Near real-time surveillance of the SARS-CoV-2 epidemic with incomplete data

When responding to infectious disease outbreaks, rapid and accurate estimation of the epidemic trajectory is critical. However, two common data collection problems affect the reliability of the epidemiological data in real time: missing information on the time of first symptoms, and retrospective revision of historical information, including right censoring. Here, we propose an approach to construct epidemic curves in near real time that addresses these two challenges by 1) imputation of dates of symptom onset for reported cases using a dynamically-estimated "backward" reporting delay conditional distribution, and 2) adjustment for right censoring using the NobBS software package to nowcast cases by date of symptom onset. This process allows us to obtain an approximation of the time-varying reproduction number (Rt) in real time. We apply this approach to characterize the early SARS-CoV-2 outbreak in two Spanish regions between March and April 2020. We evaluate how these real-time estimates compare with more complete epidemiological data that became available later. We explore the impact of the different assumptions on the estimates, and compare our estimates with those obtained from commonly used surveillance approaches. Our framework can help improve accuracy, quantify uncertainty, and evaluate frequently unstated assumptions when recovering the epidemic curves from limited data obtained from public health systems in other locations.

Near real-time surveillance of the SARS-CoV-2 epidemic with incomplete data

Designing public health responses to outbreaks requires close monitoring of population-level health indicators in real-time. Thus, an accurate estimation of the epidemic curve is critical. We propose an approach to reconstruct epidemic curves in near real time. We apply this approach to characterize the early SARS-CoV-2 outbreak in two Spanish regions between March and April 2020. We address two data collection problems that affected the reliability of the available real-time epidemiological data, namely, the frequent missing information documenting when a patient first experienced symptoms, and the frequent retrospective revision of historical information (including right censoring). This is done by using a novel back-calculating procedure based on imputing patients' dates of symptom onset from reported cases, according to a dynamically-estimated "backward" reporting delay conditional distribution, and adjusting for right censoring using an existing package, NobBS , to estimate in real time (nowcast) cases by date of symptom onset. This process allows us to obtain an approximation of the time-varying reproduction number ( R t ) in real-time. At each step, we evaluate how different assumptions affect the recovered epidemiological events and compare the proposed approach to the alternative procedure of merely using curves of case counts, by report day, to characterize the time-evolution of the outbreak. Finally, we assess how these real-time estimates compare with subsequently documented epidemiological information that is considered more reliable and complete that became available later in time. Our approach may help improve accuracy, quantify uncertainty, and evaluate frequently unstated assumptions when recovering the epidemic curves from limited data obtained from public health surveillance systems in other locations.

Coverage and predictors of influenza vaccination in patients with cystic fibrosis in a campaign with a mobile phone text messaging intervention

Annual influenza vaccination is considered the best protection against influenza infection. We analyzed the influenza vaccine coverage (IVC) in cystic fibrosis (CF) patients and evaluated the factors associated with the IVC, including the effect of text-message/SMS reminders.

We performed a cross-sectional study in the Community of Madrid (Spain) in 2015. The target population was people with CF older than 6 months of age at the beginning of the flu vaccination campaign. The IVC was calculated according to the study variables.

A total of 445 CF patients were analyzed. In 2015, IVC reached 67.9% and was higher in children and women. The main factor associated with flu vaccination was having been vaccinated in the previous campaign (aOR 14.36; IC95%: 8.48–24.32).

The probability of being vaccinated after receiving the SMS was more than twice than for those who did not receive it, although no statistical significance was reached.

In conclusion the IVC of patients with CF is high, but it still has room for improvement. SMS reminders sent to CF patients might improve influenza vaccine uptake.

Request for HIV serology in primary care: A survey of medical and nursing professionals

INTRODUCTION

In the Community of Madrid there is 42.7% late HIV diagnosis. Primary care is the gateway to the health system and the frequency of serological tests requested by these professionals is unknown. The objectives were to establish the frequency of requests for HIV serology by medical and nursing primary care professionals in the Community of Madrid and the factors associated with these requests.

