Mojo City is located in the East Shewa zone, Oromia region of Ethiopia at 74 km distance from Addis Ababa, the capital city of Ethiopia. Based on Ethiopia’s 2007 census data, the estimated population of Mojo City was 58,941, of which 51% (30,060) were females. Adults aged 30 years and older constitute 29% (17,093/58,941) of the total population. Mojo City has 1 primary hospital, 2 health centers, and 2 health posts. Mojo City is strategically located along the Ethio-Djibouti railway and was identified by the Ministry as one of the high NCD burden areas in the country.

This study used a concurrently embedded a mixed methods approach that combines both quantitative and qualitative data collection strategies at 2 sentinel sites in Mojo City, such as Mojo and Haachaltu Gudina Health Centers, and their 2 administrative institutions, such as the Mojo City health administration office and Oromia regional health Bearau. The evaluation included the performances of these sites from September 1, 2021 to July 30, 2022.

We purposively included the only available 2 sentinel sites in Mojo City, that is, Mojo and Haachaltu Gudina Health Centers. Based on the phenomenon of interest, recommendations from qualitative studies, and practical considerations such as data and resource needs, 14 key informants were purposively selected based on predefined criteria, such as training, experience, education, and position they held.

We included and reviewed the central triage register of all adults aged 30 years and older, screened for hypertension to identify patients with initially elevated BP. We included all actively followed patients with hypertension for BP control assessment. Also, we randomly took a convenient sample of 20 patients’ records to assess data completeness and the agreement between cohort register and surveillance reporting platforms. For sensitivity calculation, we compared the proportions of hypertensive cases captured by the surveillance system with those revealed by the survey findings. Finally, we took all individuals having initially raised BP and later confirmed to truly have hypertension to determine positive predictive value (PPV).

We followed the Centers for Disease Control and Prevention’s updated guidelines for evaluating public health surveillance systems to assess the usefulness and 9 attributes of the surveillance system [21]. In addition, we adapted the standards for communicable disease surveillance and response systems evaluation to assess the core and supportive surveillance functions such as case detection, investigation and confirmation, communication and reporting, routine data analysis and interpretation, the availability of surveillance platforms, training, and supportive supervision and feedback system [22].

Data were collected using a semistructured questionnaire through face-to-face interviews, observations, and record reviews. Simplicity, flexibility, stability, acceptability, and usefulness were assessed via key informant interviews. A semistructured interview guide was developed in English and translated into the local language, Oromifa. The interview guide was designed in a way that allowed free responses as well as choices from predefined categories, for example, yes or no response-option questions. After the informant consent was obtained from each key informant, the lead author (ADG) conducted 14 face-to-face interviews. Interviews were audio-recorded, notes were taken, and each interview lasted 15 to 30 minutes.

We assessed data analysis and interpretation practice, availability of surveillance guidelines and standards, supervision and feedback, and timeliness via observations. Using medical record reviews, hypertension screening and case detection, BP control, investigation and confirmation, BMI screening, data quality, PPV, and representativeness were assessed. Both medical record reviews and observations were conducted by trained Bachelor of Science (BSc) degree nurses.

Before the actual data collection, we discussed with key informants to ensure that the evaluation was able to address important questions and gather credible evidence. During each visit of data collection, we briefed participants on the purpose and objective of the evaluation, underscoring that it is not an individual performance evaluation. The entire data collection process was closely monitored by research supervisors and principal investigators.

The quantitative data were coded, cleaned, and analyzed via SPSS software version 25.0 (IBM Corp) while the qualitative data were analyzed via the following methods. First, we transcribed the audio-recorded data and carefully read the text to obtain a general sense of the content. Second, we identified condensed meaning units and developed initial code books. Code books were designed in a way that they are flexible and devised when novel themes emerged. After the open coding, we created subcategories that merged codes with similar meanings, and subcategories were then combined to establish categories. The data were subsequently analyzed manually, narrated, and summarized via thematic analysis.

The study was approved by the Afar Public Health Institute (approval APH/25/21). We also obtained a collaboration letter from Ethiopia’s Ministry of Health, Oromia’s Public Health Institute, and Mojo City’s Administration Health Office. Before commencement of interview, informed consent was obtained from the participant to ensure their understanding and agreement to the recording process. After each participant understood the purpose, potential risks and benefits of the study, they were orally consented for participation in the study. The participation process was entirely voluntary and no monetary or material compensation was provided to the participants. Such consent and compensation processes were part of the study protocol approved by the ethical approval committee. The data and all collected information were delinked, deidentified, and anonymyzed for confidentiality. They were also kept in a secure file cabinet.

The study invited 14 key informants for interviews, and all were willing to provide an interview. Of these key informants, 4 (29% held master’s degrees in public health, 4 (29%) were public health officers, and 6 (42%) were BSc degree nurses, of whom 4 (29%) were surveillance officers, and 2 (14%) were facility heads (Table 1).

