This study of childhood obesity prevention legislation includes both bills and resolutions. A bill is a proposed new law or amendment to an existing law that is presented to the legislature for consideration. To become law, bills require approval by both chambers of the legislature and by the governor. (Bills can be enacted with or without the governor's signature as long as they are not vetoed.) Bills may appropriate money, prescribe fees or penalties, repeal existing law, or take other action. A resolution is a formal expression of the will, opinion, or direction of one or both chambers of the legislature on a matter of public interest. Simple resolutions require approval only by one chamber; concurrent and joint resolutions require approval by both chambers. In general, resolutions require no action by the governor and do not have the force of law.

Different terminology is used to describe the final approval of a bill or resolution in the legislative process. For example, enact means to establish by law and refers to the final approval of bills, whereas adopt means to approve or endorse and is usually applied to amendments and resolutions (but not bills). To simplify the terminology used in this study, adoption was defined as a favorable final action (i.e., approval in the last stage of the legislative process) for both resolutions and bills. Consistent with the definitions provided above, adoption was defined differently for simple resolutions (approved in the chamber of origin), joint and concurrent resolutions (approved in both chambers), and bills (approved in both chambers and enacted into law).

We used a legislative database created by Netscan's Health Policy Tracking Service (HPTS) (16) to identify state legislation affecting nutrition, physical activity, and obesity prevention introduced in all 50 states from 2003 through 2005. HPTS performed a legislative scan for 2003 and 2004 using the same search criteria that were previously developed for their 2005 report on state nutrition, activity, and obesity legislation (17). HPTS performed separate searches on 24 topic areas (e.g., farmers' markets, nutrition standards and vending machines, BMI reporting, safe routes to school), so it was possible for a single bill or resolution to be listed in more than one topic area.

The legislative scan identified 1149 bills and resolutions (including simple, joint, and concurrent resolutions) introduced from January 1, 2003, through December 31, 2005. We excluded bills that were merged with or substituted by a similar bill that was subsequently enacted (n = 35), resulting in 965 bills and 149 resolutions for further consideration. We reduced the number of topic areas from 24 to 18 by combining similar categories and categories with small numbers. The 18 topic areas were categorized as relevant or irrelevant to childhood obesity prevention. Four topic areas were excluded because of irrelevance: 1) labeling of genetically modified food products, 2) insurance coverage of gastric bypass surgery, 3) Medicaid coverage of obesity-related treatments, and 4) restrictions on civil liability lawsuits related to obesity and food consumption.

The 14 relevant topic areas were further categorized as school-related or community-related. The 813 bills and 144 resolutions within these topic areas were examined in more detail to ensure their applicability to childhood obesity (yes or no) and direction of health impact (positive, negative, or unsure). The task of coding was divided among four members of the research team. Eighty bills were coded in duplicate to assess interrater agreement. Agreement between raters was 89% for applicability (n = 80) and 94% for health impact (n = 63) (health impact was coded only for applicable bills). We excluded from further consideration bills and resolutions that were coded as not applicable (e.g., specific to senior citizens, concerning sex education in schools) or as having a negative health impact (e.g., repealing of BMI reporting, allowing exemptions for physical education). After removal of 78 bills and 10 resolutions that were not applicable and 18 bills with a negative health impact, this study reviewed 717 bills and 134 resolutions. Legislative history was reviewed to determine whether each bill or resolution was adopted as of December 31, 2005.

A descriptive analysis was performed to examine patterns in the introduction and adoption of legislation by time, topic area, and geography. Patterns over time were described by comparing data for 2003 and 2005. Because of differences in the frequency and length of legislative sessions, fewer bills and resolutions are introduced in even years. (Six states have biennial sessions that meet only in odd years; among states that meet annually, 25 have 2-year sessions that begin in odd years [e.g., 2003–2004].)

The number introduced and adopted and percentage adopted were calculated separately for bills and resolutions for each of the 14 relevant topic areas and for each of the 50 states. In addition to quantity, the number of topic areas covered (possible range, 0-14) through introduced and adopted legislation was assessed to measure the breadth of approaches addressed within each state.

Next, we examined the geographical patterns of introduced legislation and topic areas covered by introduced legislation and compared them with obesity prevalence. Three U.S. maps were created with each variable of interest categorized into quartiles. Childhood obesity prevalence estimates were not available for all 50 states, so we used adult obesity prevalence from the 2004 Behavioral Risk Factor Surveillance System (18) as an indicator for childhood obesity prevalence. The association between adult obesity prevalence and childhood obesity legislation (number of introduced bills and resolutions) was examined dichotomously (low and high) using Pearson chi-square testing and was examined by quartiles (1 to 4) and rank order (1 to 50) using Spearman rank correlation. Analyses were performed using SPSS version 14.0 (SPSS Inc, Chicago, Ill).

We found that the number of bills and resolutions introduced and adopted increased from 2003 through 2005. Given this short time frame, there is a need for continued monitoring of nutrition, physical activity, and obesity prevention legislation to assess trends over time. Legislative tracking services and surveillance systems will be useful in all phases of policy research. As part of surveillance, it is important to develop a standardized method for identifying and cataloging legislation. This will likely prove to be a difficult task given the wide range of topic areas that fall under the umbrella of obesity prevention (e.g., urban development, transportation, farmers' markets, task forces, school nutrition, advertising).

