Model-based estimates of risks of disease transmission and economic costs of seven injection devices in sub-Saharan Africa.

Model-based estimates of risks of disease transmission and economic costs of seven injection devices in sub-Saharan Africa* Donatus U. Ekwueme,1 Bruce G. Weniger,2 & Robert T. Chen3 Objective To investigate and compare seven types of injection devices for their risks of iatrogenic transmission of bloodborne pathogens and their economic costs in sub-Saharan Africa. Methods Risk assumptions for each device and cost models were constructed to estimate the number of new hepatitis B virus (HBV) and human immunodeficiency virus (HIV) infections resulting from patient-to-patient, patient-to-health care worker, and patient-to- community transmission. Costs of device purchase and usage were derived from the literature, while costs of direct medical care and lost productivity from HBV and HIV disease were based on data collected in 1999 in Coˆte d’Ivoire, Ghana, and Uganda. Multivariate sensitivity analyses using Monte Carlo simulation characterized uncertainties in model parameters. Costs were summed from both the societal and health care system payer’s perspectives. Findings Resterilizable and disposable needles and syringes had the highest overall costs for device purchase, usage, and iatrogenic disease: median US$ 26.77 and US$ 25.29, respectively, per injection from the societal perspective. Disposable-cartridge jet injectors and automatic needle-shielding syringes had the lowest costs, US$ 0.36 and US$ 0.80, respectively. Reusable-nozzle jet injectors and auto-disable needle and syringes were intermediate, at US$ 0.80 and US$ 0.91, respectively, per injection. Conclusion Despite their nominal purchase and usage costs, conventional needles and syringes carry a hidden but huge burden of iatrogenic disease. Alternative injection devices for the millions of injections administered annually in sub-Saharan Africa would be of value and should be considered by policy-makers in procurement decisions. Keywords Disease transmission; Iatrogenic disease/epidemiology; Injections/instrumentation/economics; Needles/adverse effects/ economics; Syringes/adverse effects/economics; Injections, Jet; Hepatitis B/transmission; HIV infections/transmission; Risk factors; Costs and cost analysis; Models, Theoretical; Africa South of the Sahara (source: MeSH, NLM ). Mots cle´s Transmission maladie; Affection iatroge´nique/e´pide´miologie; Injections/instrumentation/e´conomie; Aiguille/effets inde´sirables/e´conomie; Injections flash; Seringue/effets inde´sirables/e´conomie; He´patite B/transmission; HIV, Infection/transmission; Facteur risque; Couˆt et analyse couˆt; Mode`le the´orique; Afrique subsaharienne (source: MeSH, INSERM). Palabras clave Transmisio´n de enfermedad; Enfermedad iatroge´nica/epidemiologı´a; Inyecciones/instrumentacio´n/economı´a; Agujas/efectos adversos/economı´a; Jeringas/efectos adversos/economı´a; Inyecciones a chorro; Hepatitis B/transmisio´n; Infecciones por VIH/transmisio´n; Factores de riesgo; Costos y ana´lisis de costo; Modelos teo´ricos; A´frica al Sur del Sahara (fuente: DeCS, BIREME ). Bulletin of the World Health Organization 2002;80:859-870. Voir page 867 le re´sume´ en franc¸ais. En la pa´gina 867 figura un resumen en espan˜ol. Introduction The Expanded Programme on Immunization has been increasingly successful in reducing the incidence of vaccine- preventable diseases in developing countries (1), where, unfortunately, a pattern of unsafe injection practices has been observed (2). Simonsen et al. estimated the prevalence of unsafe injections to range from 20% up to at least 50% in these countries. In 20–80% of health centres in sub-Saharan Africa there are insufficient supplies and equipment to guarantee safe injection (3). Incorrect injection practices include reuse of contaminated needles and syringes without sterilization between patients (4); incorrect disposal of used needles and syringes in the community (5); absence of swabbing with alcohol or acetone of the reusable nozzles of needle-free jet injectors between consecutive patients (6); and other unsafe practices, such as changing needles but not syringes between patients (7). When not properly sterilized, or if contaminated, needles and syringes can produce local abscesses (8, 9) and can transmit bloodborne infections between patients (10, 11). Needlestick injuries can transmit infectious agents from patients to health care workers (12–15), while incorrect disposal can transmit disease to the community as a consequence of both needlestick injuries and improper reuse (3). Hepatitis B virus (HBV) (16) and human immunodeficiency virus (HIV) (17) are two of the most important bloodborne pathogens in terms of prevalence, * The mention of trade names is for identification purposes only and does not imply endorsement by the authors, the United States Centers for Disease Control and Prevention, or the United States Department of Health and Human Services. 1 Prevention Effectiveness Fellow, National Immunization Program (NIP), Centers for Disease Control and Prevention (CDC), Atlanta, GA, USA. Currently, Senior Health Economist, Epidemiology and Health Services Research Branch, Division of Cancer Prevention and Control, National Center for Chronic Disease Prevention and Health Promotion, CDC. Correspondence should be sent to this author at Mailstop K-55, CDC, 4770 Buford Highway, Atlanta, GA 30341, USA (email: dce3@cdc.gov). 2 Assistant Chief for Vaccine Development, Vaccine Safety and Development Branch, NIP, CDC, Atlanta, GA, USA. 3 Chief, Vaccine Safety and Development Branch, NIP, CDC, Atlanta, GA, USA. Ref. No. 99-0135 859Bulletin of the World Health Organization 2002, 80 (11) morbidity, and mortality, especially in many parts of the developing world (4, 18). Complications associated with HBV infection include chronic active hepatitis, cirrhosis of the liver, primary hepatocellular carcinoma, and premature death (16). HIV infection leads to the acquired immunodeficiency syndrome (AIDS), opportunistic infections, and premature death. It is estimated that humans in health care settings receive each year between 8 and 12 billion parenteral injections, of which about one billion are for vaccines (19). In addition to routine immunizations for children, emergency campaigns in 1996 alone accounted for the administration of more than 240 million doses of vaccine (20). The plans for global measles control and eradication (21) can be expected to require billions more injections than are currently adminis- tered. As the number of vaccine injections increases, it may become increasingly difficult to ensure the safety of every injection, and thus to minimize risk for consequent iatrogenic disease (7). Since 1997, WHO, the United Nations Children’s Fund (UNICEF), and theUnitedNations Population Fund (UNFPA) have strongly recommended (22–24) the use of ‘‘auto-disable’’ needles and syringes (25) designed to prevent improper reuse. (Originally called ‘‘auto-destruct’’, these syringes were renamed because they still require proper disposal and destruction by incineration or othermeans.) The three agencies also agreed on a policy of ‘‘bundling’’, which requires donors of vaccine for developing countries also to supply a corresponding number of auto-disable needles and syringes along with ‘‘sharps’’ collection boxes to permit safe disposal. The full risks and economic costs of conventional needles and syringes and alternative injection delivery technologies have not been adequately compared. We investigated the risks of iatrogenic disease transmission and the economic costs associated with various such devices for the parenteral administration of vaccines and other medica- tions. Sub-Saharan Africa was selected as the setting for the model, because injection practices there are often unsafe, and severe financial barriers exist for the introduction of newer technologies. Methods The risk model Three major categories of transmission of bloodborne infections by injection devices were modelled. First, patient- to-patient transmission can occur when a device is reused without sterilization or when it is incorrectly sterilized and transfers infected blood between patients. Second, transmis- sion from patient-to-health care worker occurs when an accidental needlestick injury transfers infectious patient blood to the worker. Third, patient-to-community transmission may occur from improper disposal of needles and syringes, as when people scavenging waste dumps receive needlestick injuries. Devices ‘‘recycled’’ from dumps may also be reused unsterile, producing iatrogenic abscesses and transmission of pathogens. A risk model was constructed for each of these routes of transmission, building on previous models (4, 13, 26, 27), in order to estimate the number of new HBV or HIV infections that might result from seven injection technologies. The general model is represented by the following equation: In order to simplify Eq. 1 and because vaccines are administered mainly to young children, we ignored the decrease in susceptibility to HBV infection that occurs among groups of increasing age, due to immunity from incident HBV infections (as evidenced by the presence of hepatitis B core antibody). We studied the use of seven devices for the parenteral delivery of vaccine and other medications (see Box 1) (28–31). Disease costs were totalled from the economic perspectives both of the health care system (‘‘payer’s’’ direct medical costs only) and of society (directmedical and lost productivity costs). The societal perspective allows a comprehensive assessment of the overall impact of different injection technologies on the economies of the countries concerned. The perspective of the health care system focuses on the narrower impact for national health care expenditures. HBV and HIV prevalence The prevalence of carriers of HBV surface antigen in the population of vaccinees whose blood might contaminate injection equipment was estimated at 10% (the ‘‘base case’’) for countries in sub-Saharan Africa, with a lower estimate of 5% and an upper of 15% used for sensitivity analysis (16, 32, 33) (Table 1). HIV seroprevalence was also estimated at 10% on the basis of reported rates exceeding 5% but less than 15% in 16 countries in the region (17). An HIV seroprevalence range of 2–25%was used for the sensitivity analysis. The 2% rate was estimated on the basis of data from the 19 countries in the region with the lowest reported values, ranging from 0.08% in Mauritius to 4.16% in Gabon (17). The 25% rate was based on data from eight countries with values ranging from 16% in Malawi to 36% in Botswana. Transmission from patient to patient For the base case, it was assumed that after every sterile injection with either a resterilizable needle and syringe (N&S) or a disposableN&S, non-sterile reuse would occur 30% of the time (range: 15–50% for the sensitivity analysis) (2, 4, 26) (Table 1). We assumed no risk to patients of blood exposure for the auto-disableN&S, auto-shieldingN&S, and disposable- cartridge jet injector devices. For manual-shielding N&S devices, the base case assumed one non-sterile reuse 15% of the time (range: 1–30%) (34). For reusable-nozzle jet injector devices, we assumed a worst-case scenario in which health care workers did not swab the nozzle with alcohol or acetone between patients (6), contrary to the manufacturers’ recom- mendations. On this basis we estimated a 1% probability (range: 0.1–5.0%) that the device would expose the next patient to transferred blood (6, 35–37). The probability of newly acquiring HBV infection as a result of exposure to reuse of or to needlestick injury from an Eq. 1. Expected number of new cases of HBV or HIV infections = (prevalence of HBV [or HIV]) x (probability of blood exposure through: A. reuse of non-sterile needle, or B. vaccination by reusable-nozzle jet injector, or C. needlestick injury to health care worker, or D. probability of improper disposal x probability of needlestick injury or unsterile reuse in community) x (probability of transmission of infection upon blood exposure to HBV [or HIV]) x (proportion susceptible in population [1 – prevalenceHBV [or HIV]]). 