Tuberculosis Treatment Outcomes Among Prisoners and Non-Prisoners in Zomba, Malawi

Background TB remains a major global health problem. it is particularly prevalent in prisons in sub-Saharan Africa. Prisoners have experienced worse TB treatment outcomes than the general population. The researchers investigated the TB treatment outcomes and pre d ictors of unsuccessful treatment outcomes among prisoners and the general population. Methods We retrospectively reviewed TB registers of prisoners and the general population diagnosed with TB from January 2011 to December 2016 at Zomba Maximum Central Prison and Zomba Central Hospital, Malawi. The study used routinely collected data extracted from national, standardized TB treatment monitoring tools. Successful treatment outcome was classified as the total for cured and completed treatment while unsuccessful treatment outcome was classified as the total of deaths and treatment failures. We used descriptive statistics to compare the demographics and TB treatment parameters among prisoners and non – prisoners and computed multivariate analysis to predict the independent factors of unsuccessful treatment outcomes. Results Of 1652 registered cases, 27% were prisoners (all males) and 72% were non-prisoners (58 % males). The median age was 35 years (IQR: 29 - 42); 76 % were Pulmonary TB cases (78 % among prisoners vs 75% among non-prisoners); 83 % were new TB cases (77% among prisoners vs 86% among non-prisoners); and 65 % were HIV positive (50 % among prisoners vs 71% among non – prisoners). Regarding treatment outcome, 1472 (89%) were cured and/or completed treatment (93% among prisoners vs 88 % among non – prisoners), 2 (0.2%) were treatment failures, 122 (8%) died (5 % among prisoners vs 8 % among non – prisoners) and 55 (3%) were not evaluated (1% among prisoners vs 4% among non – prisoners). Unsuccessful TB treatment outcomes were associated with age greater than 35 years (aOR = 0.68: 95% C.I: 0.58 – 0.80), Extra-Pulmonary TB (aOR=1.69: 95% C.I: 1.08 – 2.63) andHIV positive status (aOR = 0.63: 95 % C.I: 0.42 – 0.94). Conclusion Maximum prisons provide a stable population that can be easily monitored during TB treatment. Good TB treatment outcomes comparable to the general population can be achieved among prisoners despite the challenging prison conditions.

Tuberculosis (TB) remains a major global health problem. Globally in 2018, there were an estimated 10 million new TB cases of which 8.6 % were among People living with HIV [1]. There were an estimated 1.4 million TB deaths worldwide with 17 % (0.2 million deaths) resulting from TB disease among people living with HIV [1]. TB is the most common severe opportunistic infection associated with HIV in many resource-limited countries [2] with 68 % of incident TB cases reported in the World Health Organization (WHO) South East Asia and African [1]. TB is more in prison populations than in the general populations [2]and HIV epidemic, poor living conditions, low socioeconomic status and poor general health, appears to have contributed directly to higher rates of TB within prisons [2].
Prisons therefore serve as reservoirs for TB infection, not only for inmates and staff but also for the general population as visitors come and go and released prisoners return to society [2].
In 2018, Malawi reported an estimated 15,000 TB case notifications (181 cases per 100,000 population) with a 48% HIV/TB co-infection rate (99% on ART{anti-retroviral therapy}) and 86% TB treatment success rate [1]. Malawi is one of the focus countries in the End TB Strategy era (2016-2035), as it is in the top 30 countries of high TB incidence among people living with HIV [1].
Data from the WHO epidemiological surveillance for TB treatment outcomes of 2017 cohort showed TSR of 76 % in the WHO Region of the Americas( due to high levels of loss to follow up and missing data), 78 % in the European Region(due to high rates of treatment failure and deaths, from high rates of MDR-TB),; 83% in South East Asia; and82 % in sub -Saharan Africa . In contrast, Prison settings have shown sub-optimal TB treatment success rates than the recommended 90 % set by the End TB Strategy. TB treatment outcomes' studies among prisoners showed TSR of 50 %, 45 %, 48 % in Brazil(2011) [3], South Africa(2009) [4] and Uganda(2012) [ The Zomba Central Prison is overcrowded (houses approximately 2500 prisoners at any point in time for a facility meant for 800 prisoners). Despite the evidence on the higher TB cases in prisons than the general population, TB TSR are not disaggregated according to prisons TB TSR serves as a proxy for the quality of TB control as a secondary prevention. Therefore, we investigated the TB treatment outcomes and predictors of unsuccessful treatment outcomes among prisoners and non -prisoners to audit the effectiveness of TB control in identifying gaps in the national treatment policy and practice in prisons to initiate evidence based practice. .

