Andrew Noymer

Vitamin D (25OHD) Serum Seasonality in the United States.

Background Vitamin D is an important micronutrient for health. Hypovitaminosis D is thought to play a role in the seasonality of a number of diseases and adverse health conditions. To refine hypotheses about the links between vitamin D and seasonal diseases, good estimates of the cyclicality of serum vitamin D are necessary. Objectives The objective of this study is to describe quantitatively the cyclicality of 25-hydroxyvitamin D (25OHD) in the United States. We provide a statistical analysis with weekly time resolution, in comparison to the quarterly (winter/spring/summer/fall) estimates already in the literature. Methods We analyzed time series data on 25OHD, spanning 287 consecutive weeks. The pooled data set comes from 3.44 million serum samples from the United States. We statistically analyzed the proportion of sera that were vitamin D sufficient, defined as 25OHD ng/mL, as a function of date. Results In the United States, serum 25OHD follows a lagged pattern relative to the astronomical seasons, peaking in late summer (August) and troughing in late winter (February). Airmass, which is a function of solar altitude, fits the 25OHD data very well when lagged by 8 weeks. Conclusions Serum vitamin D levels can be modeled as a function of date, working through a double-log transformation of minimal solar airmass (easily calculated from solar altitude, retrievable from an online solar altitude/azimuth table).

The transmission and persistence of 'urban legends': Sociological application of age-structured epidemic models

This paper describes two related epidemic models of rumor transmission in an age-structured population. Rumors share with communicable disease certain basic aspects, which means that formal models of epidemics may be applied to the transmission of rumors. The results show that rumors may become entrenched very quickly and persist for a long time, even when skeptics are modeled to take an active role in trying to convince others that the rumor is false. This is a macrophenomeon, because individuals eventually cease to believe the rumor, but are replaced by new recruits. This replacement of former believers by new ones is an aspect of all the models, but the approach to stability is quicker, and involves smaller chance of extinction, in the model where skeptics actively try to counter the rumor, as opposed to the model where interest is naturally lost by believers. Skeptics hurt their own cause. The result shows that including age, or a variable for which age is a proxy (e.g., experience), can improve model fidelity and yield important insights.

Mortality selection and sample selection: a comment on Beckett.