860 Bulletin of the World Health Organization 2002, 80 (11) Research unsterile injection device containing blood from an infected person was assumed to be 30% (range: 20–40%) (Table 1). For acquisition of HIV infection, 0.3% was used for the base case (range: 0.2–0.5%). These rates for HBV and HIV were based on empirical data from needlestick injury case series and surveillance (2, 38–43). We assumed that needles and jet injector nozzles contaminated with blood or tissue fluid from intramuscular or subcutaneous injections would transmit infection at rates similar to those observed in the above studies of injuries from needles used for drawing blood or other intravascular access. Transmission from patient to health care worker On the basis of data from the literature (2) and the observations in Coˆte d’Ivoire, Ghana, and Uganda (33, 51), we assumed a base-case frequency for needlestick injuries of 5% (range: 1– 8%) for each use of the resterilizable N&S (which requires more handling to disassemble, clean, and sterilize), and 3% (range: 2–5%) for the disposable N&S (Table 1). Table 1 also provides the modelled probabilities for needlestick injuries for other devices (2, 29, 33, 34, 37, 51). Themanual-shieldingN&S carried some needlestick injury risk because of the possibility that health care workers would intentionally not activate the safety features, in order to reuse the device. The hypothetical auto-shielding N&S and both types of jet injectors were assumed to have no risk of needlestick injury. Transmission from patient to community This route for acquiring infection is a consequence of improper disposal of sharps and needlestick injury outside the original health care setting where the device was originally used. In the model, the probabilities assumed for unsafe disposal (Table 1) are multiplied by those for needlestick injuries with various devices. Of course, the auto-shielding N&S and both types of jet injectors present no risk to the community. Because of the absence of data, we ignored possible patient-to-patient transmission from reuse of such disposed sharps salvaged in the community. Economic costs Costs of purchasing and using devices For each injection device studied, data were collected from UNICEF (44) and WHO (45–48), device manufacturers (30, 49, 50), and the literature (5) for purchase prices of the items themselves, as well as the costs of necessary equipment (e.g. sterilizers, spare parts, supplies, and other consumables, including items for proper sharps disposal). In addition, the costs of labour for maintenance of necessary equipment and for actual administration of vaccine were estimated. The value of vaccine wasted in the routine use of some devices (e.g. purging air from reusable-nozzle jet injector) was also considered. All costs, including capital costs for equipment and reusable supplies, were amortized for the expected number of injections over the lives of the equipment or supplies, and converted to cost per injection. In order to account for uncertainties about such purchase and usage costs, in the sensitivity analysis the calculated base-case values were varied by factors of 25% for the lower estimate and 200% for the upper. Table 1 summarizes the overall total of such costs for each device. The individual component costs and details of the calculations, along with reference citations to the sources used (5, 30, 44, 45, 47–50) are provided in Annex Table A (available on the Bulletin web site: http://www.who.int/bulletin). Direct medical care costs Medical care costs were based upon data fromoriginal sources in Coˆte d’Ivoire, Ghana, and Uganda collected by the first author from June to December 1999 (33, 51). These direct costs for each new HBV and HIV infection were determined by modelling the reported costs, frequencies, coverage, and duration of outpatient visits, inpatient care, diagnostic tests, and occasional antiviral therapies for HBV (i.e. interferon in Coˆte d’Ivoire only) and for HIV (i.e. zidovudine, lamivudine, and indinavir). HBV infections were assumed to have been acquired in the first year of life as a result of unsafe vaccination or other injection. It was also assumed that the resulting direct medical costs would all be incurred in the year of the average age of premature death resulting from this disease (Coˆte d’Ivoire: 43 years, Ghana: 40 years, Uganda: 41 years). These direct medical costs were then discounted at 3% to present net values, using standard methods (52). HIV infections fromunsafe injectionwere also assumed to have been acquired in the first year of life. Using models and methods described by Over et al. (53) and Mansergh et al. (54), symptomatic AIDS and death were assumed to ensue among infected infants at a rate of 10% per year until, by the age of Box 1.Descriptions of injection devices compared in this study for the parenteral administration of vaccines Resterilizable N&Sa Conventional resterilizable (detachable) steel needle and resterilizable syringe (traditionally glass, but plastic also available). Disposable N&S Conventional plastic syringe with detachable, disposable, steel needle. Auto-disable N&Sb Auto-disabling, single-use needle permanently fixed to a plastic syringe, designed to prevent inadvertent or intentional reuse. Manual-shielding N&Sb Manually activated N&S in which, after injection, the user may activate a needle-shielding or needle-retracting mechanism to prevent needlestick injuries. Auto-shielding N&Sc Hypothetical, automatically activated, single-use N&S with both needle-shielding or needle-retraction and auto-disabling features to prevent both needlestick injuries and inadvertent or intentional reuse. Reusable-nozzle jet injectorb Needle-free jet injector for high-speed vaccination which feeds vaccine from multidose vials through reusable fluid chambers, pathways, and nozzles that are in contact with consecutive patients without intervening sterilization. Disposable- cartridge jet injectorb Needle-free jet injector designed for disposable, single-use vaccine cartridges/nozzles. a N&S = needle and syringe. b Commercial examples and prototypes are mentioned in Annex Box A, available on the Bulletin web site: http://www.who.int/bulletin. c No such commercial device is known to exist. A hypothetical device was therefore modelled, assuming the same purchase cost per unit as that of the manual-shielding N&S device. 861Bulletin of the World Health Organization 2002, 80 (11) Risks and costs of injection devices in sub-Saharan Africa 10 years, all had become symptomatic and died within a year. The direct medical costs attributable to HIV infection were discounted by 3% (conversion rate of 0.806) to the present value. Foreign currencies were converted to US$ at the year 2000 exchange rates (55). The arithmetic means of the present value totals of direct medical costs for the three countries were used as single sub-Saharan Africa estimates for the model. Additional details and assumptions for the input values and calculations of direct medical costs are described in Annex Table B (available on the Bulletin web site: http://www.who.int/bulletin). Indirect costs — lost productivity Lost productivity was the sole indirect cost considered for iatrogenic HBV and HIV diseases (Table 1) and was modelled Table 1. Input parameters and assumptions for modelling costs and risks of alternative injection technologies and consequential disease from hepatitis B virus (HBV) and human immunodeficiency virus (HIV) Parameters Base case Lower estimate Upper estimate References Prevalence of infection in vaccinated population HBV (surface antigen) 0.1 0.05 0.15 16, 32, 33 HIV 0.1 0.02 0.25 17 Patient-to-patient transmissiona Probability of unsterile reuse/blood exposure [resterilizable N&Sb, disposable N&S] 0.3 0.15 0.5 2, 4, 26 Probability of unsterile reuse/blood exposure [auto-disable N&S, auto-shielding N&S, disposable-cartridge jet injector] 0 0 0 c Probability of unsterile reuse/blood exposure [manual-shielding N&S] 0.15 0.01 0.3 34 Probability of blood exposure from routine use [reusable-nozzle jet injector] 0.01 0.001 0.05 6, 35, 36 Patient-to-health care worker transmission Probability of needlestick/blood exposure from: Resterilizable N&S 0.05 0.01 0.08 2, 33 Disposable N&S 0.03 0.02 0.05 33 Auto-disable N&S 0.01 0.001 0.02 29, 33, 37 Manual-shielding N&S 0.002 0.001 0.004 c Auto-shielding N&S, reusable-nozzle jet injector, disposable-cartridge jet injector 0 0 0 c Patient-to-community transmissiona Probability of unsafe sharps disposal [resterilizable N&S, disposable N&S, auto-disable N&S, manual-shielding N&S, auto-shielding N&S] 0.05 0.01 0.1 33 Probability of unsafe sharps disposal [reusable-nozzle jet injector, disposable-cartridge jet injector] 0 0 0 c Probability of needlestick injury [resterilizable N&S] 0.0002 0 0.0005 c Probability of needlestick injury [disposable N&S, auto-disable N&S] 0.002 0.0005 0.004 33 Probability of needlestick injury [manual-shielding N&S, auto-shielding N&S, reusable-nozzle jet injector, disposable-cartridge jet injector] 0 0 0 c Probability of infection upon blood exposurea HBV [resterilizable N&S, disposable N&S, manual-shielding N&S, reusable- nozzle jet injector] 0.3 0.2 0.4 2, 38–42 HIV [resterilizable N&S, disposable N&S, manual-shielding N&S, reusable- nozzle jet injector] 0.003 0.002 0.005 2, 40–43 Lifetime direct medical care costs, per infectiond (US$) HBV 34.19 14.48 67.77 e HIV 2 532 1 436 4 139 e Cost of lifetime productivity loss, per infectiond (US$) HBV 2 575 1 016 3 637 e HIV 19 129 17 570 20 191 e Device purchase and usage costsf (US$) Resterilizable N&S 0.0697 0.0174 0.1394 g Disposable N&S 0.1016 0.0254 0.2031 g Auto-disable N&S 0.1357 0.0339 0.2713 g Manual-shielding N&S 0.5432 0.1358 1.0863 g Auto-shielding N&S 0.5432 0.1358 1.0863 g Reusable-nozzle jet injector 0.0393 0.0098 0.0786 g Disposable-cartridge jet injector 0.3595 0.0899 0.7191 g a Parameters are relevant only for the types of injection devices indicated in square brackets. See Box 1 for definitions of injection devices. b N&S = needle and syringe. c Values estimated in the absence of published data or previous sources. d Medical costs and productivity losses estimated at year 2000 prices, discounted by 3%. e Sources and calculations of lifetime direct medical costs and indirect costs (productivity losses) for HBV and HIV detailed in Annex Table B (available on the Bulletin web site: http://www.who.int/bulletin). f Base-case estimates for device purchase and usage costs were varied conservatively to 25% and 200% for the lower and upper estimates respectively. See Annex Table A (available on the Bulletin web site: http://www.who.int/bulletin) for details of calculation and ref: 5, 30, 44–50. g See Annex Table A footnotes (available on the Bulletin web site: http://www.who.int/bulletin) for references used in estimating device purchase and usage costs. 