Volume 22 N

r. 2 2000

Pharmacy W

orld & Science

Artic

le

Introduction Pharmacists in Great Britain now have a wider rangeof products to select from when recommending OverThe Counter (OTC) medicines due to the increasedre-regulation of many Prescription Only Medicines(POM) to Pharmacy Medicines (P). Other factors suchas consumer expectations, General Practitioners’increasing workload and high prescription costs havealso led to an increased demand for OTC productsand advice from the community pharmacy [1]. Tocarry out this function community pharmacists mustpossess a knowledge base of disease processes, druginformation and the ability to select a suitable remedyor recommend appropriate treatment.

The principles of clinical governance and evidence-based decision making are now well established asunderlying best practice. [2] Pharmacists must takeresponsibility for and be able to justify if necessary,their choices for recommendations of OTC productsfrom the wide range available. There are practicalguides available for the community pharmacist[3 4 5 6 7] which give information on drug treat-ments but most do not differentiate between prod-ucts. Other guides which deal exclusively with brand-ed OTC products, such as the Proprietary Association 47

An investigation of the factors affecting community pharmacists’ selection ofover the counter preparations• Emi ly Kennedy and Mar ta Moody

of Great Britain OTC Directory and the Chemist andDruggist Guide to OTC Medicines merely give infor-mation on the uses of the products and their compo-nents and do not prioritise products in relation to effi-cacy. The choices made by an individual pharmacistcould be influenced by a wide range of factors involv-ing advertising, experience, financial and social, andclinical issues. [8] As the community pharmacy oper-ates as a commercial enterprise it is of interest to dis-cover whether economic matters override profession-al decisions in any situation as has been suggested[9]. Previous studies [8 10] have identified factorsinvolved in pharmacists’ preferences for non-prescrip-tion cough suppressants and analgesics but influencesover a range of OTC medicines for a variety of condi-tions have not been investigated.

The aim of this project was to identify these influ-ences.

MethodA pilot study was carried out with a sample of 23pharmacists working in 15 community pharmacies inthe Grampian region, using semi-structured inter-views. The sample was selected in order to include asmany variables as possible in terms of type and loca-tion of pharmacy and in position and length of expe-rience of pharmacists. From the pilot study data theselection criteria for the conditions to be studied wereidentified and a questionnaire designed to be used asa postal survey. The selection criteria for the condi-tions to be studied were: regular presentation to com-munity pharmacists and eight or more products avail-able for selection. The conditions chosen were chestycough, dry cough, sore throat, haemorrhoids, hang-over and travel sickness. The pharmacist was asked tostate what first line product they would recommendfor each of these conditions in an uncomplicated situ-ation with no other relevant problems. Pharmacistswere then asked to indicate, using a five-point Likertscale [11], their level of agreement with eighteen fac-tors identified from the pilot study and a further ninefactors which had been identified additionally in aprevious study [8]. They were also asked if there wereany other factors which might affect their choices.Further questions were included to ascertain informa-tion sources and the use of complementary medi-cines. Information was also requested regarding thetype and location of the pharmacy, the position, gen-der and year of registration of the respondent andtheir participation in continuing education.

Using the Annual Register of Pharmaceutical Che-mists, 1995 [12] the type of each pharmacy was identi-fied i.e. multiple or independent. Using population den-sity, these were further divided into location groups,city, suburban, town and rural. Random numbers wereidentified by computer program and sampling for eachstratum was carried out, resulting in a stratified randomsample of 1138 registered pharmacies.

Pharm World Sci 2000;22(2): 47-52.© 2000 Kluwer Academic Publishers. Printed in the Netherlands.

Emily Kennedy and Marta Moody (correspondence):The School of Pharmacy, The Robert Gordon University,Schoolhill, Aberdeen AB10 1FR, Aberdeen

