Objective To assess the economic cost of routine Aedes aegypti control in an at-risk environment without dengue endemicity and the incremental costs incurred during a sporadic outbreak.
Methods The study was conducted in 2006 in the city of Guantanamo, Cuba. We took a societal perspective to calculate costs in months without dengue transmission (January–July) and during an outbreak (August–December). Data sources were bookkeeping records, direct observations and interviews.
Results The total economic cost per inhabitant (p.i.) per month. (p.m.) increased from 2.76 USD in months without dengue transmission to 6.05 USD during an outbreak. In months without transmission, the routine Aedes control programme cost 1.67 USD p.i. p.m. Incremental costs during the outbreak were mainly incurred by the population and the primary/secondary level of the healthcare system, hardly by the vector control programme (1.64, 1.44 and 0.21 UDS increment p.i. p.m., respectively). The total cost for managing a hospitalized suspected dengue case was 296.60 USD (62.0% direct medical, 9.0% direct non-medical and 29.0% indirect costs). In both periods, the main cost drivers for the Aedes control programme, the healthcare system and the community were the value of personnel and volunteer time or productivity losses.
Conclusions Intensive efforts to keep A. aegypti infestation low entail important economic costs for society. When a dengue outbreak does occur eventually, costs increase sharply. In-depth studies should assess which mix of activities and actors could maximize the effectiveness and cost-effectiveness of routine Aedes control and dengue prevention.
Objectif: Evaluer le coût économique du contrôle de routine de l’Aedes aegypti dans un environnement à risque sans endémicité pour la dengue et les coûts marginaux encourus lors d’une épidémie sporadique.
Méthodes: L’étude a été menée en 2006 dans la ville de Guantanamo, à Cuba. Nous avons adopté une perspective sociétale pour calculer les coûts dans les mois sans transmission de la dengue (janvier à juillet) et lors d’une épidémie (août à décembre). Les sources des données étaient les registres, des observations directes et des interviews.
Résultats: Le coût financier total par habitant (ph) par mois (pm) a augmenté de 2,76 USD au cours des mois sans transmission de la dengue à 6,05 USD au cours d’une épidémie. Au cours des mois sans transmission, le programme de contrôle de routine de l’Aedes coûtait 1,67 USD ph.pm. Les coûts marginaux lors de l’épidémie ont été engagés principalement par la population et le secteur primaire/secondaire du système de soins de santé, et à peine par le programme de lutte antivectorielle (incréments de 1,64; 1,44 et 0,21 UDS ph.pm, respectivement). Le coût total pour la prise en charge d’un cas de dengue hospitalisé a été estiméà 296,60 USD (62,0% en coûts directs médicaux, 9,0% en coûts directs non médicaux et 29,0% en coûts indirects). Dans les deux périodes, les principaux facteurs déterminant les coûts pour le programme de contrôle du genre Aedes, le système de soins de santé et la communautéétaient la valeur du personnel et le temps des bénévoles ou les pertes en productivité.
Conclusions: Des efforts intensifs pour maintenir basse l’infestation par Aedes aegypti entraînent d’importants coûts économiques pour la société. Lorsqu’une épidémie de dengue arrive éventuellement, les coûts augmentent fortement. Des études approfondies doivent évaluer quels types de combinaisons d’activités et d’acteurs peuvent maximiser l’efficacité et la rentabilité du contrôle de routine de l’Aedes et la prévention de la dengue.
Objetivo: Evaluar el coste económico del control rutinario de Aedes aegypti en un ambiente de riesgo sin dengue endémico y los costes incrementales incurridos durante un brote esporádico.
Métodos: El estudio se realizó en el 2006 en la ciudad de Guantanamo, Cuba. Se utilizó una perspectiva social para calcular los costes en meses sin transmisión del dengue (Enero a Julio) y duriante el brote (Agosto-Diciembre). Las fuentes de datos fueron libros de registro, la observación directa y entrevistas.
