Workplace interventions for preventing work disability
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
The first objective of this systematic review is to determine whether workplace interventions to prevent long‐term work disability, for employees who are already sick‐listed, are more effective than no intervention, or alternative clinical or clinic‐based interventions. Thus, this review will focus on secondary prevention of work disability and not on the initial prevention of a specific disorder. The second objective is to determine whether there are differences between the effectiveness of these interventions for musculoskeletal problems, mental problems and other problems.
Background
Work disability is a major public health problem in western industrialized countries, and is a considerable economic burden on society (Dorman 2000; Henderson 2005; OECD 2003). In the UK the total costs of work disability are estimated at 24 billion UK pounds a year (Department of Health 2004). Even though long‐term sick leave accounts for only a small fraction of sick leave taken, longer periods of absence from work account for more than a third of the total days lost and up to 75% of the costs of absence (Department of Health 2004). Work disability might have serious consequences for individuals, for example a poorer quality of life, loss of social identity, and even indefinite or permanent exclusion from work. Although work disability can be viewed as the inability of an employee to perform his or her regular work activities due to a medical condition (this means that it includes both sick leave and work limitations), the focus of this review will be on work disability where sick leave is involved.
On the one hand, return to work (RTW) is influenced by recovery from a medical condition itself, and on the other hand by various psychosocial and occupational factors that enhance the symptoms or complaints and in turn, make RTW a difficult and fearful issue (Steenstra 2005; Sullivan 2005; WHO 2001). It has been suggested that even though an employee may continue to experience symptoms, work resumption might be possible and does not necessarily contribute to unfavourable effects on these symptoms. This seems to be true not only for physical symptoms such as back pain (Anema 2007; Loisel 1997; Staal 2004), but also for disability due to mental health problems (Blonk 2006; van der Klink 2003). However, the initial RTW interventions focused on the treatment of medical conditions, expecting the alleviation of symptoms to facilitate RTW. Over the past decade, a change in paradigm has occurred from disease prevention and treatment to disability prevention and management (Loisel 2001). Especially for work‐related back pain the focus is now more on prevention of long‐term disability and prolonged absence from work, rather than treatment of the back pain itself. This change is accompanied by a lack of empirical support for a purely biomedical model of work disability and RTW (Schultz 2002; Schultz 2007). It has been suggested that a focus on RTW behaviour rather than on the medical condition of an employee is now needed in the development of interventions (Frank 1998; Shaw 2001; Waddell 2001).
Long‐term work disability is no longer seen simply as the consequence of an illness (or impairment), but rather as the result of interactions between the employee and three main systems: (1) the health care system, (2) the work environment system and (3) financial compensation systems (Franche 2002; Loisel 2001). Moreover, during the RTW process dimensions related to the work environment and the roles of the stakeholders have to be considered, regardless of the type of disorder (Baril 2003; Loisel 2001). Since a worker's disability is influenced by the stakeholders' actions and attitudes, and interactions between the stakeholders, a case management approach has been recommended. In this approach, disability management takes place in the workplace, and work adaptations will be discussed as part of the RTW process (Anema 2007; Schultz 2007; Young 2005).
Since some evidence indicates that return to work interventions should be carried out closely with the workplace and in collaboration with the key stakeholders (Franche 2005; Krause 1998), the proposed Cochrane systematic review is aimed at obtaining knowledge about the effectiveness of interventions directed at work, such as changes to the workplace or equipment, changes in work design and organisation, changes in working conditions or work environment, and/or occupational (case) management with active stakeholder involvement of (at least) the employee and the employer. These workplace interventions will be compared to no intervention, or alternative clinical or clinic‐based interventions. The proposed definition is a form of disability management in which the main objective of the interventions is the facilitation of return to work by removing the barriers for RTW, therefore these interventions will be frequently applied for sick‐listed employees.
Unlike other Cochrane reviews of RTW interventions, we will include both musculoskeletal disorders, mental health problems, and other health conditions such as cardiovascular diseases or cancer, which will allow us to draw conclusions about the effectiveness of RTW interventions for all conditions. Franche 2005 suggested that workplace interventions for low back pain should be adapted to develop suitable interventions that can be applied for other causes of disability. However, to what extent can the principles of disability management for musculoskeletal conditions be applied to other conditions, such as mental health problems or other health problems? A systematic review of the existing workplace interventions provides more insight into the need to develop and test these interventions. In the literature we found no systematic reviews that focused on disability due to musculoskeletal, mental health and other health problems.
