Socio-demographic characteristics of the respondents
This study involved 115 health facilities in 11 districts. A total of 93 health facility workers and 13 district officials were interviewed. About two-thirds (65.6%) of them were Diploma holders. Clinical officers accounted for the majority of the respondents (43.0%) followed by nurses (16.1%) and midwives (11.8%). Most (52.6%) of the respondents had over 10 years of working experience (Table 1). At the district office, 13 members of the Council Health Management Team were interviewed. They included District Medical Officers, District HMIS and Reproductive Health Focal Persons.
Utilisation of HMIS data at facility and district levels
Health facility level: About two-thirds of the facilities (60%) claimed to use the HMIS data they collect. Data were used for comparing performance in terms of services coverage (53%), determining morbidity and mortality trends over time (50%), providing public health education and promotion (55%), and for determining the requirements for drugs and other medical supplies (37%). The commonly identified types of data displayed at health facilities were top ten diseases (58%). More than half (56%) of the facilities had displayed recent data analysis outputs (during the past three months). However, most of the displayed information, did not indicate dates hence it was difficult to detect the respective period covered. Most facilities did not have statistics to compare values but rather listing the top ten diseases. Less than 10% (n=11) of the facilities were found to conduct proper analysis and use of their data.
Most facility respondents described the quality of data collected by the facilities to affect its utilisation in evidence-based decisions. However, most respondents believed that always decisions are made based on the evidence or actual needs, at both facility and district levels. These included the actual needs of the service population and considering costs or the financial capacity of the facility. Furthermore, it was found that in few instances decisions at health facilities were made based on directives from higher authorities. Few good practices on data use were claimed not to be noticed by superiors and about two-third complained that staff was not rewarded for their good work but always made responsible for their poor performance. Regarding the importance of collecting health information, health facility workers strongly agreed that they do understand why they are required to collect and submit data. However, most of them agreed that collecting data that is not used for decision making discourages them.
Nearly all (96.3%) of the health facilities reported to have routine meetings to review managerial and/or administrative matters. Most of the health facilities (46.6%) reported that the meetings were held monthly and nearly all (92.2%) maintained official records of the meetings. More than half (58%) of the facilities were able to provide copies of proceedings of their previous meetings within the respective quarters. However, when the proceedings were examined, the majority (62.8%) did not have data as an agenda for discussions. Issues related to data that were observed from the proceedings included management of HMIS (data quality and reporting) (57%) and discussion about commodity stock-out (60%). Most of them (70.9%) had made decisions based on their discussion. Only about half (49.1%) of the facility reported to participate in meetings convened by the district office to discuss HMIS data during the last three months.
District level: Data-use at the district level was reported by more than half of the respondents (9/13). The district respondents reported to use HMIS data mainly for monitoring and evaluating set district targets. They reported that their districts produce reports that contain information from HMIS data. The commonly produced reports included: annual Council Comprehensive Health Plan, quarterly Council Health Management Team reports, and Reproductive and Child Health reports. Nine of the 13 district respondents reported to systematically receive feedback on the quality of their reports from higher authorities. However, eight respondents reported that the district office does not provide feedback reports to the health facilities upon receiving the HMIS monthly reports.
Five out of 13 district respondents reported to analyse data from the HMIS. A larger proportion (10/13) of the respondents reported that the district office does not routinely hold meetings to review HMIS information as illustrated by one respondent: “We do not have routinely scheduled meetings hence no records of meetings are available. Our data are of poor quality because only a very few staff have received proper training on data management” (Tandahimba). The three reported examples of promotion and use of reported data at the district level included public health education, annual budgeting, forecasting demand, and procurement of supplies. One of the district respondent said: “The data help us to budget and forecast on-demand and procurement” (Njombe).
