Patterns and Predictors of Enactment of State Childhood Obesity Legislation in the United States: 2006–2009

Abstract

Objectives. We developed a content review for state policies related to childhood obesity, and we have quantitatively described the predictors of enactment.

Methods. We collected an inventory of 2006 through 2009 state legislation on 27 childhood obesity topics from legislative databases. We coded each bill for general information, topic content, and other appropriate components. We conducted a general descriptive analysis and 3 multilevel analyses using bill- and state-level characteristics to predict bill enactment.

Results. Common topics in the 27% of the bills that were enacted were community physical activity access, physical education, and school food policy. Committee and bipartisan sponsorship and having term limits significantly predicted enactment in at least 1 model. Bills with safe routes to school or health and nutrition content were twice as likely to be enacted. Bills containing product and menu labeling or soda and snack taxes were significantly less likely to be enacted.

Conclusions. Bipartisan and committee support and term limits are important in bill enactment. Advocacy efforts can be tailored to increase awareness and sense of priority among policymakers.

The increase in the prevalence of childhood obesity in the United States is well documented.^1–3 Surveillance estimates that 12.5 million children and adolescents are obese.⁴ Predictions of long-term outcomes of the obesity epidemic include a decline in population health because of disease and disability along with substantial societal and economic costs.⁵ In response to this increasing prevalence, there is a focus on identifying effective interventions to reverse trends by 2015.⁶ These intervention strategies include policy and environmental changes that are designed to provide opportunities, support, and cues to help people develop healthier behaviors.⁷

Policy changes, particularly at the state level, can influence access, social norms, and opportunities for better nutrition and increased physical activity among children and the general population.⁸ In recent years, states have introduced and adopted legislation that focuses either directly or indirectly on the prevention of childhood obesity.⁹ Childhood obesity policies include nutrition and physical activity topics¹⁰ and are applicable to school environments and the general community.⁹ Examples of state policies that influence physical activity include the mandate of quality physical education programs in schools as well as transportation policies that facilitate walking or reduce automobile–bicycling conflicts and result in increased bicycling.¹¹ Examples of state-level nutrition policies include school food and vending policies, farmers’ market guidelines, and soda taxes.^9,10

Although research interest in childhood obesity prevention legislation is increasing and we are gaining greater understanding of what constitutes evidence-based policy,^12,13 there remains much to be learned about the patterns of policy introduction and enactment over time and across states. Some recent studies have focused on the methods for evaluating state policies,^14,15 whereas others looked more specifically at factors influencing bill enactment. In a study of state legislation from 2003 to 2005, Boehmer et al. found that certain obesity prevention bill topics such as safe routes to school and statewide obesity taskforces or initiatives were more likely to be enacted than were others and that certain state characteristics such as a 2-year legislative session or Democratic control of both chambers can predict enactment.^9,10

We have described the quantitative portion of a larger multimethod project on childhood obesity prevention legislation, the State Childhood Obesity Policy Evaluation, the overall aim of which is to examine patterns and predictors of childhood obesity legislation at the state level through both qualitative investigation and quantitative bill content analysis. The State Childhood Obesity Policy Evaluation builds on previous methods and frameworks from the studies of Boehmer at al.^9,10 and expands the topic areas and years studied.

We sought to develop a descriptive content review for state policies relating to childhood obesity and to quantitatively describe the predictors of enactment of legislation on childhood obesity prevention.

METHODS

We used an aggregate, online legislative database (NetScan)¹⁶ to identify state legislation. The research team identified 26 legislative topic areas from the literature^9,10,17 and current obesity prevention recommendations.¹⁸ State legislation search topics included physical education, nutrition, and community trails. We developed Boolean search terms with the assistance of reference attorneys to best capture the most appropriate content in each topic area. We tested these search terms for relevance and modified them as appropriate.

We ran each search term for all 50 states for the years 2006, 2007, 2008, and 2009. We scanned all the bills returned by the search for relevance, defined as a new bill or a bill containing new language that strengthened the obesity prevention content area. The search identified 2016 bills, of which we used 1761 in the final analysis. We omitted bills for several reasons. We omitted bills prefiled for the 2010 session, amended bills that had no change in the content language of interest, and bills that merely mentioned the topic without any context.