KeywordsCommunity pharmacistInfluencesOTC drug therapyDecision makingSelection

AbstractA postal survey of pharmacists in a random sample ofcommunity pharmacies in Great Britain was carried out toascertain the factors which influenced their decisions whenrecommending a product for Over The Counter (OTC) sale.Six categories of condition which regularly present incommunity pharmacies were selected and the pharmacistswere asked to state which first line product they wouldrecommend for each condition, in an uncomplicatedsituation with no other relevant problems. Of the 1138questionnaires sent, 635 were suitable for analysis giving aresponse rate of 56.7%. The recommendations were foundto be appropriate in 99.7% of cases. Factors which had thegreatest influence on product selection were activeingredients, evidence of product efficacy, ease of use bypatient and patient characteristics such as age andconcurrent medication. Economic factors such as profit,pressures of excess stock and product promotion by salesrepresentatives were shown to have a significantly greaterinfluence on proprietor pharmacists than the othercategories of pharmacist. The results of this study suggestthat in the majority of cases pharmacists’ decisions on OTCdrug therapy are based on clinical and patient factors andwhilst commercial factors may be involved they do notcompromise the quality of the recommendations.

Accepted December 1999

The questionnaire was posted during the summer of1996 to the selected pharmacies along with anexplanatory covering letter and reply paid envelope.Each questionnaire was numbered to allow a followup questionnaire to be sent to non- responders.

As the questionnaire was sent to pharmacies select-ed on the basis of location and type the variables ofgender and years of experience of the respondentscould not be controlled. To check whether the sampleof pharmacists was representative in these two vari-ables, manpower statistics were obtained from theRoyal Pharmaceutical Society of Great Britain (RPSGB).

To allow comparison of the pharmacists` selectionwith the current recommended management of thesix selected conditions, five readily available referencesources [3 4 5 6 7] were scrutinised and the recom-mendations noted.

The data analysis was carried out using the WorldHealth Organisation Word Processing Database andStatistics Program for Public Health Epi-info v 6 soft-ware package. [13] Chi - squared tests were used anddifferences were considered to be statistically signifi-cant if probability P was less than or equal to 0.05.

ResultsOf the 1138 questionnaires sent, 19 questionnaireswere returned unopened due to a change of addressand 640 replies were received, of which 5 were notcompleted (3 required payment, 1 said it was notcompany policy and 1 ‘did not fill in questionnaires’).This gave 635 questionnaires suitable for analysis, afinal response rate of 56.7%.

Demographic details.The respondents comprised 395 (62.2%) males and240 (37.8%) females. The RPSGB statistics for 1996indicated that 53.9% of community pharmacists weremale and 46.1% were female. The location of thepharmacies was suburban 238 (37.5%), town 212(33.4%), rural 98 (15.4%) and city 82(12.9%), 5 didnot respond. The type of pharmacy was single inde-pendent pharmacy 220 (34.6%), national multiple256 (40.3%) and small multiple (2 - 9 branches) 157(24.7%), 2 did not respond.

The year of registration and gender of the respon-dents is indicated in Table 1 with the pharmacist’sposition in the pharmacy shown in Table 2.

The RPSGB`s statistics gave information on age ofpharmacists therefore a direct comparison with yearof registration and consequently, years of experiencecannot be made. The sample of respondents in thisstudy appears to be more heavily weighted to theyounger age range than the national figures.

Participation in continuing educationThe majority of respondents, 581 (91.5%) said thatthey participated in continuing education, 42 (6.6%)said they did not and 12 (1.9%) did not respond. Onehundred and fifty five (26.7%) respondents spentmore than 30 hours per year, 280 (48.2%) spentbetween 10 and 30 hours and 144 (24.8%) spent lessthan 10 hours per year.

Analysis of the results using a chi squared testshowed that the extent to which pharmacists partici-pated in continuing education had no significantinfluence on their responses (p< 0.05).

OTC selections for six conditions which regularlypresent in community pharmaciesThe therapeutic components and the OTC prepara-tions which were most frequently selected are indicat-ed in Table 3.

Sources of information which influence OTC selectionsJournal articles 514 (80.9%) and reference books 408(64.3%) were indicated as the sources of informationthat were most likely to influence OTC product selec-tion. Television/radio and wholesalers/distributorswere least likely to influence OTC product selection,with 398 (62.7%) responders indicating ‘no’ forTV/radio and 412 (64.9%) indicating ‘no’ for infor-mation from wholesalers/distributors. Neither thegender of the pharmacists, the year of registration northe position the pharmacist held in the pharmacy wasfound to have a significant influence on these results.

