Skip to main content
NCBI home page
The BMJ logoLink to The BMJ
. 2002 Mar 16;324(7338):673.

Effectiveness of guidelines on persistent glue ear in children

Authors' estimates of size of impact are probably excessive

Nick Black 1,2, Andrew Hutchings 1,2
PMCID: PMC1122587  PMID: 11895832

Editor—Mason et al have shown how the Effective Health Care bulletin on treating persistent glue ear in children was associated with a subsequent significant decline in rates of surgery for this condition.1 Their estimates of the size of the impact, however, are probably excessive. This illustrates the dangers of undertaking short time series analyses.

They compared surgical rates after publication of the bulletin with the rates in the preceding three years. They report that during the preceding years the rate rose from about 1.7 to 2.1 per 1000 children. The rise over this period, however, does not reflect the longer term trends in the surgical rate observed in children aged under 10 in the districts covered by the old Oxford and East Anglian regions: the rate peaked in the mid-1980s and then fell slowly but steadily during the following six years.2 Mason et al seem to have based their claim that the rate increased greatly during the 1980s on a paper that covered the period only up to 1982.3

The upturn during 1989-92 has to be seen in the context of an overall decline from 1985 to the present day. The reason for the short reversal in an otherwise steady decline is unclear. One clue is provided by the observation that the same happened for tonsillectomy rates, which, with the exception of an upturn in 1989-92, have continued their longstanding decline right up to the present day.

This paper illustrates the need, when undertaking time series analyses, to consider a sufficiently long period.

References

  • 1.Mason J, Freemantle N, Browning G. Impact of Effective Health Care bulletin on treatment of persistent glue ear in children: time series analysis. BMJ. 2001;323:1096–1097. doi: 10.1136/bmj.323.7321.1096. . (10 November.) [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Black NA. Surgery for glue ear: the English epidemic wanes. J Epidemiol Community Health. 1995;49:234–237. doi: 10.1136/jech.49.3.234. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Black NA. Surgery for glue ear—a modern epidemic. Lancet. 1984;i:835–837. doi: 10.1016/s0140-6736(84)92280-3. [DOI] [PubMed] [Google Scholar]
BMJ. 2002 Mar 16;324(7338):673.

Effect claimed may depend on how persistent glue ear is defined

P Coleman 1,2,3, J Nicholl 1,2,3, E Warren 1,2,3

Editor—Mason et al reported a dramatic decline in procedures to insert grommets for glue ear in the population of England aged under 15.1-1 This prompted us to re-examine our earlier analysis of the same dataset for the same condition, included in an evaluation of the reports published by the South and West Development and Evaluation Committee.1-2 The committee's recommendation that glue ear should be managed by a period of watchful waiting1-3 mirrored that of the Effective Health Care bulletin studied by Mason et al.1-4 Unlike the result reported in the bulletin, however, we concluded that the committee's report, published two years after the bulletin, had no discernible effect.

In the same way as Mason et al did, we identified all grommet procedures from hospital episode statistics OPCS-4 surgical code D15.1 for inpatient and day case admissions. Our analysis of the data for all England shows that although the annual rate for all grommet procedures carried out in the population aged under 15 declined significantly between 1992-3 and 1997-8, as Mason et al showed, those procedures carried out specifically for persistent glue ear remained relatively constant.

Limiting the number of different ICD-9 (International Classification of Diseases ninth revision) codes in the hospital episode statistics surgical code D15.1 to those containing over 100 cases in 1992 yielded 49 different codes. In 1994 it yielded 57 different codes. We selected ICD-9 codes 381.2, 381.29, and 381.20, and ICD-10 code H65.3 (for chronic mucoid otitis media (glue ear)) to isolate those procedures carried out on children whose reason for admission was “persistent glue ear” from those carried out for other conditions. The rate of grommet insertion fell by 21% (from 3.3/1000 children aged under 15 in 1992-3 to 2.6/1000 in 1997-8) for glue ear but by 35% (from 4.9/1000 to 3.2/1000) for conditions not coded specifically as glue ear.