Study design and setting
A retrospective cross-sectional study was conducted at the Zomba Central Prison (ZCP) and Zomba Central Hospital (ZCH), Malawi. The sites were purposefully selected. Zomba Central Prison is the maximum security prison with the highest prison population in Malawi. On average, it houses approximately 2500 prisoners in a facility meant for 800 prisoners. Zomba Central Hospital is a tertiary semi -urban referral hospital (catchment area 3.1 million) for the South East Health Zone in Zomba, Malawi.
The data source was the National TB register where relevant clinical information for all patients diagnosed with active TB disease are put on anti-TB treatment regimens and monitored throughout the course of their treatment.

Organization of Health and TB services in Malawi
Malawi follows a three tier health care system which is connected by a patient referral system. Primary care where the bulk of health care occurs consists of hospitals that provide out-patient services, holding wards and community based outreach services. Secondary level care is provided by district hospitals in 28 districts which provide the same basic services as the primary care facilities in addition to a few more, such as: radiology, ambulance, operating theatre and basic laboratory. The central hospitals (located in 4 urban areas) provide tertiary specialized services.
There is decentralized provision of TB and HIV services. TB diagnostic services include smear microscopy (widely available in primary care facilities) and GeneXpert MTB/RIF (district and central hospitals) on spot and morning sputum samples. The National TB Reference Laboratory provides high level diagnostic services, including solid and liquid culture and drug sensitivity testing. It is also responsible for quality assurance services to the peripheral laboratories (central and district TB services data is captured on paper based registers: chronic cough register for presumptive TB cases, district TB register for active TB cases and TB patient cards for individual monitoring of TB treatment. The central prisons have their own TB treatment registers (same as the district TB registers). All prisoners diagnosed with TB are also captured in the district TB treatment register so that the district can keep track of the prisoners. Data is collected on a quarterly basis by the National TB program through an integrated HIV/TB program supervision.

Study Participants
All prisoners and the general population aged 15 years and above who were diagnosed and put on TB treatment at the Zomba Central Prison and Zomba Central Hospital from January 2011 to December 2016 were included ( Figure 1).

Ethical consideration
Ethical approval was obtained from the College of Medicine Research Ethics Committee (COMREC Reference No: P.05/18/2394). Individual informed consent was not sought since the study used fully anonymized retrospective data. Prior written authorization was obtained from Zomba Central Hospital, Zomba District Health Office and Malawi Prison Services.

Data collection, management and analysis
We used routinely collected data extracted anonymously from the national standardized monitoring tools in ZCP and ZCH: TB register (age, sex, occupation, site of disease, history of previous treatment, type of diagnosis, treatment outcome, HIV test status, Anti -Retroviral Therapy {ART} and Cotrimoxazole Prophylactic Treatment {CPT} status). The data was collected in an anonymized way, entered into a data extraction sheet and then into the Access database then analyzed by STATA 13.
Descriptive analysis was used to determine the patient characteristics and treatment outcomes. Chi squared tests were used to compare the characteristics between the prisoners and the general population. Bivariate logistic regression was used to assess associations between patient characteristics and poor treatment outcomes. The risk factors for poor TB treatment outcomes were analyzed using multivariate analyses adjusted for potential confounders. Odd ratios were provided with 95% confidence intervals and p-values (using a statistical significance level of <0.05). Table 2 shows the baseline characteristics of the 1652 registered adult TB patients. Twenty seven percent (27%) were prisoners (all males) and 72% representeda general population (58 % males). The median age was 35 years (IQR: 29 -42); 76 % were Pulmonary TB cases (78 % among prisoners vs 75% among general population); 83 % were new TB cases (77% among prisoners vs 86% among general population); and 65 % were HIV positive (50 % among prisoners vs 71% among general population).

Trends in TB registrations and treatment outcomes
Overall all forms of TB were gradually decreasing among the prisoners from 2011 to 2013, and the to 2016. EPTB cases in the two populations remained stable over the 6 year period (Figure 2). There was a difference in the treatment outcomes among the two populations(X 2 = 33.2; P < 0.05).
In the general population, the trend for the TSR declined in 2012 and gradually increased from 2013 to 2016 while it remained fairly constant among the prisoners over the study period. (Figure 3 and Table 2)  (Table 3).

