Article Text

Significantly increasing hospital admissions for acute throat infections among children in England: is this related to tonsillectomy rates?
  1. Elizabeth Koshy1,
  2. Joanna Murray1,
  3. Alex Bottle2,
  4. Paul Aylin2,
  5. Mike Sharland3,
  6. Azeem Majeed1,
  7. Sonia Saxena1
  1. 1Department of Primary Care and Public Health, Imperial College London, London, UK
  2. 2Dr Foster Unit at Imperial College London, Department of Primary Care and Public Health, Imperial College London, London, UK
  3. 3Department of Paediatric Infectious Diseases Unit, St George's Hospital, London, UK
  1. Correspondence to Dr Elizabeth Koshy, Department of Primary Care and Public Health, Imperial College London, Reynolds Building, St Dunstan's Rd, London W6 8RF, UK; e.koshy{at}imperial.ac.uk

Abstract

Objective To examine trends in hospital admissions for acute throat infection (ATI) and peritonsillar abscess (PTA) alongside tonsillectomy trends in children.

Design We analysed Hospital Episode Statistics data to calculate annual age-standardised and age-sex specific rates for ATI, PTA and tonsillectomies in children aged 0–17 years who were admitted to hospital in England between 1 April 1999 and 31 March 2010.

Results Age-standardised admission rates for ATI increased by 76% from 107.3 (95% CI 105.3 to 109.2) to 188.4 (95% CI 185.9 to 191.0) admissions per 100 000 children. Median length of stay for ATI admissions decreased from 1 to 0 days. Admission rates for PTA remained stable at between 9.6 (95% CI 9.0 to 10.2) and 8.7 (95% CI 8.1 to 9.2) per 100 000 children in 1999/2000 and 2009/2010, respectively. Age-standardised tonsillectomy rates declined from 367.4 (95% CI 363.8 to 371.0) to 278.0 (95% CI 274.9 to 281.1) per 100 000 children between 1999/2000 and 2000/2001, respectively, increased to 322.4 (95% CI 319.0 to 325.7) in 2002/2003 and then gradually declined again to 293.6 (95% CI 290.4 to 296.8) in 2009/2010.

Conclusions ATI admission rates have increased substantially in the past decade, but the majority of children are discharged after a short stay. PTA admission rates have remained stable. This suggests the severity of throat infection has not increased. Tonsillectomy rates in England have been declining overall but do not appear to be associated with this increasing trend in ATI admissions. The increase most likely reflects changes in primary care and hospital service provision.

  • Infectious Diseases
  • Ent
  • Paediatric Practice

Statistics from Altmetric.com

What is already known on this topic

  • Acute throat infections (ATIs) range widely in severity, from self-limiting to serious infections requiring hospital admission.

  • Tonsillectomy, with or without adenoidectomy, is one of the most common operations performed in children.

  • It is not known whether recent reductions in tonsillectomy rates have led to an increase in hospital admission rates for ATIs of increased severity.

What this study adds

  • ATI hospital admission rates in England increased linearly by 76% between 1999 and 2010, but the median length of hospital stay declined.

  • Peritonsillar abscess (PTA) rates remained stable between 1999 and 2010.

  • Tonsillectomy rates in England have declined overall, but this was not associated with increased severity of ATI or PTA admissions.

Introduction

Acute throat infection (ATI) in children imposes a considerable burden on the child and their family and, in the UK, is among the most common reasons for consulting a general practitioner (GP).1 ,2 ATI encompasses acute pharyngitis and acute tonsillitis, and it can be difficult to differentiate between them clinically.1 ATI, if severe, can require hospital admission. Rarely, suppurative complications such as peritonsillar abscess (PTA) can develop, which accounts for 3–5 hospital admissions per 100 000 children in England3 and may require surgical interventions such as incision and drainage.

