Search

Background: Physical activity (PA) and discretionary screen time (DST; television and computer use during leisure) are both associated with obesity risk, but little longitudinal evidence exists on their combined influence. This study examined the independent and joint associations of changes in PA and DST with incident obesity, body mass index (BMI) and waist circumference (WC). Methods: We analysed the data of individuals aged 40–69 years from the UK Biobank, a large-scale, population-based prospective cohort study. PA was measured using the International Physical Activity Questionnaire and DST was defined as the total of daily TV viewing and non-occupational computer use. Changes in PA and DST over time were defined using departure from sex-specific baseline tertiles and categorised as worsened (PA decreased/DST increased), maintained, and improved (PA increased/DST decreased). We then used each exposure change to define a joint PA-DST change variable with nine mutually exclusive groups. We used multivariable adjusted mixed-effects linear and Poisson models to examine the independent and joint associations between PA and DST changes with BMI and WC and incident obesity, respectively. Development of a BMI ≥ 30 kg/m2 was defined as incident obesity. Results: Among 30,735 participants, 1,628 (5.3%) developed incident obesity over a mean follow-up of 6.9 (2.2) years. In the independent association analyses, improving PA (Incident Rate Ratio (IRR) 0.46 (0.38–0.56)) was associated with a lower risk of incident obesity than maintaining PA, maintaining DST, or improving DST. Compared to the referent group (both PA and DST worsened), all other combinations of PA and DST changes were associated with lower incident obesity risk in the joint association analyses. We observed substantial beneficial associations in the improved PA groups, regardless of DST change [e.g., DST worsened (IRR 0.31 (0.21–0.44)), maintained (IRR 0.34 (0.25–0.46)), or improved (IRR 0.35 (0.22–0.56)]. The most pronounced decline in BMI and WC was observed when PA was maintained or improved and DST was maintained. Conclusion: We found that improved PA had the most pronounced beneficial associations with incident obesity, irrespective of DST changes. Improvements in PA or DST mutually attenuated the deleterious effects of the other behaviour’s deterioration.

Objective there may be age-related differences in the impact of weight change on health. This study systematically reviewed the evidence on the relationship between weight change and all-cause mortality in adults aged 65 years and older. Methods MEDLINE, EMBASE and CINAHL were searched from inception to 11 June 2020, PROSPERO CRD 42019142268. We included observational studies reporting on the association between weight change and all-cause mortality in older community-dwelling adults. A random-effects meta-analysis was performed to calculate pooled hazard ratios and scored based on the Agency for Healthcare Research and Quality guidelines. Results a total of 30 studies, including 1,219,279 participants with 69,255 deaths, demonstrated that weight loss was associated with a 59% increase in mortality risk (hazard ratio (HR): 1.59; 95% confidence interval (CI): 1.45–1.74; P 0.05). Conclusion for community-dwelling older adults, weight changes (weight loss, gain or weight fluctuation) are associated with an increased risk of all-cause mortality risk relative to stable weight. Further research is needed to determine whether these associations vary depending upon initial weight, and whether or not the weight loss/gain was intentional.

Background Type 2 diabetes mellitus (T2DM) is emerging as a major public health challenge in Nepal. Behavioural, social and economic changes are likely to play a part in the rise of this chronic disease, as they are in many developing countries. A better understanding of the relationship between physical activity (PA), social and economic factors and T2DM can inform the design of prevention programs. Methods This study analysed data from 1977 Nepalese adults aged 40-69 years from the 2013 WHO STEPS survey. Socioeconomic position was determined based on education, occupation and ethnic group. Participants’ fasting blood glucose level was used to ascertain their diabetes status. Structural equation modelling was conducted to identify the regression pathways between PA, socioeconomic position, anthropometric and metabolic variables and T2DM. Results The overall prevalence of T2DM was 9% with almost double the burden among urban residents (17 %). Socioeconomic position (0.338), waist circumference (0.250), triglycerides (0.115) and hypertension (0.129) had a statistically significant direct effect on T2DM. PA had an indirect effect on T2DM mediated by waist circumference (-0.015). The indirect effects of socioeconomic position on T2DM were mediated by body mass index, waist circumference, triglycerides and total cholesterol. Conclusion Policies and programs to address T2DM in Nepal should address the factors contributing to unhealthy weight status, particularly among those of higher socioeconomic status. Key message Among Nepalese adults, higher socioeconomic position has a significant direct effect on T2DM, while both PA and higher socioeconomic position have significant indirect effects.

