AD 0 N I S 0306525191000382
Br. J. clin. Pharmac. (1991), 31, 188-189
Geographical differences in adverse drug reaction reporting rates in the Northern Region D. N. BATEMAN, A. LEE, M. D. RAWLINS & J. M. SMITH Northern Regional Adverse Drug Reaction Monitoring Centre, Wolfson Unit, The University, Newcastle upon Tyne, NE1 7RU
The reporting rates of adverse drug reactions in the Northern Region have been examined with respect to source (hospital or general practice) and district of origin for the 3 years 1986-88. Mean annual reporting rates, corrected for population, or clinical activity, varied approximately 4-fold in general practice (0.066 to 0.257 reports per 1000 resident population) and more than 12-fold in hospitals (0.14 to 1.77 reports per 1000 deaths and discharges). There was no correlation between hospital and general practice reporting rates in the same health district. Overall more reports were received from general practice (1606) than hospitals (1000), although proportionally more reports of serious adverse reactions originated from hospitals. The reasons for these differences are unclear. Keywords adverse drug reactions drug monitoring adverse drug reaction reporting pharmaco-epidemiology
Introduction
Methods and results
The United Kingdom Committee on Safety of Medicines' adverse reaction reporting scheme (the 'yellow card' scheme) has now been operating for 25 years. It is widely recognised as an important surveillance scheme for identifying and assessing drug toxicity. The system depends on the spontaneous reporting of suspected adverse reactions by medical and dental practitioners, either directly to the Committee, or indirectly through drug manufacturers. Reporting rates, however, are rarely more than 10%, and sometimes substantially less (Rawlins, 1988a). In order to increase reporting, the CSM has actively promoted the adverse reaction reporting scheme and, in recent years, yellow card reporting forms have been included in general practitioners' prescription pads (FP10), the British National Formulary, the ABPI Data Sheet Compendium and MIMS (Griffin & Weber, 1985; Rawlins, 1988a). A further attempt to improve the reporting of adverse drug reactions has been the introduction of regional monitoring centres (RMCs) in the Northern and West Midlands (NHS) Regions and in Wales; the Northern CSM monitoring centre was established in 1981. The objective of the present study was to examine differences between Health Districts in adverse reaction reporting.
Details of the clinical base (general practice or hospital) of the reporting doctor and the Health District of origin of all reports received by the Northern Regional Monitoring Centre (RMC) during 1986-88 were recorded. Our analysis has been limited to these years because we know that for this period 80% (2606 of 3257) of reports from the Region were made through the RMC, whereas the figure in earlier years was lower. We have not analysed reports sent directly to the CSM in London because of technical difficulties in obtaining these. Marked differences were observed in the absolute numbers of reports received from individual Health Districts (Table 1). These districts are, however, of different sizes and, in order to standardise for clinical activity, we have expressed the numbers of reports in relation to the size of the population (general practice reports) or for the annual number of deaths and discharges (hospital reports). When assessed in this way, hospital reporting rates vary more than 12-fold, from 0.14 (95% CI ± 0.1) to 1.77 (± 0.17) per 1000 deaths and discharges per annum (Figure 1). District 'P', the highest reporting district contains the Region's teaching hospitals. General practitioner reporting rates vary almost 4-fold, between 0.07 (± 0.03) and 0.26 (± 0.03) per 1000 resident
Correspondence: Dr D. N. Bateman, Northern Regional Adverse Drug Reaction Monitoring Centre, Wolfson Unit, Thd University, Newcastle upon Tyne NE1 7RU
188
Short report Table 1 Reporting of adverse drug reactions 1986-88 (number of reports) by general practitioners and hospital doctors in the 16 health districts of the Northern Region (ranked by hospital reporting rate adjusted for clinical activity) District
A B C D E F G H I J K L M N 0 P Regional total
Total reports General practice
Hospital
18 97 150 232 163 95 84 38 117 77 101 60 69 44 46 215 1606
8 17 61 36 71 32 22 27 32 45 39 32 46 40 41 451 1000
a 18f---
1.6 -
0
. 0.0
B
0.30 0.250.20
A B C D E F G H I J K LMN O P
District
Figure 1 a) Mean hospital reporting rate 1986-88, per 1000 deaths and discharges per annum (based on 1986 clinical activity data). b) Mean general practice reporting rate 198688, per 1000 resident population (based on 1988 population data).
189
population per annum. The variations in both general practitioner and hospital reporting rates, are statistically significant using the chi-squared test of heterogenicity (P < 0.001). Overall, more reports were received from general practitioners (1606) than hospitals (1000). In 1988 only 19% of general practitioner reports were of serious reactions (as defined by the CSM adverse reaction dictionary) but 57% of reports from hospitals were of serious reactions.
