livrepository.liverpool.ac.uklivrepository.liverpool.ac.uk/3005371/1/Bull_WHO... · Web viewJohns...
Transcript of livrepository.liverpool.ac.uklivrepository.liverpool.ac.uk/3005371/1/Bull_WHO... · Web viewJohns...
Title Page: Wingfield et al Original Research Article
Title: A household-randomized controlled evaluation of socioeconomic support to improve tuberculosis preventive therapy initiation and increase tuberculosis treatment success, Peru
Authors: Wingfield T (MRCP DTMH PhD),1,2,3,4 Tovar MA (MD MSc),2,5 Huff D (MPHTM PA-C),2,6 Boccia D
(MSc PhD),2,7 Montoya R (RGN),2 Ramos E (MSc),5 Datta S (MD),1,2 Saunders MJ,1,2,5 Lewis JJ (MSc PhD),2,7
Gilman RH (MD DTMH PhD),8 Evans CA (FRCP DTMH PhD)1,2,5
Author affiliations:
1) Infectious Diseases & Immunity, Imperial College London, and Wellcome Trust Imperial College
Centre for Global Health Research, London, UK
2) Innovación Por la Salud Y Desarrollo (IPSYD), Asociación Benéfica PRISMA, Lima, Perú
3) The Monsall Infectious Diseases Unit, North Manchester General Hospital, Manchester, UK
4) Institute of Infection and Global Health, University of Liverpool, Liverpool, UK
5) Innovation For Health And Development (IFHAD), Laboratory of Research and Development,
Universidad Peruana Cayetano Heredia, Lima, Perú
6) Tulane University School of Public Health and Tropical Medicine, New Orleans, USA
7) Department of Infectious Disease Epidemiology, London School of Hygiene & Tropical Medicine,
London, UK
8) Johns Hopkins Bloomberg School of Public Health, Baltimore, USA
Corresponding Author: Dr Tom Wingfield, NIHR Academic Clinical Lecturer and ST7 Specialist Registrar
in Infectious Diseases, Institute of Infection and Global Health, Ronald Ross Building, 8 West Derby
Road, Liverpool, L69 7BE. Tel: 0151 7959667. Email: [email protected]
Short running title: A randomized controlled evaluation of socioeconomic support to enhance TB
prevention measures and TB treatment success
Key words: TB; social protection; socioeconomic support; social determinants; poverty; conditional
cash transfers; TB control; TB prevention; TB treatment success; WHO End TB Strategy
Manuscript word count: 3046
Table: 1 Figures: 5 Box: 1
Statement of competing interests: all authors declare no conflicting or competing interests
1
Research in context
Evidence before this study: The World Health Organization (WHO) post-2015 End TB Strategy highlights socioeconomic support as a key pillar in the global response to TB. To identify existing evidence concerning such interventions, we searched PubMed using the terms “Tuberculosis/economics”[Mesh] OR “Tuberculosis, pulmonary/economics”[Mesh] OR “Tuberculosis/prevention and control”[Mesh] AND “Economic support/intervention”[Mesh] OR “Social support/intervention”[Mesh] OR “Socioeconomic support/intervention”[Mesh] OR “Cash transfers”. Articles were excluded that concerned socioeconomic support which: related specifically to HIV/ AIDS (such as the IMAGE study);1 did not have a control group or impact assessment;2,3 or provided economic support in the form of food4 rather than cash or bank transfers. This search found no randomized controlled evaluations of integrated socioeconomic support to improve TB preventive therapy or TB treatment but did find one randomized controlled evaluation of economic support alone to improve tuberculosis treatment outcomes, from South Africa. 5 The study showed no improvement in TB treatment success rates in patients who were randomized to the economic intervention, which consisted of monthly vouchers throughout treatment. However, the impact was limited by low fidelity to the intervention, which lacked a social support element. This is relevant because research in the field of HIV has suggested that economic support complemented by social support (e.g. education, information, and mutual-support groups) is likely to have a greater impact on health outcomes.6
New knowledge contributed by this study: Building on a decade of research addressing the social determinants of TB in Peru,7–11 we aimed to generate rigorous evidence on the impact of socioeconomic support on TB control to inform the WHO End TB strategy. To do this, we carried out one of the world’s first randomized controlled evaluations of a TB-specific integrated socioeconomic support intervention to improve TB prevention measures and increase TB treatment success. Our novel intervention was acceptable and feasible in the challenging setting of impoverished shantytowns of Callao, Peru.12 Moreover, the intervention had a measurable positive impact on health and economic outcomes by: reducing the likelihood of TB-affected household’s incurring catastrophic TB-related costs;13 increasing treatment success in TB patients; and improving initiation of TB preventive therapy in household contacts of these TB patients. These findings expand the limited literature supporting the use of economic support for TB-affected households and provide preliminary, rigorous evidence to inform implementation of socioeconomic support as mandated by the End TB strategy. The next step is to evaluate the impact of the intervention on TB prevention during the Community Randomized Evaluation of a Socioeconomic Intervention to Prevent TB (CRESIPT) project.
