Assessment of Public Opinion on Transparency at the US Food and Drug Administration.

Introduction

Historically, the US Food and Drug Administration (FDA) has considered much of the oversight of clinical trials, interactions with companies, and review processes to be confidential.[1] In recent years, however, the agency has faced calls for greater transparency, including in its COVID-19 pandemic response[2] and use of expedited review pathways.[3] To our knowledge, the views of the US public on transparency at the FDA have not been previously reported.

Methods

We conducted a nationally representative, cross-sectional survey of US adults from June 14 to July 2, 2021, in English and Spanish. The survey was based on a general population sample from the AmeriSpeak Panel of NORC (formerly known as the National Opinion Research Center) at the University of Chicago, which covers more than 97% of the US population.[6] Participants provided informed consent to participate in the AmeriSpeaks panel, and the study was approved by the institutional review board of the Johns Hopkins Bloomberg School of Public Health. The study followed the American Association for Public Opinion Research (AAPOR) reporting guideline on the conduct of survey studies (eFigure 1 in the Supplement).

The survey assessed public opinion about 10 potential transparency measures and was developed to be generally understandable using a 2017 report[4] and other recommendations[5] as the foundation (Table 1). Using a 5-point Likert scale, survey participants could express their level of support (with 1 for strongly agree or 2 for somewhat agree) or opposition (with 5 as strongly disagree or 4 as somewhat disagree) (eFigure 2 in the Supplement).

Table 1.

Statements on Transparency at the FDA

Statement No.	Statement
1	The FDA should share with the public the reasons for why medications are not approved
2	The FDA should share with the public the name of the company and medication when companies seek to conduct studies on human subjects
3	The FDA should share with the public when studies are suspended as well as the reasons for suspending a study
4	If the FDA permits a suspended study to restart, the FDA should share with the public the reasons for allowing the study to restart
5	The FDA should share with the public the company and medication name when applications for generic medications are filed
6	The FDA should share with the public when the agency speeds up a review as well as the scientific basis for this action
7	The FDA should share with the public how special safety measures put in place for certain medications are working
8	The FDA should publish a list of medications that drug companies have stopped developing
9	The FDA should share with the public the agency’s analyses of medications that drug companies have stopped developing
10	If drug companies provide false or misleading information about the development of medications to the public, the FDA should share the correct information with the public

Abbreviation: FDA, US Food and Drug Administration.

We conducted subgroup analyses by sex, geographic region, race and ethnicity (which were self-reported in these categories: Asian or Pacific Islander [non-Hispanic], Black [non-Hispanic], Hispanic, White [non-Hispanic], and more than 2 [non-Hispanic] or other race and ethnicity), educational status, political ideology, and voting in the 2020 presidential election. After reviewing the data to identify a threshold for outliers, we identified subgroups whose views of FDA transparency differed from those of the rest of the US population by more than 15% in at least one-half of the measures. Statistical calculations were performed in R, version 3.6.0 (R Foundation for Statistical Computing).

Results

A total of 4002 US adults were invited to participate, of whom 1035 (25.9%) completed the survey either online (n = 962) or by telephone (n = 73). The unweighted survey population had a mean (SD) age of 51.0 (17.1) years; included 476 men (46.0%) and 559 women (54.0%); and identified under Asian or Pacific Islander (19 [1.8%]), Black (94 [9.1%]), Hispanic (143 [13.8%]), White (730 [70.5%]), more than 2 (35 [3.4%]), or other (14 [1.4%]) race and ethnicity categories. The margin of error was ±4.28%. Study participants closely matched the demographic characteristics of US adults by age, race and ethnicity, sex, and educational status. The sample, compared with the general population, included a smaller proportion of people with higher household income (≥$125 000); 12.2% vs 27.1%) and with children (16.5% vs 32.5%).

Support for all transparency measures was high. Specifically, 86.2% of participants supported the disclosure of the reasons medications are not approved, 66.7% supported the release of basic information about medications in development, 77.6% supported the disclosure of the reasons clinical studies are placed on hold, and 77.2% supported the disclosure of the reasons studies are permitted to resume. There was strong support for the FDA disclosing basic information on pending applications for generic drugs (71.6%), explaining the reasons medications receive expedited review (80.5%), and sharing whether safety programs are working (83.2%). When asked whether the FDA should list medications that drug manufacturers have stopped developing, 67.6% of participants said yes, with 65.9% supporting the release of relevant FDA analyses. In addition, 90.7% of participants supported the FDA correcting the misleading information spread by drug manufacturers.