Background: With the economic and social change, participation in occupational and transport-related physical activity is declining among Nepalese adults, highlighting the growing importance of leisure-time physical activity (LTPA). Purpose: To qualitatively explore the socioecological influences upon participation in LTPA among Nepalese adults aged 40 years and above. Methods: A total of 51 adults (30 females and 21 males) participated in one of nine focus groups conducted in Kathmandu, Nepal. Data collection was guided by the social-ecological model of physical activity and analysis was done using the reflexive thematic analysis approach in NVivo 12. Results: Leisure time was mostly spent resting, socialising, or engaging in sedentary activities such as watching television. Individual-level barriers to LTPA participation included lack of knowledge, perceived lack of skill, lack of motivation, considering oneself as sufficiently active and engagement in sedentary screen activities. Family and household responsibilities, lack of support and fear of being judged by community members constituted the interpersonal barriers, while environmental barriers to LTPA included an absence of a supportive social norm, lack of open spaces, weather conditions and perceived lack of safety. Health benefits, prioritising physical activity, social support, provision of group-based activities and age-appropriate public exercise facilities were identified as major facilitators. Conclusion: Critical issues that need to be addressed to increase LTPA among Nepalese adults include traditional gender roles, family and social support, and social norms. Modifications of the built environment, such as public exercise facilities, offer further opportunities and will require coordination beyond the health sector., The Health & Fitness Journal of Canada, Vol. 14 No. 3 (2021): Proceedings from the 8th International Society for Physical Activity and Health Congress

Background and aims Associations between overweight/obesity and television viewing have been identified in high-income countries, but little evidence is available from low- and middle-income countries. This study examined the trends and correlates of television viewing and overweight/obesity among Nepalese women between 2006 and 2016. Methods and Results We analysed the data from 22,161 women aged 15-49 years who took part in Nepal Demographic Health Surveys 2006 (n=10,115), 2011 (n=5,881) or 2016 (n=6,165). Trained staff collected data on TV viewing and sociodemographic attributes from a face-to-face survey while height and weight were measured objectively to determine body mass index. Around 38% of the study participants watched TV at least once a week in 2006, which increased to 49% in 2011 and 46% in 2016. The prevalence of overweight and obesity increased from 8% in 2006 to 14.6% in 2011 and 19.8% in 2016. Compared to those who did not watch TV at all, those watching it at least once a week were at 1.54 (95% CI: 1.02-2.33), 1.79 (95% CI: 1.23-2.60) and 1.46 (95% CI: 1.13-1.88) times higher odds of being overweight/obese in 2006, 2011 and 2016 respectively. Conclusion The prevalence of TV viewing rose among women in Nepal between 2006 and 2016 and was associated with overweight/obesity, which also increased dramatically over this period. Future studies examining the use of multiple screen devices, daily usage duration and content viewed are recommended to understand better the health impacts of transitions to more sedentary living in Nepal and similar settings.

Aim: To analyze the data from the World Health Organization Nepal STEPS survey 2013 to determine the prevalence of total and domain-specific physical activity (PA) and associated factors among Nepalese adults. Methods: A multistage cluster sampling technique was used to proportionately select participants from the 3 ecological zones (Mountain, Hill, and Terai) in Nepal. The Global PA Questionnaire was used to assess PA. The data were analyzed using quantile and ordinary least square regression. Results: Only 4% of the adults did not meet the World Health Organization PA guidelines. Age had a negative monotonic association with total PA and occupational PA, with the highest difference at the upper tails of the PA distribution. Lower total PA and occupational PA were associated with secondary or higher education, being retired or in unpaid employment, living in Terai or urban areas, and nonsmoking. Age, higher education, unpaid employment, and Terai or urban residence were negatively associated, while being currently married was positively associated with transport-related PA. Conclusion: Increasing age, higher education, unpaid employment, unemployment or retirement, and urban residence were associated with lower PA, with the stronger association at the upper tails of the distribution. The correlates had dissimilar associations across the quantiles of PA distribution.

