- Email: [email protected]
Invasive meningococcal disease: Completeness and timeliness of reporting of confirmed cases in Thames Valley, 2006–2007
Public Health 123 (2009) 805–808
Contents lists available at ScienceDirect
Public Health journal homepage: www.elsevierhealth.com/journals/pubh
Original Research
Invasive meningococcal disease: Completeness and timeliness of reporting of confirmed cases in Thames Valley, 2006–2007 K. Paranthaman a, *, L. Kent a, N. McCarthy a, S.J. Gray b a b
Thames Valley Health Protection Unit, Oxford Science Park, Oxford OX4 4GP, UK Health Protection Agency Meningococcal Reference Unit, Manchester, UK
a r t i c l e i n f o
s u m m a r y
Article history: Received 15 July 2009 Received in revised form 28 September 2009 Accepted 27 October 2009
Objectives: Regular evaluation of disease surveillance systems is essential. This study assessed the completeness and timeliness of reporting of invasive meningococcal disease (IMD) in Thames Valley in 2006–2007. Study design: Retrospective review of two data sources used in disease surveillance: the list of notified cases to the Thames Valley Health Protection Unit (TVHPU) and the list of confirmed cases at the reference laboratory during 2006–2007. Methods: The datasets were compared by checking patient name, date of birth, sample date and date of onset of illness. Completeness was estimated using Tilling’s capture–recapture method. Timeliness was assessed by calculating the difference between the date of admission and the date of notification to the TVHPU. Results: The estimated completeness of reporting of IMD cases was calculated as 90.5% (95% confidence interval 88.6–92.4). Thirty-six percent of cases were notified on the day of admission, 63% were notified within 1 day and 72% were notified within 2 days (range 0–36 days). Conclusions: Timeliness and completeness of reporting of IMD was clearly suboptimal. It is critical to educate clinicians on the need to notify all suspected cases of IMD to public health authorities in a timely manner. Ó 2009 The Royal Society for Public Health. Published by Elsevier Ltd. All rights reserved.
Keywords: Meningococcal infections Disease notification Population surveillance Retrospective studies Time factors
Introduction Invasive meningococcal disease (IMD) remains a significant cause of mortality and morbidity, with over 1200 laboratoryconfirmed cases every year in England and Wales.1 Timely institution of prevention and control measures such as chemoprophylaxis and vaccination of close contacts is dependent on the completeness and timeliness of statutory notifications of suspected cases of IMD to public health authorities.2 To enable prompt public health actions, healthcare professionals are statutorily required to immediately notify clinically suspected cases of IMD to the ‘proper officer’ of the local authority under the Public Health Act 1984.3 However, all disease surveillance systems have particular strengths and limitations, leading to varying degrees of confidence in their validity and reliability. Regular evaluation is essential to confirm their public health utility, to understand the limitations and to identify priority areas for improvement.4 This study aimed to evaluate the completeness and timeliness of notifications of
* Corresponding author. Tel.: þ44 1865765144. E-mail address: [email protected] (K. Paranthaman).
confirmed IMD among residents in the Thames Valley region (covering Berkshire, Buckinghamshire and Oxfordshire; population ~2.1 million) over a 2-year period. Methods Data on confirmed IMD cases occurring between 1 January 2006 and 31 December 2007 were compared from two sources: records of notifications of clinically confirmed IMD cases to the Thames Valley Health Protection Unit (TVHPU) office, and the list of confirmed cases with samples referred to the Health Protection Agency Meningococcal Reference Unit (MRU), Manchester. Definition of cases All patients who were permanent or temporary residents of the three counties at the time of onset of illness with confirmed IMD were included. For the purpose of this study, the definition for a confirmed case was the same as the epidemiological definition used nationally: Neisseria meningitidis was isolated from blood, cerebrospinal fluid or other deep tissue; Gram-negative diplococci were seen in
0033-3506/$ – see front matter Ó 2009 The Royal Society for Public Health. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.puhe.2009.10.015
806
K. Paranthaman et al. / Public Health 123 (2009) 805–808
cerebrospinal fluid; or clinical signs of meningitis or septicaemia were accompanied by positive serology or by positive results of a polymerase chain reaction (PCR) test on cerebrospinal fluid or blood.2 TVHPU dataset The complete list of cases of IMD notified to the TVHPU office was extracted from the Notifications of Infectious Diseases surveillance system database, from which confirmed cases were identified. Manual searching of paper records and electronic files was undertaken to ensure completeness of the patient list. To assess timeliness of reporting, data from completed enhanced surveillance forms were entered on to a Microsoft Excel spreadsheet. MRU dataset The reference laboratory at MRU receives clinical samples of suspected cases for PCR investigations, and isolates of cultureconfirmed cases from acute hospitals across the country for confirmation and characterization. The list of confirmed cases from samples sent from acute hospitals in the Thames Valley region was extracted from the MRU dataset for the same 2-year period. As the catchment area of acute