Comparison with other Cochrane Systematic Reviews
In the field of occupational health, other Cochrane reviews dealing with the effectiveness of workplace interventions have been published (Aas 2005; Bruinvels 2007; Nieuwenhuijsen 2006). However, none of the study protocols included assessment of the effectiveness of workplace interventions to prevent work disability caused by both physical and mental health problems in sick‐listed employees. Whereas Aas 2005 assessed the effectiveness of workplace interventions in employees with back and neck pain, our study population is not restricted to any specific type of disorder. Furthermore, interventions that will be included in our proposal are employee‐specific. Moreover, in the studies Bruinvels 2007 and Nieuwenhuijsen 2006, the focus of the interventions was not limited to workplace interventions. In these studies a subgroup analysis of interventions aimed at the workplace was performed. Therefore this review will be unique in the field of the effectiveness of workplace interventions. By including both physical and mental health problems in our study design, we will be able to compare the effects of the treatments in these two groups.
Objectives
The first objective of this systematic review is to determine whether workplace interventions to prevent long‐term work disability, for employees who are already sick‐listed, are more effective than no intervention, or alternative clinical or clinic‐based interventions. Thus, this review will focus on secondary prevention of work disability and not on the initial prevention of a specific disorder. The second objective is to determine whether there are differences between the effectiveness of these interventions for musculoskeletal problems, mental problems and other problems.
Methods
Criteria for considering studies for this review
Types of studies
All randomised controlled trials (RCTs) concerning workplace interventions aimed at preventing work disability by means of job‐accommodation or the involvement of stakeholders will be included. The stakeholders involved in these interventions must include at least the employee and the employer.
Types of participants
The participants in the trials must be working adults who are on sick leave (full or part‐time) due to a medical condition. Regardless of the type of medical condition, employees with all types of work disability will be included. The diagnosis as provided in the studies will be registered and information about development or coexistence of secondary diagnoses will be collected.
Types of interventions
The Cochrane Occupational Health Field has classified workplace interventions as a type of intervention for disability management (Schonstein 2006). For this review, the term workplace intervention is proposed to be used for interventions focusing on changes in the workplace or equipment, changes in work design and organisation (including working relationships), changes in working conditions or work environment, and/or occupational (case) management with active stakeholder involvement of (at least) the employee and the employer (Anema 2004; Franche 2005a). Active involvement is defined as face‐to‐face conversations about RTW between (at least) the employee and the employer. This definition is a synthesis of the IEA definition (Stapleton 2000) and the Waddell et al. definition (Waddell 2001). Changes in the workplace and equipment are for instance changes in the furniture or the materials needed to carry out the work. Changes in work design and organisation concern scheduling, tasks (including training in task performance) and working relationships with supervisor and co‐workers. Changes in working conditions refer to the financial and contractual arrangements, and, finally, changes in work environment concern noise, illumination, vibration, etc. Our definition allows us to include interventions only if they are closely linked to the workplace by primarily focusing on work adaptations or the involvement of stakeholders from the work environment system. Interventions that are intended to recreate the demands of work in a clinical setting, without any change and/or involvement of the work system in the RTW process, will not be included in this review.
Studies will be excluded if the intervention is:
-
focused on primary prevention,
-
focused clinical purely,
-
group‐based rather than individual‐based,
-
focused on "just" education and advice about ergonomics,
-
aimed at body function modifications (for instance correct positions) only.
The interventions will also be excluded if the main goal of the intervention is not related to RTW. The definition of a workplace intervention is mainly inspired based on musculoskeletal disorders, however for mental health problems some examples of interventions according to this definition exist and these interventions comprise at least the preparation of return to work in contact with the workplace (Schene 2006) or advise about changes in work processes to facilitate RTW (Blonk 2006). Identified workplace interventions will be compared to no intervention (usual care), or alternative clinical or clinic‐based interventions. Interventions that include more components than needed exclusively for a workplace intervention will not be excluded, as long as the workplace intervention is a structural part of the intervention (i.e. intention to apply it to all participants in the intervention group).
Types of outcome measures
Trials that measured at least the primary outcomes of work absenteeism will be included. All studies that measure a dichotomous measure of work absenteeism only, will be excluded, because in these studies no information is available about the exact duration of work disability. The definitions of work absenteeism as given in the studies will be used for the analyses. However, when studies report the results on several work absenteeism definitions we will choose to analyse the data based on the first of the following definitions:
-
Time until lasting RTW: a period of work absence from the first day of sick leave to full return to work in own or equal work, for at least 4 weeks without dropout.