862 Bulletin of the World Health Organization 2002, 80 (11) Research using an adaptation of themethod of Over et al. for determining lost productivity from perinatal HIV transmission (53) (see footnote o in Annex Table B), available on the Bulletin web site: http://www.who.int/bulletin, for further explanation). Average annual earnings in public or private sectors collected from original sources in Coˆte d’Ivoire, Ghana, and Uganda (33) were adjusted for unemployment rates, and then applied to the years of life lost. This was calculated as the difference between average life expectancy at birth (51 years in Coˆte d’Ivoire, 57 in Ghana, 48 in Uganda) and the earlier average age of death due to HBV (Coˆte d’Ivoire: 43 years, in Ghana: 40 years, Uganda 41 years) or to HIV (6 years in all three countries). As assumed by Over et al. (53), HIV-infected infants were assumed to have only 15% of average adult income during the ‘‘lost’’ years from 6 to 15 years (e.g. for tasks such as child care, wood gathering, and other domestic chores), and full income (100%) thereafter to the age of 50 years. From the ages of 51 to 65 years, income was adjusted to 85% of the average. For both HBV and HIV, the amounts of future lost income were discounted at 3% per year, standardized to US$ for the year 2000 exchange rates, and averaged among the three countries for the regional base-case amounts shown in Table 1. To avoid counting lost productivity in full years for both the year of premature death and year of death after normal life expectancy, only half of lost income was counted in those first and last years of the discounting model. Further details, input data, and the lost productivity discounting formula are provided in Annex Table B and Annex Box B (available on the Bulletin web site: http://www.who.int/bulletin). Sensitivity analysis In order to ascertain the degree of uncertainty inherent in the point estimates for the purchase and usage costs of each injection device and for the direct medical and indirect (lost productivity) costs of HBV and HIV disease, we performed multivariate sensitivity analyses using the Monte Carlo simulation sampling method (56–58). For various base-case point estimates of input data in Annex Table B, lower and upper estimates were made and modelled in parallel runs. For example, in Uganda, the number of days of hospitalization for HIV disease averaged 14 days (lower and upper estimates 7 days and 31 days, respectively). In Ghana, the average number of follow-up doctors’ visits for HIV care varied from two to 10, around a base case of four. In Coˆte d’Ivoire, the average cost of a laboratory test for hepatitis B surface antigen was US$ 42.14, with US$ 14.05 and US$ 84.27 set as the lower and upper estimates, respectively. Calculated device purchase and usage costs (Annex Table A, available on the Bulletin web site: http://www.who.int/ bulletin) were varied by 25% and 200% to produce lower and upper estimates. Parallel runs of the model using such lower and upper cost estimates were used, along with the base- case estimates, to construct triangular probability distributions (59) for the Monte Carlo analyses (the triplicate input costs are provided in the final three sections of Table 1). The triangular probability distribution is often used in the absence of a large data set when the mean value is small and the standard deviation is large (60). The simulations were conducted using@RISK software (Palisade Corporation, Newfield, NY, USA) (61), an add-in to Excelt spreadsheet software (Microsoft Corporation, Red- mond,WA,USA). On each of 1000 simulation runs, a value for each parameter was drawn from its associated distribution and used to calculate risk and cost estimates for each injection device. For each device, the output of the simulation runs produced the mean, standard deviation, 5th, 50th (median), and 95th percentiles. Results Cost of device purchase and usage The device with the highest purchase price and usage cost was the manual-shielding N&S, at US$ 0.54 each (Table 1, with input details provided in Annex Table A (available on the Bulletinweb site: http://www.who.int/bulletin). The reusable- nozzle jet injector was the least expensive to buy and use, at US$ 0.04 per injection. The conventional disposable N&S was calculated to cost US$ 0.10 per injection. Number of disease cases produced Base-case point estimates for the predicted number of HBV and HIV infections resulting from one million injections with each of the modelled devices are shown in Table 2. The conventional resterilizable N&S caused the greatest number of iatrogenic infections per million injections (n = 9545), followed closely by the disposable N&S (n = 9002). The manual-shielding N&S incurred somewhat less then half this burden (n = 4145). In contrast, both the auto-disable N&S and reusable-nozzle jet injector produced relatively few HBV and HIV infections (n = 276 and n = 273 respectively). Of course, the reusable-cartridge jet injector and the hypothetical auto-shielding N&S produced no infections according to the model. Costs of disease The overall economic burdens of HBV and HIV disease resulting from the predicted iatrogenic infections are summar- ized in Table 3. The overall societal costs attributable to the resterilizable N&S and disposable N&S were US$ 26.71 and US$ 25.18 respectively, as the base case point estimates per injection (HBV and HIV costs combined). Each use of an auto-disable N&S was estimated to produce disease costs of US$ 0.77, which was nearly identical to the point estimates for the reusable-nozzle jet injector (US$ 0.76). The Monte Carlo sensitivity analyses of these disease costs, also in Table 3, reveal medians that vary only slightly from the point estimates of each injection device for HBV disease, but somewhat more widely for HIV/AIDS. The 5th and 95th percentiles reveal modest ranges. For example, the resterilizable N&S ranged from US$ 11.71 to US$ 40.19 for HBV disease, and fromUS$ 0.90 to US$ 4.46 for HIV/AIDS. Overall costs Combining all costs for a societal perspective — device purchase and usage, medical costs, and lost productivity — it was estimated that the most expensive technology for administering vaccines is the resterilizable N&S, at US$ 26.77 per injection (Fig. 1). The next most expensive is the disposable N&S (US$ 25.29). The lowest costs were for the disposable-cartridge jet injector and auto-shieldingN&S, at US$ 0.36 and US$ 0.54, respectively. Intermediate costs were found for the reusable-nozzle jet injector (US$ 0.80) and auto- disable N&S (US$ 0.91). Looking only from the health care payer’s perspective (Fig. 2), the relative overall costs of the various devices change. 863Bulletin of the World Health Organization 2002, 80 (11) Risks and costs of injection devices in sub-Saharan Africa The manual-shielding N&S takes the lead as the most expensive device at a median cost of US$ 0.81 per injection, followed by identical costs (US$ 0.67) for both the res- terilizable N&S and disposable N&S. The disposable-cartridge jet injector moves up to fifth in order of cost (US$ 0.37), no longer being the lowest cost. The lowest cost is now for the reusable-nozzle jet injector (US$ 0.06). The auto-disable N&S becomes next to lowest, at US$ 0.16. The multivariate sensitivity analyses found median costs to be similar to the base-case results (Table 3, Fig. 1). For example, the HBV disease cost attributable to each injection with the resterilizable N&S was US$ 24.66 in the base case and US$ 23.17 (94%) in the multivariate sensitivity analyses. For HIV disease, the corresponding values were US$ 2.05 and US$ 2.24 (109%) respectively. The 5th and 95th percentiles revealed modest ranges. For example, for the disposable N&S, the HBV disease cost ranged from US$ 11.45 to US$ 38.02 (around a base case of US$ 23.25, median US$ 22.09). Discussion Our modelling reveals that unsafe parenteral injection in sub- Saharan Africa causes a substantial health and economic burden from iatrogenic disease. Most of this cost is hidden because new infections are usually unrecognized, or cannot be linked to a causative injection, and because most of the disease sequelae are greatly delayed. We found that the most commonly used injection devices, resterilizable and disposable needles and syringes, actually cost around US$ 0.67 per injection in direct medical costs, and a staggering US$ 25 to US$ 27 in overall costs when lost productivity from premature death was included. These costs are high relative to estimated annual expenditures of US$ 33 per capita for all public and private health purposes by countries in sub-Saharan Africa (62). Our input assumptions and findings are consistent with those of previouswork on the incorrect use of injection devices (2, 4, 7, 26, 63). Another mathematical model assumed that a needle would be reused between one and four times (4). We assumed a probability of 0.3 (range: 0.15–0.5). An average of 33% of health centres in Chad, Coˆte d’Ivoire, Uganda, and Swaziland reused syringes or needles without sterilization (3). Amacrolevel analysis byKane et al. for the entire population of sub-Saharan Africa calculated the annual number of newHBV and HIV infections attributable to unsafe injection to be 780 052 and 51 208 respectively (26). Our modelling exercise is limited by the numerous assumptions and input cost estimates that must be made, as there is a paucity of published, scientifically gathered sources for such data. Nevertheless, the risk and cost estimates used here are relatively conservative. They excluded the economic consequences of disease and premature death arising from other bloodborne pathogens that can be contracted by unsafe injection, e.g. hepatitis C, Trypanosoma sp., Plasmodium sp., and agents of haemorrhagic fever. We also ignored treatment costs for opportunistic