Abstract
Purpose
To assess changes in publications of health behavior research in leading journals across a 20-years period.
Design
Systematic comparative descriptive snapshot analysis of all primary research-based articles published in the 2003 and 2023 issues of five leading health behavior journals.
Setting
Document review.
Sample
Articles (n = 1207) published in the following journals in 2003 and 2023: American Journal of Health Promotion, American Journal of Preventive Medicine, Annals of Behavioral Medicine, Health Education & Behavior, and Health Psychology.
Analysis
T-tests and chi-squared tests were used to assess differences in publication characteristics between the 2 years.
Results
The number of published studies more than doubled between the two time points. Significant changes were seen between the health behavior of interest, methodology, setting, and the use of large data sets.
Conclusions
Compared to 2003, there were twice as many articles published in 2023. Sample sizes were larger, but cross-sectional studies still predominated. Theoretical frameworks have greatly expanded, and there were no clear preferred theories.
Critical Limitations
Papers were only abstracted from five journals across 2 years. Comparisons of other journals and other volumes may yield different results.
Introduction
Over the past 2 decades, the field of health behavior research has undergone significant transformation, driven by advances in methodology, technology, and public health priorities. National Institutes of Health funding for health research increased 45% from 2003 to 2023. 1 Despite this transformation, chronic diseases and conditions such as heart disease, cancer, stroke, and diabetes remain as major contributors to the top ten causes of death in the United States. 2
Trends in the proportion of Americans engaging in key health behaviors contributing to these diseases and conditions vary, but overall, have been relatively stable. Rates of current cigarette use show the largest change over the past 20 years, decreasing from 21.5% in 2003 to 11% in 2023. 3 Although physical activity (PA) is a major contributor to health promotion and disease prevention, the number of adults who report no PA has remained relatively constant with a static one quarter of adults reporting no PA outside of work. 4 The percentage of adults and children consuming little or no fruit has increased from 33.4% between 2005-2008 to 40.3% between 2017-2020. 5 Vegetable intake among adults from 1999-2018 has remained essentially unchanged. 6
Evidence shows the importance of quality sleep for physical and mental health. 7 In 2022 Gallup survey, one-third of American adults reported their sleep as fair or poor compared to 18% from a 2005 study. 8
Poor mental health remains a public health issue and appears to be worsening. Data from the Behavioral Risk Factor Surveillance Survey (BRFSS) reported that the percentage of US adults reporting 14 or more poor mental health days per month rose from 13.8% in 2019 to 15.6% in 2024. 9 Over half (51%) of US employees now report feeling significantly stressed. Persistent feelings of sadness and hopelessness among youth increased from 30% in 2013 to 40% in 2023. 10 Certain population groups such as women, minority, low-SES and older-age groups experience a disproportionate burden of mental health concerns. 11
Novel infectious respiratory diseases appeared repeatedly throughout the past 20 years and included SARS (2002-2004), H1N1 (2009), MERS (2012), and COVID-19 (2019-2021). 12 The COVID-19 pandemic took the lives of over 1.1 million Americans and contributed to a decrease in life expectancy. 13
There has been a significant increase in the awareness of structural inequities contributing to health over the past 2 decades.11,14 The acknowledgement of “social determinants of health” has become commonplace in health behavior research, and funders are calling on investigators to promote health equity. 15 One of the overarching goals in Healthy People 2030 is specifically related to SDOH is to “create social, physical, and economic environments that promote attaining the full potential for health and well-being for all.” 16 Recognizing disparities and addressing equity has become a cornerstone of many health behavior studies.
