Discussion

Throughout the analysis period, opioid prescribing rates by primary care providers were significantly higher in nonmetropolitan counties than in metropolitan counties. Whereas the prescribing rate increased from January 2014 through January 2015 (period 1) in both micropolitan and noncore counties, those trends halted, and rates became flat or declined through mid-March 2016 (period 2). Trends in all other urban-rural categories were flat or decreasing over the same two periods. The odds of a patient receiving an opioid prescription decreased in all urban-rural county groups after the March 2016 publication of the CDC Guideline. Those trends represented significant decreases in the micropolitan and large central metropolitan categories. In the other four county groups, however, the significant decreases after March 2016 represented a continuation of previously decreasing trends.

Higher odds of opioid prescribing in nonmetropolitan counties might be attributed in part to prescription drug use and misuse at an earlier age as well as higher prevalences of chronic pain among persons living in rural areas (6,7). Nonmetropolitan counties also tend to have larger populations of older adults who have higher prevalences of conditions associated with pain (6). Opioid prescribing in rural (nonmetropolitan) areas is strongly influenced by providers’ individual relationships with their patients (8), and can be inconsistent with opioid prescribing guidelines. As well, access to medication-assisted treatment facilities and alternative therapies are limited in rural areas (8). Variations in the implementation of state-run prescription drug monitoring programs and state-based laws (9), such as the regulation of pain-management clinics, might also differ in urban and rural communities.

Despite reductions in opioid prescribing in recent years (1), opioid-involved overdose death rates have increased, largely driven by heroin and illicitly manufactured fentanyl (2). Many persons who self-report heroin use have a history of misusing prescription opioids (10). Addressing prescription opioid use is an important step in curbing opioid-involved overdose deaths. Interventions such as using Prescription Drug Monitoring Programs and practices that align with evidence-based adoption of the CDC Guideline can improve prescribing decisions.** The Guideline can help providers and patients weigh the benefits and risks of prescribing opioids according to best available evidence and individual patient needs (4). This study demonstrates that data from EHRs can effectively supplement traditional surveillance methods for monitoring trends in opioid prescribing and other areas of public health importance. The lag between the collection of the data and this analysis could potentially be reduced to a matter of weeks with optimized workflows.

The findings in this report are subject to at least three limitations. First, the conclusions drawn from the records provided by Athenahealth might not be generalizable to all patients in primary care. Second, although the data include all patients with an opioid prescription, they do not include other characteristics of each prescription, including indication (e.g., chronic versus acute pain or opioid use disorder treated with buprenorphine [although this drug accounted for a small fraction of all opioids prescribed]) and whether prescriptions were filled and taken as prescribed. Finally, this analysis does not account for differing demographic profiles across counties, such as age distributions and payer types, which could be confounded by population density in its association with opioid prescribing rates.

The percentage of patients who received an opioid prescription was lower in more densely populated counties than among less populated rural counties; however, all areas, including rural counties, experienced substantial decreases in prescribing over time. As less densely populated areas appear to indicate both substantial progress in decreasing opioid prescribing and ongoing need for reduction, community health care practices and intervention programs must continue to be tailored to community characteristics.

Discussion

Influenza activity remains elevated in the United States (4). Overall, influenza A(H1N1)pdm09 viruses have predominated in most of the country, although circulation of influenza A(H3N2) and low levels of influenza B viruses have also been observed. Effectiveness of influenza vaccines in reducing the risk for medically attended influenza illness has ranged from approximately 40%–60% across all ages during seasons when most circulating influenza viruses are antigenically like the recommended influenza vaccine components. The overall interim estimate of 47% vaccine effectiveness against influenza A(H1N1)pdm09 in all age groups is similar to that observed during the most recent A(H1N1)pdm09 predominant season (45%) in 2015–16 (3), but lower than a meta-analysis of vaccine effectiveness against A(H1N1)pdm09 since the 2010–11 season in the United States (5). This interim estimate also is lower than the recently reported interim estimates of 72% effectiveness against A(H1N1)pdm09 in Canada during the 2018–19 season (6) and 78% against A(H1N1)pdm09 in Australia during the 2018 Southern Hemisphere influenza season (7). The reasons for these differences might include limited sample size caused by low attack rates in some age groups, geographic differences in circulating viruses, and genetic variation within virus subtypes (4). Of note, vaccine effectiveness against A(H1N1)pdm09 among children and adolescents aged 6 months–17 years (62%) was similar to that observed during the 2015–16 season in this age group (49%–63%) (3). Among adults aged ≥50 years, interim estimates of effectiveness were not significant. Vaccine effectiveness against A(H3N2) virus infection was 44% (95% CI = 13%–64%) but a limited number of A(H3N2) viruses were detected. Several more weeks of influenza are likely, and CDC continues to recommend influenza vaccination while influenza viruses are circulating in the community. Vaccination can protect against infection with influenza viruses that are currently circulating, as well as those that may circulate later in the season.

Vaccination remains the best method for preventing influenza and its potentially serious complications, including those that can result in hospitalization and death. In particular, vaccination has been found to reduce the risk for influenza-associated deaths in children (8). During past seasons, including the 2017–18 season, approximately 80% of reported pediatric influenza-associated deaths have occurred in children who were not vaccinated. Vaccination also has been found to reduce the risk for influenza-associated hospitalization in pregnant women (9) and can reduce the risk for cardiac events among persons with heart disease (10). CDC recommends antiviral treatment for any patient with suspected or confirmed influenza who is hospitalized, has severe or progressive illness, or is at high risk for complications from influenza, regardless of vaccination status or results of point-of-care influenza diagnostic tests.^† Antiviral treatment also can be considered for any previously healthy symptomatic outpatient not at high risk for complications, with confirmed or suspected influenza, if treatment can be started within 48 hours of illness onset.

The findings in this report are subject to at least four limitations. First, sample sizes are smaller than in recent interim reports, resulting in wide confidence intervals, particularly in adults aged ≥50 years. The small sample size also limits the number of age groups included in this analysis. This limitation is common among interim vaccine effectiveness reports during mild or late influenza seasons. End-of-season vaccine effectiveness estimates could change as additional patient data become available or if a change in circulating viruses occurs later in the season. Second, vaccination status included self-report at four of five sites; end-of-season vaccine effectiveness estimates based on updated documentation of vaccination status might differ from interim estimates. For this reason, the type of vaccine received by participants (e.g., egg-based, cell culture–based, or recombinant antigen) is not available at this time, although this information will be updated at the end of the season. Third, an observational study design has greater potential for confounding and bias than do randomized clinical trials. However, the test-negative design is widely used in vaccine effectiveness studies and has been used by the U.S. Flu VE Network to estimate vaccine effectiveness for previous influenza seasons. Finally, the vaccine effectiveness estimates in this report are limited to the prevention of outpatient medical visits rather than more severe illness outcomes, such as hospitalization or death; data from studies measuring vaccine effectiveness against more severe outcomes will be available at a later date.

