Untangling the effects of multiple exposures with a common reference group in an epidemiologic study: A practical revisit.

When assessing multiple exposures in epidemiologic studies, epidemiologists often use multivariable regression models with main effects only to control for confounding. This method can mask the true effects of individual exposures, potentially leading to wrong conclusions. We revisited a simple, practical, and often overlooked approach to untangle effects of the exposures of interest, in which the combinations of all levels of the exposures of interest are recoded into a single, multicategory variable. One category, usually the absence of all exposures of interest, is selected as the common reference group (CRG). All other categories representing individual and joint exposures are then compared to the CRG using indicator variables in a regression model or in a 2×2 contingency table analysis. Using real data examples, we showed that using the CRG analysis results in estimates of individual and joint effects that are mutually comparable and free of each other's confounding effects, yielding a clear, accurate, intuitive, and simple summarization of epidemiologic study findings involving multiple exposures of interest.

An Epidemic of Pediatric HIV From Reuse of Infusion Equipment in Pakistan.

BACKGROUND: From April to June 2019, a total of 909 new HIV infections were identified in Larkana, Pakistan; 86% was children younger than 15 years. To identify the possible transmission links in this outbreak, a case-control study was conducted in June 2019. METHODS: For cases, we selected a systematic random sample of 100 HIV-positive children from the screening list. We chose 2 age-matched and sex-matched controls from the neighborhood of each HIV-positive case. All selected children were tested using the World Health Organization-approved rapid diagnosis test algorithm. We interviewed the parents of each selected child about previous exposures to parenteral treatment and compared exposures of case and control children using conditional logistic regression. RESULTS: The ages of the selected children ranged from 1 month to 10 years. More than 90% of both HIV+ and HIV- children had received outpatient health care from MBBS-qualified private physicians. Eighty-three percent of HIV+ children versus 46% of HIV- children had received health care from one private physician [adjusted odds ratio (aOR) = 29, 95% confidence interval (95% CI): 10 to 79]. Intravenous infusions during the last outpatient visit were reported by 29% of case versus 7% of controls (aOR 57, 95% CI: 2.9 to >1000), whereas no case children and 17% of control children had received only intramuscular injections (aOR 0, 95% CI: 0 to 41). Among cases, 94% had been given infusions through a drip set compared with 85% of control children (aOR = 7.7, 95% CI: 2.3 to 26). Infusions had been administered with reused IV drip sets in 70% of cases compared with 8% of controls (aOR = 197, 95% CI: 16 to 2400). DISCUSSION: Private physicians reusing intravenous drip sets to treat outpatients seen in private practice were responsible for this HIV epidemic. Mapping and regulation of private practitioners were suggested.

One Field Epidemiologist per 200,000 Population: Lessons Learned from Implementing a Global Public Health Workforce Target.

The World Health Organization monitoring and evaluation framework for the International Health Regulations (IHR, 2005) describes the targets for the Joint External Evaluation (JEE) indicators. For workforce development, the JEE defines the optimal target for attaining and complying with the IHR (2005) as 1 trained field epidemiologist (or equivalent) per 200,000 population. We explain the derivation and use of the current field epidemiology workforce development target and identify the limitations and lessons learned in applying it to various countries' public health systems. This article also proposes a way forward for improvements and implementation of this workforce development target.

Building Global Epidemiology and Response Capacity with Field Epidemiology Training Programs.

More than ever, competent field epidemiologists are needed worldwide. As known, new, and resurgent communicable diseases increase their global impact, the International Health Regulations and the Global Health Security Agenda call for sufficient field epidemiologic capacity in every country to rapidly detect, respond to, and contain public health emergencies, thereby ensuring global health security. To build this capacity, for >35 years the US Centers for Disease Control and Prevention has worked with countries around the globe to develop Field Epidemiology Training Programs (FETPs). FETP trainees conduct surveillance activities and outbreak investigations in service to ministry of health programs to prevent and control infectious diseases of global health importance such as polio, cholera, tuberculosis, HIV/AIDS, malaria, and emerging zoonotic infectious diseases. FETP graduates often rise to positions of leadership to direct such programs. By training competent epidemiologists to manage public health events locally and support public health systems nationally, health security is enhanced globally.

Protective Effect of Hand-Washing and Good Hygienic Habits Against Seasonal Influenza: A Case-Control Study.