Other societal factors have shaped health behaviors over time. Ninety-eight percent of American adults now own a cell phone, and smart phone ownership has substantially increased (to 91%). 17 There are now hundreds of thousands of health apps and digital health tools such as fitness trackers and other biometric technologies impact both personal health assessment and the use of this data in research. Increasing internet usage and streaming services has also impacted human behavior. 18
There has also been a shift in focus from individual approaches to behavior change to more of a socio-ecological approach which acknowledges the contribution of social, community, and policy factors. With the advancement of this approach as a means to influence and improve population health, there has also been an increase in the opposition to certain policy approaches (ie, soda taxes) to influence healthy behaviors, mainly borne from divisive political ideologies. 19 Although the social ecological model was introduced in the 1980s, it did not gain widespread adoption until after 2000. 20
As the landscape of health promotion and disease prevention has evolved over the past 2 decades, little is known about how the questions, tools, and frameworks that define health behavior research have evolved, and the platforms through which the findings are disseminated. Questions remain about how much substantive progress has been made in advancing the field’s scientific foundation, relevance, and impact.
To address this, we conducted a systematic comparative analysis of articles published in 2003 and 2023 across five leading peer-reviewed journals in health behavior research. By examining trends in research design, theoretical application, target populations, behavior change strategies, and attention to equity and context, we aimed to assess how the field has evolved and where it has remained static. This review offers a longitudinal lens through which to evaluate whether health promotion research is keeping pace with emerging health priorities and methodological innovations, or if it continues to replicate patterns established in prior decades. In doing so, we seek to inform future research directions, funding priorities, and training needs that can strengthen the rigor, relevance, and reach of health promotion in the years ahead.
Methods
Journal Selection
The research team selected five peer-reviewed journals that publish a majority of papers in the health promotion and health behavior literature. Journals were selected if they were in Q1 as identified by Scimago in both years and had an average h-index of at least 100. 21 Journals were excluded if they did not have a focus on health promotion (eg, epidemiological or health services focus), were not published in English, or did not exist at both years (2003 and 2023) of this study. Given this criteria, American Journal of Health Promotion, American Journal of Preventive Medicine, Annals of Behavioral Medicine, Health Education & Behavior, and Health Psychology were selected. Using PubMed, abstracts from all articles within the selected journals for the years of 2003 and 2023 were downloaded and imported into a database.
Article Inclusion/Exclusion Criteria
The research team established the following criteria for article inclusion. All primary research studies with a health promotion focus or behavior outcome were included. Articles were included if they reported quantitative, qualitative, or mixed- methods study design. Exclusion criteria were: reviews, commentaries, and studies that did not have a health behavior focus (eg, disease focus or health services research). To evaluate fit with the study goals according to those criteria, two members of the research team independently evaluated article information as coded it as “included” or “excluded”. If excluded, a brief statement indicating reason for exclusion was reported. The study team members met to discuss their coding and reach consensus for any discrepant inclusion or exclusions. Full text from all papers included in the selection criteria were available to the study team. There were no additional exclusion criteria.
Coding and Abstraction
Codes and Definitions Abstraction Information
Analysis
Once all abstracts were coded, the Qualtrics data was downloaded into SPSS v 26 for analysis. Data were reviewed for accuracy by a team lead, and frequencies were run to identify outliers and to determine if patterns existed for missing data. Once the dataset was prepared for analysis, the team conducted descriptive statistics for each code, for each journal, for each year. Descriptive statistics, t-tests and ANOVA were used to assess difference between selected articles published in 2003 and those published in 2023.
Results
Categories of Articles Published in Five Leading Health Behavior Focused Journals 2003 and 2023
Total number of health behavior related research articles published by journal 2003 and 2023
Article Characteristics by Year
a – Behavioral focus areas were analyzed separately due to the presence of multiple behavior focused studies.
Next, study design was assessed. Cross-sectional studies were the most common in both years (38.5% vs 39.2%). Cohort and longitudinal studies increased from 18.3% of studies in 2003 to 25.3% in 2023 while randomized clinical/community trials decreased from 16.1% to 10.1%. Qualitative studies decreased from 10.1% to 6.7%. Quasi-experimental and mixed methods studies each appeared less than 5% of the time in both time points. Among studies where an intervention was delivered, there was a significant change from in-person to virtual intervention delivery. In 2003, 78.2% of interventions were delivered in person compared to 48.8% in 2023. Table 2 contains the complete results.