The recommendations of the development and evaluation committee's report,1-3 like those of the Effective Health Care bulletin,1-4were specific to the insertion of grommets for persistent glue ear in children aged under 15, and the finding of a greater decline in other conditions suggests that it was not the committee's report that caused the decline. We may have been too restrictive in the codes that we applied to extract the data specific to glue ear, but the impact of the Effective Health Care bulletin on procedures for persistent glue ear would be better understood if the methods that Mason et al used to identify grommet procedures for this condition from others in the routine data were made explicit.

References

  • 1-1.Mason J, Freemantle N, Browning G. Impact of Effective Health Care bulletin on treatment of persistent glue ear in children: time series analysis. BMJ. 2001;323:1096–1097. doi: 10.1136/bmj.323.7321.1096. . (10 October.) [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 1-2.Nicholl J, Coleman P, Touch S, Brennan A, Dixon S. An evaluation of the impact, quality and value of the South and West development and evaluation (DEC) reports. Final report to the NHS Executive South and West. Sheffield: Sheffield School for Health and Related Research (ScHARR); 2000. [Google Scholar]
  • 1-3.Insertion of grommets for glue ear. Southampton: Wessex Institute of Public Health Medicine; 1994. . (South and West development and evaluation (DEC) report No 26.) [Google Scholar]
  • 1-4.The treatment of persistent glue ear in children. Effective Health Care 1992;No 4.
BMJ. 2002 Mar 16;324(7338):673.

Guidelines in Australia were less effective than guidelines in England

Marilyn I Rob 1,2, Johanna I Westbrook 1,2

Editor—Mason et al report that, in the four years after the distribution of an Effective Health Care bulletin, the quarterly rate of grommet insertions fell from 2.1/1000 children under 15 in 1992 (equivalent to an annual rate of 8.4/1000) to 1.5/1000 in 1996 (as calculated from the quoted decrease of 0.044/1000 per quarter).2-1 We adopted a similar method to examine the effect of the introduction of clinical guidelines on rates of myringotomy (with or without insertion of tympanostomy tubes)2-2 among children under 15 in New South Wales.

By 1992-3 (the year before dissemination of the guidelines) the annual myringotomy rate in New South Wales had reached 8.1/1000 children under 15. In the three years immediately after the guidelines were disseminated we too found that annual rates of the procedure fell (to 7.6/1000, 7.4/1000, and 6.9/1000 respectively). In the next years (1996-7, 1997-8, and 1998-9), however, rates rose again to levels similar to those in the period before the guidelines were disseminated (7.8/1000, 7.5/1000, and 7.7/1000 respectively).

Our results with regard to tonsillectomy also differed. Far from displaying an independent trend, rates of tonsillectomy moved almost in tandem with myringotomy rates, falling at the same time and then rising again to their 1992-3 level. We hypothesise that paying attention to one ear, nose, and throat procedure prompted examination of another.

The reason for such conflicting results in the two countries is unknown. If guidelines are solely responsible for the changes observed in clinicians' behaviour then perhaps the failure to sustain decreased rates in New South Wales could be due to differences in the content, presentation, or dissemination of the guideline. Given the literature on the effectiveness of guidelines on clinical practice,2-3 however, we would argue that other factors are likely to have contributed to the outcomes reported by Mason et al.

References

  • 2-1.Mason J, Freemantle N, Browning G. Impact of Effective Health Care bulletin on treatment of persistent glue ear in children: time series analysis. BMJ. 2001;323:1096–1097. doi: 10.1136/bmj.323.7321.1096. . (10 November.) [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2-2.NSW Health Department Working Party. Guidelines on the management of paediatric middle ear disease. Med J Aust. 1993;159(suppl):S1–S8. [PubMed] [Google Scholar]
  • 2-3.Gross P, Greenfield S, Cretin S, Ferguson J, Grimshaw J, Grol R, et al. Optimal methods for guideline implementation: conclusions from Leeds Castle meeting. Med Care. 2001;39(suppl 2):85–92. [PubMed] [Google Scholar]

Articles from BMJ : British Medical Journal are provided here courtesy of BMJ Publishing Group

RESOURCES