Tuberculosis (TB) remains a major public health problem in the world [1]. Detection and treatment completion of TB-infected cases are crucial for preventing and controlling this disease [1-4]. TB monitoring and controlling programs in Indonesia have been performed for a long time and refer to the WHO recommendations in the form of directly observed treatment short-course (DOTS) strategies [3, 4]. Compliance with TB treatment is the key factor that influences TB control in various countries [5, 6]. In 2008, Mateus-Solarte and Carvajal-Barona showed high non-compliance of TB medication in Colombia (65.6%) [6]. Similar results have been reported in other studies, according to which 50% of new TB patients did not complete their treatment in India [7]. Zhou reported a lower percentage of 16% for TB patients who did not complete TB treatment in China [8]. The Indonesian Tuberculosis Prevalence Survey 2013-2014 showed that 40.2% of TB patients discontinued treatment before being cured [9]. In most cases, treatment is discontinued during the continuation phase [4, 7]. In the continuation phase, most patients tend to feel healthier and some are asymptomatic. Such conditions increase the risk of non-compliance with treatment for TB [4, 7]. The incomplete TB treatment can lead to antimicrobial resistance of M. tuberculosis. This condition might contribute to the increased cost of TB elimination [5-8, 10].

Family support can influence TB treatment compliance. This is in line with the reports by Mateus-Solarteu (2008), who stated that the factors that influence treatment non-compliance include less family support and living away from the family [6]. The WHO endorsed a specific strategy named DOTS to improve compliance by requiring health workers, community volunteers, or family members to observe and record patients swallowing each dose. Several studies have shown the effectiveness of family roles in supporting successful TB treatment [11-13]. Wright et al. in 2004 proved that family members could play as good DOTS officers. In addition to reducing government costs, this can also reduce the social stigma of TB patients [11]. Similar results were reported by Newell et al. in 2006, stating that DOTS by the family and DOTS by the community had the same effectiveness in meeting TB treatment targets. DOTS by family members can enable treatment in remote places, such as the mountains, where it is otherwise difficult to perform community DOTS [12]. Research conducted by Dave et al. in 2015 also proved that DOTS conducted by family members is as effective as that performed by a trained DOTS officer in the community [13].

Family support plays a major role in determining the treatment outcomes, especially in terms of medication compliance [12, 14, 15]. Family support can be developed with intentions and practices [16]. Intention, according to Fishbein and Azjen, is a subjective probability or possibility, that is, a person's estimate of the possibility of performing a particular action [16, 17]. Intentions and practices of family support in encouraging TB treatment compliance in the continuation phase can be built through appropriate training. This is consistent with the report by Dave et al. (2015) and Shen (2017) in different settings that prove that after getting training, families can enforce DOTS effectively for TB management [13, 18].

Tuberculosis Family Support Training (TB FaST) is conducted to increase the intentions and practices of family members in encouraging compliance to the treatment of TB patients in the continuation phase. This study aimed to analyze the effect of Tuberculosis Family Support Training (TB FaST) on the intentions and practices to improve compliance in the treatment of tuberculosis patients in the continuation phase.

The characteristics of the subjects from both groups are shown in Tables 1 and 2.

Table 1 shows no significant differences in the characteristics of the family members (p > 0.05) of the intervention group and control group. The groups were homogeneous.

The results of the statistical analyses presented in Table 2 show that the two groups of TB Patients did not differ significantly in terms of age, sex, education, occupation, and marital status (p > 0.05 for each variable). The groups were also homogeneous.

As seen in Table 3, for all three parameters collected before training, there are no significant differences between the intervention and the control groups (p>0.05). These results prove that the two groups are statistically similar in the parameter measurement baseline, either in the intention, the practice done by family members, or the family support accepted by the patients. On the other hand, the results of the three parameters collected five weeks after training are varied. There is a significant difference between the two groups on the practice done by family member parameter at five weeks after training (p<0.05), which is confirmed as family support accepted by the patient with a significant result as well (p<0.05). Unfortunately, the two groups on the intention parameter at five weeks after training did not differ significantly (p>0.05).

In the intervention group, the mean of changes in the intention parameter was higher than the control group, but the statistical analysis showed there was no significant difference between the two groups (p>0.05). In contrast, in practice done by the family members parameter, the mean of changes in the intervention group was higher than in the control group significantly (p<0.05). Likewise, the family support accepted by the patient parameter, used as confirmation of the practice done by the family members parameter, also showed a significant difference (p<0.05) between the two groups.

Table 5 reveals the dynamic changes in participants’ intentions at different times of data collection. The mean of changes in family members' intention immediately after training was significantly higher statistically (p <0.05). Unfortunately, this high intention is not long-lasting. A lower mean of changes was observed five weeks after training compared to immediately after, with a significant difference (p <0.05). Despite the decrease in intentions five weeks after training, the mean of changes in intentions was still higher compared to before training, but the results of statistical tests showed no significant differences (p> 0.05).