Complementary medicines stockedThree respondents indicated that they were qualifiedhomoeopathic practitioners and overall three hun-dred and seventy nine (59.7%) pharmacists indicatedthey stocked complementary medicines. Only onepharmacist selected a complementary medicine as hisfirst line recommendation. This was for the treatment

48

Pharmacy W

orld & ScienceVolum

e 22 Nr. 2 2000

Table 1 Year of registration and gender of respondents (n=624, 11 non responders)

Year of registration Males n (%) Females n (%) Total n (%)

1986 - 1995 92 (23.7) 113 (48.1) 205 (32.9)1976 - 1985 129 (33.2) 71 (30.2) 200 (32.1)1966 - 1975 87 (22.3) 36 (15.3) 123 (19.7)1956 - 1965 68 (17.5) 15 (6.4) 83 (13.3)1946 - 1955 13 (3.3) 0 (0.0) 13 (2.1)

Table 2 Position held in pharmacy. (n= 633, 2 non responders)

Position held Frequency n (%)

Manager 305 (48.0)Proprietor 241 (38.0)Locum tenens pharmacist 35 (5.5)Consultant 27 (4.3)Second pharmacist 9 (1.4)Part time employee pharmacist 8 (1.3)Others 8 (1.3)

of haemorrhoids. Two (0.3%) stated that they wouldrecommend ‘ wrist bands’ for travel sickness.

Two hundred and sixty one respondents (68.9%)said the availability of these products did not influ-ence their product selection. Of these, 99 (37.9%)indicated that they were stocked to offer increasedchoice to meet customer demand. The remaindergave no reasons for stocking them.

Of the 118 respondents who indicated that stock-ing complementary medicines did influence their pre-scribing, 64 (54.2%) said that it allowed them a widerchoice of product, particularly in situations whereallopathic medicines were contraindicated or wereinappropriate. The other 54 respondents gave no rea-sons.

Factors which influence product selectionTable 4 shows the factors which influence productselection, in rank order, ranging from those with thegreatest positive influence on pharmacists’ selectionof OTC products to those with the least influence.

Using a chi squared test, gender and year of registra-tion had no significant influence on pharmacists’responses, however there was a wide variation in theinfluences of certain factors according to the status ofthe pharmacist. Differences in the influences of factorsbetween the groups as determined by chi-squaredtests were either found to be not statistically significant(p<0.05) or highly significant with p<0.001.

Statistically significant variations related to status areindicated in Tables 5 and 6. Level of profit was stated tobe a positive influencing factor by 265 (41.7%) phar-macists, with 157 of these being proprietor pharma-cists. When compared with their other pharmacist col-leagues, proprietor pharmacists were more influencedby the profit gained from the sale (χ2 = 87, p< 0.001)and by the fact that a product is classified as aPharmacy product (P) (χ2 = 59.8, p< 0.001). A majorityof proprietor pharmacists indicated that they would beinfluenced by financial pressures of excess stock, prod-uct promotion by sales representatives and a short sellby date on the product. In the other status groups itwas found that the numbers who agreed or stronglyagreed that these three factors influenced them were inthe minority. Comparison of the results from the pro-prietor pharmacists with those of the other pharmaciststatus groups gave the following results: ‘financial pres-sures of excess stock’ (χ2 = 22.8, p< 0.001), ‘productpromotion by sales representatives’ (χ2 = 22.95, p<0.001), ‘short sell by date’ (χ2 =17.86, p< 0.001).Employee pharmacists appeared to be significantly lessinfluenced by their ‘confidence in the manufacturer’than the other groups (χ2 = 73.4, p< 0.001).

Those pharmacists who indicated they were influ-enced by profit were more likely to select a brandedproduct to treat dry cough than their colleagues whowere not influenced by profit (54.2% compared with31.1%, λ2 = 33.4, p< 0.001). 49

Volume 22 N

r. 2 2000

Pharmacy W

orld & Science

Table 3 The therapeutic components and the OTC preparations most frequently selected by pharmacistsfor six conditions which regularly present in community pharmacy n=635

Condition Therapeutic compo- Total number of Preparation most Preparation secondnents of preparations preparations cited frequently selected most frequentlymost frequently (n,%) selected (n,%)selected (n,%)

Chesty cough guaiphenesin 18 Benylin Meltus (guaiphenesin)(369, 59.0) (guaiphenesin) (99, 15.6)

(176, 27.7)

