Criteria for considering studies for this review
This systematic review and meta-analysis will be conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [41] and detailed in Table 1. This systematic review protocol has been registered with the International Prospective Register of Systematic Reviews (PROSPERO) database (registration number: CRD42020215446).
The PRISMA checklist [41] encourages authors to describe eligibility criteria using the PICO reporting system (which describes the participants, interventions, comparisons, outcome(s), and study design of the included studies).
Types of studies
This review will include randomized controlled trials (RCTs) and quasi-experimental studies that compare light therapies with treatment as a control condition. Studies will be included regardless of light treatment intensity, duration, time of lighting exposure (e.g., morning, midday), or follow-up time.
Types of participants
Studies involving cancer patients, regardless of age, sex, tumor type, and type of treatment, will be included. Pregnant and lactating women will be excluded. Patients who had severe psychiatric disorder (e.g., severe major depressive disorder), suicidal ideations with a risk of attempting suicide, diseases contraindicating light therapies (e.g., severe cataracts, diabetes), and patients who had been taking photosensitive medication (e.g., imipramine) were excluded.
Types of interventions
Studies reporting the effects of a light intervention (e.g., bright light therapy) on multiple outcomes will be included if fatigue is one of the outcomes of interest. Studies of light therapy combined with another treatment, such as cognitive behavior therapy and the control condition, that did not include the other treatment will be considered for exclusion.
Types of comparators
The control group will be a waiting list, dim red light, usual care, psychotherapy (e.g., cognitive behavior therapy), exercise or medications.
Types of outcome measures
Primary outcomes
The included studies must have fatigue as an outcome of interest. This will incorporate fatigue measured as the main outcome or within a cluster measurement of physical symptoms or quality of life.
Studies will be included if fatigue was measured by a questionnaire specifically designed to evaluate fatigue. Fatigue subscales as sections of a broader quality-of-life measure will also be included if specific fatigue-related data are available. Fatigue may be measured in terms of characteristics such as intensity, distress, duration, or frequency or as dimensions such as physical fatigue, mental fatigue, or general fatigue.
Secondary outcomes
-
Mood disturbances (e.g., depression and/or anxiety or distress) were measured by any reliable and valid instrument.
-
Sleep difficulties and subjective data are obtained from any reliable and valid self-reported scale, and objective data are obtained from Actiwatch and/or Actigraphy.
-
Cortisol rhythms measured by saliva and/or blood samples.
-
Quality of life was measured by any reliable and valid instrument.
-
Adherence.
-
Adverse events.
Search methods for identification of studies
No date restriction will be imposed on the studies. Studies will be included if a full-text paper in English is available either through databases or through contact with the study authors. Whenever available, protocol methods will be compared with the methods and results reported in the included study.
Electronic searches
We will search the following electronic databases: PubMed, Cochrane Central Register of Controlled Trials (CENTRAL), Embase (OVID), and CINAHL (EBSCO). All databases will be evaluated from the date of creation. The same search strategies will be used with alterations as appropriate for each database interface. Details of the search strategy for PubMed are shown in Table 1. A combination of MeSH terms and keywords will be applied. We also searched grey literature in the ProQuest Dissertations & Theses databases and http://clinicaltrials.gov.
Finally, the reference lists of all relevant studies and review articles will be searched for additional studies not identified by electronic searches.
Table 1
Details of the search strategy for PubMed
Search term
|
1. (neoplasms [MeSH] OR neoplasm* [tiab] OR cancer*[tiab] OR carcino*[tiab] OR tumo*[tiab] OR malignanc*[tiab] OR oncolog*[tiab])
|
2. (phototherapy [MeSH] OR phototherap*[tiab] OR heliotherap*[tiab] OR “light treatment*” [tiab] OR “light therap*” [tiab] OR “light exposur*” [tiab] OR “artificial light” [tiab] OR “bright light” [tiab] OR sunlight [tiab])
|
3. (fatigue [MeSH] OR fatig*[tiab] OR tire*[tiab] OR lassitude[tiab] OR letharg*[tiab] OR astheni*[tiab])
|
Data collection and analysis
Selection of studies
All studies retrieved were sent into EndNote 8. We will use EndNote8 to eliminate duplicate studies. One member (XP) will initially screen study titles and abstracts and eliminate those that are unrelated to this review. Two members of the review team (DS and XJ) will then independently screen the remaining titles and abstracts for their eligibility for inclusion in accordance with the abovementioned criteria. Ineligible studies will be excluded at this stage. When the title and abstract do not provide all the information concerning the criteria, the full texts of the papers will be retrieved and screened. We will retrieve full texts of all studies if either the members who review the studies determine that the study possibly or definitely meets the inclusion criteria and need to record the reason for rejection after reading the full text. Any disagreements between the two reviewers will be resolved by discussion, with the involvement of a third member where agreement cannot be reached (ZY, LX, or DY). Multiple reports of the same study will be counted as a single study. The PRISMA template will be used to produce a flow chart (Figure 1) showing details of studies included and excluded at each stage of the study selection process.
Data extraction and management
Two members of the review team (XP and XJ) will independently extract data from the studies using a specifically designed data extraction form. The form will be piloted on a sample of three studies and then revised if required before full data extraction begins. Discrepancies will be resolved by discussion, with the involvement of a third member where necessary. Authors will be contacted to obtain any missing data. Findings will be reported regardless of their direction. Positive and negative findings must be clearly defined in the included studies. The following information will be extracted from the studies:
-
Bibliometric (author and year of publication).