Coding Bill Content

We developed a coding tool using methods from several past projects.^8,9,19–21 The tool included basic bill information (e.g., bill status, sponsor), topic content, and other components such as strength of bill language, funding, and enforcement language. To test the coding tool, 3 members of the research team coded the same 10 bills and discussed the process. They reached consensus on any disagreement and revised the coding tool and finally transferred it to an online data collection system. We coded each bill and entered the data into the online survey system, which facilitated data collection into 1 file from multiple team members.

Variables

The outcome of interest for this study was bill enactment status. We divided this variable into 2 categories: enacted or not enacted. We coded bills as enacted if they progressed through and passed both chambers (or the unicameral legislature in Nebraska) and the governor signed them into law. If the bill did not progress to this stage or the governor vetoed it, we labeled it as “not enacted.”

We selected state-level variables derived from analyzing past research and included additional variables hypothesized to influence obesity prevention legislation enactment. The variables included socioeconomic status variables (e.g., percentage of non-White population, index of income inequality, high school noncompletion rate), health variables (e.g., child and adult obesity, adolescent pregnancy rate), governmental variables (e.g., type of legislature, presence of term limits, party of governor and chambers), and other variables of interest (e.g., presence of a Centers for Disease Control and Prevention–funded Prevention Research Center in the state, governor communications about obesity). We conducted a preliminary analysis to identify variables with high correlations. Two variables (legislative session length and type of state legislature) were highly correlated (0.74), and thus we used only session length in our analysis.

We divided bill-level variables into 2 categories: general bill information and bill topic. General bill information included year introduced, originating chamber, sponsor information, appropriations, enforcement language, and whether the bill creates new legislation or revises or amends existing statutes or codes. We divided bill topics into 2 categories: school topics and community. These categories included the 26 search term topics, although we combined several in theoretical groupings (Table 1). For example, “community physical activity access” was a variable that contained park and recreation legislation and trail legislation. We created dichotomous yes or no variables for each topic. The bill content by topic was not mutually exclusive, as many bills contained more than 1 topic area.

TABLE 1—

Bill-Level Variables and Definitions: State Childhood Obesity Legislation, United States, 2006–2009

Variable	Definition
School topic
Farm to school	Supports the development or implementation of programs and initiatives for incorporating locally grown products into schools
School food policy	Related to food for school lunches or food served during the school day
Vending	Supports healthy vending in schools
PE	Related to time, instruction, or other PE topics in the school day
Safe routes to school	Related to programs, initiatives, or infrastructure to support safe routes programs or active transportation to and from school
Before and after school nutrition and PA	Supports nutrition or PA in programs at schools that provide before- or after-school care
Physical screening	Supports measurement (e.g., BMI) in schools
PA and recess	Related to PA outside PE in the school day; includes recess
Health and nutrition curriculum	Related to supportive changes in curriculum on health and nutrition topics
Community topic
Community food access	Supports community gardens, farmers markets, and grocery stores in communities
Bicycle and pedestrian safety	Related to infrastructure or programs and initiatives to facilitate safe active transportation in communities
Public transportation	Supports the development, improvement, or maintenance of public transportation programs in a state
Community PA access	Related to trails, paths, parks, or other recreation facilities that may increase community PA
Product and menu labeling	Supports posting nutritional information by restaurants or manufacturers
Soda and snack tax	Related to adding additional taxes to certain food or beverage items
Taskforce and initiative and council	Related to state-level committees on reducing the prevalence of obesity
Breastfeeding	Related to policies that facilitate breastfeeding among mothers
Childcare nutrition and PA	Supports healthy nutrition and PA regulations in childcare setting

Note. BMI = body mass index (defined as weight in kilograms divided by the square of height in meters); PA = physical activity; PE = physical education.

Statistical Analysis

We sought to identify important bill- and state-level characteristics associated with bill enactment and to describe the variations in the likelihood of bill enactment across states after accounting for these important bill- and state-level characteristics. We performed 3 separate analyses: overall pooled analysis (all bill topics), analysis for bills with school topics only, and analysis for bills with community topics only. We used similar methods for each of the 3 analyses with the 2-level mixed models.

To identify the important bill- and state-level characteristics, we adopted strategies proposed by Hosmer and Lemeshow²² in logistic regression modeling. First, we fit a simple bivariate mixed logistic regression model for each of the candidate bill- and state-level characteristics with a random intercept and a fixed parameter for each characteristic. We included characteristics with a liberal P value of .2 for further evaluation in multiple variable analyses with the mixed logistic regression model.