Support for FDA transparency exceeded opposition among participants from different sexes, races and ethnicities, geographic regions, and political ideologies (Table 2). Participants in 3 demographic groups expressed at least 15% less support for transparency than the rest of the population for at least half of the measures: those with less than a high school education (8 measures), those who did not vote in the 2020 election (7 measures), and those whose political party identification was “don’t lean, independent, or none” (5 measures). Less support was associated with increases in the neither-agree-nor-disagree response rather than opposition.

Table 2.

Support for and Opposition to Specific Statements About US Food and Drug Administration Transparency Among Representative Sample of US Adults

Variable	Percentage	Percentage
		Statement 1		Statement 2		Statement 3		Statement 4		Statement 5		Statement 6		Statement 7		Statement 8		Statement 9		Statement 10
		Support	Oppose	Support	Oppose	Support	Oppose	Support	Oppose	Support	Oppose	Support	Oppose	Support	Oppose	Support	Oppose	Support	Oppose	Support	Oppose
Total	100.0	86.2	4.2	66.7	10.9	77.6	9.0	77.2	7.0	71.6	6.6	80.5	3.4	83.2	3.6	67.6	10.5	65.9	11.5	90.7	0.6
Sex
Male	48.2	86.1	3.5	70.2	10.8	78.6	8.7	77.6	8.7	73.0	7.1	82.6	4.7	82.6	4.7	67.7	10.7	67.6	10.6	90.5	0.4
Female	51.8	86.4	4.9	63.4	10.9	76.6	9.2	76.8	5.3	70.4	6.2	78.4	2.2	78.4	2.2	67.5	10.2	64.4	12.2	91.0	0.7
Age group, y
18-29	20.2	84.1	4.0	71.9	5.7	79.5	11.2	76.4	7.3	59.7	5.5	73.4	6.5	73.4	6.5	64.0	12.8	71.1	9.2	83.8	0.4
30-44	25.3	88.4	1.5	59.3	8.7	73.7	5.8	72.9	4.3	67.7	10.1	76.4	4.2	76.4	4.2	71.0	6.1	63.5	8.7	83.7	1.0
45-59	24.2	85.5	6.4	73.5	12.3	77.7	7.2	77.6	6.8	73.1	4.6	86.1	2.3	86.1	2.3	68.1	9.2	66.8	8.0	94.4	0.4
≥60	30.3	86.4	5.0	63.9	14.9	79.5	11.6	80.9	9.2	81.7	6.1	84.0	1.6	84.0	1.6	66.9	13.5	63.9	18.0	98.3	0.5
Race and ethnicityb
Asian or Pacific Islander (non-Hispanic)	4.4	80.2	0	65.6	4.7	71.0	12.8	69.3	2.1	52.7	0	50.0	4.0	50.0	4.0	62.9	6.1	51.5	9.8	77.5	0
Black (non-Hispanic)	12.0	87.2	9.2	74.4	4.8	80.1	3.5	82.1	2.2	64.0	8.2	73.1	3.1	73.1	3.1	74.4	6.5	65.6	17.2	90.4	0.4
Hispanic	16.9	78.7	5.3	66.0	3.8	66.8	11.7	66.3	11.8	65.0	7.0	69.9	8.7	69.9	8.7	61.2	9.5	63.2	14.3	76.5	1.5
White (non-Hispanic)	62.4	88.1	3.6	64.6	14.6	79.8	9.2	79.7	7.0	75.2	7.0	86.3	2.2	86.3	2.2	67.9	12.3	67.2	10.3	95.5	0.4
Otherc	4.3	92.3	0.4	79.2	6.8	87.4	6.0	78.4	5.9	86.6	1.2	89.2	0	89.2	0	75.4	3.5	73.9	3.0	92.2	1.2
Educational status
<High school	9.2	70.3	4.5	49.2	5.4	58.8	7.0	66.8	2.7	53.1	11.4	65.3	3.4	65.3	3.4	55.1	15.4	49.3	17.1	70.2	0.5