Vaccination prevents a substantial number of influenza-related illnesses, hospitalizations, and deaths annually. However, better protection and improved vaccination coverage are needed to realize the full potential of influenza vaccines. Evaluation of influenza vaccine effectiveness is an essential component of ongoing efforts to improve influenza vaccines. Influenza activity remains elevated in the United States, highlighting the importance of vaccination. CDC will continue to monitor influenza disease throughout the season to better understand the impact of vaccination, identify factors associated with reduced protection, and support efforts to improve influenza vaccines.

Discussion

Deaths involving cocaine and psychostimulants have increased in the United States in recent years; among 70,237 drug overdose deaths in 2017, nearly a third (23,139 [32.9%]) involved cocaine, psychostimulants, or both. From 2016 to 2017, death rates involving cocaine and psychostimulants each increased by approximately one third, and increases occurred across all demographic groups, Census regions, and in several states. In 2017, nearly three fourths of cocaine-involved and roughly one half of psychostimulant-involved overdose deaths, respectively, involved at least one opioid. After initially peaking in 2006, trends in overall cocaine-involved death rates declined through 2012, when they began to rise again. The 2006–2012 decrease paralleled a decline in cocaine supply coupled with an increase in cost.^¶¶¶ Similar patterns in death rates involving both cocaine and opioids were observed, with increases for cocaine- and synthetic opioid-involved deaths occurring from 2012 to 2017. From 2010 to 2017, increasing rates of deaths involving psychostimulants occurred and persisted even in the absence of opioids. Drug overdoses continue to evolve along with emerging threats, changes in the drug supply, mixing of substances with or without the user’s knowledge, and polysubstance use (3–8). In addition, the availability of psychostimulants, particularly methamphetamine, appears to be increasing across most regions.**** In 2017, among drug products obtained by law enforcement that were submitted for laboratory testing, methamphetamine and cocaine were the most and third most frequently identified drugs, respectively.^†††† Previous studies also found that heroin and synthetic opioids (e.g., illicitly-manufactured fentanyl) have contributed to increases in stimulant-involved deaths (3,9,10). Current findings further support that increases in stimulant-involved deaths are part of a growing polysubstance landscape. Although synthetic opioids appear to be driving much of the increase in cocaine-involved deaths, increases in psychostimulant-involved deaths have occurred largely without opioid co-involvement; however, recent data suggest increasing synthetic opioid involvement in these deaths.

The findings in this report are subject to at least four limitations. First, at autopsy, substances tested for and circumstances under which tests are performed vary by time and jurisdiction. Therefore, recent improvements in toxicologic testing might account for some reported increases. Second, 15% and 12% of death certificates in 2016 and 2017, respectively, did not include mention of specific drugs involved. The percentage of death certificates with at least one drug specified varied widely by state, ranging from 54.7% to 99.3% in 2017, limiting comparisons across states. Third, potential racial misclassification might lead to underestimates for certain groups, primarily AI/AN and A/PI.^§§§§ Finally, certain trend analyses were limited, given small numbers of deaths and the inability to calculate stable rates among some stimulant-opioid drug combinations before 2003.

Preliminary 2018 data indicate continued increases in drug overdose deaths.^¶¶¶¶ The rise in deaths involving cocaine and psychostimulants and the continuing evolution of the drug landscape indicate a need for a rapid, multifaceted, and broad approach that includes more timely and comprehensive surveillance efforts to inform tailored and effective prevention and response strategies. CDC currently funds 45 states and DC for opioid surveillance***** and/or prevention activities.^{†††††} The contribution of opioids to increases in stimulant-involved overdose deaths underscores the importance of continued opioid overdose surveillance and prevention measures, including existing efforts to expand naloxone availability to persons at risk for drug overdose. CDC is expanding drug overdose surveillance efforts to include stimulants and is implementing multiple, evidence-based opioid prevention efforts, such as enhancing linkage to care, building state and local capacity, and public health/public safety collaborations.^§§§§§ Because some stimulant deaths are also increasing without opioid co-involvement, prevention and response strategies need to evolve accordingly. Increased efforts are required to identify and improve access to care for persons using stimulants, implement upstream prevention efforts focusing on shared risk and protective factors that address substance use/misuse, and improve risk reduction messaging (e.g., not using alone). Continued collaborations among public health, public safety, and community partners are critical to understanding the local illicit drug supply and reducing risk as well as linking persons to medication-assisted treatment and risk-reduction services.

Discussion

Nonsmoking workers residing in states with comprehensive smoke-free laws reported significantly lower prevalences of frequent exposure to workplace SHS. Moreover, SHS exposure among nonsmoking workers also significantly varied by industry. During 2013–2014, one in four U.S. nonsmokers reported exposure to SHS (5), and an estimated 41,000 deaths among nonsmoking adults were associated with SHS exposure (1). Furthermore, workplace SHS exposure has been recognized as one of the top occupational hazards that contributes substantially to the prevalence of occupational cancer among nonsmokers (6). During 2000–2015, the number of states with smoke-free laws that prohibited smoking in indoor areas of worksites, restaurants, and bars increased from none to 27 (4). In this report, workers residing in states with smoke-free laws in all three venue categories were least likely to report frequent exposure to workplace SHS. Previous studies have revealed that the absence of a policy restricting or prohibiting smoking at the worksite put workers at higher risk for workplace SHS exposure (7). Despite the considerable progress in implementation of smoke-free laws over the past 2 decades, this analysis found that even in states with smoke-free laws in three categories of venues, 8.6% of nonsmoking workers reported frequent workplace SHS exposure. This finding suggests that certain workplaces might be outside the scope of most smoke-free laws.

Based on NHIS data for 2014–2016, 34.3% of workers in the construction, 30.4% of workers in the mining, and 30.2% of workers in the transportation industries used some form of tobacco (8). Higher smoking prevalences among workers employed in these industries might lead to exposure of their nonsmoking coworkers to SHS. Previous findings of higher tobacco use and SHS exposure among workers in the construction industry are consistent with current findings (3,8). The industry subcategories with the highest prevalences of reported SHS exposure in this study and the industry category with the highest number of exposed workers (construction) include outdoor workplaces and other settings that are unlikely to be protected by smoke-free laws. A recent study determined that indoor workers who reported working at a worksite having a 100% smoke-free policy had significantly lower odds of smoking combustible tobacco than did those reporting a partial or no smoke-free policy (9). Enhanced and sustained efforts to protect nonsmoking workers through comprehensive smoke-free laws and implementation of smoke-free workplace policies by employers can benefit public health.

The findings in this report are subject to at least five limitations. First, all information in NHIS, including work characteristics and SHS exposure, was self-reported at the time of interview and might be subject to reporting bias. Second, although NHIS records state of residence, some workers might work outside the states in which they reside or in multiple states where smoke-free laws might differ. Third, estimates for SHS exposure for some groups were unreliable because of small sample sizes and were therefore suppressed. Small sample sizes within individual industry groups also precluded analyses that combined the state and industry variables. Fourth, the study only accounted for statewide smoke-free policies, and considerable progress has been made in implementing local level smoke-free policies in many states;^§ therefore, workers classified as being unprotected by statewide laws might have been protected by local level laws. Finally, variable distribution of industries by state might have led to some confounding.

