Article Text
PDF
Abstract
Background Exclusive breastfeeding (EBF) is a major public health problem in Ethiopia. However, the spatial variation of EBF and the associated factors have not been studied as much as we have searched. This study aimed at assessing geospatial variation and the predictors of EBF using geographically weighted regression.
Methods A cross-sectional study was conducted using the 2019 Mini-Ethiopian Demographic and Health Survey data set. The study used a total weighted sample of 548 infants. Hotspot spatial analysis showed the hotspot and cold spot areas of EBF. The spatial distribution of EBF was interpolated for the target population using spatial interpolation analysis. SaTScan V.9.6 software was used to detect significant clusters. Ordinary least squares regression analysis identified significant spatial predictors. In geographically weighted regression analysis, the effect of predictor variables on the spatial variation of EBF was detected using local coefficients.
Results The weighted prevalence of EBF in Ethiopia was 58.97% (95% CI 52.67% to 64.99%), and its spatial distribution was found to be clustered (global Moran’s I=0.56, p<0.001). Significant hotspot areas were located in Amhara, Tigray, Southern Nations, Nationalities, and Peoples’ Region, and Somali regions, while significant cold spots were located in Dire Dawa, Addis Ababa and Oromia regions. Kulldorff’s SaTScan V.9.6 was used to detect significant clusters of EBF using a 50% maximum cluster size per population. The geographically weighted regression model explained 35.75% of the spatial variation in EBF. The proportions of households with middle wealth index and married women were significant spatial predictors of EBF.
Conclusion Middle wealth index and married women were significant spatial predictors of EBF. Our detailed map of EBF hotspot areas will help policymakers and health programmers encourage the practice of EBF in hotspot areas and set national and regional programmes focused on improving EBF in cold spots by considering significant predictor variables.
- breastfeeding
- growth
- infant
- neonatology
Data availability statement
Data are available upon reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
High proportion of children didn't get exclusive breastfeeding in Ethiopia. EBF is important to prevent malnutrition, gastroenteritis, diabetes mellitus and hypertenssion. Even though it is important, evidences regarding spatial distribution of EBF in Ethiopia and its determinants are scarce.
WHAT THIS STUDY ADDS
Spatial distribution of exclusive breastfeeding practice was clustered in Ethiopia. Marital status and wealth-index were spatial determinants of exclusive breastfeeding in the country.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
This finding identified spatial distribution of exclusive breastfeeding practices in Ethiopia which can be used as baseline for policy makers to focus on cold spot areas of EBF. In addition, it is important for program planners to easily identify target areas to practice nutritional programs.
Introduction
A WHO report defines exclusive breastfeeding (EBF) for infants under 6 months of age as feeding only breast milk for the first 6 months of life and avoiding taking any other food or drink, including water, with the exception of oral rehydration solutions and drops or syrups containing vitamins, minerals and medicines.1 EBF is important in reducing the risk of childhood mortality, ear infections, type 1 diabetes mellitus, respiratory infections, asthma, gastroenteritis and obesity; developing favourable weight; stabilising metabolic levels to prevent metabolic disorders; and improving cognitive and behavioural development in children.2–4 US dietary guidelines, WHO and UNICEF recommend that infants should be fed exclusive breast milk for the first 6 months of life.5 6 In Ethiopia, the low prevalence (58%) of EBF practices is a public health problem.7 Non-EBF before the first 6 months of age results in hypertension, diabetes mellitus, allergic diseases, gastroenteritis and respiratory tract infections.8 9
A WHO report revealed that the prevalence of EBF in the USA is 24.4%.10 Additionally, the prevalence of EBF among infants younger than 6 months in the USA, in developing countries and in low-income and middle-income countries is 17%, 13.8% and 38.7%, respectively.11–13 In Sub-Saharan Africa, the magnitude of EBF is 36%.14 A systematic review of 16 studies, a meta-analysis and two multilevel findings in Ethiopia revealed that the prevalence of EBF in the country is 59.3%, 59.76% and 46.8%, respectively.15–17
Globally, the rate of EBF during the first 6 months of life is targeted to increase by at least 50% in 2025.18 UNICEF guidelines show that working mothers with infants less than 6 months old and live close to their offices are allowed 30 min of time off twice per day for breastfeeding.5 Geospatial analysis of the trends in EBF prevalence from 2000 to 2018 across 94 countries revealed a wide variation between and within 94 middle-income and low-income countries, and that only three countries are predicted to meet WHO’s global nutrition target of achieving ≥70% EBF prevalence in all its district-level units by 2030.11
