Profiles

Åke Svensson

Åke Svensson

Professor emeritus

Visa sidan på svenska
Works at Department of Mathematics (incl. Math. Statistics)
Telephone 08-16 45 51
Email akes@math.su.se
Visiting address Roslagsvägen 101, Kräftriket, hus 6
Room 315
Postal address Matematiska institutionen 106 91 Stockholm

About me

My present interest is related to biostatistics in the broad sense, spanning from genetics to clinical trials and epidemiology. In particular I am interested in questions related to studies of infectious diseases, both as regards statistical inference of data on infectious diseases and building probabilistic models for epidemic spread of infections. Recent work aims at understanding basic concepts used in epidemic modelling.

Publications

A selection from Stockholm University publication database
  • 2017. Anna Törner (et al.). Hepatology 65 (3), 885-892

    The Cancer Register (CR) in Sweden has reported that the incidence of primary liver cancer (PLC) has slowly declined over the last decades. Even though all cancers, irrespective of diagnostic method, should be reported to the CR, the PLC incidence may not reflect the true rate. Improved diagnostic tools have enabled diagnosis of hepatocellular carcinoma based on noninvasive methods without histological verification, possibly associated with missed cancer reports or misclassification in the CR. Our objective was to study the completeness and assess the underreporting of PLC to the CR and to produce a more accurate estimate based on three registers. The CR, the Cause of Death Register, and the Patient Register were investigated. Differences and overlap were examined, the incidence was estimated by merging data from the registers, and the number reported to none of the registers was estimated using a log-linear capture-recapture model. The results show that 98% of the PLCs reported to the CR were histologically verified; 80% were hepatocellular carcinoma and 20% were intrahepatic cholangiocarcinoma. Unspecified liver cancer decreased over time and constituted <10% of all reported liver cancers. The CR may underestimate the liver cancer incidence by 37%-45%, primarily due to missed cancer reports. The estimated annual number of liver cancers increased over time, but the standardized incidence was stable around 11 per 100,000. Hepatitis C-associated liver cancer increased and constituted 20% in 2010. Conclusion: There was an underreporting of PLC diagnosed by noninvasive methods; the incidence was considerably higher than estimated by the CR, with a stable incidence over time; reporting needs to improve and combining registers is recommended when studying incidence.

  • 2015. Åke Svensson. Mathematical Biosciences 270 (Part A), 81-89

    A simple class of stochastic models for epidemic spread in finite, but large, populations is studied. The purpose is to investigate how assumptions about the times between primary and secondary infections influences the outcome of the epidemic. Of particular interest is how assumptions of individual variability in infectiousness relates to variability of the epidemic curve. The main concern is the final size of the epidemic and the time scale at which it evolves. The theoretical results are illustrated by simulations.

  • 2013. David Lindenstrand, Åke Svensson. Mathematical Biosciences 246 (2), 272-279

    Data, on the number of infected, gathered from a large epidemic outbreak can be used to estimate parameters related to the strength and speed of the spread. The Malthusian parameter, which determines the initial growth rate of the epidemic is often of crucial interest. Using a simple epidemic SEIR model with known generation time distribution, we define and analyze an estimate, based on martingale methods. We derive asymptotic properties of the estimate and compare them to the results from simulations of the epidemic. The estimate uses all the information contained in the epidemic curve, in contrast to estimates which only use data from the start of the outbreak. 

  • 2013. Åke Svensson. Journal of Mathematical Biology 68 (4), 951-967

    The probability that an observed infection has been transmitted from a particular member of a set of potential infectors is calculated. The calculations only use knowledge of the times of infection. It is shown that the probabilities depend on individual variability in latent and infectious times. The analysis are based on different background information and different assumptions on the progress of infectivity. The results are illustrated by numerical calculations and simulations.