High school equivalent	28.7	87.4	4.1	75.6	9.7	80.2	9.0	77.5	8.4	70.7	8.8	74.3	6.5	74.3	6.5	64.4	13.9	66.8	13.5	90.7	1.0
Some college/associate degree	27.0	88.6	4.4	67.4	11.1	79.9	7.9	79.5	6.1	74.3	4.3	83.8	2.1	83.8	2.1	74.3	6.3	69.1	8.1	93.0	1.0
Bachelor's degree	21.6	90.1	2.4	68.4	8.6	79.6	10.2	79.8	7.8	77.8	4.2	89.2	1.1	89.2	1.1	71.3	8.3	69.8	8.5	94.9	0.1
Graduate degree	13.6	83.8	6.7	55.6	20.1	77.1	10.5	75.2	7.5	71.1	7.0	83.2	3.0	83.2	3.0	63.9	11.5	62.8	14.7	93.8	0
Household income, $
<30 000	27.4	84.2	5.8	72.8	9.0	73.8	10.0	77.8	8.6	72.1	7.9	73.3	7.7	73.3	7.7	65.4	13.9	64.7	15.6	83.7	0.8
30 000 to <60 000	27.4	87.4	3.0	66.4	10.2	76.9	10.4	73.0	8.5	65.2	8.1	78.8	1.4	78.8	1.4	67.1	12.4	67.1	10.3	94.0	1.0
60 000 to <100 000	24.9	87.4	2.5	64.6	8.3	80.3	5.6	78.4	3.1	75.7	2.5	85.1	0.8	85.1	0.8	70.1	4.2	66.8	7.9	90.4	0.5
≥100 000	20.3	85.9	5.9	61.4	17.4	80.1	9.9	80.5	7.5	74.7	7.8	86.7	3.4	86.7	3.4	68.3	10.9	65.0	11.9	96.4	0
Geographic region
New England	4.7	85.6	6.2	55.6	7.6	76.8	6.5	76.0	0.9	71.4	8.5	85.7	4.5	85.7	4.5	65.3	15.7	56.7	26.0	96.8	0
Mid-Atlantic	12.6	90.6	4.8	72.7	13.7	82.4	9.8	83.9	7.0	79.2	11.0	90.5	0	90.5	0	71.5	14.6	76.4	11.0	98.3	0.5
East North Central	14.2	91.6	3.4	71.6	9.4	81.6	8.2	77.5	8.9	72.8	3.9	82.8	1.4	82.8	1.4	73.6	7.2	74.2	3.2	95.7	0
West North Central	6.4	92.5	3.2	73.0	10.1	79.5	7.4	76.9	8.6	76.7	4.7	83.9	0	83.9	0	65.5	12.1	71.3	6.5	97.8	0
South Atlantic	20.4	78.3	7.9	63.5	11.6	72.5	10.4	72.5	9.7	66.3	9.4	73.3	7.8	73.3	7.8	67.4	9.5	57.3	17.0	83.2	0.9
East South Central	5.8	89.0	7.7	64.8	27.1	82.4	7.2	80.9	2.8	72.7	6.6	70.8	8.8	70.8	8.8	65.4	16.8	72.4	10.8	89.7	0.9
West South Central	12.0	85.8	0.3	70.3	5.2	76.6	11.4	79.5	6.8	63.6	2.7	78.5	1.7	78.5	1.7	63.0	11.0	60.5	15.5	88.9	1.3
Mountain	7.6	83.7	4.1	53.1	15.3	68.3	11.5	72.9	6.5	69.3	5.6	87.2	2.2	87.2	2.2	66.5	8.2	60.4	6.9	83.0	0
Pacific	16.2	86.3	1.5	66.8	6.5	79.5	6.2	77.0	4.8	76.2	5.7	78.4	3.2	78.4	3.2	66.0	7.5	66.3	9.3	90.8	0.9
Political party identification
Strong Democrat	18.9	89.2	4.6	69.2	11.2	81.1	7.9	79.8	7.0	73.9	6.5	81.9	3.6	81.9	3.6	75.3	6.2	72.6	9.2	96.2	0.2
Not so strong Democrat	15.3	86.5	7.6	63.1	12.7	75.7	10.2	73.6	8.1	65.8	7.8	83.7	3.3	83.7	3.3	61.3	12.5	61.5	12.7	96.6	1.2
Lean Democrat	10.5	88.9	2.8	68.8	10.7	77.1	8.9	82.3	8.4	75.5	3.2	86.6	2.6	86.6	2.6	63.9	11.2	58.1	16.7	97.0	0
Don’t lean, independent, or none	17.3	73.1	6.8	54.8	6.2	64.4	9.8	65.4	9.6	59.5	5.1	62.7	7.8	62.7	7.8	57.8	14.6	55.3	14.0	73.7	0.9
Lean Republican	11.6	93.5	2.3	72.1	10.9	86.0	6.7	80.2	3.9	78.4	4.9	82.6	3.3	82.6	3.3	76.5	9.1	72.9	9.9	92.3	0.4