In multiple variable analyses, we entered bill-level variables before state-level variables. We retained the variables with the significant influence on the likelihood of bill enactment or with influence on variation of likelihood of bill enactment across states in the model. We deleted noisy variables, which increase the variation across the states without significant influence on likelihood of outcome.

We examined statistical significance of 1 variable in the presence of others by type 3 tests of fixed effect, and we quantified the magnitude and direction of influence of the variable on the outcome by odds ratio (OR) and its 95% confidence interval (CI). We tested the random effects (variance parameter) with likelihood ratio tests.

To describe variation of bill enactment across states, we fit a random intercept-only model and obtained a variance parameter estimate, and then we fit models with important bill- and state-level characteristics and made comparisons of the change in the variance parameter estimates. We used intercorrelation coefficients to quantify the variation across states with the latent variable method, which assumes that there is a latent continuous variable that describes the propensity for a state to enact the bills. The latent variable follows the logistic distribution with a unit scale parameter, yielding the variance of π²/3 = 3.29. Therefore the intercorrelation coefficient was V_s/(V_s + 3.29), where V_s was the variance parameter estimate. We performed all analyses using SAS version 9.2 software (SAS Institute, Cary, NC).

We conducted a reliability analysis on a sample of 86 bills (5% of sample). From an alphabetical list, 2 members of the research team selected every 20th bill and coded it. We analyzed the double codes using the κ statistic, which measures percentage agreement among responses after adjusting for chance.²³ For reference, we considered items with a κ score of greater than 0.75 to have excellent agreement, those with 0.75 to 0.40 fair to good agreement, and those with less than 0.40 poor agreement.²⁴ Interrater reliability was high for basic bill information coding with κ of 0.70 to 0.85. For bill topic content, κ varied. Of the 26 categories, 4 had complete agreement (1.00), 12 were 0.75 or higher, 8 were 0.40 to 0.74, and 2 were less than 0.40. We reanalyzed and recoded all bills containing the 2 topic variables with the lowest agreement (model policies and lunch). As a result, the research team clarified the topics by combining school lunch with school nutrition and model policies with the other topics in the bill.

RESULTS

From the study period, 475 of the 1761 introduced bills (27%) in our sample were enacted. The number of introduced bills ranged from 176 in New York to 2 in South Dakota, with an average of 35 per state. The range of state enactment was 76% in Arkansas to 0% in Kansas with an average of 34% (Table 2).

TABLE 2—

Number of Bills Introduced and Enacted, by State: State Childhood Obesity Legislation, United States, 2006–2009

State	Introduced, No.	Enacted, No. (%)
NY	176	22 (12.50)
MN	96	8 (8.30)
MA	88	13 (14.80)
HI	78	11 (14.10)
CA	77	31 (40.30)
NJ	71	7 (9.90)
MD	66	22 (33.30)
WA	55	16 (29.10)
FL	53	11 (20.75)
CT	49	14 (28.60)
TX	49	13 (26.50)
OK	46	19 (41.30)
MI	45	16 (35.60)
NC	43	9 (21.00)
IL	41	12 (29.30)
AL	40	11 (27.50)
OR	40	14 (35.00)
RI	40	16 (40.00)
TN	39	6 (15.10)
MS	38	4 (10.50)
IA	37	9 (24.30)
MO	37	7 (18.90)
VT	37	15 (40.50)
WV	37	6 (16.20)
NM	34	12 (35.30)
AR	30	23 (76.70)
KY	26	7 (26.90)
AK	24	9 (37.50)
ME	23	11 (47.80)
OH	23	10 (43.50)
VA	23	10 (43.50)
AZ	21	4 (19.50)
LA	21	13 (61.90)
CO	20	16 (80.00)
PA	20	8 (40.00)
WI	17	2 (11.80)
GA	13	4 (30.80)
SC	12	3 (25.00)
IN	10	1 (10.00)
KS	10	0 (0.00)
NH	9	5 (55.60)
MT	8	2 (25.00)
DE	7	4 (57.10)
UT	7	7 (100.00)
WY	7	4 (57.10)
ND	6	3 (50.00)
NE	6	2 (33.30)
ID	2	1 (50.00)
NV	2	1 (50.00)
SD	2	1 (50.00)

Table 3 shows the number of bills introduced and enacted by coding category. More bills were introduced in the houses (n = 1028) than in the senates (n = 727), with little difference in enactment rate between the 2 chambers (26.5% and 27.7%, respectively). There was no time trend in overall bill introduction or enactment over the 4-year study period. More bills were introduced with Democratic sponsorship (n = 1163), yet enactment was higher for bills introduced with Republican sponsorship (29.6% vs 22.9%). A greater percentage of bills with bipartisan sponsors (30.6%) and bipartisan cosponsors (34.4%) were enacted compared with those with single party sponsorship. Forty-five percent of bills with committee sponsorship were enacted. There was higher enactment in bills without strong enforcement wording (e.g., recommend, suggest, encourage; 31.8%) than with stronger wording (e.g., shall, require, mandate; 21.7%).