Discussion
The overall rate of reporting of adverse drug reactions in the Northern Region is currently higher than that of the UK as a whole. Thus, in 1988 the national reporting rate of adverse reactions was 40.8 per million GP prescriptions, and in the Northern Region was 45.5 reports per million prescriptions. When examined in different Health Districts both the absolute numbers of reports, and rates adjusted for clinical activity, vary considerably. Moreover, there was no correlation between reporting rates by general practitioners and by hospital doctors within Health Districts (Spearman's r = -0.169). The variation appears to be greater in hospitals than in general practice, but this may reflect a relatively high rate of reporting from the teaching hospitals in Newcastle (1.77 per 1000 deaths and discharges per annum). It is salutary to note that even this figure, which includes nearly 50% non serious reports, is likely to be a substantial underestimate of the true incidence of adverse reactions occurring in hospitals. Thus, in a prospective study conducted in surgical wards, the Boston Collaborative Drug Surveillance Programme found 97 major life-threatening adverse reactions in 5232 patients (18.5 per 1000) (Danielson etal., 1982). Although a proportion (20% ) of adverse reaction reports is known to bypass the Northern RMC this would not have any significant effect on our conclusions. Neither could differences in GP prescribing rates, which range from 7.3 to 9.07 prescriptions per patient per annum, have any real effect on adverse reaction reporting rates for the FPCs in the Region. In conclusion, this study demonstrates considerable variation in the reporting rates of adverse drug reactions within the Northern Region, and the reasons for these differences are unclear. It is obviously important to establish factors underlying these differences and this is to be the subject of further investigations.
References Danielson, D. A., Porter, J. B., Dinan, B. J., O'Connor, P. C., Lawson, D. H., Kellaway, G. S. M. & Jick, H. (1982). Drug monitoring of surgical patients. J. Am. med. Ass., 284, 1482-1485. Griffin, J. P. & Weber, J. C. P. (1985). Voluntary systems of adverse reaction reporting-Part I. Adv. Drug React. Ac. Pois. Rev., 4, 213-230.
Rawlins, M. D. (1988a). Spontaneous reporting of adverse drug reactions I: The data. Br. J. clin. Pharmac., 26, 1-5. Rawlins, M. D. (1988b). Spontaneous reporting of adverse drug reactions II: Uses. Br. J. clin. Pharmac., 26, 7-11.
(Received 30 April 1990, accepted 17 September 1990)
Br. J. clin. Pharmac. (1991), 31, 188-189
Geographical differences in adverse drug reaction reporting rates in the Northern Region D. N. BATEMAN, A. LEE, M. D. RAWLINS & J. M. SMITH Northern Regional Adverse Drug Reaction Monitoring Centre, Wolfson Unit, The University, Newcastle upon Tyne, NE1 7RU
The reporting rates of adverse drug reactions in the Northern Region have been examined with respect to source (hospital or general practice) and district of origin for the 3 years 1986-88. Mean annual reporting rates, corrected for population, or clinical activity, varied approximately 4-fold in general practice (0.066 to 0.257 reports per 1000 resident population) and more than 12-fold in hospitals (0.14 to 1.77 reports per 1000 deaths and discharges). There was no correlation between hospital and general practice reporting rates in the same health district. Overall more reports were received from general practice (1606) than hospitals (1000), although proportionally more reports of serious adverse reactions originated from hospitals. The reasons for these differences are unclear. Keywords adverse drug reactions drug monitoring adverse drug reaction reporting pharmaco-epidemiology
Introduction
Methods and results
The United Kingdom Committee on Safety of Medicines' adverse reaction reporting scheme (the 'yellow card' scheme) has now been operating for 25 years. It is widely recognised as an important surveillance scheme for identifying and assessing drug toxicity. The system depends on the spontaneous reporting of suspected adverse reactions by medical and dental practitioners, either directly to the Committee, or indirectly through drug manufacturers. Reporting rates, however, are rarely more than 10%, and sometimes substantially less (Rawlins, 1988a). In order to increase reporting, the CSM has actively promoted the adverse reaction reporting scheme and, in recent years, yellow card reporting forms have been included in general practitioners' prescription pads (FP10), the British National Formulary, the ABPI Data Sheet Compendium and MIMS (Griffin & Weber, 1985; Rawlins, 1988a). A further attempt to improve the reporting of adverse drug reactions has been the introduction of regional monitoring centres (RMCs) in the Northern and West Midlands (NHS) Regions and in Wales; the Northern CSM monitoring centre was established in 1981. The objective of the present study was to examine differences between Health Districts in adverse reaction reporting.