2
Abstract (250 words)
Objective. The End TB Strategy advocates socioeconomic support for TB-affected households but
there is limited impact assessment. We evaluated the impact of socioeconomic support on TB:
preventive therapy initiation in contacts; and treatment success in patients, and refined the support
for the “Community Randomized Evaluation of a Socioeconomic Intervention to Prevent TB”
(CRESIPT) project.
Methods. Design: An un-blinded household-randomized controlled study.
Setting: 32 shantytown communities, Peru.
Participants: All consenting patients treated for TB disease by the Peruvian TB Program and their
contacts aged <20 years.
Randomisation: Patient households were randomly assigned to control (received Peruvian TB
Program standard of care) or intervention (additionally received socioeconomic support) arms.
Socioeconomic support: consisted of economic support (conditional cash transfers ≤230 United
States dollars) integrated with social support (household visits and community meetings)
throughout TB treatment.
Outcomes: compared intervention versus control households using logistic regression adjusting for
household-clustering.
Findings. Between February 2014 and August 2014, 90% (282/312) of patients participated, 135
randomized to intervention and 147 to control arms. Primary outcome was TB preventive therapy
initiation in contacts aged <20 years (n=410). This increased from 25% in controls to 43% in the
intervention arm (adjusted odds ratio, aOR=2.2, 95% confidence interval, CI=1.1-4.2, p=0.02).
Secondary outcome was an intention-to-treat analysis of patient TB treatment success (cure or
completed treatment). This increased from 53% in controls to 64% in the intervention arm (OR=1.6,
95%CI=1.0-2.6, p=0.05). This increase was equitable, independent of poverty.
Conclusion. A TB-specific socioeconomic support intervention improved TB preventive therapy
initiation and TB treatment success.
Funding. Joint Global Health Trials consortium of Wellcome Trust, Medical Research Council, and
Department For International Development; British Infection Association; Wellcome Trust; Bill and
Melinda Gates Foundation; and the charity Innovation For Health And Development (IFHAD)
3
Introduction
It has been estimated that one third of the world’s population has latent tuberculosis (TB) infection.
In 2015, 10.4 million people developed TB disease.14 Those at highest risk of TB disease include:
household contacts of patients with TB and people living in poverty.15 A number of trials have shown
that TB preventive therapy decreases the risk of progression to TB disease by 60-90%.15–17 Despite
this proven benefit, the global impact of preventive therapy on TB control is severely limited
because people with latent TB infection are infrequently identified18 and infrequently take TB
preventive therapy,7,19 hampering global TB control.20,21 Difficulty adhering to medicines is common
and reduces their intended benefits.19,20,22 TB patients who do not take adequate TB treatment are
more likely to: have adverse outcomes (including death, treatment failure, and TB recurrence);23
transmit TB, especially to households contacts;24 and develop multi-drug resistant TB (MDR-TB),25 an
increasing global public health threat.18
The current predominantly biomedical approach to TB control is not reducing TB incidence rates to
levels required to eliminate TB as envisioned in the World Health Organisation’ (WHO) End TB
Strategy.26,27 Enhancing access to and initiation of TB preventive therapy and TB treatment is likely to
improve TB prevention and treatment success but requires strategies complementary to biomedical
care, including socioeconomic support. Social protection interventions include conditional cash
transfers and aim to reduce poverty and vulnerability by improving people’s capacity to manage
social and/or economic risks.8,28–34 Although these and socioeconomic interventions are common
tools in HIV/AIDS and maternal health,6,35 there is extremely limited evidence of their impact on TB
care or prevention.8,29,30,36
Our research group (www.ifhad.org) has been funded to undertake the Community Randomized
Evaluation of a Socioeconomic Intervention to Prevent TB (CRESIPT) project. The planning, design,
and economic impact of the intervention have been described.12,13 Here we report the final results of
the initial phase of CRESIPT, consisting of: a household-randomized controlled study to evaluate the
impact of TB-specific socioeconomic support on TB preventive therapy initiation and TB treatment
success, and refine the intervention for CRESIPT.
4
Methods
Setting. Thirty-two contiguous shantytowns in Callao, Peru, the northern coast extension of the capital
Lima. This regions has a population of one million, considerable poverty, zones with frequent drug
addiction and gun crime, and an annual TB case notification rate of 123 new cases/100,000
people/year in 2014, the highest rate nationally (Figure 1).37
Design. This household-randomized controlled study evaluated the impact of a socioeconomic
support intervention (Box 1) on TB preventive therapy uptake by patients’ household contacts aged
<20 years and treatment success for patients with TB.
Participants. Inclusion criteria were households of patients commencing treatment for TB disease
administered by the Peruvian National TB Program (NTP) in the study communities. These
households were invited to participate with an informed written consent form that explained the
randomization process, completed by the patient on the household’s behalf. For patients who were
minors, a parent or guardian was invited to give informed written consent and the patient was
invited to assent. Exclusion criteria were inability or unwillingness to consent. During a household
visit, individuals who the patient had reported as being in the same house as the patient for over 6
hours/week in the two weeks prior to the patient’s TB diagnosis were identified and validated and
are henceforth described as “contacts”. Any contacts declared or discovered following
randomization (but not during initial recruitment) were not included in the analysis nor invited to
participate in the intervention.
