The economic burden of urinary tract infections in women visiting general practices in France: a cross-sectional survey

This initial study conducted in France on the costs of community UTIs in women over 18 years of age estimated a total cost from the societal perspective of €58 million in 2012, €44 million for direct costs and €14 million for indirect costs. Half of this cost was supported by the French National Health Insurance, and half was supported by the patients (before private complementary health insurance participation). Visits represented the largest expense item followed by sick leave and prescription drugs. Although very expensive, hospitalizations were rare and therefore represented the smallest expense item. The costs for 75 % of the women were below the mean cost. Additional visits with specialist physicians, ultrasounds, hospitalizations and sick leave concerned less than one quarter of the women. The important costs of these additional health care procedures clearly increased the mean UTI cost.

An important strength of this study was the use of a sampling design. This allowed to correct the bias due to drop-outs and geographical repartition and to generalize with caution the cost of UTIs to the population of women over 18 years of age who visit GPs for presumed UTIs. Furthermore, to estimate the possible costs according to expense items as accurately as possible, the maximum amounts of data from the French National Health Insurance and the French National Institute of Statistics and Economic Studies were used. The use of the EGB allowed for the accounting of possible differences in care consumption according to gender and excess fees according to medical specialties and French departments, particularly in terms of the costs of physician visits [37]. The systematic collection of urine sample from all women who visited their GPs for suspected UTIs permitted the distinction between clinical UTIs with positive urine cultures from clinical UTIs with negative urine cultures and the estimation of their related costs. Furthermore, to estimate the real cost, the women’ declarations were preferred to the GPs’ declaration to account only for prescriptions that were actually utilized or purchased.

This study has some limitations that might have resulted in the over- or underestimation of the costs of UTIs. First, the cost generalization should be interpreted with caution: women suffering from UTIs without dysuria, frequent or urgent of urination were not included, which could decrease the estimated cost of UTIs; patients included had more symptoms than eligible patients, which could influence the achievement of diagnostic test, especially lower back pain with suspected pyelonephritis; the socio economic status was not available for eligible women, preventing to compare eligible and included patients on this point. However, in our study, there were more workers and less non-economically active persons than in the general population. This is concordant with the French inequalities of health care recourse: unemployed and retired persons seek less care than economically active persons [38]. Second, the costs calculated in this study were for women who visited GPs and not for the general population. The estimation of the costs of UTIs among the general population would have required the estimations of the costs of women who visited hospital emergencies departments and specialists (i.e., urologists and gynecologists). Third, the data used to estimate the non-medical direct costs (i.e., time lost and monetary expenses), intangible costs (i.e., loss of well-being for the patient and her close family and friends) and presenteeism costs (i.e., the loss of productivity due to a UTI while the patient was working) were not available. Fourth, the friction cost method was chosen to determinate the indirect costs. The cost’s results should be interpreted with caution because this method is controversial in cases of short-term disease. The human capital approach overstates the production lost because the sick employee’s colleagues can offset the absence via increased productivity [27]. Consequently, the estimates of friction costs should represent the upper bound of the estimates of the short-term indirect costs [27]. However, in cases involving teamwork, the absence of an employee can also reduce the production of several employees [39]. Fifth, although the study was designed to exclude the costs of reinfection, the women who were symptomatic at 2 weeks might have experienced a reinfection between the second and the eighth weeks, and these costs were taken into account because differences between relapses and reinfections could not be identified during this period. Sixth, the costs of long-term symptomatic failure (i.e., the persistence of symptoms after 8 weeks) could not be taken into account because the follow-up period stopped after 8 weeks. Seventh, from the patient perspective, the costs could have been overestimated because some companies might have paid their employees during their sick leaves; or these costs could also have been underestimated because there were no data from which excess fees for medical imaging procedures could be assessed. Another point from the patient perspective is the costs of over the counter drugs, which could not still have been accounted for because costs differ between retail outlets. None of these data were available. As last limitation, the results of diagnostic tests prescribed by physicians were not collected.

According to Foxman [4], 10.8 % of women over 18 years of age experience at least one UTI per year, which represents more than 2.5 million people in France as of 1 January 2012 [40]. In our study, the estimated number of women who visited a GP for a UTI was estimated to 832,073 (95 % CI = 623,614–1,040,532) in 2012. However, the rate of care seeking for UTIs in France is unknown. Some women with UTI symptoms might have spontaneously recovered healed [41], visited other specialists (e.g., gynecologists and urologists) or emergency departments. The yearly number of emergency visits for UTIs is estimated to 410 000 (2.3 % of the 18 million of emergency visits), half less than in primary care [42, 43].

In Italy, the mean annual direct cystitis cost (i.e., physician visits, diagnostic tests and prescription drugs) from the Italian National Health Service (NHS) perspective was evaluated to be €229 per patient between January 2007 and December 2010 [44]. Each patient had an average of 4.5 episodes per year. The cost for the Italian NHS was higher than the cost for the French National Health Insurance. However, the women included in the Italian study visited referral centers for the treatment of cystitis, which might have resulted in the selection of more complicated UTIs. In the United Kingdom, a cost analysis was performed to evaluate the mean monthly direct UTIs costs in women between the ages of 18 and 70 years who were seen by a GP or a nurse between 2005 and 2006 from the NHS perspective. The mean monthly direct UTI costs were estimated at £30.60 when the UTIs were treated with empirical antibiotic therapy and £37.10 when urine cultures were performed [45]. The cost differences between this study and the present French study may have been related to the systematic one-month follow-up (with no differences between relapse and reinfection) used in the English study and the higher remuneration of English GPs (approximately 30 % greater than that of French GPs in 2008 [46]). In the United States in 2010, the annual direct and indirect UTI cost was estimated at $2.3 billion [47]. Considering that the American population was approximately five times greater than the French population in 2012, the American cost was six times greater than the French cost. This difference could be explained by the costs of physician visits and diagnostic tests, which are three to ten times more expensive in the United States than in France [48]. Cost-effectiveness studies have found that most cost-effective treatment is the empirical use of antibiotics that are effective against E. coli [12, 49, 50]. Furthermore, in the present French study, the mean UTI cost due to wild E. coli was not significantly different from the mean UTI cost due to antibiotic-resistant E. coli. This lack of a difference could have resulted from our use of systematic urine analyses. Even without antibiotics, 20 % of women recover from uncomplicated lower UTIs within 3 days and 26 % recover within one week [41, 51]. The effect of inadequate antibiotic prescription should be studied in greater detail and might have little influence on the consumption of care by women who visit GPs for UTIs.

Approximately one quarter of the suspected UTIs seen in general practice had negative urine cultures. This illustrates the limits of the clinical diagnosis of urinary tract infections. The probability of having an UTI for a woman with urinary symptoms is 48 % [52]; this probability increases if dysuria, urinary frequency or hematuria is present [53]. This suspected UTIs with negative urine cultures had a significant influence on cost, i.e., more than 23 % of the overall UTI cost. The environmental influence of antibiotic therapy for these women cannot be neglected due to the risk of the selection of resistant bacteria [54]. More than half of diagnostic tests performed for UTIs were prescribed outside of the recommendations as previously reported by other authors [55]. It is interesting to consider how to the prescription of potentially unnecessary and environmentally harmful treatments the administration of potentially inappropriate diagnostic tests can be prevented.