Studies have identified significant factors that might lead mothers to start weaning their babies off breast milk before the recommended age, such as sore nipples, milk insufficiency and availability of infant formulas.3 19 Predictor variables such as receiving postnatal care, maternal age, childbirth attended by a healthcare provider, marital status, rural residence, maternal education, caesarean section, wealth index, place of delivery, low birth weight and mother’s employment were also significantly associated with EBF.14 20–22
EBF varies across geographical regions in Ethiopia.16 23 Sore nipples affect the practice of EBF due to nipple pain experienced when babies suck milk.24 Mothers who delivered by caesarean section might choose to use formula feeding due to pain interference when holding their infants. Employed and urban women are often at their workplace, and are therefore separated from their infants, which could lead them to wean their babies off breast milk before 6 months of age.25 26 Mothers who delivered at a health institution, had a childbirth attended by a healthcare provider, and who attended postnatal care (PNC) and antenatal care (ANC) services might have received counselling services on infant feeding, which could have raised their awareness about the importance of EBF.27 Educated mothers might be near and easily accept child health-related information.28 Women with low birthweight infants as well as aged mothers might choose to practise formula feeding due to lack of energy in their low birthweight infants to suck breast milk and aged mothers’ unwillingness to continue breastfeeding due to lack of breast milk.29 Mothers belonging to the poorest wealth index might need to leave their infants alone at home for a long time since they need to be out for work.30 Meanwhile, married women might practise EBF because they have husbands who might help and share responsibilities with them.
Despite its importance, to our knowledge, the spatial distribution and the associated spatial predictors of EBF among infants under 6 months of age have not been studied in Ethiopia. Spatial analysis is important for area-targeted interventions as it identifies high-burden and low-burden areas. It is very important to show the geographical distribution of and the effect of predictors on EBF to apply interventions specific to the area. Therefore, this study aimed to assess the regional variation of EBF and its predictors in Ethiopia. The findings of this study will help policymakers, programme planners, researchers and clinicians facilitate EBF practices. Indeed, mapping hotspot and cold spot areas of EBF will allow the evaluation of previous intervention strategies.
Methods
Study design, subjects, setting and source
A cross-sectional study was conducted using the 2019 Mini-Ethiopian Demographic and Health Survey (Mini-EDHS) data set. Ethiopia is located at the Horn of Africa (3°–15° N latitude and 33°–48° E longitude), which covers about 1.1 million km2 and where the altitude ranges from 4620 m above sea level (Ras Dashen) to 148 m below sea level (Dallol Depression).31 It has nine regions (Tigray, Afar, Amhara, Benishangul-Gumuz, Gambela, Harari, Oromia, Somali, and Southern Nations, Nationalities, and Peoples’ Region (SNNPR)) and two city administrations (Addis Ababa and Dire Dawa).32
The analysis was done using the Mini-EDHS 2019 data set. Samples of enumeration areas were stratified and selected in two stages using the 2019 Ethiopian Population and Housing Census.33 In the first stage, 305 enumeration areas (93 urban and 212 rural areas) were selected. In the second stage, households were selected. Data were requested online from the international Demographic Health Survey (DHS) and accessed from the official website of DHS (www.dhsprogram.com). The detailed methodology of the survey is described in the Mini-EDHS report. Total weighted samples of 548 live infants less than 6 months of age who were the youngest and lived with their mother were included in the analysis.34 For women who had more than one infant during the interview, only the youngest infant was included in the study. Missing data were managed based on the DHS guideline.
Variables
Outcome variable
The outcome variable was EBF, which was defined as infants under 6 months of age who were fed only breast milk, meaning there was no need for other liquid or food except for oral rehydration solutions, drops or syrups containing vitamins, minerals and medicines.35 36 The youngest infants living with their mothers and fed only breast milk were considered to be exclusively breast fed. Therefore, infants were categorised into 1 (EBF) and 0 (non-EBF). Finally, a weighted proportion of EBF per cluster was employed for spatial analysis.
