Table 5 Modifiers of change

Forsyth et al., 2009 [64]

Gender, Age, BMI

The authors report which sub-categories have been collected, without providing any further details.

Mailey et al., 2010 [25]

Gender, Age, Ethnic group

The data have been used to describe the sample of participants.

Oeland et al., 2010 [31]

Gender, Age, BMI, VO₂ max

The authors report which subcategories have been collected. They excluded from the study those who had a BMI > 35.

Pentecost et al., 2015 [32]

Gender, Age BMI, Ethnic group, relationship status, smoking status, postcode, number of dependents and age upon leaving full-time education

Usable descriptive data were reported for 28 (47%) participants. Only 11 participants (37%) at baseline and 9 (30%) at the 4-month follow up provided data for BMI and BP.

Piette et al., 2011 [26]

Gender, Age, Ethnic Group, relationship status, Education,

Employment Status; annual household income, BMI, Diabetes Medication,

Antidepressant medication

16% of participants were ethnic minorities, however, no other information about this could have affected the uptake of PA was provided.

The authors identified that there were differences between those who provided follow-up data at the 12-month follow up; they had higher income.

Suija et al., 2009 [27]

Gender, Age, BMI, Physical Activity level, Smoking status, antidepressant medication

The baseline characteristics of participants have been reported. The authors haven’t discussed, however, how these sub-categories could have affected the uptake of PA.

Crone et al., 2008 [33]

Gender, Age

This was the UK’s PARS study.

Women made up the majority (64%) of patients referred to scheme due to mental health. The average age of “mental health participants” was significantly lower than “physical health participants” (42 ± 14 year versus 51 ± 14 years; p < 0.0001).

Fewer patients with mental health problems (60%) took up referral at the local leisure centre, compared to those with poor physical health (69%).

The authors refer to their previous studies related to the provision of PA for patients with mental health problems; where one financial constraint (sub-category – household income), was a reason for mental health patients dropping out from PARS. p. 1093.

Duda et al., 2014 [28]

Gender, Age, Ethnic Group, Qualifications, alcohol intake

The sub-categories are reported in the article, but how they might have affected the uptake of PA isn’t.

The authors state: “The city in which the trial took place has a relatively young, ethnically diverse population, with about third of the people non-white [32] and 16.5% born outside the UK at the 2001 census.” P. 4

They also add report that the recruitment to the study was challenging due to the ethnic diversity of the sample, resulting in difficulties in administering the study questionnaire to people who do not speak English.

Littlecott et al., 2014 [29]

Gender, Age, Level of Deprivation, Baseline Activity level

The baseline characteristics of participants have been reported. The authors haven’t discussed, however, how these sub-categories could have affected the uptake of PA.

This was the UK’s study that identified effects of PA for patients with CHD risk only, mediation analyses were limited to this subsample.

Pomp et al., 2013 [30]

Gender, Age, Marital Status, Educational Background, occupational status

The differences between groups (participants in the intervention and control arms) at T1 were found of physical activity and educational background. The participants did not differ with regard to sex, age, and occupational status.

The results revealed that those who continued to participate in the study were younger than those who dropped out. They did not differ, however, in terms of gender, occupational status, high school degree, partner status (between T1 and T3).