Use of online food retail platforms throughout the COVID-19 pandemic and associated diet-related chronic disease risk factors: A systematic review of emerging evidence

online food retail with weight gain and emotional eating. Further research is needed to investigate the links between online food retail and obesity.

online food retail with weight gain and emotional eating.Further research is needed to investigate the links between online food retail and obesity.

K E Y W O R D S
COVID-19, digital technology, food delivery, food environments, healthy diets, online food retail

| INTRODUCTION
The coronavirus disease 2019 (COVID-19) pandemic has inflicted a profound effect on health systems, economies, and livelihoods. 1As governments across the world actioned social-distancing measures and stay-at-home orders to mitigate the spread of the disease, many people's lifestyle behaviors were significantly disrupted. 2These measures restricted people's ability to exercise, 3 increased stress and anxiety levels, 4 and heightened demand for unhealthy foods. 5The World Health Organization (WHO) has reported that rates of weight gain have risen during the COVID-19 pandemic and that this increase was likely attributed to more "obesogenic" changes to the food environment 6 including a surge in promotion of fast-food, alcohol, and sugarsweetened beverages on social media platforms. 7,8stematic reviews on the impact of the COVID-19 pandemic on food consumption have revealed adverse changes in eating habits.
González-Monroy et al. 9 reported increased snack frequency and preference for sweets and ultra-processed food instead of fruits, vegetables, and fresh foods and a decreased adherence to healthy diets.
Similarly, Neira et al. observed an increase in food consumption, weight, and body mass index (BMI). 10A meta-analysis of six studies across Italy, Spain, Poland, Turkey, and Australia and cohorts from Asia, Africa, and Europe showed that body weight change from March to May 2020-the first phase of restricted movements during the pandemic-observed a significantly higher body weight in the postlockdown period compared to a period before the lockdown.The weighted mean between-group difference was 1.57 kg (95% confidence interval [CI]: 1.01, 2.14). 11Although this change in weight may not appear to be a clinically meaningful result of ≥2 kg, 12 it is noteworthy to consider the rapid rate at which this weight gain occurred within a few months of the pandemic.
The United Nations has also acknowledged the significant impacts of the COVID-19 pandemic in further undermining the food security and nutritional status of millions of households across the world. 13Due to food supply chain issues and restricted store opening hours, an individual's access to food was significantly altered during this time. 14In a systematic review conducted by Eliás and Jámbor, most (78%) studies reported an increase in household food insecurity during the first year of the COVID-19 pandemic. 15Research has shown that food insecurity is associated with poorer diet quality 16 and increased risk of obesity 17 and chronic diseases such as hypertension and type 2 diabetes. 18cial-distancing measures have contributed to increased demand for e-commerce and contact-free delivery 19 through online food retail platforms.A recent review outlined three distinct categories of online food retail platforms with the greatest implications for public health including online groceries for fresh produce and pantry staples (e.g., Amazon Fresh ® ), meal delivery applications (apps) for readymade takeaway meals (e.g., Uber Eats ® ), and meal kits with ingredients for cook-at-home meals (e.g., HelloFresh ® ). 20e United Nations Conference on Trade and Development (UNCTAD) reported globally that the online share of total retail sales increased from 16% to 19% in 2019-2020, a three-percentage point rise compared to a two-percentage point rise between 2018 and 2019. 21UNCTAD also indicated that throughout the COVID-19 pandemic, online retail sales increased substantially in the countries that comprise half of the global gross domestic product (GDP) including China and the United States-growing from $2 trillion USD in 2019 to $2.9 trillion in 2021. 22An analysis of Google search interest in "delivery" in the Organisation for Economic Co-operation and Development (OECD) countries such as the United Kingdom and Germany found that interest doubled in a short period before and after COVID-19 confinement measures were implemented. 19line food retail platforms were therefore likely an important method to access food for many people during the COVID-19 pandemic and potentially contributed a significant portion to the growth in total online retail sales.The full extent of the adoption of online food retail throughout the pandemic, however, remains underexamined.It is important to understand how usage of these online sources of food may have changed during lockdowns and to consider what impact these platforms may have on people's diets.There are signs from the growing body of literature that indicate that meal delivery applications are causes of concern for health due to the poor nutritional quality of foods offered online 23,24 and lack of energy labeling on menu items to inform consumers. 25Moreover, results from a recent systematic review on the evidence gaps in assessing the healthiness of online supermarkets also indicate that there is an advanced ability to disguise healthier products and promote unhealthier products through marketing strategies in an online grocery retail setting. 26 such, the primary aim of this systematic review was to exam-

| METHODS
Systematic review was conducted under the standards established in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). 27The review protocol was prospectively registered in the International Prospective Register of Systematic Reviews (PROSPERO) (Registry Number CRD42022320498).

