The present Analytical cross-sectional study was carried out in Mayurbhanj district, Odisha, India. The study period was from February to July, 2007. The study population included all New Smear Positive (NSP) pulmonary TB tribal patients, registered under DOTS in 4^th quarter from 2006 of four TB units. Patients not registered under DOTS, extra–Pulmonary TB cases and death cases from four TB units were excluded.

The operational definitions used for study were:

Patient Delay: Time interval between onset of cough or other symptoms until the first visit to a health facility. Patients Delay > 21 days, on the basis, that a person with cough for 3 weeks should be a suspect of TB [5–7].

Health System Delay: Time interval from patients 1^st visit to health system to the date of start of TB treatment. Health system delay > 7 days assuming that treatment should be started within 7 days of diagnosis of TB [5,6,8].

Total Delay: Sum total of patient delays and system delays.

All NSP pulmonary TB tribal patients from four TB units registered under DOTS in 4^th quarter of 2006 were selected as study sample. Pre-tested semi-structured questionnaire was used to conduct the interview [9,10]. Before the interview a written consent was taken from each TB patient. Ethical clearance was taken from Ethics Committee, NCDC, Delhi.

Data after collection was entered and analysis was done using Epi-info 3.3.2 version. Statistical comparisons between groups were done by using Unadjusted Odds Ratio (UOR), Chi-square test and 95% confidence intervals with statistical significance at p<0.05 to determine significant difference between groups.

A total of 261 out of 267 TB patients could be interviewed from four TB units. Six patients could not be interviewed as they succumbed to the illness. The median patient delay was 24.0 days (range1-709 days) with a mean of 36.5 days. The median health system delay was 3 days (range 1-75 days) with a mean of 11.59 days. The median total delay to confirm diagnosis was 24 days (range 1-710 days) with a mean of 37.52 days. Two hundred and fifty three (97%) TB patients had patient delay and 51 (19.5%) TB patients suffered health system delay.

The higher patient delay was observed among male TB patients 184(96.8%) compared to females.

Majority of TB patients were in the age group of 20-60 years (85.9%). Female TB patients 42 (59.2%) were highest in the age group of 20-40 years. Male TB patients 90 (47.4%) were in the age group of 40-60 years. Maximum number of tribal TB patients were illiterate 145 (55.6%), married 179 (68.6%), farmer by occupation 100 (38.3%), money spent on transport/treatment greater than four hundred rupees 224(86%), poor knowledge about TB 228 (87.4%) distance travelled to reach the DMC/ TB unit was greater than 5 kilometres 223 (85.5%), traditional healers was first choice of health care providers consulted 110 (42%) and self- medication 139 (53.3%).

The main reasons for patient delay among TB patients were lack of awareness about TB (30.5%), long distance of TU/DMC from home (29.9%) which were significantly associated [Table/Fig-1]. Other reasons were poor- socio economic condition (24.1%), loss of wages (10.7%), drinking habits (3.4%) and dependent being too old/young/severely ill (1.4%).

The reasons for health system delay were mainly (48.60%) due to administrative verification regarding presence of DOTS provider in locality for patient medicine distribution timely, (43.17%) delay due to visit to traditional healers/private practitioners after TB diagnosis which were significantly associated [Table/Fig-2]. Other reasons were health personnel absence (6.07%) and medicines out of stock (2.17%).

The present study highlights the problem of patient and health system delays and the reasons associated with delays among tribal TB patients. The median and mean patient delays were 24 days and 36.5 days respectively and system delay was 3 days and 11.59 days. Total median and mean delays were 24 and 37.52 days. Similar findings were reported in Sikkim and South India [4,10] and there are studies who have also reported higher delays in high burden countries [11–22]. Though median and mean delays in present study were less compared to other high burden countries, it must not be overlooked, as it is a tribal pocket.

In this study (73%) males patients showed higher patient delay as compared to females (27.2%), which may be because of men using various forms of substance abuse and frequent contact with infected persons where as females in rural areas especially in tribal, women with TB may not be seeking care within TB programme or not receiving treatment or they must be visiting traditional healers /private practitioners. Studies in India and Vietnam indicate that women are ostracized in the household and neighborhood [2,9,17].