Dry cough pholcodine 17 pholcodine Benylin(383, 61.3) (372, 58.6) (dextromethorphan)

(73, 11.5)

Haemorrhoids soothing agents only 8 Anusol Anodesyn(507, 81.1) (soothing agents) Soothing agent +

(483, 76.1) local anaesthetic)(56, 8.8)

Hangover analgesic plus antacid 11 Resolve paracetamol(360, 57.6) (Paracetamol + (131, 20.6)

antacid)(315, 49.6)

Sore throat benzocaine plus 15 Merocaine Strepsilsbactericide Benzocaine + (Bactericide)(334, 53.4) bactericide) (90, 14.2)

(178, 28.0)

Travel sickness cinnarizine 9 Stugeron Kwells(467, 74.7) (Cinnarizine) (Hyoscine)

(462, 72.8) (59, 9.3)

Additional factorsOne hundred and thirty two (20.8%) of pharmacistsgave information about additional factors which theyconsidered affected their choices. The comments madeby respondents were very generalised and could becategorised into two groups, namely patient orientated(105, 79.5%) and business orientated (27, 20.5%). Nospecific additional influences were identified.

Recommendations indicated in reference sourcesFour of the selected reference sources gave recom-mendations for the management of five of the chosenconditions, the exception being hangover. The refer-ence source ‘Non- Prescription Drugs’ [7] was theonly one of the five which gave recommendations forthe management of this condition.

Comparison of pharmacists selections with current recommendationsThe products chosen by pharmacists were appropriateand followed the published recommendations in all but11 of the 3810 (0.3%) cases. These were as follows:

nine pharmacists recommended products for thetreatment of chesty cough which are consideredinappropriate. Six selections were products which

included cough suppressants (codeine, pholcodine ormorphine) which are contraindicated in the treat-ment of productive or chesty cough [4] and threewere demulcents which are more appropriate for drycough.

One pharmacist recommended bismuth subsalicy-late for the treatment of hangover and one pharma-cist indicated Triludan tablets (terfenadine) as his firstline treatment for travel sickness. Neither of thesechoices are supported by current literature.

Of the 11 pharmacists who selected inappropriatetreatments 3 stated they were influenced by profit.

DiscussionIt could be argued that when completing the ques-tionnaire used in this study community pharmacistsmay have been likely to select as positive influencingfactors those which could be perceived as being ‘correct’. Validation of the OTC recommendations bycovert or semi-covert study [14] could be carried outfor a sample of pharmacists but it is not practical oreven possible to validate the influence of the 27 fac-tors, as they are not easily quantified. However thewide range of stated opinions on the influence of

50

Pharmacy W

orld & ScienceVolum

e 22 Nr. 2 2000

Table 4 The influencing factors in rank order

Factor Sum Mean Std devofscores

Score applied 2 1 0 -1 -2

Active ingredients 471 158 1 1 0 4 1099 1.73 0.47Evidence of product efficacy 416 196 15 3 0 5 1025 1.61 0.57Patient characteristics e.g. age,

other medication 374 228 25 5 2 1 967 1.52 0.65Product formulation 309 265 41 13 0 7 870 1.37 0.71Ease of use and convenience to patient 265 340 19 6 0 5 864 1.36 0.60Presence of side effects with product 284 290 41 10 5 5 838 1.32 0.74Successful self-use of product 260 296 48 22 6 3 782 1.23 0.81Customer feedback 170 403 46 11 1 4 730 1.15 0.64Product abuse potential 237 274 71 30 15 8 688 1.08 0.95Pharmacy only (P) product 232 207 94 86 13 3 559 0.88 1.11Cost of product for patients 101 364 103 53 11 3 491 0.77 0.87Time available to question 74 314 103 104 36 4 286 0.45 1.07Confidence in manufacturer 65 299 145 94 25 7 285 0.45 0.99Range of dosage forms 54 287 171 111 5 7 274 0.43 0.90Customers recognise brand 34 325 134 103 36 3 218 0.34 1.00GP recommendation 37 254 169 128 37 10 126 0.20 1.01Prescribe product by habit 51 248 110 177 39 10 95 0.15 1.10Own brand 61 232 125 145 63 9 83 0.13 1.17Profit from product 63 202 99 169 98 4 - 37 - 0.06 1.26Social status of customer 39 164 130 203 92 7 - 145 - 0.23 1.16Short sell by date 34 125 171 186 114 5 - 221 - 0.35 1.14Influenced by colleagues 4 138 150 256 84 3 - 278 - 0.44 0.99Display materials 15 112 159 232 111 6 - 312 - 0.49 1.05Financial pressure of excess stock 23 122 119 208 159 4 - 358 - 0.56 1.16Product promotion by sales reps 13 120 135 206 155 6 - 370 - 0.58 1.11Attractive packaging 9 83 130 232 176 5 - 483 - 0.76 1.04Under company/employer’s instructions 10 45 44 182 329 25 - 775 - 1.22 1.00