| Eligibility criteria
Quantitative studies including cross-sectional and longitudinal studies, qualitative studies, and mixed-methods studies published in peerreviewed journals were included.In addition, studies were included if the following criteria were satisfied: (i) conducted during the COVID-  "SARS-COV-2" AND "stay-at-home order" or "lockdown" or "restrictions" AND "digital platforms" OR "mobile applications" OR "internet" AND "online food delivery" OR "online food shopping" OR "online food ordering" OR "meal delivery."The full search query for the MEDLINE via Ovid database search is provided in Table S1.

| Selection process
Search results from all databases were retrieved and merged using Endnote X9 reference management software (Clarivate Analytics, Philadelphia, USA) and imported into Covidence (Veritas Health Innovation Ltd, Melbourne, Australia), a web-based collaboration software platform used to streamline the systematic review process.In Covidence, duplicates were removed, and studies were first selected from title and abstract screening.This was performed independently by two authors (S.S.J. and R.R.).Full texts were retrieved for the studies included from the title and abstract screening process and assessed against the inclusion criteria for a second time.Any discrepancies regarding eligibility for included studies were resolved through discussion between authors S.S.J. and R.R.If discrepancies remained unresolved, a third author (S.R.P.) was consulted.

| Data collection process
Data extraction was independently conducted by one author (S.S.J.) and validated for accuracy by another author (S.W.).

| Data items
The following data were extracted: first author details, study setting, study aims, study design, time of study, target population, and sample characteristics including sample size, age and gender of participants, and income level of participants.To address the main research questions, the following were extracted: (i) outcomes and measures used to assess change in use of online food retail, (ii) findings relating the use of online food retail to diet-related chronic disease risk factors, and (iii) themes from lived experiences of sourcing food from online platforms during COVID-19.

| Synthesis methods
The convergent segregated approach of the Joanna Briggs Institute (JBI) methodology for mixed-methods systematic reviews was used for data synthesis. 28 Diet-related chronic disease risk factors associated with the use of online food retail platforms were evaluated for their potential impact on health as either negative or positive.Evidence of dietary behaviors that could have a positive impact on health would involve orders of foods from the five food groups 29 such as vegetables and legumes, fruits, dairy products and alternatives, grains and cereals, and meat and alternatives.Conversely, orders of "discretionary foods" 29 such as fried and salty foods, sweets, and other foods high in energy and low in nutrients would be deemed to have a "negative" impact on health.

| Study risk of bias assessment
The JBI critical appraisal checklists were used to assess the quality of the included studies based on their study design.Cross-sectional studies were appraised using the relevant checklist, which encompassed 8 items 30 and 11 items for cohort studies. 31Qualitative studies were assessed against the relevant 10-item questionnaire. 32For mixedmethods studies, the JBI checklists were used to assess the quantitative and qualitative components separately.An overall "risk of bias" judgment was assigned based on the total points the studies had accumulated from their quality appraisal checklist.An answer resulting in "yes" scored 3 points, "no" scored 0 points, and "unclear" scored 1 point. 33The overall quality of the study was based on the percentage of points earned under each quality appraisal tool.A study that was of "poor quality" attained only 33% of the maximum number of points, "moderate quality" scored between 33% and 67%, and "high quality" was given to studies that scored above 67%.Two authors (S.S.J. and E.S.) independently assessed each study, and any disagreements were resolved by a third reviewer (S.R.P.).

| Deviations from PROSPERO protocol
In this review, some terminology varied from the protocol to ensure alignment with the published literature on digital food environments, which is rapidly emerging.That is, instead of online food purchasing platforms, online food retail was used as an umbrella term to describe the main categories of online food delivery.In addition, given the diverse age groups captured in each individual study, an analysis of online food retail use by age was not conducted.