Independent variables
Mothers’ age, number of infants, number of ANC visits, household wealth index (poor, middle, rich), mothers’ marital status, region, place of residence, mothers’ educational level, history of caesarean section, mothers’ PNC, babies’ PNC, type of profession (doctor, nurse, midwifery, health officer, health extension worker, traditional birth attendant) and breastfeeding counselling were candidate predictor variables in the spatial regression model.
Data management and analysis
Descriptive analyses were done using STATA V.17 software. Spatial analysis was performed using ArcGIS V.10.8. Before conducting spatial analysis, the weighted proportion of EBF and candidate predictor variables was obtained in STATA V.17 and exported to ArcGIS V.10.8.
Spatial analysis
The global Moran’s I statistics were computed to investigate the presence of spatial autocorrelation using ArcGIS V.10.8 software. Spatial autocorrelation was employed to test whether the spatial distribution of EBF was dispersed, clustered or randomly distributed across regions in Ethiopia.37 In spatial autocorrelation, the closer geographical areas are to each other, the more related they are. The global Moran’s I spatial statistics generated an output value in the range of −1 to +1. A Moran’s I value close to −1 indicates that EBF was dispersed, a value close to +1 indicates that EBF practice was clustered across regions and a Moran’s I close to 0 shows that EBF is randomly distributed. The spatial analysis decision to accept or reject the null hypothesis is based on the p value of Moran’s I output. In the global Moran’s I analysis, the null hypothesis is rejected if Moran’s I value is significant (p<0.05), which implies the presence of spatial autocorrelation (clustered or dispersed), whereas accepting the null indicates a random distribution of the outcome variable. Getis-Ord statistics were employed to identify significant hot spots and cold spots. However, it is difficult to show hot spots and cold spots in all areas throughout the country, and therefore ordinary Kriging spatial interpolation was used to generalise hotspot and cold spot areas of EBF for the whole country using sampled areas. Kulldorff’s SaTScan V.9.6 was used to show significant clusters of EBF practice using a 50% maximum cluster size per population. Infants fed breast milk exclusively were considered cases, whereas those who were non-EBF were controls.
Spatial regression
Ordinary least squares
After detecting hotspot areas, spatial regression modelling was used to identify predictors of the observed spatial clusters of EBF. OLS is a global statistical model used to test and measure the relationship between outcome variables and predictor variables. The results of ordinary least squares (OLS) regression are only reliable if all six assumptions are fulfilled. The coefficients of explanatory variables in a specified, properly constructed OLS model should be statistically significant and have a positive or negative sign. The model should be non-stationary, include key explanatory variables and be free from multicollinearity. In addition, residuals should be normally distributed, should not reveal spatial patterns and be free from spatial autocorrelation. A data mining tool was used to identify a model that fulfils the assumption of the OLS regression. Additionally, explanatory regression identified models that fulfilled the assumptions of the OLS methods and models with high adjusted R2 values. The final model was validated by internal cross-validation.
Geographically weighted regression
A strong predictor variable in one cluster may not be a strong predictor in another cluster. Such a type of cluster variation (non-stationarity) can be detected using geographically weighted regression (GWR). OLS uses a single linear regression equation for all of the data in the study area, whereas GWR creates an equation for each cluster. Therefore, the coefficients of GWR take different values for each cluster. The GWR map of the coefficients of each predictor variable guides targeted interventions. The GWR model can be written as:
where yi is the observation of response; (uivi) is the latitude and longitude; βk (ui, vi) (k=0, 1,… p) is the p unknown function of the geographical location (uivi); xik is the independent variable at location (uivi), where i is equal to 1,. 2,…; and εii is the error term/residual with zero mean and homogeneous variance σ2.
