Madam, Recently, Khawaja et al. published a cross-sectional survey to find out the prescription pattern of Benzodiazepines in a tertiary care hospital. 1 They used Chi-square test and Student's t-test to determine statistically significant observations. They have shown in Table 1 that after grouping patients into two broad categories of surgical and non-surgical patients on the basis of admitting service, the use of benzodiazepines was more frequent among surgical patients than those admitted under non-surgical service. However, when we group patients into surgical and non-surgical ones, there would be some patients who are suffering from malignancy and are admitted for surgery. These patients usually receive a shared care by oncologists and the surgeons. In such a case the indication of benzodiazepines can come through any of the two services. The methods section does not specify any algorithms to determine the indications. This creates confusion at their conclusion that "benzodiazepines were prescribed more frequently by the surgeons than the physicians, psychiatrists or oncologists". On the other hand if we assume that they used multiple response option for shared care cases, the observations do not remain independent which is violation of one of the basic assumptions of Chi-square test. Chi-square distribution assumes that the observations are independent and this limits the application of Chi-square distribution in situations where all the observations are not independent. 2 This renders the conservative interpretation to be very difficult. Zisselman et al show that the relationship between admitting service and frequency of prescriptions is statistically significant which is not exactly similar to that of Khawaja et al. 3

Hassan Khan 1 , Abdul Waheed 2
Second1 and Final 1 Year Medical Students
The Aga Khan University, Karachi.

1. Khawaja MR, Majeed A, Malik F, Merchant KA, Maqsood M, Malik R, et al. Prescription pattern of benzodiazepines for Inpatients at a Tertiary Care University Hospital in Pakistan. J Pak Med Assoc 2005;55:259-63.

2. Kuzma JW. The Chi-square test. In: Kuzma JW eds. Basic Statistics for the Health Sciences. 2nd ed. California: Mayfield Publishing Company, 1992, pp 189-90.

3. Zisselman MH, Rovner BW, Kelly KG, Woods C. Benzodiazepine utilization in a university hospital. Am J Med Qual 1994;9:138-41.