infections associated with HIV/AIDS, such as tuberculosis, as well as burial costs, which can be substantial in developing countries (64, 65). Also excluded were the indirect costs for the time of others in caring for a patient. Thus, the Bulletin results are probably underestimates of the true costs of unsafe injection. This problem is not peculiar to sub-Saharan Africa or other developing countries. A survey in Eastern Europe in 1992–93 revealed about half of health centres were adminis- tering unsafe injections (1). HBV and HIV spread widely in Romanian orphanages due to needle and syringe reuse (66–69), as did HBV in the Republic of Moldova (70). Fortunately, the problem is becoming increasingly recognized (19). In 1994, more than 50 African countries endorsed the Yamoussoukro Declaration on the safety of injections and its goal of 95% safe practice (1). As a result, the auto-disable N&S that cannot be reused is now the normative standard of care for developing country immunization programmes (22–24). We estimated US$ 0.14 per injection to buy and use them, plus an additional US$ 0.77 per injection for the medical costs of consequential needlestick injuries, which they do not prevent (Fig. 1). Reusable-nozzle jet injectors are a needle-free vaccina- tion technology formerly used inAfrica formass immunization Table 2. Numbers of iatrogenic hepatitis B virus (HBV) and human immunodeficiency virus (HIV) infections attributable to the use of injection devices, by route of transmission, per million injectionsa Type of injection device No. of infections Patient-to-patient Patient-to-health Patient-to-community All-routes totalsb transmission care worker transmission transmission HBV HIV HBV HIV HBV HIV HBV HIV Both Resterilizable N&Sc 8100 81 1350 14 <1d <1d 9450 95 9545 Disposable N&S 8100 81 810 8 3 <1d 8913 89 9002 Auto-disable N&S 0 0 270 3 3 <1d 273 3 276 Manual-shielding N&S 4050 41 54 1 0 0 4104 41 4145 Auto-shielding N&S 0 0 0 0 0 0 0 0 0 Reusable-nozzle jet injector 270 3 0 0 0 0 270 3 273 Disposable-cartridge jet injector 0 0 0 0 0 0 0 0 0 a Calculated from base-case assumptions in Eq. 1 (see text). b Sum of cases for each route of transmission may not equal all-routes totals because of rounding errors. c N&S = needle and syringe. d Less than one infection expected per million injections. 864 Bulletin of the World Health Organization 2002, 80 (11) Research campaigns, such as control of meningococcal (71) and yellow fever (72) outbreaks. We calculated they cost only US$ 0.06 per injection from the health care payer’s perspective. Such devices have delivered billions of injections in mass immuniza- tion campaigns and epidemic control activities since their introduction in the 1950s (73). However, their high initial capital cost and complex maintenance requirements make them unsuitable for routine immunization clinics. We there- fore modelled them only for mass campaigns with an assumed usage of 1000 doses on each day of use (Annex Table A, available on the Bulletin web site: http:www.who.int/bulletin). The Program for Appropriate Technology in Health (PATH) estimated that low-volume use of such devices would result in a much higher direct cost of US$ 0.20 per injection (50). In the mid-1980s, concern about the possibility of bloodborne disease transmission between consecutive vacci- nees from reusable-nozzle jet injectors (Annex BoxA, available on the Bulletin web site: http:www.who.int/bulletin) arose following an outbreak of hepatitis B in California, USA, caused by a Med-E-Jett device (35, 74, 75). A 1990s study in Brazil identified contamination in 1–6% of ejectates collected immediately after the vaccination of patients (6). The Public Health Laboratory Service of the United Kingdom of Great Britain and Northern Ireland, with assistance from WHO, pioneered an animal model to identify and quantify blood at levels theoretically sufficient to transmit HBV in succeeding injections. The Public Health Laboratory Service found contamination of ejectates and/or transmission of HBV in the succeeding injection occurred with all devices tested (36). These and other unpublished studies formed the basis for the Bulletin modelled base case assumption that these devices would transmit blood 1% of the time. In 1997, liability risk led to manufacturer withdrawal of the Ped-O-Jett from the market, followed by its recall by the United States military (76, 77). In 1998, WHO recommended that such injectors should not be used until testing demon- strated their safety (78). The United States Centers for Disease Control and Prevention recommended that public health authorities weigh the potential risk against the benefit in certain situations where the rapid vaccination of large numbers of people is required and the use of needles and syringes is not practical (79, 80). As a result, the world lacks a high-speed device of unquestioned safety for intramuscular or subcuta- neous vaccination for use in influenza pandemics, measles eradication, response to biological terrorism, or other necessary mass immunization campaigns. This vulnerability will probably disappear when high-speed jet injectors with disposable cartridges are developed. Low-workload jet injectors with disposable cartridges now exist (73, Annex Box A, available on the Bulletin web site: http:www.who.int/bulletin), but they are not affordable for use in the developing world because of the current expense of their cartridges — US$ 0.25–0.50 per injection. We modelled these at the lower cost. Also impeding their acceptance are their proprietary, non-interchangeable cartridges. Universal stan- dards for a common cartridge might enhance market demand for such technology and reduce their costs through mass production. Another obstacle is the need for end users to fill the empty cartridges manually in the clinic. Vaccine manu- facturer prefilling would be highly convenient and save users Table 3. Overall direct and indirect costsa (societal perspective) estimated for iatrogenic hepatitis B virus (HBV) and human immunodeficiency virus (HIV) disease attributable to the use of various injection devices, in US$ equivalents, by route of transmission, per injection Costs (US$) Type of Patient-to-patient Patient-to-health Patient-to-community All-route totalsb injection device transmission care worker transmission transmission HBV HIV HBV HIV HBV HIV HBV HIV Resterilizable N&Sc 21.13 (20.05)d [10.30–34.51]e 1.55 (1.94) [0.78–3.85] 3.52 (3.12) [1.41–5.68] 0.26 (0.30) [0.11–0.61] 0.001 (0.001) [<0.001–0.002] <0.001 (<0.001) [<0.001–<0.001] 24.66 (23.17) [11.71–40.19] 2.05 (2.24) [0.90–4.46] Disposable N&S 21.13 (20.05) [10.30–34.51] 1.55 (1.94) [0.78–3.85] 2.11 (2.04) [1.14–3.49] 0.15 (0.20) [0.08–0.37] 0.007 (0.007) [0.002–0.015] 0.001 (0.001) [<0.001–0.002] 23.25 (22.09) [11.45–38.02] 1.93 (2.14) [0.87–4.22] Auto-disable N&S 0 (0) 0 (0) 0.70 (0.65) [0.25–1.30] 0.05 (0.06) [0.02–0.14] 0.007 (0.007) [0.002–0.015] 0.001 (0.001) [<0.001–0.002] 0.71 (0.65) [0.25–1.31] 0.06 (0.06) [0.02–0.14] Manual-shielding N&S 10.57 (9.54) [3.56–19.42] 0.77 (0.92) [0.29–2.03] 0.14 (0.14) [0.07–0.26] 0.01 (0.01) [0.005–0.03] 0 (0) 0 (0) 10.71 (9.68) [3.63–19.69] 0.89 (0.94) [0.29–2.05] Auto-shielding N&S 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) Reusable-nozzle jet injector 0.70 (0.89) [0.22–2.34] 0.05 (0.08) [0.02–0.25] 0 (0) 0 (0) 0 (0) 0 (0) 0.70 (0.89) [0.22–2.34] 0.06 (0.08) [0.02–0.25] Disposable-cartridge jet injector 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) a Estimated costs attributable to each injection device calculated as follows: (expected new cases of HBV [or HIV] infection) x (direct medical care costs for HBV [or HIV] + lost productivity costs for HBV [or HIV]. In each cell, the figure on the left is the base case-estimate. b All-routes totals vary slightly from combined component amounts due to rounding in route-of-transmission columns. c N&S = needle and syringe. d Figures in parentheses are medians (50th percentiles) of the sensitivity analysis using 1000 Monte Carlo simulations, with input parameters in Table 1 set to triangular probability distributions. e Figures in square brackets are ranges bound by the 5th and 95th percentiles of the sensitivity analysis using 1000 Monte Carlo simulations, with input parameters in Table 1 set to triangular probability distributions. 865Bulletin of the World Health Organization 2002, 80 (11) Risks and costs of injection devices in sub-Saharan Africa 866 Bulletin of the World Health Organization 2002, 80 (11) Research the expense of purchasing empty cartridges. One such prefilled cartridge was successfully pilot tested in both industrialized and developing countries (81–84), but its further development was halted for unspecified reasons. Needlestick injuries have been a focus of concern in both developing (29) and industrialized countries (12–14, 85, 86). In the USA, occupational safety regulations now require safer injection devices, such as the needle-shielding syringes and needle-free injectors we modelled (87–89). But needle- shielding syringes remain too expensive (modelled at US$ 0.54) for developing countries. Future needle-free vaccine technologies, such as mucosal (90) or transcutaneous (91, 92) immunization, would avoid the dangers of injection. However, they will probably takemany years to be registered in developed countries and their costs may put them out of reach of developing countries for decades. The hidden disease and economic cost of unsafe injections are enormous. Health ministries in sub-Saharan Africa, and the international agencies and initiatives that promote immunization and therapeutic injections should recognize this burden. To rephrase Hippocrates’ Epidemics, in selecting injection technology, one should ‘‘do less harm’’. n Acknowledgements This research was supported in part by Dr Ekwueme’s appointment to the Research Participation Program at CDC, administered by the Oak Ridge Institute for Science and Education through an interagency agreement between the United States Department of Energy and CDC, and by a consultation on behalf of the World Health Organization. He gratefully acknowledges the invaluable contributions and assistance in the collection of original data from: Coˆte d’Ivoire (Kakau Aka, Mami Pieorno Aby-Sy, Timothy Herrick, Alassane Mahassadi, Ouattara Siguifota, Kone Souleymane, Mamadou Moustapha Sy and others from Baptist Hospital Ferke and SakassouGeneral Hospitals, the University Hospital Centers of Treichville and Yopougon, and WHO-Coˆte d’Ivoire); Ghana (Ernestina Agyepong, F. Avokey, Martin P. Mandara, AnthonyNsaiah-Asare, PaapaObimpeh, Andrew Seidu Korkor, and others at the Ghanaian Ministry of Health, UNICEF-Ghana, and WHO-Ghana); and Uganda (John Barenzi, Robert Basaza, Patrick Isoke, P.K. Kataaha, Tonny Mubiru, Grace Murindwa, Francis Omaswa, Rachel N. Ser- uyange, Susan Seruyange, Sam Zaramba, and others from the Ugandan Ministry of Health, UNICEF-Uganda, Ugandan Blood Bank Transfusion