A few tracking resources are available to the public for monitoring the introduction of obesity prevention legislation, including the Centers for Disease Control and Prevention's (CDC's) Nutrition and Physical Activity Legislative Database (19) and the National Conference of State Legislature's (NCSL's) Healthy Community Design Legislation Database (20). However, we found low concordance between the HPTS, CDC, and NCSL legislative databases in terms of the number of bills introduced and how they were categorized. A single, standardized database of all introduced legislation will assist with future research for identification of relevant policies and determinants of policy adoption.

We found that legislation within certain topic areas was more likely to be adopted than others. Additional research is needed to identify characteristics of bills that are adopted. For example, it may be that legislation related to statewide initiatives, studies, and task forces may be easier to pass because of the limited amount of resources necessary to implement such laws, and resource-intensive and revenue-restricting bills may be more difficult to pass. Understanding bill characteristics that are associated with adoption may assist with the development of model legislation and lead to more successful advocacy efforts.

Geographic comparisons also demonstrated wide variation among states in the amount of legislation introduced and the proportion of legislation adopted. Although regional geographic patterns were identified, no statistical link was found between legislative activity and adult obesity prevalence. Future research should examine why some states are more likely to introduce and adopt childhood obesity prevention legislation than other states. For example, state-level political, economic, and sociocultural factors may affect legislative priorities within state governments. As a follow-up to this study, we plan to conduct a quantitative, multilevel analysis to examine both bill-level and state-level factors associated with bill adoption. Another possibility is to conduct qualitative case studies of states that are considered high or low adopters.

Certain topic areas (e.g., school nutrition, task forces) were more commonly introduced than others and may represent early steps in the development of obesity prevention policies. For example, vending machine restrictions were first considered and adopted in California, setting an example and providing momentum for other states to follow. Additionally, bills and resolutions related to statewide initiatives, studies, and task forces likely represent a first, capacity-building step in a process leading toward more comprehensive programs and policies. Future research should investigate the process of policy development as it relates to childhood obesity prevention. This may include establishing a way to measure a state's level of readiness for developing and implementing childhood obesity prevention legislation, as well as outlining stages of progress.

The extent to which evidence guides obesity prevention policies is also important to evaluate. Researchers often assume that evidence guides policy development. However, policymakers are influenced by multiple domains (e.g., social, media-related, economic). Assessing the extent to which current policy initiatives are guided in development by multiple forms of data and the role evidence plays in that process are critical in understanding effective evaluation of policy impact (21).

Surveillance of enacted legislation will promote research on policy quality, implementation, and effectiveness at achieving desired health outcomes. For example, the National Cancer Institute's State Cancer Legislative Database program maintains a public-use, searchable database of adopted legislation related to several types of cancer (e.g., breast, cervical, prostate, skin) as well as access to care, genetics, surveillance, and tobacco control (22). This database has been used in several analytic studies to evaluate the scope and quality of enacted policies and their impact on health behaviors and outcomes, such as youth access to tobacco and clean indoor air laws (23-26). As policies related to obesity prevention continue to be introduced and adopted, a database of enacted legislation should be developed to assist with future studies examining the impact of policy on outcomes related to energy balance.

The findings of this study are subject to at least four limitations. First, the limited time period of the study prohibited the examination of trends over time. Extension of the time period was not possible, and childhood obesity-related legislation introduced before 2003 was limited. Second, this study may slightly underestimate the proportion of bills adopted, because legislation introduced in 2005 may have been carried over and adopted in 2006 in the 25 states that have 2-year legislative sessions (2005–2006). Third, the quality of data in the HPTS legislative database depends upon information available from state Web sites. Therefore, some information about current bill status (adopted or not) may have been incomplete or out-of-date. Finally, the identification and categorization of bills within topic areas were based on HPTS search criteria, which likely differ from that of other agencies (e.g., CDC, NCSL). Unfortunately, none of the available legislative databases have been formally validated or compared with one other. As a result, the quality of the HPTS legislative database is unknown, both in terms of completeness (i.e., amount of legislation introduced) and accuracy (i.e., classification of legislation into topic areas).

This study is an initial attempt to develop policy-relevant data on childhood obesity. This information can be powerful in assessing progress, identifying effective approaches, and supporting advocacy efforts to address the problem. As such, there are several implications for public health practitioners:

The process of policy development involves three key criteria: 1) sufficient evidence base, 2) development of effective coalitions, and 3) commitment of policy makers (10). Although the knowledge base for successful programs and policies is limited (8) and movement toward social consensus and public action is just beginning (29), our study shows considerable adoption of legislation targeting childhood obesity. This suggests a growing desire and dedication among state legislators. Expanded policy surveillance (including standardized identification and cataloging) of introduced and adopted legislation will enhance our ability to track progress and identify effective approaches. Future policy research should examine the determinants, implementation, and effectiveness of legislation to prevent childhood obesity.