Sexual Violence in the Media: An Exploration of Traditional Print Media Reporting in the United States, 2014–2017

Sexual violence is prevalent and, for many victims, begins early in life (1). In the United States, one in five women and one in 38 men report completed or attempted rape victimization during their lifetime, with 43.2% of female and 51.3% of male victims reporting that their first rape victimization occurred before age 18 years (1). Media have been shown to act as a socializing agent for a range of health and social behaviors (2). Media portrayals might influence, reinforce, or modify how the public responds to incidents of sexual violence and their support for prevention efforts and media might construct a lens through which the public can understand who is affected by sexual violence, what forms it takes, why it happens, and who is responsible for addressing it (3). Media portrayals of sexual violence were assessed using a systematic random sample of newspaper articles from 48 of the top 50 distributed traditional print media outlets that were examined for sexual violence content and potential differences by geographic region and year of publication. Differences by year and region in type of sexual violence covered, media language used, and outcomes reported were identified, highlighting an opportunity for public health officials, practitioners, and journalists to frame sexual violence as a preventable public health issue and to incorporate best practices from CDC and the National Sexual Violence Resource Center's Sexual Violence Media Guide (4).

Sexual violence is prevalent and, for many victims, begins early in life (1). In the United States, one in five women and one in 38 men report completed or attempted rape victimization during their lifetime, with 43.2% of female and 51.3% of male victims reporting that their first rape victimization occurred before age 18 years (1). Media have been shown to act as a socializing agent for a range of health and social behaviors (2). Media portrayals might influence, reinforce, or modify how the public responds to incidents of sexual violence and their support for prevention efforts and media might construct a lens through which the public can understand who is affected by sexual violence, what forms it takes, why it happens, and who is responsible for addressing it (3). Media portrayals of sexual violence were assessed using a systematic random sample of newspaper articles from 48 of the top 50 distributed traditional print media outlets that were examined for sexual violence content and potential differences by geographic region and year of publication. Differences by year and region in type of sexual violence covered, media language used, and outcomes reported were identified, highlighting an opportunity for public health officials, practitioners, and journalists to frame sexual violence as a preventable public health issue and to incorporate best practices from CDC and the National Sexual Violence Resource Center's Sexual Violence Media Guide (4).
Whereas numerous studies describe media portrayals of sexual violence and other forms of violence (5)(6)(7), none examined regional or temporal differences in coverage. This study used 27 sexual violence-related terms* to identify a systematic * Boolean search for each publication: (("sexual violence") OR ("sexual assault") OR ("sexual abuse") OR ("child sexual abuse") OR rape OR incest OR ("intimate partner violence") OR ("sexual exploitation") OR ("human trafficking") OR ("sex trafficking") OR prostitution OR ("sexual harassment") OR exposure OR ("unwanted penetration") OR ("unwanted sexual contact") OR ("forced oral contact") OR ("forced genital contact") OR grope OR voyeurism OR ("alleged victim") OR ("alleged perpetrator") OR perpetrator OR ("sex scandal") OR intercourse OR ("perform oral sex") OR fondle OR accuser). random sample of 2,600 articles from 48 of the top 50 traditional print media outlets distributed in the United States (8) available via electronic newspaper databases. † Outlets were stratified by regional or nationwide reach, and equal systematic samples of 130 articles were selected from each stratum for each publication year, 2014-2017. Articles were coded for strata represented and year published, type of sexual violence mentioned (sexual assault, rape, child sexual abuse, sexual exploitation, sex trafficking, prostitution, sexual harassment, or child pornography), what Sexual Violence Media Guide language was used (sex scandal/scandal, sex/intercourse, accuser, or accused) (4), and outcomes. Outcomes included perpetrator consequences (criminal justice system, civil justice system, social, or business consequences) and prevention messaging (primary, secondary, and tertiary prevention). The codebook development relied on the Sexual Violence Media Guide, which provides relevant information for effective communication about sexual violence (4). The guide is grounded in media language recommendations from the Maine Coalition Against Sexual Assault (4), CDC's Stop SV: A Technical Package to Prevent Sexual Violence (9), and past similar research (5,6). Media language considerations include suggested language (e.g., "alleged perpetrator" or "perpetrator" if convicted) and language to avoid (e.g., "accused"). Two coders were trained, and intercoder reliability was assessed on 20% of the sample, resulting in an average Kappa = 0.81, and the remaining sample was randomly split between the coders and coded. Analysis of variance (ANOVA) and post-hoc Tukey comparisons were made by article characteristic (region or year) for the type of sexual violence mentioned, media language used/language to avoid, and outcomes. Codes were not mutually exclusive.
The types of sexual violence mentioned in newspaper articles (Table 1) differed significantly by region ( Table 2). The percentage of articles within each region covering child sexual abuse was lower nationwide (28.5%) than in the Midwest (38.3%) and Northeast (42.9%) regions. National outlets published a significantly higher percentage of articles on sexual harassment (27.7%) than did media in all other regions (11.5% to 19.2%). National outlets used the term "sex scandal" or "scandal" more frequently than did media in all four regions (11.0% versus 3.5%-6.0%). The percentage of articles using the term "sex" or "intercourse" was higher in national outlets (17.1%) than in media in the Midwest (10.8%), Northeast (8.5%), and West (9.6%) regions. Inclusion of consequences for perpetrators was similar in all regions; however, calls for primary prevention of sexual violence were more frequent in national media articles (12.5%) than in those published in the Northeast (6.0%), South (6.0%), and West (7.3%).
Coverage for the types of sexual violence was similar by year, except for significant differences in reporting during 2017 for rape, sexual exploitation, sex trafficking, and sexual harassment ( Table 3). In 2017, reporting on rape and sex trafficking was significantly lower (34.9%, and 5.7%, respectively) than during 2014-2016 (46.8%-48.5% and 9.8%-10.9%, respectively; Table 3). Sexual harassment articles were more frequent in 2017 (35.7%) than in previous years (a low of 9.7% in 2014).
Newspaper coverage in 2017 differed considerably from that in other years in media language used, with significantly more coverage than all other years for use of the term "sex scandal" or "scandal" (10.9%), "accuser" (15.4%), and "accused" (37.4%).
In 2017, coverage of consequences for perpetrators (38.9%) was significantly higher than coverage in 2014 (31.5%). No significant differences by year regarding calls for primary, secondary, or tertiary prevention were found.

Discussion
Major differences in the type of sexual violence mentioned, media language used, and outcomes were identified by region, year, or both. Overall, a higher percentage of articles in national outlets than in regional outlets used sex scandal, sex/intercourse and included calls for prevention. In general, the type of sexual violence mentioned and the language used in 2017 differed from that during other years (e.g., decreased mention of rape and sex trafficking and increased mention of sexual harassment). These changes might reflect wider coverage of sexual harassment and exploitation allegations involving prominent figures in the film industry, media, state and national congresses, and technology companies, including the "#metoo" movement, which experienced a resurgence in the fall of 2017 that could have influenced article content during the last quarter of 2017. § The findings in this report are subject to at least three limitations. First, research was limited by access to electronic databases that carried traditional print media newspapers; therefore, only 48 of the top 50 distributed newspapers in the United States were accessible. Second, although outlets were identified by reach and stratified by region, how much each publication outlet encompasses rural readership is unclear, and generalizations to these populations should be made with caution. However, many print outlets are also widely available online, likely increasing their reach beyond their physical distributions. Finally, this study did not examine how audiences interact with print and electronic news media through social media. For example, social media allows users to comment on and challenge how traditional news frames sexual violence (10). Such social media interactions present an opportunity for further research and consideration in understanding the complex impact of media on public perceptions of sexual violence.
Media reporting included both suggested language (e.g., "sexual assault") and language to avoid (e.g., "sex scandal" or "scandal"), as referenced in the Sexual Violence Media Guide (4). Traditional media might have more of an impact on increasing awareness and prevention of sexual violence if their portrayals do not place blame on the victim and if they use suggested terms to describe violent acts throughout their articles. Focused dissemination of the Sexual Violence Media Guide (4) might benefit all media outlets. Outcomes including perpetrator consequences or prevention messaging generally were reported infrequently. Although outcomes might not be known at the time of reporting, traditional media might be missing an opportunity to integrate prevention messages within current or breaking news. The media can play an important role by partnering with public health organizations to ensure that their portrayals of sexual violence are factual, nonbiased, do not inadvertently blame victims, and include prevention messages in stories about sexual violence. One of the prevention strategies identified in the STOP SV technical package, which includes the best available evidence to prevent sexual violence, is promoting social norms that protect against violence (9). As an institution that can influence social norms, the media might contribute to efforts to prevent sexual violence through accurate descriptions of prevalence and impact of sexual violence, establishment of sexual violence as a public health issue, and, when possible, inclusion of messages and resources for prevention. In this way, awareness of the problem and prevention messaging might reach broader audiences.
Understanding how media outlets have historically framed sexual violence might help public health officials and practitioners work productively with journalists to identify potential unintended effects of specific language use. The Sexual Violence Media Guide (4) can be used to inform and evaluate the impact of public health and media collaborations. The media, public health practitioners, and communities can work together to incorporate language from the Sexual Violence Media Guide (4) to change public perceptions about circumstances surrounding sexual violence and encourage public health approaches to prevention.