Service, and Mulago, Rubaga, and Mildmay hospitals). The authors thank HalimaDao, Frank Ferguson, Robert Harrington, Janine Jagger, Robert E. Jones, Mikko Lainejoki, Gordon Larsen, John Livengood, John Lloyd, Larry Petersen, Rohit Patel, John Stengel, and Michel Zaffran for data, assistance, and advice. We also acknowledge Hazel Dean, Martin Meltzer, Mark Miller, and Mary McCauley for their comments and suggestions on early drafts. Conflicts of interest: none declared. Re´sume´ Estimations tire´es de la mode´lisation du risque de transmission de maladies et du couˆt e´conomique lie´s a` sept dispositifs d’injection en Afrique subsaharienne Objectif Etudier et comparer sept types de dispositifs d’injection du point de vue du couˆt e´conomique et du risque de transmission iatroge´nique de germes a` diffusion he´matoge`ne en Afrique subsaharienne. Me´thodes Des hypothe`ses de risque et des mode`les de couˆts ont e´te´ e´tablis pour chaque dispositif de manie`re a` estimer le nombre de nouvelles infections par le virus de l’he´patite B (HBV) et le virus de l’immunode´ficience humaine (VIH) a` la suite d’une transmission d’un patient a` l’autre, d’un patient a` un agent de soins de sante´ et d’un patient a` la communaute´. Les couˆts d’achat et d’utilisation des dispositifs ont e´te´ tire´s des donne´es publie´es, tandis que les couˆts des soins me´dicaux directs et de la perte de productivite´ associe´e a` la maladie dans le cas des infections a` HBV et a` VIH ont e´te´ tire´s de donne´es recueillies en 1999 en Coˆte d’Ivoire, au Ghana et en Ouganda. Des analyses multivarie´es de sensibilite´ au moyen du mode`le de Monte Carlo ont permis de caracte´riser l’intervalle d’incertitude des parame`tres du mode`le. Les couˆts ont e´te´ additionne´s du double point de vue de la socie´te´ et du syste`me de soins de sante´. Re´sultats Les aiguilles et seringues reste´rilisables et jetables avaient le couˆt global le plus e´leve´ en ce qui concerne l’achat, l’utilisation et les maladies iatroge´niques, avec un couˆt socie´tal me´dian par injection de US $26,77 pour le mate´riel reste´rilisable et US $25,29 pour le mate´riel jetable. Les injecteurs sans aiguille a` cartouche jetable et les seringues a` dispositif automatique de protection de l’aiguille avaient le couˆt le plus faible, soit respectivement US $0,36 et US $0,80. Les injecteurs sans aiguille a` buse re´utilisable et les aiguilles et seringues autobloquantes avaient un couˆt interme´diaire, soit respective- ment US $0,80 et US $0,91 par injection. Conclusion Malgre´ leur couˆt nominal d’achat et d’utilisation, les aiguilles et seringues conventionnelles comportent un risque non visible mais important de maladie iatroge´nique. D’autres dispositifs d’injection utilisables pour les millions d’injections pratique´es chaque anne´e en Afrique subsaharienne seraient inte´ressants et devraient eˆtre examine´s par les responsables de l’e´laboration des politiques lors des de´cisions d’achat. Resumen Estimaciones basadas en modelos de los riesgos de transmisio´n de enfermedades y el costo econo´mico de siete dispositivos de inyeccio´n en el A´frica subsahariana Objetivo Investigar y comparar siete tipos de dispositivos de inyeccio´n en cuanto a su riesgo de infeccio´n iatroge´nica por pato´genos de transmisio´n hemato´gena y su costo econo´mico en el A´frica subsahariana. Me´todos Se elaboraron hipo´tesis de riesgos para cada dispositivo y modelos de costos para estimar el nu´mero de nuevas infecciones por los virus de la hepatitis B (VHB) y de la inmunodeficiencia humana (VIH) debidas a la transmisio´n entre pacientes, de paciente 867Bulletin of the World Health Organization 2002, 80 (11) Risks and costs of injection devices in sub-Saharan Africa a agente de salud, y de paciente a la comunidad. Los costos asociados a la compra y el uso de los dispositivos se calcularon a partir de informacio´n hallada en la literatura, mientras que los costos de la atencio´n me´dica directa y de la productividad perdida como consecuencia de las infecciones por el VHB y el VIH se basaron en datos reunidos en 1999 en Coˆte d’Ivoire, Ghana y Uganda. Los intervalos de incertidumbre de los para´metros del modelo se determinaron mediante ana´lisis de sensibilidad multifactoriales basados en el me´todo de Monte Carlo. Se sumaron los costos obtenidos desde la perspectiva tanto de la sociedad como de los contribuyentes al sistema de atencio´n de salud. Resultados Las agujas y las jeringas reesterilizables y desecha- bles se asociaron a los costos globales ma´s altos en lo que atan˜e a la compra, el uso y las enfermedades iatroge´nicas: medianas de US$ 26,77 y US$ 25,29, respectivamente, por inyeccio´n desde el punto de vista de la sociedad. Los costos ma´s bajos correspondie- ron a los inyectores sin aguja con cartucho desechable y las jeringas con proteccio´n automa´tica de la aguja: US$ 0,36 y US$ 0,80, respectivamente. Los inyectores de presio´n con boquilla reutiliza- bles y las agujas y jeringas no reutilizables obtuvieron resultados intermedios, con US$ 0,80 y US$ 0,91, respectivamente, por inyeccio´n. Conclusio´n A pesar de su costo nominal de adquisicio´n y uso, las agujas y las jeringas convencionales comportan una carga oculta pero enorme de enfermedades iatroge´nicas. El uso de dispositivos de inyeccio´n alternativos en los millones de inyecciones que se administran anualmente en el A´frica subsahariana serı´a una medida inestimable, que deberı´a ser tenida en cuenta por los formuladores de polı´ticas en las decisiones de compra. References 1. State of the world’s vaccines and immunization 1996. Geneva: World Health Organization; 1996. p. 1-161. WHO document WHO/GPV/96.04. 2. Simonsen L, Kane A, Lloyd J, Zaffran M, Kane M. Unsafe injections in the developing world and transmission of bloodborne pathogens: a review. Bulletin of the World Health Organization 1999;77:789-800. 3. Dicko M, Oni AQ, Ganivet S, Kone S, Pierre L, Jacquet B. Safety of immunization injections in Africa: not simply a problem of logistics. Bulletin of the World Health Organization 2000;78:163-9. 4. Aylward B, Kane M, McNair-Scott R, Hu DJ. Model-based estimates of the risk of human immunodeficiency virus and hepatitis B virus transmission through unsafe injections. International Journal of Epidemiology 1995;24:446-52. 5. Aylward B, Lloyd J, Zaffran M, McNair-Scott R, Evans P. Reducing the risk of unsafe injections in immunization programmes: financial and operational implications of various injection technologies. Bulletin of the World Health Organization 1995;73:531-40. 6. Brito GS, Chen RT, Stefano IC, Campos AM, Oselka G. The risk of transmission of HIV and other blood-born diseases via jet injectors during immunization mass campaigns in Brazil [abstract no. PC-0132]. 10th International Conference on AIDS; 1994 Aug 7-12; Yokohama, Japan. AIDSLINE 10:301. Available from: URL: http://www.aegis.org/pubs/aidsline/1994/dec/ m94c3258.html (accessed on 22 March 2002). 7. Changing needles but not the syringe: an unsafe practice. Weekly Epidemiological Record 1987;62:345-52. 8. Greaves WL, Hinman AR, Facklam RR, Allman KC, Barrett CL, Stetler HC. Streptococcal abscesses following diphtheria-tetanus toxoid-pertussis vaccination. Pediatric Infectious Diseases 1982;1:388-90. 9. Soeters R, Aus C. Hazards of injectable therapy. Tropical Doctor 1989;19: 124-6. 10. Mann JM, Francis H, Davachi F, Baudoux P, Quinn TC, Nzilambi, et al. Risk factors for human immunodeficiency virus seropositivity among children 1–24 months old in Kinshasa, Zaire. Lancet 1986;2:654-7. 11. Lepage P, Van de Perre P. Nosocomial transmission of HIV in Africa: What tribute is paid to contaminated blood transfusions and medical injections? Infection Control and Hospital Epidemiology 1988;9:200-3. 12. Jagger J, Hunt EH, Brand-Elnaggar J, Pearson RD. Rates of needlestick injury caused by various devices in a university hospital. New England Journal of Medicine 1988;319:284-8. 13. Jagger J, Hunt EH, Pearson RD. Estimated cost of needlestick injuries for six major needled devices. Infection Control and Hospital Epidemiology 1990;11:584-8. 14. Armstrong SE. The cost of needlestick injuries: The impact of safer medical devices. Nursing Economics 1991;9:426-30. 15. Sagoe-Moses C, Pearson RD, Perry J, Jagger J. Risks to health care workers in developing countries. New England Journal of Medicine 2001;345:538-41. 16. Kane M, Clements J, Hu D. Hepatitis B. In: Jamison DT, Mosley HW, Measham AR, Robadilla JL, editors. Disease control priorities in developing countries. New York: Oxford University Press; 1993. p. 321-30. 17. Report on the global HIV/AIDS epidemic. Geneva: Joint United Nations Programme on HIV/AIDS (UNAIDS); 2000. pp. 1-137. UNAIDS document UNAIDS/00.13E. Available from: URL: http://www.unaids.org/epidemic_ update/report/Epi_report.pdf (accessed on 23 March 2002). 18. Hu DJ, Kane MA, Heymann DL. Transmission of HIV, hepatitis B virus, and other bloodborne pathogens in health care settings: a review of risk factors and guidelines for prevention. Bulletin of the World Health Organization 1991;69:623-30. 19. Safe injection — vital to health. Geneva: World Health Organization; 1999. Available from: URL: http://www.who.int/vaccines-access/injection_safety/ Injections_Safety/vital_to_Health/ (accessed on 23 March 2002). 20. Zaffran M, Lloyd J, Clements J, Stilwell B. A drive to safer injections. Geneva: World Health Organization; 1997. WHO document WHO/GPV/SAGE.97/ WP.05. 21. Progress towards global measles control and elimination, 1990–1996. Morbidity and Mortality Weekly Report 1997;46:893-7. 22. World Health Organization, United Nations Children’s Fund. WHO-UNICEF policy statement for mass immunization campaigns. Geneva: World Health Organization; 1997. p. 1-4. WHO document WHO/EPI/LHIS/97.04 Rev.1. 23. World Health Organization, United Nations Children’s Fund, United Nations Population Fund. Safety of injections: WHO-UNICEF-UNFPA joint statement on the use of auto-disable syringes in immunization services. Geneva: World Health Organization; 1999. p. 1-4. WHO document WHO/V&B/99.25. Available from: URL: http://www.who.int/vaccines-documents/DocsPDF99/ www9948.pdf (accessed on 23 March 2002). 24. GPV declares war on unsafe injections. Vaccine and Immunization News 1997;5:1-5. Geneva: World Health Organization. WHO document GPV/VIN/ 97.03. Available from: URL: http://www.who.int/vaccines-documents/ DoxNews/pdf-news/e_vin05.pdf (accessed on 23 March 2002). 25. Lloyd JS, Milstien JB. Auto-disable syringes for immunization: issues in technology transfer. Bulletin of the World Health Organization 1999;77:1001-7. 26. Kane A, Lloyd J, Zaffran M, Simonsen L, Kane M. Transmission of hepatitis B, hepatitis C and human immunodeficiency viruses through unsafe injections in the developing world: model-based regional estimates. Bulletin of the World Health Organization 1999;77:801-7. 27. Owens DK, Nease RF. Occupational exposure to human immunodeficiency virus and hepatitis B virus: A comparative analysis of risk. American Journal of Medicine 1992;92:503-12. 