-
Study characteristics (sample size, sample description, country, study design, study settings, the form of recruitment, recruitment time, and funding).
-
Participant characteristics (age, sex, treatment type, cancer stage, main physical symptoms that light therapy targets).
-
Intervention information for each arm of the study (light parameters, duration, the timing of the day of treatment, comparison/s).
-
Outcome (measurement methods used in the study; when the total score of the outcomes is different, the values will be converted to a scale of 0 to 100 points).
-
Timing of assessment for each outcome.
-
Results of the studies (mean and standard deviation or median and interquartile range and confidence interval).
Assessment of risk of bias in included studies
Criteria for features of the RCT design are based on those set out by the Cochrane Collaboration Risk of Bias Tool and will be considered for each of the included studies to assess the risk of bias. These criteria include the following: (1) random sequence generation, (2) allocation concealment, (3) performance bias, (4) detection bias, (5) attrition bias (bias due to the amount, nature, or handling of incomplete outcome data), (6) selective reporting bias (bias due to selective outcome reporting by comparing in-publication reporting of the outcomes of interest reported in the methods section to those reported in the results section), and (7) other sources of bias (e.g., bias due to baseline differences, the inappropriate influence of the study sponsor, and early stopping for the benefit) [42]. Each domain will be judged independently by two authors as having a high, low, or unclear risk of bias. Inconsistence will be resolved by discussion, with the involvement of a third reviewer where necessary.
Measures of treatment effect
We will use Review Manager (RevMan) 5.3 software for all analyses. For continuous data, we will report the mean differences between groups and the 95% confidence interval (95% CI). Where no standard deviations are reported, we will calculate the standard deviation using the methods described in the Cochrane Handbook for Systematic Reviews of Interventions [43]. Where the same outcome is measured using different measurement methods, we will calculate the standardized mean difference and the 95% CI for continuous data. For categorical data, we will report the odds ratios (ORs) between groups and the 95% CI. For arms with zero events, we impute a 0.5 event score.
Assessment of heterogeneity
Statistical heterogeneity will be tested using the standard chi-square tests (significance level: 0.1) and I2 statistics (0 to 40%: might not be important; 30 to 60%: may represent moderate heterogeneity; 50 to 90%: may represent substantial heterogeneity; 75 to 100%: considerable heterogeneity). Statistical heterogeneity will be regarded as substantial when the χ2 p-value is <0.1 or I2 is >50%. Meta-analyses will be conducted by using RevMan 5.3.
Assessment of publication biases
We will examine funnel plots corresponding to a meta-analysis of the primary outcomes to assess the potential for small-study effects such as publication bias if a sufficient number of studies (i.e., more than 10) are identified [44].
Data synthesis
Data will only be pooled if selected studies are sufficiently homogeneous in design and comparators. Otherwise, a narrative synthesis of the data will be conducted.
Continuous data will be combined only where (i) means and standard deviations are available or calculable and (ii) there is no clear evidence of skew in the distribution (using methods described in the Cochrane Handbook for Systematic Reviews of Interventions (2011)).
If it is possible to combine mean differences of scales measuring the same clinical outcomes in different ways, they will be standardized to combine results across scales (otherwise, weighted mean differences will be used).
We anticipate that there will be clinical heterogeneity driven by differences in light treatment parameters, study duration, and so on. Therefore, we will use a random effects meta-analysis model to produce an overall summary of the average treatment effect across the included trials.
Subgroup analysis and investigation of heterogeneity
We hypothesize that each of the factors below has the potential to have a clinically meaningful effect on the response to light therapy among fatigued cancer survivors. Therefore, if sufficient data are available, we will undertake a subgroup analysis based on the following:
Most studies have shown that white light above 2,000 lux has therapeutic effects and few effects below 500 lux [45]. Therefore, this review will compare light therapies with high intensity and those with low intensity.
-
Light treatment duration (≥ 4 weeks versus <4 weeks).
-
Low-light versus nonlight control conditions.
-
Intervention for specific cancer type only versus intervention for any cancer type.
-
Cancer population (pediatric or adolescent and young adult versus adult).
-
Fatigue as the inclusion standard of the study versus other physical symptoms as the inclusion standard of the study.
Sensitivity analysis
Trial quality
To evaluate the reliability and stability of the meta-analysis outcomes, we will conduct a sensitivity analysis based on trial quality, whereby studies of high or unclear risk of bias across different domains (see the section on “Assessment of risk of bias in included studies”) will be excluded to assess for any substantive difference to the overall effect estimates. If no substantive difference exists, the studies will be left in for the main analysis. This sensitivity analysis will be conducted for the fatigue outcomes only.
Outcome validity
Given that there is no accepted definition of CRF and no agreement on how it should be measured [46], a sensitivity analysis will be conducted based on outcome measurement. Scales may vary in the quality of psychometric properties. Studies that confirm that the scales are validated measures of fatigue will be compared to those that do not meet the following a priori criteria based on criteria outlined by Minton and Stone [46].
The paper that used the scale should have referred to at least three of the following: internal consistency, test-retest reliability, known group validity (discriminant validity), responsiveness to change, or convergent validity (against other scales). The original scale reference will be accessed if this information is not provided. The original paper will also be cross-referenced for citing articles to evaluate the frequency of scale use and the type of populations studied.
Additionally, single-parameter fatigue subscales as part of a broader quality-of-life measure will be compared to multiple-parameter scales that were specifically designed to assess fatigue.