Resultados: El coste económico total por habitante (p.h.) por mes (p.m.) aumentó de 2.76 USD en meses sin transmisión del dengue al 6.05 USD durante un brote. En meses sin transmisión el programa de control rutinario de Aedes cuesta 1.67 USD p.i. p.m. Los costes incrementales durante el brote eran principalmente para la población y para los niveles primario/secundario del sistema sanitario, casi nunca para el programa de control vectorial (1.64, 1.44 y 0.21 UDS aumento p.h. p.m., respectivamente). El coste total de manejar un caso hospitalizado con sospecha de dengue era de 296.60 USD (62.0% costes médicos directos, 9.0% costes directos no médicos y 29.0% costes indirectos). En ambos periodos, los principales factores de coste para el programa de control de Aedes, para el sistema de cuidados sanitarios y para la comunidad eran el coste del personal y el tiempo de voluntarios, o pérdidas de productividad.
Conclusiones: Los intensos esfuerzos para mantener bajos los niveles de infestación de Aedes aegypti conllevan a importantes costes económicos para la sociedad. Cuando eventualmente ocurre un brote de dengue, los costes aumentan considerablemente. Estudios en profundidad deberían evaluar cual mezcla de actividades y actores podría maximizar la efectividad y la costo-efectividad del control rutinario de Aedes y de la prevención del dengue.
Dengue, globally the main arbovirosis, has become a predominant international public health problem over the last years (Guzman 2011); 3.61 billion people are at risk of becoming infected in 124 countries worldwide and every year 500 million of them effectively contract infection (Beatty et al. 2010). This results in 50 million clinical dengue fever cases (http://www.denguevaccines.org/disease-burden), in 2.1 million cases of dengue haemorrhagic ever – dengue shock syndrome cases in need of hospitalization – and 21 000 deaths each year (Beatty et al. 2010). Aedes aegypti, a mosquito that breads in water-filled containers in the (peri-) domestic environment, is the main dengue vector (Guzman & Kouri 2003; Scott & Morrison 2010). In the absence of a vaccine, the only strategy to prevent transmission and disease is vector control: eliminating larval habitats, reducing adult mosquito abundance and life span and preventing mosquito-human contact (Eisen et al. 2009). For this daunting and costly task, combinations of indoor and outdoor application of chemical and biological larvicides and of insecticides are used, often complemented by environmental management (TDR 2006; WHO 2009).
In spite of the worldwide growth of the burden of dengue, the economic impact of the disease has been poorly documented (TDR 2006). This impact includes the burden of disease (Torrance 1987; Murray 1994) and the cost of illness (Suaya et al. 2006). The latter comprises health service cost, both for prevention and curative care, and household cost related to disease and productivity losses (Flessa 2009). Published studies on the subject mainly deal with the burden of disease and costs of illness during epidemics in particular settings (Beatty et al. 2008). The worldwide cost of case management for health systems and families was estimated to reach 1.8 thousand million USD yearly, excluding the expenses for epidemiologic surveillance and vector control (Suaya et al. 2009). Figures on the costs related to routine vector control are even scarcer. The few papers published report, each for a particular mix of activities, yearly cost per capita ranging from 0.20 USD in Cambodia, to 32.0 USD in Cuba (Suaya et al. 2007; Armien et al. 2008; Baly et al. 2009, 2011; Tun-Lin et al. 2009).
To our knowledge, no study has integrated information on the economic cost of routine dengue control in non-epidemic periods with figures on incremental costs for vector control and case management during outbreaks. We conducted research in Guantanamo Cuba, to compare economic cost of dengue outbreak with expenses during non-epidemic period.
The city of Guantánamo in east Cuba has 244 100 inhabitants and 68 648 households (Oficina Nacional de Estadisticas 2007). Administratively, it is divided into nine people’s councils and 151 circumscriptions, the lowest level of local government. There are five health areas, each with a polyclinic staffed by the basic medical specialties and with a network of family practices with GPs and nurses, and at provincial level, there is general and paediatric hospital. The average temperature is 32 °C, and rainfall (on average 610 mm/year) is concentrated in two short wet seasons from April to July and September to October. Infestation with A. aegypti is linked to a deficient water supply and lack of adequate environmental management, but the average House index (H.I.) at municipal level remained below 0.6 throughout 2003–2005. From 2006 onwards, after promising results obtained in pilot projects in Santiago de Cuba (Toledo et al. 2006; Toledo Romani et al. 2007) and Guantanamo itself (Vanlerberghe et al. 2009), community-based environmental management, intertwined with the routine vector control programme, had been gradually implemented throughout Guantanamo (Toledo et al. 2009). Nevertheless, dengue transmission was detected in the last week of July 2006 (PAHO 2006). By February 2007, the outbreak had been controlled, and no further cases were picked up by the active surveillance system.