Randomization was done following recruitment of patient’s households. Randomization was
performed using random number tables, which generated an individual household randomization
sequence for each health post restricted in blocks of 30. The randomization assigned a patient
household at a ratio of 1:1 to either the control arm in which households received Peruvian NTP
standard of care, or the intervention arm in which households additionally received the integrated
socioeconomic support package. The mechanism for allocation concealment consisted of cards
placed in pre-numbered sealed envelopes detailing allocation to the intervention or control arm.
Once patients gave informed consent to participate, a project nurse opened the envelope in front
of the patient and informed them of their household’s allocation.
Data collection. A locally-validated questionnaire7,10 was used to collect health, wellbeing, and
socio-demographic data at baseline (the time that treatment commenced) and again 24 weeks later
(28 weeks if treatment was prolonged, including due to suboptimal treatment adherence). Both
interviews included measurement of height and weight to calculate body mass index (BMI), and
assessment of socioeconomic position.10,12
5
Preventive therapy. The Peruvian NTP guidelines that applied throughout this study recommended
that for contacts of patients with pulmonary TB that was not known to be caused by MDR-TB,
preventive therapy should be provided to all contacts under five years old (unless the contact was
known to have had previous TB disease) without tuberculin skin testing, and be considered for
contacts aged five to 19 years who have a positive tuberculin skin test.10 However, tuberculin was
generally unavailable throughout this study. Preventive therapy consisted of a six-month course of
daily isoniazid collected weekly from health posts and taken unsupervised in their home.10 Data
concerning preventive therapy initiation, adherence, and completion were collected collaboratively
from Peruvian NTP records, which included the number of weeks of preventive therapy collected
(herein defined as preventive therapy taken) from the health-post for each household contact.
Treatment. The Peruvian NTP offered free TB diagnostic testing to all people with symptoms
suggestive of TB who, if diagnosed, received free anti-TB treatment with directly observed therapy
(DOT) of every dose, which were administered at health posts.10 As an incentive and benefit for
participation, all patients, regardless of their randomisation, were offered an additional sputum test
with Xpert MTB/rif™ performed by our research laboratory for rapid rifampicin susceptibility
testing, which was not otherwise routinely available. TB treatment outcome data were collected
collaboratively with the Peruvian NTP from each individual patient’s treatment card at the time of
final follow-up and were not influenced by this research.
Treatment outcome definitions. Peruvian NTP outcomes were consistent with those defined by
WHO.14 The Peruvian NTP guidelines defined “cure” for patients with bacteriologically-confirmed
drug-susceptible TB at diagnosis if they completed treatment, had a negative sputum smear during
the final month of treatment, and a favourable clinical assessment by an NTP physician who
evaluated symptoms, examination, weight, and when necessary chest radiographs and blood tests.10
Patients who completed their TB treatment course without evidence of treatment failure but did not
complete the required sputum testing and/or physician review were classified as having “completed
treatment”. Patients with “treatment success” were those who completed treatment or were cured.
Other outcomes consistent with WHO guidance were “death” (all-cause mortality before starting or
during TB treatment); and “treatment failure” (positive sputum microscopy or culture at month five
or later of treatment). Patients whose treatment was interrupted for at least 30 consecutive days or
discontinued treatment having taken less than 30 days were defined as “lost to follow-up” (that was
locally termed “abandoned” and was shorter than the two months or more treatment interruption
in the WHO definition). Treatment outcome could not be assessed in patients who had no treatment
outcome assigned including cases transferred out to another treatment unit or patients who were
6
still on treatment at our 28 week follow-up interview (including MDR-TB patients whose treatment
duration commonly extended to 24 months).
Primary study outcome was TB preventive therapy initiation in contacts aged <20 years who were
available for follow-up. This compared intervention versus control households.
Secondary study outcome was TB treatment success (cure or completed treatment, as defined
above) in patients with TB analysed on an intention-to-treat basis, including patients with an
unknown outcome, using data from NTP records. This compared intervention versus control
households.10
Sample size calculations estimated that a study with 400 household contacts would have 80%
statistical power to detect a 50% increase in the primary outcome comparing intervention versus
control households with two-sided 5% significance, based on data from the study site.7
Blinding. It was not feasible for households or the research team to be blinded to allocation.
Peruvian NTP staff were not informed of and were generally unaware of household allocation, but
were not confirmed to be blinded.
Analysis. Socioeconomic variables were combined into a composite index of household
socioeconomic position using principal component analysis, as described.10 Data at the household
and individual (patient and contact) level were used to analyse study outcomes. The primary and
secondary outcomes were analysed by crude univariable logistic regression. For the primary study
outcome, multivariable logistic regression was also used to adjust for household clustering with
robust standard errors, generating adjusted odds ratios (aOR). Time-to-event analysis was
performed to generate a crude, unadjusted log-rank value comparing the number of weeks of
preventive therapy taken by contacts from intervention versus supported households.