Mortality Selection and Sample Selection: A Comment on Beckett* ANDREW NOYMER University of CaIifornia—BerkeIey Journal of Health and Social Behavior 2001, Vol 42 (Septmber): 326-327 In an interesting article, Megan Beckett (2000) examines the important question of converging health inequalities in later life. Many studies have shown that the differences in health across socioeconomic strata narrow at older ages. Using panel data from the National Health and Nutrition Examination Survey (NHANES), Beckett shows that the converging health inequality cannot be accounted for by mortality selection. The pre- sent comment reconsiders Beckett’s approach to the selection problem, which, while creative, is open to multiple interpretations. Consider the phenomenon that would cause converging health inequalities at later ages to be an “artifact,” as Beckett puts it, of mortality selection. At younger ages, persons with high- er socioeconomic status (SES) have lower lev- els of health problems than those with lower SES. At older ages, the prevalence of health problems in the two groups is closer to parity. If patterns in morbidity are mirrored in mortal- ity, then at older ages a lower SES cohort (higher morbidity and mortality) will be small- er compared to its starting size than a higher SES cohort (lower morbidity and mortality). Since the seminal work of Vaupel, Manton, and Stallard (1979) and Keyfitz and Littman (1979), many demographers have assumed that there are different rates of “frailty” within a population, which determine an individual’s deviation, at any age, from some baseline mor- tality risk. According to the frailty hypothesis, those who die at young ages tend to have high frailty, which skews the distribution of sur- vivors to be more robust. If this condition of nonrandom mortality risks is met, then the aged low SES cohort will be more robust than ‘Address correspondence to: Andrew Noymer, Departments of Sociology and Demography, University of Califomia-—Berkeley, 2232 Pied- mont Avenue, Berkeley, CA 94720; email: andrew@demog.berkeley.edu. when it started out. This reversal of fortune over the life course is called “cohort inversion” (Hobcraft, Menken, and Preston 1982). On the other hand, the low mortality, high SES cohort will have a much less-changed frailty distribu- tion, and will experience less cohort inversion. The greater cohort inversion of the low SES cohort could be enough to overcome the mor- tality disadvantage of being low SES. This problem must be analyzed cautiously, however, as the entire framework for understanding mortality selection effects rests on a counter- factual foundation. If we hold that convergence is a result of mortality selection, we imply that an intrinsic differential persists into older ages and that we would observe it were it not for the selection efiect. On the other hand, if we hold that the convergence is either intrinsic or the result of, for example, access to Medicare (Beckett 2000), we posit that even without selection we would see convergence. In both cases, there is the troubling verb “would.” In reality, we can only observe vital rates that do occur, not those that would occur if some con- dition is met. The general problem of sample selection is encountered frequently in the social sciences (cf. Stolzenberg and Relles 1997; Winship and Mare 1992), and as Beckett (2000) notes, dif- ferential mortality is just a special case of the more general problem. Although we cannot simply “control for” (i.e., condition on) selec- tion bias the same way we would a confound- ing variable, statistical techniques do exist that try to counteract the bias. However, mortality selection is a very special case of sample selec- tion, all the more so if the dependent variable in question is itself health-related. Because of cohort inversion, sample selection due to mor- tality has causal implications beyond nonran- dom missing data in panel followups. This is what makes Beckett’s approach problematic. Consider the statistical technique used by Beckett to set up the hypothetical of no mor- 326

Influenza Mortality in the United States, 2009 Pandemic: Burden, Timing and Age Distribution

Background In April 2009, the most recent pandemic of influenza A began. We present the first estimates of pandemic mortality based on the newly-released final data on deaths in 2009 and 2010 in the United States. Methods We obtained data on influenza and pneumonia deaths from the National Center for Health Statistics (NCHS). Age- and sex-specific death rates, and age-standardized death rates, were calculated. Using negative binomial Serfling-type methods, excess mortality was calculated separately by sex and age groups. Results In many age groups, observed pneumonia and influenza cause-specific mortality rates in October and November 2009 broke month-specific records since 1959 when the current series of detailed US mortality data began. Compared to the typical pattern of seasonal flu deaths, the 2009 pandemic age-specific mortality, as well as influenza-attributable (excess) mortality, skewed much younger. We estimate 2,634 excess pneumonia and influenza deaths in 2009–10; the excess death rate in 2009 was 0.79 per 100,000. Conclusions Pandemic influenza mortality skews younger than seasonal influenza. This can be explained by a protective effect due to antigenic cycling. When older cohorts have been previously exposed to a similar antigen, immune memory results in lower death rates at older ages. Age-targeted vaccination of younger people should be considered in future pandemics.

Magnitude of Ebola relative to other causes of death in Liberia, Sierra Leone, and Guinea