Previous observational studies have reported protective effects of hand-washing in reducing upper respiratory infections, little is known about the associations between hand-washing and good hygienic habits and seasonal influenza infection. We conducted a case-control study to test whether the risk of influenza transmission associated with self-reported hand-washing and unhealthy hygienic habits among residents in Fujian Province, southeastern China.Laboratory confirmed seasonal influenza cases were consecutively included in the study as case-patients (n = 100). For each case, we selected 1 control person matched for age and city of residence. Telephone interview was used to collect information on hand-washing and hygienic habits. The associations were analyzed using conditional logistic regression. Compared with the poorest hand-washing score of 0 to 3, odds ratios of influenza infection decreased progressively from 0.26 to 0.029 as hand-washing score increased from 4 to the maximum of 9 (P < 0.001). Compared with the poorest hygienic habit score of 0 to 2, odds ratios of influenza infection decreased from 0.10 to 0.015 with improving score of hygienic habits (P < 0.001). Independent protective factors against influenza infection included good hygienic habits, higher hand-washing score, providing soap or hand cleaner beside the hand-washing basin, and receiving influenza vaccine. Regular hand-washing and good hygienic habits were associated with a reduced risk of influenza infection. These findings support the general recommendation for nonpharmaceutical interventions against influenza.

Risk factors for critical disease and death from hand, foot and mouth disease.

BACKGROUND: There has been a high mortality and morbidity rate of critical and fatal patients from hand, foot and mouth disease (HFMD) in China in recent. Causes for development of critical and fatal disease remain unclear. METHODS: We performed a case-control study to assess the association between use of drugs and development of critical disease and death from HFMD. RESULTS: We found that glucocorticoids treatment was associated with a greater incidence of severe HFMD, whereas andrographolides treatment was associated with a protective effect when they are used for treatment within 48 hours after onset or before being diagnosed as critical. CONCLUSIONS: We recommend that glucocorticoids should not be used for mild HFMD and andrographolides should undergo clinical trials for treatment of enterovirus 71 infections.

Protection by face masks against influenza A(H1N1)pdm09 virus on trans-Pacific passenger aircraft, 2009.

In response to several influenza A(H1N1)pdm09 infections that developed in passengers after they traveled on the same 2 flights from New York, New York, USA, to Hong Kong, China, to Fuzhou, China, we assessed transmission of influenza A(H1N1)pdm09 virus on these flights. We defined a case of infection as onset of fever and respiratory symptoms and detection of virus by PCR in a passenger or crew member of either flight. Illness developed only in passengers who traveled on the New York to Hong Kong flight. We compared exposures of 9 case-passengers with those of 32 asymptomatic control-passengers. None of the 9 case-passengers, compared with 47% (15/32) of control-passengers, wore a face mask for the entire flight (odds ratio 0, 95% CI 0-0.71). The source case-passenger was not identified. Wearing a face mask was a protective factor against influenza infection. We recommend a more comprehensive intervention study to accurately estimate this effect.

A cluster of primary pneumonic plague transmitted in a truck cab in a new enzootic focus in China.

We investigated a cluster of five cases of severe pneumonia from one village in Yunnan Province, China. We searched for severe pneumonia in the village and hospitals. We interviewed patients and family members about exposures. We tested acute and convalescent sera for antigen and antibody of severe acute respiratory syndrome, avian influenza, and plague. The only common exposure of the five patients was riding together in the enclosed cab of a truck for 1.5 hours while taking the first patient to the hospital. Seroconversion to plague F1 antigen confirmed plague in three survivors. Unfamiliarity of clinicians with plague and lack of sputum examination, blood culture, or postmortem examination delayed the diagnosis. No plague cases occurred among family and village contacts and health care workers. High infectivity in this cluster was limited to a crowded, poorly ventilated truck.

Hypoglycemia and death in mice following experimental exposure to an extract of Trogia venenata mushrooms.