Approvals included the ethics committees of DIRESA Callao (regional ministry of health), Imperial
College London, UK, and Asociación Benéfica PRISMA, Peru.
7
Results
Participants. Recruitment commenced 10th February 2014, the intended sample size was reached
14th August 2014, and follow-up was completed 1st June 2015. 312 households of patients with TB
were invited to participate, 90% (282/312) were recruited and randomized to the intervention
(n=135) and control (n=147) arms. 9% (24/282) of patients had MDR-TB, none of whom completed
treatment during the study. Patients from the 282 recruited households had a total of 1297
contacts (mean average five contacts per household). Of these contacts, 40% (518/1297) were aged
under 20 years old of whom 79% (410/518) completed follow-up (Figure 2). There were no
substantive imbalances found between households randomised to the intervention or control arm
(Table 1).
Intervention. 90% (122/135) of households randomized to the intervention arm received at least
one conditional cash transfer. A total of 890 conditional cash transfers were made (80% of potential
conditional cash transfers) with an average total of 520 Peruvian Soles (186 US Dollars) received per
household of a maximum 640 Peruvian Soles (230 dollars) available per household.10,12
Primary outcome of TB preventive therapy initiation in contacts aged <20 years was 25% in the
control arm and 43% in the intervention arm. This difference was statistically significant, both in
crude analysis (OR=2.2, 95%CI=1.4-3.3, p<0.001) and analysis adjusting for household clustering
(aOR=2.2, 95%CI=1.1-4.2, p=0.02), Figure 3a.
Secondary outcome intention-to-treat analysis of patient TB treatment success (cure or completed
treatment) was 53% in the control arm and 64% in the intervention arm. This difference was
statistically significant in crude analysis (OR=1.6, 95%CI=1.0-2.6, p=0.05), Figure 4a. No analysis
adjusted for household clustering was required because there was only one patient per household.
Supplementary analyses:
WHO treatment outcomes: Patients from intervention households were also more likely than those
from control households to be cured (51% [95%CI=43-60%] versus 37% [95%CI=30-45], p=0.02).
Cure and other WHO-defined treatment outcomes are shown in Figure 4b.
Adherence: The intervention was associated with a statistically significant approximate doubling in
preventive therapy initiation (aOR 2.2, primary outcome described above, figure 2a) and this
increase was maintained throughout the 24 weeks of recommended preventive therapy.
Specifically, amongst those who initiated preventive therapy, the average weeks of preventive
therapy taken were similar for intervention and control arms (18, standard deviation, SD=7.7 versus
8
18, SD=7.8 weeks respectively, p=0.9). Consequently, the overall average weeks of preventive
therapy taken was significantly greater in intervention than control arms (7.8, SD=10, versus 4.8,
SD=8.9, p=0.002, Figure 3b). Time-to-event analysis demonstrated that the intervention was
associated with greater overall preventive therapy uptake (log-rank p=0.005, Figure 3b). The study
sample size was selected to test for an effect of the intervention on the whole study population, as
described above, so the study did not have statistical power to test for effects in sub-groups. Thus,
although the intervention was associated with an approximate doubling in preventive therapy
completion (20%, 95%CI=14-25 versus 12%, 95%CI=7-16), this effect in this minority of the study
population was statistically significant only in crude analysis (OR=1.9, 95%CI-1-1-3.2, p=0.03) but
not in adjusted analysis (aOR=1.9, 95%CI=0.78-4.5, p=0.2).
Equity: To assess the equity of the intervention, we compared the study outcomes in the most
versus less vulnerable subpopulations. Vulnerability was assessed as the poorest tercile of the
population, and for preventive therapy initiation also for child contacts in the age groups <5 years
versus contacts aged 5-19 years. Figure 3b and Figure 4b demonstrate that the effect of the
intervention was similar for these subgroups that were most versus less vulnerable and that the
intervention was associated with increases in both primary and secondary outcomes in all these
subgroups. Furthermore, the statistical significance of the intervention effect on preventive
therapy uptake was independent of age group (aOR=2.2 95%CI=1.1-4.2, p=0.02) and was
independent of poverty group (aOR=2.2, 95%CI=1.1-4.1, p=0.02). The intervention was associated
with a trend towards increased treatment success after adjusting for poverty group (aOR 1.7,
p=0.07). These analyses suggest that the intervention was equitable, benefiting more as well as less
vulnerable subgroups, independent of age and poverty.
9
Discussion
During this initial phase of the CRESIPT project, we designed, implemented, refined,12 and evaluated
a novel TB-specific socioeconomic support intervention including cash transfers in a resource-
constrained setting. The intervention proved to be feasible12 and improved rates of TB preventive
therapy initiation in household contacts and TB treatment success in patients with TB.
