Bangladesh, India (Uttar Pradesh), Mozambique, and Uganda were chosen to reflect various population sizes and urbanicity among developing countries in Africa and Asia (see Table 1). Session size data were collected from representative drug discovery facilities in the four countries. IPV wastage and associated costs were examined in this paper, though our model enables users to simulate different types of vaccines in various presentation and dose schedules. Our model
uses a 1-dose schedule for IPV. This study used data on session sizes to model populations from Bangladesh, India (Uttar Pradesh), Mozambique, and Uganda. The rural data from Bangladesh originated from four clinics in the Sunamganj district, consisting of one large outpatient clinic, two union health centers, and one subcenter. The urban data from Bangladesh came from three urban subcenters, two urban HC III clinics, and three large urban clinics (“HC” stands for “health center”). The number of pentavalent vaccine doses administered between January and December 2012 were counted at each session. For India, we collected data on the number of DPT doses administered in two HC III clinics in the Basti district of Uttar Pradesh from January to February 2012. There were no data available from urban clinics in Uttar Pradesh. The data from Mozambique came from 74 Centro Salud Rural (CSR) 1 sessions, 49 CSR2 sessions, as well as 45 outreach sessions Gefitinib cell line from the Inhambane district of Mozambique in 2012. The number of
children receiving a pentavalent vaccine each day was recorded. There were also no data available from urban clinics in Mozambique.
The Ugandan data originated from the Service Provision Assessment (SPA) Survey of 2007 that was collected by Macro International [14]. After weighting, the survey provided a national representative sample of all government health care facilities in Uganda. Data were collected by site inspections and health record review from 433 facilities providing immunization at HC-IIs, HC-IIIs, HC-IVs, rural hospital settings and urban settings. Bay 11-7085 The SPA survey had sampling weights for each type of facility, so one can produce estimates of the national count of each type of facility. The counts of daily children arriving in facilities in the SPA data were based on all children, not just children requesting immunization. The estimated number of facilities in each country relied on SPA data in Uganda [18], and Bangladesh [15]. Facility count estimates for Mozambique were extrapolated on a population basis from Inhambane province to all Mozambiquan provinces. Facility count estimates for India were confined to only rural Uttar Pradesh. In each country or region, the daily session size data for each clinic type was determined by fitting the parameters of various distributions. A maximum likelihood algorithm to find parameters that minimized the root mean squared error between the data and each candidate distribution was implemented in Palisades @Risk Version 6.