TABLE 3—

Frequencies of Variables, by Bill-Coding Category: State Childhood Obesity Legislation, United States, 2006–2009

Variable	Introduced, No.	No. Enacted, No. (%)
Originating chamber
House	1028	272 (26.50)
Senate	727	201 (27.70)
Legislative	6	2 (33.30)
Total	1761	475 (27.00)
Year introduced
2006	328	107 (32.60)
2007	513	134 (26.10)
2008	351	112 (31.90)
2009	569	122 (21.40)
Primary sponsor party
Democrat	1163	266 (22.90)
Republican	355	105 (29.60)
Both Republican and Democrat	36	11 (30.60)
Committee	200	90 (45.00)
Independent	7	3 (42.90)
Cosponsor party
Bipartisan	453	156 (34.40)
Single party	486	114 (23.50)
NA	822	205 (29.50)
Funding
Appropriations amount listed	523	218 (41.70)
Mentions funding, no amount	105	31 (29.50)
No funding mentioned	1133	226 (20.00)
Type of legislature
Professional (n = 10)	611	123 (21.60)
Intermediate (n = 23)	865	250 (28.90)
Citizen (n = 17)	285	93 (32.60)
Term limits
No (n = 15)	414	167 (40.30)
Yes (n = 35)	1347	308 (22.90)
Enforcement wording
Yes (shall, require, mandate)	1317	286 (21.70)
No (recommend, suggest, encourage)	88	28 (31.80)
NA (e.g., appropriations)	356	161 (45.22)
Obesity category
Nutrition	648	154 (23.80)
PA	839	226 (27.00)
Both	219	79 (36.10)
NA	55	16 (29.10)
School topics
Farm to school	108	34 (31.50)
School food policy	283	73 (25.80)
Vending	91	15 (16.50)
Physical education	297	64 (21.50)
Safe routes to school	81	34 (42.00)
Before- and after-school PA	33	12 (36.40)
Physical screening	87	18 (20.70)
School PA and recess	197	42 (21.30)
Health and nutrition education	88	29 (33.00)
Community topics
Community food access	182	79 (43.40)
Bicycle and pedestrian safety	191	48 (25.10)
Public transportation	132	57 (43.20)
Community PA access	376	132 (35.10)
Product and menu labeling	124	6 (04.80)
Snack and soda tax	38	2 (05.30)
Taskforce and council	186	70 (37.60)
Breastfeeding	95	22 (23.20)
Childcare nutrition and PA	17	5 (29.40)
Topics in bill for school, no.
0	952	257 (27.00)
1	473	136 (28.80)
2	257	67 (26.10)
≥ 3	79	15 (19.00)
Topics in bill for community, no.
0	615	149 (24.20)
1	987	246 (24.90)
2	126	66 (52.40)
≥ 3	33	14 (42.20)

Note. NA = not available; PA = physical activity.

The representation of school topics in the bill sample varied. The most prevalent content topics were physical education (n = 297) and school food policy (n = 283). Before- and after-school physical activity was the least represented topic (n = 33). The bills containing 3 or more school topics had lower enactment rates (19.0%) than did bills with just 1 topic (27.0%). None of the school topics showed time trends during the study period.

The most prevalent community topic in the bill sample was community physical activity access (n = 376), with the least prevalent being bills containing childcare nutrition and physical activity (n = 17). Enactment rates by community topic varied. The 2 topics with the lowest enactment rates were product and menu labeling (4.80%) and snack and soda tax (5.26%). Both bicycle and pedestrian safety and taskforce and initiative topics had a consistent upward time trend over the 4-year study period.