METHOD

An 'on-line' survey was conducted, asking professionals who participated in the evaluation study of strategies to promote early diagnosis of HIV in primary care in the Community of Madrid (ESTVIH) about the number of HIV-serology tests requested in the last 12 months. The association between HIV-serology requesting and the sociodemographic and clinical practice characteristics of the professionals was quantified using adjusted odds ratios (aOR) according to logistic regression.

RESULTS

264 surveys (59.5% physicians). Eighty-two point two percent of medical and 18.7% of nursing professionals reported requesting at least one HIV-serology in the last 12 months (median: 15 and 2 HIV-serology request, respectively). The doctors associated the request with: being male (aOR: 2.95; 95% CI: 0.82-10.56), being trained in pre-post HIV test counselling (aOR: 2.42; 95% CI: 0.84-6.93) and the nurses with: age (<50 years; aOR: 2.75; 95% CI: 0.97-7.75), and number of years working in primary care (>13 years; aOR: 3.02; 95% CI: 1.07-8.52).

CONCLUSION

It is necessary to promote HIV testing and training in pre-post HIV test counselling for medical and nursing professionals in primary care centres.

Main causes of hospitalization in people with Angelman syndrome

BACKGROUND

The objective was to describe the main causes of hospitalization in people with Angelman syndrome (AS).

METHOD

Population-based cross-sectional study in the Community of Madrid (CM), Spain. The information source for AS cases was the information system for rare diseases in the CM. Variables related to hospitalization, for the period 2006-2014, were the following: number of episodes, outcome, main cause, length of stay and type of admission. Main causes of hospitalization were described by age group and sex.

RESULTS

The most frequent causes of hospitalization were the following: oral-dental care (28.9%), seizures (19.6%), orthopaedic problems (14.4%) and acute respiratory disorders (12.4%). The percentage of hospitalizations was higher for oral-dental care in women and for orthopaedic problems in men (p-value <.05). Hospitalizations for an acute respiratory disorder were higher in adults (p-value <.05).

CONCLUSIONS

Some differences in the causes of hospitalization of people with AS were observed by sex and age.

Autochthonous Crimean-Congo Hemorrhagic Fever in Spain

Crimean-Congo hemorrhagic fever (CCHF) is a widely distributed, viral, tickborne disease. In Europe, cases have been reported only in the southeastern part of the continent. We report two autochthonous cases in Spain. The index patient acquired the disease through a tick bite in the province of Ávila - 300 km away from the province of Cáceres, where viral RNA from ticks was amplified in 2010. The second patient was a nurse who became infected while caring for the index patient. Both were infected with the African 3 lineage of this virus. (Funded by Red de Investigación Cooperativa en Enfermedades Tropicales [RICET] and Efficient Response to Highly Dangerous and Emerging Pathogens at EU [European Union] Level [EMERGE].).

Age and gender differences in Clostridium difficile-related hospitalization trends in Madrid (Spain) over a 12-year period

This study aimed to analyze temporal trends by gender and age in Clostridium difficile infection (CDI)-related hospitalization rates in the Autonomous Community of Madrid (Spain) over a 12-year period. A population-based cross-sectional study of all hospital admissions with a CDI diagnosis from 2003 to 2014 was carried out. Annual age-specific hospitalization rates were calculated by gender. All the analyses were performed separately for total hospitalizations and hospitalizations with CDI as the primary diagnosis. Joinpoint regression models were used to analyze time trends. A total of 13,526 hospital discharges were identified (26.8 % with CDI as the primary diagnosis). In both sexes, a gradient in age-specific rates was observed, ranging in 2014 from 5.92 hospitalizations per 100,000 person-years in patients <15 years of age to 378.96 in patients ≥85 years of age. Since 2009, in the age group of 15-44 years, both men and women presented an increasing trend of around 18 %. A significantly increasing trend was detected in women of age 45-84 years, with an estimated annual percentage of change of 7.6 % in the age group of 45-64 years, and rounding with 4.5 % in the age group of 65-84 years. In men of age 45-64 years, the average annual percentage of increase was 4.7 %, and it was 21.1 % between 2010 and 2014 in the age group of 65-74 years. No trends were identified in the 85 years and over age group. Surveillance methods to assess trends by age group should be implemented. Preventive and therapeutic initiatives should remain a priority.