Correspondence With more than 20 000 cases reported, the outbreak of Ebola virus disease (EVD) in west Africa is by far the largest in recorded history. Despite the scale of the current outbreak, EVD is often perceived as a “small-scale killer”. 1 By comparison, malaria caused an estimated 854 000 deaths world wide in 2013. 2 However, although limited at the global level, the impact of EVD on mortality could be substantial in countries with intense transmission. We thus aimed to compare EVD with other causes of death in Liberia, Sierra Leone, and Guinea in 2014. We did an uncertainty analysis of EVD mortality (see appendix), based on two parameters: the extent of under- reporting of EVD cases and the case fatality rate (CFR)—ie, the proportion of EVD cases who die. Similar to other analyses of EVD spread, 3 we hypothesised that there were up to 2·5 times more EVD cases than reported. This factor derives from a mathematical model, which compared the reported number of EVD cases to the number of beds in use in Ebola treatment units in August, 2014. 4 We assumed that the CFR varied between 60% and 85%. The lower rate corresponds to CFRs seen among hospitalised EVD patients with known disease outcomes. 5 Lower CFRs have been documented, but only in Ebola treatment units that implement non-standard treatment protocols. 6 The upper rate corresponds to CFRs seen in non-hospitalised EVD patients. 5 We estimated the number of EVD deaths as the product of (1) the reported number of EVD cases, (2) the under-reporting factor and (3) the CFR. Based solely on confirmed and probable EVD cases, the number of EVD deaths in 2014 ranged from 2928 to 10 372 in Liberia, from 4468 to 15 824 in Sierra Leone, and from 1739 to 5548 in Guinea. www.thelancet.com/lancetgh Vol 3 May 2015 We used the most recent (2013) national estimates of non-EVD mortality, 2,7 together with projections of population growth, to calculate the expected number of deaths from non- EVD causes in Liberia, Sierra Leone, and Guinea in 2014. For all com binations of model parameters, we mapped how the estimated number of EVD deaths ranked relative to the expected number of deaths from non-EVD causes. Liberia In Liberia, for virtually all model parameters, EVD deaths exceeded the expected number of deaths due to the leading non-EVD cause of death (fi gure). In Sierra Leone, a broad range of model parameters also indicated that EVD might have killed more people in 2014 than the leading non- EVD cause of death (ie, malaria). In other sets of model parameters, EVD still killed more people than the second Sierra Leone Guinea 10k 15k 5k 10k Ratio of true: reported EVD cases Magnitude of Ebola relative to other causes of death in Liberia, Sierra Leone, and Guinea 5k Malaria See Online for appendix 2·5k LRI 5k LRI Malaria Case fatality rate among EVD cases (%) Ranking of EVD deaths relative to non-EVD causes of death EVD=1st cause of death EVD=3rd cause of death EVD=lower ranking EVD=2nd cause of death Figure: Comparison of Ebola virus disease (EVD) deaths and expected deaths from non-EVD causes in Liberia, Sierra Leone, and Guinea in 2014 LRI=lower respiratory infections. White contours represent estimated number of EVD deaths in 2014 for specifi c combinations of model parameters (where 2·5k = 2500 deaths, etc). Calculations of number of EVD deaths, and comparisons with other causes of death, are described in the appendix. Range of possible case fatality rates (CFRs) is narrower in Guinea than in the other two countries because a higher percentage of EVD deaths was recorded in that country among reported EVD cases. Lower bound of CFR is thus set at level estimated by surveillance data in that country. To aid fi gure interpretation, we illustrate the case of Sierra Leone. In the red area, there are more EVD deaths than we expect deaths from the leading non-EVD cause of death in that country—ie, malaria. Hence, in this region of the parameter space, EVD would be the fi rst cause of death. In the light blue area, there are fewer EVD deaths than deaths from the leading non-EVD cause of death, but there are more EVD deaths than deaths from the second non-EVD cause of death (ie, LRI). Hence EVD would be the second cause of death in that region of the parameter space. In the slightly darker blue area at the bottom of the graph, there are fewer EVD deaths than deaths from LRI, but more than deaths from HIV/ AIDS—ie, the third leading non-EVD cause of death (see appendix). EVD would thus be the third leading cause of death in that region of the parameter space. The boundary between the red and light blue areas (marked “malaria”) represents combinations of model parameters where the number of EVD deaths is equal to the expected number of malaria deaths. The boundary between the light blue and the darker blue areas (marked “LRI”) represents combinations for which the number of EVD deaths is equal to the expected number of LRI deaths. In Guinea, the entire graph is dark blue because the number of EVD deaths is lower than the expected number of deaths from the third non-EVD cause of death in the country for all model parameters. e255

