記住我
ORIGINAL ARTICLE
Year : 2023 | Volume
: 14
| Issue : 2 | Page : 82-87
Loss-to-follow-up rate among tuberculosis patients registered at two tuberculosis units of ballabgarh block of district Faridabad, Haryana: A desk review
Anwita Khaitan1, Sanjay K Rai2, Anand Krishnan2, Sanjeev Kumar Gupta2, Shashi Kant2, Gopi Chand Khilnani3
1 Department of Preventive and Social Medicine, JIPMER, Puducherry; Centre for Community Medicine, All India Institute of Medical Sciences, New Delhi, India
2 Centre for Community Medicine, All India Institute of Medical Sciences, New Delhi, India
3 Chairman, PSRI Institute of Pulmonary, Critical Care and Sleep Medicine; Department of Pulmonary Medicine, Critical Care and Sleep Disorders, All India Institute of Medical Sciences, New Delhi, India
Correspondence Address:
Dr. Anwita Khaitan
Department of Preventive and Social Medicine, JIPMER International School of Public Health, JIPMER, Puducherry - 605 006
India
Source of Support: None, Conflict of Interest: None
DOI: 10.4103/injms.injms_8_23
Introduction: Treatment loss-to-follow-up (LTFU) is a great obstacle to tuberculosis (TB) elimination. Our objective was to estimate its prevalence in district Faridabad, Haryana, and to see its association with selected variables. Materials and Methods: We conducted a desk review of TB notification registers of two TB units under a district TB center, Faridabad. All patients registered between October 1, 2016, and September 30, 2017, were analyzed. Following univariate analysis, factors with P < 0.2 were included in the multivariable logistic regression model. Results: Of 999 eligible patients, 742 (74.2%) had treatment outcomes available. LTFU was estimated to be 3.5%, with cumulative incidence calculated to be 6.63 LTFUs per 100 person-years. LTFU increased with increasing age. LTFU was 5.2% (9.8 LTFUs per 100 person-years) and 0.3% (0.6 LTFUs per 100 person-years) among pulmonary TB (PTB) and extra-PTB (EPTB) patients, respectively. LTFU was higher in patients with previous relapse (9% or 14.2 LTFU per 100 person-years) and previous LTFU (16% or 32 LTFUs per 100 person-years) as compared to only 2.4% (4.7 per 100 person-years) among new patients. PTB patients and patients on treatment after prior LTFU had 12 times and 5 times higher odds of LTFU than EPTB patients and new patients, respectively. Conclusion: The observed LTFU burden is lower than the <5% target set by the National TB Elimination Program. Lack of information on outcomes under the program should be cause for concern. Special attention should be paid under the program on high-risk groups identified to ensure treatment adherence.
Keywords: Haryana, loss-to-follow-up, National Tuberculosis Elimination Program, tuberculosis
India is the largest contributor to the global burden of tuberculosis (TB). It accounts for 27% of the 10 million incident TB cases, 24% of the 558,000 incident multidrug-resistant TB cases, and 27% of the 1.5 million TB-related deaths in the world.[1]
Treatment loss-to-follow-up (LTFU) is the outcome when treatment has been interrupted continuously for 1 month or more.[2] It is a major cause of recurrence of TB and emergence of different forms of drug resistance. The Government of India (GoI) recommends that LTFU be kept at a minimum, below 5% of all patients on treatment. To lower treatment interruption and LTFU, the directly observed treatment short-course (DOTS) strategy was adopted nationally in 1996, which included five components, namely political and administrative commitment, good quality diagnosis primarily by sputum smear microscopy and other microbiological tools, uninterrupted supply of quality drugs, directly observed treatment, systematic monitoring, and accountability. Further, India shifted from intermittent DOTS to daily DOTS in 2016 to further lower the rate of LTFU.
Studies from various parts of India have reported treatment LTFU rates ranging between 15% and 25%.[3],[4],[5],[6],[7],[8] Most of these studies were conducted before daily DOTS was introduced.