The success of TB treatment is influenced by treatment compliance [5, 6]. TB treatment requires a long time and involves the initial and continuation phases, and need a commitment from the patient to complete the treatment. This commitment can be well maintained by the presence of family members and their support [14]. The continuation phase of TB treatment is very challenging, where patients tend to feel healthier and may become asymptomatic. Such conditions might risk the occurrence of non-compliance with TB treatment by making the family careless about giving attention and support for treatment completion [4, 7]. This indicates that the continuation phase of TB treatment needs more attention; therefore, greater efforts are needed to prevent treatment non-compliance in this phase. The TB FaST is one of the efforts that we have developed to prevent non-compliance with TB treatment in the continuation phase that aims to increase the intentions and practices among family members to support TB treatment.

This study measures the intentions and practices of family members in giving family support to the patients. This study’s results showed changes in both the intentions and practices after being given TB FaST training. The changes obtained in this study were positively significant; it can be seen from the increase in the practices of family members after being given training and confirmed by the significant increase in family support accepted by TB patients compared to the control group. The family support consisted of four types of support: informational, emotional, instrumental, and appreciation support [20]. The increased practices that occurred after training following the results achieved by Dave (2015) and Shen (2017), as the training carried out for family members can affect the practices of family members, which means the family can function properly as a DOT supervisor in completing TB treatment [13, 18]. The role of the family as a DOT supervisor in completing TB treatment is a form of family support, particularly instrumental support.

The increase in family support practices that occur cannot be separated from the underlying intention. This study revealed that the intention could be built through the training provided, as seen from the increase in the intention measured immediately after the training. It is in line with the increase in practices measured five weeks after training. These results showed us that the intentions could be a potential predictor of the practice to be taken. Azjen and Fishbein have proven that the process of forming behavior begins with a strong intention to perform a certain behavior [16, 26]. Philipa Lally (2013) states that intentions and practices are part of the habit formation process to become a behavior [27].

TB FaST provides knowledge and education about TB disease, the importance of completing TB treatment, the role of families in preventing TB treatment non-compliance, opportunities to share experiences between families of TB patients, opportunities to practice problems identification, and means to provide family support regarding the TB treatment completion. The knowledge and experience gained from the training contributed to escalating the intentions of family members. This is in line with a previous study by Ukoli et al. (2013), Odedina et al. (2014), and Shaleh AM et al. (2015), where educational programs can broaden the knowledge and strengthen the intentions to support healthy lifestyle compliance [28-30].

The training was provided in the form of short-term education using systematic and organized methods and procedures [31, 32]. Delivery methods used in this training included lectures with a question-and-answer session, small interactive focus-group discussions (problem-based discussion), and role-play. The method of lectures with question-and-answer sessions can build a knowledge base regarding diseases and family functions in dealing with diseases. Problem-based interactive discussion methods can help in training participants to strengthen their knowledge base, shape critical thinking patterns, build an ability to communicate and express opinions, and provide opportunities for sharing experiences to overcome problems. The role-play method actively engages trainees in activities and provides an opportunity to experience and experiment with overcoming problems so that training participants are expected to have the readiness to take action in supporting treatment compliance. These three methods can synergize to develop an awareness of receiving information, responding actively, and conducting an assessment so that it can form a strong desire for action to be taken [32, 33].

The dynamic changes in the intention are observed in this study. An increase in intention immediately after training was followed by a decrease in intention five weeks after training. The intention is very dynamic and depends on the individual condition, including attitude aspects toward behavior, subjective norm, and perceived control behavior [33, 34]. This allows the intention of each individual to experience an increase or decrease. Ajzen and Fishbein explain that the dynamics that occur in an individual when taking certain actions are related to the three aspects mentioned above [26, 33, 35, 36]. The intention of some behaviors can change simply as a function of time, some depending on the emergence of new information. This raises an important implication in predicting an action by estimating intentions. When the intention predicts that a person will or will not take an action, but it shows a failure to predict the achievement of a certain action, then this is likely due to factors outside the individual's control that prevent him from carrying out his intentions [33]. Regarding our study results, further research may be needed to identify factors that may contribute to the changes in intention based on these three aspects.

In addition, as mentioned above, the dynamics of the intentions of family members show the need to perform assessments more frequently. This aims to determine the right time to re-intervene in strengthening intentions, thus, TB FaST is given at the right time interval, and so family members still manage to have a high intention in providing support for TB patients. TB FaST can affect the intention of family members if it is held at the right time, so it is important to determine the proper intention measurement.

In this study, the family practice measurement after training was performed only once; thus, it may not reflect the response to the intention measured five weeks after the training and cannot be interpreted to indicate improved practices and behaviors. This study did not assess the factors that predisposed the intention; thus, the factors that influenced a worsening of the intention were not identified.

Further studies with a bigger sample size are encouraged to determine the factors that can influence the intention to realize an action and the need to measure the intentions in stages to determine the pattern of change in intention after complete training to enable the identification of the optimal time to intervene again. For better achievement, we need to extend the use of TB FaST not only to family members but also to TB patients. This training can help the patients better understand and preserve the commitment to complete the treatment. TB FaST can be used as a sustainable health promotion program in primary health care to provide internal reinforcement in the family environment and realize family empowerment.

The present results showed that the TB FaST strongly influenced the practices of family members in increasing treatment compliance in tuberculosis patients during the continuation phase. We recommend that TB FaST or other educational tools be re-administered following the TB FaST to improve the intention.