Qualitative data.

The qualitative data were derived from two sources. The first source was from an in-depth interview set collected previously between May-September 2012 as previously described [3,4]. This set of interview data was used to identify relevant themes of DO and to inform the design of the questionnaires for the cross-sectional quantitative study and the content of the interview guides for the key informant interviews.

During the second phase, complementary qualitative study, we interviewed key informants with specialised knowledge and expertise on the realities of DO implementation and applicability for DR-TB treatment in the Mumbai context. We chose informants who would incorporate views from both the private and public health sectors. We interviewed a prominent private chest physician, a District TB Officer representing RNTCP and a high-ranked officer from the Care, Support and Treatment department of MSACS (Maharashtra State AIDS Control Society). Interviews were recorded using a digital recorder and transcribed. Notes were taken in one case where the key informant did not wish to be recorded and were expanded immediately afterwards.

Quantitative data.

We used an identical definition for DO (“a patient being directly observed by a provider swallowing his/her medications on a daily basis throughout the entire treatment course”) for patients, providers and MSF staff. Patients were defined as “following Strict DO” and “following DO” if a DOT-provider had seen the patient swallow his/her medications “every day”‘ or “most of the days” (at least 4 out of the last 7 days of the previous week) respectively. If DO was not followed, reasons were recorded through the provision of a check-list of responses and an option for ‘other’ if there was a novel reason expressed.

The study questionnaires were piloted and after finalization they were administered to three distinct groups: patients, DOT-providers and MSF health staff. All patient and MSF health staff questionnaires were administered at the MSF clinic in Mumbai. DOT-provider questionnaires were completed at each individual location of the provider. The questionnaires employed were paired (patient/DOT-provider), quantitative and descriptive. Patient and DOT-provider questionnaires were facilitated by one and two interviewers respectively, all fluent in Marathi, Hindi and English to maintain consistency: the interviewers were not directly involved in the care of the patients or previously linked to the DOT-providers. The MSF health staff questionnaires were facilitated in a group meeting and their responses were anonymous. The questionnaires were identical with the ones used for the pairs of patients/DOT-providers but the MSF staff was asked to report on the overall cohort and not on the individual patients and providers.

Both the questionnaires and the interviewer emphasized the oral medication so as to reduce automatic answers related only to seeing a health professional for receiving the daily injection in the intensive phase. To reduce the risk of social desirability bias the specific terminology of ‘DOT’ or ‘DO’ was not used. Patients were asked to mark their overall adherence on a simple scale for the IP and CP phases of the treatment. They also nominated a category for the way they took their medicine in the previous week overall, and then they were asked to recall specifically day by day in the previous week how they took their medicines. DOT-providers were asked exactly the same questions about the same time period but with the day-by-day recall omitted as we found during the pilot study that this was not practical. A cover sheet was also completed by the interviewer with information taken from the patient’s file related to demographics, treatment, HIV status and DOT-provider details. Finally we used the individual patient medical records kept at the MSF clinic to obtain demographic and clinical data as well as interim and final treatment outcomes.

Data analysis.

A thematic analysis was performed to analyse the qualitative data in the study. The data transcripts were read and reviewed manually by one of the investigators (JF). A second investigator (PI) reviewed the transcripts to enable cross-validation of the themes identified.

Differences in interpretation were discussed and resolved by the two investigators.

Quantitative data were analysed using descriptive statistics. The chi-square test was used to examine differences in characteristics between patients who followed DO and those who did not. Poisson regression models were used to explore associations between demographic, clinical and programmatic characteristics of the patients and adherence to DO.

Negative reasons.

1) Lack of accessibility: Lack of accessibility was a main reason given by participants for not following DO. A commonly reported problem was the absence of providers when patients went to receive DO.

Patient: “Yes … it has happened that the doctor (DOT-provider) did not come so I missed the injection as well as the pills.”

Interviewer: “So that was because the doctor did not come. Has it happened due to a fault of yours?”

Patient: “No I have never missed a dose on my account. I told the counsellor. He said it should not happen again. If the doctor (DOT-provider) has to be away the next day then the previous day he gives me the injection and pills. He had not done it then so I missed the dose.”—Patient 6

Several patients also reported missing doses of medications because they had to attend appointments with other physicians, including doctors for HIV and diabetes treatment. DO for TB and DR-TB was totally vertical and in isolation from any other health-service, even in clinics where HIV and TB services were totally integrated, such as the MSF clinic.