Activities of the different social actors involved in dengue control
The Cuban government accords the highest priority to Aedes control and dengue prevention and assigns the necessary resources to conduct a wide range of activities that are intensified during epidemics that are coordinated by governmental intersectoral bodies.
The A. aegypti control programme is centrally regulated by the MOH but operationally decentralized to the municipal and health area level. Dedicated vector control workers perform the main programme activities: entomological surveillance, source reduction through periodic inspection of houses, larviciding with temephos in water-storage containers, selective perifocal insecticide spraying against adult mosquitoes, health education and the enforcement of mosquito-control legislation through the use of fines. During dengue outbreaks, house inspection cycles are shortened, intra-domiciliary adulticiding (every 7 days or less) with transmission is introduced in all houses in neighbourhoods and specialized teams from provincial and national level conduct spatial spraying and reinforce quality control activities.
The polyclinics and family practices are the entry point for routine, passive, clinical-epidemiological surveillance of fever of unknown origin. When dengue transmission is suspected, active population screening for fever cases is organized with family doctors, nurses, non-health personnel from different government agencies and community organizations. All households are visited daily, and persons with fever are referred to the polyclinics and assessed. If dengue fever is suspected, they are admitted to temporary, dedicated wards in a hospital and treated following a standardized protocol. On the 6th day after symptom onset, anti-dengue IgM is determined at the Provincial Centre for Hygiene and Epidemiology of Guantanamo. Positive samples are reassessed in the laboratories of the Institute of Tropical Medicine ‘Pedro Kouri’ for confirmation as long as an outbreak has not been declared.
Various grassroot-level organizations participate in Aedes control. Additionally, in Guantanamo, Community Working Groups were created at circumscription level. By actively involving the community and securing intersectoral support links (Vanlerberghe et al. 2009), they became a driving force for the community-based environmental management strategy encompassing, among others, actions to change behaviour (cover water-storage containers and protect/destroy artificial containers), reduce intra or peridomiciliary environmental risks (evacuation of waste water) and promote grassroot-level surveillance of Aedes infestation.
The information on costs incurred in 2006 by the social actors involved in dengue control was collected using the micro costing method (Weinstein et al. 1996) except for the hospital, where we used macro costing to derive the inpatient cost per day for the wards managing dengue cases. Costs were classified according to Johns et al. (2003), first by actor/activity and subsequently as recurrent and capital costs.
For the vector control programme and the primary healthcare services, the recurrent costs included salaries, supplies and materials (larvicides and insecticides, diagnostic tests, drugs, protective clothing, gloves, office materials), operational costs (fuels and lubricants, vehicle rent, per diems and food, spare parts and maintenance of equipment, vehicles and buildings) and utilities (electricity, water, telephone). For capital means (portable fogging equipment, trucks for spatial spraying, other vehicles, laboratory equipment, furniture), we recorded their time of use for dengue control. Buildings were not included because of lack of information on replacement cost. The information on quantities consumed was extracted on a monthly basis from the bookkeeping. This was complemented in January–July (no dengue transmission) and in August–December by semi-structured interviews with health system managers and a random sample of 100 family doctors, nurses and vector control personnel.
For the community organizations and households, we did not consider capital cost. We measured the time that these actors used for Aedes control (promotion, environmental management) and for active surveillance of fever cases from activity records kept by the community participation project referred to above (Toledo et al. 2009). We also conducted, in January–July and in August–December, semi-structured interviews with 200 randomly selected household members and 100 community leaders. The same interviews permitted to estimate out-of-pocket expenditures related to supplies and materials for vector control (purchase of water-storage tanks, lids for containers).