Workplace SHS exposure is harmful for workers’ health. In this study, nonsmoking workers residing in states without comprehensive smoke-free laws and those employed in certain industries were more likely to be frequently exposed to workplace SHS. NIOSH encourages employers, especially those in industries with high prevalences of SHS exposure, to implement workplace-specific smoke-free policies to complement state and local smoke-free laws to help reduce SHS exposure among workers and protect workers’ health (10).

Discussion

Surveillance of U.S. breastfeeding duration and exclusivity, including monitoring for Healthy People 2020^¶ objectives, reports estimates among all infants, regardless of whether they had initiated breastfeeding. The findings in this report demonstrate that differences between black and white infants in any and exclusive breastfeeding at ages 3 and 6 months are caused, in part, by racial/ethnic differences in breastfeeding initiation. Interventions to improve breastfeeding initiation and support continuation among black mothers might be important to closing the black-white gap in duration.

Black mothers disproportionately experience a number of barriers to breastfeeding, including lack of knowledge about breastfeeding; lack of peer, family, and social support; insufficient education and support from health care settings; and concerns about navigating breastfeeding and employment (3). Subjective norms, or perceptions of approval from others who are important to the person (e.g., family members), are important drivers of breastfeeding behaviors, particularly among black women (3). Increasing interpersonal support for breastfeeding might help increase breastfeeding initiation and duration among black women, who might lack breastfeeding role models in their social networks and be more likely to face negative perceptions of breastfeeding among their peers and communities (3,4). For example, peer counseling might increase breastfeeding initiation and duration among black mothers (3).

In the United States, the rate of implementation of evidence-based maternity care practices supportive of breastfeeding is lower among maternity care facilities in neighborhoods with larger black populations (5). Hospitals’ use of such practices, which include helping women initiate breastfeeding within the first hour of birth and not providing breastfeeding infants with infant formula without a medical indication, increases rates of breastfeeding initiation, duration, and exclusivity (6). A recent analysis indicated that making improvements in these practices among maternity care facilities in four southern states reduced black-white disparities in breastfeeding initiation (7).

Returning to work is another major barrier to breastfeeding initiation and continuation, particularly for black women (3). A woman’s plans for returning to work are associated with her intention to breastfeed; specifically, women planning to return to work before 12 weeks postpartum, planning to work full-time, or both were less likely to intend to exclusively breastfeed, compared with women planning to return to work after 12 weeks postpartum, planning to work part-time, or both (8). Black women, especially those with a low income, return to work earlier than do women in other racial/ethnic groups and are more likely to experience challenges to breastfeeding or expressing milk, including inflexible work hours (9). Policies that enable taking paid leave after giving birth, flexible work schedules, and support for breastfeeding or expressing milk at work might help improve breastfeeding intention, initiation, and duration.**

The findings in this report are subject to at least three limitations. First, response rates averaged 53.8% for the landline sample and 28.6% for the mobile telephone sample; further, households without a telephone are not represented. The possibility exists that selection bias occurs even after adjusting weights for nonresponse and noncoverage. Second, maternal reports of breastfeeding behaviors could be subject to recall bias because mothers reported these behaviors when their children were aged 19–35 months and to social desirability bias because of a desire to provide socially acceptable responses. However, maternal recall of breastfeeding behavior has been found to be valid and reliable, especially when recalled within 3 years (10). Finally, although this report focuses only on black-white breastfeeding differences, lower rates of breastfeeding duration and exclusivity among Hispanic infants, compared with non-Hispanic white infants, have been documented (3). However, because Hispanic and white infants have similar rates of breastfeeding initiation, the methods applied in this report did not affect estimates of breastfeeding duration and exclusivity.

Breastfeeding provides optimal nutrition to infants and provides health benefits for both infants and mothers, and CDC works to increase breastfeeding rates among all mothers in the United States. In order to address disparities in breastfeeding duration, continued efforts are needed to increase rates of breastfeeding initiation and support continuation of breastfeeding among black women. Closing the black-white gap in breastfeeding duration might require efforts of multiple groups. Families, hospitals, and employers can help black women initiate and continue breastfeeding, thereby providing their infants with optimal nutrition.

Clinical Evaluation for Patients with Suspected EVALI

EVALI is considered a diagnosis of exclusion because, at present, no specific test or marker exists for its diagnosis (Box 1). Health care providers should consider multiple etiologies, including the possibility of EVALI and concomitant infection. In addition, health care providers should evaluate alternative diagnoses as suggested by clinical findings and medical history (e.g., cardiac, gastrointestinal, rheumatologic, and neoplastic processes; environmental or occupational exposures; or causes of acute respiratory distress syndrome) (6).

Patient history. Based upon medical chart abstraction data submitted to CDC, 95% (323/339) of patients diagnosed with EVALI initially experienced respiratory symptoms (e.g., cough, chest pain, and shortness of breath), and 77% (262/339) had gastrointestinal symptoms (e.g., abdominal pain, nausea, vomiting, and diarrhea). Gastrointestinal symptoms preceded respiratory symptoms in some patients (1–3). Respiratory or gastrointestinal symptoms were accompanied by constitutional symptoms such as fever, chills, and weight loss among 85% (289/339) of patients (Table).

All health care providers evaluating patients for EVALI should ask about the use of e-cigarette, or vaping, products and ideally should ask about types of substances used (e.g., THC, cannabis [oil, dabs], nicotine, modified products or the addition of substances not intended by the manufacturer); product source, specific product brand and name; duration and frequency of use, time of last use; product delivery system, and method of use (aerosolization, dabbing, or dripping). Empathetic, nonjudgmental, and private questioning of patients regarding sensitive information to assure confidentiality should be employed. Standardized approaches should be used for interviewing adolescents. Resources exist to guide patient interviews, including those of adolescents.^§ In some situations, asking questions over the course of the hospitalization or during follow-up visits might elicit additional information about exposures, especially as trust is established between the patient and clinicians.

Physical examination. For patients who report the use of e-cigarette, or vaping, products, physical examination should include vital signs and pulse-oximetry. Tachycardia was reported in 55% (169/310) of patients and tachypnea in 45% (77/172); O₂ saturation <95% at rest on room air was present for 57% (143/253) of patients reported to CDC (Table), underscoring the need for routine pulse-oximetry. Among patients identified to date, pulmonary findings on auscultation exam have often been unremarkable, even among patients with severe lung injury (personal communication, Lung Injury Response Clinical Working Group, October 2, 2019).

Laboratory testing. Laboratory testing should be guided by clinical findings. A respiratory virus panel, including influenza testing during influenza season, should be strongly considered. Additional testing should be based on published guidelines for evaluation of community-acquired pneumonia.^¶ Infectious diseases to consider include Streptococcus pneumoniae, Legionella pneumophila, Mycoplasma pneumoniae, endemic mycoses, and opportunistic infections; the likelihood of infection by any of these varies by geographic prevalence and patient medical history. Other abnormal laboratory tests reported in patients with EVALI include elevated white blood cell (WBC) count, serum inflammatory markers (C-reactive protein, erythrocyte sedimentation rate [ESR]), and liver transaminases. In a report of initial patients from Illinois and Wisconsin, 87% had a WBC >11,000/mm³ and 93% had an ESR >30mm/hr; 50% of patients had elevated liver transaminases (aspartate aminotransferase or alanine aminotransferase >35 U/L) (3). However, at this time, these tests cannot be used to distinguish EVALI from infectious etiologies. In all patients, providers should consider conducting, with informed consent, urine toxicology testing, including testing for THC.