Background Chronic diseases contribute to about half of the adult disease burden in the South Asian region. Meanwhile, physical activity levels are declining despite the global evidence of its role in the prevention of chronic diseases. While there are a plethora of systematic reviews on the effects of physical activity on chronic diseases, there has not yet been a synthesis of the evidence concerning the nature of this relationship among people living in South Asian countries incorporating multiple chronic diseases and a focus on physical activity domains. The aim of this protocol is to describe the rationale and methods for a systematic review of published research to identify the association between physical activity and selected chronic diseases and their markers and analysis of the strength of association with a focus on physical activity domains among South Asian adults 40 years and older. Methods Nine electronic databases including Medline, PsycINFO, Embase, CENTRAL, CINAHL Plus, AgeLine, SPORTDiscus, Scopus and Web of Science will be systematically searched for papers reporting the association between physical activity and selected chronic diseases (type 2 diabetes mellitus, breast cancer, colorectal cancer, coronary heart disease, stroke, vascular diseases and musculoskeletal diseases (osteoarthritis, osteoporosis, back and neck pain)) and their markers using predefined search terms. Searches will be limited to peer-reviewed, English language papers with a quantitative design. In addition, a manual search of references of relevant systematic reviews as well as citations and references of eligible studies will also be carried out. The methodological appraisal will be performed using the National Institutes of Health quality assessment checklist for observational studies and the Effective Public Health Practice Project quality assessment tool for intervention studies. The overall quality of evidence for the study outcomes across the study designs will be assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework. The review results will be presented in the form of narrative synthesis, and a random effects meta-analysis is planned depending on the nature of included studies and available data. Discussion This review will summarise the strength of the association between physical activity and selected chronic diseases and their markers among South Asian adults 40 years or older. The findings will provide an evidence base to guide public health policy and interventions in the South Asian region and to inform future research to address the rising burden of chronic diseases. Systematic review registration PROSPERO CRD42018096505 Electronic supplementary material The online version of this article (10.1186/s13643-018-0848-9) contains supplementary material, which is available to authorized users.

Background With an increasing number of institutional deliveries, the Nepalese health system faces a challenge to ensure a quality of service provision. This paper aims to identify the determinants of client satisfaction with maternity care in Nepal using data from a nationally representative health facility survey. Methods A total of 447 exit interviews, with women who had either recently delivered or who had experienced obstetric complications, were conducted across 13 districts in Nepal (87% in hospitals, 8% in Primary Health Care Centres (PHCCs), and 5% in Sub/Health Posts(S/HPs). Client satisfaction was measured using an eight item scale that covered accessibility, interpersonal communication, physical environment, technical aspect of care and decision making. A client satisfaction index was computed using ordinal principal component analysis. A multivariate probit model was used to assess the net effect of explanatory variables on client satisfaction. Results Longer waiting times and overcrowding increased the likelihood of dissatisfaction. Having an opportunity to ask questions was positively associated with client satisfaction. Respondents from hill districts and rural areas were more likely to be satisfied in comparison to respondents from mountain, terai and urban areas. Socio-demographic factors (age, parity, caste/ethnicity, education, and ecological zone) and supply side factors (the time taken to reach a facility, type of facility, payment for services, and unknown heath worker or anyone entering the delivery room) were not statistically associated with satisfaction. Conclusions The findings suggest client satisfaction with the quality of maternity services in Nepal could be improved by reducing waiting times and overcrowding, and giving the mothers adequate time to ask questions. If clients are more satisfied they are more likely to use the facility again/recommend to a friend.

Background and aim Type 2 diabetes mellitus (T2DM) is emerging as a significant public health challenge in Nepal. Behavioural, social and economic changes are likely to play a part in the rise of this chronic disease, as they are in many developing countries. A better understanding of the relationship between physical activity (PA), socioeconomic factors and T2DM can inform the design of prevention programs. This study aimed to identify the path relationships between PA, socioeconomic position, anthropometric and metabolic variables and T2DM. Methods and results This study analysed data from 1977 Nepalese adults aged 40–69 years from the cross-sectional WHO STEPS survey undertaken in 2013. The latent variable “PA” was created using the information on domains of PA while the latent variable “socioeconomic position” was created using the variables education, occupation and ethnicity. Participants’ fasting blood glucose was used to determine their diabetes status. Structural equation modelling was conducted, and correlations and adjusted regression coefficients are reported. Individuals with higher education, in paid employment and from advantaged ethnic groups were more likely to have T2DM. Waist circumference, triglycerides and hypertension were found to have a statistically significant positive direct effect on T2DM. PA had indirect effects on T2DM, mediated by waist circumference. The indirect effects of socioeconomic position on T2DM were mediated by body mass index, waist circumference, triglycerides and total cholesterol. Conclusion Among Nepalese adults, higher socioeconomic position had a significant direct effect on T2DM, while both PA and higher socioeconomic position had significant indirect effects. Policies and programs to address T2DM in Nepal should address the factors contributing to unhealthy weight status, particularly among those of higher socioeconomic status.

