Performance of decentralised facilities in tuberculosis case notification and treatment success in Armenia

Abstract

We assessed the performance of decentralised tuberculosis (TB) out-patient centres in tuberculosis (TB) case notification and treatment success in Armenia. An average threshold case notification of ⩾37/100 000 was seen in centres that had higher numbers of presumptive TB patients, where more TB was diagnosed by in-patient facilities and where TB contacts were examined. The number of doctors and/or TB specialists at centres did not influence case notification. Onsite smear microscopy was significantly associated with a treatment success rate of ⩾85% for new TB patients. Addressing specific characteristics of TB centres associated with lower case notification and treatment success and optimising their location may improve performance.

In Armenia, the tuberculosis (TB) case detection rate is estimated to be 79% of expected incidence.¹ Although it is difficult to determine data on TB incidence, annual notifications of TB cases provide a good proxy indication of incidence in countries with high-performance surveillance systems with little underreporting of diagnosed cases and where few cases go undiagnosed.¹

Treatment success among sputum smear-positive TB cases in Armenia was 63%, far below the World Health Organization (WHO) target of 85%.¹ This poor outcome might be partly explained by the high prevalence of drug-resistant forms of TB; Armenia is among the world's 27 high multidrug-resistant TB (MDR-TB) burden countries.^1,2 Prompt diagnosis of patients with presumptive TB (formerly termed ‘TB suspects’)³ is essential to find missed cases, institute early anti-tuberculosis treatment, limit TB transmission and achieve high treatment success rates.⁴

In Armenia, the responsibility for TB diagnosis and treatment lies with in-patient and decentralised out-patient TB centres (TB clinics).⁵ In 2012, there were 72 TB clinics serving 11 administrative regions. The clinics vary in a number of ways, including the population covered, setting, staffing levels, availability of treatment for drug-resistant TB, availability of on-site direct sputum smear microscopy, hospitalisation facilities and referral of individuals with presumptive TB to and from the clinics.⁵ All of these factors may influence case notification rates, which ranged from 6 to 170 per 100 000 population in 2012.⁶ These factors may also have influenced levels of treatment success, which ranged from 50% to 100% in the same year.⁶

The present study aimed to assess TB clinic performance using two DOTS strategy outcomes: levels of TB notification and treatment success. Knowledge of clinic characteristics associated with higher TB case notification rates and treatment success might help efforts to improve the performance and quality of decentralised services, but this has never been assessed in Armenia.

In a countrywide audit, we thus assessed whether specific characteristics of TB clinics were associated with higher than average TB case notification and the desired treatment success target of ⩾85%.

METHODS

Design

This was an ecological study using 2012 programme data.

Study setting

Armenia has 11 regions with both urban and rural areas, with the exception of the capital region, Yerevan. TB diagnosis and care is provided free of charge under the National TB Programme (NTP). The DOTS strategy has been implemented throughout the country since 2002.⁷ TB clinics were introduced in 2006; their locations were decided based on the population covered and geographic criteria.⁸

TB management follows WHO guidelines. Treatment outcomes are standardised and include cured, treatment completed, died, lost to follow-up, failure and not evaluated.^2,5

Data collection, sources and statistical analysis

Data related to the study objectives were extracted from the national TB electronic database. Differences between groups were assessed using the χ² or Fisher's exact test for categorical variables and Student's t-test for continuous variables. TB case notification included new and recurrent TB cases during a given year. A case notification threshold of 37/100 000 was used to compare clinics (this cut-off was chosen based on the LOWESS [locally weighted scatterplot smoothing] smoothed plot, and on the mean case notification per clinic in 2012).⁹ Analyses were conducted using STATA 10 (Stata Corp, College Station, TX, USA); the level of confidence was set at 5%.

Ethics

Ethics approval was obtained from the Institutional Review Board/Committee on Human Research, American University of Armenia, Yerevan, Armenia, and the Ethics Advisory Group of the International Union Against Tuberculosis and Lung Disease, Paris, France.

