J HEALTH POPUL NUTR 2013 Dec;31(4) Suppl 1:S17-S29 ©INTERNATIONAL CENTRE FOR DIARRHOEAL ISSN 1606-0997 | $ 5.00+0.20 DISEASE RESEARCH, BANGLADESH Estimating the Burden of Acute Gastrointestinal Illness in Grenada Lindonne M. Glasgow1, Martin S. Forde1, Samuel C. Antoine2, Enrique Pérez3, Lisa Indar4 1Department of Public Health and Preventive Medicine, St. George’s University, St. George, Grenada; 2Ministry of Health, Ministerial Complex, St. George, Grenada; 3Pan American Health Organization, Panama; 4Caribbean Public Health Agency, 16-18 Jamaica Boulevard, Federation Park, Port of Spain, Trinidad and Tobago ABSTRACT This is the first study conducted in Grenada, with a population of approximately 108,000, to quantify the magnitude, distribution, and burden of self-reported acute gastroenteritis (AGE). A retrospective popula- tion survey was conducted in October 2008 and April 2009 and a laboratory survey from October 2008 to September 2009. The estimated monthly prevalence of AGE was 10.7% (95% CI 9.0-12.6; 1.4 episodes/ person-year), with a median of 3 days of illness. Of those who reported AGE, 31% sought medical care (stool samples were requested from 12.5%); 10% took antibiotics; 45% took non-prescribed medication; and 81% reported restricted activity. Prevalence of AGE was significantly higher among children aged <5 years (23.5%, p<0.001). Of the AGE stool samples submitted to the laboratory for analysis, 12.1% were posi- tive for a foodborne pathogen. Salmonella enteritidis was the most common foodborne pathogen associated with AGE-related illness. The estimated percentage of underreporting of syndromic AGE to the Ministry of Health was 69%. In addition, for every laboratory-confirmed foodborne/AGE pathogen, it was estimated that there were 316 additional cases occurring in the population. The minimum estimated cost associated with treatment for AGE was US$ 703,950 each year, showing that AGE has a potentially significant eco- nomic impact in Grenada. Key words: Acute gastroenteritis; Burden of Illness Study; Diarrhoea; Foodborne diseases; Foodborne pathogen; Laboratory survey; Population survey; Surveillance system; Grenada INTRODUCTION Petite Martinique (area: 586 acres, population ~ 800) commonly experience harsher dry seasons, have no Grenada is the main island of the Tri-island State rivers, and surface-water is virtually non-existent. comprising the mainland and two dependency Tourism and agriculture are currently the main live- islands Carriacou and Petite Martinique, with a lihood activities and sources of generating foreign total area of 133 square miles (344 sq. km) and a exchange in the Tri-island State. The islands are still population of approximately 108,000. The main recovering from the effects of two devastating hurri- island—Grenada—is divided into six parishes: St. canes––Ivan and Emily––which shattered the island George, St. John, St. Marks, St. Patrick, St. Andrew, in 2004 and 2005 respectively and severely damaged and St. David—with St. George parish being the the agricultural sector and housing stock. most densely populated and St. Mark the least popu- lated. While Grenada is characterized by lush tropi- Acute gastroenteritis (AGE) and diarrhoea are com- cal vegetation, the much smaller islands of Carria- mon clinical outcomes of foodborne illness (1-3), cou (area: 13 square miles, population~5,000) and which pose significant public-health problems, both in developed and developing countries (4-12). Correspondence and reprint requests: Dr. Martin S. Forde The World Health Organization (WHO) has found Department of Public Health and Preventive that diarrhoeal illness cause an estimated 1.87 mil- Medicine lion deaths globally (95% CI 1.56-2.19), account- St. George’s University ing for approximately 19% of deaths in children PO Box 7 aged <5 years (13). Diarrhoeal illnesses also have St. George a huge negative impact on the tourism and agri- Grenada Email: [email protected] cultural sectors (6,12,14,15). Studies indicate that, Burden of acute gastrointestinal illness in Grenada Glasgow LM et al. in developing countries, children aged <5 years are collaborative research funding partnership of the at the greatest risk of AGE-related illness (13,16). Canadian Institutes of Health Research, the Cana- Data from the Caribbean Epidemiology Centre dian International Development Agency, Health (CAREC) indicate that AGE/diarrhoeal illnesses in Canada, the International Development Research the Caribbean region have increased from 60,574 Centre, the Public Health Agency of Canada, and cases in 2005 to 105,688 in 2006 (17). The ‘true’ the Public Health Agency of Canada (PHAC). burden of AGE in the region, however, remains