Time and cost associated with utilization of services at mobile health clinics among pregnant women

Background: Antenatal care (ANC) is provided for free in Tanzania in all public health facilities. Yet surveys suggested that long distances to the facilities limit women from accessing these services. Mobile health clinics (MHC) were introduced to address this problem; however, little is known about the client cost and time associated with utilizing ANC at MHC and whether these costs deter women from using the provided services. Methods: Client-exit interviews were conducted by interviewing 293 pregnant women who visited the MHC in rural Tanzania. Two subgroups were created, one with women who travelled more than 1.5 h to the MHC, and the other with women who travelled within 1.5 h. For each subgroup we estimated the direct cost in US$ and time in hours for utilizing services and they hinder service utilization. The Wilcoxon-Mann-Whitney rank sum test was performed to compare the differences between the estimated mean values in the two groups. Result: Total direct cost per visit was: US$2.27 (SD = 0.90) for overall, US$2.29 (SD = 1.03) for those women who travelled less than 1.5 h and US$2.53 (SD = 0.63) for those who travelled more than 1.5 h (p = 0.08). Laboratory and medicine cost accounted for 70 and 16% of the total direct cost and were similar across the groups. Total time cost per visit (in hours) was: 3.75 (SD = 1.83), 2.88 (SD = 1.27) for those women who travelled less than 1.5 h and 5.02 (SD = 1.81) for those who travelled more than 1.5 h (p < 0.01). The major contributor of time cost was waiting time; 1.89 (SD = 1.29) for overall, 1.68 (SD = 1.02) for those women who travelled less than 1.5 h and 2.17 (SD = 1.57) for those who travelled more than 1.5 h (p = 0.07). Participants reported having missed their scheduled visit due to lack of money (15%) and time (9%). Conclusion: Women receiving nominally free ANC incur considerable time and direct cost, which may result in an unsteady use of maternal care. Improving availability of essential medicine and supplies at health facilities, as well as focusing on efficient utilization of community health workers may reduce these costs. The study was carried out in the Kisarawe district in the coast region of Tanzania from November 2015 to June 2016. The district has a population of 101,598 out of which 50,967 are women and 25,779 are of reproductive age as estimated from the 2012 national population census [12]. The district belongs to the lesser developed parts of Tanzania and about 90% of its population lives in rural areas on subsistence farming [13]. The district has a total of 37 governmental dispensaries, eight private dispensaries and four health centres (including one non-governmental centre). The first level of care in the region is represented by dispensaries and health centres. MHC operates in 20 villages, which have been classified as “remote” by using certain criteria (e.g. populations residing more than 5 km from the health facility, or populations residing less than 5 km from the health facility which, however, lacks personnel and/or essential medical equipment would fall into that category) [3, 4]. We adopted a patient perspective to ascertain the cost of health care utilization. As the main aim of this study was the estimation of the client’s direct cost and time cost resulting from seeking a free ANC, it was justified to concentrate on service users rather than on doing a household survey. We used a convenient, non-probability sampling technique. Our subjects were enrolled according to their availability and accessibility. This method was selected because it is quick, inexpensive, and convenient. In our situation, the accessible population were pregnant women attending the antenatal clinic at the mobile health clinic in Kisarawe District. Therefore, within the study period, any woman who seek ANC at any mobile health clinic in Kisarawe District and who agreed to participate to the study met the eligibility criteria and were included in this study. Type, amount, and extent of the cost incurred by women were assessed by conducting key informant interviews (KII) with ten pregnant women who seek services at the mobile health clinic prior to the design of the interview guide. The information gathered from the KII together with information based on literature was used to create the structured questionnaire. The structured interview questionnaire (see Additional file 1) was administered to a total of 293 pregnant women attending the MHC from November 2015 to June 2016. Women signed an informed consent form (see Additional file 1), and all interviews were done in private with only interviewers and respondent being present. The questions focused on several aspects regarding cost and time spent: Time spent on travelling, waiting and consultation at the MHC, cash payments for services, travel, drugs and supply cost. Interview data on time spent on services was compared and verified by observing waiting and consultation time at the MHC. Information on travel cost was verified by comparing with the public transport rates in the rural areas, while information on the cost of prescribed medicines was compared with the Tanzania Medical Store Department drug cost lists. Based on recommendations from cost guidelines [14, 15], all cost data were collected in Tanzanian shillings and converted to USD for the exchange rate of the year 2016 in which 1US$ was equivalent to 2200 Tanzanian shillings [16]. We collected data on expenditures on four broad categories: Cost of visits to the clinic, informal payments paid to health workers, payment due to medicine and laboratory investigations. Expenditures associated with clinic visits were assessed on a per-visit basis. For instance, patients were asked: “Did you pay to see the health provider today? If yes, how much did you pay? What means of transport did you use to come to the clinic today? If you paid for transport, how much did you pay today?”