Timing of paediatric orchidopexy in universal healthcare systems: international administrative data cohort study

Background International guidelines in 2008 recommended orchidopexy for undescended testis at 6–12 months of age to reduce the risk of testicular cancer and infertility. Using administrative data from England, Finland, Ontario (Canada), Scotland and Sweden (with data from Victoria (Australia) and Iceland in supplementary analyses), the aim of this study was to investigate compliance with these guidelines and identify potential socioeconomic inequities in the timing of surgery before 1 and 3 years. Methods All boys born in 2003–2011 with a diagnosis code of undescended testis and procedure codes indicating orchidopexy before their fifth birthday were identified from administrative health records. Trends in the proportion of orchidopexies performed before 1 and 3 years of age were investigated, as were socioeconomic inequities in adherence to the guidelines. Results Across all jurisdictions, the proportion of orchidopexies occurring before the first birthday increased over the study period. By 2011, from 7·6 per cent (Sweden) to 27·9 per cent (Scotland) of boys had undergone orchidopexy by their first birthday and 71·5 per cent (Sweden) to 90·4 per cent (Scotland) by 3 years of age. There was limited evidence of socioeconomic inequities for orchidopexy before the introduction of guidelines (2008). Across all jurisdictions for boys born after 2008, there was consistent evidence of inequities in orchidopexy by the first birthday, favouring higher socioeconomic position. Absolute differences in these proportions between the highest and lowest socioeconomic groups ranged from 2·5 to 5·9 per cent across jurisdictions. Conclusion Consistent lack of adherence to the guidelines across jurisdictions questions whether the guidelines are appropriate.


Introduction
Undescended testis is common, affecting around 3-9 per cent of full-term boys at birth 1 . Most undescended testes spontaneously descend, but the problem persists beyond 1 year of age in around 1 per cent of individuals 1 . When uncorrected, undescended testis is associated with an increased risk of testicular cancer in adulthood as well as reduced testicular volume, sperm count and hormonal levels, with potential risks for fertility that may impact on psychosexual well-being [2][3][4][5][6][7][8][9][10][11] . Early surgical placement of the testis into the scrotum is recommended to reduce these risks.
Two European consensus statements 2,4 were published in 2008, followed by one from the British Association of Paediatric Urologists 12 in 2011, recommending surgical intervention between 6 and 12 months of age. Although guidelines from the American Urological Association 13 in 2014 recommended orchidopexy before the age of 18 months, a systematic review 14 concluded that surgery between 6 and 12 months of age may optimize fertility and protect against malignancy. Older, single jurisdiction, studies have found poor adherence to these guidelines [15][16][17][18][19][20][21][22][23] .
Many healthcare systems provide free universal care. Equity of access is a key goal of these systems. A number of investigations have found, however, that people of lower socioeconomic position (SEP) are less likely to access healthcare services after adjusting for clinical need 24-28 . There are also concerns that public health interventions that do not address root causes of inequities in access to healthcare may lead to increased health inequity overall 29 . Orchidopexy is a useful exemplar condition to assess inequities in healthcare access, because the occurrence of undescended testis is not known to be associated with SEP 30 and the condition is routinely screened for at birth.
The aims of this study were to investigate trends in the proportion of orchidopexies performed before 1 and 3 years of age in five jurisdictions (England, Finland, Ontario (Canada), Scotland and Sweden), and to investigate potential socioeconomic inequities in the proportion of boys from highversus low-SEP families having orchidopexy before 1 and 3 years of age, stratified according to the introduction of the guidelines in 2008.

Methods
Rolling yearly male-only birth cohorts were identified using administrative health data sets in five jurisdictions (England, Finland, Ontario, Scotland and Sweden) from 2003 to 2011. Owing to data limitations, the 2003 birth cohort was unavailable for Finland. All singleton live births among boys surviving to at least 6 months of age who had orchidopexy before their fifth birthday were included; the population denominator was the number of all singleton live-born males in each respective year.
The data set for each jurisdiction was a whole-population administrative data set from health services, from which it was possible to construct birth cohorts. Detail on each data set is available in Table S1 (supporting information). Each data set contained information on demographic and clinical characteristics of the patients, as well as dates and procedure codes associated with their hospital admissions. Complementary data were also supplied for Iceland and Victoria (Australia), but were excluded from the main analyses because of data limitations. In Iceland, the number of boys undergoing orchidopexy in each year was very small; in Victoria, birth cohorts could not be constructed and cross-sectional data only were provided. Direct comparisons with these jurisdictions were therefore not feasible. Results for these two regions are presented in the supporting information, as indicated in the results section.

Case definition
A case was defined as any child born in 2003-2011 with hospital records indicating a diagnosis of undescended testis and who underwent orchidopexy by the age of 5 years. Orchidopexies were identified using jurisdiction-specific procedure codes (Table S1, supporting information). As orchidopexy is also used to treat testicular torsion, patients were included only if they had an ICD-10 cryptorchidism diagnosis code (Table S1, supporting information) recorded either on the same admission record as their orchidopexy or previously. Only the first operation was counted (revisions or second stages were ignored). Patients were followed up to their fifth birthday to reduce the risk of counting acquired ascending testis, which can occur during later childhood 31 .
Boys born preterm (at less than 37 weeks' gestation) were excluded, as were those with a congenital anomaly at birth, identified by any ICD-10 diagnosis code indicating the presence of congenital anomalies in the first 3 months of life (Appendix S1, supporting information) 32 .