un- MATERIALS AND METHODS known. Burden of Illness (BOI) studies have been proposed to estimate the community prevalence of The Grenada BOI study consisted of two core com- diarrhoeal illnesses relating to food consumption, ponents: a population survey and a laboratory sur- identify the sources of disease-causing organisms, vey. The WHO definition for diarrhoea was used and develop evidence-based policies for the control in defining an AGE case as being someone who and prevention of foodborne diseases (12). had an acute (sudden) onset of diarrhoea, with or without fever (>38 °C or 100.4 °F), presenting with In all three islands––Grenada, Carriacou, and Petite three or more loose or watery stools within the past Martinique––the Ministry of Health (MOH) has the 24 hours, with or without vomiting, and/or visible responsibility for monitoring AGE-related illness. blood in the stool (19). Healthcare services are provided through three hospitals, six health centres, and 30 health stations. Ethical approval Several private health clinics also operate in Gre- nada and Carriacou. The Pathology Laboratory at The St. George’s University Institutional Review Board (IRB) and the MOH in Grenada granted ethi- the General Hospital in Grenada is the main pro- cal approval for the study (reference: 08010). Names vider of laboratory services, although limited labo- of respondents were not included in the question- ratory service is also provided by private facilities. naire. Respondents were informed of the purpose The MOH conducts surveillance through collect- of the survey, and written consent was requested ing weekly reports of syndromic AGE from public before the survey was administered. healthcare facilities and aetiologic laboratory re- ports from the Pathology Laboratory. Limited data Population survey are also collected from private health facilities. A retrospective cross-sectional population survey The Grenada National Strategic Plan for Health, was administered in two phases. The first survey 2007-2011 indicates that diarrhoeal illness is a ma- was conducted during the 44-46 weeks in October jor cause of morbidity in Grenada, with the highest 2008 to represent the low-AGE season, and the sec- rate of cases among children aged <5 years. In 2006, ond survey was conducted during the 14-16 weeks the MOH syndromic data showed 1,094 cases of in April 2009 to represent the high-AGE season. The AGE among children aged <5 years and 958 cases low- and high-AGE seasons for Grenada were based among persons aged >5 years (18). There are, how- on the trends of syndromic AGE over the period ever, limited data on the aetiology of AGE and the 2001-2003 and 2006-2007 (Figure 1). Trained in- true burden of AGE-related illness in this island. terviewers administered the surveys via face-to-face interviews during evening hours and on weekends. In 2007, St. George’s University and the Ministry of Health in Grenada collaborated with CAREC and Using the 2006 estimated population of 108,000, the Pan American Health Organization (PAHO) a sample-size of 1,057 was calculated using Epi to conduct a BOI study in Grenada as part of the Info (version 6.0, Centers for Disease Control and Caribbean Burden of Illness (BOI) Study. The goals Prevention, USA), with a 95% confidence interval, of the Grenada BOI study were to estimate the 50% prevalence, and 3% allowable error. An overall prevalence and distribution of AGE, quantify the sample-size of 1,300 was used for this study, with degree of underreporting of AGE in the population, 650 individual surveys conducted in each phase. identify key pathogens that cause AGE, estimate Two hundred and sixty-two Enumeration Districts economic impact (burden) of AGE, and provide a (EDs) were stratified by parish first, and the 1,300 better understanding of what the gaps in the sur- surveys were proportionately distributed based on veillance system were. This was the first time that the number of households in each ED. A random such a study was done in Grenada. Technical sup- numbers generator in Microsoft Excel (Microsoft port and funding for this study were also provided Corporation, Redmond, WA) was used for generat- by Teasdale-Corti grant programme of the Cana- ing household numbers for sampling in each ED. dian Global Health Research Initiative (GHRI), a Each parish, along with Carriacou, was designated 18 JHPN Burden of acute gastrointestinal illness in Grenada Glasgow LM et al. Figure 1. Total number of suspected cases of AGE reported to the National Surveillance Unit in Grenada for the period, 2001-2003 and 2006-2007* 300 s 250 e s cad 200 ted orte 150 5-year totals ecep 100 pr us 50 S 0 1 5 9 13 17 21 25 29 33 37 41 45 49 