Previous evidence assessing interventions to improve TB prevention measures and/or TB treatment
adherence has been limited by lack of randomization, small sample sizes, and/or being conducted in
high-resource settings with restricted patient groups (e.g. HIV-infected people,38 homeless people,39
migrants,40 or injecting drug users15,41). Recent systematic reviews concluded there was no existing
evidence on the impact of incentives including cash transfers on TB preventive therapy completion42
and minimal evidence to guide WHO recommendations on implementation and scale-up of TB-
specific socioeconomic support in resource-constrained settings.43 Our study informs global TB policy
and contributes new knowledge to fill this evidence gap5,7
Current shortages in the worldwide supply of tuberculin have made effective management of
household contacts of TB patients difficult to achieve44 and commercial interferon-gamma release
assays are rarely performed due to being too expensive (~200 US dollars/test in the study site),
technically demanding, and not routinely available in resource-constrained settings, including Peru.
Despite these logistical challenges in the study setting, the intervention approximately doubled TB
preventive therapy initiation, highlighting the strength of the impact of the socioeconomic support
received. Although 24 weeks preventive therapy with isoniazid is recommended in Peru, the
protective effect increases with more prolonged administration and longer courses are
recommended elsewhere.16,17 Therefore, our finding that the intervention increased the number of
weeks of TB preventive therapy taken is important because non-adherence to TB preventive therapy
is frequent20,45,46 and improving initiation whilst maintaining adherence could potentially decrease
rates of secondary TB disease. Moreover, supplementary analyses showed that this effect seemed to
be maintained in younger contacts and contacts from poorer households, suggesting that the
intervention was equitable.
While the above findings are encouraging, supplementary analyses showed that socioeconomic
support almost doubled, but did not statistically significantly increase completion of 24-weeks TB
preventive therapy. Potential reasons for this include: this was an exploratory analysis that was not
powered to assess this outcome, with a small number of contacts completing preventive therapy in
each study arm; unlike TB treatment, conditional cash transfers for preventive therapy were not
given monthly; the conditional cash transfer for preventive therapy completion were only made
10
when all eligible household contacts achieved it; and the conditional cash transfers received did not
completely mitigate direct out-of-pocket expenses, suggesting that the financial burden of TB was
still high for many intervention households.47,48 We have since optimised the economic support for
the main CRESIPT project to completely mitigate direct expenses and to offer monthly conditional
cash transfers to individual household contacts to further enhance TB prevention measures and
prevent incident TB.
This research provides evidence supporting the WHO End TB Strategy, which calls for expansion of
the existing biomedical paradigm of TB control by incorporating socioeconomic support
interventions to address poverty and social determinants, the main drivers of the global TB
epidemic.26 While these findings demonstrate that conditional cash transfers can improve patient
outcomes (for example through diminishing food insecurity and improving healthcare access), our
intervention went beyond cash transfers during household visits and participatory community
meetings by offering education and information, and promoting stigma reduction, inclusiveness, and
empowerment. This social support focused on risk factors for non-adherence to TB preventive
therapy and TB treatment such as: lack of TB-related knowledge, being female and/or
marginalized.49 The design of this study did not allow analysis of the differential impact of social
versus economic support on TB treatment and prevention outcomes. Nevertheless, in the field of
HIV, conditional cash transfers interventions have often been complemented by health education for
beneficiaries and, without education or social support, conditional cash transfers may have only
limited impact on health outcomes.6 Thus, optimising social support whilst addressing specific
aspects of poverty such as food insecurity may strengthen the impact of future socioeconomic
interventions.
Limitations of this study include, firstly, the intention-to-treat analysis meant that treatment
outcome could not be evaluated for patients still taking treatment at the final 28-week study
follow-up, including those with MDR-TB. Consequently, the proportion of patients with treatment
success was likely to have been underestimated in both intervention and, perhaps to a greater
extent, control households. However, the majority of patients recruited were HIV negative with
drug-susceptible TB and would be expected to complete treatment by 28 weeks unless there had
been treatment interruption. This supported the intention-to-treat analysis approach in which
patients still on treatment at 28 weeks were classified as not having treatment success. Secondly,
some households could have potentially declared a greater number of household contacts in the
hope of receiving more incentives, although the similar number of contacts per household
comparing the intervention with control households suggests that this did not occur. This issue was
avoided by incentives being provided at household rather than individual level and by inclusion in
11
the intervention of only those household contacts declared prior to randomization, whose validity
was confirmed with a household visit. Thirdly, the patients and study team were un-blinded to the
intervention and there was a conditional cash transfer for households in which the patient achieved
confirmed cure and the relevant contacts completed TB preventive therapy. This could have led to
bias because the patients receiving the socioeconomic intervention could have been more likely to
attend health-posts and request clinical evaluation for confirmation of cure. Furthermore, duration
of preventive therapy taken was calculated from the number of weeks of isoniazid tablets collected
from the health post, which may not necessarily have equated to adherence. Nevertheless, the
study team did not influence NTP staff to initiate such consultations, patient feedback suggested
that seeking confirmation of cure was an empowering element of the intervention,12 and in
supplementary analyses socioeconomic support significantly improved not only treatment success
but also confirmed cure. Finally, the socioeconomic support was integrated, so could only be
evaluated together, precluding assessment of the differential impact of the social versus economic
components. The authors’ original proposal of a 2x2 factorial study (evaluating the impact of
socioeconomic support versus social support only versus economic support only versus standard of
care) was too expensive due to sample size required to achieve sufficient power and instead the
integrated intervention was evaluated. Future larger studies should assess the differential impact of
social and economic support on TB prevention and treatment success and evaluate whether these
results may be replicated in and are generalizable to other patient populations, including those with
high rates of HIV-TB co-infection, rural communities, and those in low income countries.
