, doctors, measurers, or analysts) depends on the particularity of the studies. | Low: appropriate blinding method was applied; No blinding, but the outcome and the outcome measurement are not likely to be influenced by lack of blinding; | |
| Unclear: relevant information was not provided; | |||
| High: no blinding method was applied, or applied incorrectly, or ineffectively, which very likely affected the outcome. | |||
| 7. Incomplete outcome data | The authors should provide relevant information about the completeness of outcome data, including the level of incompleteness, reasons, and analytic method to tackle these data shortcomings, etc. | Low: no missing outcome data, or the reason is acceptable, or missing outcome data were appropriate analyzed; | |
| Unclear: it is hard to judge because some information was not provided; | |||
| High: missing outcome data existed and the reasons were unacceptable, and the analytic method was inappropriate. | |||
| 8. Selective outcome reporting | The authors should report all the outcomes fully. Selective reporting of part of outcomes or data for an outcome or subsets of the data or analyses using the same data and etc. should be avoided. | Low: fully reported; | |
| Unclear: it is hard to judge due to the unavailability of some original information; | |||
| High: the reports of the study suggest a high risk of selective outcome reporting. | |||
| 9. Other bias | Any other potential risk of bias that may affect the quality of cross-over studies. | Low: the study is apparently free of other problems; | |
| Unclear: whether certain problems existed and led to a risk of bias is uncertain; | |||
| High: high risk of bias existed due to evident problems. | |||
NOTE: the standard was summarized from the Cochrane Collaboration’s tool for assessing risk of bias and the Cochrane handbook`s suggestions for assessing risk of bias in cross-over studies. The assessment of some items, especially items 5–8, are almost the same as that described in Cochrane Collaboration’s tool for assessing the risk of bias.