Coverage of and factors associated with pneumococcal vaccination in chronic obstructive pulmonary disease

SETTING

Patients aged ⩾ 40 years with a diagnosis of chronic obstructive pulmonary disease (COPD; codes R95, R79 and R91 in the International Classification for Primary Care) registered in primary care clinical records in the Autonomous Community of Madrid, Spain.

OBJECTIVE

To assess pneumococcal vaccination coverage in patients with COPD and to analyse factors associated with vaccination uptake.

DESIGN

Population-based cross-sectional study in which data were collected in September 2010.

RESULTS

We found that 93,797 patients (72.0% men and 28.0% women) had COPD. Overall coverage was 65.5% (67.5% men vs. 60.4% women, P < 0.001). In patients aged 40-59 years, coverage was 19.5%, reaching 75.8% in those aged ⩾ 60 years. In patients aged <60 years, uptake was associated with a higher number of comorbidities and appropriate adherence to seasonal influenza and pandemic vaccination schedules. In patients aged ⩾ 60 years, factors associated with uptake in both sexes were older age and appropriate adherence to seasonal influenza vaccination schedules. Factors associated with uptake in men were concomitant comorbidities and pandemic vaccination.

CONCLUSION

Vaccination coverage in individuals aged <60 years with COPD is less than acceptable in Madrid. Coverage was higher in men and in patients with another chronic condition.

Dealing with Ebola virus disease in Spain: epidemiological inquiries received by the Department of Public Health Alerts, April to December 2014

We describe the inquiries regarding Ebola virus disease (EVD) received by the Department of Public Health Alerts of the Community of Madrid between April and December 2014. A total of 242 inquiries were received. Consultations were initiated most frequently by hospital clinicians (59 inquiries, 24%), private citizens (57 inquiries, 24%) and primary care physicians (53 inquiries, 22%). The most frequent topic of inquiry was possible EVD in a patient (215 inquiries, 89%). Among these, 31 persons (14%) presented both EVD-compatible symptoms and epidemiological risk factors, and 11 persons (5%) fulfilled the criteria for a person under investigation. Recent travel abroad was reported in 96 persons (45%), but only 32 (15%) had travelled to an EVD-affected area. Two high-risk and one low-risk contact were identified through these inquiries. Low specificity of the EVD symptoms led to many difficulties in protocol application. Ineffective communication with healthcare professionals and unfamiliarity with the EVD protocols caused many case classification errors. A rapid consultation service by telephone is essential for providing qualified advice during emergencies. Our experience may help other countries dimension their activities and resources for managing similar exceptional outbreaks in the future.

Sociodemographic characteristics and chronic medical conditions as risk factors for herpes zoster: a population-based study from primary care in Madrid (Spain)

The objectives of this study were to estimate incidence density rates (IDR) of herpes zoster (HZ) and to analyze the association with sociodemographic characteristics and selected chronic medical conditions. The study cohort consisted of the adult population included in the Public Health System of the Autonomous Community of Madrid, Spain on 1/10/2009 (5 244 402 persons). Data source were electronic medical records from primary care between 1/10/2009-31/12/2012. Individual socioeconomic status (SES) was inferred by geocoding. Poisson regression analyses were stratified by sex, to identify factors associated with HZ. We identified 81 541 incident cases of HZ (61.7% in women and 46.5% in the group aged 60 and over). IDR was 4.11 per 1000 person-years in men and 5.95 in women. IDR were higher with age, in autochthonous population, those with lower SES and in patients with immunodeficiencies. After adjustment, higher incidence rate ratios were found with age, autochthonous origin, lower SES, and in patients with HIV-infection/AIDS (3.20, CI95% 2.90-3.53 in men and 2.98, CI95% 2.58-3.45 in women), and other immunodeficiencies (1.57, CI95% 1.41-1.75 and 1.65, CI95% 1.50-1.80). COPD, asthma, DM, ischemic heart disease, other cardiovascular diseases, and cancer were also associated with an increased incidence of HZ. We conclude that older, autochthonous patients with lower SES and with certain underlying medical conditions had a higher probability of suffering HZ. Electronic databases are useful for estimating the incidence of HZ, and for finding associations with sociodemographic and clinical characteristics. Identifying unrecognized risk factors for HZ, such as asthma or cardiovascular diseases, is crucial to interpret the epidemiology of HZ, to target vaccination programs and to monitor their effect.