28. Nelson CM, Sutanto A, Suradana IG. Use of SoloShot autodestruct syringes compared with disposable syringes, in a national immunization campaign in Indonesia. Bulletin of the World Health Organization 1999;77:29-33. 29. Steinglass R, Boyd D, Grabowsky M, Laghari AG, Khan MA, Qavi A, et al. Safety, effectiveness and ease of use of a non-reusable syringe in a developing country immunization programme. Bulletin of the World Health Organization 1995;73:57-63. 868 Bulletin of the World Health Organization 2002, 80 (11) Research 30. Device achieves breakthrough in battle against spread of AIDS, other blood- borne diseases [press release]. Little Elm (TX): Retractable Technologies Inc.; 1997. 31. Younger B, Hunt EH, Robinson C, McLemore C. Impact of a shielded safety syringe on needlestick injuries among health-care workers. Infection Control and Hospital Epidemiology 1992;13:349-53. 32. Coursaget P, Kane MA. Overview of clinical studies in developing countries. In: Ellis RW, editor. Hepatitis B vaccines in clinical practice. New York: Marcel Dekker; 1993. p. 209-28. 33. Ekwueme DU. Comparative cost analysis of injection delivery technologies in sub-Saharan Africa, 28 December 1999. p. 1-128. Consultant’s report submitted to the World Health Organization, 1999. 34. Charney W. Retractable safety syringe activation study. Journal of Health Care Safety, Compliance and Infection Control 1998;2:413-5. 35. Canter J, Mackey K, Good LS, Roberto RR, Chin J, Bond WW, et al. An outbreak of hepatitis B associated with jet injections in a weight reduction clinic. Archives of Internal Medicine 1990;150:1923-7. 36. Hoffman PN, Abuknesha RA, Andrews NJ, Samuel D, Lloyd JS. A model to assess the infection potential of jet injectors used in mass immunisation. Vaccine 2001;19:4020-7. 37. Grabowsky M, Hadler SC, Chen RT, Bond WW, De Brito G. Risk of transmission of hepatitis B virus or human immunodeficiency virus from jet injectors and from needles and syringes. Atlanta (GA): Centers for Disease Control and Prevention, 1994. Unpublished document. 38. Grady GF, Lee VA, Prince AM, Gitnick GL, Fawaz KA, Vyas GN, et al. Hepatitis B immune globulin for accidental exposures among medical personnel: final report of a multicenter controlled trial. Journal of Infectious Diseases 1978;138:625-38. 39. Seeff LB, Wright EC, Zimmerman HJ. Type B hepatitis after needlestick exposure: prevention with hepatitis B immune globulin. Annals of Internal Medicine 1978;88:285-93. 40. Recommendations for preventing transmission of human immunodeficiency virus and hepatitis B virus to patients during exposure-prone invasive procedures. Morbidity and Mortality Weekly Report Recommendations and Reports 1991;40(RR-8):1-9. 41. Bell DM, Shapiro CN, Culver DH, Martone WJ, Curran JW, Hughes JM. Risk of hepatitis B and human immunodeficiency virus transmission to a patient from an infected surgeon due to percutaneous injury during an invasive procedure: estimates based on a model. Infectious Agents and Disease 1992;1:263-9. 42. Culver J. Preventing transmission of blood-borne pathogens: a compelling argument for effective device-selection strategies. American Journal of Infection Control 1997;25:430-3. 43. Henderson DK, Fahey BJ, Willy M, Schmitt JM, Carey K, Koziol DE, et al. Risk for occupational transmission of human immunodeficiency virus type 1 (HIV-1) associated with clinical exposures. Annals of Internal Medicine 1990;113:740-6. 44. Supply catalogue. Copenhagen: United Nations Children’s Fund; 1997. 45. World Health Organization, United Nations Children’s Fund. EPI product information sheets: 1997 edition. Geneva: World Health Organization. WHO document WHO/EPI/LHIS/97.01. 46. World Health Organization, United Nations Children’s Fund. EPI Product information sheets: 1998 supplement. Geneva: World Health Organization. WHO document WHO/EPI/LHIS/98.03. 47. EPI syringe and sterilizer study. Geneva: World Health Organization; 1985. p. 1-28. WHO document EPI/CCIS/85.6. 48. Choice of syringes for the EPI. Weekly Epidemiological Record 1986;61:41-3. 49. Keystone Dental Group. Ped-O-Jet January 1997 price list. Cherry Hill (NJ): Ped-O-Jet International, Keystone Industries; 1997. p. 1-4. 50. Cost analysis of various vaccine delivery systems. In: HealthTech Product Development Plan - MEDiVAXO ` low-workload jet injector. Seattle (WA): Program for Appropriate Technology in Health; 1996. Unpublished document. 51. Department of Vaccines and Biologicals. Assessing the whole cost of injection technologies: case studies in Africa. In: Technet consultation, Harare, 6-10 December 1999. Geneva: World Health Organization, 2000. WHO document WHO/ATT/TECHNET.99/Session 3/WP.3. Available from: URL: http://www.who.int/vaccines-documents/DocsPDF00/www536.pdf (accessed on 23 March 2002). 52. Drummond MF, O’Brien BJ, Stoddart GL, Torrance GW. Methods for the economic evaluation of health care programmes. 2nd ed. Oxford: Oxford University Press; 1997. 53. Over M, Bertozzi S, Chin J. Guidelines for rapid estimation of the direct and indirect costs of HIV infection in a developing country. Health Policy 1989;11:169-86. 54. Mansergh G, Haddix AC, Steketee RW, et al. Cost-effectiveness of short-course zidovudine to prevent perinatal HIV type 1 infection in a sub-Saharan African developing country setting. JAMA 1996;276:139-45. 55. The world factbook 2001. Virginia: McLean (VA): Central Intelligence Agency; 2001. Available from: URL: http://www.cia.gov/cia/publications/factbook/ (accessed on 30 May 2002). 56. Critchfield GC, Willard KE. Probabilistic analysis of decision trees using Monte Carlo simulation. Medical Decision Making 1986;6:85-92. 57. Dobilet P. Begg CB, Weinstein MC, Braun P, McNeil BJ. Probabilistic sensitivity analysis using Monte Carlo simulation: a practical approach. Medical Decision Making 1985;5:157-77. 58. Dittus RS, Roberts SD, Wilson JR. Quantifying uncertainty in medical decisions. Journal of the American College of Cardiology 1989;14(3 Suppl A):23A-28A. 59. Evans M, Hastings N, Peacock B. Statistical distributions. 2nd ed. New York: John Wiley and Sons; 1993. 60. Oakshott L. Business modeling and simulation. Washington (DC): Pitman Publishing; 1997. 61. Palisade Corporation. Guide to using @RISK: risk analysis and simulation add- in for Microsoft Excel or Lotus 1-2-3, Windows version. Newfield (NY): Palisade Corporation; 2000. 62. World Bank. Health expenditure, services and use. In: World development indicators 2000. Washington (DC): World Bank; 2000. p. 92. 63. Patient exposures to HIV during nuclear medicine procedures. Morbidity and Mortality Weekly Report 1992;41:575-8. 64. Davachi F, Baudoux P, Ndoko K, N’Galy B, Mann J. The economic impact on families of children with AIDS in Kinshasa, Zaire. In: Fleming A, Carballo M, FitzSimons DW, Bailey MR, Mann J, editors. The global impact of AIDS. New York: Alan R. Liss; 1988. p. 167-9. 65. Foster S. The economic impact of AIDS in Zambia. Washington (DC): World Bank; 1993. Unpublished paper prepared for the Human Resources and Poverty Division. 66. Hersh BS, Popovici F, Apetrei RC, Zolotusca L, Beldescu N, Calomfirescu, A, et al. Acquired immunodeficiency syndrome in Romania. Lancet 1991; 338:645-9. 67. Hersh BS, Popovici F, Jezek Z, Satten GA, Apetrei RC, Beldescu N, et al. Risk factors for HIV infection among abandoned Romanian children. AIDS 1993;7:1617-24. 68. Feilden R. Final report: Choosing a sustainable policy for giving safe injections in Romania. New York: United Nations Children’s Fund; 1993. Unpublished document. 69. Holding R, Carlsen W. Deadly needles: fast track to global disaster. San Francisco Chronicle 1998 Oct 27. p. A-1 et seq. Available from: URL: http:// sfgate.com/cgi-bin/article.cgi?file=/chronicle/archive/1998/10/27/ MN52NEE.DTL (accessed on 24 March 2002). 70. Hutin YJF, Harpaz R, Drobeniuc J, Melnic A, Ray C, Favorov M, et al. Injections given in healthcare settings as a major source of acute hepatitis B in Moldova. International Journal of Epidemiology 1999;28:782-6. 71. Spiegel A, Moren A, Varaine F, Baudon D, Rey M. Aspects e´pide´miologiques et controˆle des e´pide´mies de me´ningite a` me´ningocoque en Afrique [Epidemiologic aspects and control of meningococcal meningitis epidemics in Africa]. Cahiers Sante´ 1994;4:231-6. In French. 72. Artus JC. Vaccination de masse par le vaccin souche Rockefeller 17D au Se´ne´gal. Utilisation des ‘‘Ped-O-Jet’’ [Mass vaccination with the Rockefeller 17D strain in Senegal. Use of the ‘‘Ped-O-Jet’’]. Me´decine Tropicale 1966;26:527-36. In French. 73. Reis EC, Jacobson RM, Tarbell S, Weniger BG. Taking the sting out of shots: control of vaccination-associated pain and adverse reactions. Pediatric Annals 1998;27:375-86. 74. Hepatitis B associated with jet gun injection — California. Morbidity and Mortality Weekly Report 1986;35:373-6. 75. Transmission of hepatitis B associated with jet gun injection. Weekly Epidemiological Record 1986;61:309-11. 76. United States Department of Defense. Automatic jet hypodermic injection units/ withdrawal (DPSC 970147). Fort Detrick (MD): United States Army Medical Material Agency; 1997. Medical material quality control document MMQC- 97-1169. Available from: URL: http://review.detrick.army.mil/usamma/ mmqc_messages/Q971169.txt (accessed on 27 March 2002). 869Bulletin of the World Health Organization 2002, 80 (11) Risks and costs of injection devices in sub-Saharan Africa 77. Centers for Disease Control and Prevention (CDC). Needle-free injection technology: policies (WHO, CDC, and DoD). Atlanta (GA): National Immunization Program, CDC. Available from: URL: http://www.cdc.gov/nip/ dev/jetinject.htm#policies (accessed on 4 March 2002). 78. Safety of injections in immunization programmes: WHO recommended policy. Geneva: World Health Organization; 1998. p. 1-11. WHO document WHO/ EPI/LHIS/96.05 Rev.1. Available from: URL: http://www.who.int/gpv- documents/DocsPDF/www9665.pdf (accessed on 27 March 2002). 79. Centres for Disease Control and Prevention. Jet injectors. In: General recommendations on immunization. Recommendations of the Advisory Committee on Immunization Practices (ACIP). Morbidity and Mortality Weekly Report, Recommendations and Reports 1994;43(RR-1):1-38. 80. Centres for Disease Control and Prevention. Jet injectors. In: General recommendations on immunization. Recommendations of the Advisory Committee on Immunization Practices (ACIP) and the American Academy of Family Physicians (AAFP). Morbidity and Mortality Weekly Report, Recom- mendations and Reports 2002;51(No. RR-2):1–38. 81. Galy M, Genet A, Saliou P. Un progre`s dans le domaine de l’injection sans aiguille: le syste`me Imulet [Progress in the field of needle-free injection: the Imulet system]. STP Pharma Pratiques 1992;4:261-6. In French. 82. Fisch A, Cadilhac P, Vidor E, Prazuck T, Dublanchet A, Lafaix C. Immunogenicity and safety of a new inactivated hepatitis A vaccine: a clinical trial with comparison of administration route. Vaccine 1996;14:1132-6. 