Results
Characteristics of participants
A total weighted sample of 548 infants under 6 months old participated in this study. About half (278, 50.7%) of the infants were female and 214 (39%) were in the 4–5 months age group. Among the mothers, nearly a quarter (24.8%) were in the 30–34 years age group. Most (232, 42.3%) of the study participants were from Oromia, while the lowest number of participants (2, 0.36%) was from the Harari region. About 262 (47.63%) and 101 (18.43%) respondents belonged to the poor and middle household wealth index, respectively (table 1).
Weighted proportion of exclusive breastfeeding by place of residence and region in Ethiopia, Mini-EDHS 2019
Prevalence of EBF in Ethiopia
In this study, the overall prevalence of EBF was 58.97% (95% CI 52.67% to 64.99%), and the highest percentage of EBF (256, 79.26%) was found in rural areas (figure 1).
Prevalence of exclusive breastfeeding among infants below 6 months old across regions in Ethiopia, Mini-EDHS-2019. Mini-EDHS, Mini-Ethiopian Demographic and Health Survey; SNNP, Southern Nations, Nationalities, and Peoples’ Region.
Spatial autocorrelation
The output of the spatial analysis in this study showed that the distribution of EBF was clustered in Ethiopia, which implies that there is a significant spatial variation in EBF throughout the country, with a global Moran’s I value of 0.56 (p<0.001) (figure 2A).
(A) Spatial autocorrelation and (B) hotspot and cold spot areas of exclusive breastfeeding among infants less than 6 months old in Ethiopia, Mini-EDHS 2019. Mini-EDHS, Mini-Ethiopian Demographic and Health Survey; SNNPR, Southern Nations, Nationalities, and Peoples’ Region.
Hotspot (Getis-Ord Gi*) and cold spot analysis
Statistically significant hotspot areas of EBF were the Amhara, Tigray, SNNPR and Somali regions, while the Dire Dawa, Addis Ababa, and Oromia regions were the significant cold spots (figure 2B).
Kriging interpolation
The results of the spatial Kriging interpolation showed that EBF practice was high in the Amhara, Benishangul-Gumuz, Tigray and SNNP regions, and low in the Afar and Somali regions (figure 3A).
Results of (A) Kriging interpolation analysis and (B) SaTScan analysis of EBF among infants under 6 months old in Ethiopia, Mini-EDHS 2019. EBF, exclusive breastfeeding; Mini-EDHS, Mini-Ethiopian Demographic and Health Survey; SNNPR, Southern Nations, Nationalities, and Peoples’ Region. Foot notes; RR: Relative Risk, LLR: Log-Likelihood Ratio
SaTScan statistical analysis
In this study, the results of the SaTScan analysis of EBF showed that, out of a total of 305 clusters, 46 were significant. Out of 46 clusters, 32 were primary clusters located at 8.156190 N and 38.664590 E, covering a 199.59 km radius in the Amhara region. Infants in this area had 0.44 times less practice of EBF than outside the window (Log-Likelihood Ratio (LLR)=13.19, Relative Risk (RR)=0.44, p=0.003). For detailed information, see figure 3B and table 2.
Significant clusters of exclusive breastfeeding practices among infants under 6 months old in Ethiopia, Mini-EDHS 2019
Factors affecting spatial variation of EBF
OLS analysis
The results of the OLS model revealed that about 32.3% of the EBF variation was explained (adjusted R2=0.323). The variance inflation factor value for the OLS model of this study was less than 7, which indicates there was no multicollinearity. The joint Wald statistics were statistically significant (p<0.001), which indicates that the overall model was significant (table 3). The Jarque-Bera statistics were not significant (p=0.16), which indicates that the model residuals were normally distributed. The Koenker statistics test was found to be statistically significant and showed a non-stationary relationship between the predictor variables and EBF (p<0.001). This recommends the need to conduct GWR to consider the relationship between the predictor variables and EBF by obtaining a local coefficient for every explanatory variable. The proportion of the middle wealth index and the proportion of married women were predictors of EBF.