Imaging. Radiographic findings consistent with EVALI include pulmonary infiltrates on CXR and opacities on chest computed tomography (CT) scan (1,7). A CXR should be obtained on all patients with a history of e-cigarette, or vaping, product use who have respiratory or gastrointestinal symptoms, particularly when accompanied by decreased O₂ saturation (<95%). Chest CT might be useful when the CXR result does not correlate with clinical findings or to evaluate severe or worsening disease, complications such as pneumothorax or pneumomediastinum, or other illnesses in the differential diagnosis, such as pneumonia or pulmonary embolism. In some cases, chest CT has demonstrated findings such as bilateral ground glass opacities despite a normal or nondiagnostic CXR (3). Among patients with abnormal CXR findings and a clinical picture consistent with EVALI, a chest CT scan might not be necessary for diagnosis. The decision to obtain a chest CT should be made on a case-by-case basis depending on the clinical circumstances.

Consultation with specialists. Consultation with several specialists might be necessary to optimize patient management. For patients being evaluated for possible EVALI, consideration should be given to consultation with a pulmonologist, who can help guide further evaluation, recommend empiric treatment, and review the indications for bronchoscopy. The decision to perform bronchoscopy and bronchoalveolar lavage (BAL) to rule out alternative diagnoses such as pulmonary infection should be made on a case-by-case basis. The value of staining BAL cells or fresh lung biopsy tissue for lipid-laden macrophages (e.g., using oil red O or Sudan Black) in the evaluation of EVALI remains unknown. In addition, there should be a low threshold for consulting with critical care physicians, because, based upon data submitted to CDC, 47% (159/342) of patients were admitted to an intensive care unit and 22% (74/338) required endotracheal intubation and mechanical ventilation (Table); critical care physicians should be consulted to determine optimal management of respiratory failure. Consultation with medical toxicology, infectious disease, psychology, psychiatry, addiction medicine, and other specialists should be considered as warranted by patient circumstances.

Management of Patients with Suspected EVALI

Admission criteria and outpatient management. Several factors should be considered when deciding whether to admit a patient with potential EVALI to the hospital (Box 2). Among 1,002 cases reported to CDC with available data as of October 8, 96% of patients were hospitalized. Patients with suspected EVALI should be admitted if they have decreased O₂ saturation (<95%) on room air, are in respiratory distress, or have comorbidities that compromise pulmonary reserve. Consider modifying factors such as altitude to guide interpretation of measured O₂ saturation.

Outpatient management of suspected EVALI might be considered on a case-by-case basis for patients who are clinically stable, have less severe injury, and for whom follow-up within 24–48 hours of initial evaluation can be assured. Candidates for outpatient management should have normal O₂ saturation (≥95%), reliable access to care, and strong social support systems. For these patients, empiric use of antimicrobials, including antivirals, if indicated, should be considered. Some patients who initially had mild symptoms experienced a rapid worsening of symptoms within 48 hours. In Illinois and Wisconsin, 72% of patients had either an outpatient or emergency department visit before seeking additional medical care that resulted in hospital admission (3). Health care providers should instruct all patients to seek medical care promptly if respiratory symptoms worsen.

Medical treatment. Corticosteroids might be helpful in treating this injury. Several case reports describe improvement with corticosteroids, likely because of a blunting of the inflammatory response (3–5). In a series of patients in Illinois and Wisconsin, 92% of 50 patients received corticosteroids; the medical team documented in 65% of 46 patient notes that “respiratory improvement was due to the use of glucocorticoids” (3). Among 140 cases reported nationally to CDC that received corticosteroids, 82% of patients improved (Table). However, the natural progression of this injury is not known, and it is possible that patients might recover without corticosteroids or by avoiding use of e-cigarette, or vaping, products. In some circumstances, it would be advisable to withhold corticosteroids while evaluating patients for infectious etiologies, such as fungal pneumonia, that might worsen with corticosteroid treatment. Nevertheless, because the diagnosis remains one of exclusion, aggressive empiric therapy with corticosteroids, antimicrobial, and antiviral therapy might be warranted for patients with severe illness. A range of corticosteroid doses, durations, and taper plans might be considered on a case-by-case basis. Whenever possible, decisions on the use of corticosteroids and dosing regimen should be made in consultation with a pulmonologist.

Early initiation of antimicrobial treatment for community-acquired pneumonia in accordance with established guidelines** should be strongly considered given the overlapping of signs and symptoms in these conditions. During influenza season, health care providers should consider influenza in all patients with suspected EVALI. Antivirals should be considered in accordance with established guidelines.^†† Decisions on initiation or discontinuation of treatment should be based on specific clinical features and, when appropriate, in consultation with specialists.

Follow-up from hospital admission. Patients discharged from the hospital after inpatient treatment for EVALI should have a follow-up visit no later than 1–2 weeks after discharge that includes pulse-oximetry, and clinicians should consider repeating the CXR. Additional follow-up testing 1–2 months after discharge that might include spirometry, diffusion capacity testing, and CXR should be considered. Long-term effects and the risk of recurrence of EVALI are not known. Whereas many patients’ symptoms resolved, clinicians report that some patients have relapsed during corticosteroid tapers after hospitalization, underscoring the need for close follow-up (personal communication, Lung Injury Response Clinical Working Group, October 2, 2019). Some patients have had persistent hypoxemia (O₂ saturation <95%), requiring home oxygen at discharge and might need ongoing pulmonary follow-up. Patients treated with high-dose corticosteroids might require care from an endocrinologist to monitor adrenal function.

It is unknown if patients with a history of EVALI are at higher risk for severe complications of influenza or other respiratory viral infections if they are infected simultaneously or after recovering from lung injury. Health care providers should emphasize the importance of annual vaccination against influenza for all persons >6 months of age, including patients with a history of EVALI. In addition, administration of pneumococcal vaccine should be considered according to current guidelines.^§§

Addressing exposures. Advising patients to discontinue use of e-cigarette, or vaping, products should be an integral part of the care approach during an inpatient admission and should be re-emphasized during outpatient follow-up. Cessation of e-cigarette, or vaping, products might speed recovery from this injury; resuming use of e-cigarette, or vaping, products has the potential to cause recurrence of symptoms or lung injury. Evidence-based tobacco product cessation strategies include behavioral counseling and FDA-approved cessation medications.^¶¶ For patients who have addiction to THC-containing or nicotine-containing products, cognitive-behavioral therapy, contingency management, motivational enhancement therapy, and multidimensional family therapy have been shown to help, and consultation with addiction medicine services should be considered (8–10).

Special considerations for groups at high risk. Patients with certain characteristics or comorbidities, including older age, history of cardiac or lung disease, or pregnancy, might be at higher risk for more severe outcomes. Among reported cases (Table), patients aged >50 years experienced the highest percentage of endotracheal intubation and mechanical ventilation (54%) and the longest mean inpatient stays (15 days). The mean first recorded O₂ saturations among those who did and did not require intubation were 87% and 92%, respectively (data not shown). Among those with and without past cardiac disease, 31% and 21%, respectively, required intubation (Table). Special consideration might need to be given to patients aged >50 years, because these patients might require longer duration of hospitalization and have a higher risk of intubation (Figure). Rapid identification of exposure, a high index of suspicion of EVALI, initiation of corticosteroids, and specialist consultations might be lifesaving in this patient population.