-
Time until first RTW: a period of work absence from work preceded and followed by a period of at least one day at work (de Vet 2002).
-
Cumulative duration of work absence: total days of sick leave during the follow‐up period (regardless if resulting from one of more absence spells).
-
Recurrences of work absence: The frequency and duration of recurrent episodes of sick leave.
Secondary outcomes are:
-
Functional status.
-
Quality of life.
-
Direct and indirect costs of work disability.
These outcomes are likely to be meaningful for employees who are on sick leave, their employers, their care‐providers (such as treating and occupational physicians), insurers, as well as the policymakers who are involved in the decisions.
Search methods for identification of studies
We will search the following databases:
-
EMBASE.com (Medline and Embase combined)
-
The Cochrane Central Register of Controlled Trials (CENTRAL)
-
PsycINFO
-
Database of the Cochrane Occupational Health Field
We will search for studies in EMBASE.com by combining the following:
1. RTW interventions
(vocational‐rehabilitation/exp OR occupational‐intervention OR disability‐prevention OR disability‐management OR 'disability'/de/dm_dm,dm_pc,dm_th OR 'work disability'/de/dm_dm,dm_pc,dm_th OR occupational‐rehabilitation/exp OR workplace‐intervention OR modified‐duty OR modified‐duties OR vocational‐guidance OR case‐manager OR case‐management OR ergonomics OR 'ergonomic *3 approach' OR 'ergonomic *3 training' OR 'ergonomic *3 education' OR 'ergonomic *3 counselling' OR job‐accommodation OR on‐the‐job‐program OR workplace‐accommodation OR modified‐work OR supported‐employment OR work‐reintegration‐plan OR light‐duty OR work‐site‐visit OR work‐visit OR work‐adjustment OR solution‐focused‐intervention OR 'vocational *3 counselling' OR 'vocational *3 placement' OR 'vocational *3 training' OR 'occupational disease'/exp/dm_dm,dm_th)
2. Methodological filter and exclusion of chemicals and drugs
(random*:ti,ab OR clinical‐trial OR clinical‐trials OR health‐care‐quality/exp) NOT ('chemicals and drugs'/exp/mj)
The methodological filter we will use is a best sensitive methodological filter for EMBASE.com, to identify a set of relevant RCTs that is as complete as possible (Wong 2006).
3. #1 AND #2
Outcome terms are important for this review. However, Emtree terms such as work‐disability and disease‐duration are very broad, and should be combined with population terms. On the other hand, specific terms for outcome are suitable to be incorporated without the population terms. Therefore we will search with two combinations.
4. Specific terms for outcomes
(absenteeism/exp OR (((worktime OR work‐time) OR workday*) AND (loss OR lost)) OR return‐to‐work OR returns‐to‐work OR sick‐leave OR work‐resumption/de OR sick‐absence OR sickness‐absence OR lost‐workdays OR sick‐listed OR work‐resumption OR duration‐of‐absence OR work‐reentry‐rate OR time‐loss‐from‐work OR time‐lost‐from‐work)
5. More general terms for outcomes, if used singly, are too broad, and we will therefore use them in combination with terms for population.
(absenteeism/exp OR (((worktime OR work‐time) OR workday*) AND (loss OR lost)) OR return‐to‐work OR returns‐to‐work OR sick‐leave OR (work AND limitation*) OR job‐performance/de OR work‐resumption/de OR sick‐absence OR sickness‐absence OR 'disease duration'/exp OR work‐disability/de OR work‐disability OR disability‐prevention OR disability/de OR disability‐management OR employment‐after‐rehabilitation OR (regain AND (employment OR work)) OR lost‐workdays OR (compensation AND cost*) OR work‐resumption OR duration‐of‐absence OR work‐reentry‐rate OR time‐loss‐from‐work OR time‐lost‐from‐work) AND (employee/exp OR employee* OR employer/exp OR employer* OR worker/exp OR worker* OR workman* OR work‐site OR worksite OR workman‐compensation/de OR workers‐compensation OR benefit‐duration OR time‐on‐benefits OR workplace/de OR workplace OR work‐environment/de OR supervisor*)
The two searches in EMBASE.com are:
#3 AND #4
#3 AND #5
All references from the EMBASE.com search will be combined in a Reference Manager database and further checked for double references by means of a duplicate search. Each double reference we find will be deleted.