Summary
What is already known about this topic? Sexual violence media portrayals can influence public perceptions, which can affect social norms and behavior.
What is added by this report?
Examination of articles from traditional print media outlets found regional and temporal differences in types of sexual violence covered, media language used, and outcomes reported in news story coverage in 2017, compared with that from 2014 to 2016.
What are the implications for public health practice?
Through cross-sectoral collaboration and use of the Sexual Violence Media Guide language suggestions, media, public health practitioners, and communities can work together to effectively use best practices to report on sexual violence, emphasize sexual violence as preventable, and frame sexual violence as a public health issue.
Most persons infected with SARS-CoV-2, the virus that causes coronavirus disease 2019 (COVID-19), develop virus-specific antibodies within several weeks, but antibody titers might decline over time. Understanding the timeline of antibody decline is important for interpreting SARS-CoV-2 serology results. Serum specimens were collected from a convenience sample of frontline health care personnel at 13 hospitals and tested for antibodies to SARS-CoV-2 during April 3-June 19, 2020, and again approximately 60 days later to assess this timeline. The percentage of participants who experienced seroreversion, defined as an antibody signal-tothreshold ratio >1.0 at baseline and <1.0 at the follow-up visit, was assessed. Overall, 194 (6.0%) of 3,248 participants had detectable antibodies to SARS-CoV-2 at baseline (1). Upon repeat testing approximately 60 days later (range = 50-91 days), 146 (93.6%) of 156 participants experienced a decline in antibody response indicated by a lower signal-to-threshold ratio at the follow-up visit, compared with the baseline visit, and 44 (28.2%) experienced seroreversion. Participants with higher initial antibody responses were more likely to have antibodies detected at the follow-up test than were those who had a lower initial antibody response. Whether decay in these antibodies increases risk for reinfection and disease remains unanswered. However, these results suggest that serology testing at a single time point is likely to underestimate the number of persons with previous SARS-CoV-2 infection, and a negative serologic test result might not reliably exclude prior infection.
Once infected with SARS-CoV-2, most persons develop virus-specific antibodies within 2-3 weeks (2,3). Serology tests are now being used widely in seroprevalence studies to understand patterns of viral spread, cumulative incidence of SARS-CoV-2 infection, and pandemic trajectory (4)(5)(6). Further, serologic testing has been proposed as a way to identify persons who might have developed immunity through a previous infection. Understanding how rapidly SARS-CoV-2 antibody levels decline after seroconversion is critical for interpreting serology results. A limited number of studies have * Wesley H. Self and Mark W. Tenforde contributed equally to this report; Manish M. Patel and Natalie J. Thornburg contributed equally to this report.
found declines in SARS-CoV-2 antibody levels over time (7)(8)(9), but the frequency and timing of seroreversion (the decline in antibody levels below the positivity threshold after initial seroconversion) remains largely unknown. The Influenza Vaccine Effectiveness in the Critically Ill (IVY) Network, a collaboration of academic medical centers in the United States that studies influenza and COVID-19 (1), enrolled a convenience sample of frontline health care personnel at 13 centers in 12 states, † with a target of 250 participants per center. Health care personnel were eligible if they reported regular direct contact with COVID-19 patients and worked in the emergency department, intensive care unit, or other hospital-based unit that cared for patients with COVID-19. Participants underwent two study visits: a baseline visit (conducted April 3-June 19, 2020) and a follow-up visit approximately 60 days after the baseline visit. At both visits, blood was collected for SARS-CoV-2 antibody testing, and participants were questioned about demographic characteristics, underlying medical conditions, signs or symptoms of an acute viral infection from February 1, 2020, until the visit date, § and any previous SARS-CoV-2 testing (e.g., reverse transcription-polymerase chain reaction [RT-PCR]) for acute infection. Blood specimens collected at the baseline and followup visits were tested for SARS-CoV-2 antibodies at CDC using an enzyme-linked immunosorbent assay (ELISA) against the extracellular domain of the SARS-CoV-2 spike protein (4). The assay detects all SARS-CoV-2 immunoglobulin (Ig) types (IgA, IgM, or IgG). Specimens were considered reactive with a signal-to-threshold ratio >1.0 at a background corrected serum dilution of 1:100, with higher ratios indicating higher antibody titers. The assay has a sensitivity estimated at 96% and specificity at 99% (4).
The change in signal-to-threshold ratio between the baseline visit and follow-up visit was quantified, and the percentage of participants who experienced seroreversion was reported. Logistic regression was used to evaluate the association between baseline signal-to-threshold value and seroreversion, adjusting for age, sex, race/ethnicity, number of days between the baseline and follow-up visit, and presence of one or more chronic medical condition. Analyses were conducted using Stata (version 16; StataCorp). The project was determined to be nonresearch public health surveillance by participating institutions and CDC and was conducted consistent with applicable federal law and CDC policy. ¶ Among 3,248 health care personnel, 194 (6.0%) had antibodies to SARS-CoV-2 at the baseline visit (1). Among these, 156 (80.4%) returned for the follow-up visit around 60 days later (range = 50-91 days). Among these 156 participants with a positive baseline serology and follow-up antibody testing performed, median age was 38 years (interquartile range [IQR] = 30-48 years), 94 (60.3%) were female, and 108 (69.2%) reported one or more symptoms of an acute infection consistent with COVID-19 between February 1, 2020 and the baseline visit. Among the 108 participants who reported symptoms, the median interval between symptom onset and baseline serology testing was 30 days (IQR = 19-40 days). Participants who reported symptoms of an acute viral illness since February had higher baseline signal-to-threshold ratios (median = 3.6; IQR = 3.1-3.9) than did those who did not report symptoms (median = 2.5; IQR = 1.5 to 3.6) (p<0.001). Among these 156 participants, 72 (46.2%) reported past RT-PCR testing for SARS-CoV-2, 46 (63.9%) of whom had positive test results; no hospitalizations were reported.

Discussion
In this study of 156 frontline U.S. health care personnel who received positive SARS-CoV-2 antibody test results in spring 2020 and returned for follow-up testing approximately 60 days later, 146 (93.6%) had a decline in antibody levels between baseline and follow-up, and 44 (28.2%) had complete seroreversion, i.e., a decline of antibody to levels below the threshold for positivity. A higher percentage of those with low baseline antibody levels seroreverted (64.9%) than did those with high baseline titers (7.1%). These results suggest that a substantial proportion of persons infected with SARS-CoV-2 might have negative serologic test results in the months following infection. This has several important implications. Crosssectional seroprevalence studies that estimate the number of  . ¶ All variates in table were included in multivariable logistic regression model. ** This measured the odds ratio for seroreversion associated with a 1-unit difference in signal-to-threshold ratio value (e.g., 5 versus 4), comparing the higher ratio to the lower ratio. † † This measured the odds ratio for seroreversion associated with a 10-year difference in age (e.g., 60 years versus 50 years), comparing the higher age to the lower range. § § This measured the odds ratio for seroreversion associated with a 1-week difference in time to follow-up (e.g., 9 weeks versus 8 weeks), comparing the later follow-up time to the earlier follow-up time. ¶ ¶ Medical conditions included one or more of the following: asthma, chronic obstructive pulmonary disease, other chronic lung condition, chronic heart failure, coronary artery disease, diabetes mellitus, hypertension, chronic renal disease (dialysis), autoimmune disease, active cancer (not in remission), immunosuppression (undergoing active chemotherapy or taking a medication to suppress the immune system).
persons who have been infected with SARS-CoV-2 will likely underestimate incidence because a proportion of previously infected persons will likely serorevert and thus not be counted as having been previously infected. In addition, these results challenge the notion of using serologic testing results at an individual level to designate previous SARS-CoV-2 infection. COVID-19 convalescent plasma is widely being used as a treatment for COVID-19, including through a Food and Drug Administration Emergency Use Authorization in the United States (10); these results demonstrate that the optimal window for collecting convalescent plasma with high levels of SARS-CoV-2 antibodies from donors who have recovered from COVID-19 might be short because of substantial