BACKGROUND: Clusters of sudden unexplained death (SUD) in Yunnan Province, China, have been linked to eating Trogia venenata mushrooms. We evaluated the toxic effect of this mushroom on mice. METHODS: We prepared extracts of fresh T. venenata and Laccaria vinaceoavellanea mushrooms collected from the environs of a village that had SUD. We randomly allocated mice into treatment groups and administered mushroom extracts at doses ranging from 500 to 3500 mg/kg and water (control) via a gavage needle. We observed mice for mortality for 7 days after a 3500 mg/kg dose and for 24 hours after doses from 500 to 3000 mg/kg. We determined biochemical markers from serum two hours after a 2000 mg/kg dose. RESULTS: Ten mice fed T. venenata extract (3500 mg/kg) died by five hours whereas all control mice (L. vinaceoavellanea extract and water) survived the seven-day observation period. All mice died by five hours after exposure to single doses of T. venenata extract ranging from 1500 to 3000 mg/kg, while the four mice exposed to a 500 mg/kg dose all survived. Mice fed 2000 mg/kg of T. venenata extract developed profound hypoglycemia (median= 0.66 mmol/L) two hours after exposure. DISCUSSION: Hypoglycemia and death within hours of exposure, a pattern unique among mushroom toxicity, characterize T. venenata poisoning.

Clusters of sudden unexplained death associated with the mushroom, Trogia venenata, in rural Yunnan Province, China.

INTRODUCTION: Since the late 1970's, time-space clusters of sudden unexplained death (SUD) in northwest Yunnan, China have alarmed the public and health authorities. From 2006-2009, we initiated enhanced surveillance for SUD to identify a cause, and we warned villagers to avoid eating unfamiliar mushrooms. METHODS: We established surveillance for SUD, defined as follows: sudden onset of serious, unexplained physical impairment followed by death in <24 hours. A mild case was onset of any illness in a member of the family or close socially related group of a SUD victim within 1 week of a SUD. We interviewed witnesses of SUD and mild case-persons to identify exposures to potentially toxic substances. We tested blood from mild cases, villagers, and for standard biochemical, enzyme, and electrolyte markers of disease. RESULTS: We identified 33 SUD, a 73% decline from 2002-2005, distributed among 21 villages of 11 counties. We found a previously undescribed mushroom, Trogia venenata, was eaten by 5 of 7 families with SUD clusters compared to 0 of 31 other control-families from the same villages. In T. venenata-exposed persons SUD was characterized by sudden loss of consciousness during normal activities. This mushroom grew nearby 75% of 61 villages that had time-space SUD clusters from 1975 to 2009 compared to 17% of 18 villages with only single SUD (p<0.001, Fisher's exact test). DISCUSSION: Epidemiologic data has implicated T. venenata as a probable cause of clusters of SUD in northwestern Yunnan Province. Warnings to villagers about eating this mushroom should continue.

Early use of glucocorticoids was a risk factor for critical disease and death from pH1N1 infection.

BACKGROUND: Glucocorticoids increase the risk of developing critical disease from viral infections. However, primary care practitioners in China use them as antipyretics, potentially exposing hundreds of millions to this risk. METHODS: We enrolled all patients with confirmed pandemic influenza A (pH1N1) virus infection aged ≥3 years with available medical records at 4 Shenyang City hospitals from 20 October to 30 November 2009. A critical patient was any confirmed, hospitalized pH1N1 patient who developed ≥1 of the following: death, respiratory failure, septic shock, failure or insufficiency of ≥2 nonpulmonary organs, mechanical ventilation, or ICU admission. In a retrospective cohort study, we evaluated the risk of developing critical illness in relation to early (≤72 hours of influenza-like illness [ILI] onset) glucocorticoids treatment. RESULTS: Of the 83 hospitalized case-patients, 46% developed critical illness, 17% died, and 37% recovered and were discharged. Critically ill and other patients did not differ by underlying conditions and severity, median temperature at first clinic visit, and other measured risk factors. Of 17 patients who received early glucocorticoid treatment, 71% subsequently developed critical disease compared with 39% of 66 patients who received late (>72 hours) or no glucocorticoid treatment (RR(M-H) = 1.8, 95% CI = 1.2-2.8, after adjusting for 2 summary variables; ie, presence of underlying diseases and presence of underlying risk factors). Proportional hazards modeling showed that use of glucocorticoids tripled the hazard of developing critical disease (hazard ratio [HR] = 2.9, 95% CI = 1.3-6.2, after adjusting for the same summary variables). CONCLUSIONS: Early use of parenteral glucocorticoids therapy for fever reduction and pneumonia prevention increases the risk for critical disease or death from pH1N1 infection. We recommend that guidelines on glucocorticoid use be established and enforced.