hospitals might vary from that of the TVHPU catchment area and due to incomplete recording of residence data on the MRU records, non-residents (i.e. not living within the TVHPU catchment area) were excluded by checking on the National Health Service Open Exeter patient database. Finally, duplicate records and duplicate samples were removed. Statistical analysis The datasets were compared by checking patient name, date of birth, sample date and date of onset of illness. Confirmed cases were those who fulfilled the criteria for laboratory confirmation as given in the case definition above. Cross-checking and matching of cases in both datasets were performed manually. The completeness of reporting was estimated using the capture–recapture method described by Tilling, which provides an estimate of the total number of reported and unreported cases in the community by comparing cases in either and both systems.5 Cases that could not have been captured by either system, primarily non-residents, were excluded. Confirmed cases that were not notified directly to the TVHPU by the usual notification systems, but were reported by the MRU to the TVHPU following laboratory confirmation of diagnosis were excluded when assessing completeness of notifications. Timeliness of notification was assessed by calculating the difference between the date of admission to hospital and the date of notification to the TVHPU from enhanced surveillance forms completed for each case. In most cases, the actual times of admission and notification were not recorded, so analysis was restricted to any difference in the number of calendar days. The job designation of the person notifying the case was extracted to identify if they constituted statutory or other forms of ascertainment. To assess whether a case was notified during office hours or to the outof-hours (OOH) team, information on the person completing the form (TVHPU or OOH staff), time of notification where recorded and other notes made in the case record were used. Results Completeness of notifications Seventy-seven confirmed cases were present in both the MRU and the TVHPU datasets. The TVHPU dataset had 79 cases,
including two positive cases not present in the MRU dataset, that fulfilled the criteria for confirmed case definition. One case had gram-negative diplococci on cerebrospinal fluid, but the PCR test performed by the MRU was negative; and the other case was confirmed by PCR test at the local hospital, but no sample was sent to the MRU. In addition to the 77 cases present in both datasets, eight cases were confirmed at the MRU and were not directly notified by clinicians to the TVHPU. Overall, 87 cases were confirmed by laboratory investigations (Table 1). The TVHPU was notified of 79 of the 87.2 confirmed cases, thus giving an estimated completeness of reporting of 90.5% (95% confidence interval 88.6–92.4). Timeliness of reporting Assessment of the timeliness of notification of confirmed cases from completed enhanced surveillance forms revealed that 36% were notified on the day of admission, 63% were notified within 1 day and 72% were notified within 2 days (range 0–36 days). Twenty-six cases were laboratory confirmed at the time of notification to the TVHPU. Among the 85 confirmed cases notified to the TVHPU, 53 cases were notified to the TVHPU office which operates 40 h/week, 24 cases were notified to the OOH team which covers the remaining 128 h/week, and information was unclear in eight cases. The time of notification to public health authorities is dependent on the time of presentation of the case to health services, which was not recorded adequately in the enhanced surveillance forms, as mentioned previously. However, considering the longer duration of operation of the OOH team and the fact that IMD cases present to health services as emergency admissions without following strict time patterns, the low notification rate (24/77, 31.2%) to the OOH team suggests that clinicians delayed notification until office hours. Of the 87 cases, 67 were notified formally by clinicians, four were notified by other persons, six were notified by laboratory reporting after diagnosis confirmation by the MRU, and information was unclear in eight cases. The remaining two cases found in the MRU dataset were not notified to the TVHPU by statutory or laboratory reporting. Secondary cases Secondary cases may indicate onward transmission to exposed close contacts due to a failure or delay in reporting of the index case. During this 2-year period, one case was linked to a previously
Table 1 Estimated completenessb of reporting for confirmed meningococcal disease using Tilling’s capture–recapture method: Thames Valley, 2006–2007. TVHPU dataset
MRU dataset
Cases reported Cases not reported All cases
Cases reported
Cases not reported
All cases
77a 2 (C) 79 (T1)
8 (B) 0.2 (X) 8.2 (B þ X)
85 (T2) 2.2 (C þ X) 87.2 (N)
TVHPU, Thames Valley Health Protection Unit; MRU, Meningococcal Reference Unit. A, number of cases identified in both TVHPU and MRU datasets; B, number of cases only identified in MRU dataset; C, number of cases only identified in TVHPU dataset; X, estimated number of confirmed cases not present in TVHPU or MRU datasets, calculated as X¼(B C)/A; T2, number of cases identified in MRU dataset; T1, number of cases identified in TVHPU dataset; N, estimated total number of cases of meningococcal disease in Thames Valley region during 2006–2007, calculated as N¼A þ B þ C þ X. b Completeness of reporting was calculated as T1/N and the 95% confidence interval was calculated as T1/N þ 1.96 Var (N)1/2 where Var (N)¼(T1 T2 x B C)/A3.
