Variability of adenoidectomy/tonsillectomy rates among children of the Veneto Region, Italy
© Fedeli et al. 2009
Received: 18 September 2008
Accepted: 07 February 2009
Published: 07 February 2009
Skip to main content
© Fedeli et al. 2009
Received: 18 September 2008
Accepted: 07 February 2009
Published: 07 February 2009
Despite national guidelines in 2003 aimed at limiting the recourse to tonsillectomy and/or adenoidectomy (A/T), the latter are among the most frequent pediatric surgeries performed in Italy. Aim of the study is to investigate variability of A/T rates among children of the Veneto Region, Italy.
All discharges of Veneto residents with Diagnosis-Related Groups 57–60 and ICD9-CM intervention codes 28.2 (tonsillectomy), 28.3 (adenotonsillectomy), 28.6 (adenoidectomy) were selected in the period 2000–2006 for a descriptive analysis. A multilevel Poisson regression model was applied to estimate Incidence Rate Ratios (IRR) with 95% Confidence Intervals (CI) for A/T surgery among children aged 2–9 years in 2004–2006, while taking into account clustering of interventions within the 21 Local Health Units.
Through 2000–2006, the overall number of A/T surgeries decreased (-8%); there was a decline of adenoidectomies (-20%) and tonsillectomies (-8%), whereas adenotonsillectomies raised (+18%). Analyses on children aged 2–9 resulted in an overall rate of 14.4 surgeries per 1000 person-years (16.1 among males and 12.5 among females), with a wide heterogeneity across Local Health Units (range 8.1–27.6). At random intercept Poisson regression, while adjusting for sex and age, intervention rates were markedly lower among foreign than among Italian children (IRR = 0.57, CI 0.53–0.61). A/T rates in the 10–40 age group (mainly tonsillectomies) computed for each Local Health Unit and introduced in the regression model accounted for 40% of the variance at Local Health Unit level of pediatric rates (mainly adenoidectomies and adenotonsillectomies).
A/T rates in the Veneto Region, especially adenoidectomies among children aged 2–9 years, remain high notwithstanding a decrease through 2000–2006. A wide heterogeneity according to nationality and Local Health Units is evident. The propensity to A/T surgery of each Local Health Unit is similar in different age groups and for different surgical indications.
Tonsillectomies and/or adenoidectomies (A/T) are among the most frequent pediatric surgeries performed in Italy ; in 2001 they accounted for 28% of surgical hospitalizations in children aged 0–14 . National guidelines issued in 2003 restricted the surgical option mainly to: children with significant obstructive apnea (adenotonsillectomy), children with recurrent otitis media and ventilation-tube placement or with chronic/recurrent sinusitis and failure of appropriate antibiotic therapy (adenoidectomy), children and adults with severe acute recurrent tonsillitis (tonsillectomy) .
The first claim of high tonsillectomy rates and geographical variability unexplained by medical indications was issued in England and Wales as early as in 1938 . 70 years later, wide variation between and within countries continues to be registered . The adoption of guidelines with clear indications for surgery should overcome such differences in surgical rates, at least in an area with mild heterogeneity in the delivery of health care such as the Veneto Region, Italy. Aim of the study is to investigate variability of A/T rates among children of the Veneto Region.
Population, otorhinolaryngological units and hospital beds, number of tonsillectomies (T), adenotonsillectomies (T+A), adenoidectomies (A) performed on children aged 2–9, years 2004–2006.
Population 2–9 yrs
% non-Italians, age class 2–9 yrs
n. otorhinolaryngological units
Total hospital beds available in otorhinolaryngological units
T (2–9 yrs)
T+A (2–9 yrs)
A (2–9 yrs)
Data on the mid-year population by sex, age, Local Health Unit and citizenship were computed from data provided by the National Institute of Statistics . Since the pediatric population is rapidly changing in our Region mainly due to immigration, and data from the National Institute of Statistics on non-national population is available only for more recent years, further analyses were restricted to children aged 2–9 operated in 2004–2006. Crude rates of A/T surgery as a whole, adenoidectomy, tonsillectomy with or without adenoidectomy were computed for each Local Health Unit. A random-intercept Poisson regression model was applied to estimate Incidence Rate Ratios (IRR) with 95% confidence Intervals (CI) for A/T surgery associated with age, sex, and citizenship (Italian and non-Italian), while taking into account clustering of interventions within Local Health Units. Finally, a cluster-level variable (overall A/T rates of each Local Health Unit in the 10–40 year age class) was introduced in the hierarchical model to assess if it could account for part of the variability between Local Health Units in pediatric A/T rates.
Age-adjusted Incidence Rate Ratios (IRR) with 95% Confidence Intervals (CI) for tonsillectomy and/or adenoidectomy (A/T) among children aged 2–9 years.