Estimates of out-of-pocket expenditure and productivity losses linked to dengue illness were obtained through semi-structured interviews with one-third of the adult cases discharged between September and October 2006. We did not collect information on expenditure incurred by fever cases that were not clinically suspected of dengue.
Entomological indicators (monthly number of A. aegypti positive houses) and information on output of the Aedes control programme (houses inspected, houses sprayed) and the healthcare services′ output (number of febrile cases detected, laboratory test performed, patients hospitalized) were compiled from routine reports and registers.
Hospital indicators (number of admissions because of dengue, number of discharges, average length of stay, number of diagnostic tests performed) were compiled from the records of accountancy department of the general provincial hospital. We collected from the records of the Provincial Centre for Hygiene and Epidemiology information on the number of intersectoral activities during the outbreak. Unfortunately, detailed costs of these were unavailable.
All economic costs were analysed for two periods of the year, January–July (non-epidemic period) and August–December (outbreak period), taking a societal perspective (Gold 1996). We calculated for the two periods the average total cost p.m. for: each actor/activity, for the community and the health system and the corresponding costs p.i. p.m., as well as their difference and ratio.
The mean hospitalization cost per case was calculated by multiplying the average length of stay in each of the two provincial hospitals with the average cost per dengue suspect inpatient day in the general hospital (i.e. assuming an equivalent cost per patient discharged for the paediatric hospital). Economic productivity losses of cases were estimated by multiplying the average length of stay by the minimum government pension rate. The time invested by caregivers, as well as the time used in voluntary community work for dengue control, was valued at the same rate. We did not value time of children.
All costs were analysed at constant prices and converted at the 2006 official exchange rate of 1 peso = 0.92 USD. We used the actual gross salary costs to deduct the value of the personnel time devoted to dengue control activities. Supplies and materials were costed at local market prices. We valued the annual depreciation of the capital means by the linear method (Haddix & Teuctsch 2003), based on an interest rate of 6%, normal length of life and market price replacement cost. The depreciation share was allocated proportional to the time of use for dengue control.
The study protocol was approved by the Ethical Committee of the Institute of Tropical Medicine ‘Pedro Kouri’ and by the Provincial Health Authorities of Guantanamo. We did not access data from individual patients. All people interviewed provided informed consent.
In the months January–July of the years 2003–2005 on average, 395 houses were positive for A. aegypti (H.I. 0.3%); in the months August–December of those years, the number rose to 607 on average (H.I. 0.5%) (Figure 1). In 2006, this trend was inversed. In January–July, an average of 490 positive houses was detected p.m. (H.I. 0.4%), against 216 (H.I. 0.1%) in August–December. Dengue suspected cases started appearing in July 2006. In that month, 29 cases were reported and the outbreak reached a peak of 1433 cases in September 2006. It was controlled by February 2007.
The major resources used during the whole year, their absolute quantities and their percentage increase during outbreak control are reported in Table 1. The number of full time equivalent (FTE) personnel used by the A. aegypti control programme was stable over both periods, but their activities shifted drastically. The use of resources and the number of provided services increased sharply, particularly house and spatial spraying (+106.6% and +184.5%, respectively). At the primary healthcare level, the FTE personnel used in dengue control activities increased dramatically from 134.2 to 2437.2 p.m., mainly for the search for and the screening of fever cases. The average number of hospital beds made available during the epidemic was 161; the occupancy rate was 85% and 72.8 FTE doctors and nurses were mobilized p.m. to take care of a total of 3549 hospitalized suspected dengue cases, of whom 2937 were eventually confirmed. The average hospitalization stay was 5.8 days (range 2–15).
Table 1. Quantities (range) of resources used and services provided for dengue control, by actor, in periods without (January–July) and with (August–December) transmission. Guantánamo, Cuba. 2006
*Full Time Equivalent (FTE) in person month or equipment month (8 h/day).