Coverage and predictors of vaccination against 2012/13 seasonal influenza in Madrid, Spain: analysis of population-based computerized immunization registries and clinical records

We aim to determine 2012–13 seasonal influenza vaccination coverage. Data were analyzed by age group and by coexistence of concomitant chronic conditions. Factors associated with vaccine uptake were identified. We also analyze a possible trend in vaccine uptake in post pandemic seasons.

We used computerized immunization registries and clinical records of the entire population of the Autonomous Community of Madrid, Spain (6 284 128 persons) as data source.

A total of 871 631 individuals were vaccinated (13.87%). Coverage for people aged ≥ 65 years was 56.57%. Global coverage in people with a chronic condition was 15.7% in children and 18.69% in adults aged 15–59 years.

The variables significantly associated with a higher likelihood of being vaccinated in the 2012–13 campaign for the age groups studied were higher age, being Spanish-born, higher number of doses of seasonal vaccine received in previous campaigns, uptake of pandemic vaccination, and having a chronic condition.

We conclude that vaccination coverage in persons aged &lt;60 years with chronic conditions is less than acceptable. The very low coverage among children with chronic conditions calls for urgent interventions. Among those aged ≥60 years, uptake is higher but still far from optimal and seems to be descending in post-pandemic campaigns. For those aged ≥65 years the mean percentage of decrease from the 2009/10 to the actual campaign has been 12%.

Computerized clinical and immunization registers are useful tools for providing rapid and detailed information about influenza vaccination coverage in the population.

Negative trends from 2008/9 to 2011/12 seasons in influenza vaccination coverages among high risk subjects and health care workers in Spain

OBJECTIVES

We aim to describe influenza vaccination coverage for the Spanish population across four consecutive campaigns (2008/2009 to 2011/2012). The data was analyzed by high risk groups and health care workers (HCWs). Also, coverage trends were analyzed to assess uptake in post-pandemic seasons.

METHODS

We used data from two nation-wide representative health surveys namely the 2009/10 European Health Interview Survey for Spain (N=22,188) and the 2011-12 Spanish National Health Survey (N=21,007) Influenza vaccination status was self-reported. We analyzed influenza vaccine coverage by age, sex, number of chronic conditions, being a heath care worker (HCWs) and nationality. Time trends for campaigns among high risk groups were estimated by a multivariate logistic regression model.

RESULTS

We analyzed data from 43,072 subjects aged ≥ 16 years. As a whole, coverage decreased by 3.31% (22.57-19.26%) between the 2008/2009 and 2011/2012 campaigns with a significant decreasing trend (OR 0.92; 95% CI: 0.90-0.94). Coverage in people under 60 years with a chronic disease decreased significantly (OR 0.92: 95% CI: 0.85-0.99) during the analyzed period from 21.02% in 2008/2009 to 17.40% in 2011/2012. Among HCWs, the highest influenza vaccination coverage was achieved in 2009/2010 (31.08%) in the latest campaign coverage has almost halved (17.88%). For the 2011/2012 season and for all age groups the variables associated with a higher probability of having received the influenza vaccine were older age and presence of associated chronic conditions. Among those aged ≥ 60 years, immigrants had lower uptake (OR 0.60; 95% CI: 0.32-0.99).

CONCLUSIONS

Seasonal influenza vaccine uptake rates in the recommended target groups in Spain are unacceptably low and seem to be decreasing in the post pandemic seasons. Further studies are necessary to precisely identify reasons for non-compliance and barriers to influenza vaccination. Meanwhile urgent strategies to improve seasonal vaccination uptake must be discussed and implemented.

