|
nif:isString
|
-
Ethical approval for this study was obtained from the Conjoint Health Research Ethics Board at the University of Calgary (E-24432), the Clinical Research Ethics Board at the University of British Columbia (HII-00947), and Institutional Review Committees at the Mbarara University of Science and Technology (March 11, 2011). The study was registered at clinicaltrials.gov (NCT02046018).
This study was conducted in Kyabugimbi sub-county (population 37,200), where the iCCM intervention occurred. Located in Bushenyi District in southwestern Uganda, Kyabugimbi sub-county is a rural area with steep, mountainous terrain and few roads, mostly unpaved. Most families in the area rely on subsistence farming for survival and few have access to electricity, running water or modern sanitation systems. Access to government health centres is limited, and health services at existing facilities are challenged by staffing, equipment shortages and limited infrastructure. In 2010, 198 CHWs were selected and trained in 98 villages in the intervention sub-county. Selection of CHW's occurred through a nomination process (show of hands) at community meetings. Community members were asked to select CHW's based on the individual's age (18+years), perceived trustworthiness, demonstrated willingness to volunteer, years of residency in the community, and the ability to be available in the community. Level of education was not a factor for CHW selection. Each village selected 2 to 5 CHWs depending on the population size of the community; these individuals attended a five-day basic health promotion course, conducted by local health centre staff and in accordance with Ministry of Health guidelines, which followed a standard curriculum set by the Government of Uganda [18]. All CHWs provided health education and promoted healthy practices within their respective communities. In April 2012, two CHWs from each village were selected (by community nomination) to receive additional iCCM training. Following five days of iCCM training, these CHWs began assessing and providing pre-packaged drugs to treat uncomplicated illnesses—antimalarials (coartem) for fever, antibiotics (amoxicillin) for presumed pneumonia, and oral rehydration salts/zinc for diarrhea—to sick children under five years of age in their communities. About half of the CHWs trained in iCCM were also enrolled in an enhanced study arm and were provided with mobile phones to supplement iCCM provision. CHWs selected for the cell phone arm of the study were nominated into the role at community meetings using the same guidelines employed for the selection of non-cell phone CHWs. CHWs with study-supplied phones used a mobile application to report the age of the child and his/her symptoms; the application then guided the CHW through a treatment algorithm that suggested appropriate treatment. Data entered into study-supplied phones was transmitted immediately to a database located at a local health facility; health centre staff reviewed the database daily. All CHWs were supervised and supported by local health centre staff who reviewed CHW reports, either paper or cell phone generated, met monthly with CHW's to discuss the program and address problems, and communicated with CHWs regarding referrals or follow-up of patients discharged from hospital. Health centre staff were also available to consult with CHW's regarding complicated cases as needed.
This qualitative study was part of a larger research effort, which included a quantitative component with household surveys in intervention and control communities, as well as a review of operational data. Further details on the larger study are published elsewhere [20] Focus group discussions (FGDs) and key informant interviews (KIIs) were designed using a phenomenology framework [22] and were conducted post-intervention with key stakeholders who were involved with the HCU iCCM project in intervention communities. FGDs and KIIs tools were semi-structured, developed in English, translated into the local language (Runyankore) and then back-translated into English to ensure accuracy of translation. Interview tools included open-ended questions that addressed stakeholders impressions of the iCCM by CHW intervention; access to care; perceived satisfaction and quality of care; referral; patient encounters; supervision; and drug supply. A different interview tool was created for each stakeholder group. Sample FGD/KII questions are presented in Table 1. FGDs commenced directly after completion of household surveys, which permitted preliminary results from the household survey to inform the development of questions for the FGDs and KIIs.
Table data removed from full text. Table identifier and caption: 10.1371/journal.pone.0098610.t001 Sample of Questions from Focus Group Discussions and Key Informant Interviews. Note: In Uganda CHWs are called Village Health Teams (VHTs) and are referred to as VHTs in these questions.