“When I go there (MSF clinic for my HIV appointment), it is hard to take first the injection and pills and then go there. I feel dizzy and nauseous; I vomit sometimes and I have to travel far to get here. So on that day I miss the dose.”—Patient 13

Patients often lamented the possibility of resuming a normal life and work, especially given the requirement for DO; the administration of drugs became a daily burden and the daily life and work routine had to be built around the time of DO.

“I left home in this morning, travelled to come here, will take medicines at [the DOT facility] and then go to work directly. Then I will return home at midnight. See how difficult it will be for a man to take the medicines and work.”—Patient 9

The absence of a provider was a common enough occurrence that many patients made “contingency plans” for taking treatment when their DOT-providers were not around. Such contingency plans are common among HIV patients on antiretroviral treatment but they are not part of any national TB or DR-TB treatment programme strategies.

“An extra packet is kept here at home which the doctor has given for the times when he is not there so that he can at least take the pills. The injection is not given but a pill packet is given.”—Family care provider to Patient 1

Of note, patients reported travelling all the way to the clinic to receive DO but instead got self-administered therapy while sitting in the clinic. Reporting to a clinic or other facility was seen by DOT-providers as a substitute for DO. One should question the “honesty” of such a strategy that requires patients to go to a facility for a “daily drug prescription”.

“No the doctor is around there. He has to see other patients so he gives the medicine and I have to sit outside and eat it. He also has a business to run and cannot sit around so long for me. He says eat the pills and that’s all.”—Patient 2

2) Promotion of stigma through DO: Patients, providers, and family members all reported that DO was difficult because they felt it reinforced or perpetuated stigma. This was particularly felt among HIV/DR-TB co-infected patients who had experienced stigma earlier as HIV patients at the HIV clinics, and then again while receiving treatment for DR-TB. The interviews often revealed embarrassment and loss of self-respect among patients taking DO at a doctor’s clinic:

“The way to take medicines is that I go to the doctor, if there are 4–5 customers over there, and I would take medicines in front of them then they will question me what is this medicine for? The doctor will tell them that this man has HIV, then the way other persons look at you changes, people try to keep away from you and sometimes I would vomit, or bring out sputum then they would pinch their noses, so I thought it is better that I took medicines at some distant place”—Patient 8

The perception that DO enforces stigma was also reported by family members or other caregivers who were dismayed by the reactions of neighbours, relatives, and other patients at the waiting areas of private clinics which served as DO facilities.

“Yes. He does not eat them there because many of our neighbours are around in the clinic and if he vomits or is dizzy after taking the pill it will be embarrassing for him.”—Family care provider to Patient 1

3) Concerns about managing adverse events: Several patients and providers reported that DO was not done because it made things difficult for the patient if there were adverse events. Adverse events, especially nausea, vomiting and dizziness which were extremely common, especially during the first period of the treatment, made the DR-TB patients “visible” to other patients at the DO facilities and induced stigma as earlier described. But adverse events, especially the ones that occur directly after the swallowing of the pills were moreover disabling and incapacitating. In several circumstances the DOT-providers and patients “agreed” to skip DO as it was not easy for the patient.

“My house is on a hill and the dispensary at the bottom and it takes me 10 minutes to get there. It tires me to go there so I get the injection at the doctor’s and then take the pills home and to take them later otherwise I get dizzy and it gets late to return home. I take the medicines after I reach home. The doctor has helped me a lot by doing this. Otherwise they say the pills have to be taken at the doctors place itself. They say that. There are patients in the clinic, every day if you come there and take the pills and vomit, it does not look nice so he gave me the pills to take at home.”—Patient 6

This was also reported by providers as a reason for not giving DO.

“But he takes the medicines home—he had told me that he would like to take these at home because he might vomit etc. He informs me immediately if he vomits after taking medicines. But I see the result of the medicines.”—DOT-provider 2

Positive Reasons.

1) Flexibility and understanding among DOT-providers and programme-managers: The key informants who were interviewed for the study were aware of “alternatives to DO” practices in the field and in fact were very open and understanding. They often used the terms “flexibility” and “support” to describe what they thought was needed to ensure adherence to treatment. The need for counselling was a theme that was recurrently emerging during the interviews.