Summary of OLS results for exclusive breastfeeding in Ethiopia, Mini-EDHS 2019
GWR results
Although OLS analysis predicted hotspot areas, it is a global model that assumes the relationship between each predictor variable and the outcome variable is stationary across the study area. A strong predictor variable in one cluster may be a weak predictor in another cluster. GWR is a local model that can identify this type of cluster variation. In the current study, the Koenker statistics were significant (p<0.001), which violated the OLS assumption of stationary independent variables across the study area. GWR improves the model fit when the relationship between the predictor variables and EBF is non-stationary. Therefore, the GWR model was run to show the effect of every significant predictor across the country using local coefficients of predictors. In this study, GWR analysis revealed that the global model had improved significantly. The Akaike’s information criterion value decreased from 221.64 in the OLS model to 201.81 in the GWR model. In addition, the adjusted R2 obtained from OLS increased from 32.30% to 35.75%, which implies that GWR analysis improved the model’s ability to predict EBF better than the OLS model (table 4). In the GWR model, being married and households that belong to the middle wealth index were statistically significant predictors of EBF.
GWR model for exclusive breastfeeding in Ethiopia, Mini-EDHS 2019
Married women had a positive association with EBF practice in most parts of Ethiopia, while it had a negative relationship in some areas. Specifically, married women had a positive association with EBF practice in most areas of the SNNPR, Amhara, North Tigray and some parts of the Benishangul-Gumuz region. On the other hand, married women had a significantly negative association with EBF practice in most parts of the Oromia region and some parts of the Dire Dawa and Afar regions (figure 4A).
Map of geographical weighted regression coefficients: (A) EBF coefficients for married women and (B) EBF coefficients for middle wealth index among mothers of infants in Ethiopia, 2019. EBF, exclusive breastfeeding; SNNPR, Southern Nations, Nationalities, and Peoples’ region.
The wealth index of households was both negatively and positively associated with EBF practice in different parts of Ethiopia. As the middle wealth index of households increased, EBF intervention increased in most areas of Benishangul-Gumuz, SNNP, and some parts of Amhara and Tigray, while the practice decreased in Somalia, Oromia, Afar and some areas of the Benishangul-Gumuz region (figure 4B).
Discussion
This study aimed to assess the geographical variation of EBF and its predictors among infants less than 6 months of age in Ethiopia using GWR. The findings of the current study showed that the overall prevalence of EBF was 58.97% (95% CI 52.67% to 64.99%). The findings are consistent with previous findings in East Africa and a meta-analysis in Ethiopia using the Ethiopian Demographic and Health Survey 2016 data set.15 38 On the other hand, the results of this study were higher than the findings conducted in Sub-Saharan Africa, Brazil, Nepal and in developing countries.14 39–41 Possible justification for this discrepancy might be cultural beliefs, such as the breast milk spoiling if the mother spends time away from infants, and sociocultural differences across countries.42 43
In this study, the spatial distribution of EBF was clustered across regions in Ethiopia, with a global Moran’s index value of 0.56 (p<0.001). Hotspot areas of EBF were observed in the Amhara, Tigray, SNNPR and Somali regions, while cold spots areas were seen in the Dire Dawa, Addis Ababa and Oromia regions. This might be due to the higher performance of the health extension package, including infant and child nutrition, in the Amhara and SNNP regions than the other regions in Ethiopia.44 In addition, the EBF prevalence report for the Somali region was omitted in the Mini-EDHS 2019 report due to the low proportion of EBF (<50%) in the region.45 Therefore, the high proportion of EBF in the Somali region might be due to overestimation of EBF using non-representative samples in the region. Being a married woman had a positive effect on EBF in most areas of SNNPR, Amhara, Northern Tigray, and some parts of the Benishangul-Gumuz region, whereas it had negative relationship in eastern Oromia and Southern Tigray.
In the current study, middle wealth index and married women were identified as predictors of EBF hotspot areas. In the local GWR analysis, significant predictors of EBF were middle wealth index and married women. The proportions of infants from middle-income households and married households were predictors of EBF hotspot areas. The GWR coefficients for married women and middle-income households ranged from −15.79 to 13.38 and from −102.4 to 114.1, respectively. Positive GWR coefficients indicated that EBF increased, while negative coefficients revealed that EBF practice decreased in the located areas.