National guidelines and reviews of effectiveness recommend (adeno)tonsillectomy for children with throat infections which recur frequently (based on the Paradise criteria)4 or become chronic, and adversely affect a child's well-being.1 ,5 ,6 Although tonsillectomy can prevent tonsillitis, the impact on pharyngitis is less predictable.5 Tonsillectomy, with or without adenoidectomy, is one of the most common operations carried out in children and the majority of tonsillectomy operations are performed for recurrent or chronic throat infections.7 However, the efficacy of tonsillectomy and evidence about which children are most likely to benefit is debated.1 ,5 The potential benefits of tonsillectomy in reducing recurrent or chronic throat infection need to be weighed against operative and post-operative complications, including haemorrhage and infection,8 and the possibility that the throat infections might spontaneously resolve without intervention.5 ,9 Evidence from systematic reviews and randomised controlled trials in children suggest only modest reductions in throat infection episodes following (adeno)tonsillectomy, predominantly in children who are more severely affected at baseline.5 ,4 ,10 ,11

These findings have led to a drive to curtail the number of operations performed.1 The UK has one of the lowest tonsillectomy rates in Europe12 and there are concerns that this may leave some individuals exposed to avoidable morbidity.7 As tonsillectomy rates continue to fall, the hospital admission rate for tonsillitis and its complications may rise.7 We hypothesised that the decrease in tonsillectomies may be associated with an increase in hospital admission rates for ATI of increased severity and for PTA in children. We examined national data on hospital admissions for children with ATI and PTA, and trends in tonsillectomy rates, from 1999/2000 to 2009/2010.

Methods

Data source and definitions

The Hospital Episode Statistics (HES) database provides information on all episodes of National Health Service (NHS) hospital care in England.13 The data are recorded as completed consultant episodes, defined as the time during which a patient is under the care of a particular hospital consultant. The main reason for hospital admission (‘primary diagnosis’) is coded using the International Classification of Diseases, 10th revision (ICD-10),14 as are up to 19 secondary diagnoses. All procedures or operations during the hospital admission are coded using the OPCS-4 system (Office of Population, Censuses and Surveys: Classification of Interventions and Procedures, 4th Revision).

We obtained data on all emergency hospital admissions for children (0–17 years) where ATI or PTA were recorded as the primary diagnosis (see online supplementary data). Acute pharyngitis and acute tonsillitis are difficult to differentiate clinically and so we grouped associated codes (J02.0, J02.8, J02.9, J03.0, J03.8 and J03.9) into a single outcome category of ‘ATI’. In the UK, PTA is managed in a hospital setting and the codes we searched for were J36 and F36.3. We also investigated viral upper respiratory tract infections (URTI) and total emergency hospital admissions in children to study secular trends. We studied admissions for elective and emergency tonsillectomy and adeno-tonsillectomy, where tonsillectomy was the main operation recorded. The codes (F34) were adapted from those used by the UK Department of Health.15

Outcome measures and analysis

The main outcome measures were age-standardised, age-specific and age-sex specific rates for ATI, PTA and tonsillectomies in children aged <18 years between 1 April 1999 and 31 March 2010. We also calculated these rates separately for boys and girls aged <1, 1–4, 5–9, 10–14 and 15–17 years, reflecting important stages in childhood development. The median length of stay (LOS) for admission for throat infection was calculated to provide a proxy measure of severity of illness. A LOS of zero means that the patient was admitted and discharged on the same day without an overnight stay. We categorised admissions according to LOS <2 days or ≥2 days to separate minor from more severe infections. We calculated admission and tonsillectomy rates as the total number of admissions or operations divided by the Office for National Statistics (ONS) mid-year child population estimates.16 Admission data were for UK financial years (April–March), whereas the population estimates were for calendar years, so there was a consistent overlapping period between them for all study years. For age-standardisation, the rates were directly standardised to the 2004 ONS age structure of the English population, representing the mid-point of the study period, to enable comparisons across all years. We explored cohort effects to examine the potential long-term effect of tonsillectomy on subsequent admission for tonsillitis. We assumed that the population of children who underwent tonsillectomy aged 1–4 years between 1999/2000 and 2005/2006 would have been aged 5–9 years 4 years later, between 2003/2004 and 2009/2010. Similarly, children who had a tonsillectomy aged 5–9 years between 1999/2000 and 2004/2005 would have been aged 10–14 years between 2004/2005 and 2009/2010.