Volunteers – search for febrile cases activities [FTE]
Inhabitants – activities of cooperation with the vector control workers [FTE]
Vector control programme
Quality control (inspection of houses and public places)
Gasoline and fuel oil [l]
26 412.6 (19 471–29 618)
81 604.1 (25 558–100 797)
Portable spraying equipment [FTE]
Spatial (heavy) spraying equipment [FTE]
Inspected houses [number]
122 196.1 (114 734–125 869)
140 637.8 (122 930–196 101)
Sprayed houses [number]
80 863.9 (58 624–111 190)
167 081.2 (53 687–219 190)
Space spraying with heavy equipment or airplane[number]
Primary health care
Dedicated to active febrile cases search (including persons from other government sectors) [FTE]
Febrile cases detected [number]
Processed blood samples (IgM) [number]
Doctors and nurses [FTE]
Beds made available [number]
Hospitalizations on suspicion of dengue [number]
Diagnostic tests [number]
10 400.2 (3111–21 340)
Between July and December, the community-based organizations and households more than doubled the volunteer time on vector control activities, environmental sanitation and the active search for febrile cases. Further intersectoral activities included collection of 4 880 181 m3 of solid waste, broadcasting 41 mass media programmes, publishing six articles on dengue in local print media and organizing more than 67 000 meetings in work places.
Compared to the January to July 2006 period, the average total cost per inhabitant (p.i.) p.m. for dengue control more than doubled during the August to December outbreak: from 2.76 USD to 6.05 (Table 2). In absolute terms, the average monthly cost increased with 803 658.0 USD. This resulted in a total cost of 48.57 USD p.i. for the year 2006. Cuba’s yearly GDP was 4699.0 USD per capita (Oficina Nacional de Estadisticas 2009). The dengue control cost p.m. in Guantanamo amounted to 0.7% of the country’s monthly GDP in the period without transmission and to 1.5% in the period with transmission.
Table 2. Cost [USD]* for dengue control, by actors and differences and ratios between periods without (January–July) and with (August–December) transmission. Guantanamo, Cuba, 2006
Difference between periods p.i.†
Ratio between periods
Absolute average p.m.
Average p.i.† p.m. (%)
Absolute average p.m.
Average p. i.† p.m. (%)
*All numbers are rounded.
Grass root organizations
Promotion and communication
Search for febrile cases
Cooperation with vector control workers
Out-of-pocket expenses for vector control
Out-of-pocket expenses for illness
Productivity losses due to illness
Vector control programme
Primary health care
Others (all other costs)
Hospitalization of suspected cases of Dengue
1 477 617
The costs for the A. aegypti control programme increased only modestly: from 1.67 USD to 1.88 USD p.i. p.m., or in absolute terms from 408 281.80 USD to 459 406.0 p.m. The increase is almost exclusively attributable to the intensive use of insecticides against adult Aedes stages. The picture is quite different for the other health system actors. The primary healthcare cost for dengue control almost rose by a factor of 5 from 0.25 USD to 1.60 p.i. p.m. (an absolute increase of 222 687.80 USD p.m.). This was primarily because of a boost in the consultation rate of fever cases and to more intensive surveillance and health promotion activities. The average monthly cost for hospitalization of suspected dengue cases amounted to 130 515.60 USD (0.53 USD p.i. p.m.). The hospital cost per hospitalized patient was 183.90 USD, or 31.70 USD per day of which 20% were indirect costs, mainly for support services.
At the community level, the average economic cost p.m. per inhabitant tripled from 0.84 USD to 2.48, constituting 40.9% of all costs during the outbreak. Between July and December, the community invested p.m. the equivalent of 0.63 USD per inhabitant in volunteer time for active surveillance of fever cases, and it incurred out-of-pocket expenditure and productivity losses because of the hospitalization of 0.08 USD and 0.25 USD p.i. p.m., respectively. Per suspect case, the average out-of-pocket expenditure was 26.70 USD (72% for additional food purchases, 13% for transportation, 7% for medicines and 8% for other expenses), and the productivity loss was 85.90 USD. This brings the total cost per hospitalized case up to 296.60 USD, of which 62.0% are direct medical costs, 9.0% direct non-medical costs and 29.0% indirect cost.