As the proportion of married women increased, the practice of EBF also increased in most areas of SNNPR, Amhara, Northern Tigray and in some parts of the Benishangul-Gumuz region. Possible justification for this could be that married women may get support from their husbands, who are critical partners and who provide support to activities such as with breastfeeding, household work and childcare tasks, as well as psychological support such as love, affection and motivation by encouraging the woman to see breastfeeding as a team sport.7 46–48
On the other hand, the proportion of EBF practices decreased as the proportion of married women increased in most parts of the Oromia region and some parts of the Dire Dawa and Afar regions. A possible reason could be the low PNC service coverage, including breastfeeding counselling, in the Oromia and Dire Dawa regions.49 Additionally, most people in the Afar region are pastoralists who do not have a permanent place of residence. Mothers and small children remain at home, while fathers and other family members, including non-lactating women, move seasonally from place to place to look for grassy land and water for their cattle. Health services, including infant feeding counselling, are rarely available in these temporary grassy areas.50 In addition, married women who have migrated with their husbands might not attend and access healthcare services due to lack of caregivers for their other children at home. These might result in lower rates of EBF practices.
As the proportion of middle-income households increased, the proportion of EBF also increased in most areas of Benishangul-Gumuz, SNNP, and some parts of Amhara and Tigray. On the other hand, the proportion of EBF decreased as the proportion of households with a middle wealth index increased in Somalia, Oromia, Afar and some areas of the Benishangul-Gumuz region. This might be because most people living in the Afar region are nomadic and live in an area only for a temporary period, and hence might face lack of access to health institutions. In addition, households with a middle-income index can afford infant formula, which can easily substitute breastfeeding. Also, infants stay with their caregivers when their mothers are away for work and so breastfeeding might be easily replaced with other foods. In addition, after delivery, only 4 months of rest time are given to government-employed mothers in Ethiopia, resulting in early separation of the infants from their mothers. As strength, the current study is the first to use nationally representative data to visualise the spatial heterogeneity factors associated with EBF. In addition, a weighted sample was used to get reliable estimates. As a limitation, the geographical locations of enumeration areas were displaced by up to 2 km in urban areas and 5 km in rural areas for data privacy issues, which may affect the estimated effects of spatial analysis in clusters. Other important possible predictors that could affect EBF practice, such as mothers’ occupation, breastfeeding cultural practices and other healthcare access-related factors like travelling distance and availability of transport to reach health institutions, were not included in the study due to incomplete information in the data set (these variables were not available in the Mini-EDHS 2019 data set). Social desirability and recall bias might also be possible limitations of the study.
Conclusion
EBF was spatially varied across regions in Ethiopia. Significant hotspot areas of EBF were the Amhara, Tigray, SNNPR and Somali regions. In the GWR analysis, households belonging to the middle wealth index and married women were significant predictors of EBF. Our detailed map of EBF hotspot areas will help policymakers and health programmers encourage the practice of EBF in hotspot areas and set national and regional programmes focused on improving EBF in cold spots by considering significant predictor variables.
Data availability statement
Data are available upon reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
Informed consent from the participants was not applicable because secondary data analysis was conducted using the EDHS data set. Data requests and approval for access were obtained from DHS International, and all data were fully anonymous before they were accessed with informed consent.
Acknowledgments
We would like to express our deepest gratitude to the DHS programme for allowing access to EDHS data set for further analysis.
Footnotes
X @DersehNebiyu
Contributors Conceptualisation: GAY, HWA, NMD, MCA, TKT. Formal analysis: GAY, TKT, HWA, NMD, MCA, MAA. Investigation: GAY, TKT. Methodology: GAY, MAA, TKT. Software: GAY, HWA, MCA, NMD, MAA. Supervision: GAY, HWA. Validation: GAY, NMD, MCA, MAA, HWA. Visualisation: GAY, MCA, TKT, NMD, HWA. Writing—review and editing: GAY, MCA, NMD, MAA, TKT, HWA. The Guarantor of the study is GAY.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Map disclaimer The inclusion of any map (including the depiction of any boundaries therein), or of any geographic or locational reference, does not imply the expression of any opinion whatsoever on the part of BMJ concerning the legal status of any country, territory, jurisdiction or area or of its authorities. Any such expression remains solely that of the relevant source and is not endorsed by BMJ. Maps are provided without any warranty of any kind, either express or implied.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.