The mention of theoretical frameworks or behavioral theory in the articles was uncommon. Behavioral theory was mentioned slightly more frequently in 2003 (35.1%) than in 2023 (32.7%, P < .001). Seventy-four articles mentioned use of a theory or a theoretical framework in 2003. These included the Transtheoretical Model (17 studies; 23%), social ecological model (12; 16.2%), Health Belief Model, Social Cognitive Theory and Theory of Reasoned Action/Planned Behaviors (9 each; 12.2%). In 2023, 161 articles mentioned a theoretical framework or behavioral theory: Social Cognitive Theory (19 studies, 11.8%), Health Belief Model (14, 8.7%), Theory of Reasoned Action/Planned Behavior (10; 6.2%), social ecological model (9, 5.6%) and the Transtheoretical Model (3; 1.9%). In 2023, there were also studies reporting unique behavioral theories or theoretical frameworks that were only mentioned in one article. Examples of these unique theories were the anti-expert model, the conspiracy beliefs model, necessity-concerns framework, and the precaution-adoption process model.
Next sample characteristics and study settings were assessed. The percent of females in studies significantly decreased from 67.8% in 2003 to 60.9% in 2023, P < .001. Non-significant differences were seen for the percent of the sample over 65, which decreased from 12.6% to 9.5% and the percent of the sample that was non-White from 50.7% to 46.5%. The settings where the study took placed changed significantly between 2003 and 2023 (x2 (5) = 12.8, P < .05). At both time points, many of the studies focused on individual level (59.2 vs 62.0). School (12.4 vs 8.0), healthcare (8.7% vs 5.3%) and workplace (2.8% vs 1.5%) decreased while community increased slightly (12.8 vs 13.7%).
Sample size changed dramatically with the use of big data sets. In the 624 studies with sample size reported, 103 had sample sizes larger than 10 000. Of these 18 (9.7%) were published in 2003 compared to 85 (19.3%) in 2023. All 24 studies with sample larger than 250 000 were published in 2023. After removing studies with a sample size larger than 10 000, sample sizes were significantly larger (M = 1,479, SD = 1519) in 2023 compared to 2003 (M = 970, SD = 2219).
Data collection methods were then assessed. Secondary, non-survey data (eg, content analysis of social media) increased significantly from 6.0% of studies in 2003 to 23.9% of studies in 2023 (P < .001). Interviews decreased significantly from 18.3% to 9.7% (P < .001), while surveys remained the most common way to collect data (61.9% vs 56.2%) and observational data remained rare (3.2% vs 3.4%).
Discussion
In the 20 years between 2003 and 2023, many aspects of research changed, yet many stayed the same. The first and most obvious change is the number of papers published by these five journals. With the move toward publishing on-line only and less restrictive page limits, the journals included in this study expanded in volume publishing more than double the number of papers they did in 2003. However, newer on-line journals publish significantly more articles than the journals assessed in our study. The rise of mass publishers including MDPI and Frontiers that require pay for open access has proliferated. For example, The International Journal of Environmental Research and Public Health published 894 articles in January 2023 more than all five of the journals reviewed here published that year. This trend has also affected the journals we reviewed. In 2014, Preventive Medicine launched their cascade journal, Preventive Medicine Reports. In 2023, Preventive Medicine Reports published 6 volumes with 137 papers appearing in the December volume alone. The number of journals accepting health promotion related papers has also grown, and include newer journals founded close to or after 2023 such as BMC Psychology, BMC Public Health, Frontiers in Public Health. This increase in publishing has been noted across scientific fields and creates challenges for scientists in reviewing and keeping current with the literature. 22 The increased availability of publishing opportunities arguably has the potential to facilitate the more rapid dissemination of the results of rigorous research, but may also unintentionally lower the bar for what is ultimately published. It also creates challenges for the tenure and promotion and academic recognition process. It is common for new faculty to inquire about the expected number of publications for promotion and tenure, yet standards for these expectations have changed over time. 23 Because publication outlets were far more limited in 2003 with fewer journals and restricted space in monthly or quarterly issues, achieving 50 publications at that time required navigating a much more constrained and competitive publishing landscape than achieving the same number in 2023. Amid evolving expectations of increased numbers of publications, faculty may be focused on quantity over quality.
