The adherence to policies of the National TB Control Programme (NTP) to manage a case of tuberculosis (TB) is a fundamental step to a successful programme in any country. Since the initiation of the campaign to control tuberculosis, a positive change has been observed in managing TB cases in the Eastern Mediterranean region [1] with e.g Pakistan reaching 100% DOTS coverage in year 2005 meaning all districts having a TB control setup. Currently the NTPs in the high TB burden countries are dealing with the components of the TB control strategy within which management of childhood TB remains an essential component for those to be offered care [2].

The diagnosis of pediatric tuberculosis is usually made on the basis of clinical evidence and is difficult particularly in infants and young children in whom the symptoms are often non-specific [3]. Several studies have documented the poor diagnosis and treatment practices of clinicians while managing a case of tuberculosis [4–6]. The World Health Organization (WHO) has published guidelines on the diagnosis and management of TB in children in 2006 [7]. There are several diagnostic criteria that have been detailed in the guidelines including the use of clinical symptoms, sputum microscopy, chest radiograph (CXR), tuberculin skin test (TST) and scoring charts. Childhood TB services face a challenge of case management due to difficulty with diagnosis in clinical practice [8] and the recent introduction of international policies [7].

For the control of childhood TB in Pakistan the NTP has developed and piloted its guidelines in 2006-2007 in the public sector hospitals in ten selected districts. The policy guidelines were developed in collaboration with the Pakistan Pediatric Association and consist of protocols with a purpose to improve care of children with TB [9]. The guidelines were introduced by training pediatricians in the pilot districts, supplying anti-TB drugs, introduction of TST and monitoring support. A retrospective review of the piloting experience in three of these districts was conducted to compare the case notification and treatment outcomes of children with TB in 2004-5 with 2006-7 [10]. The objective of this study was to compare the documented case management practices of pediatricians and its impact on the outcome before and after introducing the NTP policy guidelines. Our hypothesis was that the implementation of new NTP policy guidelines would improve the case management practices and the treatment outcomes among children with tuberculosis.

In our study we have shown that there was a substantial increase in the childhood TB case notification from 189 pre-intervention to 731 post-intervention after the implementation of new policy guidelines. The NTP provided the training on policy guidelines for childhood TB and logistic support including TST and anti-TB drugs to initiate the childhood TB care in the public sector district level hospitals. Before these NTP policy guidelines adult anti-TB drugs were provided free of cost and the individual dosage was given per kg bodyweight. Later the child specific anti-TB drugs were also provided along with TST. Based on available information we cannot say how much of the increase in TB notification in children is due to guidelines or logistic support or other issues. The training was provided to the pediatrician and not to the DOTS facilitator (paramedics and nurses) who were supposed to be involved in the R&R activity. This resulted in weak implementation of R&R with incompleteness of records. The use of tuberculin skin test after introduction of the NTP guidelines increased. The TST was not initially available from the programme and after the implementation of the new policy its use by the pediatricians in the programme conditions was encouraging. The TST as a tool to support diagnosis of TB among children is well established and still considered to be the best available method to detect TB infection [12]. It can be used as an adjunct in diagnosing TB disease [13] and a regular stock of tuberculin should be maintained in hospitals involved with managing childhood TB [14].

The documented use of chest radiographs to assist in diagnosis increased among both pulmonary and extra-pulmonary TB cases. The use of CXR in making the diagnosis of children with TB has been supported internationally and a majority of cases with children with pulmonary TB have CXR changes suggestive of TB [13]. The CXR should always be read by a trained health care provider. Since NTP adopted the DOTS strategy in 1995 the emphasis remained on the use of sputum smear microscopy which downplayed the importance of CXR in supporting the diagnosis of smear negative TB. In a study in Malawi over 80% of the childhood TB cases the CXR finding were consistent with TB [4]. The increased use of CXR emphasizes the need of structured training of heath care providers on reading the CXR among children with TB; if not done it may have consequences on the quality of diagnosis and care. The NTP policy guidelines had section on the use of X-ray to help the pediatrician diagnose a child with suspected TB. The quality control mechanism was not well defined with a possibility of missed cases or over diagnoses. To address this limitation the NTP Pakistan has recently developed a training module having sections on reading chest X-ray in children, adapted from international literature, which will address this important issue related to case management. There are also international guidelines available to assist the health care workers in low income countries to interpret chest radiographs of children suspected of having TB [15]. The diagnosis of TB in low-income countries needs to be improved based on the available technology and resources [14].

