Legislators’ Behavioral Health Research-Seeking Practices and Dissemination Preferences: Variations by Political Party

Abstract

Objective:

This study sought to characterize the behavioral health research-seeking practices and dissemination preferences of US state legislators and assess differences by political party.

Methods:

A cross-sectional survey of state legislators was conducted in 2017 (N=475). Items assessed where legislators seek, and the most important features of, behavioral health research. Bivariate analyses and multivariate logistic regression were conducted.

Results:

Advocacy organizations (53%), legislative staff (51%), and state agencies (48%) were identified as sources of behavioral health research most frequently. Universities were only identified by 27% of legislators and significantly more frequently by Democrats than Republicans (34% vs. 19%; aOR=1.79, 95% CI=1.13–2.85). Data about budget impact and cost-effectiveness were rated as “very important” most frequently, but significantly less frequently by Democrats than Republicans (77% vs 87% and 76% vs 89%, respectively).

Conclusions:

To reach legislators and satisfy their information preferences, behavioral health researchers should target diverse audiences, partner with intermediary organizations, and craft messages that include economic evaluation data.

There are 7,383 state legislators in the United States and their decisions have tremendous influence on mental health and substance use outcomes (hereafter referred to as behavioral health). For example, through decisions related to insurance coverage and workforce development, legislators influence access to behavioral health services.¹ Through reallocation of tax revenue (e.g., supportive housing) and regulation of behavior (e.g., recreational marijuana laws), legislators influence the social determinants of behavioral health.² For these reasons, state legislators have been identified as an important audience to target when disseminating behavioral health research findings.^3,4 However, as demonstrated by a 2016 systematic review,⁴ there is little empirical guidance about how to most effectively disseminate behavioral health research to state legislators, or policymakers more broadly.

At least three important knowledge gaps exist. First, little is known about where legislators seek behavioral health research when making policy decisions. Legislators typically lack in-depth expertise on specific issues (e.g., behavioral health) and seek research as issues emerge on policy agendas.⁵ Universities are major producers of behavioral health research and are largely perceived as credible,⁶ but studies suggest that legislators might not perceive universities as possessing information that is relevant to the policy decisions they face.⁷ An understanding of where legislators turn for behavior health research can inform researchers’ dissemination strategies.

Second, it is not clear which features of behavioral health research are most important to legislators. Such features might include the format in which findings are presented (e.g., a concise policy brief) and the information provided about implications (e.g., budget impact). A 2012 survey found that state legislators who prioritized behavioral health issues had stronger opinions about features of research than legislators who did not prioritize these issues.⁶ While this work, and other studies,⁵ indicate legislators’ preferences for research in general, little is known about their preferences for behavioral health research in particular. This distinction is important because misconceptions about behavioral health issues, and stigma towards people with behavioral health conditions, are pervasive.^8,9 Furthermore, information about legislators’ specific preferences for behavioral health research can provide the most concrete guidance for behavioral health advocates.

Third, studies of legislators’ preferences for disseminated research have not examined differences by political party affiliation. Polarization is increasing in US state legislatures¹⁰ and it is plausible that Democrat and Republican legislators seek behavioral health research from different sources and value different features of research. Such differences could have important implications because opinions about behavioral health issues, and policy solutions to address them, vary by political party. Republicans generally have more stigma towards people with behavioral health conditions than Democrats and are also less supportive of government spending on behavioral health.^8,9 Information about how Democrat and Republican legislators’ preferences for behavioral health research differ could inform dissemination strategies (e.g., tailored materials based on party affiliation).

This exploratory study was conducted to inform dissemination strategies that aim to enhance the use behavioral health research by state legislators. The objectives were to: 1) determine where state legislators seek behavioral health research, 2) identify the features of behavioral health research that are most important to legislators, and 3) assess differences in these research-seeking practices and dissemination preferences between Democrats and Republicans.

METHODS

A cross-sectional, multi-modal (post-mail, e-mail, telephone) survey of state legislators was conducted between March-September 2017. Complete details about the recruitment process and sample frame are provided in the published study protocol.¹¹ In short, the survey was sent to a state-stratified random sample of 2,902 legislators who were in office as of January 15, 2017. Each legislator was post-mailed an invitation to complete the survey online, sent two paper versions of the survey, called up to 15 times, and received ten e-mail invitations to complete the survey online.