The second area of change is in behavioral content. Not surprisingly, the percentage of articles around infectious diseases and vaccines increased following the COVID-19 pandemic. Notably, the SARS outbreak in 2002 did not seem to spur research in health promotion journals. This may be due to the SARS outbreak being more quickly contained, being geographically limited, and occurring before today’s rapid, high-volume publishing systems existed. In contrast, COVID-19 became a prolonged global crisis in a very different scientific and publishing era. The decline in physical activity research is unexpected given the relatively small changes seen in population prevalence. 4 One possible explanation is the launch of physical activity-specific journals (eg, Journal of Physical Activity and Health launched in 2004) resulting in authors shifting their journal target. We also found a low number of published studies related to alcohol and other drugs. Drug overdose deaths accounting for over 100 000 deaths among Americans in 2023 with the rate quadrupling since 2003. Yet there was almost no increase in published papers in this area. 24
The findings show continued focus on historical behavioral topics such as tobacco use and less attention to emerging health behavior issues. This may be partially driven by funding priorities at the Federal Government or large philanthropic organizations. Also, the average age that investigators receive their first R01 is now over 41. 25 Investigators with decades of experience in a particular research area may be less likely to change the behavioral focus of their research based on public health needs. Flexibility in funding priorities based on emerging public health threats may help match the science with anticipated needs.
The third issue is study design. More than the third of published studies were cross-sectional. A high-rate and stability of cross-sectional studies shows little progress in advancing the rigor of studies in behavioral science. These studies are typically inexpensive to implement and easy to conduct but cannot assess causality or temporal order. 26 In contrast, the strongest design -- randomized clinical/community trials -- only account for 10% of papers in 2023, a 6% decrease from 20 years earlier. Quasi-experimental studies were almost absent from these journals. Given the limited number of papers published in these highly-rated health behavior journals, editors should consider prioritizing more randomized trials and quasi-experimental studies and deemphasizing cross-sectional studies.
The fourth area is behavioral theories. In 2003, the big 4 theories (Social Cognitive Theory, Health Belief Model, Theory of Planned Behavior/Reasoned Action and the Transtheoretical Model) were the most prevalently used theories/models in the field. In 2023, many unique theories or frameworks only appearing in one paper across all five journals emerged. Theories moving from broad to niche could have various explanations. While potentially suited for more specific behaviors and populations, these theories may be more applicable in intervention development. However, it is also likely that most scientists and practitioners will not be intimately familiar with the theory, and many theories may be used sporadically with limited efficacy testing. 27 There is a broad debate in psychology on the proliferation of theories. Critics argue that this increase has led to conceptual redundancy, whereby similar constructs are just labeled with different names. They also note that many newer theories are too complex for rigorous testing and application and go unused in research. 28 Some have argued that the proliferation of theories is a positive development that offers better specificity and adds opposing ideas for a broader perspective. 29 Despite the debate, the general consensus is that theories should be more specific in their pathways, prediction, and measurement, be falsifiable and be rigorously tested.28,30
The next area of interest was sample characteristics. Despite increasing diversification of the racial and ethnic identities of the U.S. population, there were more white participants in studies published in 2023 than in 2003. 31 Similarly, while the older adult population of the U.S. grew from just over 12% in 2003 to 17.7% in 2023, the proportion of studies published on this age group declined from 2003 to 2023. The rise in large data was notable, with 24 studies having a sample size of over 250 000 data points. Many of these were social media scapings, geolocation data, or large databases. The connection between these study types and potential behavior interventions is still emerging. There was also a decrease in studies conducted at the organizational level of the social ecological model. The number of studies in schools, healthcare settings, and worksites decreased between the two time periods. This decline may not necessarily reflect less interest in these settings, but rather a shift toward community, policy, and technology-based studies that are more scalable, fundable, and relevant in today’s environment. Due to the advances in technology and the swift shift to virtual implementation during COVID-19 pandemic, there was an increase in articles describing behavioral interventions delivered virtually. With continued technological advancement, a more detailed exploration of apps, telehealth, and online platforms is warranted.