We will search for studies in CENTRAL by combining the following areas:
1. Terms for population/place of application of intervention
(employee* OR employer* OR worker* OR manpower OR "work site" OR worksite OR "workman compensation" OR "workers' compensation" OR workplace OR "work environment" OR "work capacity" OR supervisor*)
2. Terms for outcome
(absenteeism OR ((worktime OR workday*) AND (loss OR lost)) OR "return to work" OR "returns to work" OR "sick leave" OR "job performance" OR "work resumption" OR "sick absence" OR "sickness absence" OR "disease duration" OR "work disability" OR "disability prevention" OR disability OR "disability management" OR "employment after rehabilitation" OR "regain employment" OR "regain work" OR "lost workdays" OR "duration of absence" OR "work reentry rate" OR "time loss from work" OR "time lost from work")
The search in CENTRAL is:
#1 AND #2
This search will be restricted to the Cochrane Central Register of Controlled Trials (Clinical Trials).
PsycINFO
We will search for studies in PsycINFO by combining the following areas:
1. Terms for intervention
(DE=("vocational rehabilitation" or "supported employment" or "vocational evaluation" or "work adjustment training" or "occupational adjustment" or "disability management" or "case management") or KW=("workplace intervention*" or "job accommodation*" or "workplace accommodation*" or "modified work" or "work site visit" or "ergonomic*" or "occupational intervention" or "disability prevention" or "occupational rehabilitation" or "workplace intervention" or "modified duty" or "light duty" or "modified duties" or "vocational guidance" or "case manager" or "on the job program" or "work reintegration plan" or "solution focused intervention" or "vocational counseling"))
2. Terms for outcome
((DE=("employee absenteeism" or "reemployment" or "employee leave benefits")) or KW=("return to work" or "returns to work" or "work disability" or "employment after rehabilitation" or "time loss from work" or "time lost from work" or "work rehabilitation" or "absenteeism" or "work resumption" or "sick leave" or "sick listed" or "sick absence*" or "sickness absence*" or "absenteeism" or "worktime loss" or "work time loss" or "workday loss" or "work resumption" or "lost workdays" or "duration of absence" or "work reentry rate" or "time loss from work" or "time lost from work"))
The search in PsycINFO is:
#1 AND #2.
The database of the Cochrane Occupational Health Field will be searched by combining:
1. A code for research design: RCT‐study (all non‐indexed fields).
2. A code for outcome: disability‐outcome (all non‐indexed fields).
Trials will also be identified by hand searching the reference lists of relevant review articles and eligible studies, and through personal contact with experts in the field of occupational health. We will consider articles in all languages and translate articles if necessary.
Data collection and analysis
The methods of this review will follow the Cochrane Handbook for Systematic Reviews of Interventions (Version 4.2.6) (Higgins 2006).
Selection of studies
The title and abstract (if available) of all identified studies will be stored in a new database in Reference Manager. After removing the double references, a bibliography will be generated, including the title, keywords and abstract of each reference found.
The study selection will be completed in two steps. In step 1, two authors (SHO and MTD) will screen the title, keywords and abstracts of all references retrieved by the literature search, to determine whether each article meets the inclusion criteria. The inclusion criteria are: study design is an RCT, participants are sick‐listed employees, the intervention investigated meets with the definition of a workplace intervention, and work absenteeism is measured. A standardized form with inclusion criteria will be developed for that purpose. In step 2 the full article will be retrieved for studies about which no decision could be made by screening in step 1, and these will be fully reviewed. A consensus procedure will be used to solve disagreements about the selection of RCTs, and a third author (JRA) will be consulted if the disagreement persists. The reason for exclusion and whether exclusion was based on the title, keywords and abstract only, or based on the full‐text, will be reported on a standardized form for each excluded study.
Data extraction and management
We will (SHO and MTD) independently extract the data onto a pre‐designed data extraction form. This form includes at least essential study information about the participants, interventions, outcome measures and results. A small sample of the articles will be used to test whether the form is feasible according to the two authors.
If there are any disagreements about the data extraction, consensus will be achieved by discussion among the authors. A third author (JRA) will be consulted if the disagreements persist. If certain articles do not contain sufficient information, the authors of the articles will be contacted.