All rates were calculated with associated 95% CI using Poisson approximation. Statistical analysis was performed using Stata/SE V.11.1.17 We have ethics approval for this study from the South East Research Ethics Committee.

Results

There were a total of 193 973 admissions for ATI and PTA among children aged 0–17 years between 1999/2000 and 2009/2010 (see online supplementary data), 2% of all emergency admissions among children over this time frame.

Acute pharyngitis admissions accounted for 10% (18 640) of all ATI admissions over the decade. The age-standardised admission rates for ATI increased linearly by 76% from 107.3 (95% CI 105.3 to 109.2) to 188.4 (95% CI 185.9 to 191.0) admissions per 100 000 children between 1999/2000 and 2009/2010, respectively. The highest rates were among children aged 1–4 years, followed by children aged <1 year (figure 1). There was an increasing pattern of admissions for ATI for boys and girls in each age group over the decade. The largest absolute and relative increases in admission rates during the study period were among children aged <5 years (see online supplementary data). Boys were more likely to be admitted with ATI than girls among children aged <5 years, whereas girls were more commonly admitted among the 10–17-year age range. Admission rates for PTA remained stable at between 9.6 (95% CI 9.0 to 10.2) and 8.7 (95% CI 8.1 to 9.2) per 100 000 children in 1999/2000 and 2009/2010, respectively. Admission rates for PTA increased with age and were highest in the 15–17-year age group. Girls aged 10–17 years were more likely to be admitted with PTA than boys of the same age. Age-standardised total emergency admission rates increased linearly by 19% over the decade from 5828 (95% CI 5813 to 5842) to 6941 (95% CI 6926 to 6957) per 100 000 children. By contrast, trends in viral URTI showed year-to-year variability over the study period, ranging from 79 (95% CI 77 to 81) per 100 000 children to 124 (95% CI 122 to 126) per 100 000 children.

Figure 1

Annual admission rates for acute throat infections for boys and girls by age group.

Age-standardised admission rates for ATI with a LOS of <2 days increased from 73 (95% CI 72 to 75) to 157 (95% CI 155 to 160) per 100 000 children between 1999/2000 and 2009/2010, respectively. Although there was an absolute increase in admissions for ATI, the median LOS for all ATI admissions decreased from 1 day in 1999/2000 to 0 days (ie, no overnight stay) in 2008/2009. By contrast, the corresponding rates for a LOS of ≥2 days fell from 34 (95% CI 33 to 35) per 100 000 children in 1999/2000 to 31 (95% CI 30 to 32) in 2009/2010. The median LOS for PTA and tonsillectomy remained at 2 days and 1 day, respectively.

The modal ages for tonsillectomies were 4 and 5 years, which represented 25% (94 006/378 861) of all tonsillectomies over the decade. Tonsillectomy trend patterns did not reflect those for ATI and PTA. Age-standardised tonsillectomy rates declined by 24% from 367.4 (95% CI 363.8 to 371.0) to 278.0 (95% CI 274.9 to 281.1) per 100 000 children between 1999/2000 and 2000/2001, respectively. The rate then increased by 16% to 322.4 (95% CI 319.0 to 325.7) in 2002/2003, followed by a slight downward trend to 293.6 (95% CI 290.4 to 296.8) in 2009/2010. Tonsillectomy rates were higher among boys aged <5 years compared with girls and were lower among boys aged 5–17 years compared with girls, particularly in the 10–17-year range (see online supplementary data). Tonsillectomy rates among children aged 5–17 years declined between 1999/2000 and 2009/2010 (figure 2). The largest absolute reduction in procedure rate was among children aged 5–9 years. By contrast, there was an absolute and relative increase in the tonsillectomy rate for children aged 1–4 years from 2001/2002 onwards.