Summary
What is already known about this topic?
Most persons develop virus-specific antibodies to SARS-CoV-2 after infection; however, the timeline of antibody decline over time is uncertain.
What is added by this report?
Among 156 frontline health care personnel who had positive SARS-CoV-2 antibody test results in spring 2020, 94% experienced a decline at repeat testing approximately 60 days later, and 28% seroreverted to below the threshold of positivity. Participants with higher initial antibody responses were more likely to have antibodies detected at the follow-up test than were those who had a lower initial antibody response.
What are the implications for public health practice?
SARS-CoV-2 antibodies decline over weeks following acute infection. Negative SARS-CoV-2 serologic results do not exclude previous infection, which has significant impacts on how serologic studies are interpreted.
decline in antibody levels within 60 days. Whether decline in SARS-CoV-2 antibodies increases risk for reinfection and disease in humans remains unknown. Humoral immunity to primary infections from a novel virus might not be as durable or strong as that to secondary infections, but memory B-cell and T-cell responses might reduce the severity of illness with repeat exposure or infection. The findings in this report are subject to at least four limitations. First, the timing of the baseline serologic test relative to symptom onset was not standardized, which might affect baseline signal-to-threshold ratios, particularly for recently acquired infection in which antibody levels might still have been increasing. Second, the study population was derived from a convenience sample, which might result in nonrepresentativeness. Third, 38 (20%) participants were lost to follow-up, limiting size of the study population. Finally, misclassification of antibody status was possible; however, this was considered to be unlikely because of the high sensitivity and specificity of the ELISA.
In this study of frontline health care personnel at 13 medical centers who received positive SARS-CoV-2 antibody test results in spring 2020, more than one quarter were seronegative approximately 60 days after testing. Because SARS-CoV-2 antibody levels might decline in a proportion of persons following primary infection, a negative serology test does not reliably exclude previous infection. These antibody declines might not equate to loss of protective immunity or increased risk for reinfection; this was not assessed in this study. Cross-sectional seroprevalence studies to evaluate population immunity are likely to underestimate rates of previous infection because antibodies appear to only be detectable for a discrete period of time following infection. US Department of Health and Human Services/Centers for Disease Control and Prevention Abbreviations: COVID-19 = coronavirus disease 2019; ELISA = enzyme-linked immunosorbent assay. * Antibody response was categorized as high or low based on signal-to-threshold ratio of panimmunoglobulin reactivity to SARS-CoV-2 full length S protein ELISA at baseline visit. † Signs and symptoms included one or more of the following reported between February 1, 2020, and the date of baseline study visit: fever (temperature >99.5°F [37.5°C]), cough, shortness of breath, myalgias, sore throat, vomiting, diarrhea, change in or loss of taste or smell, and chest tightness.  Breastfeeding has health benefits for both infants and mothers and is recommended by numerous health and medical organizations* , † (1). The birth hospitalization is a critical period for establishing breastfeeding; however, some hospital practices, particularly related to mother-newborn contact, have given rise to concern about the potential for mother-to-newborn transmission of SARS-CoV-2, the virus that causes coronavirus disease 2019 (COVID-19) (2). CDC conducted a COVID-19 survey (July 15-August 20, 2020) among 1,344 hospitals that completed the 2018 Maternity Practices in Infant Nutrition and Care (mPINC) survey to assess current practices and breastfeeding support while in the hospital. Among mothers with suspected or confirmed COVID-19, 14.0% of hospitals discouraged and 6.5% prohibited skin-to-skin care; 37.8% discouraged and 5.3% prohibited rooming-in; 20.1% discouraged direct breastfeeding but allowed it if the mother chose; and 12.7% did not support direct breastfeeding, but encouraged feeding of expressed breast milk. In response to the pandemic, 17.9% of hospitals reported reduced in-person lactation support, and 72.9% reported discharging mothers and their newborns <48 hours after birth. Some of the infection prevention and control (IPC) practices that hospitals were implementing conflicted with evidence-based care to support breastfeeding. Mothers who are separated from their newborn or not feeding directly at the breast might need additional postdischarge breastfeeding support. In addition, the American Academy of Pediatrics (AAP) recommends that newborns discharged before 48 hours receive prompt follow-up with a pediatric health care provider.
Peripartum practices, including immediate maternal-newborn skin-to-skin contact, enabling mothers and newborns to room-in together, and teaching mothers to breastfeed (e.g., demonstrating good positioning and attachment and supporting mothers to express breast milk if they are temporarily separated from their newborns), are recommended worldwide for implementation in birth hospitals (3,4). These practices have a positive impact on both short-and long-term breastfeeding outcomes, which in turn benefit maternal and child health (5,6). Guidance on care during the birth hospitalization has evolved with the pandemic and at times has varied across public health and professional medical organizations. * https://www.cdc.gov/breastfeeding/about-breastfeeding/why-it-matters.html. † https://apps.who.int/nutrition/topics/exclusive_breastfeeding/en/index.html. CDC's mPINC survey is a census of all birth hospitals in the United States and its territories. § The 2,039 hospitals that completed the 2018 mPINC survey (70% response) were sent an e-mail link to a 13-item COVID-19 survey, conducted using Research Electronic Data Capture (REDCap) (version 10.0.08; Vanderbilt University). The survey was open during July 15-August 20, 2020 and asked about current hospital practices. Maternity services had been closed at 22 hospitals, and 130 e-mailed surveys were undeliverable (i.e., e-mails bounced back). Overall, 1,344 hospitals completed the survey (66.6% overall response rate; 71.2% among delivered surveys). Hospitals were asked about their actual or planned approach to managing maternity patients with suspected or confirmed COVID-19 (as defined by the hospital), and the approximate number of these patients they had cared for. Descriptive analyses were conducted using SAS software (version 9.4; SAS Institute). This activity was reviewed by CDC and was conducted consistent with applicable federal law and CDC policy. ¶ Data on annual births and hospital type were obtained from the 2018 mPINC survey.
Among 1,343 hospitals with available information, 724 (53.9%) had cared for one to 19 newborns whose mothers had confirmed COVID-19; 152 (11.4%) had cared for 20 or more, and 457 (34.0%) had not cared for any (Table 1). Approximately one half of the hospitals reported fewer than 1,000 annual births, 42.0% reported 1,000-4,999, and 2.8% reported 5,000 or more. The majority (78.8%) of hospitals were nonprofit. Among 1,344 birth hospitals, 1,211 (90.2%) reported having enough COVID-19 tests, and 864 (64.3%) were performing universal COVID-19 testing of women admitted to labor and delivery ( Table 2). Few hospitals (4.8%) reported separating all mothers and newborns until the mother received a negative test result, and 28.6% separated newborns and mothers if the mother was symptomatic or had a known exposure until test results were obtained; 24.2% separated mothers and newborns only if the mother received a positive test result.
Because of the COVID-19 pandemic, 239 (17.9%) hospitals reported decreased access to in-person lactation support and 72.9% discharged mothers and newborns <48 hours after birth. After discharge, 802 (59.7%) and 655 (48.7%) hospitals offered in-person and virtual breastfeeding consultations, respectively. Since the start of the pandemic, 924 hospitals (68.9%) reported that their exclusive breastfeeding rates during hospitalization had stayed about the same, and similar percentages reported increases (11.3%), compared with decreases (12.2%).

Discussion
During a 5-week period (July 15-August 20) when many areas of the country were experiencing substantial community transmission of SARS-CoV-2, hospitals were implementing a variety of practices intended to balance evidence-based maternity care with COVID-19-related IPC. Hospital practices are likely evolving along with the pandemic (7), potentially driven by multiple factors, including level of community transmission, guidance from public health and medical professional organizations, and a hospital's own experience in preparation for or caring for pregnant women and newborns with COVID-19.
Various organizations have promulgated COVID-19 guidance on care for pregnant women and newborns, which at times has been conflicting. The World Health Organization recommends that mothers with COVID-19 be able to practice skin-to-skin care, rooming-in, and direct breastfeeding while wearing a mask, unless they are too ill to do so**; similar guidance is supported by the American Academy of Family Physicians, † † and the American College of Obstetricians and Gynecologists § § promotes shared decision-making with the mother and the health care team. On the other hand, when this survey was launched, CDC and AAP recommended temporary separation of newborns from mothers with suspected or confirmed COVID-19 (8). During data collection for this survey, CDC (August 3 ¶ ¶ ) and AAP (July 22***) updated their guidance, supporting maternal autonomy in decision-making. Changes in guidance reflect evolving knowledge about the virus and its potential impact on newborns. To date, there has been no definitive evidence of transmission of SARS-CoV-2 through breast milk. In addition, there have been reports of secretory immunoglobulin A against SARS-CoV-2 in breast milk samples of women with COVID-19, suggesting that breastfeeding might be particularly important for newborns of mothers with COVID-19 (9).
One third of hospitals in this study reported not having cared for any neonates born to mothers with COVID-19; however, others had extensive experience, including 17 hospitals reporting caring for at least 100 of these newborns. Follow-up of mothers with COVID-19 delivering at three large New York birth hospitals found reduced breastfeeding rates both in the hospital and after returning home among mothers who had been separated from their newborns (7). After identification of this finding, and the observed stress among mothers and ** h t t p s : / / w w w. w h o . i n t / n e w s -r o o m / c o m m e n t a r i e s / d e t a i l / breastfeeding-and-covid-19. † † https://www.aafp.org/about/policies/all/breastfeeding-covid19.html. § § https://www.acog.org/en/Clinical/Clinical%20Guidance/Practice%20 Advisory/Articles/2020/03/Novel%20Coronavirus%202019. ¶ ¶ https://www.cdc.gov/coronavirus/2019-ncov/hcp/caring-for-newborns.html. *** https://services.aap.org/en/pages/2019-novel-coronavirus-covid-19-infections/ clinical-guidance/faqs-management-of-infants-born-to-covid-19-mothers/. newborns as a result of separation, the hospital system revised its policy and began to allow asymptomatic mothers with laboratory-confirmed COVID-19 to room-in and breastfeed. Nearly one in five hospitals in this study reported that inperson lactation support had decreased during the pandemic. Approximately 60% of hospitals in this study reported offering in-person breastfeeding consultations postdischarge, compared with 69% of hospitals reporting offering this in the 2018 mPINC survey (CDC, unpublished data, 2020). Lactation specialists working in health care settings should follow recommended IPC measures for those settings. † † † Nearly one half of † † † https://www.cdc.gov/coronavirus/2019-ncov/hcp/infection-controlrecommendations.html. hospitals reported offering virtual breastfeeding consultations; however, no data are available on this practice before the pandemic. Notably, these changes in lactation support affect newborns broadly, not just those born to mothers with COVID-19. Approximately equal numbers of hospitals reported that their exclusive breastfeeding rates had increased and decreased; the majority reported the rate had stayed approximately the same. The reasons for these changes are unknown. However, the pandemic could contribute to reduced breastfeeding as a result of maternal/newborn separation and reductions in lactation support. On the other hand, the visitor restriction policies implemented by many hospitals could potentially provide more opportunities for breastfeeding and for a mother

Summary
What is already known about this topic?
Evidence-based hospital practices supporting breastfeeding have sometimes conflicted with COVID-19 infection prevention and control measures.
What is added by this report?
During summer 2020, hospitals implemented a variety of practices intended to balance evidence-based maternity care with infection prevention and control. Because of the pandemic, 17.9% of hospitals reported that in-person lactation support had decreased, and 72.9% reported discharging mothers and their babies <48 hours after birth.
What are the implications for public health practice?
Additional postdischarge breastfeeding support and newborn follow-up might be needed during the COVID-19 pandemic. Longer-term monitoring of exclusive breastfeeding rates at hospital discharge will be important for assessing this aspect of the pandemic on infant health.
to learn her newborn's feeding cues. § § § Longer-term monitoring of exclusive breastfeeding rates at hospital discharge will be important to assess as one measure of the impact of the pandemic on infant health.
Approximately three fourths of hospitals reported discharging mothers and their newborns <48 hours after birth because of the pandemic. No standard length of stay for the birth hospitalization exists, but the Newborn Mothers' Health and Protection Act of 1996 prohibits the restriction of benefits to <48 hours for a vaginal delivery or <96 hours for a cesarean section. ¶ ¶ ¶ AAP describes discharge <48 hours after delivery as a "shortened hospital stay" and notes that although it can be accommodated for healthy term newborns, it is not appropriate for all mothers and infants (10). AAP also recommends that all newborns discharged <48 hours after birth be evaluated by a pediatric health care provider within 48 hours of discharge (10). This visit, in part, assesses effective feeding, which might be more critical during the pandemic when breastfeeding mothers might be receiving reduced hospital lactation support.
The findings in this report are subject to at least three limitations. First, although mPINC is a census of all birth hospitals, this survey was only sent to the hospitals that completed the 2018 mPINC survey. Second, hospitals used their own definitions for suspected and confirmed COVID-19, and those definitions might have been different. Finally, data captured in this survey represent a single point in time. Policies and practices likely will continue to change as the pandemic evolves. § § § https://www.medscape.com/viewarticle/933964. ¶ ¶ ¶ https://www.cms.gov/CCIIO/Programs-and-Initiatives/Other-Insurance-Protections/nmhpa_factsheet. Women with suspected or confirmed COVID-19 who are separated from their newborns and whose newborns are not feeding directly at the breast might need timely, professional, breastfeeding support.**** In addition, AAP advises that infants discharged <48 hours after delivery receive prompt follow-up with a pediatric health care provider to ensure optimal feeding.
On November 20, 2020, this report was posted as an MMWR Early Release on the MMWR website (https://www.cdc.gov/mmwr).
During August 7-16, 2020, a motorcycle rally was held in western South Dakota that attracted approximately 460,000 persons from across the United States to numerous indoor and outdoor events over a 10-day period. During August-September 2020, the Minnesota Department of Health (MDH) investigated a coronavirus disease 2019 (COVID-19) outbreak associated with the rally in Minnesota residents. Fifty-one primary event-associated cases were identified, and 35 secondary or tertiary cases occurred among household, social, and workplace contacts, for a total of 86 cases; four patients were hospitalized, and one died. Approximately one third (34%) of 87 counties in Minnesota had at least one primary, secondary, or tertiary case associated with this rally. Genomic sequencing supported the associations with the motorcycle rally. These findings support current recommendations for mask use, physical distancing, reducing the number of attendees at gatherings, isolation for patients with COVID-19, and quarantine for close contacts to slow the spread of SARS-CoV-2 (1). Furthermore, although these findings did not capture the impact of the motorcycle rally on residents of other states, they demonstrate the rationale for consistent mitigation measures across states.