There was a need to estimate an updated rate of LTFU. Hence, we conducted this study with the objectives to estimate the prevalence of treatment LTFU in TB unit (TU) Ballabgarh, and TU Mohna, in district Faridabad, Haryana, and to find out the association between treatment LTFU and selected sociodemographic variables.
Materials and MethodsThe TU is the nodal point for TB control activities in the subdistrict.[2] Each TU caters to a population of 150,000–250,000 and is made up peripheral health institutions (PHIs). A PHI is the lowest rung in the organogram at which treatment can be initiated.
The study was conducted through a desk review/secondary data analysis in Ballabgarh block of district Faridabad, Haryana. This block comprised both urban and rural areas. Two out of the nine TUs under the district TB center, Faridabad, were purposively selected – one rural (TU Mohna) and one urban (TU Ballabgarh). TU Ballabgarh is located within the premises of the subdistrict hospital (SDH), Ballabgarh. It comprises three PHIs - SDH Ballabgarh, urban health center (UHC) Hari Vihar, and UHC Partapgarh - and serves a predominantly urban population. TU Mohna is located within the premises of primary health center (PHC) Mohna. It comprises 4 PHIs – PHCs Mohna, Chhainsa, Dayalpur, and Panhera Khurd – and it serves a rural population.
The National Tuberculosis Elimination Programme (NTEP) TB notification registers maintained at the TUs were the reference data sources. They had a standard format as recommended by the NTEP training modules.[3] The registers maintained at the TU had all the data collated from the PHIs. The TU-level register is routinely maintained by the senior treatment supervisor (STS), and the PHI-level register is maintained by the TB-health visitor (TB-HV).
Through personal visits, the investigator (AK) abstracted relevant data from the register. The authors tried to gather the missing data by examining other possible sources of data at the facility, namely NTEP PHI-level notification registers, NTEP laboratory registers, and NTEP treatment cards.
The desk review was conducted in May–June 2018, which gave sufficient time for the outcomes to have occurred and get recorded. The data were cleaned and checked for duplicate records before subjecting them to analysis. The data of all patients registered between October 1, 2016, and September 30, 2017, were entered in the MS Excel database. The data were cleaned and checked for duplicate records before subjecting them to analysis. The study period was selected to ensure that the treatment outcome could be declared.
Analysis was done using MS Excel and STATA version 12 StataCorp's Stata statistical software version 12 (College Station, Texas : StataCorp LP; 2011). Continuous data were summarized using means and standard deviations, or medians and interquartile ranges, as applicable. Categorical data were summarized as proportions. The strength of associations between the treatment outcome and other variables was expressed as unadjusted/crude odds ratios with 95% confidence intervals (CIs) derived from univariate logistic regression. All univariate associations with P < 0.2 were included in a multivariable logistic regression model to obtain adjusted odds ratios with 95% CIs. P < 0.05 was considered statistically significant. We included all factors that had P < 0.2 in the multivariable logistic regression model. The stepwise backward elimination technique was used.
ResultsA total of 999 patients were started on treatment (749 of TU Ballabgarh and 250 of TU Mohna). We selected a reference period that would have allowed the closure of all patients at the time of data collection.
The data in the TU-level registers were incomplete - treatment outcome data were available only for 74% of the patients. Key informant interviews (KIIs) with the senior TB laboratory supervisors (STLS), the STS, and the Designated Microscopy Centres' Laboratory Technicians/Technologists (DMC LTs) shed light on the possible reasons. One reason for incomplete data entry was the vacancy of key posts in both the TUs. The post of STS was vacant at TU Ballabgarh, hence the STLS had to maintain the TU-level register there. Further, since the TB-HV posts were vacant at all but one PHI, the PHI-level registers were maintained by multipurpose health workers, who were trained by the STS/STLS in this task. The KIIs also revealed that since 2012, there had been a tendency to occasionally enter data first into the Nikshay database, and then copy it into the TB register later. This was because, since May 2012, TB notification had become mandatory.