This study has several limitations. First, only five health behavior journals were assessed. While these would be considered premier journals, they do not capture health behavior research published in behavior- or setting- specific journals (eg, Journal of Physical Activity and Health; Journal of School Health). Because we focused on comparing 2003 to 2023, any journals founded after 2003 were not included. Expanding this type of study to more journals or timepoints might require the use of artificial intelligence programs to help manage and assess the large number of published papers. Given that all the journals were ranked as first quartile journals at both time points, we would expect the papers in these journals to be of higher methodological rigor than papers published in other outlets. One could speculate that cascade journals (ie, Preventive Medicine Reports), which receive papers that were not deemed high enough impact by other journals, and publications ranked in the bottom three quartiles might be more likely to have high rates of cross-sectional studies and low rates of randomized trials. Sample size might also be lower and, given the high number of published papers in those journals, the proliferation of theories would also be expected. This hypothesis warrants further systematic assessment.
Conclusions
This study examined five quartile 1 health behavior focused journals at two time points twenty-years apart. Since only 2 years were assessed, we cannot definitively state which results are shocks to the system (ie, COVID-19 pandemic) vs those that appear to be long-term trends (ie, number of articles published per journal per year). We can review the findings and speculate on these and their implications for the field. First, the increase in the number of published papers both in these journals and throughout the field appears to be a strong trend. In many ways this is a positive development. Before the development on on-line publications, many journals had more than a year gap between acceptance of a paper and print publication. Print copies also limited access to research findings to institutions and individuals that had access. Today research is much more accessible globally. This rise in publications has almost made it difficult to keep up with the emerging literature, even within narrow topics. This has also affected the tenure and promotion process, exacerbating the constant pressure faculty feel to publish more articles rather than focus on quality.
This rise in large sample sizes also appears to be a trend. Big data sources including electronic medical records, cell phone scraping, monitoring of environmental conditions, and many others are now readily available. Artificial Intelligence is increasingly being harnessed to help with analysis of big data. While 20 years ago, there were concerns about null findings with small sample sizes, with many samples in research studies today topping 1 000 000 observations, a focus on effect size is imperative. 32
The proliferation of theories also appears to be a trend that is well documented in the psychological literature. The large number of infectious disease papers published in 2023 is most likely a shock related to the COVID-19 pandemic and the changes in the systems for publishing. A more in-depth analysis is needed to assess the decline in physical activity papers and the stability in tobacco-related papers is consistent across years, in conjunction with an assessment of papers in content-specific journals over time. The So What? section should be between 100-150 words. Individually list and answer the following questions: Health behavior research has grown over the past 2 decades with an increased focus on technological interventions. This article assesses changes in the research publications in five leading health behavior journals related to the number of articles published, study design, area of focus, sample size and use of theory. Editors should actively seek high-quality papers and strive to publish less correlational, cross-sectional studies. Funders should focus on both current and emerging health behavior issues and reassess funding for areas where the potential impact on population health is lower. Health behavior researchers should be actively engaged in the debate over theory proliferation and in multi-level, longitudinal research that involves truly representative samples and considers system and social influences.So what?
What is already known on this topic?
What does this article add?
What are the implications for health promotion practice or research?
Footnotes
Ethical Considerations
There are no human participants in this article and informed consent is not required.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
Data is available by request from the corresponding author.