12
Conclusions
During this initial household-randomised phase of the CRESIPT project, a novel socioeconomic
support intervention in an impoverished setting was feasible, acceptable, enhanced TB preventive
therapy initiation and improved TB treatment success. These findings support the new global End
TB Strategy and highlight the need for larger-scale evaluations, including CRESIPT, to determine the
impact of socioeconomic support on TB care, prevention, control, and potentially TB elimination.
13
References
1. Kim, J. et al. Assessing the incremental effects of combining economic and health
interventions: The IMAGE study in South Africa. Bull. World Health Organ. 87, 824–832
(2009).
2. Richter, L. M. et al. Economic support to patients in HIV and TB grants in rounds 7 and 10
from the global fund to fight AIDS, tuberculosis and malaria. PLoS One 9, e86225 (2014).
3. Tanimura, T., Jaramillo, E., Weil, D., Raviglione, M. & Lönnroth, K. Financial burden for
tuberculosis patients in low- and middle-income countries: a systematic review. Eur. Respir. J.
43, 1763–75 (2014).
4. Martins, N., Morris, P. & Kelly, P. M. Food incentives to improve completion of tuberculosis
treatment: randomised controlled trial in Dili, Timor-Leste. Bmj 339, b4248–b4248 (2009).
5. Lutge, E., Lewin, S., Volmink, J., Friedman, I. & Lombard, C. Economic support to improve
tuberculosis treatment outcomes in South Africa: a pragmatic cluster-randomized controlled
trial. Trials 14, 154 (2013).
6. Heise, L., Lutz, B., Ranganathan, M. & Watts, C. Cash transfers for HIV prevention :
considering their potential. J. Int. AIDS Soc. 16, 1–5 (2013).
7. Rocha, C. et al. The Innovative Socio-economic Interventions Against Tuberculosis (ISIAT)
project: an operational assessment. Int. J. Tuberc. Lung Dis. 15 Suppl 2, S50-7 (2011).
8. Boccia, D. et al. Europe PMC Funders Group Cash transfer and microfinance interventions for
tuberculosis control : review of the impact evidence and policy implications. 15, 1–21 (2011).
9. Wingfield, T. et al. The seasonality of tuberculosis, sunlight, vitamin d, and household
crowding. J. Infect. Dis. 210, 774–83 (2014).
10. Wingfield, T. et al. Defining Catastrophic Costs and Comparing Their Importance for Adverse
Tuberculosis Outcome with Multi-Drug Resistance: A Prospective Cohort Study, Peru. PLoS
Med. 11, (2014).
11. Saunders, M. J. & Evans, C. A. Fighting poverty to prevent tuberculosis. Lancet Infect. Dis.
3099, 10–11 (2015).
14
12. Wingfield, T. et al. Designing and implementing a socioeconomic intervention to enhance TB
control: operational evidence from the CRESIPT project in Peru. BMC Public Health 15, 810
(2015).
13. Wingfield, T. et al. The economic effects of supporting tuberculosis-affected households in
Peru. Eur. Respir. J. 48, 1396–1410 (2016).
14. The World Health Organisation Global Tuberculosis Report 2015. doi:10.1037/e530172011-
002
15. Smieja, M., Marchetti, C., Cook, D. & Fm, S. Isoniazid for preventing tuberculosis in non-HIV
infected persons ( Review ). (2010).
16. Sterling, T. R. et al. Three months of rifapentine and isoniazid for latent tuberculosis infection.
N. Engl. J. Med. 365, 2155–66 (2011).
17. Martinson, N. A. et al. New regimens to prevent tuberculosis in adults with HIV infection. N.
Engl. J. Med. 365, 11–20 (2011).
18. World Health Organisation. World Health Organisation Global Tuberculosis Report 2014.
(2014).
19. World Health Organisation. Adherece to long-term therapies: Evidence for action. (2003).
20. Sk, S., Sharma, A., Kadhiravan, T. & Tharyan, P. Rifamycins ( rifampicin , rifabutin and
rifapentine ) compared to isoniazid for preventing tuberculosis in HIV-negative people at risk
of active TB ( Review ). (2013).
21. Saunders, M. J. & Datta, S. Contact Investigation: A Priority for Tuberculosis Control
Programs. Am. J. Respir. Crit. Care Med. 9, 1049–51 (2016).