Age distribution pattern in the present study was similar to other studies conducted in India [2,9,23,24,31] which have reported peak levels of delaying was in the age group of 20-40 years and 40-60 years. Most of tribal TB patients were illiterate (55.6%), farmers (38.3%) and daily wagers (30.7%) by occupation. Studies conducted in India have reported similar findings [23–25]. It is obvious that since most of tribal TB patients are in their productive years of life which is definite to affect the economic condition of family in case of males who do not intend to lose daily wages where as in females it severely affects their social and personal life and as well as hampers their reproductive cycle.

Our study reveals that majority of tribal TB patients (87.4%) have poor knowledge about TB, (98.7%), lived at a distance of more than 5km from the health facility and (42.1%) first consulted traditional healers or private doctors (34.1%). Over half of patients (53.3%) were on self medication such as home remedies and cough syrups during the initial stage of health problems. Some patients in tribal areas had to travel nearly 40 km to reach health center with microscopic facilities, for which patient delay is obvious. In our study majority of patients were on self-medication which caused patient and health system delays but were statistically insignificant. Studies carried out in Vietnam, USA and India have shown that self-medication is a predictor of delays in diagnosis and treatment of TB [1,4,9,10,12,19–22,24–27].

Patient delay was high in our study, which may due to various problems faced by tribal patients. The most common self reported reasons for patient delay were lack of awareness about TB which was significantly associated and was nearly equal in either sex which may be ascribed to illiteracy or due to limitations of non tribal health workers in motivating the tribal patients in their preferred local language. Tribal community leaders appointed as DOTS providers can be an effective connecting link between the programme and the tribal community which needs to be looked upon.

In our study patient delay was significantly associated with other factors such as illiteracy, caste, cost of treatment/ transport and distance. Mode of transport/treatment too influenced patient delay, which is directly linked with money spent as patients had to come from distant places. Although “direct costs” of diagnosis and treatment were significant for poor families, the greatest economic loss occurs as a result of “indirect” costs, such as loss of employment, travel to health facilities, sale of assets to pay for treatment-related costs, and in particular, lost productivity from illness and premature death. Similar findings were reported from studies in Malaysia, Lusaka, Vietnam, rural China and India [4,10,12,16,24,27–29]. Factors like old age, loss of wages, poor socio-economic condition and drinking habits also contributed to patient delay but were statistically insignificant.

The most common reasons for health system delay was (48.60%) due to administrative verification regarding DOTS provider presence in the area and delay due to visit to traditional healers/private practioners (43.1%) after diagnosis. Both the factors were significantly associated. Studies conducted in the other countries have also highlighted the weak links at the point of first contact such as traditional healers in Tanzania, Nepal, Ethiopia, Peru and private practioners in Ghana and Sikkim [10,13,18,30,31]. The concept of private public mix though has been successfully exemplified in non- tribal areas but seem to be non existence in this tribal region as there is minimal presence of private sector. So efforts should be taken to encourage NGOs and set up more health facility in the area. Besides the above other socio demographic factors that were found significantly associated with health system delay were family size, occupation, distance and monthly income [24].

The present study had some limitations regarding the duration of patient delay which may have been little underestimated as new definition for “TB suspect with 2 weeks duration of cough” (2009) had not been taken into account and another was of poor patient recall of onset of symptoms, their duration and date of the first health-seeking encounter.

The period before diagnosis and start of treatment is important since most of disease transmission occur during this time. However, a big chunk of tribal TB patients had delayed for more than 21 days. Although most patient characteristics were statistically insignificant with patient and health system delays. Illiteracy, lack of awareness about TB, long distance, cost of treatment/ transports were strongly associated with patient delay. The choice of traditional healers /private practitioners, occupation, and administrative verification were strongly associated with health system delay. To reduce patient delay it is crucial to improve TB knowledge, address TB misconceptions in tribal dialect, ensure the involvement of traditional healers and pay closer attention to setting up of treatment/microscopic centers at accessible distance. Health system delay can be reduced by developing public private partnerships and flexible administrative approach in hard core reach areas.