K. Paranthaman et al. / Public Health 123 (2009) 805–808
confirmed case. The index case was notified promptly; however, the secondary case did not fulfil the criteria for close contact and was therefore not given prophylaxis. Further information on this unusual transmission has been described elsewhere.6 Discussion This study assessed the completeness and timeliness of reporting of confirmed IMD cases to the public health authorities in the Thames Valley region in 2006–2007. The majority of confirmed cases were notified to the TVHPU by statutory notification or other means of ascertainment. However, 10% of cases confirmed at the MRU were not notified by clinicians. Three-quarters of confirmed cases were notified formally by clinicians, while the rest were ascertained by other methods such as laboratory reporting or lay public enquiries. Previous studies in England have documented a variable notification rate of 40–95% using different approaches such as retrospective case ascertainment, case comparison and capture–recapture methods, and this ascertainment rate is in line with similar countries in Europe.7–12 It must be recognized that laboratory-confirmed cases are an underestimate of the true burden of the disease in the community.7 Laboratory confirmation of IMD may not always be possible due to the lack of a clinical specimen, prior antibiotic therapy or limited sensitivity of available diagnostic tests.13–15 Inclusion of clinically diagnosed cases without laboratory confirmation by extracting Hospital Episodes Statistics appropriately coded for meningococcal disease might be useful to assess completeness, although they are likely to include a substantial proportion of non-cases.11,16 Timeliness of reporting of confirmed cases was unsatisfactory, with 64% of cases notified to the TVHPU at least 1 day after admission to hospital. This finding is of great concern because interventions such as chemoprophylaxis and vaccination are most effective if given early after diagnosis in the index case.2,17 A previous study found that 91% of cases were notified within 24 h, although this is not directly comparable to this study due to lack of information on time of presentation and notification.11 Avoidable delays might result in secondary cases of this serious disease where outcomes can be fatal. Although it is reassuring to note the absence of avoidable secondary cases in this study, this should not be a reason for complacency. The delays in notification, together with the failure of clinicians to notify one-quarter of confirmed cases, emphasize the importance of early notification of suspected cases to public health authorities. A prospective study to evaluate the timeliness of notification of IMD cases with capture of the time of presentation to hospital and notification and the factors leading to delays in notification would provide a better understanding of the problem, and will inform targeted approaches to improve timely notification. Six confirmed cases were reported to the TVHPU by the MRU leading to appropriate public health actions, which would otherwise have been missed completely, although two further cases were not reported. Despite the inherent delays caused by transport of specimen to the reference laboratory and diagnosis confirmation, this practice of laboratory reporting identified six additional cases that were not ascertained by traditional surveillance and led to institution of appropriate public health interventions. Indeed, laboratory reporting of infectious diseases has been shown to improve completeness and timeliness of surveillance systems.18,19 This mechanism of back reporting of confirmed cases to the relevant health protection unit by the MRU needs to be strengthened. This study has some limitations. Capture–recapture analysis is only reliable when the two datasets are independent of each other, and each individual in the catchment population has an equal chance of capture without any bias towards particular groups. In
807