. We also asked questions on overnight stays in which women were asked if they had to pay for accommodation to stay the night nearby. We also measured expenses incurred for medicine and laboratory investigations by asking a question referring to the visit made during the current pregnancy: “Did you receive all needed medicines and investigations?”. If the answer was “no”, the follow-up question was asked whether they have to go to the private provider and drug outlet to buy the medicine they were prescribed or to do the investigations which they could not find at the mobile clinic. A similar approach was taken to measure the amount of money they have used for food by asking about the events that happened during their visits during this pregnancy. We also asked about the frequency of these events; although and because our unit of estimation was per visit, we did not take into account the frequency of these payments. Data were also collected on the time-related cost associated with clinic visits. Data were collected on time (in hours) spent travelling to the clinic, and time spent at the clinic by asking questions like: “How long did it take to travel from your home to this clinic? How long in total does it take for you to finish all that you need here at the clinic, from seeing the health provider and taking the needed medication to investigations?”. Total time cost accounted only for waiting time, consultation time and travel time for a one-way journey. Travel cost was estimated as one-way because interviews revealed that women utilized the time and costs while travelling back to their houses by going to the market or to attend other social activities, hence it seems appropriate to account only for the travel cost and time for only one way. Patients were asked whether time and cost prevented them from utilizing health care, using questions like: “In the last 2 months, have you ever missed your scheduled visit due to lack of money? If yes, how often?” and “In the last 2 months, have you ever missed your scheduled visit due to lack of time? If yes, how often?”. Similar questions were asked for unscheduled visits. We constructed a dichotomous variable “cost as a barrier”, which took the value of 1 if individuals reported missing either their scheduled or unscheduled visit due to lack of money, or 0 for those who answered no. We also generated a dichotomous variable of “time as a barrier” which took the value of 1 if individuals reported missing either their scheduled or unscheduled visit because of time, or 0 for those who answered no. A double data entry was done using EpiData software version 3.1. Data were cleaned and extracted in STATA version 12. Data were grouped into women who travelled for more than 1.5 h and the ones who travelled less than 1.5 h to allow comparison of the mean time utilized, mean cost and mean time cost between the groups. The cut point of 1.5 h was based on recent findings on a study done in Tanzania that modelled the geographic access of emergency obstetrics and neonatal care [17]. In the study, it was observed that only 13% of women can reach health care facilities within 2 h on foot and almost 32% of live births were among women residing in areas where it is impossible to reach facilities within 2 h [17]. That indicated that the World Health Organisation(WHO) [18] optimal travel distance of 2 h is in fact not perceived as “optimal” for the majority, especially in rural and remote settings like in Tanzania. Based on that information and the information on the mode of transport of our participants, 1.5 h were taken as a compelling cut point for travel time. The demographic summary statistics such as proportions of women belonging to similar occupations, education levels, and parity levels and their percentages were computed. Summary statistics on cost and time such as mean and their standard deviation were also computed and compared by the group. The Wilcoxon–Mann–Whitney rank sum test was performed to compare the differences between the estimated mean values in the two groups because the data were positively skewed. A p-value of less than 0.05 was considered statistically significant. However, and because no authoritative reference for setting the significance level exists [19–22], we also reported the real p-values and standard deviations of our estimates. Unlike in other medical research, presenting the original data in cost analysis and their distribution is argued to give the reader a more accurate understanding of the similarities and differences in cost than focusing on p values alone [14, 15, 23, 24].