We first analyzed all bill- and state-level variables using binary analysis in which we included all variables in 3 different multilevel models. Only 4 state-level variables emerged with a P value less than or equal to .2 in all 3 models in this first step of analysis. These variables were political party of sponsor, funding, term limits, and high school noncompletion rate. Other state-level variables carried forward to the latent variable method in step 2 of the analysis included party of cosponsors (models 1 and 2), obesity topic and session length (model 1), and predominant party in the senate and non-White population (model 2). Neither adult nor child obesity rates were significant predictors in any of the 3 models, and thus we excluded them. Table 4 shows the ORs of the variables that we carried to the second step of the analysis in each of the models.

TABLE 4—

Multilevel Model Results: State Childhood Obesity Legislation, United States, 2006–2009

Model	Odds Ratio or Significant Odds Ratio (95% CI)
Model 1: general final model variables
Topic of bill
School topic (Ref)	1.00
Community topic	1.35a (1.02, 1.78)
Both categories	1.97a (1.30, 2.98)
Sponsor
Republican (Ref)	1.00
Democrat	0.69a (0.50, 0.96)
Committee	5.13a (3.04, 8.66)
Both Republican and Democrat	0.90 (0.38, 2.16)
Cosponsor
Single party (Ref)	1.00
Bipartisan	1.48a (1.06, 2.07)
Funding
No funding mentioned (Ref)	1.00
Mentions funding but no amount	2.93a (2.20, 3.91)
Amount of funding listed	1.49 (0.88, 2.51)
Obesity category
Nutrition only (Ref)	1.00
PA only	0.93 (0.64, 2.58)
Both	1.66 (0.70, 1.24)
Term limit
Yes (Ref)	1.00
No	0.43a (0.23, 0.80)
High school noncompletion rateb	1.28 (1.00, 1.61)
Session lengthb	0.98 (0.96, 1.00)
Model 2: school final model variables
Safe routes to school
Yes	2.02a (1.10, 3.71)
No (Ref)	1.00
Health and nutrition education
Yes	2.00a (1.09, 3.61)
No (Ref)	1.00
Sponsor
Republican (Ref)	1.00
Democrat	0.82 (0.49, 1.37)
Committee	8.16a (3.70, 18.10)
Both Republican and Democrat	1.16 (0.31, 4.20)
Cosponsor
Single party (Ref)	1.00
Bipartisan	1.83a (1.09, 3.09)
Funding
No funding mentioned (Ref)	1.00
Mentions funding but no amount	3.64a (2.33, 5.70)
Amount of funding listed	1.37 (0.61, 3.07)
Dominant party in senate
Republican (Ref)	1.00
Democrat	2.25a (1.07, 4.69)
Both Republican and Democrat	4.53a (1.01, 4.90)
Term limit
Yes (Ref)	1.00
No	0.49 (0.23, 1.02)
High school noncompletion rateb	1.19 (0.96, 1.47)
% Non-White populationb	0.97 (0.95, 1.00)
Model 3: community final model variables
Product and menu labeling
Yes	0.18 (0.07, 0.42)
No (Ref)	1.00
Soda and snack tax
Yes	0.20 (0.05, 0.88)
No (Ref)	1.00
Sponsor
Republican (Ref)	1.00
Democrat	0.69a (0.42, 0.97)
Committee	2.36a (1.26, 4.38)
Both Republican and Democrat	1.16 (0.26, 2.56)
Funding
No funding mentioned (Ref)	1.00
Mentions funding but no amount	2.15a (1.49, 3.11)
Amount of funding listed	0.96 (0.48, 1.92)
Term limit
Yes (Ref)	1.00
No	0.46a (0.23, 0.93)
High school noncompletion rateb	1.20 (0.96, 1.00)

Note. CI = confidence interval; PA = physical activity

^aSignificant odds ratio.

^bContinuous variable.

Model 1 included all bills categorized as school topics or community topics. In addition to the 4 variables common to all 3 models that were carried forward to the next step in analysis, we retained number of topics per bill and presence of nutrition topics, physical activity topics, or both as significant variables for step 2. In step 2, bills were more likely to be enacted if they had community topics or both school and community topics versus solely school topics. Sponsorship was a significant predictor in bill enactment. Bills were 5 times more likely to be enacted if sponsored by a committee (OR = 5.13; 95% CI = 3.04, 8.66) and less likely with Democrat sponsorship (OR = 0.69; 95% CI = 0.50, 0.96) compared with Republican sponsorship. Bills were also more likely to be enacted with bipartisan cosponsorship (OR = 1.48; 95% CI = 1.06, 2.07) than were bills with sponsors and cosponsors from a single party. Term limits remained a significant predictor in this model. Bills were less likely to be enacted if a state did not have term limits (OR = 0.43; 95% CI = 0.23, 0.80). High school noncompletion rates did not retain significance in step 2 of the model.