A community outbreak of rotavirus diarrhea associated with exposures in a hospital outpatient department in South China.

INTRODUCTION: Between October 1 and December 10, 2006, Bao'an, China had a 5-fold increase in acute rotavirus gastroenteritis (ARGE) cases compared with the same time in 2005. To identify the risk factors for ARGE during this outbreak, we conducted a case-control study among children ≤ 24 months old from the most heavily affected area. METHODS: We defined an ARGE case as diarrhea, with group A rotavirus antigen detected from fecal samples by the enzyme-linked immunosorbent assay test. In the case-control study, 86 confirmed ARGE cases were identified. We enrolled 98 healthy control children matched by age. We administered questionnaires about exposures through telephone interviews. RESULTS: Of 8 general exposure types, we identified the following as being associated with ARGE: visiting outpatient services of hospital X, odds ratio (OR) = 7.1 (95% confidence interval [CI], 2.2-26); contact with other children with diarrhea, OR = 2.1 (95% CI, 1.1-3.7); and hand-washing before eating, OR = 0.48 (95% CI, 0.27-0.82). After improvements in the outpatient department of hospital X, ARGE incidence in the community during the ARGE transmission season (October through December) decreased from 4.3/10000 in 2006 to 1.4/10000 in 2009. CONCLUSIONS: Outpatient services in hospital X may have contributed to the transmission of ARGE and improvements in infection control practices in this setting were associated a marked decrease incidence of ARGE in this community.

Paraplegia and paraparesis from intrathecal methotrexate and cytarabine contaminated with trace amounts of vincristine in China during 2007.

PURPOSE: The production and administration of drugs used intrathecally requires special care to prevent contamination with neurotoxic agents. In 2007, we investigated a widespread outbreak of paraplegia and paraparesis among Chinese patients who received intrathecal drugs to identify the presumed contaminant and its source to prevent further cases. PATIENTS AND METHODS: We defined a case as onset from January 1 to October 31, 2007, of bilateral flaccid paraparesis or paraplegia or retention and incontinence of stool or urine, in a patient receiving intrathecal drugs. Using a retrospective cohort approach, we selected 12 hospitals from all hospitals that had reported cases. In these hospitals, we identified all 448 patients (including 107 cases) who received intrathecal chemotherapy or chemoprophylaxis in 2007. We calculated attack rates and Mantel-Haenszel adjusted risk ratios for intrathecal drug type and lot. RESULTS: All 12 hospitals used intrathecal methotrexate or cytarabine produced by one pharmaceutical plant. Only two lots of each drug were associated with cases. Lot-specific attack rates ranged from 42% to 100% (risk ratio, ∞; lower confidence bounds, 1.8 to 7.3). Vincristine production had immediately preceded production of the implicated lots on the same equipment. By using ultra performance liquid chromatography, we detected vincristine (0.28 to 18 µg) in unused vials from implicated lots of methotrexate and cytarabine. CONCLUSION: Trace amounts of vincristine that contaminated intrathecal drugs caused a large outbreak of severe neurologic damage. Vincristine and other neurotoxic drugs should not be produced on any equipment that is also used for producing drugs that are to be administered intrathecally.

Risk factors for hand, foot, and mouth disease and herpangina and the preventive effect of hand-washing.

BACKGROUND: Hygiene and social distancing are recommended control measures for hand, foot, and mouth disease (HFMD) and herpangina. However, empirical data to support this recommendation are limited. METHODS: During an outbreak of HFMD and herpangina due to infection by the human enterovirus 71, we defined a case as a vesicular papular rash on the hands, feet, buttocks, or oral mucosa and onset from April 30 to June 26, 2008. We selected 176 HFMD and herpangina case-children and a stratified random sample of 201 asymptomatic control-children; frequency matched according to residency status. We administered a questionnaire to the parents about their children's exposures and hygienic behaviors. RESULTS: Risk factors for HFMD and herpangina included playing with neighborhood children (odds ratio [OR]: 11 [95% confidence interval (CI): 6.2-17]), visiting an outpatient clinic for another reason ≤ 1 week before onset (OR: 20 [95% CI: 5.0-88]), and community exposures to crowded places (OR: 7.3 [95% CI: 4.1-13]). By using a score summarizing responses to 4 hand-washing questions, we found that 50% of the case-children and 2.5% of control-children had a poor score of 1 to 3, whereas 12% of the case-children and 78% of control-children had a good score of ≥ 7 (OR: 0.00069 [95% CI: 0.0022-0.022]) after we adjusted for residency, age, and community exposures by using logistic regression. CONCLUSIONS: Hand-washing by preschool-aged children and their caregivers had a significant protective effect against community-acquired HFMD and herpangina from the human enterovirus 71 infection.