Not so strong Republican	13.0	85.1	2.2	69.7	12.4	79.1	11.6	78.2	5.6	67.4	13.6	81.2	0.7	81.2	0.7	67.0	11.8	66.1	11.6	86.8	1.3
Strong Republican	13.4	91.3	1.3	73.1	12.9	83.3	7.7	85.3	5.3	85.7	4.8	90.3	0.9	90.3	0.9	72.6	8.1	75.2	7.0	95.8	0
Political ideology
Very liberal	10.1	81.9	10.7	69.6	14.1	74.5	15.5	79.3	10.8	79.2	2.7	81.9	6.0	81.9	6.0	76.7	2.9	71.0	5.7	97.8	0
Somewhat liberal	16.4	88.5	2.9	65.8	11.0	79.3	11.5	82.1	12.4	73.4	4.7	91.2	0.3	91.2	0.3	69.3	12.3	65.1	15.0	94.0	0.7
Moderate	36.4	86.8	3.9	65.3	10.1	78.9	5.8	75.4	6.0	66.5	6.1	77.5	3.3	77.5	3.3	64.3	10.2	66.2	8.9	90.2	0.5
Somewhat conservative	24.3	88.6	2.5	64.4	12.6	78.6	8.1	76.0	4.0	71.5	9.6	74.7	6.3	74.7	6.3	67.6	11.8	61.1	14.6	89.6	0.6
Very conservative	11.6	81.9	4.1	72.5	7.7	72.8	10.8	75.7	6.1	76.8	8.6	86.6	0	86.6	0	68.3	12.0	69.6	14.2	86.5	1.3
2020 presidential vote
Donald Trump/Mike Pence	31.2	88.5	2.4	69.2	11.8	81.5	10.5	76.3	6.6	78.1	6.7	84.2	1.0	84.2	1.0	70.3	8.0	69.1	9.9	91.2	0.6
Joe Biden/Kamala Harris	48.2	88.4	5.6	64.9	12.1	78.4	7.9	79.2	7.6	71.7	5.5	82.6	3.6	82.6	3.6	70.1	9.1	66.3	11.0	95.7	0.6
Someone else	2.5	95.2	1.7	76.2	2.2	86.2	1.7	89.3	4.2	78.8	0	91.9	1.7	91.9	1.7	82.3	5.1	80.7	8.0	95.9	0
Did not vote in 2020 race	17.1	74.6	4.5	64.4	7.3	67.6	10.3	71.7	6.2	58.5	11.1	67.2	7.5	67.2	7.5	53.9	19.9	56.8	16.2	76.4	0.5

Statements are shown in Table 1. The introductory text for each statement can be found in eFigure 2 in the Supplement. Survey data were weighted to the US population. The margin of error was ±4.28%. Percentages do not add to 100% because some respondents chose “neither agree nor disagree” or “don’t know” or did not answer.

Race and ethnicity data were self-reported.

Other included more than 2 (non-Hispanic) or other race and ethnicity.

Discussion

Amid the COVID-19 pandemic, which has been characterized by rancorous political division, there is broad support in the US population for greater transparency at the FDA. This support may reflect the interest of the public in understanding the drug development process and has implications for the ongoing pandemic response and beyond.

A limitation of this study is that, because the survey did not inquire about other possible transparency measures, there may be more or less public support for other ideas. Another limitation is that the survey did not seek to educate by providing arguments for or against transparency. The data reflect the initial reactions of the US public, which might change with additional information.