| Study selection
In total, 53 studies were included in the review following full-text screening as displayed in the PRISMA flow diagram (Figure 1).

| Study characteristics
The characteristics of the included studies are summarized in Table 2.
There were 46 cross-sectional studies, 4 qualitative studies, 2 longitudinal cohort studies, and 1 mixed-methods study.For cross-sectional and cohort studies, 34% (21/62) had a sample size between 1000 and 4000 participants.Sample sizes for qualitative studies ranged between 12 81 and 34. 82The COVID-19 time periods for when studies were conducted were documented and categorized into quarterly periods starting from 2020.Nearly 40% of outcomes (21/53, 39.6%) were from studies conducted during the second quarter of 2020 from April to June.

| Geographical region and countries
Of the 53 included studies, six studies were multi-country [34][35][36][37][38] and reported outcomes for each individual country, thus resulting in a total of 64 outcomes.As shown in Table 2, study outcomes from East Asia and the Pacific comprised 27% of the total (17/64), North America comprised 22%, and Europe and Central Asia made up 20.6%, followed by the Middle East and North Africa, Latin America, and South Asia.
None of the included studies were conducted in sub-Saharan Africa.
More than half (55.6%) of the included studies were conducted in high-income countries according to the World Bank classification, and no studies were conducted in low-income countries.

| Participant characteristics
Of the 52 studies, three studies reported participant characteristics for the different countries sampled, 34,37,76 three studies each reported results for two groups of participants, 35,36,78 while two studies 54,58 shared the same group of participants.As such, there were a total of

| Change in frequency of use of online food retail platforms during COVID-19
The use of online food retail platforms was measured by the change in frequency during the pandemic compared to before the pandemic.
The results are summarized in Table 3 and described below.Collectively, there were 45 reported outcomes on the usage of online groceries, 40 outcomes on the use of meal delivery apps, and only 2 outcomes on meal-kit delivery services.As each study may have reported multiple outcomes on the use of online food retail platforms, the number of outcomes outnumbers the referenced studies in Table 3.
Of the 45 total outcomes that noted a change in frequency of online grocery use during COVID-19, 43 outcomes suggested an increase in usage compared to before COVID-19, 1 outcome implied varied use, and 1 outcome showed a decrease in usage during COVID-19 (Table 3).Varied use indicated that among the participants sampled for the study, there were both similar proportions of participants reporting a decrease and an increase in use.
For meal delivery applications, there were 22 outcomes that found a decrease in use compared to 18 outcomes that demonstrated an increase in use.The two outcomes on meal-kit delivery services both indicated that the impacts of COVID-19 lockdowns led to an increase in use.

| Changes in the use of online food retail platforms throughout different time periods of the COVID-19 pandemic
This review showed that online groceries and meal delivery apps varied in use across different time periods of the COVID-19 pandemic.
During the initial outbreak in the first quarter of 2020 (Q1 2020), two studies from China found that there was an increase in use of online groceries. 61,6280,81 While four studies reported an overall increase in the use of meal delivery apps, 35,52,78,81 more studies saw a decrease in use. 36,37,40,47,54,66,75In the third quarter of 2020 (Q3 2020), the pattern of increased online grocery shopping was maintained 42,44,45,63,70,74,84,86 with only one study that noted a decrease. 85Meal delivery apps during this period saw more mixed results with a similar number of studies, which reported decreased usage 42,45,64,70 compared to an overall increase in use. 44,65,84,85The same trends were observed in the fourth quarter of 2020 (Q4 2020).The studies conducted during this COVID-19 period all observed an increase in online groceries. 34,38,46,48,55,57,67More studies saw decreased use of meal delivery apps 34,38,46,48,57 compared to an increase in use. 38,49,55From the second half of 2021 to the last quarter of 2022, all study outcomes showed an increase in use of online groceries and meal delivery apps.3.5 | Themes from qualitative studies relating to the use of online food retail platforms Findings from the four qualitative studies included in this review suggest that online platforms offered an alternate way to source food during the COVID-19 lockdowns.These platforms were important in rural communities and areas where it was challenging to venture outdoors according to two studies conducted in India and the United States. 82,83Two qualitative studies also suggest that ordering takeaway meals via meal delivery apps was often a treat or "reward" for families during lockdown. 81,84One study mentioned the benefits of meal delivery applications in alleviating cooking duties for female gatekeepers who had already experienced high levels of stress from working from home and caring duties. 88Some negatives of online grocery shopping were highlighted in another study.Interviews revealed that participants thought online grocery shopping was cumbersome as orders needed to be prepared ahead of time and, at times, were incorrect with items missing in the delivered package. 81