22. Haynes, R., Ackloo, E., Sahota, N., Mcdonald, H. & Yao, X. Interventions for enhancing
medication adherence ( Review ). Cochrane Database Syst. Rev. 2008, (2008).
doi:10.1002/14651858.CD000011.pub3
23. Kliiman, K. & Altraja, a. Predictors and mortality associated with treatment default in
pulmonary tuberculosis. Int J Tuberc Lung Dis 14, 454–463 (2010).
24. Burman, W. J. Noncompliance With Directly Observed Therapy for Tuberculosis:
15
Epidemiology and Effect on the Outcome of Treatment. CHEST J. 111, 1168 (1997).
25. Pritchard, A., Hayward, A., Monk, P. & Neal, K. Risk factors for drug resistant tuberculosis in
Leicestershire - poor adherence to treatment remains an important cause of resistance. 481–
483 (2003).
26. Uplekar, M. et al. WHO’s new End TB Strategy. Lancet 1799–1801 (2015). doi:10.1016/S0140-
6736(15)60570-0
27. Dye, C., Lönnroth, K., Jaramillo, E., Williams, B. & Raviglione, M. Trends in tuberculosis
incidence and their determinants in 134 countries. Bull. World Health Organ. 87, 683–691
(2009).
28. United Nations Research Institute for Social Development (UNRISD). Combating poverty and
inequality: structural change, social policy and politics. (2010).
29. Chatham House. Social protection interventions for tuberculosis control: the impact, the
evidence, and the wat forward. Meeting summary. (2012).
30. Lagarde, M., Haines, A. & Palmer, N. The impact of conditional cash transfers on health
outcomes and use of health services in low and middle income countries. Cochrane Database
Syst. Rev. (2009).
31. Evans, David B; Elovainio R; Humphreys, G. World Health Organisation report. Health systems
financing: the path to universal coverage. (2010).
32. UNAIDS Expanded Business Case. Enhancing Social Protection. Geneva: Joint United Nations
Programme on HIV/AIDS. (2010).
33. Social protection floor of a fair and inclusive globalization. Geneva: International Labour
Organisation. (2011).
34. Doetinchem, O., Xu, K. & Carrin, G. Conditional cash transfers: What’s in it for health?
Technical Briefing Papers, 1. (2008).
35. Pettifor, A., Macphail, C., Nguyen, N. & Rosenberg, M. Can money prevent the spread of HIV?
A review of cash payments for HIV prevention. AIDS Behav. 16, 1729–1738 (2013).
36. Torrens, A. W. et al. Effectiveness of a conditional cash transfer programme on TB cure rate: a
16
retrospective cohort study in Brazil. Trans. R. Soc. Trop. Med. Hyg. 110, 199–206 (2016).
37. Ministerio de trabajo y promoción del empleo de Perú: Diagnóstico socioeconómico laboral de
la región Callao. (2012).
38. Rai, C. et al. Adherence to tuberculosis preventive therapy among. (1996).
39. Tulsky, J. P. et al. Adherence to Isoniazid Prophylaxis in the Homeless. Arch. Intern. Med. 160,
(2000).
40. Ailinger, R. L., Martyn, D., Lasus, H. & Lima Garcia, N. The effect of a cultural intervention on
adherence to latent tuberculosis infection therapy in Latino immigrants. Public Health Nurs.
27, 115–20 (2010).
41. Malotte, C. K., Hollingshead, J. R. & Larro, M. Incentives vs outreach workers for latent
tuberculosis treatment in drug users. Am. J. Prev. Med. 20, 103–107 (2001).
42. Ee, L., Cs, W., Se, K. & Volmink, J. Material incentives and enablers in the management of
tuberculosis ( Review ). (2012).
43. Cobelens, F., van Kampen, S., Ochodo, E., Atun, R. & Lienhardt, C. Research on
implementation of interventions in tuberculosis control in low- and middle-income countries:
a systematic review. PLoS Med. 9, e1001358 (2012).
44. Tebruegge, M., Bogyi, M., Soriano-Arandes, A. & Kampmann, B. Shortage of purified protein
derivative for tuberculosis testing. Lancet 384, 2026 (2014).
45. Horsburgh, C. R. et al. Latent TB infection treatment acceptance and completion in the United
States and Canada. Chest 137, 401–9 (2010).
46. World Health Organisation. Guidelines on the management of latent tuberculosis infection.
(2015).
47. Williamson, J., Ramirez, R. & Wingfield, T. Health, Healthcare Access, and Use of Traditional
Versus Modern Medicine in Remote Peruvian Amazon Communities: A Descriptive Study of
Knowledge, Attitudes, and Practices. Am. J. Trop. Med. Hyg. (2015). doi:10.4269/ajtmh.14-
0536
48. Wingfield, T. et al. The CRESIPT project: community feedback and practical challenges of
17
conditional cash transfers for TB-affected households in Peru. Invited oral abstract
presentation. Int. J. Tuberc. Lung Dis. Supplement, (2014).
49. Garie, K. T., Yassin, M. a & Cuevas, L. E. Lack of adherence to isoniazid chemoprophylaxis in
children in contact with adults with tuberculosis in Southern Ethiopia. PLoS One 6, e26452
(2011).