In our study some children had sputum smear done, although this did not increase with introduction of the new policy guidelines. Very few of the cases below 5 years of age got sputum smear done which shows the challenge faced by the health care providers in getting smear done in very young children within routine programme conditions. The NTP childhood TB case management policy and involving pediatricians in TB control activities were recent activities. We observed an increase in case notification during post-intervention phase but the linkage between various investigation facilities to operationalize diagnostic algorithms within the hospital was limited. The role of laboratory services in childhood TB diagnosis including sputum smear microscopy and specimen collection by gastric lavage, was not well defined. The operational linkages between the pediatricians and laboratory services need to be strengthened in order to enhance sputum microcopy among children suspected with TB. Similar operational issues are relatively common in the start of a new initiative and gradually improve with programme implementation. More training is required for introducing techniques such as gastric lavage and induced sputum within routine programme situation in order to increase sputum smear microscopy among children. The use of the score chart was introduced by the new guidelines. The score chart helps suspect screening [7] and the NTP policy-aim is to have a score of each case diagnosed. The use of a score chart has been suggested in the situations where accurate diagnosis of childhood TB is complicated by the lack of diagnostic tools and facilities [16]. The diagnostic criteria documented in the score chart are simple and well structured (Figure 2) and can be applied in a resource poor setting like the health facilities in Pakistan [17]. More attention needs to be paid by the health care workers to document the use of score charts.

The study has shown that the proportion of cases with correct regimen has not changed from pre-intervention to post-intervention. There is wide variation in the practices of pediatricians regarding prescribing anti-TB drugs to the diagnosed children and none of the doses prescribed were found to be recorded correctly. To have good treatment outcomes the application of standardized treatment regimen needs to be according to the relevant diagnostic category [18]. This suggests that NTP could benefit from more stringent process monitoring of the childhood TB care to improve practices. The treatment support and treatment outcome were recorded in a low proportion of the children and the proportion of cases documented as cured even lower in the post-intervention, than in the pre-intervention period. The purpose of our study was to compare practices before and after introduction of guidelines, and not the difference between districts or hospitals; enhanced childhood TB case finding has shown variations across the three districts studied, which may related to difference of individual practices. We have shown in a previous paper that the high number of cases notified and the high number of unknown outcome (pre-intervention 21.7%, post-intervention 73.3%) were from one of the three districts because of absence of adequate measures to case follow-up [10]. In studies conducted in Malawi and Papua New Guinea unknown outcome was reported in 21% and 23% of notified cases [19, 20]. The training and logistic inputs remained standard for all these districts in the post-intervention period. The incomplete reporting of treatment outcomes may be due to the practitioner's level of commitment, work load and limited experience with the new intervention. Similarly the patient/families may have failed to adhere to treatment due to socio-economic reasons. These issues are more common at the start of the programme and gradually settle down with more structured monitoring of process of care and outputs, involving the community volunteers in treatment support, awareness and social mobilization.

Patient treatment cards are recommended for documenting treatment adherence [18] and children with TB should always be included in the routine NTP recording and reporting system [21]. It is an advantage for the childhood TB control that adequate attention should be given on the R&R system. Pediatricians should be engaged in diagnosis and treatment. The already trained and available personnel should be oriented and involved in R&R to capture childhood TB cases, with a provision of additional human resource where required, so that all clinicians dealing with childhood TB are trained to do it. This further supports the arguments that with systematic improvement of childhood TB case management practices in the public sector we could achieve good results.

A strength of this retrospective study is that it is based on the review of available patient records under routine programme conditions at nine selected hospitals in public sector in three districts with previous adult TB care experience and good collaboration with NTP. One major limitation is that many patient records were incomplete and it was not possible to get more information on each of the patients from the hospitals.

The study has shown that pediatricians have started following parts of the national policy guidelines for management of childhood TB. In the diagnosis of childhood TB the documented use of TST, CXR and score chart increased but the same was not observed for sputum smear which actually declined. Record keeping of case management practices including diagnosis and treatment support remains inadequate. The impact of these practices on treatment outcomes was scanty with too many unknown outcomes. Perhaps the absence of operational tools to guide the process of care and monitoring to improve the case management practices and outcomes confines the achievement of desired results. Development and implementation of standardized operational tools and regular monitoring system is required to improve the performance and quality of services.