Study area

The study was conducted in Kampala, the Capital city of Uganda. In mid-2020, it had a population of 1,680,600 as estimated by UBOS [17]. Kampala was selected because PHPs make up 98% of all health providers in Kampala due to the high level of private health care attendance and a high TB burden. Uganda had an estimated TB incidence rate of 200 per 100,000 [18] with Kampala’s incident TB cases at an estimated 8,380 people [16] in 2020.

There were 1395 private clinics, HC IIs and HC IIIs registered by MoH as of 2020 in all the five (5) divisions of Kampala. Lower-level health facilities (Clinics to Health centre IIIs) serve populations of up to 20,000 while higher-level facilities serve bigger populations between 100,000 to 10,000,000. These facilities may be Private, Faith-based, Community-based or Non-governmental Organization owned facilities and the majority of these are Private for Profit (PFP). The Private for Profit mainly consist of lower-level facilities that offer preventive, promotive, outpatient curative health services, outreach, laboratory and maternity services. Whereas higher-level facilities are mostly Private not for Profit and Government facilities that offer, in addition to lower-level facility health services, in-patient, radiology, specialist services, surgery and research. This study was conducted among selected private facilities across selected divisions of Kampala.

Study population

This study was conducted in Kampala district; in private (PFP&PNFP) clinics, HCIIs and HCIIIs that are reporting in DHIS2. All health facilities are expected to report data on TB screening, TB cases diagnosed and TB treatment at their facility. Hospitals and Health Centre IVs were not included because, based on DHIS2 data, TB case notification is especially a problem in lower-level facilities like Clinics, Health Centre IIs and some Health Centre IIIs and conducting the study among them would provide more relevant data for our study. The target population was the above groups of private facilities in Kampala City registered in DHIS2. Although laboratories may do bacteriological diagnosis of TB, they were left out of this study since they do not report into DHIS2.

Study design

This was a cross-sectional study. We employed a mixed methods approach in which quantitative and qualitative data were collected and analyzed concurrently to gain a better understanding of the identified factors that act as barriers and facilitators of TB case notification. We also sought to capture qualitative information explaining some nuances observed at facility level as well as to obtain a deeper understanding of facilitators and barriers to TB notification at the facilities with high or low TB case notification rates.

Sample size

Using the sample size estimation formula below by Kish Leslie [19], the sample size was calculated using a 95% confidence interval and 5% precision, and the expected prevalence of TB notification by private facilities in Uganda which is estimated to be 21% based on unpublished data from the NTLP. The resulting sample size was adjusted for a finite population and non-response rate of 10% catered for to obtain our final sample size of 238 private facilities.

Sampling procedure

Out of the five (5) administrative divisions of Kampala, we purposively selected three (3)—Central, Nakawa and Kawempe—because they are the most populated in Kampala and have higher poorer populations; conditions that pose the highest risk for TB transmission. Study units were sampled from the list of private facilities in DHIS2 which are expected to notify TB cases. Using the line list of DHIS2 registered private facilities in the study area for each of the three divisions, a number proportionate to the DHIS2 registered facilities in the division was randomly selected with the aid of computer-generated numbers Fig 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Sample selection flowchart.

https://doi.org/10.1371/journal.pone.0315402.g001

For the survey, at each facility we interviewed one (1) TB focal person at the facility and, in his/her absence, the substitute.

Participants for the KIIs and IDIs were selected using purposive sampling. The facilities for the In-depth interviews were selected according to high and low facility TB case notification: in each division, we selected one TB focal person from each of two (2) facilities with the highest case notification. Similarly, we had one TB focal person from each of the two facilities with the lowest TB case notification rates in each of the selected divisions. High and low notification was based on the facility’s TB screening numbers reported in DHIS2 between October and December 2021.

Two (2) KI interviews were conducted among two (2) NTLP staff including the Kampala region TB focal person and the PPM officer, and two (2) KI interviews were conducted among Kampala Capital City Authority TB focal persons from Central and Kawempe divisions. The Key Informants included individuals who are responsible for overseeing TB case management in private health facilities in Kampala Capital City and also have expert knowledge on and active involvement in support of TB case notification by PHPs.

