Background
Rabies remains a major public health threat in Asia with an estimated 39,000 deaths annually, mostly due to spill over from the canine reservoir [
1]. Wider use of post-exposure prophylaxis (PEP) might reduce human mortalities in this region of the world [
2]. On the other hand, the escalating cost of life-saving PEP represents a major burden to both national economies and families, mostly in poor rural communities [
1,
3,
4].
In Bhutan, the number of reported animal rabies cases was stable in the decade 1996 to 2005 but increased during 2006 to 2008, mostly in 4 districts in southern Bhutan bordering India [
5]. Maintenance of rabies in the canine reservoir in southern Bhutan was likely due to low coverage of dog vaccination programs. Since 2009, mass dog sterilization and vaccination programs contributed to a decline in the incidence of canine rabies [
6]. However, the disease remains endemic in southern Bhutan and in some pockets in eastern Bhutan. All 13 human deaths due to rabies recorded in Bhutan between 2009 and 2017 were reported from southern districts where the estimated average annual incidence was 0.4 deaths/100,000 population [
7,
8]. Past studies estimated that PEP intervention effectively averted about 15 human deaths annually in rabies endemic areas of Bhutan [
9]. Hence the provision of free PEP is an important public health policy implemented by the Government of Bhutan.
PEP is a very significant on-going cost for the Government of Bhutan. Between 2009 and 2016, the annual number of reported dog bite victims presenting to health centers for PEP increased from 1000 to over 7000 [
7]. Ministry of Health records show that about Nu. 9.2 million (USD 143,000) was spent for procurement of anti-rabies vaccines and immunoglobulin in the fiscal year 2016–17, representing approximately 6% of the essential medicines budget (Ministry of Health Medical procurement records 2016–17, unpublished). While PEP is an essential component of rabies control in Bhutan, inappropriate use of PEP can result in substantial costs to the health sector [
10].
Implementation of PEP is guided by the Ministry of Health’s World Health Organization (WHO)-adapted National Rabies Management Guidelines 2014 (NRMG) which provides clinicians with criteria for categorizing rabies risk in potentially exposed people and recommends approaches to management of cases for the three risk categories. Prescription of Anti-Rabies Vaccine (ARV) is recommended for cases with a moderate or severe risk (NRMG Categories 2 and 3), while additional Rabies Immunoglobulin (RIG) administration is recommended for the highest risk category 3 exposures. To date there has been no evaluation of the management of human cases potentially exposed to rabies and the implementation of PEP in Bhutan. An evaluation would identify opportunities for improving the cost-effectiveness of PEP prescription to prevent human rabies cases whilst at the same time reducing wastage through unnecessary or inappropriate prescription of PEP in people that have a negligible risk of being exposed to the virus. Hence, the objective of this study was to evaluate clinicians’ management of human cases potentially exposed to rabies and PEP prescribing practices in the rabies-endemic areas of southern Bhutan. The results of this study were used to improve the cost-effectiveness with which this important rabies management policy is implemented in Bhutan to achieve the national goal of zero human rabies deaths in Bhutan by 2030 [
11].
Discussion
This is the first study to evaluate clinicians’ management and PEP practices for human cases potentially exposed to rabies in Bhutan. All 50 clinicians working in the 13 health centers in the high rabies risk areas of southern Bhutan were evaluated in the study, including 2 regional hospitals, 6 district hospitals and 2 BHU-Is. Amongst the 50 clinicians, 42% (21/50) were medical doctors with a MBBS degree, 10% were clinical officers who held a diploma and 48% were health assistants of whom 8% held a diploma and 92% held a certificate (Table
2). Doctors had a median of only two years of experience, while clinical officers and heath assistants had a median of 21 and 19 years’ experience respectively. Each clinician conducted a median of four consultations (range 1–19). Doctors and health consultants conducted a similar proportion of consultations, 45 and 42% respectively, while clinical officers conducted only 13%. Clinical officers practiced only in BHU while health assistants and doctors practiced both in BHU and hospitals.