this instance, the two databases are not entirely independent as the TVHPU and the MRU are important partners in disease control, particularly during outbreaks. The MRU data also showed (data not presented in this paper) that children were more likely to be tested but not notified, and therefore might have a higher likelihood of being on the MRU dataset. Another possibility that might lead to overestimation of completeness is that some cases might have been confirmed at local hospitals without being notified to the TVHPU or samples referred to the MRU. However, detection of such cases would require a complete review of all laboratory results for confirmed IMD cases from all acute hospitals in the region. This study did not include cases that might have been referred to neighbouring hospitals. However, this is unlikely to be significant enough to bias the results and conclusions. Despite these limitations, capture–recapture methods provide a rough estimate of completeness and are useful to understand the utility of single sources of surveillance data. The lack of complete information on the time of case presentation to health services and notification to public health authorities meant that timeliness could only be assessed as the difference in the number of days between the two events. In addition, the study assessed the timeliness of notification of confirmed cases of IMD alone, and therefore suspected but unconfirmed cases were excluded. Conclusions This study has shown that the timeliness and completeness of reporting of confirmed cases of IMD were suboptimal. Despite the risk of failure in achieving full compliance, it is critical to educate clinicians on the need to notify all suspected cases of IMD to the public health authorities in a timely manner. Laboratory reporting of confirmed IMD cases to health protection units identified additional cases that were not notified by the usual methods of case ascertainment. In light of this, it would be appropriate to consider the need for wider implementation of this practice.
Ethical approval None sought.
Funding None declared.
Competing interests None declared. References 1. Health Protection Agency. Laboratory confirmed cases of all invasive meningococcal disease by age and calendar year, England and Wales 1999 to 2008 (provisional). Available at: http://www.hpa.org.uk/webw/HPAweb&HPAwebStandard/HPAweb_C/ 1234510034636?p¼1201094595391; 2009 (last accessed 30/06/2009). 2. Health Protection Agency Meningococcus Forum. Guidance for public health management of meningococcal disease in the UK. Available at: http://www. hpa.org.uk/web/HPAwebFile/HPAweb_C/1194947389261; 2006 (last accessed 30/06/2009). 3. Public Health (Control of Disease) Act. London: HMSO; 1984. 4. Teutsch SM, Churchill RE. Principles and practice of public health surveillance. New York: Oxford University Press; 1994. 5. Tilling K. Capture–recapture methods – useful or misleading? Int J Epidemiol 2001;30:12–4. 6. Petsas A, Sharma A, Aghadiuno O, Abid M, Paranthaman K. A secondary case of meningococcal disease in an ambulance worker, Berkshire, November 2007. Euro Surveill 2008;13. pii ¼ 8020.
808
K. Paranthaman et al. / Public Health 123 (2009) 805–808
7. Davies LA. Assessing the value of different sources of information on meningococcal disease. Community Med 1989;11:239–46. 8. Wylie PA, Stevens D, Drake 3rd W, Stuart J, Cartwright K. Epidemiology and clinical management of meningococcal disease in west Gloucestershire: retrospective, population based study. BMJ 1997;315:774–9. 9. Fortnum H, Mason K. Reliability of notification data for childhood bacterial meningitis. J Public Health Med 1993;15:315–9. 10. Davison KL, Crowcroft NS, Ramsay ME, Begg NT, Kaczmarksi EB, Stuart JM, White JM, Orr H. Enhanced Surveillance of Meningococcal Disease Project Group. Enhanced surveillance scheme for suspected meningococcal disease in five regional health authorities in England. Commun Dis Public Health 1998;2002(5):205–12. 11. Breen E, Ghebrehewet S, Regan M, Thomson AP. How complete and accurate is meningococcal disease notification? Commun Dis Public Health 2004;7:334–8. 12. Trotter C, Samuelsson S, Perrocheau A, de Greeff S, de Melker H, Heuberger S, Ramsay M. Ascertainment of meningococcal disease in Europe. Euro Surveill 2005;10. pii ¼ 586.