RESULTS

Table 1 shows the characteristics of TB clinics associated with high or low TB case notification. A high case notification was seen in clinics that had higher numbers of presumptive TB patients, where more TB was diagnosed by in-patient facilities and where more TB contacts were examined. Both clinics with high and low case notification rates had similar patterns of referral of presumptive TB cases to in-patient facilities. Clinics located in the capital city of Yerevan and in neighbouring areas also had higher case notification rates. The number of physicians present and the presence of a TB specialist at clinic level did not influence case notification. A total of 12 (17%) clinics had <5 new TB cases registered in 2012.

TABLE 1.

Characteristics of TB clinics associated with a TB case notification rate of <37/100 000 population or ⩾ 37/100 000 population in Armenia, 2012

Table 2 shows the characteristics of TB clinics associated with treatment success (<85% or ⩾85%). The presence of an onsite smear microscopy service was significantly associated with treatment success ⩾85%.

TABLE 2.

Characteristics of TB clinics associated with a treatment success rate of <85% or ⩾85% for new TB cases in Armenia, 2012

DISCUSSION

This first countrywide evaluation of decentralised TB clinics in Armenia has shown that differences in TB case notification and treatment success were associated with specific characteristics of TB clinics. The low number of TB cases (<5) in several clinics brings into question their suitability in terms of numbers and location.

Increased case notification was significantly associated with clinics that had received more presumptive TB cases and which examined more TB contacts, implying that a minimum number of referrals of such cases is required if a minimum case notification rate is to be achieved by all centres. This highlights the importance of contact screening and improving referrals of presumptive TB from primary health centres to TB clinics. Low case notification rates may be due to a real difference in TB incidence and/or to missed cases due to limitations in the diagnostic and reporting pathway. In addition, acceptability of health services by patients and access issues may influence case notification; however, these aspects could not be assessed in this study.

Achieving a treatment success of 85% was associated with the presence of an onsite TB sputum smear microscopy facility, which may indicate earlier TB diagnosis and treatment initiation; as with case notification, acceptability of clinic services by patients, attitudes of health staff and access hurdles may also affect treatment outcomes. Qualitative research is needed to gain greater insight into the role played by these factors. Interestingly, neither the number of physicians nor the presence of a TB specialist in the clinics affected case notification or treatment success. This again implies that high-level staffing without the implementation of adequate referral and screening standards is not enough.

The strengths of the study are that data from all TB clinics in the country were included and the findings come from routine settings, and are thus likely to reflect the real situation in the field. This study also responds to operational research priorities expressed in the Armenian NTP review.² The limitations of the study were that statistical power was low due to the small number of clinics (the unit of study), and this may have compromised finding statistically significant differences between groups. Furthermore, the study period was limited to 1 year, and the variables analysed were limited to routinely collected data. The ecological study design is also a weakness. Future studies may consider assessing TB prevalence by clinic catchment area, and subsequently clinic performance in detecting these cases. Finally, we do not know if differences in TB notification were influenced by the prevalence of human immu-nodeficiency virus infection.

Despite these limitations, there are a number of policy and practice implications. First, the fact that 12 TB clinics diagnosed fewer than five new TB cases per year raises the question as to whether or not the clinics are located in the right places and/or if there are too many of them. Other models of care, including community-based case finding and ambulatory treatment, may be justified, although this needs operational evaluation. Second, primary health care services maximise their referrals of presumptive TB cases to TB clinics. Increased vigilance in screening presumptive TB cases and contacts and enhancing the diagnosis of TB at out-patient facilities is also needed. As the overall expected incidence of TB in Armenia lies below expected targets, it is logical to think that there might also be underlying problems and/or other factors such as the acceptability of health services by patients, access issues or limitations in the quality of TB screening, referral and diagnosis, that may have influenced this parameter. This merits further qualitative and quantitative research. Areas further from the capital city and rural areas also need specific attention.

In conclusion, we have identified a number of TB clinic characteristics associated with lower case notification and treatment success. For TB case notification, there seems to be a benefit in contact screening and the need to investigate the higher TB burden in the capital. For treatment success, an investigation of diagnostic and treatment delays and the impact on treatment outcomes seems warranted. Such information can inform efforts towards optimising clinic location and improving TB clinic performance in Armenia.