Model 2 included topic-specific school bills. Five topics (vending, physical education, safe routes to school, before- and after-school; also in left column nutrition and physical activity, health and nutrition curriculum) were statistically significant in the step 1 bivariate analysis. Of these, 2 remained significant predictors. If a bill contained safe routes, it was twice as likely to be enacted (OR = 2.02; 95% CI = 1.10, 3.71) than if it did not contain this topic. Similarly, bills containing the topic of health and nutrition content were twice as likely to be enacted (OR = 2.00; 95% CI = 1.09, 3.61). Committee and bipartisan sponsorship were also significant predictors. High school noncompletion rates and percentage of non-White population did not retain significance in step 2 of the analysis.

Model 3 contained community-specific bill content areas, 5 of which were significant in step 1 (community food access, public transportation, community physical activity access, product and menu labeling, and soda and snack tax). Of these 5, 2 retained significance. If the bills contained product and menu labeling (OR = 0.18; 95% CI = 0.07, 0.42) or soda and snack tax (OR = 0.20; 95% CI = 0.05, 0.88), they were much less likely to be enacted. As in models 1 and 2, bills with committee sponsorship were more likely to be enacted. As in model 1, bills with Democratic sponsorship were less likely to be enacted. If funding was mentioned with no stated amount, bills were more than twice as likely to be enacted (OR = 2.15; 95% CI = 1.49, 3.11). Presence of term limits was also a significant predictor in this model. As in model 2, high school dropout rate did not retain significance.

DISCUSSION

Our analysis of obesity prevention bills resulted in several important findings. First, the number of bills introduced from 2006 to 2009 with obesity prevention content is encouraging, as is the enactment rate of these bills. The enactment rate (27%) was considerably higher during this period than was the enactment rate during the earlier study, which analyzed bills from 2003 to 2005. Boehmer et al. found an enactment rate of 17% using similar obesity prevention topics and methodology.⁹ Another study on physical education legislation from 2000 to 2007 showed an enactment rate of 21%.²¹ In a study of community trails legislation, 29% of the bills were enacted during an 8-year time frame.²⁰ As a reference, overall bill enactment for all 50 states during the 2006 to 2009 study period was 17.1%.²⁵ Continued research on these categories of bills can establish the presence of an upward trend in enactment of obesity prevention legislation.

These findings can help inform advocacy for legislative obesity prevention efforts. For example, legislative support is necessary to ensure the passage of policy.¹⁵ Opponents of term limits suggest that representatives should serve for a long enough term that they would have the incentives and ability to pursue complex, long-term policies.²⁶ We hypothesized that support for obesity advocacy efforts would be developed as relationships with state legislators were nurtured over the course of their legislative careers and that term limits might be a barrier to developing this support. Although previous research found that shorter terms in state senates were correlated with a significantly lower number of bill introduced,²⁶ our findings suggest that having term limits predicted obesity prevention bill enactment. Studies in political science show that term limits can have a measurable impact on behavior and priorities.^27,28 Legislators in term-limited states may spend less time on activities related to reelection and more time on a broader focus on state interests, rather than those that are expressly constituent priorities.^27,29,30

Systemic consequences of term limits may include legislators who are less knowledgeable about both legislative process and policy matters and lessened power of the legislature relative to other actors in the policy process.²⁷ Advocacy groups can be an important resource to the legislators who have this knowledge gap. Term-limited legislators who support broader state interests can be of benefit to obesity prevention initiatives. Many topics of obesity prevention legislation are focused statewide, and it can be theorized that policymakers supporting these may be more willing to accept the challenging content and broader focus than would policymakers with no term limits.

The type of bill sponsor was a significant factor in enactment. As in earlier research,⁹ bipartisan and committee sponsorship predicted enactment. This is an important factor to consider when soliciting support for obesity legislation. Encouraging support from a wide range of policymakers may increase chances of bill enactment. Establishing this support is also important to bill implementation. Laws with bipartisan support are more likely to protect the initiative from change in the majority party.^31,32 Because many of the initiatives in our study need to be implemented long term to show effectiveness for obesity prevention, facilitating strong support and commitment from a variety of legislators may need to be an advocacy focus.