FGDs and KIIs were conducted in December 2012, following eight months of iCCM intervention. Participants were caregivers of children under five years, CHWs, health centre staff, local leaders and district government officials. A caregiver was defined as any adult who looks after the day-to-day needs of the child (provides food, shelter, safety). KII participants were purposely selected due to their involvement in or exposure to the intervention. FGD CHW participants were randomly selected using Rand's random digit sampling strategy [23] based on CHW lists. Caregivers (M/F) were eligible for participation in a FGD if they had a child under five years of age and had consulted an ICCM CHW for a sick child during the previous six months. Eligible caregivers were identified by CHWs. Due to small numbers and close proximity, all local leaders (12) and health centre staff (13) involved in the study were invited to FGDs. Government officials (5) were selected for a KII due to their involvement in or knowledge about the program. The sampling method for FGDs is detailed further in Table 2.
Table data removed from full text. Table identifier and caption: 10.1371/journal.pone.0098610.t002 Sampling Strategy for Focus Group Discussions. Focus Group Discussions were held in nine intervention parishes in Kyabugimbi Sub-county in November and December 2012.
Experienced facilitators (2 males, 1 female) from Mbrara University of Science and Technology facilitated the FGD's and KIIs. None of these facilitators were previously involved in the project implementation or had a vested interest in the outcomes of the study. FGDs with caregivers, CHWs and local leaders were conducted in communities, within walking distance of the homes of participants, at local churches or community gathering places. Locations provided privacy and no non-participants were present. The FGD with health workers was held at a health center in a private meeting room. KIIs were held at offices or in private residences and in privacy. All FGDs included both male and female participants together except the caregiver groups, which were separated by gender. All FGDs and KIIs lasted approximately one hour. At the end of data collection, the facilitators felt that saturation was achieved. FGDs were conducted Runyankore. KIIs were conducted in either English or Runyankore, based on participant preference. All FGDs and KIIs were audio recorded. Field notes were recorded by a HCU employee (1 female) who accompanied the facilitators to FGDs and KIIs. Recordings were transcribed by an experienced local transcriber (1 female) who previously preformed similar duties for other Mbrara University of Science and Technology projects with good results. The transcriber listened to the recordings in Runyankore and directly translated into English. On two occasions, HCU staff checked the quality of transcription by selecting a transcript and re-listening to the audio recording while reading the transcribed text. On both occasions HCU staff were satisfied with the quality of transcription. English transcripts were coded in Nvivo 9 and then analyzed by the main author, who was not involved with the study intervention. Thematic analysis strategies were used to code the data, which involved familiarization with data, identification of the emerging thematic framework, memos, and mapping [24]. Data were further conceptualized through the development of word frequencies, “clouds” and models. Emergent themes were tested through queries and classification of data into gender (male, female, mixed data), type of respondent (CHW, caregiver, health worker, government), and type of data (FGD, KII) from which additional queries were run. The FGD facilitators, who were also skilled in qualitative data analysis, cross-checked results and provided input. Informed consent was obtained from all participants. Due to the low literacy level of CHWs and community members and some local leaders, the study was explained in detail by the facilitators before commencing the FGD or KII. The informed consent form was then read to the individual or group, and each participant was asked to either sign or provide a thumbprint at the bottom of the consent document to indicate his/her consent. Other literate participants were also verbally informed about the objectives of the study and then invited to read and sign the consent form The purpose of this study was to contextualize a larger quantitative study and provide insights into stakeholders' perceptions of the iCCM project and its impact on access to health care. While every attempt was made to include respondents who represented the population in the study area, our ability to engage community members was limited by the need to have CHWs identify and invite participants to FGDs, which opens the opportunity for selection bias. Moreover, only community members who sought help from a CHW were included in FGDs. Opinions from community members who did not access CHW services are not included in this study. Due to study limitations it was not possible to share transcripts with participants and gather feedback. Results from this study cannot be generalized beyond the study group, although results are similar to other qualitative studies that examined perceptions of iCCM, most notably the study by Callagan-Koru [9].
|
![[This page as RDF]](http://data.gesis.org/somesci/static/rdf-icon.gif){kind=icon}