83. Parent du Chaˆtelet I, Lang J, Schlumberger M, Vidor E, Soula G, Genet A, et al. Clinical immunogenicity and tolerance studies of liquid vaccines delivered by jet-injector and a new single-use cartridge (Imulet): comparison with standard syringe injection. Vaccine 1997;15:449-58. 84. Schlumberger M, Parent du Chaˆtelet I, Lafarge H, Geneˆt A, Gaye AB, Monnereau A, et al, Couˆt de l’injection d’anatoxine te´tanique par injecteur sans aiguille (Imulee) lors d’une vaccination collective au Se´ne´gal : comparaison avec l’injection par seringues et aiguilles reste´rilisables [Cost of tetanus toxoid injection by needle-free injector (Imulee) in a mass vaccination in Senegal: comparison with resterilizable syringes and needles]. Cahiers Sante´ 1999;9: 319-26. In French. 85. Laufer FN, Chiarello LA. Application of cost-effectiveness methodology to the consideration of needlestick-prevention technology. American Journal of Infection Control 1994;22:75-82. 86. Holding R, Carlsen W. Epidemic ravages caregivers. Thousands die from diseases contracted through needlesticks. San Francisco Chronicle 1998 April 13. p. A-1 et seq. Available from: URL: http://www.sfgate.com/cgi-bin/ article.cgi?file=/chronicle/archive/1998/04/13/MN64658.DTL (accessed on 24 March 2002). 87. International Health Care Worker Safety Center. State legislation on needle safety, updated 01/24/02. Charlottesville (VA): University of Virginia Health System; 2002. Available from: URL: http://www.med.virginia.edu/medcntr/ centers/epinet/statelist.html (accessed on 24 March 2002). 88. Voelker R. Needlestick bill passes [Quick uptakes]. JAMA 2000;284:2585. 89. Occupational Safety and Health Administration. Occupational exposure to bloodborne pathogens; needlestick and other sharps injuries; final rule. Washington (DC): Federal Register; 2001. 29 CFR Part 1910. 66:5318-25. Available from: URL: http://www.osha-slc.gov/FedReg_osha_pdf/FE- D20010118A.pdf (accessed on 27 March 2002). 90. Levine MM, Dougan G. Optimism over vaccines administered via mucosal surfaces. Lancet 1998;351:1375-6. 91. Tang D-C, Shi Z, Curiel DT. Vaccination onto bare skin. Nature 1997;388: 729-30. 92. Guerena-Burgueno F, Hall ER, Taylor DN, Cassels FJ, Scott DA, Wolf MK, et al. Safety and immunogenicity of a prototype enterotoxigenic Escherichia coli vaccine administered transcutaneously. Infection and Immunity. 2002; 70:1874-80. 870 Bulletin of the World Health Organization 2002, 80 (11) Research Annex Box A. Examples of commercial devices or prototypes modelled in the study Auto-disable needle and syringe (N&S) Examples of auto-disabling, single-use needles permanently fixed to plastic syringes designed to prevent inadvertent or intentional reuse that meet WHO, the United Nations Children’s Fund (UNICEF), and the United Nations Population Fund (UNFPA) criteria for safe injections in developing countries are: DestroJectt (DestroJect GmbH Medical Devices, Neumu¨nster, Germany); SoloShote (28, 29) (Becton-Dickinson and Co., Franklin Lakes, NJ, USA); and UNIVEC Rx Ultrae (Univec Inc., Farmingdale, NY, USA), among others. Generic products are also available from: Atlas Medical Resources Corp. Inc., Ottawa, Ontario, Canada; Com Pro, Paris, France; and Pharmaplan GmbH, Bad Homburg, Germany, among others. UniJecte is an equivalent auto-disable device containing prefilled vaccine in a plastic blister with a fixed needle (Becton-Dickinson and Co.). An updated list of auto-disable syringe manufacturers whose products are preliminarily approved by WHO and UNICEF is available from: URL: http://www.who.int/vaccines-access/injection_safety/Injections_Safety/Injection_Technology/ADsyringes_manu.html (accessed on 17 March 2002). Manual-shielding N&S The manually activated, needle-shielding N&S for the prevention of needlestick injuries modelled in the study was the VanishPointt (Retractable Technologies Inc., Little Elm, TX, USA) (30). Similar devices are the SafetyGlidee and Safety-Loke (Becton-Dickinson and Co.), MonoJecte Safety Syringes (31) (Sherwood-Davis and Geck, St. Louis, MO, USA), and Needle-Prot (Portex, Inc., Keene, NH, USA), among others. All these devices require the health worker to perform an additional step (e.g. further depressing the plunger after injection is completed) in order to shield the needle and disable the device. A database maintained by the State of California, USA, of needle-shielding devices and manufacturers is available from: URL: http:// www.dhs.ca.gov/ohb/sharps/disclaim.htm (accessed on 17 March 2002). Reusable-nozzle jet injector The multiple-use nozzle jet injector devices modelled were the Ped-O-Jett (Keystone Industries, Cherry Hill, NJ, USA) and the identical Am-O-Jete (American Jet Injector, Lansdale, PA, USA). Similar devices include the Med-E-Jett (Evans Enterprises, Mayfield Heights, OH, USA), DermoJett (Socie´te´ AKRA DermoJet, Pau, France), Im-O-Jett (Aventis Pasteur, formerly Institut Me´rieux, Lyon, France), and BI-100e (Felton International, Lenexa, KS, USA, and CADB/Medequip, Voronezh, Russian Federation). Disposable-cartridge jet injector The disposable-cartridge needle-free jet injector modelled represented a composite of the features of various marketed and investigational devices, including the Biojector 2000t (Bioject Inc., Portland, OR, USA); the INJEXe (Equidyne Systems Inc., San Diego, CA, USA); the MEDiVAXe (Vitajet Corporation and Program for Appropriate Technology in Health, Seattle, WA, USA); the SensaJete (Genesis Medical Technologies Inc., Denver, CO, USA); and the LectraJete (DCI Inc., East Syracuse, NY, USA). A list maintained by the Centers for Disease Control and Prevention, Atlanta, GA, USA, of both reusable-nozzle and disposable-cartridge jet injectors is available from: URL: http://www.cdc.gov/nip/dev/jetinject.htm#devices (accessed on 17 March 2002). ABulletin of the World Health Organization 2002, 80 (11) Annex Annex Table A. Purchase and usage cost estimatesa for seven vaccine delivery technologies Device typeb and component costs Cost per unit No. of injections Cost per (US$) per unit injection (US$) Resterilizable needle and syringe (N&S) Glass or (sterilizable) plastic syringe purchasec 0.730 100 0.007 30 Needle purchased 0.050 45 0.001 11 Sterilizer purchasee 107.760 50 000 0.002 16 Hard-water pad for sterilizerf 20.730 25 000 0.000 83 Spare parts for sterilizerg 14.420 15 000 0.000 96 Steam sterilization indicatorsh 261.190 750 000 0.000 35 Cleaning and sterilizationi 0.016 1 0.016 00 Resharpening needlej 0.075 10 0.007 50 Safety box for used N&Sk 0.966 10 000 0.000 10 Incineration of safety boxl 0.850 10 000 0.000 09 Labourm 1.500 45 0.033 33 Total cost Base case 0.069 72 Lower estimaten 0.017 43 Upper estimaten 0.139 44 Disposable N&S Disposable syringe purchaseo 0.039 1 0.039 40 Disposable needle purchasep 0.019 1 0.019 00 Safety box for used N&Sk 0.966 100 0.009 66 Incineration of safety boxl 0.850 100 0.008 50 Labourm 1.500 60 0.025 00 Total cost Base case 0.101 56 Lower estimaten 0.025 39 Upper estimaten 0.203 12 Auto-disable N&S Auto-disable N&S purchaseq 0.093 1 0.092 50 Safety box for used N&Sk 0.966 100 0.009 66 Incineration of safety boxl 0.850 100 0.008 50 Labourm 1.500 60 0.025 00 Total cost Base case 0.135 66 Lower estimaten 0.033 92 Upper estimaten 0.271 32 Manual-shielding N&S and Auto-shielding N&S Device purchaser 0.500 1 0.500 00 Safety box for used N&Sk 0.966 100 0.009 66 Incineration of safety boxl 0.850 100 0.008 50 Labourm 1.500 60 0.025 00 Total cost Base case 0.543 16 Lower estimaten 0.135 79 Upper estimaten 1.086 32 Reusable-nozzle jet injector Device purchases 2 300.000 1 000 000 0.002 30 Spare parts kitt 275.000 50 000 0.005 50 Cleaning and sterilizationu 3.250 1 000 0.003 25 Routine maintenancev 0.250 200 0.001 25 Overhaulw 153.000 100 000 0.001 53 Trainingx 45.000 100 000 0.000 45 Three-dose vaccine waste on purgey 0.900 60 0.015 00 Labourz 1.500 150 0.010 00 Total cost Base case 0.039 28 Lower estimaten 0.009 82 Upper estimaten 0.078 56 B Bulletin of the World Health Organization 2002, 80 (11) Risks and costs of injection devices in sub-Saharan Africa Device typeb and component costs Cost per unit No. of injections Cost per (US$) per unit injection (US$) Disposable-cartridge jet injector Device purchaseaa 250.000 25 000 0.010 00 Maintenance and cleaningbb 3.000 200 0.015 00 Safety box for used cartridgesk 0.966 400 0.002 42 Incineration of safety boxl 0.850 400 0.002 13 Labourcc 1.500 60 0.025 00 Vaccine vial transfer devicesdd 0.550 10 0.055 00 Disposable cartridgesee 0.250 1 0.250 00 Total cost Base case 0.359 54 Lower estimaten 0.089 89 Upper estimaten 0.719 08 a Lower and upper estimates of costs were varied from 25% to 200% of the base-case calculation. Sources: ref. 44–46, unless indicated otherwise. b See Box 1 for classification and description of device types. c Resterilizable syringe modelled: glass, 2 ml capacity (United Nations Children’s Fund (UNICEF) catalogue item no. 078 3500). An average of 100 uses were assumed before breakage or disposal (47). d Resterilizable needle modelled: stainless steel, 0.7 mm x 32 mm (UNICEF catalogue item no. 075 0500). An average of 45 uses were assumed before disposal (47). e Sterilizer modelled: double rack, steam pressure, fuel, kit B (with accessories), 84-syringe/100-needle capacity (UNICEF catalogue item no. 990 8100). Assumed a useful life of 10 years, 100 sterilizations per year at 60% of capacity (50 syringes). Useful life estimate source: M. Lainejoki, J. Lloyd, M. Zaffran, personal communications, 1998, 1999. f Hard-water pad modelled: item 015 7118. Average 5-year life assumed, 100 sterilizations per year at 60% of capacity (50 syringes). Useful life estimate source: M. Lainejoki, M. Zaffran, personal communications, 1998. g Replacement parts modelled: gasket (UNICEF catalogue item no. 015 7115, US$ 4.01), handle (UNICEF catalogue item no. 015 7117, US$ 3.10), pressure valve (UNICEF catalogue item no. 015 7113, US$ 4.36), safety valve (UNICEF catalogue item no. 015 7116, US$ 2.95). Average 3-year life assumed, 100 sterilizations per year at 60% of capacity (50 syringes). h Sterilization confirmation indicators modelled: Temperature-Steam-Time (TST) indicator spots, at US$ 261.19 for 50 boxes with 300 spots per box (plus record sheets), assuming one test spot per load at 60% of capacity (50 syringes). i Includes cost of disinfectants for cleaning and fuel for sterilizer. Cost estimate source: ref. 48. j Assumed sharpening required for 3 minutes (at US$ 1.50/hour labour) after every 10 injections. Time estimate source: ref. 50. k Sharps box modelled: UNICEF catalogue item no. 078 2208, capacity 100 syringes (or 400 disposable needle-free cartridges of approximately 25% of size of average N&S). Resterilizable syringes deposited after average of 100 injections each; other syringes and needle-free cartridges deposited after being used once. l Assumed cost of US$ 0.85 to incinerate one safety box. Cost source: ref. 5. m Assumed 80 seconds