Study variables

The dependent variable in this study was notification of TB cases by private facilities. This was a binary variable defined by the number of PHPs that notify TB cases and the number of PHPs that do not notify TB cases. The number of PHPs that notify TB cases were obtained from the number of facilities of those recorded/registered in DHIS2 in the study area that notified TB cases. The number of PHPs that do not notify TB cases was obtained from the number of facilities in the study area that are recorded in DHIS2 but did not notify TB. This data was retrieved from the 033b Health Management Information System data in DHIS2 for the last quarter of 2021.

The independent variables included factors that influence the outcome (notification of TB cases) by private healthcare providers. These variables included: Level of knowledge of TB symptoms; degree of suspicion for TB; sufficient time for provider-patient interaction; willingness to notify a presumptive TB case; availability of guidelines for TB screening at the facility; availability of sufficient health care providers at the facility, level of turnover of health care workers at the facility, willingness of health care providers to engage in programmatic TB management, availability of diagnostic equipment and supplies, provision of TB case finding guidelines/tool kits by the NTLP, availability of specialized TB training for private health providers, TB stigma and health worker cadre.

Data collection and management

Data was collected between March and June 2022. For the survey, we interviewed health workers in sampled PHFs using a structured questionnaire programmed in Open Data Kit (ODK) software platform.

We conducted in-depth interviews using in-depth interview guides and similarly used Key Informant interview guides to interview KIs. After making an appointment with the participants, an interview was held in which information on all objectives was gathered.

Five (5) research assistants were recruited and given rigorous training in collection of the quantitative data for this study, and they were supervised during the entire period of data collection. The filled tools were reviewed for completeness and any possible errors found in the data were corrected to enhance validity of the data. We also reviewed collected data to ensure that variables needed for analysis were captured in the data collection procedures.

A copy of the electronic data was saved as password protected and stored on an external storage device for safety and ethical purposes. The cleaned data was downloaded into Microsoft excel 2019 and imported into STATA 14 for statistical analysis.

Data analysis

Based on DHIS2 data for the study facilities, we generated the outcome variable for notification in STATA and it was defined as the proportion of PHPs (private facilities) doing TB case notification: with “Yes” for those facilities that do (TB notification determined to be 15%) and “No” for those that do not do TB notification (85%).

Individual associations of each independent variable in the study with the outcome variable were tested using Fischer’s exact analysis in STATA. All independent variables whose p-value was less than 0.25, and other important variables with higher p-values were included in the model. We also tested for multicollinearity among the model variables and those whose correlation coefficient was more or equal to 0.4 were left out of the model for multivariable analysis. Since the prevalence of TB case notification among private facilities was 15%, we employed Modified Poisson for the multivariable analysis and, using parsimony, finally came up with a simple model with five (5) predictor variables. P-values for the adjusted Prevalence Ratios (PR) of the tested association of the predictor variables with TB case notification were obtained and used to determine the significance of the findings.

After interviews, audio recordings from the IDIs and KIIs were stored on an external storage device for safety and ethical purposes. We conducted thematic analysis, employing mainly a deductive process to analyze the data and using ATLAS.ti software to ease the analysis. Codes were pre-identified from the framework and then transcripts were reviewed as existing and emerging codes were identified and used to develop a codebook to guide the analysis in ATLAS.ti. The codes were arranged and merged to develop themes and sub-themes relating to common notification practices among the PHPs as well as the facilitators and barriers to TB case notification.

Technical capacity

Key informants indicated that they lacked sufficient capacity to notify TB cases. They mentioned that efforts to build capacity of TB notification by the NTLP encourage them to get involved in programmatic TB management and actively do TB case notification. However, they noted that Support is largely target public facilities and the private are left out.

“I have been here for almost 2 years. I have not had any team from either Ministry of health to talk about TB or to tell me anything about TB, to try to do anything about TB, nothing. And surprisingly, I’m here 5 days (a week) so I [would] know.” (IDI-10)

A key informant at the national level acknowledged the lack of capacity building for PHF (Private Health Facility) providers, seeing it as a significant obstacle to improving TB case notification. They indicated that the resources allocated for capacity building are only for public facilities, as per the government’s plans. Additionally, they noted that the tools used for case notification are frequently updated, requiring ongoing capacity building to maintain effective notification.