Ston

gly

agre

e

Agr

ee

Unc

erta

in

Dis

agre

e

Stro

ngly

Dis

agre

e

No

resp

onse

those factors, apparently less professionally accept-able than others, suggests a high level of honesty bythe respondents. This can be further substantiated byexamining the recommendations made by pharma-cists. Six hundred and twenty nine (99.1%) pharma-cists agreed or strongly agreed that active ingredientsinfluenced their product selection. When the prod-ucts selected were examined there were only 11(0.3%) inappropriate recommendations made out ofa possible 3810. No pharmacist made more than oneinappropriate selection with 624 (98.3%) respon-dents providing products which followed currentguidelines in all of the six conditions and 100% wereproviding appropriate therapy in five of the condi-tions.

The number of products cited for each conditionranged from 8 to 18 but when the active ingredientsof each product was examined there was a clear con-sistency in the selection of active ingredient, if not theproduct. This may be due to a number of reasons butcould be taken as evidence that pharmacists’ recom-mendations are made on the basis of efficacy. This isfurther supported by the finding that factors such asefficacy and incidence of side effects, which areimportant considerations when developing a formu-lary, also rank highly in the positive influences. Thisconcurs with the findings of a previous study [10]which showed that clinical influences are the mostimportant factors when selecting the active ingre-dients for treatment of headache.

Patient characteristics also featured as positiveinfluences along with product formulation and easeof use, indicating that pharmacists are consideringthe patient as a whole and making an appropriateselection based on these facts. The degree of impor-tance accorded to customer feedback and self use isinteresting in the light of the current emphasis on evi-dence based medicine. The professional integrity and

independence of pharmacists is emphasised by thenegative influence exerted by company or employers’instructions and promotional activities of sales repre-sentatives. The fact that colleagues’ influence ranks solow may be due to the fact that pharmacists frequent-ly work in isolation and make independent judge-ments. The positive influence exerted by the P classifi-cation of a product is complex. It could be postulatedthat P products are more likely to contain more phar-macologically efficacious drugs than productsallowed to be sold from other outlets and this maycontribute to the positive influence but the financialimplications of recommending pharmacy restrictedmedicines may also be important.

The position the pharmacist held in the pharmacydid have an influence on product selection with themost significant variation being in the area of profitfrom the product. The majority of proprietor pharma-cists indicated that profit was an influencing factorwhile only a minority of the other status groups wereinfluenced by profit. This indicates that economic andcommercial factors are a consideration in certainmembers of the profession. This can be illustrated bythe fact that pharmacists influenced by profit wereless likely to recommend the unbranded pholcodinefor the treatment of dry cough and more likely to rec-ommend branded products which are more expen-sive and will therefore yield more profit. Howeverfrom the findings of this study although the profitmotive does influence some pharmacists, particularlyin the proprietor group, it does not appear to com-promise the appropriateness of the treatment beingrecommended.

The sample of respondents was not totally repre-sentative of the profession as a whole in relation togender and age. However as neither gender nor yearsof experience were shown to influence product selec-tion this may not have biased the results.

