Abstract
Introduction. Private practitioners (PPs) in Indonesia, and other settings, are often the first point of contact with the health system for individuals with presumptive tuberculosis (TB). However, PPs tend to practice substandard delivery of TB care and not to notify TB patients that they diagnose, making evaluation of the quality and outcome of care difficult. The thesis sought to understand how PPs in Indonesia manage TB cases, to evaluate whether a public health intervention package can improve TB case notification from this provider, and to understand how notification systems can be designed more in line with provider’s preferences.
Methods. This thesis employs several methodological approaches. First, standardised patients (SPs) were used to assess PP’s adherence to the TB care guidelines in Bandung, Indonesia. From July 2018 to April 2019, twelve trained SPs presented four TB case scenarios over 341 interactions with PPs and providers at the Community Health Centres (CHCs). Second, a cluster-randomised controlled trial was conducted from July 2020 to August 2022 in 30 CHC areas in Bandung to evaluate the effect of implementing a public health intervention package to PPs on area-level TB notification rates. Lastly, an internet-based survey was conducted in October 2023 to assess physician preferences, as measured by their willingness-to-accept (WTA), for features of a TB notification system, estimated using a discrete-choice experiment (DCE) approach.
Results. The SP study revealed that utilisation of sputum examination for TB diagnosis was low among private general practitioners (GPs) [16/52(31%)] and private specialists [3/15(20%)]. By contrast, sputum examination was requested in 87% (26/30) of encounters with CHCs. PPs preferred the use of chest X-ray (CXR) to screen patients with presumptive TB. CXR were requested in 60% (31/52) of encounters with private GPs, 87% (13/15) in private specialist encounters, and in 8% of CHC encounters. We also found incorrect prescription of TB drugs in 7% and 13% of encounters with private GPs and specialists, respectively.
In the randomised trial, 105 PPs in 13 CHC areas in Bandung received the intervention package. Over an 18-month follow-up observation period, 312 presumptive and confirmed TB patients were reported to the study; 86 (28%) of whom were eventually notified as TB cases. However, the intervention did not significantly increase the TB notification rate (Rate Ratio (RR):0.9; 95%CI 0.6-1.5) or increase the proportion of bacteriology positive cases (Proportion Odds Ratio: 1.0; 95%CI 0.8-1.4). An increase in notification rates occurred during the observation period for both study arms, especially during the second half that coincided with reduced Covid-19 transmission in the country (first half RR 1.2, 95%CI 0.8-1.8; second half RR 2.0, 95%CI 1.4-3.0).
The DCE assessed preference for four TB notification system features: notification time burden, potential penalty for failure to notify, amount of continuing medical education (CME) credits, and monetary incentive awarded for notifying cases. A total of 210 physicians participated in the internet-based survey, 29 (21%) of whom were excluded due to possible non-attentiveness. PPs saw TB case notification as important and that, in economic terms, notifying TB cases to the NTP gives a positive utility for the respondents. However, notifying TB cases was also associated with the highest marginal willingness-to-accept (mWTA), suggesting a high preference for remaining in the status quo, i.e., not notifying cases, in contrast to relatively low mWTA for other notification features. Overall, minimum amount of monetary compensation that the survey participants were willing to accept for initiating notification was Indonesian Rupiah (IDR)435,000, IDR57,000 for a 10-minute notification burden, IDR53,000 to forego 1 CME credit, and IDR 71,000 to accept a system that imposed a penalty for failure to notify. Significant heterogeneity in preferences was identified.
Conclusions. Indonesian PPs show a preference for CXR over microbiological testing for triaging presumptive TB patients, and inappropriate TB drug prescriptions are not uncommon. This provides evidence of an increased risk of receiving substandard TB care associated with visiting PPs. However, interventions targeting PPs may not significantly impact case notification rates due to their low TB patient volume. In addition, increasing participation of PPs in notifying TB cases may require substantial incentive to compensate for their efforts. These findings suggest actions to increase awareness among PPs about the applicable standards of TB care and interventions aimed at increasing their participation in TB case notification should be designed and implemented strategically to maximise their impacts.