The overall total costs breakdown per period, actor and component, is presented in Figure 2. In both periods and for all actors, salaries or the value of non-remunerated voluntary labour makes up the major cost component. There was an increase in most components because of the control of the outbreak, but for the vector control programme predominantly in materials and supplies and for the primary healthcare services and the community mostly in labour.
The economic cost for A. aegypti control and dengue prevention in Guantanamo city is substantial: 2.76 USD per inhabitant p.m. in non-epidemic periods. The main cost driver for the A. aegypti control programme and for the healthcare system and the community is the value of the time invested to get the work done. The intensive routine control efforts result in non-endemicity and low A. aegypti infestation levels, but not quite low enough to avoid occasional dengue outbreaks. When these occur, additional efforts to further reduce vector density, active surveillance and case management lead to a cost explosion and expenses reach 6.05 USD p.i. p.m. The incremental costs were mainly incurred by the population and the primary and secondary healthcare level and hardly by the vector control programme. Nonetheless, the latter substantially increased the use of insecticides and fuel, which have to be purchased in hard currency.
Our integrated account of the expenditure of all actors involved in dengue control, both in non-transmission and epidemic periods, allows a global view of the relative weight of the activities of the different actors and the resulting costs. However, we did not consider the costs incurred by the MOH and other sectors at supra-provincial level, in particular by the National Unit for Vector Surveillance and Control and by the national reference laboratory of ‘Pedro Kourí’ Institute. This should not meaningfully affect our results and, in any case, leads to underestimating the expenses during the epidemic period. On the other hand, the set up of the Cuban Health Care System, its reliable routine information flow, the active surveillance during dengue outbreaks and the systematic hospitalization of suspect cases guarantee minimal underreporting.
During the non-epidemic period, the costs for the routine A. aegypti control programme in Guantánamo (1.67 USD p.i. p.m.) are of the same order of magnitude as those reported for Santiago de Cuba in 2004 (Baly et al. 2009), but they exceed figures published for other countries. In Cambodia, Suaya et al. (2007) reported a total cost of 0.02 USD p.i. p.m. for larval control campaigns with temephos, twice a year and covering 23% of accessible water-storage containers. Orellano and Pedroni (2008) pointed out that preventive intra and extra domiciliary adulticide spraying, at 7-day intervals, combined with larval control, can cost 0.67 USD p.i. p.m.. Baly et al. (2011) estimated the cost of routine Aedes control – larviciding and reactive adulticiding – at 0.05 USD p.i. p.m. in Venezuela and at 0.08 USD p.i. p.m. in Thailand. Tun-Lin et al. (2009), in a study in eight countries in Asia, Africa and Latin America with widely varying control strategies, report costs ranging from 0.05 USD p.i. p.m. in The Philippines to 0.66 p.i. p.m. in Kenya. However, none of the above programmes attain the intensity and the coverage of the Cuban routine programme where, to start with, all premises are inspected, temephos is applied to water holding containers at least once a month, and detection of A. aegypti infestation systematically leads to perifocal adulticiding.
During the outbreak, the cost p.i. for the vector control programme only increased slightly, with 0.21 USD, to 1.88 USD p.i. p.m. The explanation lays in the intensity of the routine activities already implemented in non-epidemic periods. The incremental cost relates to the more intensive pro-active use of insecticides for adult vector control, essentially with the existing workforce, which is high in comparison with other countries.
Even so, 1.88 USD p.i. p.m. is the highest expenditure to control a dengue epidemic ever reported in the scanty international literature. Armien et al. (2008) estimated 0.13 USD p.i. p.m. for vector control during the 2005 epidemic in Panamá. However, differences in dengue epidemiology and control programmes structure and activities make it difficult to compare costs across countries. At the national level, the cost of vector control in the 1997 epidemic in Santiago de Cuba reached 1.6 USD p.i. p.m. (Valdes et al. 2002). The Cuban Programme structure did not change since 1997, and the main control actions stayed similar.
One should add to the above figure the (opportunity) cost of the involvement of the organized community in vector control, which was valued at 0.82 USD p.i. p.m. No other reports of this cost component are available in the literature, nor on the community and health services costs linked to active search for fever cases during a dengue epidemic.