The LTFU rate (among the patients whose data were available in the registers) was estimated to be 3.5% with a CI of 2.3% to 5.1%. The cumulative incidence was calculated to be 6.63 LTFUs per 100 person-years.
The distribution of baseline characteristics was compared between lost-to-follow-up and not lost-to-follow-up [Table 1]. Those not lost-to-follow-up comprised all patients whose treatment outcome was any one of the following: cured, treatment complete, failure, death, transferred out, switched to category intravenous (i.e., treatment for drug-resistant TB), or missing.
Table 1: Baseline characteristics of tuberculosis patients lost-to-follow-up and not lost-to-follow-upWe found that 4.8% of males were LTFU (9.8 LTFUs per 100 person-years), compared to only 1.6% of females (3.1 per 100 person-years). Additionally, a trend of increasing LTFU with increasing age was seen - 2.7% of adolescents versus 3.5% of adults versus 5.8% of the elderly were lost to follow-up (5.1 LTFUs per 100 person-years versus 7.0 LTFUs per 100 person-years versus 13.1 LTFUs per 100 person-years).
Pulmonary TB (PTB) patients had a higher rate of LTFU (5.2%; 9.8 LTFUs per 100 person-years) compared to extra-PTB (EPTB) patients (0.3%; 0.6 LTFUs per 100 person-years). Seven percent of the previously treated patients were LTFU, compared to 2.4% among new patients. Around 9% of previously treated patients who had relapsed were LTFU in the current round of TB treatment, with a cumulative incidence of 14.2 LTFU per 100 person-years. The LTFU rate among TB patients with a history of LTFU and who were restarted on TB treatment was even higher (16%; 32 LTFUs per 100 person-years). This contrasted with new patients, of whom only 2.4% were lost to follow-up (4.7 LTFUs per 100 person-years).
On univariate regression, higher odds of LTFU (P < 0.2) were observed for male patients, patients on retreatment (either relapsed or with prior LTFU), patients with PTB, and patients with high sputum smear positivity (3+) [Table 2]. On multivariable logistic regression, we found that patients with PTB had 12 times higher odds of LTFU compared to patients with EPTB. Re-treatment patients with a history of LTFU had five times higher odds of LTFU compared with new TB patients [Table 3].
Table 2: Univariate logistic regression to see the association of treatment loss-to-follow-up with sociodemographic and selected variables
Table 3: Multivariable logistic regression for association of treatment loss-to-follow-up with sociodemographic and selected variables Discussion This estimated LTFU rate is lower than that reported by the Central TB Division, Ministry of Health and Family Welfare, GoI, which reports treatment LTFU between 2% and 6% in Haryana,[9] where this study was conducted.
Our finding is based on official programmatic data. If we compare it with the findings of Kant et al.[3] who also conducted a desk review of official data from the same study setting, we find that they reported a LTFU rate of 14.3% based on the TB register data of 8 years (2007–2014) for two PHCs which fall under TU Mohna, district Faridabad, Haryana. In their study of 885 patients, the treatment outcome was available for 595 patients, hence only 67% had outcome data available. The reported LTFU rate was higher than that reported by us, and this can be because of many possible reasons – since the data spanned 8 years, in the initial years, the LTFU rate may have been higher, and as the program matured and the health-care services improved, LTFU dropped. However, the initial years of high LTFU led to an overall LTFU rate which was high. Another possibility is that because the data for one-third of the patients were missing, an LTFU rate of 14.3% was obtained – the missing data may have tilted the estimate toward an overestimation. However, this cannot be stated with certainty.
Mundra et al.[4] reported an LTFU rate of 8.6% among 503 patients of a TU in Wardha, Maharashtra, in 2017, with outcomes missing for only 1% of the patient participants.
A multicentric cross-sectional study which reported an LTFU rate closest to the recommended target of <5% was conducted by Vijay et al.[5] – they reported a LTFU rate of 6.4% among 10639 new smear-positive (NSP) patients from twenty districts of six states of India in 2004–2005, with no missing outcomes reported. The large sample may have led to an overall relatively low LTFU rate estimation; the pockets with high occurrence of LTFU might have been masked by the regions with low LTFU rates.