51

Volume 22 N

r. 2 2000

Pharmacy W

orld & Science

Table 5 The relationship between status of pharmacist1 and positive influences of factors on OTC recommendations

Factors where respondents Proprietors Managers Employee2 Locumsagreed or strongly agreed n =241 (%) n = 305 (%) pharmacists n = 35 (%)

n = 44 (%)

Pharmacy only product 211 (87.5) 175 (57.1) 27 (61.3) 22 (62.9)Confidence in the manufacturer 168 (69.7) 158 (51.8) 12 (27.3) 21 (60.0)Profit 157 (65.2) 90 (29.5) 7 (15.9) 8 (22.8)

1 ‘other’ category of pharmacist is excluded from this table n=82 employee pharmacists include consultants (27),second pharmacists (9) and part time employees (8)

Table 6 The relationship between status of pharmacist and negative influences of factors on OTC recommendations.1

Factors where respondents Proprietors Managers Employee2 Locumsdisagreed or strongly disagreed n = 241 (%) n = 305 (%) pharmacists n = 35 (%)

n = 44 (%)

Financial pressures of excess stock 111 (46.0) 197 (64.6) 36 (81.8) 18 (51.5)Product promotion by sales reps 108 (44.8) 193 (63.3) 30 (68.2) 24 (68.5)Short sell by date 88 (36.5) 161 (52.8) 26 (59.1) 20 (57.1)

1 ‘other’ category of pharmacist is excluded from this table n=82 employee pharmacists include consultants (27),second pharmacists (9) and part time employees (8)

The commercial and professional activities of commu-nity pharmacists have long been perceived, in somequarters, as being in conflict [9]. This study has gonesome way to addressing some of these arguments.From the findings it would appear that clinical, pro-fessional decisions are foremost with the commercialaspects being given low priority. This, coupled withthe findings of a previous study [15], where custom-ers expressed satisfaction with the advice and servicethey received from pharmacies suggests that mostpharmacists are providing appropriate advice, basedprimarily on clinical decisions.

AcknowledgementsThe authors would like to thank Miss Maeve O’Brienwho carried out the pilot work for this study, MissFiona Robertson who created the database of the ran-dom sample of community pharmacies, Mrs MoiraInnes for clerical support and inputting the data andDr Denise Hansford for invaluable support, help andadvice. Thanks are also due to the pharmacists whocompleted the questionnaire.

References1 Matheson C, Bond C.M. The community pharmacist’s devel-

oping OTC role : Implications for primary health care. Reportto Grampian Health Board. March 1995.

2 Medicines, Ethics and Practice. A Guide for Pharmacists.Number 22. The Royal Pharmaceutical Society of GreatBritain, July 1999.

3 British National Formulary.London: British MedicalAssociation and the Royal Pharmaceutical Society of GreatBritain. March 1996.

4 Blenkinsopp A and Paxton P 1995 Symptoms in thePharmacy, A Guide to the Management of Common Illness,2nd edn. Blackwell Science, Oxford.

5 Harman RJ ed 1990 Handbook of Pharmacy Health- Care,Diseases and Patient Advice The Pharmaceutical Press,London.

6 Edwards C and Stillman P 1995 Minor Illness or MajorDisease, Responding to Symptoms in the Pharmacy 2nd edn.The Pharmaceutical Press, London.

7 Li Wan Po A 1990 Non- Prescription Drugs 2nd edn. BlackwellScience ,Oxford.

8 Emmerton L, Benrimoj SI. Factors influencing pharmacists’preferences for non-prescription cough suppressants. J SocAdmin Pharm. 1994;11:78-85.

9 Chappell NL and Barnes GE. Professional and business roleorientations among practicing pharmacists. Soc Sci Med1984;18(2):103-10.

10 Roins S et al. Factors used by pharmacists in the recommen-dation of the active ingredient(s) and brand of non-prescrip-tion analgeics for a simple, tension and migraine headacheInt. J. Pharm. Pract. 1998;6:196-206.

11 Likert R. A technique for the measurement of attitude.Archives of Psychology 1932;140:1-55.

12 The Annual Register of Pharmaceutical Chemists 1996, TheRoyal Pharmaceutical Society of Great Britain, London.

13 Epi-info produced by The Division of Surveillance andEpidemiology, Epidemiology Program Office, Centers forDisease Control and Prevention (CDC), Atlanta, Georgia3033 http://www.cdc.gov/epo/epi/software.htm October1999.

14 Krska J, Kennedy EJ. An audit of responding to symptoms incommunity pharmacy. Int J Pharm Prac 1996;4:129-35.

15 Krska J and Kennedy EJ. Expectations and experiences of cus-tomers purchasing over-the-counter medicines in pharma-cies in the north of Scotland. Pharm J 1996;256:354-6.

52

Pharmacy W

orld & ScienceVolum

e 22 Nr. 2 2000