| Associations between the use of online food retail platforms and diet-related chronic disease risk factors
Nine of the 52 included studies reported on diet-related outcomes associated with meal delivery app use (Table 4).The potential impacts on health were categorized as either positive or negative.Eight outcomes were associated with negative diet-related chronic disease risk factors, 38,41,49,53,64,66,76,87 and only one outcome was perceived as promoting health. 86Negative impacts on health relating to use of meal delivery apps included associations with perceived obesity status, 41 weight gain, 38,49 emotional eating, 53,65 snacking behavior, 66 and a shift to unhealthier food orders. 66,76One study showed that more frequent users of meal delivery apps had lower vegetable intake and higher sugar-sweetened beverage consumption compared to less frequent users. 87Another study that appeared to associate online food ordering with increased dietary diversity and higher dietary quality was considered to positively influence health. 86

| Risk of bias
The risk of bias for 46 of the 53 studies was assessed using the JBI critical appraisal checklists for cross-sectional studies.The JBI checklists for qualitative studies were used to assess 4 of the 53 studies, and the JBI checklists for cohort studies were used for 2 of the 53 studies.
The JBI checklists for qualitative studies and cross-sectional studies were used to assess the quality of the one mixed-methods study.A total of 33 of the 46 cross-sectional studies received a "high" score, 12 of the 46 were given a "moderate" score, and a "low" score was determined for one study. 69All results from the critical appraisal for cross-sectional studies are presented in Figure S1.
T A B L E 4 Diet-related chronic disease risk factors associated with the use of online food retail platforms and potential impacts on health.Frequent users of meal delivery apps-those that used it at least twice per week reported a 12% lower frequency of consuming two servings of vegetables per day compared to those who used meal delivery apps less than once a week (aPR: 0.88, 95% CI: 0.63, 1.24).Compared to infrequent meal delivery app users (less than once a week), those who used meal delivery apps at least twice a week and those who used meal delivery apps once a week were 11% and 38%, respectively, more likely to drink sugar-sweetened beverages Negative Varese 2023 76 Before the pandemic, meals such as pizza, fruits, and vegetables were ordered more in Poland.After COVID-19, hamburgers were more often ordered in Poland Negative Zhang 2020 86 Participants who depended mostly on online food ordering and delivery had significantly higher Household Dietary Diversity Score than those who only did in-person grocery shopping or relied on a combination of in-person shopping and in-house storage supplies Positive Abbreviations: aPR, adjusted prevalence ratio; CI, confidence interval; COVID-19, coronavirus disease 2019; OR, odds ratio.
Of the four qualitative studies, two were rated as "high" quality and two as "moderate."The two studies that were given a moderate rating lacked a statement locating the researcher culturally or theoretically, and the influence of the researcher on the research, or vice versa, was not adequately addressed (Figure S2).Both longitudinal cohort studies were given a "moderate" score (Figure S3).The one mixed-methods study included in this review received a "high" score as a cross-sectional study and as a qualitative study, which reflected a good integration of both methods.