Figure 2

Annual tonsillectomy rates for boys and girls by age group.

Tonsillectomy rates decreased among children aged 1–4 years between 1999/2000 and 2001/2002. Four years later admission rates for ATI in children aged 5–9 years increased (see online supplementary data). A similar relationship was present among children aged 5–9 years who underwent tonsillectomy and admission rates for ATI among children aged 10–14 years, 5 years later. After these time points the trends in tonsillectomy rates and subsequent ATI rates closely mirrored each other.

Discussion

Our study shows a 76% linear increase in ATI admission rates in children between 1999/2000 and 2009/2010. The majority of these children were admitted for less than a day and PTA admission rates have remained stable, suggesting that the severity of illness has not increased. Tonsillectomy rates varied over the decade with an overall downward trend. There was no evidence of an association between tonsillectomy rates and severity of ATI or PTA admission rates.

We were unable to identify any recent studies reporting on trends in emergency admission for ATI in children. PTA admission rates are consistent with those previously reported, given our wider target age range and that we included the surgical procedural code for incision and drainage of tonsillar abscess.3 Our tonsillectomy numbers are comparable with numerator data in the Department of Health report on ‘Trends in Children's Surgery 1994–2005’15 and our tonsillectomy rates are also of similar magnitude to those reported in a retrospective analysis between 2000 and 2005 in England (304 per 100 000 children aged <15 years).18

Our study population is large and our findings of major increasing ATI trends were distinct from viral URTI fluctuations and independent of secular trends showing a modest 19% increase in total emergency admissions in children. Thus, the observed increases are unlikely to have arisen by chance. Increases in ATI admissions cannot be explained by falls in community antibiotic prescribing. Although antibiotic prescribing rates by GPs in the UK decreased by a quarter between 1996 and 2000,19 they subsequently increased by 40% from 2000 to 2007.20 We found no other evidence that the incidence of ATI has increased. General practice consultations for ATI and URTI appear to be stable or declining in the UK and the Netherlands.21–24 An alternative explanation may be increased virulence of organisms responsible for ATI, such as group A-β-haemolytic streptococcus (GABHS).25–27 However, if increased virulence of GABHS was responsible, we might have seen a concurrent increase in the median LOS for ATI and PTA admission rates.

ATI admissions increased linearly during the entire study period despite variability in tonsillectomy rates during the early phase of the study period. This suggests that tonsillectomy trends are not associated with trends in ATI admissions. The increase in tonsillectomy rates among children aged 1–4 years from 2001/2002 onwards may reflect increased throat infections in this age group or perhaps increased referral rates for children with obstructive sleep apnoea, which is more common in younger children.28

There were potential cohort effects during the early years of the study suggesting an inverse relationship between the number of tonsillectomies and ATI admissions. However, in this ecological study, we cannot ascertain if the children who were admitted with ATI had previously undergone tonsillectomy or not. Longitudinal studies with individual-level data would ideally be needed to explore this further.

The selective increased proportion of short stay (<2 days) admissions and reduction in median LOS from 1 to 0 days could indicate a trend towards treatment in hospital rather than in the community for ATI. Our findings suggest that the increase in very short stay admissions for ATI is a possible explanation for the observed increase in admissions, as opposed to a real increase in ATI morbidity. This is supported by our previous study,29 which found an overall increase in short stay admissions for minor illnesses in children in recent years and we suggested it may reflect changing thresholds for admission.