Additional data might identify other groups at high risk, provide important information about disparities in outcomes, and help guide clinical care. Certain patients, such as adolescents and young adults, might benefit from specialized services, such as addiction treatment services and providers who have experience with counseling and behavioral health follow-up.

Clinical Care and Public Health Recommendations

Reporting cases to state, local, territorial, or tribal health departments is critical for accurate surveillance of EVALI. Reporting cases and obtaining and sending products, devices, and clinical and pathologic specimens for testing, can help health departments and CDC determine the cause or causes of these lung injuries.*** CDC is developing International Classification of Diseases, Tenth Edition, Clinical Modification coding guidance for health care encounters related to EVALI. Updates, when available, can be found at https://www.cdc.gov/lunginjury (Box 3).

Public health recommendations. At this time, FDA and CDC have not identified the cause or causes of the lung injuries among EVALI cases, and the only commonality among all cases is that patients report the use of e-cigarette, or vaping, products. This outbreak might have more than one cause, and many different substances and product sources are still under investigation. To date, national and state data suggest that products containing THC, particularly those obtained off the street or from other informal sources (e.g., friends, family members, or illicit dealers), are linked to most of the cases and play a major role in the outbreak (11,12). Therefore, CDC recommends that persons should not use e-cigarette, or vaping, products that contain THC. Persons should not buy any type of e-cigarette, or vaping, products, particularly those containing THC, off the street. Persons should not modify or add any substances to e-cigarette, or vaping, products that are not intended by the manufacturer, including products purchased through retail establishments.

Given that the exclusive use of nicotine-containing products has been reported by a small percentage of persons with EVALI, and that many persons with EVALI report combined use of THC- and nicotine-containing products, the possibility that nicotine-containing products play a role in this outbreak cannot be excluded. Therefore, at present, CDC continues to recommend that persons consider refraining from using e-cigarette, or vaping, products that contain nicotine. If adults are using e-cigarette, or vaping, products to quit cigarette smoking, they should not return to smoking cigarettes; they should use evidence-based treatments, including health care provider counseling and FDA-approved medications.^††† If persons continue to use these products, they should carefully monitor themselves for symptoms and see a health care provider immediately if symptoms develop. Irrespective of the ongoing investigation, e-cigarette, or vaping, products should never be used by youths, young adults, or women who are pregnant. There is no safe tobacco product, and the use of any tobacco products, including e-cigarettes, carries a risk. Therefore, persons who do not currently use tobacco products should not start using e-cigarette, or vaping, products.

This investigation is ongoing. CDC will continue to work in collaboration with FDA and state and local partners to investigate cases and to update guidance, as appropriate, as new data emerges from this complex outbreak.

Lung Injury Response Clinical Working Group

Anne Griffiths, MD, , Pediatric Pulmonary Medicine, Children’s Minnesota; Annette Esper, MD, Emory University; Carolyn S. Calfee, MD, Pulmonary and Critical Care Medicine, University of California, San Francisco; Don Hayes, Jr., MD, Nationwide Children’s Hospital and The Ohio State University; Devika R. Rao, MD, Department of Pediatrics, Division of Respiratory Medicine UT Southwestern Medical Center; Dixie Harris, MD, Intermountain Healthcare; Lincoln S. Smith, MD, University of Washington and Seattle Children’s Hospital; Scott Aberegg, MD; Sean J. Callahan, MD, University of Utah

Lung Injury Response Epidemiology/Surveillance Group

Rashid Njai, Office of the Director, Deputy Director for Non-Infectious Diseases, CDC; Jennifer Adjemian; Macarena Garcia; Kathleen Hartnett; Kristen Marshall; Aaron Kite Powell, Center for Surveillance, Epidemiology, and Laboratory Services, CDC; Adebola Adebayo; Minal Amin; Michelle Banks; Jordan Cates, National Center for Immunization and Respiratory Diseases, CDC; Maeh Al-Shawaf; Lauren Boyle-Estheimer; Peter Briss; Gyan Chandra; Karen Chang; Jennifer Chevinsky; Katelyn Chiang; Pyone Cho; Carla Lucia DeSisto; Lindsey Duca; Sumera Jiva; Charlotte Kaboré; John Kenemer; Akaki Lekiachvili; Maureen Miller; Yousra Mohamoud; Cria Perrine; Mays Shamout; Lauren Zapata, National Center for Chronic Disease Prevention and Health Promotion, CDC; Francis Annor; Vaughn Barry; Amy Board; Mary E. Evans; Allison Gately; Brooke Hoots; Cassandra Pickens; Tia Rogers; Alana Vivolo-Kantor, National Center for Injury Prevention and Control, CDC; Alissa Cyrus, Office of Minority Health and Health Equity, CDC; Tegan Boehmer; Emily Glidden; Arianna Hanchey; Angela Werner; Shideh Ebrahim Zadeh, National Center for Environmental Health, CDC; Donna Pickett, National Center for Health Statistics, CDC; Victoria Fields; Michelle Hughes; Varsha Neelam; Kevin Chatham-Stephens, National Center on Birth Defects and Developmental Disabilities, CDC; Kevin O’Laughlin; Mary Pomeroy, National Center for Emerging and Zoonotic Infectious Diseases, CDC; Sukhshant K. Atti, Agency for Toxic Substances and Disease Registry, CDC and Emory University School of Medicine; Jennifer Freed; Jona Johnson; Eva McLanahan, Agency for Toxic Substances and Disease Registry; Kate Varela, National Institute for Occupational Safety and Health; Jennifer Layden, Illinois Department of Public Health; Jonathan Meiman, Wisconsin Department of Health Services; Nicole M. Roth, Eagle Medical Services; Diane Browning, Northrop Grumman; Augustina Delaney; Samantha Olson, G2S Corporation; Dessica F. Hodges, Student Worksite Program volunteer; Raschelle Smalley, Student Worksite Experience Program volunteer; Council of State and Territorial Epidemiologists Vaping-Associated Pulmonary Injury (VAPI) Epidemiology Task Force

Discussion

Although lung cancer incidence rates declined among males and females living in nonmetropolitan and metropolitan areas during 2007–2016, the smallest decrease occurred among females living in nonmetropolitan counties, who also experienced high incidence in some age groups. During this 10-year period, the highest overall lung cancer incidence rates were observed among males in nonmetropolitan counties. National Health Interview Survey 2017 data indicate that, compared with adults living in metropolitan areas, those living in nonmetropolitan areas reported a higher prevalence of current cigarette smoking (23% versus 13%) and a lower prevalence of quit attempts (50% versus 56%) and successful cessation (5% versus 9%) (4).