  • 2011. Anna Törner (et al.). American Journal of Epidemiology 174 (8), 969-76

    Selection bias and confounding are concerns in cohort studies where the reason for inclusion of subjects in the cohort may be related to the outcome of interest. Selection bias in prevalent cohorts is often corrected by excluding observation time and events during the first time period after inclusion in the cohort. This time period must be chosen carefully-long enough to minimize selection bias but not too long so as to unnecessarily discard observation time and events. A novel method visualizing and estimating selection bias is described and exemplified by using 2 real cohort study examples: a study of hepatitis C virus infection and a study of monoclonal gammopathy of undetermined significance. The method is based on modeling the hazard for the outcome of interest as a function of time since inclusion in the cohort. The events studied were "hospitalizations for kidney-related disease" in the hepatitis C virus cohort and "death" in the monoclonal gammopathy of undetermined significance cohort. Both cohorts show signs of considerable selection bias as evidenced by increased hazard in the time period after inclusion in the cohort. The method was very useful in visualizing selection bias and in determining the initial time period to be excluded from the analyses.

  • 2010. Anna Törner (et al.). American Journal of Epidemiology 171 (5), 602-608

    Selection bias is a concern in cohort studies in which selection into the cohort is related to the studied outcome. An example is chronic infection with hepatitis C virus, where the initial infection may be asymptomatic for decades. This problem leads to selection of more severely ill individuals into registers of such infections. Cohort studies often adjust for this bias by introducing a time window between entry into the cohort and entry into the study. This paper describes and assesses a novel method to improve adjustment for this type of selection bias. The size of the time window is decided by calculating a standardized incidence ratio as a continuous function of the size of the time window. The resulting graph is used to decide on an appropriate window size. The method is evaluated by using the Swedish register of hepatitis C virus infections for 1990-2006. The complications studied were non-Hodgkin lymphoma and liver cancer. Selection bias differed for the studied outcomes, and a time window of a minimum of 2 months and 12 months, respectively, was judged to be appropriate. The novel method may have advantages compared with an interval-based method, especially in cohort studies with small numbers of events.

  • 2010. Gianpaolo Scalia Tomba (et al.). Mathematical Biosciences 223 (1), 24-31

    The generation time of an infectious disease is usually defined as the time from the moment one person becomes infected until that person infects another person. The concept is similar to ""generation gap"" in demography, with new infections replacing births in a population. Originally applied to diseases such as measles where at least the first generations are clearly discernible, the concept has recently been extended to other diseases, such as influenza, where time order of infections is usually much less apparent. By formulating the relevant statistical questions within a simple yet basic mathematical model for infection spread, it is possible to derive theoretical properties of observations in various situations e.g. in ""isolation"", in households, or during large outbreaks. In each case, it is shown that the sampling distribution of observations depends on a number of factors, usually not considered in the literature and that must be taken into account in order to achieve unbiased inference about the generation time distribution. Some implications of these findings for statistical inference methods in epidemic spread models are discussed.

  • 2009. Patricia Geli (et al.). Journal of Theoretical Biology 256 (1), 58-64

    A multi-type branching process with varying environment was used to construct a pharmacokinetic/pharmacodynamic (PK/PD) model that captures the postantibiotic effect (PAE) seen in bacterial populations after exposure of antibiotics. This phenomenon of continued inhibition of bacterial growth even after removal of the antibiotic from the growth medium is of high relevance in the context of optimizing dosing regimens. The clinical implication of long PAEs lies in the interesting possibility of increasing the intervals between drug administrations.

    The model structure is generalizable to most types of antibiotics and is useful both as a theoretical framework for understanding the time properties of PAE and to explore optimal antibiotic dosing regimens. Data from an in vitro study with Escherichia coli exposed to different dosing regimens of cefotaxime were used to evaluate the model.

  • 2009. Shaban Nyimvua (et al.). Biometrical Journal 51 (3), 408-419

    A Markovian susceptible → infectious → removed (SIR) epidemic model is considered in a community partitioned into households. A vaccination strategy, which is implemented during the early stages of the disease following the detection of infected individuals is proposed. In this strategy, the detection occurs while an individual is infectious and other susceptible household members are vaccinated without further delay. Expressions are derived for the influence on the reproduction numbers of this vaccination strategy for equal and unequal household sizes. We fit previously estimated parameters from influenza and use household distributions for Sweden and Tanzania census data. The results show that the reproduction number is much higher in Tanzania (6 compared with 2) due to larger households, and that infected individuals have to be detected (and household members vaccinated) after on average 5 days in Sweden and after 3.3 days in Tanzania, a much smaller difference.