Investigation and Findings
On August 21, 2020, MDH identified confirmed COVID-19 cases in persons who reported attending the motorcycle rally in the neighboring state of South Dakota. A primary, event-associated case was defined as an illness in a person who reported attending the rally or who traveled to western South Dakota by motorcycle during August 7-16 and who had symptom onset or specimen collection before August 30 (within 14 days after the end of the rally). Reverse transcription-polymerase chain reaction testing for SARS-CoV-2, the virus that causes COVID-19, was used to confirm cases. All confirmed cases among Minnesota residents were reported to MDH. MDH or local public health department staff members interviewed patients with confirmed SARS-CoV-2 infection to identify exposures and persons who might have been in contact with patients during their infectious period (2 days before through 10 days after symptom onset).* To assess exposures, interviews included questions about travel and being in specific settings, such as bars or restaurants, schools, health care facilities, or events or social gatherings in the 14 days before symptom onset. During August-September 2020, MDH and local health department staff members interviewed >80% of patients with a confirmed SARS-CoV-2 infection.
Secondary and tertiary cases were identified from case interview data. Confirmed secondary cases were defined as laboratory-confirmed infections in persons who did not attend the rally but who received SARS-CoV-2-positive test results after having contact with a person who had a primary case during their infectious period. Tertiary cases were laboratoryconfirmed cases in persons who had contact with a person who had a secondary case during their infectious period. Likely event-associated secondary cases were confirmed infections in patients who had contact with a person who had symptoms of COVID-19 and had attended the motorcycle rally but who were not tested. Likely event-associated tertiary cases were confirmed infections in patients who had contact with persons who had a likely event-associated secondary case during their infectious period.
To investigate genomic similarity among COVID-19 cases, available SARS-CoV-2 RNA-positive clinical specimens were obtained from clinical laboratories, and whole genome sequencing was conducted at the MDH Public Health Laboratory on 38 specimens using previously described methods (2). Phylogenetic relationships, including distinct clustering of viral whole genome sequences, were inferred based on nucleotide differences via IQ-TREE † using general time reversible substitution models (3) as a part of the Nextstrain § workflow (4). This activity was reviewed by CDC and was conducted consistent with applicable federal law and CDC policy. ¶ * The infectious period was estimated to begin 2 days before symptom onset and end 10

Summary
What is already known about this topic?
Gatherings present an opportunity for rapid spread of COVID-19.
What is added by this report?
Following a 10-day motorcycle rally in South Dakota attended by approximately 460,000 persons, 51 confirmed primary event-associated cases, 21 secondary cases, and five tertiary cases were identified in Minnesota residents. An additional nine likely rally-associated secondary or tertiary cases occurred. Four patients were hospitalized, and one died. Genomic sequencing supported the associations with the motorcycle rally.
What are the implications for public health practice?
The impact of gatherings as a source of virus transmission underscores the importance of reducing the number of attendees at gatherings, using face masks, and encouraging physical distancing to prevent ongoing transmission of SARS-CoV-2. Furthermore, these findings demonstrate the rationale for consistent mitigation measures across states.
This investigation identified 86 cases, including 51 (59%) primary event-associated cases,** 26 (30%) confirmed secondary and tertiary cases, and nine (10%) likely event-associated secondary or tertiary cases. Four patients were hospitalized, and one died (Table). The median interval between specimen collection and interview was 3 days (range = 1-13 days). Overall, 64 (74%) patients were symptomatic, including 39 (76%) of 51 patients with a primary case and 25 (71%) of 35 patients with secondary and tertiary cases. Among patients with primary cases and symptom onset after the start of the rally, onset dates ranged from August 8 to August 26 ( Figure 1). Two patients reported symptom onset before the event and attended the rally during their infectious period. Among primary patients, the median age was 44 years (range = 26-76 years), and 31 (61%) were male. Sixteen (33%) of 48 interviewed patients reported working while infectious, including five who worked at the rally and four who worked in health care after returning from the rally.
Forty-one (80%) interviewed patients with primary eventassociated COVID-19 reported having close contact † † with others during their infectious period, with an average of 2.5 close contacts per patient (range = 1-8). Overall, 36 (75%) of 48 interviewed patients with primary event-associated cases ** One patient reported attending the rally but refused interview. Two additional patients who refused to be interviewed were identified as having attended the rally through secondary case interviews in which other patients reported them as primary event-associated contacts. † † Close contact was defined as being within 6 feet of a patient with laboratoryconfirmed COVID-19 infection for ≥15 minutes. https://www.cdc.gov/ coronavirus/2019-ncov/php/contact-tracing/contact-tracing-plan/contacttracing.html.
reported having close contact with persons in their household while infectious, and 17 (35%) reported having other (social/ workplace) close contacts while infectious. Patients reported a total of 59 household contacts (range = 0-4 per patient) and 43 social/workplace contacts (range = 0-6 per patient). Among the 35 patients with confirmed or likely eventassociated secondary/tertiary COVID-19, 25 (71%) were symptomatic, with symptom onset dates during August 12-29 (Table). The median age was 32 years (range = 1-83 years), and 13 (37%) were male. Fifteen (43%) persons with secondary or tertiary COVID-19 were household contacts of a person with a primary or secondary infection, 12 (34%) were social contacts, and eight (23%) were workplace contacts. Secondary transmission from this rally occurred via two workplace outbreaks, one wedding outbreak, and one funeral outbreak. Approximately one third (34%) of Minnesota's 87 counties had at least one primary, secondary, or tertiary case associated with this rally.

Whole Genome Sequencing
Specimens were obtained from 52 (60%) patients. Among these, 38 (73%) specimens (23 [61%] from primary and 15 [39%] from secondary and tertiary cases) were successfully sequenced, covering at least 98% of the SARS-CoV-2 genome. Six genetically similar clusters with known epidemiologic links were identified (i.e., cases in patients who were close contacts or who had common exposures at the rally), five of which demonstrated secondary or secondary and tertiary transmission. Cluster A (Figure 2) included genetically similar specimens for seven primary cases and one secondary case (specimen MN-MDH-1710). Among primary cases, specimens were collected from two patients who reported working at the rally, including one who worked at a restaurant. Two other patients in this cluster reported visiting that restaurant. Another patient who attended the rally also reported visiting the same restaurant; this patient was a household contact of the patient with specimen MN-MDH-1710. Cluster B represented a chain of transmission in a workplace setting that included five cases. The secondary case in this cluster (with specimen MN-MDH-STU0004) occurred in a workplace contact of a motorcycle rally attendee (specimen MN-MDH-STU0001) and a social contact of one of the persons with a tertiary case (specimen MN-MDH-STU0008). Another secondary case § § in this cluster was in a workplace contact of the rally attendee and was a household contact of two of the three patients with tertiary cases in this cluster (specimens MN-MDH-1708 and MN-MDH-1709). Cluster C represented secondary transmission from a rally attendee (specimen MN-MDH-1651) to a household contact (specimen MN-MD-1705). Cluster D Specimen MN-MDH-1569 was from a musician who performed at the rally and later at another concert with a different band whose members did not attend the rally. Primary event-associated cases with specimens MN-MDH-1571 and MN-MDH-1572 had no known connection to each other or identified common exposure at the rally; however, the reported symptom onset dates and travel dates for these cases were identical. Similarly, primary event-associated cases with specimens MN-MDH-1568 and MN-MDH-1707 had no known epidemiologic link but had identical symptom onset dates. This might indicate a common exposure that was not identified through epidemiologic evidence.

Public Health Response
On August 5, MDH recommended through media events that motorcycle rally attendees quarantine for 14 days upon return and be tested 5-7 days later even if they were asymptomatic. Attendees and their close contacts with confirmed COVID-19 were instructed to self-isolate. Contacts of patients with confirmed COVID-19 were instructed to quarantine.