Epidemiological week *Data for 2004 and 2005 show sharp rise in the number of cases following Hurricanes Ivan and Emily respectively; therefore, data are not consistent with trends of the other years as a health region. Petite Martinique was included Additional questions were included pertaining to as an ED of Carriacou. The routes for survey admin- sociodemographic and economic factors, second- istration were marked on ED maps provided by the ary symptoms, healthcare-seeking behaviours, Central Statistics Office (CSO) in Grenada. A stan- specimen submissions, use of medications, impact dardized questionnaire developed by the Caribbean of illness, perceived cause of illness, and hygiene BOI steering team was slightly modified and used practices. Incentives, monetary or others, were not as the population survey instrument. A number provided for participation in this survey. was assigned to each questionnaire and was used Estimation of the burden and underreporting for identifying the respondent in the database. of AGE Persons who met the criteria (with the next birth- The burden and the level of underreporting syn- day falling before the day of the survey) were se- dromic AGE and laboratory-confirmed AGE/food- lected for interview in the sampled households. borne diseases were calculated by comparison of Written consent was required for the interview. The the syndromic and laboratory-confirmed AGE data neighbouring house was selected if any individual reported to the Ministry of Health for the period declined or was not available after 3 attempts to October 2008–September 2009 with that collected conduct the interview. If the selected individuals through the population and laboratory surveys. were aged <12 years, consent was required from For syndromic AGE, the BOI pyramid was defined their parents or guardians who answered the ques- using the percentage of self-reported cases who tions on their behalf. If the selected individuals sought medical care to estimate underreporting were between 12 and 17 years of age, consent was relative to syndromic AGE (Figure 2). For laborato- required from the parents or guardians, and the ry-confirmed AGE/foodborne pathogens, the BOI selected individual answered the survey questions. pyramid was defined using the percentage of AGE Persons who were aged >18 years were eligible to cases who sought medical care and submitted stool give consent and also answer the survey questions. samples, the percentage of stool samples tested, Individuals who were aged <1 year, unwilling or samples testing positive for foodborne pathogen, unable to participate in the survey, not physically and reported to the national surveillance. These present in the country at the time of survey, were proportions were then used in calculating multipli- <18 years old but did not receive parental consent, ers for each level, and these multipliers were then those who were prisoners, residents in hotels, guest- used in estimating the underreporting of laborato- houses and inns, and were mentally disabled, were ry-confirmed AGE (Figure 3). excluded from this study. Estimation of economic burden of AGE Respondents were asked if they had experienced diarrhoea in the past 4 weeks before the interview. Economic burdens of AGE estimates were calculat- Volume 31 | Number 4 (Suppl 1) | December 2013 19 Burden of acute gastrointestinal illness in Grenada Glasgow LM et al. Figure 2. Estimation of underreporting and the burden of syndromic acute gastrointestinal illness in Grenada, 2008-2009 Cases reported to the surveillance 1,669 Cases who sought medical care 100/31=3.23 (%) 31.0% Burden 3.23*1,669 5,390.87 The number of syndromic AGE cases reported to the Ministry of Health for October 2008−September 2009 was 1,669. Hence, the estimated burden is 1,669x3.23=5,390.77. This showed an underreporting factor of 3.3 (5,391/1,669). Hence, only 31% (1,669/5,360)x100) of the syndromic AGE cases were reported, and 69% were not reported Figure 3. Estimation of the burden and underreporting of laboratory-confirmed AGE Number of lab-confirmed AGI reported to the Ministry of Health (Ministry of Health surveillance data) 12 Proportion of laboratory-confirmed AGE cases reported to the 70.59 Ministry of Health (Ministry of Health surveillance data) Proportion of positive/lab-confirmed AGE samples (Laboratory data) 12.06 17/141=12.06%; Multiplier=100/12.06=8.29 95.92 Proportion of AGE samples tested at laboratory (Laboratory data) 100 Proportion of AGE samples submitted (Population survey) 12.5 Proportion of AGE samples requested (Population survey) Proportion of AGE cases seeking medical care (Population survey) 31 40/129=31.00%; Multiplier=100/31=3.23 Number of lab-confirmed