Districts were asked if the distribution of resources take into account facilities that excel in HMIS performance. HMIS focal persons from 9 districts (Kinondoni, Tandahimba, Mbulu, Hai, Mbinga, Nkasi, Njombe, Igunga, and Dodoma urban), reported that priority is given in the allocation of resources to facilities with good performance. Only two districts (Kahama, Kibaha) reported not to consider facility performance in the allocation of resources. Other criteria used for resource allocation in the annual plans included the size of the facility, the number of clients served, the number of services offered, disease burden, geographical location, and service priorities. However, only Kinondoni and Igunga districts were able to provide performance improvement tools such as flow charts and control charts to monitor the performance of their facilities. More than half (7/11) of the districts reported having a mechanism for generating funds for HMIS. Half of the respondents reported having a long-term financial plan for supporting HMIS activities while the other half reported having no such plans.
Factors influencing HMIS performance
Inadequate human and financial resources: The number of staff specifically responsible for HMIS activities differed from one district to another. All the facility respondents reported on the inadequacy of human resources for data management. The majority of those responsible for HMIS activities reported having received short training on data management focusing only on data collection with very elementary analysis. However, few claimed that no specific training has been provided to them except on job training. A relatively large proportion (41.4%) of the facility’s personnel had not received any training on data management related to HMIS in the past 12 months. On average there were 5 staff involved in HMIS activities per facility while about 21 staff were trained during the previous 12 months (Table 2). One main challenge emphasized by district key informants was limited human capacity in HMIS to apply the analytical tools and methods to synthesize information for decision-making. This was attributed to poor/inadequate training. Lack of financial resources was attributed to inadequate in-service training of staff on data management.
All district HMIS focal persons reported having HMIS training manuals. Some reported having received on-job training (6/11) and a few reported to have district training plans (3/11). The duration of the training (usually 2 -14 days) was described to be adequate by the majority of the district respondents. More than half of district respondents (n=7) reported that those responsible for HMIS activities were also responsible for other activities.
Supervision: Less than half (42%) of the health facilities reported that members of the district office provided supportive supervisory visits to the facilities during the three months’ before this assessment. However, over half (52.4%) reported that the district supervision team did not have a supervision checklist. The majority of the facility respondents (72.1%) reported that the district team had conducted an audit of their data quality. Two-thirds of the facility respondents reported that the district team discussed the performance of the health facility based on the data provided. Only a few (9.6%) of the health facilities reported to have a clear schedule for district supervisory visits. Some respondents (37.5%) explained that most of the time district teams just pay visits without informing the respective facilities. Less than half of the respondents (44.9%) were able to provide the district supervisory reports (Table 3). In contrary to the facility report, 9/11 districts reported to have checklists, visit schedules, and 10/11 to have supervisory reports.
Standard procedures for data management: Different types of procedures and logs for receiving reports which resulted in poor filing and loss of transmitted reports from lower levels were common. Some districts did not have a proper mechanism of receiving reports from health facilities. In such districts, any person who happened to be present at the district office when a report is delivered could receive the report without verification of proper filling or even document acknowledging its receipts. Feedback on reporting and quality of reports from the district office to health facility workers was mentioned by only 54% of respondents, and this was mostly oral feedback. Such practices and communication gaps provided room for the incorrectly/poorly filled and incomplete reports from a lower level to be received by the district and the inability to mark late reports. During interviews with district officials, almost all districts (90%) reported that there were no written procedures in place to address late, incomplete, inaccurate, and missing reports including follow-up with the facility on data quality issues, as one respondent said: “There were no written procedures, just verbal discussions were provided some times through phone calls” (Kinondoni).
Registers, tally sheets, or report forms with no identification of health facility, a period of use, were common. The registers, though filled with data, could not be used because the data could not be linked with an appropriate monthly report during the verification exercise. Only six facilities were found to have a logbooks available for marking and reporting submitted reports and only two districts reported to stamp the forms to indicate the receiving date. Only half of the districts reported having had quality control in place for data entry from paper-based to computer DHIS-2. Different versions of HMIS tools were used by different facilities. Some of the versions did not have the variable of interest.