In this study, the management of acute watery diarrhoea more frequently included the use of ORS (58%) as compared with a previous Indian study, where ORS was prescribed to only 22% of the children . The correct treatment according to the guidelines is ORS with Zinc, without the use of probiotics, antibiotics, racecadotril or antiemetics (with the exception of Domperidone for severe vomiting). This was prescribed in only six out of 843 prescriptions. ORS with zinc was prescribed in 22% of cases, but these prescriptions also contained other drugs (antibiotics, probiotics, antiemetics and racecadotril), which are not recommended in the guidelines. The improvement in the proportion of cases in which ORS was prescribed may be due to the increasing awareness amongst the practitioners and the community about its importance in treating diarrhoea. This increased awareness may be due to mass media campaigns about the importance of ORS. These campaigns have not focussed on the benefits of zinc, so there may be less awareness of these among practitioners as well as the community. This issue warrants further exploration by both qualitative and quantitative studies.
Randomised controlled trials from low and middle income countries have shown that zinc is a cost effective intervention, along with ORS, in reducing morbidity from diarrhoea , but knowledge of health care providers about its proper use is lacking. A retrospective study  carried out in a private tertiary care hospital of Chennai, India, to study the pattern of prescribing of zinc and antibiotics for acute watery diarrhoea showed that the use of zinc had increased to 75% over a three-year period. This was accompanied by a decline in the use of antibiotics to below 30%, which was achieved through education of health care workers on the use of zinc. This suggests the need for education of the benefits of using zinc in acute diarrhoea to decrease the duration and severity of diarrhoea, as well as appropriate antibiotic use. In the present study antibiotics were prescribed for 71% of patients, compared with the WHO recommendations of around 20%.
There was deviation from the guidelines with regard to probiotic and racecadrotil use in our study. Probiotics were prescribed in 68% of cases, and racecadotril in 19% of cases; according to the guidelines neither of these should have been prescribed at all.
Some possible reasons for the high rate of prescribing of drugs not recommended in the guidelines are: first, the natural history of diarrhoeal episode in which treatment with ORS neither shortens the duration of diarrhoea nor decreases the stool volume. There is a problem with the 'image' of ORS and zinc as effective treatment options. Up to 40% of children revisit the doctor as they have significant stool frequency for more than 4 days, despite taking ORS . As ORS does not reduce diarrhoea duration, private paediatricians and other private health care workers look for alternatives to reduce the duration of diarrhoeal episodes [17, 18]. Second, vomiting is often an important issue for parents as well as physicians during ORS therapy (ORT). The cause of vomiting in acute diarrhoea is usually hypokalaemia. Vomiting can be both caused and aggravated by incorrectly prepared ORS (hyperosmolar ORS). Parents may be discouraged to continue ORT because of this vomiting, leading to a failure of oral rehydration . Severe vomiting should lead to prescription of domperidone and not other drugs according to the IAP guidelines. However, the perceived ineffectiveness of ORT may have lead to a growing interest in adjunctive treatments such as antibiotics, probiotics and racecadotril . These drugs are also highly marketed by pharmaceutical companies, with incentives for practitioners, playing a major role in prescribing. Third, the presence of fever increases antibiotic prescribing even though most diarrhoeal episodes are of viral origin . Consequently fever remains an important and unjustified reason for prescribing antibiotics for many self-limiting infections. Fourth, rural patients contributed to hospital admissions for acute diarrhoea more frequently in our study. They received 61% of the total antibiotic prescriptions and 71% of the total intravenous rehydration prescribed. This may reflect: a) referral of sicker patients to the city; b) demand for parenteral therapy for acute diarrhoea by patients; and c) the practitioners' preference to admit rural patients in order to monitor them. Starting IVFs and antibiotics for these patients can 'justify' the admission . Fifth, most prescriptions that contained ORS with zinc were from practitioners working in the free hospital attached to medical college. Interestingly it was also the hospital with the least antibiotic prescribing (7.6%). However, there was deviation from the guidelines in terms of antiemetic (ondensetron) prescribing. Most of the prescriptions in the free hospital attached to the medical college came from fresh medical graduates or post-graduate practitioners working in the paediatric department. Working in an academic institution with access to updated information might be responsible for the low antibiotic prescribing seen . Practitioners working in the free and the charitable hospitals were more likely to prescribe in accordance with the guidelines than those in the private sector. A reason might be that the private physicians needed to provide symptomatic relief faster for the fear of losing patients to another physician, and therefore used more drugs for fever, stomach pain, vomiting and antibiotics . This is illustrated in the present study by the high ICC for antibiotic prescribing, which indicates the homogeneity within each cluster.
The AYUSH practitioners and the IHPs prescribed ORS with zinc less frequently and were found to prescribe antibiotics more frequently, underlining that these practitioners need to be made aware of standard treatment guidelines if they have to practice modern medicine, and must not depend upon the information provided by pharmaceutical companies . Continuous professional development programmes, consisting of educational meetings alone or combined with other interventions, have been shown to improve clinical practice and patient outcomes . However, whether these interventions work for the practitioners of alternative systems of medicine remains to be documented. There is evidence from India  that education regarding zinc use can improve prescribing of ORS and zinc together. Thus, to improve adherence to treatment guidelines we recommend updating the national medical curriculum to incorporate these guidelines in all the systems of medicine. Changing behaviour by utilising private health care providers and pharmaceutical representatives, as trainers in workshops, may be critical to achieve training needs .
This study included the prescriptions from providers of alternate systems of medicine and the IHPs along with the practitioners' of modern medicine to understand the treatment pattern for acute diarrhoea in children. The pharmacy assistants carried out the data collection in pharmacies and resident medical officers in the hospitals. This was done to minimise the Hawthorne effect, whereby the prescriber would have been biased in an interview or if directly observed. Pharmacies were considered the most appropriate site for data collection as most of the outpatient prescriptions reach the pharmacies for purchase of drugs. Also, pharmacy assistants have the basic information on commercial names for drugs. We built on this basic information to easily train them in completing data collection forms.
The unexpectedly high design effect affected the power of our statistical analysis, and this may have resulted in us being unable to identify other important variables, which affected the prescription patterns. We did not analyse data at the level of the individual prescriber, as the name of the prescribers was not collected for ethical reasons. We could not examine the children clinically for the degree of dehydration, systemic diseases and the status of nutrition, so we cannot comment on whether the admission or the prescription of intravenous fluids was appropriate. The study was not able to take into account the 'prescriber dispensing' of medicines at their clinics.