AGE for the period is 17. Overall underreporting multiplier: 1.42x8.29x 1.04x1x8x3.23=316.35. There is, therefore, an underreporting factor of 316.35. For every lab- confirmed case of AGE reported to the Ministry of Health, there are 316 additional cases in the population. The estimated true burden of AGE was 12x316.35=3,796.2 cases 20 JHPN Burden of acute gastrointestinal illness in Grenada Glasgow LM et al. ed to assess the potential economic burden on pa- underrepresentation of individuals aged 5-14 years tients associated with accessing healthcare and the (survey 11.9%, population 23.0%) in this study. treatment for AGE by multiplying the estimated Magnitude and distribution of self-reported episodes in the population per year with the cost cases of AGE of treatment. A total of 129 persons (10.7%) (95% CI 9.0-12.6) Laboratory survey reported that they had sudden onset of diarrhoea The laboratory survey was conducted from Oc- with three or more watery or loose stools within tober 2008 to September 2009 at the Pathology 24-hour period with or without fever, dehydration, Laboratory which is the primary laboratory that vomiting, or visible blood in the stool within 4 weeks prior to the interview. They were, therefore, conducts microbial analysis for public and private classified as self-reported AGE cases. Using the cal- health facilities in Grenada. All stool samples were culated AGE prevalence of 10.7% and the formula transported to the Pathology Laboratory within 4 outlined in the Appendix, the rate per person-year hours after collection and processed for selected was calculated to be 1.4 episodes per person-per foodborne pathogens within 2 hours after receipt. year. A similar monthly prevalence was observed Of the stool samples submitted to the Pathology for the identified low- and high-AGE seasons (low- Laboratory from patients with a history of AGE, AGE season: 10.6%; high-AGE season: 10.8%). 96% were tested for Salmonella, Campylobacter, and Table 1 provides information on the sociodemo- Shigella, and 11% were tested for norovirus. Stool graphic characteristics of the self-reported cases in samples were usually processed within 2 hours after this study. receipt at the laboratory. Standard cultural identifi- cation methods were used for processing the stool The highest monthly prevalence of AGE was found samples for Salmonella, Campylobacter, and Shi- among children aged <5 years (23.5%), followed gella bacterial pathogens as are used in routine by those between 5 and 14 years of age (14.6%). diagnostic laboratories, using selective and enrich- The lowest reported prevalence of AGE was found ment culture techniques; a commercial ELISA kit among persons aged 45-64 years (5.7%) and ≥65 (IDEIA Norovirus Elisa Kit) was used for detection years (5.2%). AGE among children aged <5 years of norovirus. statistically differed from all other age-groups (p<0.001-0.003), except those between 5 and 14 Statistical analysis years (p=0.113). The difference in prevalence be- tween females (11.7%) and males (9.4%) was not Data were manually entered into a database created statistically significant. in EpiData (Lauritsen JM & Bruus M. EpiData, Ver- sion 3) and analyzed using Epi Info (version 6.0), By health region, the highest monthly prevalence using a 95% CI. Two-sample Independent t-tests of self-reported cases was found in St. John (17.4%), were conducted to analyze statistically significant which was significantly higher than the propor- differences in the results. tion of self-reported cases found in St. Patrick and St. Andrew (p=0.018 and p<0.001 respectively); the RESULTS lowest monthly prevalence was found in St. Patrick (4.3%), St. David (9.1%), and St. Andrew (9.8%). Response rate and characteristics of Table 2 provides information on the distribution of respondents self-reported cases of AGE in the health regions in A total of 1,300 surveys were attempted but 1,232 Grenada. (94.8%) individuals were successfully contacted The majority of self-reported AGE cases (66.0%) and invited to participate in the survey. Of these, attributed their illness to consumption of food or 25 surveys were rejected due to incompleteness drink. About half (54.3%) of the respondents re- or inclusion of ineligible persons, leaving a total ported that they washed hands either always or of 1,207 (92.8%) surveys which were used in the sometimes before eating meals, and 43.9% used analysis—594 in Phase 1 (October 2008) and 613 soap. Three-quarters (75.2%) of respondents self- in Phase 2 (April 2009). Comparison of the de- reported that they washed hands either sometimes mographic profiles of survey respondents with or always after using the toilet, and 61.4% self- residents in the general population indicated that reported that they used soap to wash hands after there was overrepresentation of individuals aged using toilet. 