The topics included in the obesity prevention bills affected enactment. As in previous studies,^9,33 bills with safe routes to school content were more likely to be enacted than were other topics. One reason for this may be the awareness and support of the national Safe Routes to School program. This program, administered through the Federal Highway Administration of the US Department of Transportation,³⁴ has been successful in engaging states in the promotion of and improvements in active transportation. Additionally, federal funding (often with states matching funds) goes to each state for activities that fall in the realm of the Safe Routes to School program’s goals. Another bill topic that significantly predicted enactment was health and nutrition education. Many of these bills merely added curriculum components on nutrition and physical activity and did not require complex implementation or funding, which may have made them easier for legislators to support.

Product and menu labeling and snack and soda tax were 2 highly regulatory bill topics that were barriers to enactment. Although this may seem discouraging to advocates for these initiatives, the fact that these bills are introduced indicates that at least they are being considered in state legislation. There were 38 bills on this topic introduced from 2006 to 2009, an increase from only 10 bills introduced from 2003 to 2005.⁹ Proponents of these bills compare potential health effects to those occurring from increased tobacco regulation. Taxes on tobacco products have been highly effective in reducing consumption.^35,36 Similarly, soda taxes have been proposed as a way to decrease consumption, particularly in children.³⁷ Revenue from soda taxes has also been recommended as a source of funding for obesity prevention programs.³⁸ Similarly, the number of introduced product- and menu-labeling bills increased. Although advocates maintain that these bills promote increased decision making, less consumer confusion, and reduction in the toll from poor diets,³⁹ they can be viewed as highly regulatory, with complex implementation and governance issues. Increasing evidence on the benefits of both snack and soda tax and product and menu labeling for obesity prevention efforts and developing consistent implementation strategies may facilitate a cultural shift and legislative support. Advocates can be an important conduit for disseminating the evidence to legislators and promoting recommended strategies.^40,41

In addition to the significant predictors of enactment, the variables that were not significant warrant discussion. In our study and others,^9,33 state-level variables such as high school dropout rate and percentage non-White population did not significantly correlate with obesity prevention bill enactment. Neither childhood obesity rates nor adult obesity rates predicted enactment of obesity prevention bills in our study. We hypothesized that states with higher rates of both adult and child obesity would be motivated to enact legislation to address this issue. However, perhaps because many of the obesity prevention topics may be viewed as indirectly related to obesity (e.g., physical activity access), we did not find this correlation. National focus on childhood obesity prevention may also have increased the awareness of the problem, because all states showed an unhealthy increase in obesity prevalence during the past decade.¹

Limitations

Although this study contributes to the body of literature on state legislation and obesity prevention, several limitations warrant mention. First, we developed our bill search to capture a comprehensive list of bills on the basis of key search terms. Despite this, some bills may have been missed (e.g., bills with terms or combinations of terms not in our search string). Additionally, appropriations bills that included already codified programs may use section or program numbers from previous legislation instead of the terms for which we searched. However, by using expert input in developing the search terms and using bills from a 4-year period, we compiled a very broad list of bills for analysis.

Second, we did not examine state statutory law that reflects the codification of all enacted laws. Thus, it may be that although states did not enact legislation on many topics included in our study, they already had statutory laws.

Third, several states had a small number of bills that might have limited the identification of state-level predictors of enactment.

Conclusions

We have highlighted the complex and multifaceted strategies for obesity prevention and findings that can inform researchers, advocates, and policymakers in facilitating enactment. There are many potentially influential legislative characteristics to consider when advocating these initiatives. Aspects of the bill and the legislature can affect enactment. This is important to consider in bill development and promotion. Because awareness and support from legislators is a key piece in initiative success, advocacy efforts should focus on promoting the cause to a wide variety of state legislators to increase long-term bipartisan support.

Some obesity prevention bill topics were more likely to be enacted than were others. This does not mean that efforts should be focused on only those more likely to be enacted, but work should also be done to identify and disseminate model legislation and effectiveness research on topics that were less likely to be enacted. Also, more focus should be placed on the development and advocacy for evidence-based legislation. For some topics (e.g., physical education), there are policy elements that are known to be effective, yet the evidence base for other topics is still emerging. Focusing on what is known or likely to be effective should be a priority.

Childhood obesity is a national concern, as indicated by federal efforts, media campaigns, and political interest. Identifying and applying effective strategies to influence the legislative process may increase the success of obesity prevention legislation.