Other studies from India reported relatively higher LTFU rates on average, ranging between 6% and 30%.[6],[7],[8] These LTFU rates are slightly higher than the NTEP recommendation of <5% but much lower than those reported in the abovementioned studies. It is noteworthy that most of these studies did not have/report missing data – only Mundra et al. reported outcomes missing for 1% of their participants.
Association between loss-to-follow-up and selected variables
We had found that males, the elderly, and those on retreatment after relapse or LTFU had a higher proportion of LTFU. Previous studies have also reported a higher rate of LTFU among the elderly,[4],[8],[10],[11],[12] males,[5],[6],[12],[13],[14] and retreatment patients.[8],[11],[15],[16],[17],[18],[19],[20]
In our study, PTB patients had higher odds of LTFU compared to EPTB patients. It is possible that EPTB was perceived to be a more serious disease condition compared to PTB. Hence, a lower rate of LTFU was observed among EPTB patients. They are usually more symptomatic and thus the addition of drug side effects is likely to be more troublesome. We do not have any direct evidence for our speculation. Therefore, in the future, a separate study assessing the perception of severity of disease and treatment compliance would be helpful to generate evidence on this issue.
Limitations
The usual limitations of studies based on review of records were present in our study as well. Significant proportions (26%) of TB patients were excluded from the analysis because their treatment outcome was not available. We do not know if those not included in the study were systematically different from those that were indeed included. If those excluded were systematically different, then the estimated LTFU rate could also change substantially.
Since treatment outcome data were available only for 74% of the patients, the actual LTFU rate could lie anywhere between 2.6% and 26.3%. This level of missing data on treatment outcome impedes our efforts for accurate and precise estimation of the LTFU rate.
As mentioned, a limitation is the high proportion (26%) of data on treatment outcome missing from the notification registers. Since our data are a subset of official NTEP data, one may need to conduct further studies to assess whether the lower LTFU rates officially reported for the North Indian states in the India TB Report 2022[4] are also influenced by missing data rather than being a true reflection of the level of LTFU in this part of the country.
The total number of lost-to-follow-up patients was small (n = 26). Therefore, the 95% CI around the estimated odds ratio was wide. In fact, some of the factors where no significant statistical association was observed could be due to small sample size.
Treatment LTFU remains a major obstacle to achieving the National Strategic Plan for Tuberculosis Elimination (2017-2025) (NSP) targets. It is still significantly high in various pockets in India. It is crucial that newly recruited DOTS providers and other health-care providers are aware of the factors associated with treatment interruption or LTFU, such as male sex and prior history of leaving treatment prematurely. This information may be made part of their orientation training sessions. Even though patient profiling should always be done with the caveat that the profile may not always hold true, it is helpful in a high-burden situation such as that of India, so that one can pay special attention during counseling to those who seem more vulnerable to LTFU. Repeated counseling of at-risk patients as well as their caregivers, with content that is scientifically sound as well as customized to the unique situation of each patient, is the need of the hour since many patients leave treatment as soon as they feel better, or because they feel worse because of side effects of drugs. The DOTS strategy should be rigorously followed for such patients to ensure adherence and completion of treatment since LTFU increases the risk of adverse outcomes such as death, drug resistance, and continued spread of infection. The onus of motivation and ensuring TB treatment compliance lies on the shoulders of all health-care providers so as to prevent adverse outcomes for the TB patient, for his family, and for the community. Higher incentives for DOTS providers and regular disbursement of funds are important to motivate the DOTS providers.
ConclusionsThe observed LTFU burden is lower than the <5% target set by the NTEP. Lack of information on outcomes under the program should be cause for concern. Special attention should be paid under the program on high-risk groups identified to ensure treatment adherence.
Financial support and sponsorship
None.
Conflicts of interest
There are no conflicts of interest.
References
留言 (0)