15-24 years (survey 27.5%, population 18.2%) and females (survey 55.9%, population 50.2%), and an While almost all respondents living on the main- Volume 31 | Number 4 (Suppl 1) | December 2013 21 Burden of acute gastrointestinal illness in Grenada Glasgow LM et al. Table 1. Sociodemographic characteristics of self-reported AGE cases (n=129) in Grenada Residents Respondents Number and per- 95% Confidence Variable (N) (n) centage of AGE cases interval Gender (n=1,207; p=0.20) Male 54,481 532 50 (9.4) 7.1-12.3 Female 52,151 675 79 (11.7) 9.4-14.4 Age (completed years) (n=1,207; p<0.001) 1-4 9,948 81 19 (23.5) 14.8-34.2 5-14 20,189 144 21 (14.6) 9.3-21.4 15-24 20,624 332 36 (10.8) 7.8-14.8 25-44 27,938 322 35 (10.9) 7.8-14.9 45-64 17,485 212 12 (5.7) 3.0-9.7 ≥65 6,448 116 6 (5.2) 1.9-10.9 Cultural group (n=1,207; p=0.86) African/Black NA 1,087 116 (10.7) 8.9-12.7 Indian NA 78 10 (12.8) 6.3-22.3 Asian NA 0 0 (0) 0.0-84.2 European NA 3 0 (0) 0.0-7.0 South American NA 3 0 (0) 0.0-70.8 North American NA 4 1 (25.0) 0.6-80.6 Monthly income (EC$) (n=900; p=0.18) Low income (0-1,000) NA 345 42 (12.2) 9.0-16.2 Medium income (1,001-2,000) NA 343 47 (13.7) 10.3-17.9 High income (>2,000) NA 212 18 (8.5) 5.1-13.1 Education of mother (n=1,206; p=0.57) Primary NA 414 38 (9.2) 6.7-12.5 Secondary NA 474 59 (12.4) 9.7-15.8 Certificate/Diploma NA 122 16 (13.1) 7.7-20.4 Undergraduate/Graduate NA 35 3 (8.6) 1.8-23.1 Postgraduate NA 14 0 (0) 0.0-23.2 Education of father (n=1,206; p=0.30) Primary NA 408 36 (8.8) 6.8-12.1 Secondary NA 330 46 (13.9) 10.5-18.3 Certificate/Diploma NA 85 13 (15.3) 8.4-24.7 Undergraduate/Graduate NA 25 2 (8.0) 1.0-26.0 Postgraduate NA 15 1 (6.7) 0.2-31.9 Health region (HR)/Parish HR 1/St. Andrew 26,435 285 28 (9.8) 6.9-13.9 HR 2/St. Patrick 10,461 138 6 (4.3) 1.6-9.2 HR 3/St. George 37,403 421 49 (11.6) 8.8-15.2 HR 4/St. Mark 4,346 49 7 (14.3) 5.9-27.2 HR 5/St. John 8,405 92 16 (17.4) 10.3-26.7 HR 6/St. David 12,859 143 13 (9.1) 4.9-15.0 Carriacou and Petite Martinique 5,633 79 10 (12.7) 6.2-22.0 NA=Not available 22 JHPN Burden of acute gastrointestinal illness in Grenada Glasgow LM et al. land Grenada use pipe-supplied chlorinated water Healthcare-seeking behaviours and as their main source of drinking-water, all respon- treatments dents from Health Region 7—Carriacou and Petite Of the 129 self-reported AGE cases, 40 (31%) re- Martinique—used harvested rainwater as their ported seeking medical care, of whom only 12.5% main source of drinking-water. Of those living in were asked by their physicians to submit a stool Carriacou and Petite Martinique, 58.2% treated sample. Antibiotics were prescribed for 10.0%; the water before drinking it. The most common treatment methods used were boiling (65.2%) and oral rehydration fluid, which is the WHO-recom- adding chlorine tablets/bleach (19.6%). There was mended therapy for persons with AGE, was pre- no significant association between drinking-water scribed for 65.0%. Of those who received prescrip- source and water treatment with being a case of tion for medication, 94.4% took the medication AGE (p=0.37). as prescribed by the physicians. Almost half of the self-reported cases (45.0%) reported using non– Symptoms and severity prescribed treatments for the diarrhoeal illness; of them, 19.1% used bush medicine, i.e. any part of a The median episodes of diarrhoea among self- plant used for medicinal purposes without a phy- reported AGE cases in 24 hours was 4 (minimum 3, sician’s prescription, and 23.6% used oral rehy- maximum 10), the number of days with AGE was dration fluid. Table 4 provides information about 3 (minimum 1, maximum 14), and days spent at healthcare-seeking behaviours of the self-reported home without doing routine activities due to the AGE cases. illness was 2 (minimum 1, maximum 14). Of the 129 cases, 83.7% said they experienced abdominal Laboratory diagnosis of selected foodborne pain, 35.7% experienced vomiting, 27.1% experi- pathogens in stool samples enced headache, and 29.5% experienced nausea. Table 3 provides information on symptoms expe- From 1 October 2009 to 30 September 2009, rienced by self-reported AGE cases. 