THE MARCH OF DIMES

 LETTERS  and the Institute for Social Research, University of Michi- gan, Ann Arbor. Requests for reprints should be sent to Nancy Krieger, PhD, Department of Health and Social Behavior, Harvard School of Public Health, 677 Huntington Ave, Boston, MA 02115 (e-mail: nkrieger@hsph.harvard.edu). THE MARCH OF DIMES The August 2001 Images of Health article, “ ‘. . . So That Others May Walk’: The March of Dimes,” 1 is fascinating. The March of Dimes is an excellent example of institutional adaptation in the face of structural change. Vaccination greatly reduced the incidence of poliomyelitis in the 1950s and eventually led to the elimination, in 1991, of polio virus from the Americas. In response to this situa- tion, the anti-polio foundation had 3 options: to declare its mission fulfilled and close its doors, to focus on combating polio overseas, or to change its mission. The foundation chose the third path, and it is now officially the March of Dimes Birth Defects Foundation. The poster that is depicted appears to be more recent than the text’s topic of the found- ing of the March of Dimes in the 1930s. The nurse in the background has the words “polio vaccine volunteer” on her sleeve. Thus the “Help me, too” would seem to refer to those children who missed out on vaccination and contracted polio. The poster could be from an anti-complacency campaign conducted in the 1950s. After widespread vaccination was in- troduced in 1955, the public, and especially new parents, began to believe that the advent of vaccination meant that polio was van- quished. As the history of the March of Dimes attests, this period was the beginning of a major institutional adaptation. Andrew Noymer, MSc About the Author the August 2001 issue of the Journal. Noymer’s suggestion that the image dates from the mid-1950s seems highly plausible. The postcard in the National Library of Medi- cine collection is undated, but as the Poster Child campaign was introduced in 1946, the image must have been created after that date. It does seem likely that it was created in con- junction with the anti-complacency campaign. For a good sampling of the March of Dimes materials, interested readers can consult A Paralyzing Fear: The Triumph Over Polio in America, by Nina Gilden Seavey, Jane S. Smith, and Paul Wagner (New York, NY: TV Books; 1998), a companion volume to the film of the same title. Requests for reprints should be sent to Andrew Noymer, MSc, University of California at Berkeley, Department of Demography, 2232 Piedmont Ave, Berkeley, CA 94720 (e-mail: andrew@demog.berkeley.edu). Elizabeth Fee, PhD Theodore M. Brown, PhD Reference About the Authors 1. Helfand WH, Lazarus J, Theerman P. “. . . So That Others May Walk”: the March of Dimes. Am J Public Health. 2001;91:1190. The authors are Contributing Editors of the Journal. Eliz- abeth Fee is with the History of Medicine Division, Na- tional Library of Medicine, National Institutes of Health, Bethesda, Md. Theodore M. Brown is with the Depart- ments of History and of Community and Preventive Medi- cine, University of Rochester, Rochester, NY. Requests for reprints should be sent to Elizabeth Fee, PhD, National Library of Medicine, History of Medicine Division, 8600 Rockville Pike, Bethesda, MD 20894 (e-mail: elizabeth_fee@nlm.nih.gov). FEE AND BROWN RESPOND We were pleased to see the letter from Andrew Noymer about the Images of Health article in ERRATUM In a Letter to the Editor (Gori GB. Individualized or population risks: what is the argument? Am J Public Health. 2001;91:1919), I erro- neously quoted Colin B. Begg as stating, in his March 2001 article (Begg CB. The search for cancer risk factors: when can we stop looking? Am J Public Health. 2001;91:360–364), that “ ‘the primary purpose of epidemiology is to determine individual risks.’ ” In fact, this state- ment is not a direct quote. Begg states that “a primary purpose of epidemiologic research is to determine the extent to which this [unique individual] risk varies from person to person and the factors that explain this variation” (p 361). Gio B. Gori, The Health Policy Center, Bethesda, MD 20816. 158 | Letters American Journal of Public Health | February 2002, Vol 92, No. 2