50. Mauch, V. et al. Assessing access barriers to tuberculosis care with the tool to Estimate
Patients’ Costs: pilot results from two districts in Kenya. BMC Public Health 11, 43 (2011).
51. Floyd, K. Costs and effectiveness—the impact of economic studies on TB control. Tuberculosis
83, 187–200 (2003).
52. Floyd, K. Financial resources required for tuberculosis control to achieve global targets set for
2015. Bull. World Health Organ. 86, 568–576 (2008).
18
Box 1: Description of the socioeconomic support intervention12
Related research: The project was informed by the findings of our group’s Innovative Socioeconomic Interventions Against TB (ISIAT) study,7 two systematic reviews of cash transfer interventions,8,12 expert consultations,29 and feedback from civil society and leaders of the Peruvian TB Program.48
Outputs: Socioeconomic support targeted outcomes on the TB causal pathway and promoted equitable access to TB program activities including: i. screening for TB in contacts; ii. initiation of TB preventive therapy and completion of TB treatment; and iii. engagement with social support activities.
Socioeconomic support intervention has been described.12 Briefly, it constituted an integrated package of:
- Social support consisting of household visits and participatory community meetings that aimed to provide information, mutual support, empowerment, and stigma reduction. The household visits were made shortly after patients commenced treatment and provided education on TB transmission, treatment, preventive therapy, and household finances. The community meetings took place monthly in each study site community, were each attended by ~15 patients and their household contacts, and cost ~189 United States (US) dollars per meeting (~13 US dollars per patient per meeting).12 The meetings re-emphasized the educational themes of the household visit and developed “TB Clubs” during which participants shared TB-related experiences within a mutually supportive group (to be reported elsewhere). All household members were invited and encouraged to participate in household visits and community meetings.
- Economic support consisting of conditional cash transfers throughout treatment to defray average household TB-related costs thereby reducing TB risk factors whilst also incentivizing and enabling care. Economic support was designed with the intent that direct out-of-pocket expenses would be completely defrayed in patients who achieved all conditional cash transfers. Such direct out-of-pocket expenses had previously been found to equate to 10% of annual household income in the study setting, 10 equivalent to approximately 230 US dollars. We hypothesized that defraying these direct expenses would decrease a TB-affected household’s financial burden, likelihood of incurring catastrophic costs, and - when combined with integrated social support - enhance access to TB care and improve TB outcomes. During planning of the intervention, it had been estimated that if the intervention were implemented nationally the additional cost per patient would increase Peru’s TB Program budget by approximately 15% per patient.10 Focus group discussions with key stakeholders suggested that such increased expenditure was locally appropriate and affordable.10,12,50 Moreover, review of the relevant literature suggested that interventions that increased per patient TB Program budget by ≤50% and reduced incident TB by one third would likely be cost-effective and sustainable.51,52
19
Table 1: Baseline demographic characteristics of the study population.
Footnote: Sputum smear results were defined as positive if acid-alcohol fast bacilli were visualised in NTP reference laboratory and/or our research team’s laboratory from sputum samples prior to starting TB treatment.
20
Figure 1: Map of study site including 32 health post regions and numbers of study participants from each region.
Footnote: the location of the principal Peruvian international airport is shown adjacent to community 21.
21
Figure 2: Study schematic showing participant recruitment, randomisation and participation.
22
Figure 3a: Primary outcome analysis of TB preventive therapy initiation in contacts <20 years old from intervention and control households.
Footnote: Error bars are 95% confidence intervals. Crude analysis was the odds of preventive therapy initiation comparing intervention versus control households without adjustment for household clustering. Adjusted analysis was the odds of preventive therapy initiation comparing intervention versus control households adjusting for household clustering.
23
Figure 3b: Supplementary analysis of number of weeks of TB preventive therapy taken by contacts from intervention and control households.
Footnote: the p value shown below average weeks of preventive therapy taken in the final column of the data table is the difference between the mean average number of weeks of preventive therapy taken by contacts aged <20 years in intervention versus control households compared by Mann Whitney U test.
24
Figure 3c: Supplementary analysis of initiation of TB preventive therapy by household contacts across age groups and poverty levels
Footnote: Error bars are 95% confidence intervals. “Poorer households” were those in the lowest poverty score tercile and “less poor” households were the remaining households.
25
Figure 4a: Secondary outcome analysis of TB treatment success in patients with TB from intervention versus control households.
Footnote: Error bars are 95% confidence intervals. Likelihood of treatment success was analysed by univariable logistic regression of association of being a patient from an intervention household with TB treatment success.
26
Figure 4b: Supplementary analysis of TB treatment outcomes in patients with TB from intervention versus control households across poverty levels
Footnote: Percentages in brackets Error bars are 95% confidence intervals. Treatment outcomes were those recorded by the Peruvian NTP in line with WHO guidance and reporting: cured, treatment completed, treatment failed, died, lost to follow-up (locally termed abandoned), and not evaluated. 14
27