Changing healthcare provision in England in recent years may have increased hospital attendance rates for conditions formerly seen in primary care. Our findings may also, in part, represent changes in the definition of hospital admission over this decade, reflecting the increased emergence of short stay observation wards to enable adherence to the 4 h waiting time target in Accident and Emergency (A&E) departments which was introduced in 2004. Total emergency admissions increased by 18% (from 708 000 to 834 000 admissions) amongst children aged 0–19 years between 1996/1997 and 2006/2007.30 This increased workload may have contributed to increasing numbers of young children being admitted through A&E with a diagnosis of tonsillitis who would previously have been managed in the community. There may also have been potential changes in diagnostic coding in hospitals to help justify hospital admission, which could explain our findings. For example, tonsillitis may have become increasingly recorded as the reason for admission as it is considered a more serious diagnosis than a suspected viral URTI. Further research is needed to investigate the wide spectrum of ATI presentations at the primary and secondary care interface, including primary care and out-of-hours settings, and A&E.

Tonsillectomy trends varied over the decade, although they declined overall. The sudden decline in rates between 1999/2000 and 2000/2001 may be related to the publication of national guidelines on sore throat and tonsillectomy in 1999.1 However, the rationale for performing tonsillectomy operations is multi-factorial. In addition to evidence and guidance, the number of procedures performed is strongly influenced by parental views, the attitudes and referral patterns of GPs, locally commissioned pathways and the threshold for intervention of ear, nose and throat (ENT) surgeons.

Our study is among the first to examine trends in ATI admissions and tonsillectomies among children in England. The major strength of this study is the large, representative population-based sample. The main limitations relate to the quality of the HES data and misclassification bias in the use of diagnostic codes as proxy measures for infectious clinical outcomes in children and LOS as a marker of disease severity. Admission data are collected by clinical staff who record the reason for admission and administrative staff who subsequently code medical information from clinical records on discharge following admission. However, the quality of HES has improved considerably in recent years.31 HES data only capture NHS activity and so we were unable to monitor trends in tonsillectomies in the private sector, which accounts for around 15% of ENT activity in the UK.18 Hence, we are likely to have under-estimated tonsillectomy rates. Finally, as this is an ecological study, we cannot ascertain whether the increase in ATI is due to a greater number of children being admitted or an increase in the number of admissions per child, nor can we infer causality for the dramatic increase to any single factor.

Major increases in ATI admissions in children have not been accompanied by any increase in severity of illness or PTA complications. Therefore, the concerns that declining tonsillectomy rates result in increased admissions for more severe ATI and PTA do not appear to be justified. However, longitudinal studies are needed at an individual level to establish if any association does exist. In the meantime, it seems reasonable for clinicians to maintain a high clinical threshold for performing tonsillectomy for recurrent throat infections, so that it is restricted to those children with severe recurrent infections who are the most likely to benefit.

Acknowledgments

This study represents independent research commissioned by the National Institute for Health Research Service Delivery and Organisation programme. The interpretation and conclusions contained in this study are those of the authors. The Department of Primary Care and Public Health at Imperial College London is grateful for support from the National Institute for Health Research Biomedical Research Centre scheme and the National Institute for Health Research Collaboration for Leadership in Applied Health Research and Care scheme. We are also grateful for support from the Medical Research Council and the Engineering and Physical Sciences Research Council.

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:

Footnotes

  • Contributors EK and SS conceived the idea for the study. JM extracted the data. EK analysed the data and drafted the paper. All authors contributed to the writing and the interpretation of the data, and approved the final version.

  • Funding EK is funded by a National Institute for Health Research Doctoral Research Fellowship. SS is funded by a National Institute for Health Research Senior Researcher fellowship. JM is funded through the National Institute for Health Research (NIHR) under a Programme Grant for Applied Research. The Dr Foster Unit at Imperial College London is principally funded via a research grant by Dr Foster Intelligence, an independent health care information company and joint venture with the Department of Health. The Dr Foster Unit is affiliated with the Centre for Patient Safety and Service Quality at Imperial College Healthcare NHS Trust, which is funded by the National Institute of Health Research.

  • Ethics approval This study was approved under Section 251 (formerly Section 60) granted by the National Information Governance Board for Health and Social Care (NIGB, formerly the Patient Information Advisory Group). We also have approval for using these data for research from the South East Research Ethics Committee.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Atoms
    R Mark Beattie