Glucocorticoid and pyrazolone treatment of acute fever is a risk factor for critical and life-threatening human enterovirus 71 infection during an outbreak in China, 2008.

BACKGROUND: Human enterovirus 71 (HEV71) causes outbreaks of life-threatening diseases throughout the world. The genesis of these severe diseases is unknown. METHODS: During an outbreak of HEV71 infection, we investigated risk factors for critical illness. We developed a modified pediatric index of mortality (mPIM) incorporating heart rate, temperature, white blood cell count, respiratory rate, chest infiltrates, skin color, reflexes, responsiveness, and mobility. We calculated the mPIM for 103 patients (22 deaths) using complete scoring criteria in the medical record. In a case-control study, we compared cases (mPIM > or =10 or death) with controls (mPIM = 0-9) by drugs received within 96 hours after onset of fever, initial temperature, age, and nutritional anthropometry. RESULTS: About 66% (68/103) of the patients with an mPIM score (28 cases and 40 controls) had data on initial exposures. About 50% of the 28 cases and 18% of the 40 controls received an injection to treat fever during the first 96 hours after onset (Odds ratio [OR] = 7.0, 95% confidence interval [CI]: 1.8-28). Injections containing exclusively glucocorticoids (OR = 4.8, 95% CI: 1.2-21) or pyrazolones (OR = 4.1, 95% CI: 0.91-19, P = 0.047) were risk factors for severe HEV71 infection. About 25% of cases and 5% of controls received both drugs parenterally while 7% of cases and 30% of controls received neither (OR = 21, 95% CI: 1.8-305). Conversely, cases and controls had identical average initial temperature, and did not differ significantly by age, sex, nutritional measurements, use of other drugs, or timeliness of medical care received. CONCLUSION: Fever treatment with glucocorticoids and/or pyrazolones is a risk factor for life-threatening HEV71 infection.

Mycobacterium abscessus post-injection abscesses from extrinsic contamination of multiple-dose bottles of normal saline in a rural clinic.

BACKGROUND: We investigated an outbreak of gluteal abscesses following intramuscular (IM) injections given at a clinic in rural China to identify the causative agent, source, and method of exposure. METHODS: We defined a case as an abscess that appeared at the site of an injection given since June 1, 2006. We compared case rates by injection route, medication, and diluents. We reviewed injection practices, and cultured abscesses and environmental sites for mycobacteria. RESULTS: From October through December 2006, 5.8% (n=35) of 604 persons who had received injections at the clinic developed a case. All 35 cases occurred in 184 patients (attack rate=19.0%) who had received IM injections with various drugs that had been mixed with normal saline (NS); risk ratio=infinity; p<0.0001. No cases occurred in the absence of NS exposure. We identified Mycobacterium abscessus from eight abscesses and from the clinic water supply, and observed the inappropriate reuse of a 16-gauge needle left in the rubber septum of 100 ml multiple-dose bottles of NS in the clinic. Fourteen percent (n=527) of the 3887 registered residents of this village had been treated with IM drugs over a three-month period, often for minor illnesses. CONCLUSIONS: This outbreak of M. abscessus occurred from exposure to extrinsically contaminated NS through improper injection practices. Frequent treatment of minor illnesses with IM injections of antibiotics was likely an important contributing factor to the size of this outbreak.

[Retrospective study on 116 unexpected sudden cardiac deaths in Yunnan, China].