13. Cartwright K, Reilly S, White D, Stuart J. Early treatment with parenteral penicillin in meningococcal disease. BMJ 1992;305:143–7. 14. Nadel S, Levin M, Habibi P. Treatment of meningococcal disease in childhood. In: Cartwright K, editor. Meningococcal disease. John Wiley & Sons; 1995. p. 207–43. 15. Bohr V, Rasmussen N, Hansen B, Kjersem H, Jessen O, Johnsen N, Kristensen HS. 875 cases of bacterial meningitis: diagnostic procedures and the impact of preadmission antibiotic therapy. Part III of a three part series. J Infect 1983;7:193–202. 16. Ackman DM, Birkhead G, Flynn M. Assessment of surveillance for meningococcal disease in New York stat. Am J Epidemiol 1991;1996(144):78–82. 17. Fraser A, Gafter-Gvili A, Paul M, Leibovici L. Antibiotics for preventing meningococcal infections. Cochrane Database System Rev 2006;4. CD004785. 18. Nguyen TQ, Thorpe L, Makki HA, Mostashari F. Benefits and barriers to electronic laboratory results reporting for notifiable diseases: the New York City Department of health and mental hygiene experience. Am J Public Health 2007;97(Suppl. 1):S142–5. 19. Wurtz R, Cameron BJ. Clin Infect Dis 2005;40:1638–43.
Contents lists available at ScienceDirect
Public Health journal homepage: www.elsevierhealth.com/journals/pubh
Original Research
Invasive meningococcal disease: Completeness and timeliness of reporting of confirmed cases in Thames Valley, 2006–2007 K. Paranthaman a, *, L. Kent a, N. McCarthy a, S.J. Gray b a b
Thames Valley Health Protection Unit, Oxford Science Park, Oxford OX4 4GP, UK Health Protection Agency Meningococcal Reference Unit, Manchester, UK
a r t i c l e i n f o
s u m m a r y
Article history: Received 15 July 2009 Received in revised form 28 September 2009 Accepted 27 October 2009
Objectives: Regular evaluation of disease surveillance systems is essential. This study assessed the completeness and timeliness of reporting of invasive meningococcal disease (IMD) in Thames Valley in 2006–2007. Study design: Retrospective review of two data sources used in disease surveillance: the list of notified cases to the Thames Valley Health Protection Unit (TVHPU) and the list of confirmed cases at the reference laboratory during 2006–2007. Methods: The datasets were compared by checking patient name, date of birth, sample date and date of onset of illness. Completeness was estimated using Tilling’s capture–recapture method. Timeliness was assessed by calculating the difference between the date of admission and the date of notification to the TVHPU. Results: The estimated completeness of reporting of IMD cases was calculated as 90.5% (95% confidence interval 88.6–92.4). Thirty-six percent of cases were notified on the day of admission, 63% were notified within 1 day and 72% were notified within 2 days (range 0–36 days). Conclusions: Timeliness and completeness of reporting of IMD was clearly suboptimal. It is critical to educate clinicians on the need to notify all suspected cases of IMD to public health authorities in a timely manner. Ó 2009 The Royal Society for Public Health. Published by Elsevier Ltd. All rights reserved.
Keywords: Meningococcal infections Disease notification Population surveillance Retrospective studies Time factors
Introduction Invasive meningococcal disease (IMD) remains a significant cause of mortality and morbidity, with over 1200 laboratoryconfirmed cases every year in England and Wales.1 Timely institution of prevention and control measures such as chemoprophylaxis and vaccination of close contacts is dependent on the completeness and timeliness of statutory notifications of suspected cases of IMD to public health authorities.2 To enable prompt public health actions, healthcare professionals are statutorily required to immediately notify clinically suspected cases of IMD to the ‘proper officer’ of the local authority under the Public Health Act 1984.3 However, all disease surveillance systems have particular strengths and limitations, leading to varying degrees of confidence in their validity and reliability. Regular evaluation is essential to confirm their public health utility, to understand the limitations and to identify priority areas for improvement.4 This study aimed to evaluate the completeness and timeliness of notifications of
* Corresponding author. Tel.: þ44 1865765144. E-mail address: [email protected] (K. Paranthaman).