Outcomes
Trends in the proportion of cases by age 5 years, where the first orchidopexy procedure was performed before 1 and 3 years of age, were plotted. Inequity ratios were then calculated for boys born before (2003)(2004)(2005)(2006) and after (2008-2011) the introduction of European guidance recommending orchidopexy before 1 year of age. Inequity ratios were based on the distribution of births into quintiles (or quartiles for Finland) according to SEP using either household-or area-level measures recorded in the birth or subsequent admission records.

Ethical approval
Access to data was granted following appropriate approvals in each jurisdiction. In England, researchers had a data-sharing agreement with National Health Service Digital to use a deidentified extract of Hospital Episode Statistics linked to Office for National Statistics death registration data, so that ethical approval to use English data sets was not required. In Iceland, the study was approved by the Data Protection Authority and National Bioethics Committee (7 March 2017, VSN-17-044), the Directorate of Health (20 January 2017, 1701096/5.6.1) and Statistics Iceland (24 April 2017, 2017/01). No study permission was required in Finland, as only aggregated data were provided for the study group. The use of encoded Ontario data, accessed at ICES in this project, was authorized under section 45 of Ontario's Personal Health Information Protection Act, which does not require review by a Research Ethics Board. The Public Benefit and Privacy Panel for Health and Social Care (reference number 1516-0405) and the Privacy Advisory Committee (number XRB13020) provided permission in Scotland. The Swedish part of this study was approved by the Regional Ethics Committee in Stockholm in January 2016 (dnr 2016/5:1), and in Victoria this was covered by RCH HREC 37164.
Small cell counts were suppressed in accordance with the requirements of each jurisdiction.

Statistical analysis
Trends based on the annual cumulative incidence of orchidopexy by age 5 years in each jurisdiction were plotted over time, based on year of birth. Trends in the proportion of patients undergoing orchidopexy before 1 and 3 years of age in each region by year of birth were also plotted. These proportions are cumulative: the proportion having orchidopexy by 3 years of age includes those who had it by 1 year.
Inequity in age at orchidopexy was measured by taking the proportion of patients in the three highest (2 highest in the case of Finland) SEP groups over the proportion of patients in the two lowest SEP groups who received their first orchidopexy by 1 or 3 years of age. These ratios are termed inequity ratios. A ratio of 1 indicates equality between the patients with higher and lower SEP. Ratios above 1 indicate inequities in favour of patients with a higher SEP, and those below 1 indicate inequities in favour of patients with a lower SEP. Ninety-five per cent confidence intervals were also calculated. To assess inequities before and after the publication of the guidelines, inequity ratios were calculated for boys born in 2003-2006 (2004-2006 in Finland), and again for those born in 2008-2011.
Following up boys to their fifth birthday may have introduced bias into the analyses as later operations may nonetheless be for primary undescended testis. Sensitivity analyses were therefore carried out by extending follow-up to the tenth birthday. Data were available for these analyses only from England, Ontario, Scotland and Sweden.
Analyses were conducted in each region independently, based on a detailed specification of the cohort and codes required. The statistical software used in each region included R (R Foundation for Statistical Computing, Vienna, Austria) (England, Iceland, Victoria), Stata ® (StataCorp, College Station, Texas, USA) (Scotland) and SAS ® (SAS Institute, Cary, North Carolina, USA) (Finland, Ontario and Sweden). Aggregate results were entered into Microsoft Excel ® (Microsoft, Redmond, Washington, USA) for further analysis.   information). In 2011, 331 104 boys were born in England, 69 177 in Ontario, 54 400 in Sweden, 30 566 in Finland, and 28 099 in Scotland. Further descriptive data on the distribution of maternal age, birthweight and the numbers excluded due to death at less than 6 months of age are given in Appendix S2 (supporting information). Fig. 1 and Appendix S3 (supporting information) show the cumulative incidence of cases over time in each region. Across all jurisdictions, a minority of boys had orchidopexy before the recommended age of 1 year, but a large   (Fig. 3a) and 2008-2011 (Fig. 3b). The ratios for the 3-year threshold are shown in Fig. 3c,d. There was limited evidence of socioeconomic inequity in age at orchidopexy for births between 2003 and 2006, with small relative and absolute differences between boys with higher and lower SEP. For births between 2008 and 2011, however, there was consistent evidence across all jurisdictions of inequities for the first birthday guideline, favouring boys with a higher SEP. Boys born in 2008-2011 with a higher SEP were 25-48 per cent more likely to have had their operation by 1 year of age, depending on region, compared with those with a lower SEP. The absolute differences were small, and ranged from 2⋅5 to 5⋅9 per cent across each region.