The survey was completed by 475 legislators. The response rate was 16.4%, which is acceptable for state legislators and higher than the response rates of recent legislator surveys.^12,13 Respondents were statistically significantly more likely than non-respondents to be Democrat (49% vs. 42%, p=.001), female (33% vs. 23%, p<.001), and from the Midwest (31% vs. 23%, p<.001). To adjust for these differences, a post-stratification approach was used in which sample non-response weights that accounted for political party, gender, and geographic region were calculated and applied. Institutional Review Board approval was obtained.

Legislators’ behavioral health research-seeking practices and dissemination preferences were assessed using items adapted from Bogenschneider’s studies of state legislators.⁵ The sources from which legislators seek behavioral health research were assessed by asking “who they would turn to” if they “were going to seek out mental health/substance abuse research to make a policy decision” and instructing them to select up to three sources from a list of nine options. The features of behavioral health research that legislators perceived as most important were assessed by asking legislators to rate “how important would it be” for “mental health/substance abuse research [they receive]” to have seven features on a five-point Likert scale (1= not important, 5= extremely important). Ratings of 1, 2, and 3 were coded as “not very important” and ratings of 4 and 5 were coded as “very important.”

Political party affiliation was the primary independent variable and obtained from the National Conference of State Legislatures (NCSL). Information on gender was also obtained from NCSL and information on highest level of education was obtained through the survey and coded as ≤ college degree or ≥ postgraduate degree.

Descriptive statistics characterized the sample and χ² tests were used to examine differences between Democrats and Republicans. We did not include Independent and nonpartisan legislators in comparisons because they only accounted for 2% of the sample. Multivariate logistic regression was used to produce adjusted odds ratios (aORs) of a legislator selecting each source of research and rating each feature of research as “very important” given their political party, controlling for gender and education.

RESULTS

When all legislators were analyzed together, behavioral health advocacy organizations (53%), legislative staff (51%), and state behavioral health agencies (48%) were most frequently identified as sources legislators turn to for behavioral research (Table 1). However, the proportion of legislators identifying advocacy organizations was significantly higher among Democrats than Republicans (65% vs. 40%, aOR=2.25, 95% CI=1.49–3.39). Only 27% of legislators identified universities as a source they would turn to for behavioral health research. The proportion of legislators who identified universities was significantly higher among Democrats than Republicans (34% vs. 19%, aOR=1.79, 95% CI=1.13–2.85). Industry was identified as a source least frequently among all legislators (11%), but the proportion was significantly lower among Democrats than Republicans (6% vs. 16%, aOR=.33, 95% CI=.16.68).

Table 1:

Behavioral Health Research-Seeking Practices and Dissemination Preferences among US State Legislators, Stratified by Political Party, 2017, N= 475

	All		Democrats		Republicans		χ2 p-value1	aOR2	95% CI
	n	%	n	%	n	%
Sources where legislators seek behavioral health research
Legislative staff	228	51	129	56	99	45	.027	1.45	.97–2.17
Advocacy organizations (e.g., NAMI)	238	53	151	65	87	40	< .001	2.25	1.49–3.39
State mental health/substance abuse agencies	218	48	104	45	114	52	.125	0.71	.47–1.05
Legislator assistance organizations (e.g., NCSL)	169	38	79	34	90	41	.122	0.75	.50–1.13
Mental health/substance abuse societies (e.g., APA)	135	30	72	31	63	29	.599	1.01	.65–1.56
University researchers	121	27	79	34	42	19	< .001	1.79	1.13–2.85
Industry (e.g., insurance or pharmaceutical companies)	49	11	13	6	36	16	< .001	0.33	.16- .68
Features of behavioral health research rated as “very important”3
Provides data on budget impact	365	82	176	77	189	87	.007	0.45	.27–.76
Provides data on costeffectiveness	368	82	174	76	194	89	< .001	0.30	.17–.54
Presented in a brief, concise way	362	82	185	81	177	81	.913	0.87	.52–1.47
Relevant to my constituents	340	76	176	77	164	75	.687	1.04	.65–1.66
Tells a story of how an issue affects my constituents	300	67	162	71	138	63	.082	1.30	.85–1.99
Delivered by someone I know or respect	278	63	138	60	140	65	.322	0.79	.52–1.20
Presents implications that are politically feasible	229	52	117	51	112	51	.990	0.86	.57–1.29

aOR= adjusted odds ratio. CI= confidence interval. NAMI= National Alliance on Mental Illness. NCSL= National Conference of State Legislatures. APA= American Psychological Association.