Temporal trends in incidence rates of herpes zoster among patients treated in primary care centers in Madrid (Spain), 2005-2012

OBJECTIVES

This study aimed to estimate total and age-specific incidence rates of HZ with data from electronic clinical records in primary care (ECRPC) and to analyze trends by sex and age.

METHODS

Descriptive cross-sectional study covering the incident HZ episodes registered in the ECRPC of the Madrid Regional Public Health System in 2005-2012. Annual crude and age-adjusted incidence rates were calculated. Differences by sex and age were assessed by poisson regression. The annual percentage of change (APC) of incidence rates and 'breakthrough points' of the time trends were determined with the Joinpoint Regression Program.

RESULTS

211,650 episodes of HZ were identified (60.6% women, 52.2% > 55 years). The incidence rate increased from 363.21 to 481.92 per 100,000 person-year in 2005-2012. Rates were higher among women and increased with age. The APC for the period was 3.59% in men and 3.67% in women (p < 0.05). Age-specific rates increased in patients over 14 years. The APC in the 25-44 age group was 7.4% since 2007. The incidence rate ratio (women/men) was highest in this group.

CONCLUSIONS

The incidence of HZ presents an upward trend in 2005-2012 in adults and the elderly. Monitoring the incidence and age-specific rates, will help to detect changes in trends.

The association between in-patient death, utilization of hospital resources and availability of palliative home care for cancer patients

BACKGROUND

The impact of palliative home care programs on in-patient admissions and deaths has not been appropriately established.

AIM

The main objectives of this study have been to evaluate the frequency of in-patient hospital deaths and the use of hospital resources among cancer patients in two areas of the Madrid Region, as well as to assess differences between one area with and one without a palliative home care team (PHCT) in those variables.

DESIGN AND SETTING

We conducted a population-based study comparing two adjacent metropolitan areas of approximately 200,000 inhabitants each in the Madrid Region, Spain, measuring in-patient deaths, emergency room admissions and in-patient days among cancer patients who died in 2005. Only one of the two areas had a fully established PHCT.

RESULTS

524/549 cancer patients (95%) had an identified place of death: 74% died in hospital, 17% at home, 6% in an in-patient hospice and 3% in a nursing home. The frequency of hospital deaths was significantly lower among patients of the PHCT area (61% versus 77%, p < 0.001), as well as the number of patients using emergency and in-patient services (68% versus 79%, p = 0.004, and 66 versus 76%, p = 0.012, respectively). After adjusting for other factors, the risk of hospital death was lower among patients older than 80 (OR, 95% CI, 0.3, 0.1-0.5), higher among patients with hematological malignancies (OR 6.1, 2.0-18.9) and lower among patients of the PHCT area (OR 0.4, 0.2-0.6).

CONCLUSIONS

Our findings suggest that a PHCT is associated with reduced in-patient deaths and overall hospitalization over the last two months of life.

Pandemic influenza A (H1N1) 2009 in Madrid, Spain: incidence and characteristics in immigrant and native population

BACKGROUND

Some immigrants and refugees might be more vulnerable than other groups to pandemic influenza because of pre-existing health and social disparities, migration history and living conditions. The objective of this study was to compare, between the immigrant and autochtonous population, the incidence and characteristics of influenza cases consulting in primary care (PC) and severe influenza cases.

METHODS

Descriptive cross-sectional study of influenza episodes registered in PC and severe influenza cases reported between 1 May 2009 and 22 May 2010, by gender and origin. Age-adjusted rates were calculated and the association between origin and chronic pathology, pregnancy, delay in admission to hospital and admission to intensive care units (ICU) was analyzed by logistic regression and generalized linear models.

RESULTS

The influenza rate in PC, adjusted by age, was lower for immigrant population (2396.3, 95% confidence interval (95% CI) 2362.5-2430.0 vs. 2795.9, 95% CI 2780.4-2811.5 per 100, 000). The difference between severe influenza rates by origin was not statistically significant. Chronic conditions were less common in immigrant population. In severe influenza cases, pregnancy was more common in immigrant women, and the probability of admission to ICU was higher in men from Central and Eastern Europe (prevalence ratio (PR) 8.44, 95% CI 2.81-25.40) and North African women (PR 3.30, 95% CI 1.09-10.05).