Lung cancer prevention and control is a comprehensive approach and includes strategies such as screening for tobacco dependence, promoting tobacco cessation, implementing comprehensive smoke-free laws, testing all homes for radon and using proven methods to lower high radon levels, and reducing exposure to lung carcinogens such as asbestos (1). The U.S. Preventive Services Task Force recommends that clinicians screen all adults for tobacco use at each office visit and refer or provide behavioral and pharmacotherapy smoking cessation interventions as indicated.^†† Lung cancer screening is recommended for adults at high risk for developing lung cancer because of their age and cigarette smoking history. Screening efforts can identify lung cancer in its early stages and provide an important opportunity to promote tobacco smoking cessation. However, access to these preventive services might be more limited in nonmetropolitan areas, where a higher percentage of residents aged <65 years report being uninsured compared with those in metropolitan areas (4).

CDC’s National Comprehensive Cancer Control Program^§§ funds state, tribal, local, and territorial comprehensive cancer control programs that pool resources to lower the number of persons affected by types of cancer with the highest burden in a given community, including lung cancer. These programs advance their priorities through evidence-based interventions that include primary prevention and early detection. Examples of lung cancer prevention strategies are promoting tobacco-free living for all persons (5) and reducing exposure to indoor radon (6). An important step in implementing interventions for the early detection of lung cancer is assessing a community’s capacity to meet screening needs. For example, Maine’s Comprehensive Cancer Control Program identified lung cancer screening facilities in nonmetropolitan and metropolitan areas and is working to address screening barriers (7). Another approach is using patient navigators and community health workers to address health care barriers (e.g., financial hardships, lack of or inadequate health insurance coverage, and lack of transportation) (8). CDC, along with the Appalachian Regional Commission, has funded research to more fully understand how patient navigation can help cancer survivors in nonmetropolitan areas have better access to cancer care,^¶¶ which can then inform the development of culturally relevant training for patient navigators.

Although cigarette smoking is the primary cause of lung cancer, other risk factors, which may differ by geographic region, include use of other smoking tobacco products and exposure to secondhand smoke, indoor radon, and asbestos (1). In some states, rural areas may be less likely to have strong smoke-free laws or barrier-free access to tobacco cessation programs.***

Approximately 10%–15% of lung cancers are estimated to occur among persons who have never smoked cigarettes (9). Regardless of smoking status, lung cancer survivors might experience blame, stigma, and other negative reactions associated with their lung cancer diagnosis (10). A qualitative analysis found that lung cancer survivors believed the stigma translated into a lack of public empathy, and they desired increased public support (10). Public health programs such as CDC’s National Comprehensive Cancer Control Program are focused on cancer survivorship and can work to reduce stigma by educating the public and implementing programs to address the needs of lung cancer survivors.

The findings in this report are subject to at least two limitations. First, delays in cancer reporting might result in an underestimation of incidence. Second, incidence was not determinable by county classification for all states; therefore, these results might not apply to states excluded from the analyses.

During 2007–2016, lung cancer incidence rates declined overall in nonmetropolitan and metropolitan counties; however, rates decreased more in metropolitan than in nonmetropolitan counties, more among males than among females, and more among persons aged 35–54 years than among those aged ≥55 years. As a result, differences in lung cancer incidence rates between males and females narrowed with decreasing age, but disparities by rural-urban status persisted. A comprehensive approach to lung cancer prevention and control includes such population-based strategies as screening for tobacco dependence, promoting tobacco cessation, implementing comprehensive smoke-free laws, testing all homes for radon and using proven methods to lower high radon levels, and reducing exposure to lung carcinogens such as asbestos (1). Increasing the implementation of proven population-based lung cancer prevention and control strategies, particularly among persons living in nonmetropolitan areas, might help to reduce disparities in the decline of lung cancer incidence.

Immunization Activities

WHO and the United Nations Children’s Fund (UNICEF) use data from administrative records and vaccination coverage surveys reported annually to estimate MCV1 and second dose (MCV2) coverage through routine immunization services.^¶ During 2000–2018, estimated MCV1 coverage increased globally from 72% to 86% (Table), although coverage has remained at 84%–86% since 2010, with considerable regional variation. Since 2016, MCV1 coverage has remained relatively constant in the African Region (AFR) (74%–75%), the Eastern Mediterranean Region (EMR) (82%–83%), and the South-East Asia Region (SEAR) (88%–89%); and it has remained constant since 2008 in the European Region (EUR) (93%–95%) and in the Western Pacific Region (WPR) (95%–97%). Estimated MCV1 coverage in the Region of the Americas (AMR) decreased from 92% in 2016 to 88% in 2017 and increased to 90% in 2018.

Globally, 118 (61%) countries achieved ≥90% MCV1 coverage in 2018, an increase from 86 (45%) countries in 2000, but a decrease from 126 (65%) countries during 2012–2013. In 2018, MCV1 coverage was ≥95% nationally in 78 (40%) countries and ≥80% in all districts in 57 (29%) countries.** In 2018, 19.2 million infants worldwide did not receive MCV1 through routine immunization services. The six countries with the most unvaccinated infants were Nigeria (2.4 million), India (2.3 million), Pakistan (1.4 million), Ethiopia (1.3 million), Indonesia (1.2 million), and the Philippines (0.7 million).

Estimated MCV2 coverage increased globally from 18% in 2000 to 69% in 2018, largely because of an increase in the number of countries providing MCV2 from 98 (51%) in 2000 to 171 (88%) in 2018 (Table). Four countries (Bolivia, the Dominican Republic, Honduras, and the Solomon Islands) introduced MCV2 in 2018.

In 2018, approximately 346 million persons received measles vaccination during 45 supplementary immunization activities (SIAs)^†† in 37 countries; India’s 2018 SIA accounted for 47% of all persons vaccinated in SIAs worldwide. An additional 13 million persons were vaccinated during measles outbreak response activities.

Reported Measles Incidence

In 2018, all 194 WHO member countries conducted measles surveillance, and 191 (98%) had access to standardized quality-controlled laboratory testing through the WHO Global Measles and Rubella Laboratory Network. However, surveillance remains weak in many countries, and only 84 (55%) of 152 countries that reported surveillance indicators achieved the sensitivity indicator target of ≥2 discarded measles and rubella^§§ cases per 100,000 population.

Countries report the number of incident measles cases^¶¶ to WHO and UNICEF annually using the Joint Reporting Form.*** During 2000–2018, the number of reported cases decreased 59%, from 853,479 in 2000 to 353,236 in 2018, and measles incidence decreased 66%, from 145 to 49 cases per million population (Table). However, compared with the reported number of cases (132,413) and incidence (19 cases per million) in 2016, both cases and incidence increased in 2018, the highest levels since 2011 (Figure 1). Compared with 2016, the number of measles cases increased 167% globally, including increases of 246% in AFR, 16,732% in AMR, 931% in EMR, 1,791% in EUR, and 26% in SEAR.^††† In WPR, the number of measles cases decreased 49%, primarily because of decreased cases in China. In 2018, five (3%) of 179 reporting countries (Democratic Republic of the Congo, Liberia, Madagascar, Somalia, and Ukraine) had measles incidences >600 per million and accounted for 45% (157,239 cases) of all reported cases worldwide. The percentage of reporting countries with annual measles incidence of <5 cases per million population increased from 38% (64 of 169) in 2000 to 70% (125 of 178) in 2016, then decreased to 54% (96 of 179) in 2018 (Table) (Figure 1).

