The Pharmaceutical Journal
Vol 274 No 7351 p644
28 May 2005

This article
Reprint   Photocopy

PDF 40K, Acrobat Reader

Research is not just for academics

By John Wilson

Last year, while browsing through my archives (piles of junk, according to my wife), I came across some old supplements to the PJ which carried abstracts of the practice research papers and posters from the British Pharmaceutical Conference. Looking at these in more detail, I realised that almost all of the abstracts came from either universities or hospitals. Very few were from individual community pharmacists.

A quick inspection of the 2004 BPC abstracts, using the internet, showed that out of almost 100 practice research contributions there were two from community multiples, one from a practice pharmacist at a GP surgery, one from a group of six locums, two from a pharmacy research network in Yorkshire and just one from an individual community pharmacist.

This set me thinking — why are there so few contributions to original research from community pharmacy? The relentless pressure of work may be one reason, but then hospital pharmacies are under pressure, too. Perhaps there is not much opportunity to develop ideas which could be turned into viable research projects and thence produce papers? However, some pharmacists have managed to do it.

Community pharmacies dispense over 600 million prescriptions each year and the number is rising. This means that over 600 million different items of data pass over our dispensing benches in the course of a year, and these could be a fruitful resource for research. I decided to put this to the test in a small study. My first research question was: “Could a locum pharmacist collect sufficient data from dispensed prescriptions to carry out a useful investigation?”

Antibiotics are frequently dispensed, and here was the “germ” of an idea. Most primary care trusts, like hospitals, probably now have antibiotic policies or guidelines. How far do prescribers adhere to them? Is there any pattern in terms of age/gender to antibiotic dose and, equally importantly, course length? Did the antibiotic courses prescribed match those in the various antibiotics guidelines? These were research questions that might be answered simply by collecting data from FP10 prescription forms.

I advertised in the “Wants” column of the PJ for copies of primary care antibiotic guidelines, and received only two. One had fairly specific recommendations as to course length for various antibiotics, in a range of clinical conditions, so I decided to adopt this as the standard.

I collected data from over 700 prescriptions for antibiotics in a solid oral dosage form and from 200 prescriptions for children’s antibiotic mixtures. This was done in the course of my locum duties at about 14 different pharmacies over several months. As each of the above prescriptions was dispensed, the data taken were: age and gender of patient; drug; dosage form; strength; daily dose and course length. In some pharmacies, at quiet periods, I was also able to collect data from the filed prescriptions waiting to go to the pricing bureau. Antibiotic courses which were clearly long-term, such as two months’ supply of oxytetracycline for acne, were omitted from the study. The pharmacist member of a hospital ethics committee advised me that such data collection would not require ethics committee approval.

Interesting analysis

Analysis of the data proved interesting. For the solid oral dosages, there were 18 different antibiotics prescribed. The drug with highest usage (almost 40 per cent of the total number of courses) was amoxicillin, followed by flucloxacillin, cefalexin, erythromycin, trimethoprim and penicillin V, in that order. The other 12 antibiotics together accounted for only 5 per cent of the total.

Dental prescriptions accounted for 10 per cent of the total, while 61 per cent of all antibiotic prescriptions were for females and 39 per cent for males. In general, a higher proportion of women in the younger age groups received antibiotics, while in the older age groups there was a higher proportion of men. This is slightly at variance with results from a national study using the General Practice Research Database, but this is not altogether surprising considering the small size of my sample.

To examine course length, I took, first of all, amoxicillin as the study drug. Course length was recorded as “<7 days” or “>=7 days”. For GP prescriptions, the course length for different dosages are shown in Table 1.

There was an overall preponderance of “>=7 days”. However the difference between course length with the 250mg and 500mg strengths was marked (x² = 9.10, P lies between 0.01 and 0.001 and is, therefore, significant). I concluded that although there was a preponderance of “>=7 days” for both strengths, the difference is marked for 250mg. Thus, prescribers tend to give longer courses of 250mg than of 500mg.

With trimethoprim, frequently used for uncomplicated cystitis, the usual recommended course length is three days. In my survey, less than 40 per cent of the courses of trimethoprim were for the recommended three days.

The situation with regard to antibiotic mixtures for children was different. Out of a total of 200 courses, 189 were for one 100ml bottle. The course length was therefore pre-determined (five days for penicillin V and flucloxacillin, seven days for amoxicillin) by the dosage frequency and the bottle size. The 11 exceptions were mostly for flucloxacillin and were for 140ml (seven days) necessitating the use of two bottles.

This small study shows what could be achieved by a locum pharmacist working on occasional days at a number of locations. I accept that with the small number of prescriptions the study does not lend itself to statistical analysis. Also, data collection may have been biased since the pharmacies that acted as the data sources were determined only by my locum bookings. However, the results point to a possible problem with antibiotic prescribing. Surely it is not rational to vary the dose and course length for adults, while using, in the main, the same dose and course lengths for children of varying ages, just because of the bottle size?

Other possible studies based on the analysis of prescriptions spring to mind. For example, there has recently been a great deal of concern over medicines for children. What, exactly, is prescribed for children in the community (other than antibiotics and Calpol)? Data could easily be collected on a large number of children’s prescriptions if a group of pharmacists each collected data in the manner described above and pooled the results. What other drugs do patients taking warfarin also take? If a group of, say, 100 pharmacists each checked the patient medication records for 10 warfarin patients and pooled the results, we would have data from 1,000 patients. This would be enough for a valuable study which, to my knowledge, has not been done before. The possibilities are limited only by our imagination.

In conclusion, my first research question appears to have been answered firmly in the positive. I have to confess, however, that I have now joined forces with an academic unit to take the antibiotics study further. Thus, my question has led to further work. That is how knowledge progresses — one small step at a time.