The direct hospitalization costs for suspect dengue cases reported here, 31.70 USD per day, are within the range of the ones during the 1997 and 2006 epidemics in Santiago de Cuba, 28.50 USD and 46.06 USD, respectively (Valdes et al. 2002; Arias 2009). Torres and Castro (2007) reports a direct cost of USD 130.0 per day for dengue patients hospitalized in Nicaragua in 1994. The total cost per hospitalized case in our study is 296.6 USD, the average length of stay 5.8 days. Suaya et al. (2009) reported for five countries in Latin America and three in Asia an average cost of 571.0 USD and on average 3.8 hospitalization days, ranging from 418.0 USD (6.4 days, Guatemala) to 1065.0 USD (5.2 days, Panama). The average number of hospitalization days ranged from 2.8 (Malaysia) to 6.4 (Guatemala). The overall average direct medical costs, direct non-medical costs and indirect costs made up 68%, 9% and 23%, respectively. This, as well as the average number of hospitalization days, roughly coincides with our findings, but our absolute cost is lower. In Cuba, all clinically suspect dengue cases are hospitalized, which could lead to a different case mix. This is not necessarily reflected in the length of stay because hospital admission of non-severe cases occurs early after onset of symptoms. Furthermore, the costs for initial ambulatory care for hospitalized cases are not included in our figures. However, these are limited to one patient contact at the primary care level, low in comparison to other countries and, in comparison, already partially included in the hospital costs (e.g. the cost of serological tests). The main explanation for the difference in the total cost per hospitalized case lies in the lower wage rates, the main cost driver of the direct medical as well as the indirect cost. In Cuba, patients hardly incur any direct medical cost, apart from 1.8 USD for non-essential medicines, or costs for days of work lost, because these are compensated by the public social security system.
The epidemic was controlled after 7 months, but the monthly expenditure was more than doubled the already substantial cost for dengue prevention in non-epidemic periods. At the start of the outbreak, the intensification of activities brought Aedes infestation down to levels that were, perhaps surprisingly, well below the ones in preceding years. This all but implies that sustained routine dengue prevention efforts are useless and a waste of resources. But without them, the disease could well become endemic, as is the case in large parts of Latin America and Caribbean, profoundly affect human well-being and cause haemorrhagic manifestation and deaths because of cocirculation of dengue strains. A recent simulation of the potential consequences of the introduction of the dengue virus in Cairns, Australia, indicates that control cost as well as number of cases during the outbreak increases when mosquito surveillance and early case detection are halted (Vazquez-Prokopec et al. 2010). This cannot be directly extrapolated to the epidemiological and ecological context of our study, but illustrates our point. Lack of political commitment was certainly not the case in Guantanamo, but it remains to be assessed whether – given the number of people introducing the virus in the population, the low herd immunity and the high temperature – the investment in routine control was insufficient to prevent dengue transmission, or whether the deployed strategy was suboptimal.
Further and more in-depth studies are needed, in Cuba and wherever dengue control efforts are made, to assess which mix of activities and actors maximizes the effectiveness and cost-effectiveness in terms of morbidity and mortality averted. Simulations by modelling some of the data available today could possibly deliver the first outline of an answer to this question, but will have to be complemented by carefully designed and conducted prospective research. Finally, decision makers should be aware that it entails substantial costs to implement a successful routine vector control programme and to promote community participation to keep Aedes infestation levels low. The resources needed to control a dengue outbreak represent an opportunity costs and could, perhaps, be better invested in preventive measures that address risk factors outside the scope of action of Aedes vector control programmes. Improving the deficient solid waste collection and water supply system or tackling other structural problems could, furthermore, not only reduce dengue transmission but also the incidence of other diseases associated with sanitary and environmental conditions.
We gratefully acknowledge the role played by the health sector and provincial vector control programme staff involved in the dengue prevention and control activities. The study was partially funded through the framework agreement between the Institute of Tropical Medicine, Antwerp, and the Belgium Directorate – General for Development Co-operation (project 95900).