Cross-validation of an algorithm detecting acute gastroenteritis episodes from prescribed drug dispensing data in France: comparison with clinical data reported in a primary care surveillance system, winter seasons 2014/15 to 2016/17

Abstract

Background This study compares an algorithm to detect acute gastroenteritis (AG) episodes from drug dispensing data to the validated data reported in a primary care surveillance system in France. Methods We used drug dispensing data collected in a drugstore database and data collected by primary care physicians involved in a French surveillance network, from season 2014/15 to 2016/17. We used an adapted version of an AG discrimination algorithm to identify AG episodes from the drugstore database. We used Pearson’s correlation coefficient to evaluate the agreement between weekly AG signals obtained from the two data sources during winter months, in the overall population, by specific age-groups and by regions. Results Correlations between AG signals for all ages were 0.84 [95%CI 0.69; 0.92] for season 2014/15, 0.87 [95%CI 0.75; 0.93] for season 2015/16 and 0.94 [95%CI 0.88; 0.97] for season 2016/17. The association between AG signals estimated from two data sources varied significantly across age groups in season 2016/17 ( p -value < 0.01), and across regions in all three seasons studied ( p -value < 0.01). Conclusions There is a strong agreement between the dynamic of AG activity estimated from drug dispensing data and from validated primary care surveillance data collected during winter months in the overall population but the agreement is poorer in several age groups and in several regions. Once automated, the reuse of drug dispensing data, already collected for reimbursement purposes, could be a cost-efficient method to monitor AG activity at the national level. Electronic supplementary material The online version of this article (10.1186/s12874-019-0745-5) contains supplementary material, which is available to authorized users.

Association Between Acute Gastroenteritis and Continuous Use of Proton Pump Inhibitors During Winter Periods of Highest Circulation of Enteric Viruses

“…The detection of AGE episodes from prescribed drug dispensing data was done by using an algorithm that identifies drug dispensing events corresponding to AGE treatments, taking into account patient characteristics, the delay between the prescription and drug dispensing event, and the type, number, and quantity of drugs dispensed. 29,30 The types of drugs potentially prescribed to treat an AGE episode included oral rehydration salts, antiemetics, probiotic antidiarrheals, intestinal antipropulsives, intestinal absorbents, intestinal anti-infectious agents, and antispasmodic agents. Coprescriptions of these drugs and antibiotics, antineoplastic agents used in cancer therapy, gastric antacids, drugs used for inflammatory bowel diseases, or antiemetics in injectable form were considered to treat a disease other than AGE.…”

“…The agreement was strong in all age subgroups; the highest correlation was observed among adults aged 15 to 64 years ( r ≥ 0.87) and the lowest among elderly patients ( r ≥ 0.75). 30…”

Further Evidence to Monitor Long-term Proton Pump Inhibitor Use

“…Incident AGE was defined using a drug dispensing algorithm previously found to correlate with an indicator for AGE in French general practitioner records. 6 Vilcu and colleagues 5 found a significant association between PPI use and AGE (relative risk, 1.81; 95% CI, 1.72-1.90), translating to a number needed to harm of 153 patients. Of note, the authors also found a significant association between AGE and histamine 2 receptor antagonists in a sensitivity analysis (adjusted relative risk, 2.08; 95% CI, 1.27-3.39).…”

Syndromic Surveillance of Acute Gastroenteritis Using the French Health Insurance Database: Discriminatory Algorithm and Drug Prescription Practices Evaluations

“…This difference reduces the specificity (PPV) in comparison of our algorithm. The authors have found a strong agreement in the dynamic of mAGE activity in winter, between the estimate derived from the LTD data and the reporting of a primary care surveillance system (GP Sentinelles network) with a correlation coefficient between 0.84 and 0.94 on the mAGE weekly rate [19]. The work of Vilcu et al covers an estimation of regional or national mAGE rate during winter seasons with an exclusion of overseas departments and is not suitable for a local mAGE surveillance.…”