Assessment of risk of bias in included studies
The methodological quality of the RCTs will be assessed independently by two authors (SHO and MTD). Study quality will be assessed by an adapted version of the checklist recommended by the Back Review Group (van Tulder 2003), which is in accordance with the guidelines contained in the Cochrane Reviewers Handbook (Higgins 2006). Blinding of intervention providers and participants was not used as a criterion because of the nature of these interventions. Criteria in the checklist include randomization, concealment of treatment allocation, differences between groups at baseline, blinding of outcome assessor, avoidance of co‐interventions, compliance, dropout rate, timing of outcome assessment and intention‐to‐treat analysis (Table 1). The criteria will be scored as positive, negative or unclear ("yes", "no" and "don't know").
| Criteria | Operationalization |
| A. Was the method of randomization adequate? | A random (unpredictable) assignment sequence. Examples of adequate methods are computer generated random number table and use of sealed opaque envelopes. Methods of allocation using date of birth, date of admission, hospital numbers, or alternation should not be regarded as appropriate. |
| B. Was the treatment allocation concealed? | Assignment generated by an independent person not responsible for determining the eligibility of the patients. This person has no information about the persons included in the trial and has no influence on the assignment sequence or on the decision about eligibility of the patient. |
| C. Were the groups similar at baseline regarding the most important prognostic factors? | In order to receive a 'yes', groups have to be similar at baseline regarding demographic factors, duration and severity of complaints, percentage of patients with neurological symptoms and value of main outcome measure(s). |
| D. Was the outcome assessor blinded to the intervention? | The reviewer determines if there was enough information about the blinding of the outcome assessor to score a 'yes'. |
| E. Were co‐interventions avoided or similar? | Co‐interventions should either be avoided in the trial design or comparable between the index and control groups. |
| F. Was the compliance acceptable in all groups? | The reviewer determines if the compliance to the interventions is acceptable, based on the reported intensity, duration, number, and frequency of sessions for both the index intervention(s) and control intervention(s). |
| G. Was the drop‐out rate described and acceptable? | The number of participants who were included in the study but did not complete the observation period or were not included in the analysis must be described and reasons given. If the percentage of withdrawals and drop‐outs does not exceed 20% for short‐term follow‐up and 30% for long‐term follow‐up and does not lead to substantial bias, a 'yes' is scored. |
| H. Was the timing of the outcome assessment in all groups comparable? | Timing of outcome assessment should be identical for all intervention groups and for all important outcome assessments. |
| I. Did the study include an intention to treat analysis? | All randomized patients are reported/analyzed in the group they were allocated to by randomization for the most important moments of effect measurement (minus missing values) irrespective of non‐compliance and co‐interventions. |
A consensus method will be used to solve disagreements, and a third author (JRA) will be consulted if the disagreements persist. The authors will pilot‐test the methodological quality assessment on three articles not included in the review (regarding another intervention) to ensure that the appraisal criteria are applied consistently. For articles that do not contain sufficient information on (one or more of) the methodological criteria (score "don't know"), the authors will be contacted for additional information. We will try to locate the authors' current working address through their most recent publication in PUBMED, through the Internet or through a request for information sent to the address listed on the article. If the authors cannot be contacted, or the information is no longer available, the criteria will be scored as "unclear".
Data synthesis
The authors will decide whether the studies are sufficiently (clinically) homogeneous with regard to types of interventions, types of control groups and types of outcomes, to perform a meta‐analysis. A quantitative and qualitative analysis could also be applied (van Tulder 2003). However, the nature of the research in this field is marked by highly heterogeneous studies with respect to interventions and outcomes. This review focuses on workplace interventions only, in order to minimise the heterogeneity of the interventions. If heterogeneity is present, qualitative or quantitative analyses will be considered.
Clinical heterogeneity can be investigated by performing subgroup analyses. This review will investigate subgroups of the type of disability: disability due to musculoskeletal, mental health or other health conditions.
Meta‐analysis
Work absenteeism is frequently presented as time to RTW, and in such cases survival analysis is often performed to calculate a hazard ratio. The event is defined as successful or unsuccessful RTW. Usually, Cox proportional hazard model is used to analyse time‐to‐event data. In this approach, participants who do not RTW during the entire follow‐up period are censored to be sure that the total follow‐up period is analysed as sick leave. When comparing treatments in a meta‐analysis, a simplifying assumption is often made that the hazard ratio is constant across the follow‐up period, even though the hazards themselves may vary continuously. This is known as the proportional hazards assumption (Higgins 2006). Otherwise, work absenteeism could be presented as continuous outcomes (cumulative duration of work absenteeism and recurrences). Outcomes such as functional status, quality of life, and costs can be presented as continuous, categorical or dichotomous outcomes.