confirmed IMD among residents in the Thames Valley region (covering Berkshire, Buckinghamshire and Oxfordshire; population ~2.1 million) over a 2-year period. Methods Data on confirmed IMD cases occurring between 1 January 2006 and 31 December 2007 were compared from two sources: records of notifications of clinically confirmed IMD cases to the Thames Valley Health Protection Unit (TVHPU) office, and the list of confirmed cases with samples referred to the Health Protection Agency Meningococcal Reference Unit (MRU), Manchester. Definition of cases All patients who were permanent or temporary residents of the three counties at the time of onset of illness with confirmed IMD were included. For the purpose of this study, the definition for a confirmed case was the same as the epidemiological definition used nationally: Neisseria meningitidis was isolated from blood, cerebrospinal fluid or other deep tissue; Gram-negative diplococci were seen in
0033-3506/$ – see front matter Ó 2009 The Royal Society for Public Health. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.puhe.2009.10.015
806
K. Paranthaman et al. / Public Health 123 (2009) 805–808
cerebrospinal fluid; or clinical signs of meningitis or septicaemia were accompanied by positive serology or by positive results of a polymerase chain reaction (PCR) test on cerebrospinal fluid or blood.2 TVHPU dataset The complete list of cases of IMD notified to the TVHPU office was extracted from the Notifications of Infectious Diseases surveillance system database, from which confirmed cases were identified. Manual searching of paper records and electronic files was undertaken to ensure completeness of the patient list. To assess timeliness of reporting, data from completed enhanced surveillance forms were entered on to a Microsoft Excel spreadsheet. MRU dataset The reference laboratory at MRU receives clinical samples of suspected cases for PCR investigations, and isolates of cultureconfirmed cases from acute hospitals across the country for confirmation and characterization. The list of confirmed cases from samples sent from acute hospitals in the Thames Valley region was extracted from the MRU dataset for the same 2-year period. As the catchment area of acute hospitals might vary from that of the TVHPU catchment area and due to incomplete recording of residence data on the MRU records, non-residents (i.e. not living within the TVHPU catchment area) were excluded by checking on the National Health Service Open Exeter patient database. Finally, duplicate records and duplicate samples were removed. Statistical analysis The datasets were compared by checking patient name, date of birth, sample date and date of onset of illness. Confirmed cases were those who fulfilled the criteria for laboratory confirmation as given in the case definition above. Cross-checking and matching of cases in both datasets were performed manually. The completeness of reporting was estimated using the capture–recapture method described by Tilling, which provides an estimate of the total number of reported and unreported cases in the community by comparing cases in either and both systems.5 Cases that could not have been captured by either system, primarily non-residents, were excluded. Confirmed cases that were not notified directly to the TVHPU by the usual notification systems, but were reported by the MRU to the TVHPU following laboratory confirmation of diagnosis were excluded when assessing completeness of notifications. Timeliness of notification was assessed by calculating the difference between the date of admission to hospital and the date of notification to the TVHPU from enhanced surveillance forms completed for each case. In most cases, the actual times of admission and notification were not recorded, so analysis was restricted to any difference in the number of calendar days. The job designation of the person notifying the case was extracted to identify if they constituted statutory or other forms of ascertainment. To assess whether a case was notified during office hours or to the outof-hours (OOH) team, information on the person completing the form (TVHPU or OOH staff), time of notification where recorded and other notes made in the case record were used. Results Completeness of notifications Seventy-seven confirmed cases were present in both the MRU and the TVHPU datasets. The TVHPU dataset had 79 cases,
including two positive cases not present in the MRU dataset, that fulfilled the criteria for confirmed case definition. One case had gram-negative diplococci on cerebrospinal fluid, but the PCR test performed by the MRU was negative; and the other case was confirmed by PCR test at the local hospital, but no sample was sent to the MRU. In addition to the 77 cases present in both datasets, eight cases were confirmed at the MRU and were not directly notified by clinicians to the TVHPU. Overall, 87 cases were confirmed by laboratory investigations (Table 1). The TVHPU was notified of 79 of the 87.2 confirmed cases, thus giving an estimated completeness of reporting of 90.5% (95% confidence interval 88.6–92.4). Timeliness of reporting Assessment of the timeliness of notification of confirmed cases from completed enhanced surveillance forms revealed that 36% were notified on the day of admission, 63% were notified within 1 day and 72% were notified within 2 days (range 0–36 days). Twenty-six cases were laboratory confirmed at the time of notification to the TVHPU. Among the 85 confirmed cases notified to the TVHPU, 53 cases were notified to the TVHPU office which