Ebola virus disease in West Africa - The first 9 months: To the editor [2]

The n e w e ng l a n d j o u r na l of m e dic i n e c or r e sp ondence Ebola Virus Disease in West Africa — The First 9 Months To the Editor: The World Health Organization (WHO) Ebola Response Team (Oct. 16 issue) 1 predicted that the current Ebola epidemic would claim a dreadful 20,000 combined cases by early November 2014, assuming no change in the con- trol measures applied in West Africa. The threat that Ebola poses to national public health and social, economic, and security foundations may worsen if a secondary epidemic eventually ex- plodes in the region. Since June 2014, nearby Ghana has been affected by a serious cholera epidemic that was responsible for 12,622 cases as of September 6. 2 Current cholera and Ebola zones are separated by Ivory Coast, a frequent crossing point for commuters traversing West Africa. To effectively control cholera epidemics, specialized treatment centers, access to potable water, sani- tation, and community hygiene awareness are critical. However, in Ebola-affected areas, quaran- tine units are overwhelmed, many health facili- ties are dysfunctional after the desertion by staff members who fear viral contamination, and it has become increasingly dangerous to conduct awareness campaigns owing to violence against health and humanitarian workers accused of this week’s letters 188 Ebola Virus Disease in West Africa — The First 9 Months 189 Goal-Directed Resuscitation in Septic Shock 191 Malpractice Reform and Emergency Department Care 193 Physiological Approach to Acid–Base Disturbances 196 Inefficacy of Platelet Transfusion to Reverse Ticagrelor spreading Ebola. Likewise, neglecting to rapidly control this cholera epidemic in Ghana could have unpredictable yet potentially devastating consequences. Stanislas Rebaudet, M.D., Ph.D. Assistance Publique–Hopitaux de Marseille Marseille, France Sandra Moore, M.S., M.P.H. Renaud Piarroux, M.D., Ph.D. Aix–Marseille University Marseille, France renaud.piarroux@ap-hm.fr No potential conflict of interest relevant to this letter was re- ported. 1. WHO Ebola Response Team. Ebola virus disease in West Africa — the first 9 months of the epidemic and forward projec- tions. N Engl J Med 2014;371:1481-95. 2. Cholera outbreak in the West and Central Africa: regional update, 2014 (week 35). New York: UNICEF (http://www.unicef .org/cholera/files/Cholera_regional_update_W35_2014_West_ and_Central_Africa.pdf). DOI: 10.1056/NEJMc1413884 To the Editor: The WHO Ebola Response Team describes the epidemiology of Ebola virus dis- ease (EVD) in West Africa using anonymized pa- tient-level data generated from EVD surveillance in multiple countries. These data document the demographic profile of patients with EVD, their risk factors, and the course of their illness. We regret that the WHO neither makes this data set publicly available nor provides an interface to ex- tract customized tabulations. Such data sharing could accelerate the discovery of key factors in the epidemic and could yield insight into the eco- nomic and demographic drivers of the outbreak. It would also permit a better assessment of pos- sible control scenarios. Current models of EVD transmission 1,2 are parameterized with the use of outdated data from much smaller Central Afri- can outbreaks, which limits their applicability to West Africa. Some patient-level data sets collect- n engl j med 372;2 nejm.org january 8, 2015 The New England Journal of Medicine Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on January 11, 2015. For personal use only. No other uses without permission. Copyright