| DISCUSSION
Key findings from this systematic review demonstrate a strong shift to using online platforms to purchase food across many countries during COVID-19 restrictions.There was widespread uptake of online grocery shopping across the majority of studies included in this review, with 43 of the 45 (96%) outcomes supporting an increase in use.Unexpectedly, 22 of the 40 (55%) study outcomes reported that many participants decreased their use of meal delivery apps for takeaway meals.
These outcomes, however, were counteracted by other studies that reported a high proportion of participants who had increased their use of meal delivery apps.Compared to positive dietary behaviors, more studies reported that the use of online food retail was associated with negative dietary chronic disease risk factors such as weight gain and emotional eating.Taken together, the emerging evidence suggests that while there was considerable uptake of online food retail platforms as a source of food throughout the pandemic, these platforms may promote unhealthy behaviors that contribute to overweight and obesity.
Studies in this review consistently reported an increase in use of online grocery delivery services throughout the COVID-19 pandemic.
A potential reason for using online grocery delivery may have been linked to participants' fear of contracting COVID-19, as outlined in a study included in this review. 59Moreover, a recent meta-analysis conducted by Tyrväinen and Karjaluoto further investigated the purchase intentions behind online grocery shopping during the pandemic. 89The analysis found that consumers did not make online grocery purchases because of their perceived usefulness or positive attitude towards these services, but rather because the abrupt changes instigated by COVID-19 restrictions "forced" their shopping to move online. 89ere are several potential implications of increased online grocery shopping on food choices, nutrition, and health.Recent research from the United Kingdom during COVID-19 suggests that while online shopping was an alternative way to purchase fruit and vegetables during the pandemic, the expenditure for processed products surpassed the expenditure for fresh products. 90This is potentially concerning as a growing body of evidence has linked the consumption of ultra-processed foods to increased risk of chronic diseases. 91,92A study from the United States has shown that consumers from lowincome households may be buying less fresh fruits and vegetables online compared to in-store. 93In addition, a virtual supermarket program implemented in the United States in another study also found that although nearly half of participants increased their purchases of healthier products, an equally large proportion reported buying more juices and sugar-sweetened beverages. 94It is evident that online grocery purchases may have a substantial impact on diet quality, and while there is room to nudge consumers towards healthier choices, there is a matched capacity for the purchase of unhealthier choices.
Many studies included in this review reported a decrease in participants' use of meal delivery apps throughout the COVID-19 pandemic.
These results, while unexpected, were similar to the findings from a foods (47%) compared to 23% of participants who reported an increase in meal delivery. 95Despite this, these results differ from market research, which reported that meal delivery apps experienced an overall surge in use. 96,97A report from Accenture also found that after lockdown, customers of Uber Eats had spent three times more compared to pre-pandemic levels. 98Timing may have also impacted the results.
In this review, it was shown that, compared to the first half of 2020 in the early stages of the pandemic, more studies in the fourth quarter of 2020 observed a decrease in use of meal delivery apps.Research from the United States showed that spending at full-service restaurants returned slowly in 2021, 99 suggesting that some individuals and households returned to dining in at restaurants once restrictions were eased.
Concerningly, of the nine diet-related outcomes associated with online food ordering during COVID-19, eight outcomes linked the use of these services to weight gain, emotional eating, snacking behavior, and a shift to unhealthier food orders.Another cross-sectional study from Australia echoes these results, as respondents who engaged in meal delivery applications were more likely to be of higher BMI and have higher levels of sugary drink consumption and fast-food restaurant patronage. 100Alternatively, one study noted that ordering groceries online was associated with improvements to dietary quality and dietary diversity.In spite of this, the benefits of dietary diversity greatly depend on the constituents and proportions of the foods that make up the diet.Higher intakes of processed foods, refined grains, and sugar-sweetened beverages negate lower intakes of minimally processed foods, and this has been observed in the literature. 101reover, a study revealed that youth in China were ordering spicy hot pot meals, fried foods or hamburgers, and cakes/pastries the most via meal delivery applications during COVID-19. 102This again highlights that while there is a great potential for healthy choices to be made through online food retail platforms, there is a greater propensity towards unhealthier choices.
During crises, food acquisition can be challenging amidst food production and supply chain issues.Qualitative studies included in this systematic review indicate that online food retail platforms were an important alternative to sourcing food during the pandemic.Participants in rural communities and areas particularly valued the option of getting food delivered.A study from Nanjing, China, has similarly demonstrated the crucial role of online food access for food acquisition during COVID-19 lockdowns. 103Researchers from this study suggested that food insecurity issues were avoided by most survey participants due to the high integration of physical and digital food outlets in Chinese urban communities.The accessibility of foods from physical food outlets was supplemented by digital food outlets, which offered a delivery service for foods. 103Despite the promising prospects of online food retail in easing food security concerns, the extra delivery costs for online forms of food retail may be a disadvantage for some households, 104 and during the pandemic, online food delivery was often unavailable due to high demands. 52To ensure all individuals have safe and secure access to food during future emergencies and crises, it is imperative that online food retail settings improve and promote more healthful options and become more affordable and available, especially in rural areas.