South Asia specific reviews on the role of physical activity (PA) domains on chronic disease prevention are lacking. This study aimed to systematically review published literature to identify the association between PA domains and chronic diseases and to provide summary estimates of the strength of association. Nine electronic databases were searched using the predefined inclusion criteria which included population (South Asian adults 40 years or older), exposure (PA or sedentary behaviour) and outcome (type 2 diabetes mellitus, breast cancer, colorectal cancer, coronary heart disease, stroke, vascular disease and musculoskeletal diseases and their markers). A random-effects meta-analysis was carried out for cardiometabolic outcomes whereas narrative synthesis was completed for other outcome variables. Inactive or less active South Asian adults were at 31% higher risk of being hypertensive. Likewise, the risk of cardiometabolic outcomes was 1.34 times higher among inactive adults. Household PA was found to have a protective effect on breast cancer risk. Total and leisure time PA had a protective effect on osteoporosis among males and females respectively. Contemporary studies with a longitudinal design, representative samples, valid and reliable assessment of different domains are needed to establish the role of PA in chronic disease prevention in the region.

ObjectiveTo determine the prevalence and sociodemographic factors associated with active commuting and leisure-time physical activity (LTPA) among adults in western Nepal.DesignCross-sectional study.SettingAdults from semiurban areas in western Nepal.Participants2815 adults aged 25–65 years who participated in the ‘Community-Based Management of Hypertension in Nepal (COBIN)’ Study. Multiple logistic regression analysis was used to identify the sociodemographic factors associated with active commuting and LTPA.OutcomeSelf-reported participation in active commuting and LTPA.ResultsMost study participants (96%) commuted actively (walked or cycled) from one place to another. Our results showed that only a small proportion (3.7%) of participants engaged in moderate or vigorous LTPA. Compared with those in paid employment, the odds of commuting actively were higher among people working in agriculture or as labourers (OR: 4.57, 95% CI: 2.46 to 8.48), those retired/unemployed (OR: 2.98, 95% CI: 1.42 to 6.25) and those in unpaid employment (OR: 1.85, 95% CI: 1.06 to 3.22). Adults who were overweight or had obesity were less likely to commute actively. Compared with adults aged 25–34 years, older adults were less likely (OR: 0.35, 95% CI: 0.17 to 0.72) to engage in LTPA. Women were 0.46 times less likely to engage in LTPA compared with men.ConclusionMost adults engaged in active commuting for work or travel. Less than 5% participated in any form of moderate or vigorous LTPA. Longitudinal studies incorporating objective assessment of physical activity and a range of individual, interpersonal, and environmental factors will help understand how to promote active commuting and LTPA among Nepalese adults.Trial registration number: ClinicalTrials.govNCT02428075.

Background:This study was carried out to assess physical activity level and identify associated factors among higher secondary school students in Banke district, Nepal.Methods:A school-based, cross-sectional descriptive study was conducted among 405 students studying in grades 11 and 12 in 7 higher secondary schools selected randomly. A self-administered questionnaire based on the International Physical Activity Questionnaire was used to measure physical activity level.Results:Only 5% of students were found to be inactive, and domestic and transport-related activities were major contributors to total physical activity score. No significant difference existed for total physical activity and domain-specific and activity-specific scores across different age groups when males and females were tested separately. Being male (P = .046), lower economic status (P = .026), living at a distance of less than 30 minutes (P = .007), walking/cycling to school (P < .001), and studying in government school (P < .001) were associated with increased physical activity scores on multivariate logistic regression analysis.Conclusion:Routine activities such as transport and household chores contributed greatly to total physical activity among students. This study highlights the need for physical activity promotion interventions at school addressing the associated factors and a need for greater focus on leisure-time physical activities.