If studies are considered to be clinically homogeneous, and sufficient data are available, the data will be pooled with RevMan 4.2.8 software. The results of each RCT will be plotted as a point estimate with corresponding 95% confidence interval (CI). We will identify statistical heterogeneity by quantifying heterogeneity (I2). For studies that are statistically heterogeneous, a random‐effects model will be used. As no random‐effects model is available in RevMan for time‐to‐event data, a fixed‐effect model will be used. For continuous data (such as work absenteeism, recurrences and costs), standardized mean differences (with a 95% CI) will be used as the summary effect measure, since these outcomes are likely to be measured with different scales in the varying studies. For dichotomous outcomes, risk ratios (RR, including a 95% CI) will be presented.
Qualitative analysis
Regardless of whether or not there are sufficient data available to apply quantitative analyses to summarize the data, we will assess the overall quality of the evidence for each outcome. To accomplish this, we will use an adapted GRADE approach (GRADE working group 2004), as recommended by the Cochrane Back Review Group (Furlan 2007). The quality of the evidence for a specific outcome is based on the study design, methodological quality (van Tulder 2003), consistency of results, directness (generalizability), precision (sufficient data) and reporting of the results across all studies that measure that particular outcome. The quality starts at high when high quality RCTs provide results for the outcome, and reduces by one level for each of the factors that are not met.
-
High quality evidence = there are consistent findings among at least two (high quality) RCTs with low potential for bias that are generalizable to the population in question. There are sufficient data, with narrow confidence intervals. There are no known or suspected reporting biases.
-
Moderate quality evidence = one of the factors is not met.
-
Low quality evidence = two of the factors are not met.
-
Very low quality evidence = three of the factors are not met.
-
Conflicting evidence = may be moderate or lower, depending on the number of factors met.
-
No evidence = no evidence from RCTs.
The outcome of the studies will be considered to be 'consistent' if at least 75% report statistically significant results in the same direction. For a sensitivity analysis, the results will be analysed again, including only high quality RCTs. The aim of the sensitivity analysis is to find out whether quality level leads to changes.
| Criteria | Operationalization |
| A. Was the method of randomization adequate? | A random (unpredictable) assignment sequence. Examples of adequate methods are computer generated random number table and use of sealed opaque envelopes. Methods of allocation using date of birth, date of admission, hospital numbers, or alternation should not be regarded as appropriate. |
| B. Was the treatment allocation concealed? | Assignment generated by an independent person not responsible for determining the eligibility of the patients. This person has no information about the persons included in the trial and has no influence on the assignment sequence or on the decision about eligibility of the patient. |
| C. Were the groups similar at baseline regarding the most important prognostic factors? | In order to receive a 'yes', groups have to be similar at baseline regarding demographic factors, duration and severity of complaints, percentage of patients with neurological symptoms and value of main outcome measure(s). |
| D. Was the outcome assessor blinded to the intervention? | The reviewer determines if there was enough information about the blinding of the outcome assessor to score a 'yes'. |
| E. Were co‐interventions avoided or similar? | Co‐interventions should either be avoided in the trial design or comparable between the index and control groups. |
| F. Was the compliance acceptable in all groups? | The reviewer determines if the compliance to the interventions is acceptable, based on the reported intensity, duration, number, and frequency of sessions for both the index intervention(s) and control intervention(s). |
| G. Was the drop‐out rate described and acceptable? | The number of participants who were included in the study but did not complete the observation period or were not included in the analysis must be described and reasons given. If the percentage of withdrawals and drop‐outs does not exceed 20% for short‐term follow‐up and 30% for long‐term follow‐up and does not lead to substantial bias, a 'yes' is scored. |
| H. Was the timing of the outcome assessment in all groups comparable? | Timing of outcome assessment should be identical for all intervention groups and for all important outcome assessments. |
| I. Did the study include an intention to treat analysis? | All randomized patients are reported/analyzed in the group they were allocated to by randomization for the most important moments of effect measurement (minus missing values) irrespective of non‐compliance and co‐interventions. |