Males vs females
Non-Italian vs Italian children
A/T rates (×1000) in each LHU, age class 10–40 yrs
The regional archive of hospital discharge records includes hospitalizations in both public and private hospitals, as well as hospitalizations of Veneto residents in facilities outside the region. Since A/T surgeries are performed in hospital (mainly as day-case, or with a single overnight stay) and not in outpatient settings in Italy, the hospital discharge records archive allows a complete analysis of these interventions.
The age and sex distribution of A/T surgeries found in the Veneto region is similar to that reported from other European countries [6, 7]; the observation of higher intervention rates in male children is consistent since the first paper issued in 1938 .
Usually age and sex standardized intervention rates are reported separately for tonsillectomy (with or without adenoidectomy) and for adenoidectomy. We maintained the above distinction; however since there are two distinct peaks, around age 4 for adenotonsillectomy and adenoidectomy and among adolescents-young adults for tonsillectomy, we preferred to restrict the age range of most analysis and deal with a homogenous population in terms of type of surgery and demographics (e.g. relevance of the immigrant population).
A/T rates show a wide variability between countries [8, 9], but it has been long recognized that, despite differences in average rates across countries, even more striking variations always exist from area to area of the same country [10–13]. A wide small-area variation in intervention rates suggests that there is a high degree of "clinical uncertainty" about the indications . In Italy, high heterogeneity in tonsillectomy rates between regions, only in part related to minimum temperatures registered in regional capitals, was observed in 2000 . After the introduction of national guidelines for A/T surgery in 2003, a reduction in the number of interventions and in the inter-regional variability of rates has been registered . However, data from the Veneto Region suggest that the decline in the number A/T surgeries (mainly adenoidectomies) started before guidelines were issued, and disappeared in recent years. Moreover, wide small-area variation in intervention rates still persists. In our analyses we did not take into account factors that may be related to surgery rates at the Local Health Unit level, such as accessibility of primary health care or resource available in otorhinolaryngological units. The hierarchical model applied to our data shows a direct relationship between A/T rates in children aged 2–9 years and in adolescent-young adults aged 10–40 years, even though main surgical indications and type of operation are different between these age groups. This finding suggests that variability between Local Health Units is better explained by surgeons' general propensity to tonsillectomy and adenoidectomy procedures rather than by different underlying disease rates. All the above indicate that national guidelines did not have a major impact on the selection of children for surgery, similarly to that observed in other countries . An update of national guidelines has been issued in 2008, further restricting indications to ventilation-tube placement and adenoidectomy in recurrent otitis , but strategies should be implemented for enhancing adherence among physicians, who could see guidelines both as a threat to their decisional autonomy  and as a cost-cutting tool for health care. Few surgeons are hesitant believers in the necessity and efficacy of the operations they perform, yet feedback of population-based surgical rates showing how much such certainties vary from place to place  should convince of the need to reevaluate patterns of patient selection at the local level .
The association between socio-economic status and A/T rates has been investigated mostly at an ecological level, with some studies showing higher rates related to a lower socioeconomic level , including one investigation carried out in 1997 in Rome, Italy, claiming that disadvantaged groups are more vulnerable in receiving unnecessary surgical treatments . The hospital discharge records archive did not permit to include socioeconomic indicators in our analysis. However, the relation between socioeconomic status and intervention rates may change over time and space; the steep increase of the immigrant population in Italy in the last few years (about threefold from 1997 to 2006) could partially overcome earlier observations. In our study A/T rates are strikingly lower in the non-national (or foreign) population, which consists of persons who still have the nationality of their home country and may include persons born in Italy. Separate analyses for foreign children born within or outside the Veneto Region could have been of great interest but were not feasible due to the lack of denominator data (population by citizenship: children born in Italy from foreign parents maintain their parents' citizenship). Nonetheless, the proportion of foreign children born in the Veneto Region is rapidly growing (about 18% of newborns in Veneto are currently non-Italians), and -except for housing conditions-shares most environmental risk factors (climate, air pollution, schooling) with the Italian children. Still, being born in a foreign country or having foreign parents can be considered as a proxy variable for the socioeconomic condition. Differences in A/T rates between Italian and non-national residents could be due to different rates of the underlying diseases, but could be better explained by different patterns of diagnosis and care. The latter hypothesis is reinforced by investigations on other types of discretionary surgeries in Italy (e.g. elective cesarean section), showing lower rates among the foreign population . Lower tonsillectomy rates in children from neighborhoods with larger immigrant population has been also observed in Ontario, Canada, with differences possibly arising at the level of the primary care physician, the specialist, or parental attitudes .
Despite the introduction of national guidelines, a wide heterogeneity in A/T rates according to nationality and Local Health Unit is evident in the Veneto Region. The propensity to A/T surgery of each Local Health Unit is similar in different age groups and for different surgical indications. These findings uncover issues both of variability in medical practices and equity in the access to health care.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.