Discussion
In this 6 year retrospective study, the results show differences in the trend of all forms of TB and TB treatment success rates among the prisoners and the general population. . Age greater than 35 years, HIV positive status and extra -Pulmonary TB were independently associated with unsuccessful TB treatment outcomes (death and treatment failure). In addition, it was found that the prison only registered male TB patients over the course of the 6 years.  Hospital[11] and a Central Hospital [12] in Malawi and Ethiopia showed an 86 % TSR ( [13]. This is also higher than the 73 % TSR [14] from 27 prisons in Malawi in 2007, 66% TSR reported in Zambian prison [15] in 2010 -2011 and 48 % TSR in Ugandan prisons [5] from 2011 to 2012. However, the TSR is comparable to recent studies done in prisons in Ethiopia [16] (89%, 90 %) but slightly lower than the study done in South Africa [4] (92 %) and Ethiopia [16] (94 %). The higher TSR [15]. The varied intra prison differences may be due to the high turnover of prisoners and remandees(they are not yet sentenced and can be transferred out to another facility or released into the general public where they are lost to follow up) [15]. There can be poor linkage within prison facilities and the communities as once prisoners are released into the community, they are not adequately actively followed up for continuity of [17,18].
This study showed an overall 2.6 % loss to follow-up which was lower than in other studies reported in similar limited resource setting in prisons in Uganda (43.0%) [5], and Brazil (13.0%) [3]. The lower rate of loss to follow up is attributed to the stable prison population at the Zomba Central Prison that is serving long term sentences hence not easily transferred or released [15].
The 5.5 % death rate is comparable to the death rate observed in Uganda prison[5] but higher than the 1.4 % death rate recorded in Ethiopia [16], 1.8 % death rate recorded in South Africa [4]and 2 % death rate recorded in Brazil [3]. This high rate could be attributed to the poor prison living conditions -overcrowding [17], poor nutrition [18] and possibly to the rate of HIV/TB co-infection without use of antiviral therapy (due to the lower coverage of antiretroviral therapy use during this time period), which has been shown to be associated with unsuccessful treatment outcome [11,19]. This study also showed that unsuccessful TB treatment outcome was associated with age greater than 35 years old.
Similar studies in Botswana and Nigeria have shown that older age may be confounded by the high risk behaviors .e.g. alcohol and drug use which are common in prison settings and general population especially among men. This leads to poor adherence which result to poor outcomes.
A commonisk factor for unsuccessful TB treatment outcome, Extra Pulmonary TB was also found in this study. This might be due to the severe nature of the , delays in diagnosing EPTB [20] duelimited diagnostic capacity and lack of treatment monitoring tests for EPTB cases [21].
In this study, gender comparisons were not possible among prisoners and the general population because the prison registered male TB patients only in the whole study period. Similar studies globally have shown unevenly distributed high proportions of male prisoners than female prisoners( [18,22,23]).Globally, there has been no universally accepted explanation to this disparity and most criminology studies point to the socialized gender roles and different expectations of male and female behaviours [24].
The general findings from this study shows that prisoners can achieve good TB treatment outcomes which are comparable to the general population. The long serving prisoners in a maximum security prison provides an opportunity of uninterrupted treatment since they are not easily transferred out to other facilities. This calls for continued support to to the prison health programs [25]. Several strategies can be combined to yield successful efforts in the fight against TB in the prison setting. TB mass screenings can be used to increase TB case detection [23]. Since the TB treatment outcomes are worsened by HIV positive status [4,19,26] scaling up of the 3 phase integrated screening and treatment for HIV, TB and nutrition during entry, stay and exit of prison could increase access to HIV, TB and Nutritional screening and subsequent linkage to appropriate treatment services.
The strengths of this study include the following that we used routine program based data that is collected from the facility national TB registers which are used for TB patient registration, monitoring and evaluation from within a national public system. . We studied a large number of TB patients and the results are representative of the national TB TSR The study adhered to STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines for the reporting of observational data.
Some noted limitations include the use of retrospective data in which other important risk factors for unsuccessful TB treatment outcomes were not assessed (smoking status, alcohol status, body mass index and TB drug side effects). The analysis also excluded patients who were transferred out and those not evaluated which might posed a bias.

Availability of data and materials
The datasets used during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no conflict of interest.

Funding
Dignitas International's work at the Zomba Central Hospital and Zomba Central Prison was funded by USAID. USAID had no role in the study design, data collection, data analysis, decision to publish or preparation of the manuscript.

Authors' contributions
VS and EK designed the study and analyzed the data. VS and EK wrote the first draft of the manuscript. LC and WC contributed to subsequent drafts and have reviewed and agreed with the content of the final manuscript. Treatment failed TB patient whose sputum smear or culture is positive at month 5 or later durin Died TB patient who dies for any reason before starting or during the course of trea Loss To Follow Up TB patient who did not start treatment or whose treatment was interrupted fo Not Evaluated TB patient for whom no treatment outcome is assigned. This includes cases " whom the treatment outcome is unknown to the reporting unit. * The percentage in the 5 year trend section is provided as a percentage of the total number over 6 years.