  • 2009. S Boqvist (et al.). Epidemiology and Infection 137, 897-905

     

    Young children account for a large proportion of reported

    Yersinia enterocolitica infections in Sweden with a high incidence compared with other gastrointestinal infections, such as salmonellosis and campylobacteriosis. A case-control study was conducted to investigate selected risk factors for domestic sporadic yersiniosis in children aged 0–6 years in Sweden. In total, 117 cases and 339 controls were included in the study. To minimize exclusion of observations due to missing data a multiple non-parametric imputation technique was used. The following risk factors were identified in the multivariate analysis : eating food prepared from raw pork products (OR 3.0, 95% CI 1.8–5.1) or treated sausage (OR 1.9, 95% CI 1.1–3.3), use of a baby’s dummy (OR 1.9, 95% CI 1.1–3.2) and contact with domestic animals (OR 2.0, 95% CI 1.2–3.4). We believe that the importance of Y. enterocolitica

    infection in children has been neglected and that results from this study can be used to develop preventive recommendations.

  • 2009. Martin Camitz, Åke Svensson. Bulletin of Mathematical Biology 71 (8), 1902-1913

    In elaborating a model of the progress of an epidemic, it is necessary to make assumptions about the distributions of latency times and infectious times. In many models, the often implicit assumption is that these times are independent and exponentially distributed. We explore the effects of altering the distribution of latency and infectious times in a complex epidemic model with regional divisions connected by a travel intensity matrix. We show a delay in spread with more realistic latency times. More realistic infectiousness times lead to faster epidemics. The effects are similar but accentuated when compared to a purely homogeneous mixing model.

  • 2008. Ann-Sofi Duberg (et al.). J Viral Hepat 15 (7), 538-50
  • 2008. Shaban Nyiumvua (et al.).
  • 2008. Shaban Nyiumvua (et al.). Mathematical Biosciences 216, 1-8
  • 2008. Anders Ternhag (et al.). Emerging Infectious Diseases 14 (1), 143-8

    During 1997–2004, microbiologically confirmed gastrointestinal infections were reported for 101,855 patients in Sweden. Among patients who had Salmonella infection (n = 34,664), we found an increased risk for aortic aneurysm (standardized incidence ratio [SIR] 6.4, 95% confidence interval [CI] 3.1–11.8) within 3 months after infection and an elevated risk for ulcerative colitis (SIR 3.2, 95% CI 2.2–4.6) within 1 year after infection. We also found this elevated risk for ulcerative colitis among Campylobacter infections (n = 57,425; SIR 2.8, 95% CI 2.0–3.8). Within 1 year, we found an increased risk for reactive arthritis among patients with Yersinia enteritis (n = 5,133; SIR 47.0, 95% CI 21.5–89.2), Salmonella infection (SIR 18.2, 95% CI 12.0–26.5), and Campylobacter infection (SIR 6.3, 95% CI 3.5–10.4). Acute gastroenteritis is sometimes associated with disease manifestations from several organ systems that may require hospitalization of patients.

  • 2007. Åke Svensson. Math Biosci 208 (1), 300-11
  • 2007. Birgitte Freiesleben de Blasio, Åke Svensson, Fredrik Liljeros. Proc Natl Acad Sci U S A 104 (26), 10762-7
  • 2006. Katarzyna Grabowska (et al.). BMC Infect Dis 6, 58
  • 2005. Lena Jörgensen (et al.). Scand J Public Health 33 (4), 285-91
  • 2005. Anders Ternhag (et al.). BMC Infect Dis 5, 70
  • 2004. Karin Nygård (et al.). Epidemiol Infect 132 (2), 317-25
  • 2003. Tommi Asikainen, Johan Giesecke, Åke Svensson. Lakartidningen 100 (40), 3126-30
Show all publications by Åke Svensson at Stockholm University

Last updated: May 26, 2020

Bookmark and share Tell a friend