For the assessment of the DTTB Program the researchers had the opportunity to meet with all the current DTTBs (n=71) on the occasion of a forthcoming national CME session, and to a DTTB alumni database of the 452 former Program participants – although, as explained below, the contact information was later found to be outdated which severely compromised its usefulness for random sampling and tracing purposes. Given this constellation, the changes in the design of the Program that had taken place during the years of its existence and constraints of time and other resources, a mixed methods approach was selected: a self-administered questionnaire for all current DTTB, and oral interviews by telephone with available former DTTBs. The findings from both methods were analysed separately and later consolidated for the final interpretation
The development of the questionnaire for current DTTBs was based on a review of formally published as well as grey literature on staff retention in rural areas. A review of policy documents from the Philippines and the DTTB Program was also conducted. These included the DOH Revised Operational Guidelines for the Implementation of the DTTB Program and the Memorandum of Agreement between the DTTB, the LGU, and the DOH. Moreover, explorative interviews were conducted with key informants including the Director of the DOH Health Human Resource Development Bureau, and the past and present DTTB Program coordinators in order to guide the development of the questionnaire.
The questionnaire used in this study was inspired by The ‘Stayers Questionnaire’ used in a health workforce assessment manual in Uganda
. Permission was kindly given by the author, Emily Bancroft of the University of Washington, to adapt this tool for the purpose of the assessment. The tool covers the constructs: personal and job satisfaction, career advancement, working environment, living conditions, and compensation. These were all issues that the key informants had raised as being relevant for the assessment. In addition constructs dealing with local politics and DOH support were added. Each of the constructs comprises a list of statements. The respondents were asked to use a 5-point Likert scale to indicate whether they strongly agreed, agreed, were neutral, disagreed or strongly disagreed with the statements. 41 of the 77 statements from the original ‘Stayers Questionnaire’ were used in the version employed for this assessment.
The variables retained and explored from the demographics section of the original Stayers Questionnaire were sex, marital status, number of dependents, and city/municipality of residence. After perusing the responses from the key informant interviews it was decided to expand these variables to include whether respondents graduated from a medical school within or outside the National Capital Region (NCR), whether their main reason to join the Program was to return service for a scholarship or not, whether their main reason for joining was an interest in Public Health/Community Medicine or other reasons, and whether they planned to remain in the area of assignment for more than one year or indefinitely after the Program, or alternatively to leave.
To reach the current DTTBs in an efficient manner, advantage was taken of a CME session to request their participation in the assessment. All of the current 71 DTTBs were in attendance and all agreed to take part. Time was allocated at the session to introduce the study and the use of the Likert scale. Completing the questionnaire took approximately 30 minutes. In all but one case, the participants responded to all items of the questionnaire.
Epi Info was used to analyse data obtained by the questionnaires from current DTTBs. The descriptive aspect of the analysis measured the frequencies by which the respondents gave a rating of “agree strongly” or “agree” to a specific question and the corresponding mean score. For the inferential component of the analysis, independent two-sample t-tests were run for the statements and questions, and the differences between the mean scores were tested for statistical significance (p-value ≤ 0.05).
The topic guideline for interviews with former DTTBs was developed to broadly align with the sections of the self-administered questionnaire and made use of open questions so as to obtain further, clarifying and explanatory information. Both tools were tested by a small panel including two former DTTBs, a former DTTB Program Coordinator and the researchers conducting the assessment. The topic guideline provided the key themes for the analysis of the interviews with former DTTBs. Sub-themes emerging from the responses were then identified.
Tracing the former DTTBs proved to be extremely difficult, as the contact information in the Alumni Database had not been maintained. Snowball sampling was, therefore, used over a one-month period to generate a productive sample
, that included DTTBs who chose to remain in their rural position after the Program, some who left, and others who remained in rural service, but preferred to re-enter the Program in a different geographical area. An effort was also made to include former DTTBs from different cohorts since the Program’s inception.
Former DTTBs (n=26) were interviewed either face-to-face or, in the majority of cases for practical reasons, by phone. All the interviews were conducted by the same interviewer (JAL). None of those approached refused to be interviewed. The interviews lasted an average of 20 minutes. The responses were translated into English and directly transcribed by the interviewer.
This assessment was suggested and facilitated by the Human Resources Development Bureau which obtained ethical clearance from the Department of Health in the Philippines. Informed consent was gained from all those who agreed to take part in either the questionnaire or the interviews. Respondents were allowed to withdraw at any time or skip questions without having to give a reason. However, no one made use of this possibility. All data were de-identified during the analysis so that responses could not be traced back to a particular informant.