Characteristics of boys born in 2011 in each jurisdiction are given in
There was limited evidence of very small inequities in the 3-year threshold for births in 2003-2006 or 2008-2011. Results of the sensitivity analyses are given in Appendices S5 and S6 (supporting information). There were no substantive differences between the main and sensitivity analyses in inequity ratios.

Discussion
Across all jurisdictions, a small but growing minority (between 7⋅6 and 27⋅9 per cent in 2011) of boys had orchidopexy before the recommended age of 1 year, although most underwent the operation by 3 years. There was evidence of socioeconomic inequity in the proportion of boys born after the introduction of guidelines in 2008 who received orchidopexy by 1 year of age, favouring those of higher SEP. There was no clear evidence of inequity for those born before 2008 or for orchidopexy by 3 years of age before or after the introduction of guidelines. The findings showed remarkable consistency across regions.
Each jurisdiction had its own measure of SEP (Table S1, supporting information) 33,34 . However, relative comparisons between quintiles within countries should be stable, rendering within-country estimates reliable. There were small numbers of cases within each quintile, hence the necessity to dichotomize SEP, which may have masked more subtle gradations. The administrative data resources do not have adequate data on referrals and so it was not possible to analyse separate parts of the pathway between referral and treatment. It may, for example, be that any observed inequities are due to later diagnosis, referral or decision to treat. Despite possible different referral pathways in each jurisdiction or in the availability of operating theatre space, these results were consistent across jurisdictions, suggesting that results may be similar in other countries.
Undescended testis is a relatively common condition, which was clearly and consistently coded within jurisdictions. It is screened for at birth and that process is not associated with SEP (after excluding preterm births and boys with congenital anomalies). A limitation in the diagnostic system used, however, is that it was not possible to separate congenital from acquired cryptorchidism. In this study, operations beyond 5 years of age were censored to minimize the inclusion of acquired cryptorchidism, but it seems probable that such patients may explain some of the high rates of late operation. This study spanned a sufficiently long period to cover the introduction of international guidelines in 2008 and therefore captured what should have been a wave of innovation adoption across all SEP groups.
Other studies [15][16][17]19 have noted that guidelines are not being met routinely. In the present study, it was observed not only that this was a persisting issue but that, by using comparable methods across jurisdictions, it was consistent in different countries. There are several potentially overlapping reasons for non-adherence. Operating after 1 year of age may result from delayed diagnosis 15 or lack of knowledge about the recommended age of orchidopexy, thus causing later referral and older age at orchidopexy 14 . Healthcare systems must balance the need to perform early orchidopexy against the need to carry out other procedures, where there may also be delays and non-adherence to timing guidelines 35 . Evidence about long-term fertility and malignancy risk, particularly regarding whether orchidopexy should be performed at less than 1 year or later, is still uncertain given the limitations of existing studies, with small sample sizes or proxy outcomes. In a systematic review 14 evaluating the optimal age for orchidopexy, only two 36,37 of 24 studies examined fertility as opposed to a proxy. As for the association with malignancy, current literature has focused mostly on children undergoing orchidopexy at a later age (comparing children having the procedure at age above 10 or 13 versus less than 10 or 13 years) 7 , and it remains unclear whether findings showing that earlier surgery reduces risk for children aged under 10 years can be extrapolated to infants. The present findings of overwhelming non-adherence might in part be due to these limitations in the evidence, particularly if practitioners are weighing the risks of long-term harm with the risks and contraindications of early surgery 38 . This should therefore lead to the generation of stronger and more robust evidence of the balance of harms and benefits according to the timing of orchidopexy.
This study also showed socioeconomic inequities, which occurred most clearly after introduction of the guidelines. This suggests possible inequities in the adoption of the 'innovation' regarding surgery in the first year. Diffusion of innovation in healthcare takes time 39 , and it may be that those in better socioeconomic circumstances, as a result of either their own resources or interactions with the healthcare system, are more likely to be 'early adopters' through better access to the healthcare system or being more informed or questioning parents 29 . It has been suggested 28 , in the context of waiting times, that people with a higher SEP, by having higher levels of education and social capital, may be better equipped to articulate their needs and engage actively in waiting list systems. As these people are likely to have better working conditions, it might also be easier for them to attend appointments 28 . This is not to suggest that people with a lower SEP are less likely to seek healthcare, as there is evidence to the contrary 24 , but that they may be at a disadvantage when dealing with a complex system. If the findings of the present study are the result of diffusion of innovation, it may be that those with a lower SEP will 'catch up' and close the currently modest inequities in age at orchidopexy.
Across five jurisdictions, orchidopexy was not performed consistently before a child's first birthday, as recommended by international guidelines, for the majority of patients. Although lack of adherence to recommended practice may reflect balancing the risks of delaying surgery with those of performing an operation under general anaesthesia early in life, there may be systems-level factors, such as limited availability of hospital space, that cause delay. Evidence of long-term harms associated with performing orchidopexy before or after 1 year of age, however, remains very weak. Long-term monitoring of outcomes and further research into the precise timing of events within the diagnosis-referral-treatment pathway still need to be carried out. National and international guidelines should be reassessed to account better for the risks of early versus delayed surgery.