Guidelines for management of rabies cases is provided in the NRMG (2014). This guideline is based on WHO recommendations. The NRMG recommend a rabies risk assessment is conducted for all cases exposed to “suspected or confirmed rabid animals” by collecting epidemiological information on the exposure history of the case and provides guidance for PEP for each risk category. The NRMG provides guidelines for classifying rabies risk into three categories: none, moderate or severe risk, and recommends prescription of ARV for those in the moderate and severe categories. Additional RIG is recommended for cases in the severe risk category. Where possible, we assessed the accuracy of clinicians’ risk assessments by comparing these with an independently classified risk assessment using the NRMG guidelines, based on the patient’s interview. We also evaluated the appropriateness of the PEP prescription by comparing the clinicians’ prescription against their own risk assessment and against the independent risk assessment. Clinicians recorded the rabies risk classification for only 71% of cases. However, they did collect some information on the type of exposure for nearly all cases, which could have contributed to risk categorization. Clinical officers did not record the risk category for 44% of the cases compared with 35% for health assistants and 19% for doctors (Table
4). It is not known if clinicians did not record the risk category because they did not make a decision about the risk category or if they did make a decision but did not record this. There is no clear definition for a “suspected” animal rabies case in the NRMG, which may contribute to clinicians not categorizing the rabies risk of a case. Since rabies is endemic in southern Bhutan, all animals involved in an exposure event should be suspected of being infected with rabies, with appropriate risk assessment performed and documented and PEP prescribed accordingly.
The clinicians’ rabies risk categorization showed a low level of agreement with the independent assessment using the NRMG (kappa = 0.203). Of the 194 cases for which clinicians recorded a rabies risk category, only 53% were correctly classified. Nearly all rabies risk assessments took into account the type of exposure (Fig.
4), the latter being clearly outlined in the national guidelines (Table
1). However, they often omitted or ignored relevant epidemiologic information necessary to classify the risk appropriately. As a result, clinicians tended to underestimate the exposure risk. It is a concern that clinicians mis-classified 13% of cases as having no risk, while they were independently assessed as having a moderate or severe risk. Overall, male health assistants were the most likely group of clinicians to make an accurate risk assessment, while female health assistants were least likely (Table
6). Surprisingly, doctors did not perform as well as male health assistants, similar to findings from an Indian study [
14]. Clinicians from district or regional hospitals were more likely to perform better than clinicians in BHUs (Odds Ratios of 7.8 and 17.6, respectively) regardless of clinician type. These results possibly reflect greater opportunities for clinicians in hospitals to participate in capacity building training programs for rabies conducted by the Ministry of Health, as observed in another study in Haiti [
15]. The attitude of clinicians towards training might also differ depending on their level of qualification. Health assistants are more readily available to attend such trainings. Most doctors, however, are unable to attend or oblivious to training sessions, possibly due to high workloads. The apparently better performance of male health assistants compared to male doctors may reflect the impact of such continuing education opportunities, rather than initial education level. The poorer performance of female health assistants compared to male health assistants could arise from fewer opportunities to participate in continuing education training by female staff. Additionally, male health assistants had significantly more work experience (median of 29 years) compared with male and female doctors (median of 2 years) and female health assistants (median of 10 years). The poorer performance of clinicians in BHUs compared to hospitals, independent of clinician type, might be associated with less training and poorer awareness of the NRMG guidelines amongst BHU-level staff. This might also partially reflect clinicians in BHUs having fewer years of clinical experience, even though the coefficients for type of health center were virtually unchanged when adding this variable to the model, after accounting for type of clinician. Nevertheless, junior clinicians are often posted to lower level health facilities rather than hospitals, as per government policies. Given the majority of rabies assessments (78%) occurred in hospitals rather than BHUs, there may be less opportunity for junior clinicians in BHUs to gain experience if this pattern represents the general pattern of consultations.
The inaccuracy of clinicians’ risk assessments was compensated for by the prescription of ARV for the majority of cases (91.6%, Table
7). The rabies risk in all cases but one was independently classified as moderate or severe, hence ARV was prescribed for most cases who needed this, based on their independent risk classification. However, clinicians prescribed ARV for 10 of 26 cases whom they classified as having no rabies risk. These results are reflective of findings of a nationwide study conducted between 2005 and 2008, which reported frequent PEP administration in category I exposures [
10]. The clinicians may be erring on the safe side with respect to administering ARV; they may also be under pressure from their patients to prescribe ARV. Alternatively, some clinicians may just be using ARV in the absence of conducting a rabies risk assessment. Such practices are likely to result in unnecessary costs associated with implementation of PEP to prevent human cases of rabies. On the other hand, clinicians did not prescribe ARV for 2.4% of cases whom they had classified as moderate risk and 1.2% of cases classified as severe risk. This could potentially represent a public health threat.