The hypertension surveillance system in the Mojo City is a multitiered framework involving various health care providers and administrative personnel. Public health officers and BSc nurses are responsible for identifying and diagnosing hypertension, and initiating antihypertensive treatment. Health information technicians collect and report data, while the administrative staffs oversee data aggregation, analysis, and interpretation and use for action.

Based on the national clinical guidelines, surveillance system in Mojo City defines hypertension with a systolic blood pressure (SBP) of ≥140 mmHg or a diastolic blood pressure (DBP) of ≥90 mmHg in 2 consecutive visits. BP control is defined as achieving an SBP of <140 mmHg and a DBP of <90 mmHg in 2 consecutive visits, with the specific targets of <130/90 mmHg in patients having comorbidities such as diabetes, chronic kidney disease, and cardiovascular diseases.

The surveillance system uses both electronic and paper-based reporting platforms to facilitate efficient data flow between health care providers and surveillance officers. The information flow follows a bottom-up approach, with data transmitted from health centers to the city health office and then to the regional health bureau (Figure 1).

Between September 1, 2021, and July 30, 2022, a total of 3760 adults aged 30 years and older were screened for hypertension. Among those screened, 553 (14.7%) individuals were presented with raised BP, and 512 (13.6%) individuals were confirmed to have hypertension. Of the patients confirmed with hypertension, 76 (14.8%) patients were lost, while 436 (85.2%) were receiving antihypertensive treatment. Only 214 (49%) patients who were on antihypertensive follow-up had achieved controlled BP.

A systematic approach was employed to confirm individuals with initially raised BP. Nurses and health officers oversee BP screening and BP confirmation. All patients with raised BP were reassessed for confirmation according to the national clinical guidelines. However, the initiation of antihypertensive therapy lacked essential investigations. Specifically, only 9.7% (50/512) of patients received a baseline renal panel, despite it being a crucial step outlined in the national clinical guidelines and patients may be at higher risk of undiagnosed kidney damage, which could lead to worse outcomes, including kidney failure and cardiovascular complications. The test is essential for evaluating kidney function, detecting early signs of kidney damage, assessing potential comorbidities related to hypertension, such as chronic kidney disease, tailoring treatment plans, and preventing the progression of hypertension-related renal complications. Furthermore, BMI screening, another essential component of hypertension management was not systematically conducted for all patients, with only 25.8% (132/512) of patients with hypertension undergoing this assessment.

The surveillance system uses Health Management Information System (HMIS) framework of reporting diseases in Ethiopia. Health centers submit reports to Mojo city administration Health Office every 21st-26th of each month. Similarly, Mojo city’s administration Health Office subsequently compiled and sent to Oromia Regional Health Bureau within the same time frame. The Oromia Regional Health Bureau further transmits reports to the Federal Ministry of Health within the same reporting schedule. This type of surveillance system was fully implemented throughout the year and maintained through both electronic and paper-based reporting system (Figure 1).

Health facilities did not conduct routine data analysis, resulting in hypertension data that lacked specific information regarding time, place, and person characteristics. In addition, there were no established action thresholds for the interpretation of data or for formulating appropriate responses.

The surveillance system in the Mojo City ensured the continuous supply of essential surveillance forms, standards, and guidelines. Both health facilities had sufficient back up store of tally sheets, intake forms, follow-up forms, referral forms, and cohort registers. However, health facilities had the shortage of appointment forms and tablets for SIMPLE App (Resolve To Save Lives), a smart app designed for tracking patient follow-up.

Neither Ethiopia’s Federal Ministry of Health nor the Oromia Regional Health Bureau provided written feedback regarding the performance of health facilities. Only 6 out of 14 (43%) of key informants received hypertension training. In addition, the heads of health facilities expressed concerns about inadequate training and backup staff to manage staff turnover and annual leave.

We followed the Centers for Disease Control and Prevention’s updated guidelines for evaluating public health surveillance systems to assess on whether the existing surveillance system in Mojo City operated effectively and efficiently. The usefulness and 9 system attributes were streamlined in this evaluation. Accordingly, the surveillance system in Mojo city was useful, simple to operate, flexible to changing needs, acceptable by surveillance staff, has higher PPV but less sensitive, unrepresentative, and unstable.

While the surveillance system in Mojo City was fully operational throughout the year, it lacked regular supportive supervision and feedback mechanisms. There is need to institutionalize periodic evaluations, enhance the system’s functionality, and foster a continuous improvement. Public health surveillance system should be periodically evaluated to provide scientifically sound and epidemiologically interpreted feedbacks for performance and further for its improvement [22,23].

All surveillance staff expressed willingness to implement the system and noted that it was both useful and simple. Half perceived the surveillance system stable, while 71% (10/14) thought that it was flexible. Similarly, the performance of public health surveillance system in Dangila, Ethiopia, was found flexible and useful [24]. Furthermore, a hypertension surveillance system evaluation conducted in Indonesia at 2 health centers and 1 administrative office reported an overall low stability, simplicity, acceptability, and data quality, but the researchers were unable to assess flexibility [25]. However, unlike this study, researchers relied on data collection, storage, and reporting to assess stability, rather than using key informant interviews [25].