“[We need to] orient them regularly on the new developments because we change tools all the time.… As we go, as a district, probably to supervise public DTUs, we should be able to supervise the private.” (KI -2)

Many health workers reported that missing data or gaps in the tools were a common occurrence in their facilities which sometimes was a result of insufficient knowledge on how to fill the tools, shortage of human staff and lack of time.

“We need to be inducted in these books. So, you find people will miss some [information]…create gaps when we are filling in those tables. So, when the supervisors come, you will find some missing information; that was actually the biggest challenge. And there was no proper person responsible for that reporting—who follows those books closely.” (IDI-6)

Similar to the above was the inconsistency in reporting as expressed below:

“The challenge that we get, usually, (is) the reporting; the report is really big… sometimes they miss out many things, sometimes they don’t report at all” (KI-1)

The health workers concurred on the importance of incorporating TB in the facility CMEs since many health workers lack sufficient knowledge on diagnosis and reporting of TB. The Key Informants and private health workers also proposed health education among cases, their families and the communities to increase TB case finding from the community.

Logistics and resources for TB case notification

The health workers emphasized that the provision of funding for the case finding activities of community health workers helps them to be able to do TB case notification. Some reported that this support was previously provided by IPs but it had been withdrawn, and it was difficult for the facility health workers to go to the communities to do contact tracing or screen for TB among the people.

“Ideally [for] private clinics, one of the things I have seen, [is] if there could be a way that they can be motivated… A motivation can be airtime; because some of them can [say] like, ‘I wanted to call you but I had not airtime. We are lacking this and that…’ And remember that person is telling you when you are going to compile a report, meaning that we are not going to get a correct report.” (KI -3)

Another health worker also referred to the challenge the lack of funding for contact tracing saying:

“A client comes in, [and] we just ask them, “how many people do you stay with? Is there anybody coughing?” (Some) come, others don’t come so we miss out them. That’s the challenge also to us because we don’t have facilitation for contact tracing and community mobilization,…” (IDI-2)

Some private health workers also identified the shortage of staff to carryout notification (screening, data recording) and community case finding activities as a barrier to TB notification. Furthermore, unlike public facilities, private facilities were found to have a high staff turnover which results in a shortage of TB trained staff requiring regular training and CMEs.

“There is also another challenge. We keep on changing these staff. We don’t have a fixed person, that is, the person [in-charge of] TB screening. So, if you inducted one person, the next month the boss may bring another person who doesn’t even know how to notify TB, [or to] use registers.” (IDI-6)

Some health workers further reported the need for the government to provide the reporting tools, guidelines and IEC materials. In addition to this, they emphasized the challenge of lacking TB diagnostic equipment such as microscopes and especially GeneXpert.

“I think if we can get our own GeneXpert machine it could be better. because we can have many samples. ZN also takes long due to few staff we have in the lab. And even the X-ray, the money for X-ray some people cannot afford to go for X-ray and even scan. But if we had some funding like for x-ray or scan, maybe the notification would be more. We shall be having more cases.” (IDI-5)

Interestingly, the health workers also proposed digitalization of TB reporting systems as a way to improve TB case finding. This is because recording of TB data was felt to be tedious for the health workers and required employment of extra personnel, an extra expense for the facility.

Health workers noted that sometimes delays in Turn-Around-Time for GeneXpert test results led in Loss to Follow Up because the case had eventually moved from the area or even death because the case had approached the facility in a critical state. The shortage of time in some facilities also limited the ability to conduct screening appropriately on all people.

Limitations and future study

In this study, only Ministry of Health accredited facilities which report in DHIS2 were studied, other studies may be conducted among unaccredited facilities which also need to notify TB cases. In the survey, the observation method was not employed thus limiting the possibility of getting data on the presence or absence of equipment and infrastructure that can act as facilitators or barriers to TB case notification. Data collected on TB screening, diagnosis and reporting practices was self-reported which may have given rise to some reporting bias.