INFLUENZA ANALYSIS SHOULD INCLUDE PNEUMONIA

LETTERS nonphysician sources for Compazine (9%), Diuril (11%), Valium (18%), and Darvon Goldsworthy et al. 1 correctly noted a number of consequences of medication-sharing behav- iors, to which we would add an additional consequence: pharmacoepidemiology studies of drug safety increasingly rely on exposure information drawn from computerized pre- scription or pharmacy records (often as part of HMO or claims data). Our earlier conclusion, now supported by the work of Goldsworthy et al., is therefore even more relevant today: [I]n addition to concern that [information from such sources] includes false positive ‘expo- sures’—prescriptions issued or filled but not consumed—our findings suggest that these data sources may also include appreciable numbers of false negative exposures—drugs consumed but obtained from nonidentified sources. 3(p675) Appreciable exposure misclassification would have very serious consequences when it comes to using prescription data sources to identify adverse effects of medication expo- sures; any such efforts must take into consid- eration the extent to which studied medications are shared. j Allen A. Mitchell, MD About the Author The author is with the Slone Epidemiology Center, Boston, MA. Requests for reprints should be sent to Allen A. Mitchell, MD, Director, Slone Epidemiology Center, 1010 Commonwealth Ave, Boston, MA 02215 (e-mail: amitchell@slone.bu.edu). This letter was accepted June 4, 2008. doi:10.2105/AJPH.2008.144261 References 1. Goldsworthy RC, Schwartz NC, Mayhorn CB. Be- yond abuse and exposure: framing the impact of pre- scription-medication sharing. Am J Public Health. 2008; 2. Mitchell AA, Rosenberg L, Shapiro S, et al. Birth defects related to Bendectin use in pregnancy. I. Oral clefts and cardiac defects. JAMA. 1981;245: 3. Mitchell AA, Cottler LB, Shapiro S. Effect of ques- tionnaire design on recall of drug exposure in pregnancy. Am J Epidemiol. 1986;123:670–676. GOLDSWORTHY RESPONDS Mitchell’s letter concerning our study raises an important public health issue related to prescription medication sharing that our study—focused as it was on direct effects on consumers—did not consider: the impact of such sharing on research and evaluation pred- icated upon prescription data sources. Mitchell is absolutely correct in asserting that there is considerable potential for consumer loaning and borrowing of prescription medication to affect the accuracy of traditional adverse effects reporting systems. We agree that efforts using these sources, whether for ongoing sur- veillance or as part of clinical trials, may need to take into consideration the potential underreporting of adverse effects that may be present when prescription medications are shared. j Richard Goldsworthy, PhD About the Author The author is with Academic Edge, Inc, Bloomington, IN. Requests for reprints should be sent to Richard Goldsworthy, Academic Edge, Inc, 108 E 14th St, Bloomington, IN 47408 (e-mail: rick@academicedge.com). This letter was accepted July 15, 2008. doi:10.2105/AJPH.2008.144584 INFLUENZA ANALYSIS SHOULD INCLUDE PNEUMONIA Doshi 1 makes a number of interesting points, including a trenchant observation about the sharing of surveillance data collected under public auspices. Doshi’s conclusion that ‘‘the next influenza pandemic may be far from a catastrophic event,’’ 1(p944) may turn out to be true. However, the author’s optimism is not supported by his analysis, which omits pneumonia. Doshi contends that the mortality conse- quences of influenza pandemics have been overstated. Epidemiologists, demographers, and vital statisticians analyze mortality from influenza and pneumonia combined as a sin- gle cause. This is because influenza kills through pneumonia and is coded as such on death certificates. Concern about ‘‘labora- tory confirmation’’ 1(p939) of influenza may be germane in the clinic, but vis-a`-vis historical time series, it is nothing more than splitting hairs. Influenza pandemics occur when new strains of influenza A virus emerge. Consider the most November 2008, Vol 98, No. 11 | American Journal of Public Health recent pandemic in 1968–1969, when H3N2 influenza emerged. Figure 1 shows time series data on mortality in the United States, from the multiple cause of death data files, 2 for 60 months encompassing the 1966–1967 to 1970–1971 flu seasons. It is easy to see that the 1968–1969 pandemic brought not just increases in influenza mortality, but also pneumonia mortality. The resemblance be- tween the 2 panels is uncanny, but pneumonia kills more people. Looking at influenza deaths alone misses most of the story. How much are influenza and noninfluenza pneumonia in lockstep? Figure 2 is a scatter plot of influenza deaths (x-axis) versus non- influenza (viz., no explicit mention of influenza) pneumonia deaths (y-axis), on a log–log scale, with linear fit. The data are the same as those from the 60 months plotted in Figure 1. The linear fit is very strong (R 2 = 0.89). This could be dismissed as mechanical correlation, be- cause both causes are cyclical and in phase. However, we see from Figure 1 that, even comparing only winter peaks, whenever in- fluenza rises, so does pneumonia. This may be attributed to the 1968–1969 pandemic—but that’s the point. Some pneumonia is caused by respiratory syncytial virus, and others like it, but why does pneumonia mortality spike during influenza pandemics and covary so strikingly with in- fluenza mortality? Is it plausible that respira- tory syncytial virus becomes more deadly when new strains of influenza emerge? A more likely explanation is that many influenza deaths are being coded as pneumonia deaths. Thus, the usual practice of looking at pneu- monia and influenza mortalities combined is wiser than looking at influenza mortality, sensu stricto. The 1977–1978 flu season serves as a clos- ing vignette for Doshi. The 1977 reemergence of H1N1 was not a pandemic because it was not an antigenic shift. It was believed to have escaped from a laboratory and was identical to strains from 1950, and the world population 20 years and older was largely immune. 5 Doshi uses the occasional aberrant classification of 1977 to argue that there is no ‘‘a priori connection’’ 1(p944) between pandemics and mortality. With the use of properly classified pandemics, the connection is an empirical one, not a priori. The connection is strong, and Letters | 1927