147 stool samples were submitted to the Pa- Table 2. Distribution of self-reported cases of AGE by health region in Grenada Overall number and percentage of cases per Health region/parish 95% CI health region/parish Health Region 1: St. Andrew 28 (9.8) 6.9-13.9 Health Region 2: St. Patrick 6 (4.3) 1.6-9.2 Health Region 3: St. George 49 (11.6) 8.8-15.2 Health Region 4: St. Mark 7 (14.3) 5.9-27.2 Health Region 5: St. John 16 (17.4) 10.3-26.7 Health Region 6: St. David 13 (9.1) 4.9-15.0 Carriacou and Petite Martinique 10 (12.7) 6.2-22.0 Table 3. Secondary symptoms experienced by self-reported AGE cases (n=129) in Grenada Secondary symptom experienced by % 95% CI self-reported cases of AGE in Grenada Headache 27.1 19.7-35.7 Fever (measured) 16.7 15.0-31.2 Fever (not measured) 23.2 10.6-24.3 Nausea 29.5 21.8-38.1 Vomiting 35.7 27.4-44.6 Runny nose 7.0 3.2-12.8 Abdominal pain 83.7 76.2-89.6 Bloody diarrhoea 3.1 0.9-7.7 Cough 18.6 12.3-26.4 Sneezing 10.1 5.5-16.6 Sore throat 11.6 6.7-18.5 Volume 31 | Number 4 (Suppl 1) | December 2013 23 Burden of acute gastrointestinal illness in Grenada Glasgow LM et al. thology Laboratory from patients with history defined to calculate the burden of illness for syn- of diarrhoeal illness, of which 141 (96%) were dromic AGE and laboratory-confirmed AGE (for analyzed for foodborne pathogens–Salmonella, foodborne pathogens) in Grenada. Table 5 con- Shigella, Campylobacter, and norovirus. Overall, tains information required to calculate the burden 12.6% tested positive for a foodborne patho- of syndromic AGE and laboratory-confirmed AGE/ gen. Salmonella (9.2%) and norovirus (1.4%) foodborne pathogens. were the most commonly-identified patho- The estimated burden of syndromic AGE in Gre- gens. nada for the one year period (September 2008– Estimation of underreporting of AGE to the October 2009) was calculated as follows: MOH Surveillance Unit Number of syndromic AGE cases reported to the Figure 2 and 3 show the surveillance pyramids Ministry of Health for the specified period (1,669)x Table 4. Healthcare-seeking practices among self-reported cases of AGE in Grenada Healthcare-seeking behaviour % 95% CI Sought medical care (n=129) 31.0 23.2-39.7 Was asked to submit specimen (n=40) 12.5 4.2-26.8 Submitted specimen upon request (n=5) 100 100.0-100.0 Took antibiotics (prescribed) (n=40) 10 8.0-12.0 Took oral rehydration fluid (prescribed) (n=32) 65.6 46.8-81.4 Took non-prescribed medications (n=129) 45.0 36.2-54.0 Table 5. Information for calculation of the burden of acute gastrointestinal illness in Grenada, October 2008–September 2009 Information Source of data Formula and value Number of syndromic AGE reported to Ministry of Health 1,669 the Ministry of Health for the period Surveillance Unit October 2008–September 2009 Proportion of laboratory-confirmed Ministry of Health Laboratory-confirmed/positive AGE AGE cases reported to the Ministry of Surveillance Unit reported to the Ministry of Health= Health for the period October 2008– 80.0%. Total laboratory-confirmed September 2009 AGE reported to the Ministry of Health =70.59% (12/17) Number of lab-confirmed AGE cases ac- Laboratory survey Actual number of samples positive for tually isolated at the lab for the period AGE=17 October 2008–September 2009 Proportion of positive/laboratory- Laboratory survey Number of samples positive for AGE/ confirmed AGE (of AGE samples number of AGE samples tested=12.06% tested, what proportion was positive) (17/141) Proportion of AGE samples tested at Laboratory survey Number of samples tested for AGE/ laboratory (how often do the lab tests number of samples received by lab for a pathogen (at least 1 pathogen) =96% (141/147) Proportion of physicians requesting AGE Population survey Number of samples reported submitted/ samples submitted to the laboratory number of samples requested=100% Proportion of AGE samples requested Population survey Number of samples requested/number (of cases who sought medical care, pro- of cases seeking medical care=12.5% portion requested to submit sample) Proportion of AGE cases seeking medi- Population survey Number of AGE cases seeking medical cal care care/number of AGE cases=31% Number of AGE cases in population Population survey Number of ill persons meeting AGE case survey (meeting AGE case definition) definition=129 24 JHPN Burden of acute gastrointestinal illness in Grenada Glasgow LM et al. underreporting multiplier for seeking medical care countries, such as Canada (1.3 episodes per person- (100/31 or 3.2)=5,390.80. This showed an underre- year for the period February 2001 to February 2002) porting factor of 3.3 (5,391/1,669). Hence, only 31% (20), Ireland (0.60 episodes per person-year for the (1,669/5,360)*100) of the syndromic AGE cases were period December 2000 to November 2001) (4), and reported, and 69% were not reported (Figure 2). Australia (0.90 episodes per person-year for the period September 2001 to August 2002) (21). The The burden of laboratory-confirmed AGE/food- prevalence of AGE in Grenada, however, is com- borne pathogens was calculated as follows: parable with