Genotypes of viruses isolated from measles cases were reported by 95 (73%) of 131 countries reporting at least one measles case in 2018. Among the 24 recognized measles virus genotypes, 11 were detected during 2005–2008, eight during 2009–2014, six in 2016, five in 2017, and four in 2018 (4). In 2018, among 7,155 reported virus sequences, 3,011 (42%) were genotype B3; 20 (0.3%) were D4; 3,774 (53%) were D8; and 350 (5%) were H1.

Measles Case and Mortality Estimates

A previously described model for estimating measles cases and deaths was updated with new measles vaccination coverage data, case data, and United Nations population estimates for all countries during 2000–2018, enabling derivation of a new series of disease and mortality estimates (5). For countries with anomalous estimates in previous iterations, the model was modified slightly to generate mortality estimates consistent with observed case data. Based on the updated data, the estimated number of measles cases decreased 65%, from 28,219,100 (95% confidence interval [CI] = 20,141,900–65,455,000) in 2000 to 9,769,400 (95% CI = 6,446,900–40,538,500) in 2018. During this period, estimated measles deaths decreased 73%, from 535,600 (95% CI = 363,400–901,700) to 142,300 (95% CI = 93,600–387,900) (Table) (Figure 2). During 2000–2018, compared with no measles vaccination, measles vaccination prevented an estimated 23.2 million deaths globally.

Regional Verification of Measles Elimination

By the end of 2018, 82 (42%) countries had been verified as having eliminated measles. Austria, Bahrain, North Korea, Oman, Singapore, Switzerland, and Timor-Leste were verified as having achieved elimination during 2018. No AFR country had yet been verified as having eliminated measles. In the AMR, a region that had achieved verification of measles elimination in 2016, endemic measles transmission was reestablished in Venezuela in 2018 and in Brazil in 2019. In EUR, endemic measles transmission was reestablished during 2018 in Albania, Czechia, Greece, and the United Kingdom.

Discussion

During 2000–2018, increased coverage with MCV1 and MCV2, widespread SIAs, and other elimination efforts contributed to a 66% decrease in reported measles incidence, a 73% reduction in estimated measles mortality, and a reduction in the number of circulating measles virus genotypes worldwide. Despite this progress, the 2015 global milestones were not met: MCV1 coverage has stagnated for nearly a decade, MCV2 coverage is only 69%, and suboptimal surveillance limits data-driven actions. Reported measles incidence has increased in five regions since 2016 and estimated global measles mortality has increased since 2017. Increased measles cases and outbreaks occurred mostly among unvaccinated persons, including school-aged children and young adults.

The causes of the measles resurgence during 2017–2018 are multifactorial and vary by country. Large sustained outbreaks in a few countries with weak immunization systems accounted for most reported measles cases during this time. In addition, unidentified or unaddressed immunity gaps in older children and adults, because of historically weak routine immunization programs and inadequate SIA coverage, led to sustained transmission in some countries that previously had low incidence or had eliminated measles (6). As well, international travel by infected persons, including both unimmunized foreign visitors and unimmunized residents traveling abroad and returning home, facilitated international spread of measles. For example, in 2018, Israel experienced nearly 100 measles importations from multiple countries including Philippines, Ukraine, and the United Kingdom; and importations from Israel and Ukraine led to outbreaks in the United States (7). Sustaining elimination in the face of frequent importations and gaps in vaccination coverage presents challenges. For example, after having experienced >100 importations in 2018 as a consequence of inadequate vaccination coverage, endemic measles virus transmission has been reestablished in the United Kingdom. Countries such as Cambodia, which, through sustained efforts, identified and closed immunity gaps to achieve elimination, but which border countries with ongoing endemic transmission, must remain vigilant to identify and stop measles outbreaks rapidly. Before international travel, travelers from all countries should ensure they have been appropriately vaccinated against measles. Progress toward measles elimination will regress without a unified effort by all communities and countries.

Evaluations of routine immunization programs to identify barriers to vaccination indicate that children miss MCV1 and MCV2 doses for many reasons, including families’ limited awareness of the need for vaccination, limited access to or financial barriers to receiving vaccination; vaccine stock-outs; political instability; and vaccine hesitancy and misinformation. WHO’s Global Routine Immunization Strategies and Practices and The Guide to Tailoring Immunization Programmes provides guidance on identifying demand and supply barriers to routine vaccination and strengthening immunization programs (8,9). Outbreaks should serve as opportunities to investigate underlying causes of undervaccination and to design specific routine immunization strengthening activities to prevent future outbreaks. In addition, population immunity gaps should be identified through triangulation of data, including surveillance and vaccination coverage data, and should be targeted by vaccination activities.

The findings in this report are subject to at least two limitations. First, large differences between estimated and reported incidence indicate overall low surveillance sensitivity, making comparisons between regions difficult to interpret. Second, the measles mortality model estimates might be affected by biases in model inputs, including vaccination coverage and surveillance data.

The trends of increasing measles incidence and mortality are reversible; however, further progress toward achieving elimination goals will require 1) resource commitments to strengthen routine immunization systems, close historical immunity gaps, and improve surveillance to rapidly detect and respond to cases, and 2) a new perspective to use measles as a stimulus and guide to improving immunization programs. To achieve measles elimination, all communities and countries need coordinated efforts aiming to reach ≥95% coverage with 2 doses of measles vaccine.

As the period covered by the Global Vaccine Action Plan 2012–2020 approaches its end, a new vision and strategy for accelerated progress on immunization for 2021–2030 is being developed by countries and stakeholders (10). Pillars of this evolving strategy include commitment and demand, research and innovation, life course and integration, and supply and sustainability; all of these are vital to achieving and maintaining measles elimination. This new agenda should be used to secure the necessary resource commitments to improve coverage and equity substantially and, in so doing, further progress toward achieving the measles elimination goals.

Discussion

This report used data from 32 states to provide updated population-level suicide rates for major occupational groups and new information on suicide rates for major industry groups and detailed occupational groups. Estimates for most major occupational groups are similar, although not directly comparable, to previous estimates that were based on 2015 NVDRS data from 17 states (4). Recent NVDRS expansion to 50 states might facilitate direct comparisons over time by industry and occupation nationwide. These findings highlight opportunities for targeted prevention strategies and further investigation of work-related factors that might increase risk of suicide. Previous research indicates suicide risk is associated with low-skilled work (5), lower education (6), lower absolute and relative socioeconomic status (7), work-related access to lethal means (8), and job stress, including poor supervisory and colleague support, low job control, and job insecurity (9). Industry, labor, and professional associations, as well as employers, and state and local health departments can use this information to focus attention and resources on suicide prevention. Future research might examine these and other risk factors among the industries and occupations identified in this report as having high suicide rates.

This report estimated suicide rates comprehensively for industry and occupational groups meeting sample size criteria and identified groups with rates higher than the study’s population rate. Although relative comparisons of suicide rates in this manner are useful for prevention purposes, these results should not overshadow the essential fact that the suicide rate in the U.S. working-age population overall has increased by 40% in less than 2 decades. Therefore, all industry sectors and occupational groups can contribute to reducing suicide incidence.