Immigrant Health Around the World: Evidence from the World Values Survey

We describe the relationship between immigrant status and self-rated health around the world, both in raw descriptive statistics and in models controlling for individual characteristics. Using the World Values Survey (1981–2005), we analyze data from 32 different countries worldwide. We estimate four regression models per country. The basic model tests mean differences in self-rated health. Additional models add demographic and social class controls. Introduction of control variables (most particularly, age) changes the results dramatically. In the final model, net of controls, only two countries show poorer immigrant health and three countries show better immigrant health. The multivariate regression models net of controls show few differences in health status between immigrants and the native born. The age structure of immigrant populations is an important mediator of differences in health status compared to the native-born population.

Assessing the direct effects of the Ebola outbreak on life expectancy in Liberia, Sierra Leone and Guinea

Background: An EVD outbreak may reduce life expectancy directly (due to high mortality among EVD cases) and indirectly (e.g., due to lower utilization of healthcare and subsequent increases in non-EVD mortality). In this paper, we investigated the direct effects of EVD on life expectancy in Liberia, Sierra Leone and Guinea (LSLG thereafter). Methods: We used data on EVD cases and deaths published in situation reports by the World Health Organization (WHO), as well as data on the age of EVD cases reported from patient datasets. We used data on non-EVD mortality from the most recent life tables published prior to the EVD outbreak. We then formulated three scenarios based on hypotheses about a) the extent of under-reporting of EVD cases and b) the EVD case fatality ratio. For each scenario, we re-estimated the number of EVD deaths in LSLG and we applied standard life table techniques to calculate life expectancy. Results: In Liberia, possible reductions in life expectancy resulting from EVD deaths ranged from 1.63 year (low EVD scenario) to 5.56 years (high EVD scenario), whereas in Sierra Leone, possible life expectancy declines ranged from 1.38 to 5.10 years. In Guinea, the direct effects of EVD on life expectancy were more limited (<1.20 year). Conclusions: Our high EVD scenario suggests that, due to EVD deaths, life expectancy may have declined in Liberia and Sierra Leone to levels these two countries had not experienced since 2001-2003, i.e., approximately the end of their civil wars. The total effects of EVD on life expectancy may however be larger due to possible concomitant increases in non-EVD mortality during the outbreak.