“Even I observed, counselling is the only way. It is the only tool to decrease everything. It is the tool to make the people adhere to treatment, the tool to reduce defaulters, the tool to do infection control. Otherwise it is not possible only.”—Key Informant 2

“It (DO) is not necessary if the patient is counselled properly. Maybe it is relevant in the beginning, until they understand their medication and are used to the routine, and then no need.”—Key Informant 1

“…there should be some kind of flexibility, because people are not one size fits all. There are different levels of commitment in adherence in different people. That’s why counselling as such, when we have an experience in HIV/TB also, here DOTS is just an extended therapy. But in HIV you have to have a life-long therapy. When we look at these things, one, it is not possible to give a life-long therapy as directly observed, secondly even if you do try to give directly observed in such extended situations in life, it is very difficult to maintain.”—Key Informant 3

The extended period required for DR-TB treatment makes daily DO unrealistic. Life events were often described as examples of such unrealistic demand being put on patients, providers and systems. In fact the strictness of DO and the lack of supportive counselling were considered important reasons for the high lost-to-follow-up (“default”) rates observed among DR-TB patients. Thus a need for a flexible approach was often expressed as being essential for adherence to treatment.

“Another reason for defaulter in Mumbai—there is a sudden death in a family member, they don’t have time to come {…} to tell they are going to leave to their native place. This type of people default {…} we can tell to send their relatives to take 7 days medication which can be carried to the native place. So that we can at least help that person to decrease the number of days of defaulting.{…} The system says don’t give medication. System is telling. But what about defaulting? If you give medication with them, patient won’t default. So, there has to be flexibility in the programme, otherwise we will not be able to control this default rate because of uncontrolled circumstances which happens in life. Sudden death, marriage ceremony, festivals.”—Key Informant 2

2) Patient “trusted” to perform self-administered therapy: A patient-centred approach was spontaneously adopted by some DOT-providers and this practice was supported by patients, family members and programme managers. The “one-size-fits-all” strategy was questioned and as Key Informant 3 said:

“If there is specialized attention given for these people I think it would definitely give a better outcome.”—Key Informant 3

Patients were sometimes trusted to be given the medicines and DO was not considered relevant any more. Adherence to treatment became more of an agreement, a contract between the patient and the caregiver.

“It will depend on his rapport with the particular DOT provider. If the DOT provider feels that this patient is very sincere, honest, and he will take the medication, then he may give him for 7 days, or whatever it is, or 2 days or 3 days. {…}That can be arranged, it is a local arrangement which is done.”—Key Informant 2

Private practitioners in Mumbai do not require DO; instead the focus is on counselling and patient education. This might be one of the reasons for the “appeal” of the private sector to patients.

“When starting treatment for a patient with DR-TB, we give them the full background of the importance of taking treatment, adherence and side effects. We spend extra time with them at the start of treatment to ensure they understand this.”—Key Informant 1

3) Family member or “treatment-buddy” provided support: One additional reason why DO was not done by providers was that family members were entrusted with the job of ensuring that a patient took his or her medications. This was reported by several family members. An enabling family environment or the support by a ‘treatment-buddy’ were considered as more sustainable and convenient practice for treatment adherence, despite a reluctance by TB programmes to accept “family DO” as alternative to “facility DO” or “community DO”.

“I make her sit in front of me and give her the medicines. I take her with me to see the doctor and get her to take her day time dose and the night time dose and the injection and then I go to work. By the time I come home it is 11pm.”—Family care provider to Patient 5

As one key informant emphasized, the patients who are treated by private practitioners are in fact supported by family members or friends. Another key informant who represented the private sector went as far as to compare treatment outcomes between the two sectors and questioned the efficacy of DO and its association with better outcomes.

“If the dose could be given to the family to observe at home—almost 40% of patients are from private sector, 60% from public sector. 40% patients—how they are supervised? Doctor doesn’t supervise, their family members are supervising. They are taking medication, it is not that they are defaulting. Why this cannot be followed in the public sector?”—Key Informant 2

“With counselling the DOT is not necessary. If you look at statistics from the national program and from the private sector for DR-TB the outcomes are about the same. 25% defaulters, 17% death, 60% completed/cured. The default rate is comparable, but the reasons are different. In the private sector it is because they are shopping for other services due to lack of finance or migrating. In the public sector it is due to dissatisfaction with the service, not being treated kindly, inconvenience.”—Key Informant 1