that found in Cuba (Grenada 10.7%, Cuba 10.6%) (22). A total of 129 AGE cases were identified among the 1,207 respondents who took part in the population In Grenada, the prevalence of AGE among children survey. Of them, 40 (31%) sought medical care. From aged <5 years was significantly higher than among these 40 presenting cases, 5 (12.5%) were asked to other age-groups. This finding is consistent with submit stool samples, and all 5 complied. From the that from other studies and the WHO statements laboratory survey, a total of 147 stool samples were of a global pattern that establishes AGE as high- received, of which 141 (95.9%) were analyzed. Of est among children aged <5 years. The Ministry of these 141 analyzed samples, 17 (12.1%) contained Health (MOH) in Grenada has listed AGE-related an identified AGE pathogen but only 12 (70.6%) illness as a common cause of morbidity in chil- of these were reported to the Ministry of Health. dren on the island. Younger children are at high- Therefore, the overall underreporting multiplier fac- er risk for AGE as an outcome of their tendencies tor for laboratory-confirmed AGE pathogens was to practise poor hygiene. Results from the Global 1.42x8.29x1.04x1x8x3.23=316.35. Thus, of the 12 School Health Survey conducted in Grenada in the laboratory-confirmed AGE cases that were reported first half of 2008 showed that 11.9% of students to the MOH over the same period that the popula- in secondary schools never or rarely washed their tion survey was conducted, the estimated true bur- hands before eating; 3.7% also indicated that they den of AGE was 12x316.35=3,796.2 cases (Figure 3). never or rarely washed their hands after using a toilet or latrine (23). While these percentages rep- Socioeconomic costs resent a small number of children, disease-causing organisms can be transferred to a larger number The estimated costs of seeking and receiving medical of children and result in higher incidence of re- care for treating an AGE case in Grenada ranged from lated illnesses. This study also found that the level US$ 15 to 52, depending on the type of transporta- of handwashing before meals and after using the tion, physician (public or private), laboratory service toilet, particularly with soap, was not very high, used (public or private), and medications prescribed. which inadvertently contributes to the spread of Using this range of costs, the annual economic bur- disease-causing organisms. den on patients associated with the treatment for AGE was estimated to range from US$ 703,950 to In this study, a similar monthly prevalence for the 2,440,360. It should be noted that an estimate of the low- and high-AGE seasons (10.6% and 10.8% individual burden due to the lost quality of life is respectively) was found, which might have been not included in the costs of socioeconomic burden. indicative of outbreaks occurring during the low Additionally, the AGE cases who were unable to per- season. Syndromic data for the period 2001-2003 form routine activities lost, on an average, between and 2005-2007 also showed sporadic AGE peaks US$ 11 and 30 (minimum wage range/day) for each occurring during the low season. These outbreaks, day that they were unable to attend work. Using the which usually occur in isolated locations, should calculated median of 3 days without performing be carefully monitored by the MOH, and necessary routine activities found in this study, the productiv- controls and preventive responses should be imple- ity losses potentially ranged from US$ 33 to 90 per mented. person due to AGE-related illness. Harvested rainwater is the main source of drink- DISCUSSION ing-water in Carriacou and Petite Martinique, and the dry season (high-AGE season) has been noted This study indicates that diarrhoeal illness is com- for increased diarrhoea and vomiting, particularly mon in Grenada, with approximately 11% of the among children on this small island. The outbreaks population being affected and an incidence rate of are assumed to be associated with water being 1.4 episodes per person per year. This incidence rate drawn from low-level cisterns, which may contain was higher than that found in several developed higher concentrations of microbial contaminants. Volume 31 | Number 4 (Suppl 1) | December 2013 25 Burden of acute gastrointestinal illness in Grenada Glasgow LM et al. The highest reported rates of AGE in Carriacou and Great River passes through. The Great River is used Petite Martinique typically corresponded with the both as a dump as well as the main source of wa- harshest and the longest dry seasons. In this study, ter for domestic and hygienic purposes for several however, a lower proportion of