Statistical Recognition of Trends in Health Monitoring Systems

The determination of when trends are present in an active health monitoring system is considered. The type of data collected is often voluntary response data usually of unknown, or perhaps of low quality or reliability. Often, even if the data themselves are perfectly reliable, the different monitoring stations are usually not comparable in size or scope, so aggregrate measures would tend to mask rather than detect trends for the whole system.

Examples of such monitoring systems are the World Health Organization’s Research Center for International Monitoring of Adverse Reactions to Drugs and the »Programa de Investigacion de Modelos Operacionales de Prestacion de Servicios de Salud« (PRIMOPS) operating in Cali, Columbia. We study a »Center-Batch matrix« by using a transformation to a matrix of ranks. It incorporates most of the relevant information. A relatively simple statistical technique is presented for generating a warning signal whenever a pattern of increasing adverse events does occur. This rank Center-Batch method avoids some of the pitfalls of the previous methods used and in fact is often quite superior.

Es wird die Möglichkeit erörtert festzustellen, wann in einem aktiven Gesundheitsüberwachungssystem Trends vorhanden sind. Bei den gesammelten Daten handelt es sich häufig um Angaben, die freiwillig gemacht wurden und deren Qualität und Zuverlässigkeit gewöhnlich nicht bekannt oder vielleicht unbefriedigend ist. Selbst wenn die Daten vollkommen zuverlässig sind, sind die verschiedenen Über-wachungsstelleii in der Regel in Größe und Umfang nicht vergleichbar, so daß aggregierte Maßzahlen eher dazu neigen, Trends für das ganze System zu verschleiern als aufzudecken.

Beispiele für solche Uberwachungssysteme sind das World Health Organization’s Research Center für International Monitoring of Adverse Reactions to Drugs und das Programa de Investigacion de Modelos Operacionales de Servicios de Salud (PRIMOPS) in Cali, Columbia.

Wir untersuchen eine »Center-Batch Matrix«, indem wir eine Transformation in eine Rangmatrix benutzen. Sie enthält den größten Teil der relevanten Information. Eine relativ einfache statistische Technik wird präsentiert, mit deren Hilfe ein Warnsignal erzeugt werden kann, wenn immer ein Muster zunehmender Nebenwirkungen auftritt. Diese Rang-Center-Batch-Methode vermeidet einige der Fallen früher angewandter Methoden und ist diesen in der Tat oft überlegen.

Key-Words:

Drug Monitoring - Adverse Reactions - Trend Signalling - Center-Batch Matrix - Non-parametric Analysis - Rank Tests

Schlüssel-Wörter:

Arzneimittelüberwachung - Nebenwirkungen - Trendanzeige - Center-Batch Matrix - niohtparametrisohe Analyse - Rangtests

• References

• 1 Aiyae R. T, Guillien C. L, Albers W. Asymptotic relative efficiencies of rank tests for trend alternatives. J. Amer. statist. Ass 1979; 74: 226-231.
• 2 Cox D. R, Stuart A. Some quick sign tests for trend in location and dispersion. Biometrika 1955; 42: 80-95.
• 3 Finney D. J. Statistical aspects of monitoring for dangers in drug therapy. Meth. Inform. Med 1971; 10: 1-8.
• 4 Finney D. J. Statistical logic in the monitoring of reactions to therapeutic drugs. Meth. Inform. Med 1971; 10: 237-245.
• 5 Finney D. J. The detection of causation of adverse events. Paper delivered to the 39th session of the International Institute, Vienna 1973
• 6 Finney D. J. Systematic signalling of adverse reactions to drugs. Meth. Inform. Med 1974; 13: 1-10.
• 7 Finney D. J. Problems, data and inference. J. roy. statist. Soc. A 1974; 137: 1-23.
• 8 Finney D. J. Personal communication.
• 9 Glasses C. J, Winter R. Critical values of the coefficient of rank correlation for testing the hypothesis of independence. Biometrika 1961; 48: 444-448.
• 10 Hajek J, Sidak Z. Theory of Rank Tests. New York: Academic Press; 1967
• 11 Kendall M. G, Stuart A. The Advanced Theory of Statistics. Vol. 2. 3rd edition. New York: Hafner Press; 1972
• 12 Kendall M. G, Stuart A. The Advanced Theory of Statistics. Vol. 3. 3rd edition. New York: Hafner Press; 1976
• 13 Lai T. L. Control charts based on weighted sums. Ann. Statist 1974; 2: 134-147.
• 14 Lehmann E. L. Nonparametrics : Statistical Methods Based on Ranks. San Francisco: Holden Day; 1975
• 15 Levine A, Mandel S. P. H, Santamaria A. Pattern signalling in health information monitoring systems. Meth. Inform. Med 1977; 16: 138-144.
• 16 Mandel S.P.H. Report on Statistical Monitoring of Adverse Reactions to Drugs in the Research Center for International Monitoring of Adverse Reactions to Drugs. Report prepared for WHO, DEM/74.7. Geneva: 1974
• 17 Mandel S. P. H, Levtne A, Beleno G. E. Signalling increases in reporting in international monitoring of adverse reactions to therapeutic drugs. Meth. Inform. Med 1076; 15: 1-10.
• 18 Mann H. B. Non-parametric tests against trend. Econo- metrica 1945; 13: 245-259.
• 19 Owen D. B. Handbook of Statistical Tables. Reading/Mass.: Addison-Wesley; 1962
• 20 Page E. S. Control charts with warning lines. Biometrika 1955; 42: 243-257.
• 21 Patwaby K. M. Report on Statistical Aspects of the Pilot Research Project for International Drug Monitoring. Report prepared for WHO. Geneva: 1969
• 22 Royall B. W. International aspects of the study of adverse reactions to drugs. Biometrics 1971; 27: 689-698.
• 23 Royall B. W, Venulet J. Methodology for international drug monitoring. Meth. Inform. Med 1972; 11: 75-86.
• 24 Stuart A. The efficiencies of tests of randomness against normal regression. J. Amer. statist. Ass 1956; 51: 285-287.
• 25 Stuart A. Asymptotic relative efficiencies of distribution-free tests of randomness against normal alternatives. J. Amer. statist. Ass 1954; 49: 147-157.
• 26 Venttlet J. Adverse reactions to drugs. Int. J. clin. Pharmacol 1973; 7: 253-264.
• 27 Zar J. H. Significance testing of the Spearman rank correlation coefficient. J. Amer, statist. Ass 1972; 67: 578-580.