Discussion
Eighty-six Minnesota COVID-19 cases were associated with the South Dakota motorcycle rally; approximately one third of counties in Minnesota reported at least one case epidemiologically linked to this event. These findings highlight the far-reaching effects that gatherings in one area might have on another area. The motorcycle rally was held in a neighboring state that did not have policies regarding event size and mask use, underscoring the implications of policies within and across jurisdictions. The findings suggest that this rally not only had a direct impact on the health of attendees, but also led to subsequent SARS-CoV-2 transmission among household, social, and workplace contacts of rally attendees upon their return to Minnesota. Whole genome sequencing results supported the finding of secondary and tertiary transmission associated with this rally.  and had symptom onset or specimen collection within 14 days of the end of the rally. § Laboratory-confirmed infection in a person who had contact with a primary case during that person's infectious period. ¶ Laboratory-confirmed infection in a person who had contact with a secondary case during that person's infectious period. ** Laboratory-confirmed infection in a person who had contact with a symptomatic person who attended the rally but was not tested. † † Laboratory-confirmed infection in a person who had contact with a likely secondary case. § § Symptom status was unknown for three patients with a primary case, three with a secondary/tertiary case, and one with a likely event-associated secondary/tertiary case.
The findings in this report are subject to at least three limitations. First, despite in-depth epidemiologic investigation, the findings represent an underestimate of the motorcycle rally's impact in Minnesota and did not capture the impact within South Dakota or other states. Case interviews were voluntary, and patients could choose not to respond to certain questions. Ten patients reported having close contacts but refused to disclose additional details regarding these contacts. Therefore, it was not possible to identify all contacts of patients who attended the rally. Second, attendees and their contacts might not have been tested for SARS-CoV-2. Two rally attendees indicated that their contacts had COVID-19-like symptoms but did not plan to be tested. As such, the findings underrepresent the number of cases, close contacts, and secondary and tertiary cases. Finally, only 52 specimens were received at the MDH Public Health Laboratory because many testing laboratories do not retain or store specimens long-term. Among these specimens, only 38 were successfully sequenced. The lack of whole genome sequencing data from all cases hindered establishment of complete genetic relatedness for epidemiologic investigation.
A large event in a neighboring state triggered chains of SARS-CoV-2 transmission within Minnesota. Other studies have shown that chains of transmission associated with gatherings are not uncommon within the United States (5,6). Despite underascertainment of the rally's full impact in Minnesota and other states, these findings highlight the importance of reducing the number of attendees at gatherings and emphasizing mask use, physical distancing, isolation for patients with COVID-19, and quarantine for close contacts as strategies for reducing the spread of COVID-19. Furthermore, these findings demonstrate the rationale for consistent mitigation measures across states. On November 20, 2020, this report was posted as an MMWR Early Release on the MMWR website (https://www.cdc.gov/mmwr).

FIGURE 2. Phylogenetic tree* showing genetic distance between available † SARS-CoV-2 virus specimens collected from South Dakota motorcycle rally attendees and their contacts (N = 38) -Minnesota, August 2020
Wearing masks is a CDC-recommended* approach to reduce the spread of SARS-CoV-2, the virus that causes coronavirus disease 2019 (COVID-19), by reducing the spread of respiratory droplets into the air when a person coughs, sneezes, or talks and by reducing the inhalation of these droplets by the wearer. On July 2, 2020, the governor of Kansas issued an executive order † (state mandate), effective July 3, requiring masks or other face coverings in public spaces. CDC and the Kansas Department of Health and Environment analyzed trends in county-level COVID-19 incidence before (June 1-July 2) and after (July 3-August 23) the governor's executive order among counties that ultimately had a mask mandate in place and those that did not. As of August 11, 24 of Kansas's 105 counties did not opt out of the state mandate § or adopted their own mask mandate shortly before or after the state mandate was issued; 81 counties opted out of the state mandate, as permitted by state law, and did not adopt their own mask mandate. After the governor's executive order, COVID-19 incidence (calculated as the 7-day rolling average number of new daily cases per 100,000 population) decreased (mean decrease of 0.08 cases per 100,000 per day; net decrease of 6%) among counties with a mask mandate (mandated counties) but continued to increase (mean increase of 0.11 cases per 100,000 per day; net increase of 100%) among counties without a mask mandate (nonmandated counties). The decrease in cases * https://www.cdc.gov/coronavirus/2019-ncov/more/masking-science-sars-cov2.
html?fbclid=IwAR28PppCa6x2uxwO8Z2baHM0KHS4JXx0inzzMQs3zRH V1qql_0a8mxZfpCw. https://www.cdc.gov/coronavirus/2019-ncov/preventgetting-sick/cloth-face-cover-guidance.html. † https://governor.kansas.gov/wp-content/uploads/2020/07/20200702093130003.pdf. § Allen, Atchison, Bourbon, Crawford, Dickinson, Douglas, Franklin, Geary, Gove, Harvey, Jewell, Johnson, Mitchell, Montgomery, Morris, Pratt, Reno, Republic, Saline, Scott, Sedgwick, Shawnee, Stanton, and Wyandotte counties. Data on county orders were collected through point-in-time surveys of local health department and other county officials and were supplemented with online searches for published orders and announcements on social media and local news sites. Text in the county orders was analyzed to determine whether mask mandates were in place as of August 11, 2020. Counties that took no official action to opt out of the state mask mandate or adopted their own mask mandate shortly before or after the state mandate were considered to have a mask mandate in place. Counties were considered to not have a mask mandate in place if they took official action to opt out of the state mask mandate and did not adopt their own mask mandate or if their official action used only the language of guidance (e.g., "should" or "recommend"). among mandated counties and the continued increase in cases in nonmandated counties adds to the evidence supporting the importance of wearing masks and implementing policies requiring their use to mitigate the spread of SARS-CoV-2 (1-6). Community-level mitigation strategies emphasizing wearing masks, maintaining physical distance, staying at home when ill, and enhancing hygiene practices can help reduce transmission of SARS-CoV-2.
The Kansas mandate requiring the wearing of face coverings in public spaces became effective July 3, 2020. Data on county mask mandates were obtained from the Kansas Health Institute. ¶ A Kansas state law** enacted on June 9, 2020, authorizes counties to issue public health orders that are less stringent than the provisions of statewide executive orders issued by the governor, which allowed counties to opt out of the state mask mandate. For this study, counties in Kansas that, as of August 11, 2020, did not opt out of the state mandate or adopted their own mask mandate were considered to have a mask mandate in place; those that opted out of the state mandate and did not adopt their own mask mandate were considered to not have a mask mandate in place.
Daily county-level COVID-19 incidence (cases per 100,000 population) was calculated using case and population counts accessed from USAFacts † † for Kansas counties during June 1-August 23. § § Rates were calculated as 7-day rolling averages. Segmented regression ¶ ¶ was used to examine changes in COVID-19 incidence before and after July 3, 2020, among mandated and nonmandated counties. Mandated and nonmandated counties were compared to themselves over time, ¶ https://www.khi.org/policy/article/20-25. https://www.khi.org/assets/uploads/ news/15015/august_11_update1105.pdf. ** https://ag.ks.gov/docs/default-source/documents/addendum-3-to-march-24law-enforcement-duties-and-authorities-memo.pdf?sfvrsn = d088af1a_3. † † https://usafacts.org/visualizations/coronavirus-covid-19-spread-map. § § August 23, 2020, was selected as the study end date because most Kansas counties had already started or were about to begin school the week of August 24, 2020. The implementation of in-person schooling would have signified an important change in events influencing COVID-19 incidence rates after the executive order. ¶ ¶ Generalized estimating equation regression modeling with an autoregressive correlation variance structure was used to estimate trends over time within counties. Trends in 7-day rolling average of daily COVID-19 incidence among mask mandated counties and among non-mask-mandated counties were analyzed separately before (June 1-July 2, 2020) and after (July 3-August 23, 2020) the governor's executive order requiring masks, effective July 3.
allowing for the control of constant county-related characteristics (e.g., urbanicity or rurality) that might otherwise confound a comparison between mandated and nonmandated counties. Sensitivity analyses were also conducted by 1) examining incidence trends after July 3 separately among mandated counties with and without other public health mitigation strategies and 2) recategorizing nonmandated counties that included cities mandating masks (n=6) as mandated counties. Analyses were conducted using SAS software (version 9.4; SAS Institute). As of August 11, 24 (23%) Kansas counties had a mask mandate in place, and 81 did not. Mandated counties accounted for two thirds of the Kansas population (1,960,703 persons; 67.3%)*** and were spread throughout the state, although they tended to cluster together. Six (25%) mandated and 13 (16%) nonmandated counties were metropolitan areas. † † † Thirteen (54%) mandated counties and seven (9%) nonmandated counties had implemented at least one other public health mitigation strategy not related to the use of masks (e.g., limits on size of gatherings and occupancy for restaurants). During June 1-7, 2020, the 7-day rolling average of daily COVID-19 incidence among counties that ultimately had a mask mandate was three cases per 100,000, and among counties that did not, was four per 100,000 (Table). By the week of the governor's executive order requiring masks (July 3-9), COVID-19 incidence had increased 467% to 17 per 100,000 in mandated counties and 50% to six per 100,000 among nonmandated counties. By August 17-23, 2020, the 7-day rolling average COVID-19 incidence had decreased by 6% to 16 cases per 100,000 among mandated counties and increased by 100% to 12 per 100,000 among nonmandated counties.

Discussion
After implementation of mask mandates in 24 Kansas counties, the increasing trend in COVID-19 incidence reversed. Although rates were considerably higher in mandated counties than in nonmandated counties by the executive order, rates in mandated counties declined markedly after July 3, compared with those in nonmandated counties. Kansas counties that had mask mandates in place appear to have mitigated the transmission of COVID-19, whereas counties that did not have mask mandates continued to experience increases in cases.
The findings in this report are consistent with declines in COVID-19 cases observed in 15 states and the District of Columbia, which mandated masks, compared with states that did not have mask mandates (7). Mask requirements were also implemented as part of a multicomponent approach in Arizona, where COVID-19 incidence stabilized and then decreased after implementation of a combination of voluntary and enforceable community-level mitigation strategies, including mask requirements, limitations on public events, enhanced sanitation practices, and closures of certain services and businesses (8). The combining of community-level mitigation strategies including physical distancing and enhanced hygiene practices, in addition to consistent and correct use of masks, is a CDC-recommended approach. § § § The decreased COVID-19 incidence among mask-mandated counties in Kansas occurred during a time when the only other state mandates issued were focused on mitigation strategies for schools as they reopened in mid-August. In at least 13 (54%) of the 24 mandated counties, the mask mandates occurred alongside other county-level recommended or mandated mitigation strategies (e.g., limits on size of gatherings and occupancy for restaurants), facilitating a potential synergistic effect resulting from combining community mitigation strategies. However, in sensitivity analyses, similar decreases in COVID-19 incidence after July 3 were observed among mandated counties with and without other mitigation strategies. Therefore, although implementing multiple mitigation strategies is the recommended approach, strategies related to mask use mandates appear to be important. Additional information on the utility and acceptability of mask mandates in public settings could help further inform health education campaigns aimed at increasing proper use of masks and strengthening mandate adherence.
The findings in this report are subject to at least four limitations. First, the ecologic design of this study and limited information on community mask-wearing behaviors and county implementation and enforcement provisions of mask mandates limit the ability to determine the extent to which the countywide mask mandates accounted for the observed declines in COVID-19 incidence in mandated counties. Second, this analysis did not account for mask ordinances in six cities in non-mask-mandated counties. However, in sensitivity analyses recategorizing nonmandated counties that included cities mandating masks as mandated counties, results were consistent with those in primary analyses, although they were attenuated. In those analyses, after the governor's executive order, COVID-19 incidence among mandated counties stabilized rather than decreased, and incidence continued to increase among nonmandated counties. Third, although the design of this study limits potential confounding from constant county-related characteristics, the findings in this report are conditional on the absence of any time-varying factors (e.g., mobility patterns, changes in other community-level mitigation strategies, and access to testing) within counties before and after July 3. Nonetheless, in additional analyses examining testing data among Kansas counties during the study period, testing rates were observed to increase overall over time. Therefore, despite increases in testing during this period, decreases in COVID-19 incidence were observed in mandated counties after July 3. Finally, counties in Kansas with a mask mandate might not be representative of other U.S. counties. However, the findings are consistent with observations from other states that mask mandates are associated with declines in COVID-19 cases (7).