self-reported AGE households along the riverbanks. The town is also cases were found in Carriacou and Petite Marti- characterized by a large number of small, make- nique during the dry season than in the wet season. shift food outlets. Many of these outlets on the This finding requires further investigation to deter- main streets are owned and operated by persons mine how interactions between environmental without food vendors’ license, which increases the and other factors, such as the condition and treat- likelihood of improper food handling. While many ment protocols employed in harvested rainwater self-reported AGE cases attributed their illness to systems, contribute to or reduce the incidence of consumption of food and drink (66.0%), the ques- AGE among residents on the islands. tionnaire did not ask respondents if the food they consumed was prepared domestically or outside the While this study indicates that AGE is causing sig- home. Further research may be useful to compare nificant morbidity in Grenada, the use of home- between the risk associated with consumption of made remedies to treat this illness may have likely foods prepared at home and those purchased from contributed to the high levels of underreporting outside the home. observed, resulting in further limitations on aeti- ology, trends, and impact of AGE at the national Limitations level. While the MOH has determined high- and There were several limitations in this study. There low-AGE seasons in Grenada, this is the first study was an overrepresentation of females and persons undertaken to obtain evidence of the likely burden aged 15-24 years and underrepresentation of per- of the disease in the Tri-island State. An under- sons aged 5-14 years in this study. These biases may reporting factor of 316 for laboratory-confirmed be due to practices of convenient sampling by sur- foodborne AGE pathogens indicated that, for every veyors. To minimize this from happening, survey- case currently recorded, there were 316 additional ors were re-trained before conducting the second cases occurring but not reported to the MOH Sur- phase of the survey. Given that the respondents veillance Unit. This high rate of underreporting were asked to recollect if they had any diarrhoeal highlights an urgent need to address the potential- illness that matched our AGE definition over the ly huge burden of AGE in Grenada and the need for past four weeks, it is unlikely that many would have adjustments in medical practice and improvements forgotten this for the severity of typical AGE symp- in the surveillance system. Other notable findings toms. Also, as is expected in retrospective studies, from this study were the very low number of stool the potential of recall bias cannot be ruled out. Re- specimens requested by physicians to confirm AGE call bias is a limitation of retrospective studies. Ret- (12.5%) and the incomplete reporting of AGE cas- rospective methods may be subject to more recall es by the laboratory to the MOH (only 70.59% of bias than prospective studies. Nevertheless, similar confirmed AGE cases were reported to the MOH). methods to other retrospective studies were used, Furthermore, while there was a general increase in thereby enabling comparisons between studies. the number of samples submitted during the high- AGE season, non-adherence to sampling protocol Conclusions and recommendations by healthcare workers (for example, not storing samples at 4 °C during transportation and deliv- Diarrhoeal illness is common in Grenada, with ap- ering samples 4 hours after collection) may have proximately 11% of the population being affected. contributed to the proliferation of other organisms Salmonella enteritidis is the most common patho- in the samples, besides the organism responsible gen associated with the illness, which also has ma- for causing AGE, thereby affecting isolation in the jor economic impact. Underreporting is high, and laboratory. there is a need to improve the surveillance system to monitor and control AGE in Grenada. The literature shows that AGE is closely associated with conditions that exist as a result of poverty. Ad- Some recommendations which may be useful for ditionally, urbanization has been identified as a key monitoring and controlling AGE in Grenada are: (i) factor that may also contribute to the condition. introducing public education programmes which Gouyave town of St. John parish is characterized by promote and encourage proper hygiene practices; crowded living conditions, particularly in the cen- (ii) implementing more robust surveillance systems tre of the town and along which a river called the of street-based food vendors; and (iii) strengthen- 26 JHPN