| Strengths and limitations
To our knowledge, this systematic review is the first to comprehen- Finally, it was likely that many households supplemented their instore shopping with online grocery delivery, and this dual modality of food shopping must be considered when interpreting the results from this review.It is also an indication that future policies should promote healthy food choices in both physical and digital food settings, given the likelihood of consumers mixing their modes of food shopping.
The lack of studies from low-income countries is an important research gap to address in future studies.The digital food environment must be assessed and monitored in low-income as well as higher income countries.Current online food retail behaviors and associated diet-related chronic disease risk factors following the pandemic should also be further investigated to examine the extent of the COVID-19 pandemic's impact on population diets and health.

| CONCLUSION
Restrictions resulting from the COVID-19 pandemic have threatened food accessibility, and online food retail platforms have provided an alternate option for sourcing food.Understanding how populations across the world have used these online food retail platforms and their associated diet-related chronic disease risk factors may provide more clarity on their risks to population health.This review demonstrates a surge in the adoption of online food retail platforms.Importantly, the small body of evidence presented in this review indicates that the use of online food retail is associated with negative impacts on diet-related chronic disease risk factors.It is critical to further investigate policies and measures that will help improve the healthfulness of online food retail settings.Given the rapid adoption of digital technology and its largely unmeasured impacts on behaviors and health outcomes, the digital food environment requires further investigation and monitoring.
ine people's change in use of online food retail platforms during the COVID-19 pandemic, including their lived experiences and their access to food during COVID-19 lockdowns.A secondary aim was to identify diet-related chronic disease risk factors including dietary intake, eating behaviors, and/or weight status associated with the use of online food retail platforms during the pandemic.Unhealthy dietary behaviors developed during COVID-19 may elevate the risk of obesity and associated chronic diseases.Thus, the findings from this review have the potential to inform policy planning for future pandemics in terms of dietary behaviors related to ordering food from online platforms.

2. 3 |
Search strategyAn experienced academic librarian was consulted to formulate the search strategy.In consideration of the small volume of literature around online food retail during the COVID-19 pandemic, a broad search strategy was applied.The following search term categories, combinations, truncations, and synonyms were used: "COVID-19" or Quantitative findings were narratively summarized from the data extraction table, which gathered data on study characteristics and results for the outcomes of interest including the use of online food retail and its association with diet-related chronic disease risk factors.The use of online food retail platforms was measured by the change in frequency during the pandemic compared to before the pandemic.A content analysis approach was used to synthesize findings from qualitative studies on participants' use of online food retail during COVID-19-including their reasons for use and experiences.Concepts were formed based on the frequency of terms.Data on the use of online food retail during COVID-19 were further synthesized by the COVID-19 time period.
62 outcomes based on participant characteristics.Most studies had F I G U R E 1 Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram using the PRISMA 2020 statement. 27T A B L E 2 Study characteristics of the included studies (n = 53).
rapid assessment online survey of over 2000 individuals from 119 countries during April 2020, the second month of the COVID-19 pandemic.The study conducted by O'Meara et al. demonstrated that a larger proportion of participants decreased their delivery of takeaway sively synthesize published peer-reviewed studies to encapsulate the use of online food retail platforms during the COVID-19 pandemic and associated chronic disease risk factors.Using nine electronic databases, a comprehensive search was conducted.In addition, this review has also captured data from both quantitative and qualitative studies, which is important for illuminating the experiences of individuals during the pandemic.Several limitations are, however, noted.It was not possible to assess the change in usage as nearly 90% of the included studies were cross-sectional, across different countries and time points, and used different survey tools.Cross-sectional studies have inherent biases due to self-reported measures and reliance on retrospective recall of participants' behaviors during COVID-19.Despite this, the COVID-19 pandemic was an unexpected event, and as such, collecting objective in-person data amid restrictions was often unachievable and limited the research conducted during this time.
Population, Intervention, Comparison, Outcome, and Study Design (PICOS) inclusion and exclusion criteria.
19 pandemic (from January 1, 2020, onwards) 1 and (ii) covered the subject of online food retail platforms including online groceries, meal ,44-49,52,56,61,63-65,68,72,75,78,85,86Changes in the frequency of use of online food retail platforms before and during COVID-19.Note: N indicates the number of individual outcomes supporting the change in use; outcomes may outnumber the referenced studies as studies may have reported more than one relevant outcome.