OBJECTIVE: To identify the epidemiological and clinical features of unexpected sudden cardiac deaths (SUD) in Yunnan. METHODS: Choosing the old SUD cases from Xiangyun, Heqing, Nanjian and Dayao counties and using the standardized verbal autopsy Form, we interviewed the family members of the cases, witnesses and doctors as well as reviewing their medical files to get relative information. RESULTS: We identified 116 SUDs in 21 villages from 1984 to 2004. The village-specific annually standardized incidence rates were ranged from 0.2/1000 to 8.9/1000 (median = 0.8/1000). 66% and 29% of the SUDs occurred in July and August respectively. The incidence rates of SUD were higher (1.6/1000, chi(2) = 16, P < 0.01) in 10 - 39 year-olds, and higher in females than in males (RR = 1.6, 95% CI: 1.1 - 2.3). Seventy percent of SUD occurred in families having clustering nature and 60% of the additional cases in the family were occurred within 24 hours (median = 20 hours) after the first SUD identified in the family. SUD occurred in 23 families followed the first affected family in a village during the same season. In these 23 families, 61% of the first SUD occurred within 8 days after the first SUD in the first affected family. 68% and 66% of the SUDs did not have any complaints or signs during the last 3 weeks or from 3 weeks to 2 days prior to the onset of the disease. 63% of the SUDs had cardiac symptoms within the last 2 days prior to the onset with major symptoms as dizziness, nausea, faintness, unconsciousness, weakness and palpitation. The median duration from acute onset to death was 2 hours. CONCLUSIONS: The extreme time-space clustering of SUD in families and in villages suggested that the risk factors occurred in specific time and location. Familial clustered SUD cases had common exposure pattern. Sudden onset of acute cardiac symptoms often followed by sudden death. Epidemiological study on new cases was necessary to identify risk factors and to develop hypothesis for causation. In July 2005, we instituted a special SUD surveillance system for all the affected counties together with 10 counties which had no reported cases.

[A field epidemiological study on the risk factors of injury caused by typhoon].

OBJECTIVE: To determine the risk factors involved in the typhoon episodes and to put forward and evaluate the intervention measures. METHODS: We defined a confirmed injury case as: 'a person with fall,scalpel and stab, collision, drowning, injuries and trauma due to flying debris and building collapse, asphyxiation due to entrapment in collapsed buildings by typhoon from 0 am,August 12 to 6 pm, August 14 2004' and a death case as: 'a person with fall, scalpel and stab, collision, drowning, injuries and trauma due to flying debris and building collapse, asphyxiation due to entrapment in collapsed buildings by typhoon from 0 am, August 12 to 12 am, August 18 2004'. We investigated all hospitalized injured cases in ten hospitals and telephoned to those who were not hospitalized and the cases of death. We did case-control study with 1 pair versus 2 cases. 74 cases were selected in ten hospitals. The controls were neighbors of the controls matched by occupation, sex, village, and within 5 years of age without injury in this typhoon. We asked the cases and the controls on their alertness regarding typhoon and what actions taken when typhoon arrived. RESULTS: There were 392 injury cases in all ten hospitals and 50 death cases. The attack rate of injury was 27.3 per 100 000. The fatal rate was 11.3% with the death rate 3.1 per 100 000. We investigated 209 injury cases and 31 death cases. The number of cases who were injured from 1 to 6 hours before typhoon landing accounted for 64.6% (155) of all cases. The peak of epidemic curve was 4 hours before the landing of typhoon. Data on the analysis of 74 cases and 148 controls revealed that 42% (31) of the cases were outside their homes before and during typhoon compared to 15% (22) of the controls (OR = 3.9, 95% CI: 1.9-7.7). Compared with 20% (30) control persons (OR = 17,95% CI: 4.2-68). 28% (21) cases did not receive the alert of typhoon before it arrived compared with 18% (27) control persons (OR = 3.3, 95% CI:1.3-8.6). 53% (39) of the cases did not pay attention to the alert of typhoon before typhoon arrived. CONCLUSION: Staying outdoor, not receiving or did not take seriously about the alert of typhoon seemed to be the risk factors of injury by the typhoon episode, suggesting that the government should increase the emergency preparedness and to raise the awareness on risks associated with typhoon.

Plague from eating raw camel liver.

We investigated a cluster of 5 plague cases; the patients included 4 with severe pharyngitis and submandibular lymphadenitis. These 4 case-patients had eaten raw camel liver. Yersinia pestis was isolated from bone marrow of the camel and from jirds (Meriones libycus) and fleas (Xenopsylla cheopis) captured at the camel corral.