^1.χ² comparing Democrats and Republicans. Df= 1

^2.Binary logistic regression adjusting for gender and education. Each source and feature was the dependent variable in a separate model, reference category= Republicans

^3.Importance rating of 4 or 5 on 5-point Likert scale

Among all legislators, the features of behavioral health research that were rated as “very important” most frequently were the inclusion of data on budget impact (82%), cost-effectiveness (82%), and research being presented in a brief, concise way (82%). However, the proportion of legislators who identified budget impact data as “very important” was significantly lower among Democrats than Republicans (77% vs 87%, aOR=.45, 95% CI=.27-.76) as was the proportion who identified cost-effectiveness data as “very important” (76% vs 89%, aOR=.30, 95% CI=.17-.54). Political feasibility was identified as “very important” least frequently among both Democrats (51%) and Republicans (51%).

DISCUSSION

This study characterized state legislators’ behavioral health research-seeking practices and dissemination preferences and found that some of these varied significantly between Democrats and Republicans. Democrats most frequently identified advocacy organizations as a source they turn to for behavioral health research and Republicans most frequently identified state behavioral health agencies. Among both Democrats and Republicans, legislative staff were identified second most frequently. This finding is consistent with prior studies which indicate that staffers are an important audience to target when disseminating research to legislators⁵ and evidence suggesting that intermediary organizations play a critical role in translating health research into policy.¹⁴

Universities were not a primary source from which legislators reported seeking behavioral health research, especially among Republicans. The proportion of Republican legislators that identified universities (19%) as a source was similar to the proportion of Republicans that identified industry (16%). These findings are consistent with prior research which suggests that many legislators perceive the evidence produced and possessed by university researchers as lacking relevance to “real world” policy decisions.^5,7 However, these findings are interesting when contrasted with the results of a 2012 survey of state legislators which found that universities were rated as the most credible source of research.⁶ Considered together, these findings support the notion that legislators seek research from sources that are perceived as having information that is most relevant to the policy decisions they face, even if these sources are not perceived as most credible.⁷ This underscores the importance of university-based behavioral health researchers investigating policy-relevant questions. These findings also suggest that strategies supporting the dissemination of research to trusted intermediaries may be particularly impactful as legislators turn to these sources when making policy decisions.

Data on budget impact and cost-effectiveness were the features of behavioral health research that legislators most frequently identified as “very important.” This highlights the value of conducting economic evaluations of behavioral health interventions that are relevant to contemporary legislative decisions. However, there is a paucity of such studies. For example, a 2016 review of economic evaluations of opioid use disorder interventions found ample evidence about the cost-effectiveness of methadone maintenance treatment, but relatively few economic evaluations focused on opioid interventions that are being considered in state legislatures (e.g., increasing access to buprenorphine and naloxone).¹⁵ The finding that data on budget impact was frequently rated as “very important” signals the need to conduct behavioral health economic evaluations from narrow state government perspectives (e.g., Medicaid, criminal justice) in addition to broader societal perspectives.

Inclusion of economic evaluation data could be especially important when disseminating behavioral health research to Republican legislators. Republicans were significantly more likely than Democrats to identify cost-effectiveness and budget impact data as “very important” and public opinion surveys indicate that Republicans are less supportive of spending on behavioral health services than Democrats.^8,9 Thus, economic evaluations that demonstrate cost-savings could foster support among Republican legislators for laws that increase access to behavioral health services.

Our study has limitations. Although a response rate of 16% is higher than response rates of recent legislator surveys,^12,13 it is low compared to standards for patient and provider surveys. Information about the political party, gender, and geographic region of non-respondents allowed for sample non-response weights to be calculated and applied. This increases confidence that the survey results approximate the behavioral health research use practices and dissemination preferences of the larger population of state legislators, but it is plausible that legislators who completed the survey were systematically different than those who did not across these domains.

CONCLUSIONS

Entities who want behavioral health research to reach and influence state legislators should disseminate findings to diverse sources—specifically legislative staff, behavioral health advocacy organizations, and state behavioral health agencies. Dissemination materials should be concise and, when possible, contain information about cost-effectiveness and budget impact.