ObjectiveThis study is a systematic review of the peer-reviewed literature to identify the correlates of mobile screen media use among children aged 8 years and less.SettingHome or community-based studies were included in this review while child care or school-based studies were excluded.ParticipantsChildren aged 8 years or less were the study population. Studies that included larger age groups without subgroup analysis specific to the 0–8 years category were excluded. Eight electronic databases were searched for peer-reviewed English language primary research articles published or in press between January 2009 and March 2017 that have studied correlates of mobile screen media use in this age group.Outcome measureMobile screen media use was the primary outcome measure. Mobile screen media use refers to children’s use of mobile screens, such as mobile phones, electronic tablets, handheld computers or personal digital assistants.ResultsThirteen studies meeting the inclusion criteria were identified of which a total of 36 correlates were examined. Older children, children better skilled in using mobile screen media devices, those having greater access to such devices at home and whose parents had high mobile screen media use were more likely to have higher use of mobile screen media devices. No association existed with parent’s age, sex and education.ConclusionLimited research has been undertaken into young children’s mobile screen media use and most of the variables have been studied too infrequently for robust conclusions to be reached. Future studies with objective assessment of mobile screen media use and frequent examination of the potential correlates across multiple studies and settings are recommended.Trial registration numberThis review is registered with PROSPERO International Prospective Register of Ongoing Systematic Reviews (registration number:CRD42015028028).

Background Youth population (10-24 years of age) comprises more than 30% of the total population in Nepal. This is the age when values are formed, many become sexually active and begin to develop a pattern of risky sexual behaviour. This makes youth more vulnerable to contacting sexually transmitted diseases including HIV/AIDS. The main objective of this study was to assess the sexual behavior among school youths of a rural far-western district, Achham, Nepal. Methods A cross sectional study was conducted among school youths aged 15 to 24 years studying at grade 11 and 12 in Achham district. Data were collected from February to April, 2011 using self-administered questionnaires from 229 boys and 156 girls of randomly selected eight higher secondary schools of the district. Results The study found that over a quarter of the respondents were sexually active before marriage, boys five times more compared to girls. More than one-fourth (28.2%) of sexually active young boys and girls were found to have engaged in multiple sexual relations; of which peers were the most common (71.1%) sexual partners. Overall, three out of every four sexually active school youths were found to have practised risky sexual behavior. The study found age of the respondent, gender, education and age at first sexual intercourse were significantly associated with the sexual behavior of the school youths. Conclusions Young people at schools were found to be engaged in unsafe sexual activities in study district. They were found initiating multiple sexual relations and pre-marital sexual relations. The use of condom was found declining in subsequent sexual contacts. Such risky sexual behaviours were putting the young population at more vulnerable to the risks of any sexually transmitted infections. Sex education would therefore be imperative to focus on sexual and reproductive health interventions targeted for the young people. DOI: http://dx.doi.org/10.3126/hprospect.v12i2.9873 Health Prospect Vol.12(2) 2013: 37-41

Background Infant mortality is one of the priority public health issues in developing countries like Nepal. The infant mortality rate (IMR) was 48 and 46 per 1000 live births for the year 2006 and 2011, respectively, a slight reduction during the 5 years’ period. A comprehensive analysis that has identified and compared key factors associated with infant mortality is limited in Nepal, and, therefore, this study aims to fill the gap. Methods Datasets from Nepal Demographic and Health Surveys (NDHS) 2006 and 2011 were used to identify and compare the major factors associated with infant mortality. Both surveys used multistage stratified cluster sampling techniques. A total of 8707 and 10,826 households were interviewed in 2006 and 2011, with more than 99% response rate in both studies. The survival information of singleton live-born infants born 5 years preceding the two surveys were extracted from the ‘childbirth’ dataset. Multiple logistic regression analysis using a hierarchical modelling approach with the backward elimination method was conducted. Complex Samples Analysis was used to adjust for unequal selection probability due to the multistage stratified cluster-sampling procedure used in both NDHS. Results Based on NDHS 2006, ecological region, succeeding birth interval, breastfeeding status and type of delivery assistance were found to be significant predictors of infant mortality. Infants born in hilly region (AOR = 0.43, p = 0.013) and with professional assistance (AOR = 0.27, p = 0.039) had a lower risk of mortality. On the other hand, infants with succeeding birth interval less than 24 months (AOR = 6.66, p = 0.001) and those who were never breastfed (AOR = 1.62, p = 0.044) had a higher risk of mortality. Based on NDHS 2011, birth interval (preceding and succeeding) and baby’s size at birth were identified to be significantly associated with infant mortality. Infants born with preceding birth interval (AOR = 1.94, p = 0.022) or succeeding birth interval (AOR = 3.22, p = 0.002) shorter than 24 months had higher odds of mortality while those born with a very large or larger than average size had significantly lowered odds (AOR = 0.17, p = 0.008) of mortality. Conclusion IMR and associated risk factors differ between NDHS 2006 and 2011 except ‘succeeding birth interval’ which attained significant status in the both study periods. This study identified the ecological region, birth interval, delivery assistant type, baby’s birth size and breastfeeding status as significant predictors of infant mortality.