Summary
What is already known about this this topic?
Wearing face masks in public spaces reduces the spread of SARS-CoV-2.
What is added by this report?
The governor of Kansas issued an executive order requiring wearing masks in public spaces, effective July 3, 2020, which was subject to county authority to opt out. After July 3, COVID-19 incidence decreased in 24 counties with mask mandates but continued to increase in 81 counties without mask mandates.
What are the implications for public health practice?
Countywide mask mandates appear to have contributed to the mitigation of COVID-19 transmission in mandated counties. Community-level mitigation strategies emphasizing use of masks, physical distancing, staying at home when ill, and enhanced hygiene practices can help reduce the transmission of SARS-CoV-2.
Masks are an important intervention for mitigating the transmission of SARS-CoV-2 (1-6), and countywide mask mandates appear to have contributed to the mitigation of COVID-19 spread in Kansas counties that had them in place. Community-level mitigation strategies emphasizing use of masks, physical distancing, staying at home when ill, and enhanced hygiene practices can help reduce the transmission of SARS-CoV-2.

The Advisory Committee on Immunization Practices' Ethical Principles for Allocating Initial Supplies of COVID-19 Vaccine -United States, 2020
Nancy To reduce the spread of SARS-CoV-2, the virus that causes coronavirus disease 2019 (COVID-19) and its associated impacts on health and society, COVID-19 vaccines are essential. The U.S. government is working to produce and deliver safe and effective COVID-19 vaccines for the entire U.S. population. The Advisory Committee on Immunization Practices (ACIP)* has broadly outlined its approach for developing recommendations for the use of each COVID-19 vaccine authorized or approved by the Food and Drug Administration (FDA) for Emergency Use Authorization or licensure (1). ACIP's recommendation process includes an explicit and transparent evidence-based method for assessing a vaccine's safety and efficacy as well as consideration of other factors, including implementation (2). Because the initial supply of vaccine will likely be limited, ACIP will also recommend which groups should receive the earliest allocations of vaccine. The ACIP COVID-19 Vaccines Work Group and consultants with expertise in ethics and health equity considered external expert committee reports and published literature and deliberated the ethical issues associated with COVID-19 vaccine allocation decisions. The purpose of this report is to describe the four ethical principles that will assist ACIP in formulating recommendations for the allocation of COVID-19 vaccine while supply is limited, in addition to scientific data and implementation feasibility: 1) maximize benefits and minimize harms; 2) promote justice; 3) mitigate health inequities; and 4) promote transparency. These principles can also aid state, tribal, local, and territorial public health authorities as they develop vaccine implementation strategies within their own communities based on ACIP recommendations.
The ACIP COVID-19 Vaccines Work Group has met several times per month (approximately 25 meetings) since its establishment in April 2020. Work Group discussions included * The ACIP includes 15 voting members responsible for making vaccine recommendations. Fourteen of the members have expertise in vaccinology, immunology, pediatrics, internal medicine, nursing, family medicine, virology, public health, infectious diseases, and/or preventive medicine; one member is a consumer representative who provides perspectives on the social and community aspects of vaccination. In addition to the 15 voting members, ACIP includes eight ex officio members who represent other federal agencies with responsibility for immunization programs in the United States, and 30 nonvoting representatives of liaison organizations that bring related immunization expertise. https://www. cdc.gov/vaccines/acip/members/index.html.
review of the epidemiology of COVID-19 and consultation with experts in ethics and health equity to inform the development of an ethically principled decision-making process. The Work Group reviewed the relevant literature, including frameworks for pandemic influenza planning and COVID-19 vaccine allocation (3)(4)(5)(6)(7)(8); summarized this information; and presented it to ACIP. ACIP supported four fundamental ethical principles to guide COVID-19 vaccine allocation decisions in the setting of a constrained supply. Essential questions that derive from these principles can assist in vaccine allocation planning (Table 1).

Maximize benefits and minimize harms.
Allocation of COVID-19 vaccine should maximize the benefits of vaccination to both individual recipients and the population overall. These benefits include the reduction of SARS-CoV-2 infections and COVID-19-associated morbidity and mortality, which in turn reduces the burden on strained health care capacity and facilities; preservation of services essential to the COVID-19 response; and maintenance of overall societal functioning. Identification of groups whose receipt of the vaccine would lead to the greatest benefit should be based on scientific evidence, accounting for those at highest risk for SARS-CoV-2 infection or severe COVID-19-related disease or death, and the essential role of certain workers. The ability of essential workers, including health care workers and non-health care workers, to remain healthy has a multiplier effect (i.e., their ability to remain healthy helps to protect the health of others or to minimize societal and economic disruption). Some of these workers are at increased risk for SARS-CoV-2 infection because of their limited ability to maintain physical distance in the workplace or because they do not have consistent access to recommended personal protective equipment.
Promote justice. Inherent in the principle of justice is an obligation to protect and advance equal opportunity for all persons to enjoy the maximal health and well-being possible. Justice rests on the belief in the fundamental value and dignity of all persons. Allocation of COVID-19 vaccine should promote justice by intentionally ensuring that all persons have equal opportunity to be vaccinated, both within the groups recommended for initial vaccination, and as vaccine becomes more widely available. This includes a commitment to removing unfair, unjust, and avoidable barriers to vaccination that disproportionately affect groups that have been economically or What groups are essential to the COVID-19 response?
What groups are essential to maintaining critical functions of society?
What are the important characteristics of these groups (e.g., size or geographic distribution) that might inform the magnitude of benefit based on the amount of vaccine available or its characteristics?

Promote justice
Does the allocation plan result in fair and equitable access of the vaccine for all groups?
How do characteristics of the vaccine and logistical considerations affect fair access for all persons?
Does allocation planning include input from groups who are disproportionately affected by COVID-19 or face health inequities resulting from social determinants of health, such as income and health care access?

Mitigate health inequities
Does the plan identify and address barriers to vaccination among any groups who are disproportionately affected by COVID-19 or who face health inequities resulting from social determinants of health, such as income and health care access?
Does the allocation plan contribute to a reduction in health disparities in COVID-19 disease and death?
What health inequities might inadvertently result from the allocation plan, and what interventions could remove or reduce them?
Is there a mechanism for timely assessment of vaccination coverage among groups experiencing disadvantage and the possibility for course correction if inequities are identified?

Promote transparency
How does development of the allocation plan include diverse input, and if possible, public engagement?
Are the allocation plan and evidence-based methods publicly available?

Is the allocation plan clear about what is known and unknown and about the quality of available evidence?
What is the process for revision of allocation plans based on new information?
Is there a mechanism to report demographic data elements for vaccine recipients (e.g., age, race/ethnicity, and occupation) to support equitable vaccination coverage?
socially marginalized, as well as a fair and consistent implementation process. Input from a range of external entities, partners, and community representatives is particularly important in developing and assessing allocation plans. Mitigate health inequities. Health equity is achieved when every person has the opportunity to attain his or her full health potential and no one is disadvantaged from achieving this potential because of social position or other socially determined circumstances. † Disparities in the severity of COVID-19 and COVID-19-related death, as well as inequities in social determinants of health that are linked to COVID-19 risk, such as income or health care access and utilization, are well documented among certain racial and ethnic minority groups (9). Vaccine allocation strategies should aim to both reduce existing disparities and to not create new disparities. Efforts should be made to identify and remove obstacles and barriers to receiving COVID-19 vaccine, including limited access to health care or residence in rural, hard-to-reach areas.
Promote transparency. Transparency relates to the decision-making process and is essential to building and maintaining public trust during vaccine program planning and † https://www.cdc.gov/chronicdisease/healthequity/index.htm.
implementation. The underlying principles, decision-making processes, and plans for COVID-19 vaccine allocation must be evidence-based, clear, understandable, and publicly available. To the extent possible, considering the urgency of the COVID-19 response, public participation in the creation and review of the decision-making process should be facilitated. In addition, when feasible, tracking administration of vaccine to the groups recommended for initial vaccine allocation can contribute to transparency and trust in the process. In an ongoing public health response, the situation continually evolves as new information becomes available. Transparency includes being clear about the level of certainty in the available evidence and communicating new information that might change recommendations in a timely fashion.
For the period when the supply of COVID-19 vaccine will be limited, ACIP has considered four groups for initial vaccine allocation. These include health care personnel, other essential workers, adults with high-risk medical conditions, and adults aged ≥65 years (including residents of long-term care facilities) ( Table 2). These groups were selected based on available scientific data, vaccine implementation considerations, and ethical principles. The principle of transparency is applied across the entirety of the vaccine allocation decision-making process. Abbreviations: COVID-19 = coronavirus disease 2019; HCP = health care personnel. * Health care personnel: paid and unpaid persons serving in health care settings who have the potential for direct or indirect exposure to patients or infectious materials; other essential workers: person who conduct operations vital for continuing critical infrastructure, such as food, agriculture, transportation, education, and law enforcement; adults with high risk medical conditions: adults who have one or more high-risk medical conditions, such as obesity, diabetes, and cardiovascular disease; adults aged ≥65 years: includes adults living at home and approximately 3 million living in long-term care facilities. There is considerable overlap between groups, for example, many adults aged ≥65 years also have high-risk medical conditions. † Essential workers during the COVID-19 response have been defined by the U.S. Department of Homeland Security Cybersecurity and Infrastructure Security Agency. https://www.cisa.gov/sites/default/files/publications/Version_4.0_CISA_Guidance_on_Essential_Critical_Infrastructure_Workers_FINAL%20AUG%2018v2_0.pdf. § Medical conditions considered high-risk are updated routinely based on the best available scientific data: https://www.cdc.gov/coronavirus/2019-ncov/needextra-precautions/people-with-medical-conditions.html. ¶ The ability of one or more groups to remain healthy helps protect the health of others and/or minimize disruption to society and the economy. ** As of October 31, 2020, nearly 90% of persons with COVID-19-associated hospitalizations have at least one high-risk condition. Data are routinely updated through COVID-19-Associated Hospitalization Surveillance Network (COVID-NET) (https://gis.cdc.gov/grasp/COVIDNet/COVID19_5.html); in-hospital deaths reported to COVID-NET during March-May, 2020 were associated with certain underlying medical conditions (https://academic.oup.com/cid/advance-article/doi/10.1093/ cid/ciaa1012/5872581). † † As of November 12, 2020, 80% of COVID-19 deaths were among adults aged ≥65 years. Data are routinely updated through CDC case-based surveillance (https:// covid.cdc.gov/covid-data-tracker/#demographics); long-term care residents account for a large proportion of deaths among adults aged ≥65 years (https://data. cms.gov/stories/s/COVID-19-Nursing-Home-Data/bkwz-xpvg/). § § Influenza vaccination coverage is low among many non-health care essential workers; such coverage is lowest among construction workers (