The findings in this report are subject to at least five limitations. First, this study did not address confounding factors that might account for different suicide rates among and within industry or occupational groups. Second, it did not address suicide among unemployed decedents, military or unpaid workers, or those aged >64 years (9). Third, the numerator and denominator data were not a direct match for calculating rates; death certificates reflect decedents’ usual industry and occupation, and available population size data refer to the number of persons by current job. Fourth, the results are based on data from 32 states and are therefore not nationally representative. Finally, three states contributing to the 2016 NVDRS did not collect data on all violent deaths. Other limitations of NVDRS analysis using death certificate industry and occupation data have been described previously (4).

All industries and occupations can benefit from a comprehensive approach to suicide prevention. CDC’s Preventing Suicide: A Technical Package of Policy, Programs, and Practices (1) provides strategies with the best available evidence to prevent suicide and can serve as a resource for communities and employers. Workplace-specific strategies include 1) promoting help-seeking; 2) integrating workplace safety and health and wellness programs to advance the overall well-being of workers; 3) referring workers to financial and other helping services; 4) facilitating time off and benefits to cover supportive services; 5) training personnel to detect and appropriately respond to suicide risk; 6) creating opportunities for employee social connectedness; 7) reducing access to lethal means among persons at risk; and 8) creating a crisis response plan sensitive to the needs of coworkers, friends, family, and others who might themselves be at risk (1,10). Other community-based strategies include strengthening economic supports, strengthening access and delivery of care, teaching coping and problem-solving skills, and responsibly reporting suicide (e.g., not providing details) (1). Further workplace prevention resources are available at https://workplacesuicideprevention.com/external icon and https://theactionalliance.org/communities/workplaceexternal icon and help is available at 1-800-273-TALK (8255).

Discussion

From 2002 to 2015, the annual incidence of both type 1 and type 2 diabetes increased at constant rates among persons aged <20 years in selected counties and Indian reservations in the United States. Rates of increase in incidence were higher for type 2 diabetes (4.8% per year) than for type 1 (1.9%). Since 2012, the rate of increase in type 2 diabetes has not changed, and has also remained constant for type 1 diabetes, except among Asians and Pacific Islanders. These findings provide indicators of the number of new cases of type 1 and type 2 diabetes among U.S. youths and identify groups with increased incidences of both type 1 and type 2 diabetes. Diabetes is a chronic disease that requires lifelong treatment and management. Better understanding of the number of new cases of diabetes among youths helps in planning for health care needs and resources.

The findings in this report are subject to at least two limitations. First, a small number of cases was ascertained across years, in subgroups by diabetes type, and especially across racial/ethnic groups, possibly leading to less precision in the annual rates. Second, these findings might not be generalizable to other populations because SEARCH was not designed to be nationally representative; it includes populations from five U.S. sites. A major strength of this study is that data come from a complete, population-based registry covering approximately a decade, including both type 1 and type 2 diabetes in persons aged <20 years across multiple racial/ethnic groups.

The incidence of type 1 diabetes continues to increase in U.S. youths, with steeper increases observed in black and Hispanic youths. Since 2011, the incidence of type 1 diabetes has also significantly increased among Asians and Pacific Islanders. Reasons for this recent increase are unknown. In parallel with increased obesity prevalence in U.S. youths (9), the incidence of type 2 diabetes among adolescents has increased at a higher rate than that of type 1 diabetes, especially among racial/ethnic minority youths. There are no known prevention interventions for type 1 diabetes; in adults the onset of type 2 diabetes can be prevented or delayed with lifestyle changes programs, such as the National Diabetes Prevention Program (https://www.cdc.gov/diabetes/prevention/index.html). Although the effectiveness of these programs among youths is unknown, promoting healthy eating and lifestyles provides many health benefits (https://www.cdc.gov/diabetes/prevent-type-2/type-2-kids.html). One program targeting the prevention of type 2 diabetes in American Indian youths is the Native Diabetes Wellness Program (https://www.cdc.gov/diabetes/ndwp/index.html). This collaboration between CDC and other partners provides resources to promote healthy eating and physical activity in American Indian and Alaska Native youths. To assess public health needs and prevention efforts for type 1 and type 2 diabetes among youths, it is important to enhance and continue surveillance efforts to monitor incidence in this population.

Discussion

The annual percentage of U.S. adults who had an opioid prescription filled decreased by 31% during 2008–2018. This decline might be attributed to implementation of several opioid prescribing guidelines, enhanced prescription drug monitoring programs, and other quality improvement initiatives (5). Percentages of persons with at least one opioid prescription filled were the highest among adults aged ≥65 years. These persons might have higher frequency, longer duration, or greater intensity of chronic pain, which might contribute to higher prescription fill rates (6). Some researchers have described a stable trend from 2007 to 2016 among commercially insured and Medicare Advantage beneficiaries in opioid prescription fill rates (7), whereas the findings in this study indicated a decline. Although the reasons for this discrepancy are not clear, the patient population of the current study is different from that of the study of Medicare Advantage beneficiaries and includes all classes of payers.

Higher opioid prescription fill rates among older adults is particularly worrisome because they are more likely to have an adverse event, even death, from taking an opioid medication (8). Older adults might also be less aware of the number of doses taken, have problems with balance or gait, experience a drug interaction with another medication used to treat a chronic condition, or have reduced opioid excretion resulting from age-related changes in liver and renal function (8). The percentage decline of opioid prescriptions filled by patients aged ≥65 years was the smallest of any age group, only 16% over 11 years.

Compared with men, women in all age groups had higher odds of having an opioid prescription filled. This might be partly explained by physical differences in how women process pain (9), higher likelihood of having a diagnosis of a mental health disorder, greater use of health care, or higher prevalence of certain chronic health conditions for which opioids are commonly prescribed (e.g., arthritis and fibromyalgia) compared with that of men (10). In addition, younger women might receive opioids during their childbearing years for painful reproductive disorders (e.g., dysmenorrhea or endometriosis) (10). However, the extent to which these conditions are driving these differences is unknown.

The findings in this report are subject to at least five limitations. First, only those prescriptions filled by retail pharmacies were considered; data were not available from other sources. Second, analyzing dosage, duration, or type of formulation was beyond the scope of this study. Third, information was not available on prescriptions that were written but not filled, whether any or all of the prescription was taken by the patient, and whether the prescription was new versus a refill. Fourth, this report did not assess drug diversion, which could result in prescription opioids being obtained through illicit sources (2). Finally, the efficacy of the prescription relative to the medical condition and severity could not be determined.

Those age groups among both sexes with the highest prescription fill rates warrant special attention to understand whether and how prescribing might be reduced. Optimal prescribing for these groups might differ from that of other groups because best practices for treating pain vary by medical condition and pharmacokinetics, and the prevalence of medical conditions varies by age group and sex (4). Additional research could help better identify patient needs and effective multimodal approaches to pain management, particularly among women and persons aged ≥65 years, the groups with higher opioid prescription fill rates. This in turn could help to establish the extent to which the observed differences in fill rates are relevant and lead to optimal prescribing for all subpopulations.