Background: Leisure-time physical activity is essential for healthy and physically active life; however, this domain of physical activity is less common in developing countries. Information on leisure time physical activity and sedentary behaviour among Nepalese population is not available. The study was carried out to assess leisure time physical activity and sedentary behaviour among high school adolescents and identify the associated factors in Nepal. Methods: A cross-sectional descriptive study was carried out in Banke district, Nepal in 2013 among higher secondary school students using self-administered questionnaire based on International Physical Activity Questionnaire. A sample of 405 students, 178 females and 227 males, of the age–group 15 to 20 years from seven schools were included in the study. Multivariate logistic regression analysis was carried out to identify factors associated with participation in leisure time physical activity and sedentary behaviour. Results: Engagement of female in leisure time physical activity was lower but mean time spent on sitting per day was higher. Students who walked to school and have playground/parks near home, younger females (OR = 3.09, 95% CI: 1.18-8.08), females living in nuclear families (OR: 2.16, 95% CI: 1.01-4.62) and males who cycled to school (OR: 8.09, 95% CI: 2.35-27.80) and have provision of extra-curricular activities (OR: 2.49, 95% CI: 1.04-5.97) were more likely to be engaged in leisure time physical activity. On the other hand, students who did not have playground in school and lived in rural areas were more likely to sit for more than 6 hours a day. Likewise, male students of private school (OR: 6.41, 95% CI: 2.89-14.21), who used vehicle to reach school (OR: 5.90, 95% CI: 1.26-27.75) and have no provision of extra-curricular activities (OR: 2.98, 95% CI: 1.09-8.07) had longer sitting time. Conclusion: Difference in leisure time physical activity and sedentary behaviour was found among male and female school adolescents. Interventions are needed not only to promote leisure time physical activity but also to reduce sedentary behaviour among this group.

Background Childhood is a crucial period for shaping healthy behaviours; however, it currently appears to be dominated by screen time. A large proportion of young children do not adhere to the screen time recommendations, with the use of mobile screen devices becoming more common than fixed screens. Existing systematic reviews on correlates of screen time have focused largely on the traditional fixed screen devices such as television. Reviews specially focused on mobile screen media are almost non-existent. This paper describes the protocol for conducting a systematic review of papers published between 2009 and 2015 to identify the correlates of mobile screen media use among children aged 0–8 years. Methods A systematic literature search of electronic databases will be carried out using different combinations of keywords for papers published in English between January 2009 and December 2015. Additionally, a manual search of reference lists and citations will also be conducted. Papers that have examined correlates of screen time among children aged 0–8 will be included in the review. Studies must include at least one type of mobile screen media (mobile phones, electronic tablets or handheld computers) to be eligible for inclusion. This study will identify correlates of mobile screen-viewing among children in five categories: (i) child biological and demographic correlates, (ii) behavioural correlates, (iii) family biological and demographic correlates, (iv) family structure-related correlates and (v) socio-cultural and environmental correlates. PRISMA statement will be used for ensuring transparency and scientific reporting of the results. Discussion This study will identify the correlates associated with increased mobile screen media use among young children through the systematic review of published peer-reviewed papers. This will contribute to addressing the knowledge gap in this area. The results will provide an evidence base to better understand correlates of mobile screen media use and potentially inform the development of recommendations to reduce screen time among those aged 0–8 years. Systematic review registration PROSPERO CRD42015028028. Electronic supplementary material The online version of this article (doi:10.1186/s13643-016-0272-y) contains supplementary material, which is available to authorized users.