to assemble, fill, and administer each injection for resterilizable syringes (45/hour at US$ 1.50/hour labour), 60 seconds for disposable, auto-disable, and retractable-needle syringes (60/hour). Time estimates source: ref. 50. n Lower and upper estimates used in Monte Carlo sensitivity analyses. o Disposable syringe modelled: plastic, 5 ml capacity (UNICEF catalogue item no. 078 2405, US$ 3.94 per 100). p Disposable needle modelled: 0.7 mm x 32 mm (UNICEF catalogue item no. 074 7440). q Auto-disable N&S device modelled: 0.5 ml capacity, fixed needle (UNICEF catalogue item no. 078 2207). r Retractable needle syringe modelled: VanishPointt (see Annex Box A) (30). s Reusable-nozzle jet injector modelled: Ped-O-Jett, model POJ (foot-operated) unit (49). Assumed 50 days per year of use for mass immunization campaigns, 1000 vaccinations per day, device lifespan 20 years. t Spare parts kit no. J6000A includes sufficient replacement seals, springs, valve balls, and other wearable items needed for field maintenance during 50 000 injections (one year) (49 ; J. Stengel, R. Harrington, personal communications, 1998). u Assumed 1.5 hours per day of 1000 vaccinations (at US$ 1.50/hour labour) for device disassembly, sterilization, and reassembly, plus US$ 1.00 for autoclave sterilization, including fuel. v Assumed maintenance of 10 minutes per 200 vaccinations for clearing clogged nozzles, freeing jammed check valves, etc. (at US$ 1.50/hour labour). Time estimate source: ref. 50. Not included are costs of a second backup injector kept on hand to keep a mass vaccination programme on schedule in case of failure of the primary device. w Assumed major rebuild overhaul every 100 000 injections (2 years) by trained national technician (1 hour at US$ 3.00/hour labour), plus average US$ 150.00 cost of replacement parts and factory shipping. x Assumed training time of 10 hours each for trainee (at US$ 1.50/hour labour) and trainer (at US$ 3.00/hour labour), required biannually (every 100 000 injections). Time estimate source: ref. 50. y Reusable-nozzle jet injectors feed vaccine from multi-dose vials into internal fluid chambers, and must be purged of air when changing vaccine vials and at end of vaccination session. Assumed loss of three doses (at US$ 0.30 each) for every 60 injections. z Assumed average 150 injections per operator per hour (at US$ 1.50 per/hour labour). aa Low workload, disposable-cartridge/nozzle jet injector modelled as a composite of various commercial and prototype models (see Methods and Annex Box A). Assumed 20 injections/day for 250 days per year for 5 years useful lifespan. Cost estimate source: ref. 50. bb Assumed 2 hours labour (at US$ 1.50/hour) required every 200 injections for routine maintenance and cleaning. cc Assumed 60 injections per operator per hour (at US$ 1.50 per/hour labour), to transfer vaccine from vial to cartridge, load injector, and administer injection. dd Disposable transfer device (or other system) attaches to multi-dose or unit-dose vials to measure and transfer vaccine into injection cartridge, without needing additional syringe or needle. Assumed transfer vial cost of the INJEXt vial adaptor (see Annex Box A), i.e. US$ 0.55 each (quantities above 5000; L. Petersen, personal communication, 1999). Assumed transfer device disposed after use with one 10-dose vial. Other injector systems may have a different means of effecting transfer into disposable cartridges, if not prefilled by vaccine manufacturer. ee Assumed disposable cartridges to be filled in clinic immediately before injection (instead of prefilling by vaccine manufacturer); base price set at actual price for INJEXt brand of ampoules (see Annex Box A), i.e. US$ 0.25 each for quantities exceeding 5000 (L. Petersen, personal communication, 1999). Reduction in price to the lower (25%) estimate of US$ 0.065 each is considered feasible if several millions of ampoules were used per year. CBulletin of the World Health Organization 2002, 80 (11) Annex, cont. Annex Table B. Input data and cost assumptions,a attributable to HBVb and HIVc disease, as used for base-case estimates of corresponding direct medical costsd and indirect (lost productivity)e costs in representative sub-Saharan African countries Costsf (US$) Coˆte d’Ivoire Ghana Uganda HBV HIV HBV HIV HBV HIV Direct medical costs Outpatient care Average cost of first doctor visit 7.02 7.02 0.59 0.85 3.04 6.92 Average cost of follow-up visits 3.51 2.81 0.28 0.34 1.82 2.74 Average number of follow-up visits per lifetime 4 4 3 4 4 4 Lab test(s) for HBsAgg or HIV 42.14 3.51 0.94 0.94 6.08 2.13 Total travel costs for care 2.81 2.81 0.47 0.47 2.13 2.13 Disease-related drugs Average cost of interferon for HBV, lifetimeh 7.94 NAi NA NA NA NA Average cost for 1 month’s drug(s) for HIVj NA 173.96 NA 308.99 NA 308.99 Average months/year on HIV therapyk NA 4.31 NA 1.50 NA 1.50 Average number of years on HIV therapy NA 5 NA 5 NA 5 Inpatient hospitalization Average cost per day hospitalized 27.29 35.68 1.84 2.58 6.34 15.60 Average days hospitalized per lifetime 7.0 20.0 9.5 21.0 8.4 14.4 Subtotal: lifetime direct costs, undiscountedl 265 4487 20.34 2375 71.88 2564 Total: lifetime direct costs, 3% discountedm, n 75.00 3616 6.24 1914 21.33 2067 Indirect (lost productivity) costs Average life expectancy from birth (in years) 51 51 57 57 48 48 Average age at death from disease acquired in infancy (in years) 43 5 40 5 41 5 Average annual earnings 3 259 3 259 79.40 79.40 1 105 1 105 Average unemployment rate (%) 10 10 20 20 20 20 Adjusted annual adult earnings 2 933 2 933 64 64 884 884 Subtotal: productivity loss, undiscountedo 23 463 109 545 1 080 2 696 6 187 30 495 Total: productivity loss, 3% discounted 5 814 43 779 248 1 014 1 664 12 595 a Data collected in late 1999 from original sources in the countries listed (33, 51). b HBV = hepatitis B virus. c HIV = human immunodeficiency virus. d Lifetime direct medical care costs = (initial outpatient visit cost) + ((follow-up visit cost) x (number of follow-up visits)) + (lab tests cost) + (travel costs) + [(lifetime HBV therapy cost) or (1-month HIV drug costs) x (months per year on HIV drugs) x (years on HIV drugs)] + [(daily inpatient hospital cost) x (average inpatient days)]. e Lifetime indirect (lost productivity) costs = sum of adjusted expected earning in future years between the average age of premature death from HBV or HIV and the life expectancy otherwise, adjusted for unemployment and age group. See Annex Box B for calculation formula. f Costs collected in local currencies have been converted to US$ at the following year 2000 exchange rates for US$ 1.00: Coˆte d’Ivoire, 712 CFA; Ghana, 5322 Cedis; Uganda, 1645 Ugandan shillings (55). g HBsAg = hepatitis B surface antigen. h Reported per-patient cost in Coˆte d’Ivoire of 1 130 400 CFA (US$ 1588) for interferon treatment for chronic HBV infection was reduced by the assumed proportion among all HBV-infected who become chronic carriers (10%) and then by the estimated proportion of such carriers who will receive such interferon later in life (5%). This yields an average cost of US$ 7.94 for each new HBV infection. i N/A = not applicable. j Assume average child weight of 18 kg and surface area of 0.72 m2. Coˆte d’Ivoire: paediatric triple-drug therapy modelled using available zidovudine (AZT, 200 mg/day), lamivudine (3TC, 140 mg/day) and indinavir (1080 mg/day), representing 40%, 48%, and 45% respectively of adult doses, and costing a proportionate US$ 173.96 of the total adult dosage costs of 286 290 CFA (US$ 402) per month. Ghana: HIV drugs available at the time only from Ugandan source; Ugandan costs applied. Uganda: same drugs and dosages assumed as in Coˆte d’Ivoire; zidovudine (three 200 ml bottles per month of 10 mg/ml syrup, costing 40 500 Ugandan shillings (US$ 24.62) per bottle); lamivudine (1.75 240 ml bottles per month of 10 mg/ml solution, costing 63 000 Ugandan shillings (US$ 38.30) per bottle); and indinavir (45% (US$ 168.10) of adult monthly dosage cost of 614 500 Ugandan shillings (US$ 373.56)). k Estimated average months per year for what ideally should be year-round, triple-drug, antiretroviral therapy. This reflected both intermittent receipt and non-receipt of drug therapy by substantial proportions of HIV-infected children. l Subtotals vary slightly from individual components listed due to rounding. m HBV direct costs are assumed incurred during the final year of life at the average age of (premature) death resulting from infection acquired from unsafe injection, and discounted back at 3% to the first year of life (year 0). Totals for each country produce an arithmetic mean amount of US$ 34.19 used in the model for each new HBV infection (see Table 1). n HIV direct costs discounted at 3% back to year 0 from the assumed probability of 10% per year of developing AIDS and dying within 12 months among those HIV-infected during the first year of life, using a discounting factor of 0.801 (53, 54). Totals for each country produce arithmetic mean amount of US$ 2533 used in the model for each new HIV infection (see Methods and Table 1). o HBV: adjusted annual adult earnings x number of years between average age of death due to HBV disease and average life expectancy of uninfected people. HIV: adjusted annual adult earnings x number of years from age 16 years to average life expectancy of uninfected people, plus 15% of annual adult earnings from the average age of death (6 years) to the age of 15 years. D Bulletin of the World Health Organization 2002, 80 (11) Risks and costs of injection devices in sub-Saharan Africa Annex Box B. Formula used for calculating indirect costs of iatrogenic HBVa or HIVb disease resulting from lost productivity T–1 L = 0.5 (Et Wgt (1+r) -t ) + S En Wgn (1+r) -n + 0.5 (ET WgT (1+r) -T) n = t+1 where: L = loss of earnings attributable to premature death resulting from HBV or HIV disease n = future year of age for which the value in the first year of life (age 0) is being determined t = assumed age of premature death, in years, for HBV or HIV infections acquired in infancy T = assumed average age of death (life expectancy at birth), in years En = average per capita annual earnings of employed adults in public and private sectors, adjusted for unemployment, in the year n (t, t +1,..., T-1, T). (The current year 0 value for earnings is assumed for each future year n.) Wgn = age-weighting factor for adjustment of earnings of individuals of age group g in the year of age n (Wgn = 0 when age group g is 0–5 years, Wgn = 0.15 when age group g is 6–15 years, Wgn = 1.0 when age group g is 16–50 years, Wgn = 0.85 when age group g is 51–65 years, Wgn = 0.25 when age group g is 566 years) r = assumed annual interest rate 3% for discounting the future value of money to the present time (1+r)-n = discount factor for calculating the present value of future earnings a HBV = hepatitis B virus. b HIV = human immunodeficiency virus. EBulletin of the World Health Organization 2002, 80 (11) Annex, cont.