CONCLUSION

Differences in influenza rates were detected by origin. This information could be useful for new pandemic wave management purposes, in addition to targetting future investigations. Pandemic influenza preparedness and response plans should incorporate specific actions to improve immigrants' access to health services and to decrease cultural barriers.

Coverage and predictors of vaccination against 2009 pandemic H1N1 influenza in Madrid, Spain

This study aimed to ascertain the coverage of vaccination against pandemic influenza in individuals aged over 6 months for whom vaccination is indicated due to a chronic health condition using as data source clinical information recorded in the primary care clinical history. Of all those for whom vaccination was indicated (1,114,632), 14.6% (162,616) finally received the vaccine. There were statistically significance differences in coverage for sex (16.5% for men and 13.1% for women), age groups (5% for people under 30 years and 20% for those over 60), number of chronic conditions (11.1% for one condition, 22.5% for two conditions, and 31.3% for three or more conditions) and depending on the chronic health condition considered. The probability of being vaccinated increased with male sex, age, number of indications, type of medical card (lower among no income) and having been vaccinated against 2009 season influenza. We concluded that the coverage finally reached for those people with an indication due to chronic health condition in the H1N1 campaign was much lower than expected and wished. It is essential to investigate the different factors that could have intervened in the behavior of the population so that more efficient approaches can be adopted in future influenza pandemics.

Epidemiological usefulness of population-based electronic clinical records in primary care: estimation of the prevalence of chronic diseases

BACKGROUND

Chronic health problems are the main cause of disease, disability and death in developed countries, and their prevalence is increasing.

OBJECTIVE

To estimate the prevalence of selected chronic illnesses based on electronic clinical records in primary care (ECRPC) and to assess its usefulness for epidemiological research, by comparing ECRPC data against those reported by a contemporary health survey.

DESIGN

Descriptive cross-sectional study.

SETTING

All primary care medical consultations in the Madrid Regional Public Health System (PHS).

SUBJECTS

A total of 23 535 182 ECRPC-registered episodes of illness, generated by PHS patients over 15 years of age seeking medical care during 2005-06.

MAIN OUTCOME MEASURES

Prevalences of chronic diseases estimated on the basis of medically examined cases registered in ECRPC and morbidity as reported by a contemporary health survey covering the same geographic area.

RESULTS

A total of 52.5% of the adult population had some chronic health problem. The highest overall prevalences were hypertension (14.8%), mental disorders (12.0%) and allergy (11.6%). Prevalences were generally highest among women, elderly and the native population. Depending on the specific disease, ECRPC-based prevalences were similar to (e.g. diabetes), higher (e.g. chronic skin problems) or lower (e.g. asthma and dyslipidaemia) than those reported by surveys, with certain age- and sex-related variations.

CONCLUSIONS

Prevalences estimated from ECRPC and survey data present variations depending on the disease, age and sex. Both data sources provide complementary information about chronic disease prevalence. ECRPC have the advantage of generating an ongoing standardized register and entailing no additional effort for health professionals.

[Prevalence of diagnosed chronic disorders in the immigrant and native population]

OBJECTIVE

To estimate the prevalence rates of chronic disorders in immigrants and to compare them with those in the native population, based on electronic clinical records in primary care (ECRPC).

METHODS

We performed a descriptive cross-sectional study in patients aged 16 and over included in the Madrid Regional Public Health System. Age-adjusted prevalence rates for each sex and region were estimated on the basis of medically examined cases registered in the ECRPC with any new data entry made in 2005 or 2006.

RESULTS

After age-adjustment, a total of 36.8% immigrants had some chronic health problem (vs. 55.3% natives). These disorders were more frequent among women and among the population from Africa and Latin America. The highest overall prevalence rates in the foreign population were allergy (10.2% crude rate), low-back pain (9.1%), chronic skin problems (6.8%) and mental disorders (6.4%).

CONCLUSIONS

The prevalence rate of chronic disease is lower in the foreign population and differs according to sex and country of origin.