Details of the clinical base (general practice or hospital) of the reporting doctor and the Health District of origin of all reports received by the Northern Regional Monitoring Centre (RMC) during 1986-88 were recorded. Our analysis has been limited to these years because we know that for this period 80% (2606 of 3257) of reports from the Region were made through the RMC, whereas the figure in earlier years was lower. We have not analysed reports sent directly to the CSM in London because of technical difficulties in obtaining these. Marked differences were observed in the absolute numbers of reports received from individual Health Districts (Table 1). These districts are, however, of different sizes and, in order to standardise for clinical activity, we have expressed the numbers of reports in relation to the size of the population (general practice reports) or for the annual number of deaths and discharges (hospital reports). When assessed in this way, hospital reporting rates vary more than 12-fold, from 0.14 (95% CI ± 0.1) to 1.77 (± 0.17) per 1000 deaths and discharges per annum (Figure 1). District 'P', the highest reporting district contains the Region's teaching hospitals. General practitioner reporting rates vary almost 4-fold, between 0.07 (± 0.03) and 0.26 (± 0.03) per 1000 resident
Correspondence: Dr D. N. Bateman, Northern Regional Adverse Drug Reaction Monitoring Centre, Wolfson Unit, Thd University, Newcastle upon Tyne NE1 7RU
188
Short report Table 1 Reporting of adverse drug reactions 1986-88 (number of reports) by general practitioners and hospital doctors in the 16 health districts of the Northern Region (ranked by hospital reporting rate adjusted for clinical activity) District
A B C D E F G H I J K L M N 0 P Regional total
Total reports General practice
Hospital
18 97 150 232 163 95 84 38 117 77 101 60 69 44 46 215 1606
8 17 61 36 71 32 22 27 32 45 39 32 46 40 41 451 1000
a 18f---
1.6 -
0
. 0.0
B
0.30 0.250.20
A B C D E F G H I J K LMN O P
District
Figure 1 a) Mean hospital reporting rate 1986-88, per 1000 deaths and discharges per annum (based on 1986 clinical activity data). b) Mean general practice reporting rate 198688, per 1000 resident population (based on 1988 population data).
189
population per annum. The variations in both general practitioner and hospital reporting rates, are statistically significant using the chi-squared test of heterogenicity (P < 0.001). Overall, more reports were received from general practitioners (1606) than hospitals (1000). In 1988 only 19% of general practitioner reports were of serious reactions (as defined by the CSM adverse reaction dictionary) but 57% of reports from hospitals were of serious reactions.
Discussion
The overall rate of reporting of adverse drug reactions in the Northern Region is currently higher than that of the UK as a whole. Thus, in 1988 the national reporting rate of adverse reactions was 40.8 per million GP prescriptions, and in the Northern Region was 45.5 reports per million prescriptions. When examined in different Health Districts both the absolute numbers of reports, and rates adjusted for clinical activity, vary considerably. Moreover, there was no correlation between reporting rates by general practitioners and by hospital doctors within Health Districts (Spearman's r = -0.169). The variation appears to be greater in hospitals than in general practice, but this may reflect a relatively high rate of reporting from the teaching hospitals in Newcastle (1.77 per 1000 deaths and discharges per annum). It is salutary to note that even this figure, which includes nearly 50% non serious reports, is likely to be a substantial underestimate of the true incidence of adverse reactions occurring in hospitals. Thus, in a prospective study conducted in surgical wards, the Boston Collaborative Drug Surveillance Programme found 97 major life-threatening adverse reactions in 5232 patients (18.5 per 1000) (Danielson etal., 1982). Although a proportion (20% ) of adverse reaction reports is known to bypass the Northern RMC this would not have any significant effect on our conclusions. Neither could differences in GP prescribing rates, which range from 7.3 to 9.07 prescriptions per patient per annum, have any real effect on adverse reaction reporting rates for the FPCs in the Region. In conclusion, this study demonstrates considerable variation in the reporting rates of adverse drug reactions within the Northern Region, and the reasons for these differences are unclear. It is obviously important to establish factors underlying these differences and this is to be the subject of further investigations.
References Danielson, D. A., Porter, J. B., Dinan, B. J., O'Connor, P. C., Lawson, D. H., Kellaway, G. S. M. & Jick, H. (1982). Drug monitoring of surgical patients. J. Am. med. Ass., 284, 1482-1485. Griffin, J. P. & Weber, J. C. P. (1985). Voluntary systems of adverse reaction reporting-Part I. Adv. Drug React. Ac. Pois. Rev., 4, 213-230.
Rawlins, M. D. (1988a). Spontaneous reporting of adverse drug reactions I: The data. Br. J. clin. Pharmac., 26, 1-5. Rawlins, M. D. (1988b). Spontaneous reporting of adverse drug reactions II: Uses. Br. J. clin. Pharmac., 26, 7-11.
(Received 30 April 1990, accepted 17 September 1990)