In this study, 8 cases were exposed to laboratory-confirmed rabid cattle or buffaloes through handling the carcasses and/or drinking milk from these animals. This exposure is considered a moderate exposure risk in the NRMG, requiring ARV treatment but not RIG. One of these 8 cases was incorrectly classified by the clinician as having no risk and the case did not receive the required ARV. Another two of these cases were incorrectly classified as severe risk. Both received ARV, and one also received RIG, which is not consistent with the NRMG. An additional 8 cases were exposed to animals reported as showing typical clinical signs of rabies that had not been confirmed by laboratory diagnosis; included 6 cases exposed to dogs (5 dog bites), 1 to a cat and 1 to cattle/buffalo. All 5 cases that were bitten by dogs exhibiting pathognomonic rabies symptoms received ARV, but no RIG. One of these cases should have been categorized as severe risk, given the presence of puncture wounds, and should have received RIG. Another 76 cases were exposed to animals that clinicians classified as “suspected of rabies” including 62 dog bites and 47 with bleeding wounds. According to the NRMG these cases should have been classified as severe risk and should have been prescribed RIG. Overall, RIG was prescribed to only 3 cases (1%), one exposed to a laboratory-confirmed rabid cow and the other 2 bitten by dogs for which no pathognomonic signs of rabies were reported. This concurs with the results of the earlier study that RIG was not regularly administered to dog bite victims in Bhutan [
10]. While WHO recommends RIG is prescribed for patients bitten by suspected rabid animals, the availability of expensive RIG is very limited in Bhutan, as is true in many rabies endemic countries [
16]. These results indicate the need for clearer guidance in the NRMG regarding prescription of RIG. Given the limited availability of RIG in Bhutan it could be valuable to define criteria for an “extra-severe” risk category to prioritize for RIG prescription. Another area of the NRMG that is not clearly understood by clinicians is the criteria for risk categorization associated with exposure to dairy products and meat consumption. Criteria for risk categorization of these exposures are not clearly outlined in the WHO guidelines, and were added to the NRMG due to the high prevalence of animal product consumption during animal rabies outbreaks in Bhutan. The above results indicate that there is a need to improve the clarity of guidelines in some areas of the NRMG and to train clinicians in the interpretation of these guidelines.
This study was limited by a small sample size and the limited 2-month study-period in late winter and early spring. Previous studies show that PEP prescriptions follow a seasonal pattern, with an increase in winter and spring [
10]; a very similar pattern is also observed in dog bites in Bhutan [
9]. Therefore, this study might not be representative of year-round case demographics. However, seasonality is unlikely to affect clinicians’ rabies risk assessment. Prior communication with clinicians requested them to manage cases as they would normally. Observational biases associated with the interviewers were mitigated by providing comprehensive training on information recording. However, clinicians may have been influenced towards conducting more rigorous risk assessments during the study. Despite these limitations, the results of the study are consistent with those of a previous study in Bhutan [
10].
Good clinical judgment is essential to prevent human rabies [
17]. Other studies on clinician’s knowledge and attitudes conducted in the USA showed an unsatisfactorily low level of compliance with national guidelines resulting in inappropriate PEP treatment [
18‐
20]. Greater compliance with guidelines is important to achieve more cost-effective PEP use [
21,
22]. For example, a study in a low rabies-risk area in Massachusetts, USA highlighted how large amounts of rabies PEP could be wasted in patients with low or non-existent risk of rabies exposure [
23]. Studies in countries free of rabies [
24] and in endemic areas [
25‐
27] similarly reported insufficient rabies risk assessment due to clinicians’ lack of familiarity with the recommendations, highlighting the need to update clinicians’ knowledge [
14]. Discordant rabies risk assessment and PEP practice was also apparent in our study. In contrast, public health physicians in Israel showed a very high level of compliance with PEP guidelines [
28]. This could be due to different public health policies, better training or expertise of clinicians undertaking rabies assessments and better access to immunoglobulin in Israel compared with Bhutan.
This study highlighted important gaps in clinicians’ management of human cases potentially exposed to rabies in high rabies risk areas of Bhutan. The cost-effectiveness of applying PEP to prevent rabies in these areas could be improved by reviewing and updating the NRMG and providing more training for clinicians in using the NRMG to manage cases potentially exposed to rabies. Our findings indicate that rabies training should target medical doctors, female health assistants and clinicians in BHUs and take into account the availability and the motivation of clinicians. In addition, the WHO recommends a One Health approach and the use of Integrated Bite Case Management integrating animal health and public health sectors [
4]. Better public health outcomes and concurrent PEP savings could be achieved by systematic notification to the animal health sector of human cases that have been exposed to suspected rabid animals. In turn, appropriate quarantine and surveillance of the animals involved can be conducted. If the animal is confirmed free of rabies there is no need for further administration of ARV for the exposed case(s), which can reduce the expenditure associated with PEP.
Based on the results of this study, there is scope for significant improvements of national rabies management policies and clinician training in Bhutan. This may contribute to reaching the national goal of eliminating dog-mediated human rabies by 2030.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.