Given that 3760 adults aged ≥30 years were screened for hypertension out of a corresponding population of 17,000 inhabitants, the surveillance system in a Mojo City demonstrated low representativeness, having reached only 22% of the population. However, in a hypertension surveillance evaluation conducted in Indonesia, they were unable to assess representativeness of surveillance system [25]. The low representativeness in the present surveillance system may be attributed to low community awareness and routine BP screening, and regular medical checkup tradition, which could leave significant numbers of undiagnosed patients in the community. In this system, health centers collect data and report it to the Mojo City Health Office and the Oromia Regional Health Bureau. However, health facilities did not conduct routine data analysis or interpretation, nor did they use the information to inform action. Data collection necessitates analysis because the goal of surveillance is to use the information for action, reduce morbidity and mortality, and to improve overall health [22,23]. If the collected data are not followed by regular analysis and the application of information for action, it results in the mere consumption of resources [26].

Public health surveillance systems should operate in a manner that facilitates effective communication and dissemination of health and health-related data, enabling decision-makers at all levels to readily understand the implications of the information. Timeliness of reporting is a crucial factor for the effective communication and dissemination of health and health-related data. The report timeliness in the current surveillance system was 100%, which aligns with the Ethiopian national target of 95% [27] and the WHO target of 80% [28]. In addition, data completeness was 98%, consistent with the national target of over 95% for report completeness [27] and the WHO target of 80% [28].

The sensitivity and PPV of a surveillance system depend on the design of the case definition and its clarity among stakeholders [29]. The surveillance system in Mojo City exhibited a sensitivity of 45.3%, capturing only 45.3% of hypertension cases and more than half (54.7%) of the individuals with hypertension may be missed by the system. It is less sensitive than the hypertension surveillance system in Bogor City, Indonesia, which has a sensitivity of over 80% [30]. The low sensitivity may be attributed to low health care service utilization and suboptimal BP screening practice, as only 22% of the surveillance population was screened for BP. In a hypertension surveillance system evaluation at the Jombang District Health Office, Indonesia, they were unable to calculate sensitivity due to the lack of baseline survey data for comparison. However, the PPV of our surveillance system was 92.6%, indicating that the system correctly identifies 92.6% of true cases of hypertension. It is more predictive when compared with other hypertension surveillance systems in Dangila, Ethiopia which has PPV of 11% [24] and Jombang District Health Office of the other city in Indonesia which has PPV of 59.4% [25].

The limitations of this evaluation are as follows. First, the results may have been subject to potential social desirability bias, as surveillance stakeholders may have provided favorable responses especially qualitative attributes such as acceptability and simplicity. The surveillance stakeholders may have been inclined to provide more favorable responses to questions about these attributes, potentially overestimating the true level of satisfaction or ease with the surveillance system. To mitigate this, we briefed them on the purpose and objectives of the evaluation at each visit before commencing data collection. Second, reliance on a convenience sample of 20 patients may not have provided a representative assessment of data validity. Third, although this evaluation followed the Centers for Disease Control and Prevention’s updated guidelines for evaluating public health surveillance systems as applied in previous studies conducted in different settings to gather credible evidence on the performance of the surveillance system, it is primarily relied on responses from key informant interviews rather than objective quantitative data collection, storage, and reporting, unlike other studies. Finally, the availability of the limited data in the field led to insufficient understanding of surveillance systems in similar contexts and a few of interpretations were made compared with surveillance systems in other country which may have structural and contextual difference with our surveillance system.

The sentinel hypertension surveillance system in Mojo demonstrated effective implementation throughout the year. The surveillance staff felt that it was helpful in determining the magnitude of hypertension in the area as well as for improving documentation and patient record keeping practices. It was found to be simple, flexible, acceptable, and have a higher PPV. The data quality and reporting timeliness were consistent with national and WHO surveillance standards. However, it was less sensitive, less representative, and unstable.

The findings underscore the need for institutionalization of periodic evaluations and supportive supervision mechanisms to assess the system’s performance, provide feedback, and identify areas for improvement. The system should implement routine data analysis and interpretation mechanisms based on epidemiological principles to interpret the collected data, identify trends, and target tailored public health interventions. Continuous staff training on data collection, surveillance protocols, and the operation of the surveillance system should be designed to enhance stability. Furthermore, designing of government-driven sustainability strategies, such as establishment of dedicated public health financing system and consistent resource allocation and capacity building, and cross-sectoral collaborations with local governments, could reduce reliance on external donors and enhance the sustainability of the surveillance system. Efforts should also focus on community awareness creation activities to improve sensitivity and representativeness of the surveillance system. Finally, if resources permit, scaling the system into a nationwide program could significantly enhance hypertension surveillance across Ethiopia.