Summary
What is already known about this topic?
During the period when the U.S. supply of COVID-19 vaccines is limited, the Advisory Committee on Immunization Practices (ACIP) will make vaccine allocation recommendations.
What is added by this report?
In addition to scientific data and implementation feasibility, four ethical principles will assist ACIP in formulating recommendations for the initial allocation of COVID-19 vaccine: 1) maximizing benefits and minimizing harms; 2) promoting justice; 3) mitigating health inequities; and 4) promoting transparency.
What are the implications for public health practice?
Ethical principles will aid ACIP in making vaccine allocation recommendations and state, tribal, local, and territorial public health authorities in developing vaccine implementation strategies based on ACIP's recommendations.
ACIP's meetings are open to the public, meeting minutes and archived webcasts are available online, and data (including data from vaccine clinical trials) and analytic methods used in developing ACIP recommendations are publicly available. § Members of the public are invited to submit written comments to the Federal Register or provide oral comment during ACIP meetings. ACIP's 30 nonvoting representatives from liaison organizations facilitate engagement with professional medical and public health organizations and other stakeholders and partners. All four groups proposed for initial allocation of COVID-19 vaccine merit strong consideration from an ethical perspective. Current planning scenarios estimate, however, that the expected number of doses during the first weeks of vaccine distribution might only be sufficient to vaccinate approximately 20 million persons. ¶ Although there is considerable overlap between groups ** (10), the initial supply will not be adequate to vaccinate the entirety of all four groups; for example, there are approximately 100 million health care personnel and essential workers ( Table 2). Published frameworks for COVID-19 allocation and ACIP discussions indicate a clear consensus that the first allocation of COVID-19 vaccine supplies should be directed to health care personnel (1,(5)(6)(7)(8); discussion of allocation to the other three groups is ongoing. As additional vaccine supplies become available, other groups may be vaccinated concurrent with health care personnel. § https://www.cdc.gov/vaccines/acip/index.html. ¶ https://www.cdc.gov/vaccines/imz-managers/downloads/COVID-19-Vaccination-Program-Interim_Playbook.pdf. ** There is overlap among these four groups. For example, in one analysis, among the 3.8% of U.S. adults who work directly with patients as health care workers, 38.6% have high-risk medical conditions or are aged >65 years.

Discussion
During a pandemic, ethical guidelines can help steer and support decisions around prioritization of limited resources (3,4). Consideration of ethical values and principles has featured prominently in discussions about allocation of COVID-19 vaccines. This consideration is particularly relevant because the COVID-19 pandemic has highlighted long-standing, systemic health and social inequities. Although various frameworks for COVID-19 vaccine allocation demonstrate differences in their structure (e.g., based on varying combinations of different goals, objectives, criteria, and other structural elements) and emphasis (e.g., inclusion of global and national considerations), nearly all reference values and principles similar to those which ACIP considers fundamental (5)(6)(7)(8). ACIP viewed the following characteristics as critical for its ethical approach to COVID-19 vaccine allocation when supply is limited: simplicity in structure and definitions; acceptability to stakeholders; and ease of application, both at the national and state, tribal, local, and territorial levels.
Allocation of limited vaccine supplies is complicated by efforts to address the multiple goals of a vaccine program, most notably those related to the reduction of morbidity and mortality and the minimization of disruption to society and the economy. If the goals of a pandemic vaccination program are not clearly articulated and prioritized, drawing distinctions between groups that merit consideration for allocation of vaccine when supply is constrained can become difficult. The unanimity in opinion for early vaccination of health care personnel indicates that maintenance of health care capacity has emerged as a high priority in the context of a severe pandemic. This perspective aligns with ethical considerations for pandemic influenza planning (3,4). If vaccine supply remains constrained, it might be necessary to identify subsets of other groups for subsequent early allocation of COVID-19 vaccine. At the national, state, tribal, local, and territorial levels, such decisions should be guided, in part, by ethical principles and consideration of essential questions, with particular consideration of mitigation of health inequities in persons experiencing disproportionate COVID-19 morbidity and mortality. In the setting of a constrained supply, the benefits of vaccination will be delayed for some persons; however, as supply increases, there will eventually be enough vaccine for everyone.
In addition to ethical considerations, ACIP's recommendations regarding receipt of the initial allocations of COVID-19 vaccine during the period of constrained supply will be based on science (e.g., available information about the vaccine's characteristics such as safety and efficacy in older adults and epidemiologic risk) and feasibility of implementation (e.g., storage and handling requirements). Thus, ACIP's allocation recommendations will be made in conjunction with specific recommendations for the use of each FDA-authorized or licensed COVID-19 vaccine. Although the ethical principles in this report are fundamental for stewardship of limited vaccine supply, they can also be applied when COVID-19 vaccines are widely available, to ensure equitable and just access for all persons.

Timing of Introduction of Complementary Foods -United States, 2016-2018
Katelyn V. Chiang, MPH 1,2 ; Heather C. Hamner, PhD 1 ; Ruowei Li, MD 1 ; Cria G. Perrine, PhD 1 The American Academy of Pediatrics (AAP) recommends introducing complementary foods (i.e., any solid or liquid other than breast milk or infant formula) to infants at approximately age 6 months (1). Although a consensus on ideal timing is lacking, most experts agree that introduction of complementary foods before age 4 months is too early because of infant gastrointestinal and motor immaturity (1,2). In addition, early introduction prevents exclusively breastfed infants from reaching the recommended 6 months of exclusive breastfeeding (1) and might be associated with increased risk for overweight and obesity (3). Nationally representative data on complementary feeding are limited; state-level estimates have been previously unavailable. CDC analyzed 2016-2018 data from the National Survey of Children's Health (NSCH) (N = 23,927) to describe timing of complementary feeding introduction and prevalence of early introduction of complementary foods before age 4 months (early introduction) among children aged 1-5 years. Prevalence of early introduction was 31.9% nationally and varied geographically and across sociodemographic and infant feeding characteristics. These estimates suggest that many infants are introduced to complementary foods before they are developmentally ready. Efforts by health care providers and others who might influence infant feeding practices could help decrease the number of infants who are introduced to complementary foods too early.
NSCH is funded and directed by the Maternal and Child Health Bureau of the Health Resources and Services Administration. It is an annual web-and paper-based survey that collects information from parents and caregivers on their children's physical and emotional health, including infant nutrition, and is representative of noninstitutionalized U.S. children aged 0-17 years. During 2016-2018, the overall weighted response rate ranged from 37.4% to 43.1%. Missing data for race/ethnicity (1.3%) and household income relative to the federal poverty level (FPL) (16.3%) were imputed using hot-deck and sequential regression imputation methods, respectively (4).
Timing of introduction of complementary foods was assessed by asking respondents with children aged 0-5 years "How old was this child when he or she was first fed anything other than breast milk or formula" (4). To ensure that children had sufficient time to have been introduced to complementary foods, analysis was restricted to children aged 1-5 years. Participants with reported introduction to complementary foods at age ≥12 months (887) and those with other implausible feeding patterns (recalled breastfeeding duration, infant formula introduction, and complementary feeding introduction indicated ≥2 months with no source of nutrition; 101) were excluded from analyses. The percentage of children who were introduced to complementary foods before age 4 months (early introduction) was calculated overall, at the state and regional levels, and by sociodemographic and infant feeding characteristics using SAS-callable SUDAAN (version 11.0; RTI International). Two-sample t-tests were used to identify statistically significant (p<0.05) differences across subgroups.
Among 23,927 children aged 1-5 years, the mean age at introduction of complementary foods was 4.7 months, with 31.9% of children introduced at <4 months, 51.0% at 4-6 months, and 17.1% at 7-11 months ( Figure 1). Prevalence of early introduction varied across sociodemographic groups. Prevalence of early introduction was significantly higher among non-Hispanic Black (Black) children (40.5%), compared with all other racial/ethnic groups and significantly lower among non-Hispanic Asian children (23.8%), compared with all other groups except Hispanic children (29.9%). Prevalence of early introduction was significantly lower among children living in households at ≥400% of the FPL (28.5%) and whose mothers had a bachelor's degree or higher (27.7%), compared with all other household FPL and maternal education groups. Early introduction also differed significantly by infant milk feeding status at age 4 months: prevalence of early introduction was 18.5% among children receiving only breast milk for milk feeds, 32.1% among those receiving breast milk and infant formula, and 41.6% among those receiving only infant formula for milk feeds (Table). At the state level, prevalence of early introduction ranged from 18.0% in New Mexico to 49.0% in Mississippi. In 34 states, ≥30% of children were introduced to complementary foods before age 4 months, including 14 states in which prevalence of early introduction was at least 35% (Supplementary Table, https://stacks.cdc.gov/view/ cdc/95035) (Figure 2).

