Factors influencing the collection of information by community health workers for tuberculosis contact tracing in Ekurhuleni, Johannesburg
Maruma, Thabang Wellington
Background: Surveillance structures for tuberculosis (TB) contact tracing are not well integrated into routine national reporting structures. The implementation of reingineering of primary health care through ward based outreach teams (WBOTs) is a step towards equitable primary health care. Data and information collected by WBOTs for household TB contact tracing is an integral part of the implementation model of primary health care reengineering. The quality of patient record documentation becomes even more vital in light of the increased focus on process and outcome measures in health programmes and as a result, careful consideration be given to the WBOT data collection system used by community health workers (CHWs). In order to contribute to efforts of developing an optimised model for household contact tracing, the acceptability of the current paper-based data collection system needs to be assessed in order to develop a comprehensive monitoring & evaluaiton (M&E) framework for an optimsed model for household tuberculosis contact tracing. Methods: The current cross sectional research project is nested within a project that aims to develop an optimised model for household TB contact tracing. In this nested mixed methods study; the exploratory sequential design was used to explore the facilitators and barriers to completing the current data collection tools used by CHWs. The study had two components, firstly three focus group discussions (FGDs) were conducted in the three Ekurhuleni health sub-districts (Northern, Eastern and Southern) in three purposively selected primary health clinics and secondary data analysis of the main study`s FGDs was also conducted. Manual coding and QDA Miner software was used for coding and all qualitative analysis. Emerging themes were identified through inductive thematic analysis using the constant comparison analysis framework. The results informed the quantitative data collection and analysis. Following qualitative analysis; a close ended questionnaire was refined and informed by the results of the qualitative inquiry. CHWs were recruited using targetted sampling techniques from 6 primary health care facilities located in the different sub-districts in order to administer the questionnaire. The four point Likert Scale questionnaire was developed using theoritical framework for acceptability (TFA) constructs to asses the level of acceptability of the current data collection tools used to document tuberculosis contact tracing activities. Univariate and multivariate linear regression models were fitted to examine significant relationships between the composite acceptability scores and several predictors. All quantitative analysis was perforned on STATA version 14 (StataCorp College Station, Texas 77845 USA). Results: A total of five FGDs were conducted; two that were conducted as part of the main study supplemented the data from the three that were conducted (one in each Ekurhuleni health Sub-district). The total of 54 CHWs participated in all the five FGDs with 89% being female. Average age of all CHWs was 34.41 years [mean (sd): 34.41(8.16)]. Five broad themes emerged including inadequate CHW training, WBOT programme integration with other health and social care service providers, challenges with the WBOT data collection system, community access issues and preference for a digital based data collection system. Data related barriers identified included limitations with the current paper based data collection system such as insufficient competency assessments about the different data collection tools, lack of a specific tool to capture TB contact tracing activities, incomplete referral forms due to clinic staff not completing them, patients providing wrong information, too many papers to complete. Those that were related to the WBOT actvities included lack of community acceptance, resource constraints, violent patients and community members, community members that are not welcoming . Facilitators included motivated CHWs. 94 CHWs were enrolled for the quantitative survey with 90 (95.74%) females. From the total, 35% of the CHWs were from the Ekurhuleni health southern subdistrict, 34% and 31% were from the eastern and northern sub-districts respectively. The overall median (IQR) composite acceptability scores from all sub-districts was 48 (45 51), with the highest scores observed in the Eastern sub-district 49 (45 46) . In the overall study population, the acceptability of the current WBOT data collection tools was low. Conclusions: Main findings pertaining to CHW training indicate that the different phases of the Primary Health Care (PHC) reingeering WBOT trainings were inconsistent. There is also a lack of acknowledgement of attendance as CHW expressed their dissatisfaction in not receiving certifications which resulted in low morale for conducting outreach activities. The sub-optimal integration of the WBOT programme into the primary health care system results in a patchy referral system characterised by incomplete back referrals resulted as referral forms remain incomplete. Communication between the primary health care facility staff and WBOT CHWs needs to be strengthened in order to strengthen the referral linkages with other health and social care service providers. Funding models for WBOT programme need to be reviewed to ensure that resources needed for optimal WBOT functioning are secured. Restricted access to some communities, patients providing wrong addresses, violent and unwelcoming household members and lack of WBOT safety were barriers to accessing TB patients during outreach activities; thus leading to incomplete and innacurate data. The limitations posed by the current paper-based data collection system have been acknowledged and the CHWs preference for a digital based system highlights the need for the evaluation of the current mobile data collection technologies in other regions in order to inform nationwide scale-up. Recommendations: The implementation of the WBOT programme is still in its infancy and in order to improve the data collection processes of the programme, more research on CHW post-training competence is needed to determine the effectiveness of the wide array of training programs. Moreover, the implementation of CHW program should be coordinated among the different training providers including government, civil society organizations and NGOs. To improve the quality of the CHW training delivery and content, CHW feedback should be sought through pre-and post-assessments. There is a need to focus efforts on coordinating and strengthening the different PHC reengineering streams and integrate them into the primary health care system. This will likely strengthen the referral system between the WBOT programme and PHC facilities. The current M&E policy needs to be reviewed and special consideration should be given to TB contact tracing related indicators. This should also be accompanied by an adjustment of the current WBOT data collection tools to better reflect the agreed upon TB contact tracing indicators. The study further recommends further research in the form of economic evaluations to determine the cost effectiveness of scaling up current digital based data collection methods to inform nationwide scale up. Key words: Ward Based Outreach Teams, data collection system, data collection tools, community health workers, TB contact tracing, Community Based Information System, acceptabiltity, mHealth
A research report submitted to the Faculty of Health Sciences, University of The Witwatersrand, in partial fulfilment of the requirements for the degree of Master of Science Epidemiology in the Field of Implementation Sciences. May 2018 Johannesburg, South Africa.