Discussion
Nearly one in three (31.9%) U.S. infants is introduced to complementary foods before age 4 months, with a higher prevalence of early introduction among Black infants and infants of mothers and households at lower socioeconomic status. Reasons for early introduction to complementary foods are not fully understood; however, many early introducing mothers have reported believing that their infant was old enough to begin consuming solids (5). This suggests a perception of infant readiness for complementary feeding before the infant is actually ready and a potential lack of awareness of feeding recommendations, health effects associated with early introduction, and signs of developmental readiness. In general, infants show outward signs of readiness for complementary feeding when they can sit up on their own with good head control, show interest in mealtimes, are hungry in between feedings, and no longer have "tongue-thrust" or extrusion reflex, usually at approximately age 4-6 months (2).
Not only do younger infants lack the physiologic development to safely consume complementary foods, infants who are introduced to complementary foods too early have increased risk for multiple associated health conditions (1). Early introduction to complementary foods prevents infants from meeting the recommended 6 months of exclusive breastfeeding, decreasing the benefits both mothers and infants derive from exclusive breastfeeding. Compared with exclusive breastfeeding for 6 months, exclusive breastfeeding for 3-4 months followed by mixed breastfeeding and complementary feeding is associated with increased risk for gastrointestinal infection and slower maternal weight loss after birth (6). Further, limited evidence also suggests introduction to complementary foods before age 4 months might increase later overweight and obesity risk (3).
Health care providers can help increase awareness of recommended timing of introduction of complementary foods by employing consistent messaging in accordance with AAP recommendations and stressing the importance of developmental readiness when discussing complementary feeding with families (1). Resources are available to help health care providers engage with and educate families to better navigate the introduction (16.3%) than the estimate from NSCH participants (31.9%); however, similar patterns in early introduction by sociodemographic and infant feeding status characteristics were seen across both studies (8). The questions used to identify timing of complementary feeding introduction were the same for both studies. Some of the discrepancy might be explained by inherent differences in the surveys including representativeness of participants and response rates or by the different age ranges of studied children (6-36 months in NHANES compared with 1-5 years in NSCH). Differences might also reflect changes over time in parental attitudes toward complementary feeding or health care provider advice because participants included in the NHANES and NSCH analyses were born during 2006-2014 and 2010-2018, respectively. Over the past decade, there has been growing awareness of the benefits of not delaying introduction of allergenic foods among children at high risk for food allergies to prevent the development of food allergies (9). It is possible that research might have been misinterpreted by parents, caregivers, and health care providers, leading to increases in early introduction to complementary foods in recent years. Further education on correct timing of introduction of allergenic foods and identification of children at high risk for food allergies might be needed to improve adherence to feeding recommendations.
The findings in this report are subject to at least four limitations. First, data might be affected by information bias. Though maternal recall of breastfeeding has been shown to have high validity and reliability, recall of solid and other liquid feeding

Summary
What is already known about this topic?
The American Academy of Pediatrics recommends introducing complementary foods at approximately age 6 months. Introduction before age 4 months is too early because infants are not developmentally ready for complementary foods. Early introduction prevents infants from reaching the recommended 6 months of exclusive breastfeeding.
What is added by this report?
Nearly one in three infants is introduced to complementary foods before age 4 months; prevalence of early introduction varies geographically and across sociodemographic and infant feeding characteristics.
What are the implications for public health practice?
Increasing awareness of and adherence to feeding recommendations could help reduce early introduction. Health care providers and others who might influence infant feeding practices should educate families on recommended timing of introduction of complementary foods.
might not be as reliable (10). However, participants with implausible feeding patterns were removed from the sample to account for potential misreporting of infant feeding information. Second, although multiply imputed, household FPL data might be misclassified. Third, data do not allow for analysis of types, amounts, or frequency of complementary foods offered; these are important markers of early child nutrition. Finally, small sample sizes limited the ability to conduct further sociodemographic analyses at the state level.
Introduction of complementary foods at the recommended time could help improve infant health and might play a role in prevention of overweight and obesity; however, nearly one third of infants are introduced to complementary foods too early. Early introduction also varies geographically and across sociodemographic and infant feeding characteristics, placing some infants, such as Black infants and infants of mothers and households of lower socioeconomic status, at increased risk for potential poor health outcomes related to early introduction of complementary foods. Increased education on complementary feeding recommendations, including the possible effects of early introduction and signs of developmental readiness, might help decrease the number of infants who are introduced to complementary foods too early. federal poverty level; adults with health insurance; unemployed adults; and adults residing in the Northeast, the Midwest, or the South (Table). Among those without serious psychological distress, current cigarette smoking among adults with active epilepsy was higher (22.4%) than it was among adults without epilepsy (15.7%).
Current cigarette smoking prevalence among adults with inactive epilepsy was higher than that among adults with no history of epilepsy in many of the same subgroups. Current cigarette smoking was also higher among adults with inactive epilepsy than it was among those without epilepsy for any age group; those with <12 or >12 years of education; those with family incomes <100% or >300% of the federal poverty level; among both the insured or the uninsured; the employed; the retired; those in other employment categories (e.g., students); and those residing in all U.S. regions. Among adults without serious psychological distress, cigarette smoking prevalence among those with inactive epilepsy was higher than that among those without epilepsy.

Discussion
During the 4 survey years (2010, 2013, 2015 and 2017), approximately one in four U.S. adults with active or inactive epilepsy currently smoked cigarettes. This finding reinforces the importance of efforts to reduce cigarette smoking among all adults, especially those with any epilepsy.
Differences in current smoking among adults with epilepsy within subgroups generally paralleled those in the general U.S. adult population, with higher prevalences among some racial/ ethnic minorities and those with lower income, a disability, or serious psychological distress (2). Like the general population, adults with epilepsy have reported challenges in maintaining healthful behaviors but might benefit from interventions that increase skills for adopting and maintaining healthy behaviors (8). Cigarette smoking is especially complex in epilepsy because nicotine and tobacco smoke have both proconvulsant effects (e.g., reducing the anticonvulsive effects of certain antiseizure drugs) and anticonvulsant effects, which has been demonstrated in various human studies and animal models (5,9). Although one study found that smokers with epilepsy were approximately four times more likely to have experienced a

Summary
What is already known on this topic?
Studies have shown that cigarette smoking is as common, and sometimes more so, among adults with a history of epilepsy as it is among those without a history of epilepsy.
What is added by this report?
During 2010-2017, one in four adults with active or inactive epilepsy were current smokers, compared with one in six persons without epilepsy. Although fewer adults with active epilepsy smoked cigarettes in 2017 than in 2010, this difference was not statistically significant.
What are the implications for public health practice?
Epilepsy health and social service providers should promote smoking cessation resources to adults with active epilepsy who smoke cigarettes to help them quit smoking and reduce their risk for smoking-related disease and death. seizure in the past year than were nonsmokers with epilepsy, further research is needed to identify associations between seizure control, current smoking, and smoking cessation in representative samples of persons with active epilepsy (5,9).
Possible differences in smoking trends between adults with active and inactive epilepsy might be associated with differences in overall health status, work limitations, and quality of life between these two groups (7), but this will require further study. Encouraging smoking prevention and cessation among all adults, including those with epilepsy and other population groups with disproportionately higher prevalences of smoking, is critical to reducing their risk of smoking-related disease and death.
The findings in this report are subject to at least six limitations. First, because epilepsy and smoking status were self-reported and not validated by clinical chart review or biochemical testing, these classifications are subject to social desirability bias, interviewer effects, and misclassification of epilepsy. Second, because NHIS excludes institutionalized populations such as the military, detained or incarcerated persons, and nursing home residents, results are not generalizable to these groups. Third, assessment of subgroup differences within a relatively small sample of adults with epilepsy can obscure differences within these populations. Fourth, the NHIS survey response rate has declined from 72% in 2010 to 53% in 2017, resulting in increasing nonresponse bias, which might result in less representative samples of U.S. adults with epilepsy participating in NHIS over time. Fifth, assessment of trends in smoking prevalence among those with active epilepsy might have been underpowered because of sample size limitations. Finally, this study assessed cigarettes only, and not other forms of tobacco products; given that nicotine has been found to have both proconvulsant and anticonvulsant effects, further research on any relationship between epilepsy and the use of noncigarette tobacco products, including e-cigarettes, is warranted. Health and social service providers who interact with persons with active epilepsy should ensure that smoking cessation information and resources are available to them and should encourage persons who smoke to use these resources to help them quit smoking and to reduce their risk of smoking-related disease and death. Funding state tobacco control programs, including state quit lines, at CDC-recommended levels, increasing tobacco prices, implementing comprehensive smoke-free policies, conducting antitobacco mass media campaigns, and enhancing access to quitting assistance could increase tobacco cessation and reduce tobacco-related disease and death among all adults, including those with epilepsy † † (1,10). Insurers and employers could improve coverage and increase use of cessation treatment, and health systems can integrate cessation interventions into clinical care (1). US Department of Health and Human Services/Centers for Disease Control and Prevention Age group (yrs) 100 * With 95% confidence intervals shown with error bars. † Based on the response to a survey question that asked "In the past 3 months, how often did you have pain? Would you say never, some days, most days, or every day?" Chronic pain was defined as pain on most days or every day in the past 3 months. § Estimates are based on household interviews of a sample of the civilian, noninstitutionalized U.S. population and are derived from the National Health Interview Survey Sample Adult component.
During 2019, military veterans aged ≥20 years were more likely to have chronic pain than were nonveterans (31.5% versus 20.1%). By age group, the likelihood of having chronic pain was higher among veterans than nonveterans for those aged 20-34 years (27.1% versus 9.4%), 35-49 years (27.7% versus 17.7%), and 50-64 years (37.2% versus 26.3%). Among those aged ≥65 years, prevalence of chronic pain did not differ significantly by veteran status (30.8% among veterans versus 31.0% among nonveterans). Among nonveterans, the prevalence of chronic pain increased with age. Among veterans, those aged 50-64 years had the highest prevalence of chronic pain. For more information on this topic, CDC recommends the following link: https://www.cdc.gov/drugoverdose/prescribing/guideline.html.