operates 40 h/week, 24 cases were notified to the OOH team which covers the remaining 128 h/week, and information was unclear in eight cases. The time of notification to public health authorities is dependent on the time of presentation of the case to health services, which was not recorded adequately in the enhanced surveillance forms, as mentioned previously. However, considering the longer duration of operation of the OOH team and the fact that IMD cases present to health services as emergency admissions without following strict time patterns, the low notification rate (24/77, 31.2%) to the OOH team suggests that clinicians delayed notification until office hours. Of the 87 cases, 67 were notified formally by clinicians, four were notified by other persons, six were notified by laboratory reporting after diagnosis confirmation by the MRU, and information was unclear in eight cases. The remaining two cases found in the MRU dataset were not notified to the TVHPU by statutory or laboratory reporting. Secondary cases Secondary cases may indicate onward transmission to exposed close contacts due to a failure or delay in reporting of the index case. During this 2-year period, one case was linked to a previously
Table 1 Estimated completenessb of reporting for confirmed meningococcal disease using Tilling’s capture–recapture method: Thames Valley, 2006–2007. TVHPU dataset
MRU dataset
Cases reported Cases not reported All cases
Cases reported
Cases not reported
All cases
77a 2 (C) 79 (T1)
8 (B) 0.2 (X) 8.2 (B þ X)
85 (T2) 2.2 (C þ X) 87.2 (N)
TVHPU, Thames Valley Health Protection Unit; MRU, Meningococcal Reference Unit. A, number of cases identified in both TVHPU and MRU datasets; B, number of cases only identified in MRU dataset; C, number of cases only identified in TVHPU dataset; X, estimated number of confirmed cases not present in TVHPU or MRU datasets, calculated as X¼(B C)/A; T2, number of cases identified in MRU dataset; T1, number of cases identified in TVHPU dataset; N, estimated total number of cases of meningococcal disease in Thames Valley region during 2006–2007, calculated as N¼A þ B þ C þ X. b Completeness of reporting was calculated as T1/N and the 95% confidence interval was calculated as T1/N þ 1.96 Var (N)1/2 where Var (N)¼(T1 T2 x B C)/A3.
K. Paranthaman et al. / Public Health 123 (2009) 805–808
confirmed case. The index case was notified promptly; however, the secondary case did not fulfil the criteria for close contact and was therefore not given prophylaxis. Further information on this unusual transmission has been described elsewhere.6 Discussion This study assessed the completeness and timeliness of reporting of confirmed IMD cases to the public health authorities in the Thames Valley region in 2006–2007. The majority of confirmed cases were notified to the TVHPU by statutory notification or other means of ascertainment. However, 10% of cases confirmed at the MRU were not notified by clinicians. Three-quarters of confirmed cases were notified formally by clinicians, while the rest were ascertained by other methods such as laboratory reporting or lay public enquiries. Previous studies in England have documented a variable notification rate of 40–95% using different approaches such as retrospective case ascertainment, case comparison and capture–recapture methods, and this ascertainment rate is in line with similar countries in Europe.7–12 It must be recognized that laboratory-confirmed cases are an underestimate of the true burden of the disease in the community.7 Laboratory confirmation of IMD may not always be possible due to the lack of a clinical specimen, prior antibiotic therapy or limited sensitivity of available diagnostic tests.13–15 Inclusion of clinically diagnosed cases without laboratory confirmation by extracting Hospital Episodes Statistics appropriately coded for meningococcal disease might be useful to assess completeness, although they are likely to include a substantial proportion of non-cases.11,16 Timeliness of reporting of confirmed cases was unsatisfactory, with 64% of cases notified to the TVHPU at least 1 day after admission to hospital. This finding is of great concern because interventions such as chemoprophylaxis and vaccination are most effective if given early after diagnosis in the index case.2,17 A previous study found that 91% of cases were notified within 24 h, although this is not directly comparable to this study due to lack of information on time of presentation and notification.11 Avoidable delays might result in secondary cases of this serious disease where outcomes can be fatal. Although it is reassuring to note the absence of avoidable secondary cases in this study, this should not be a reason for complacency. The delays in notification, together with the failure of clinicians to notify one-quarter of confirmed cases, emphasize the importance of early notification of suspected cases to public health authorities. A prospective study to evaluate the timeliness of notification of IMD cases with capture of the time of presentation to hospital and notification and the factors leading to delays in notification would provide a better understanding of the problem, and will inform targeted approaches to improve timely notification. Six confirmed cases were reported to the TVHPU by the MRU leading to appropriate public health actions, which would otherwise have been missed completely, although two further cases were not reported. Despite the inherent delays caused by transport of specimen to the reference laboratory and diagnosis confirmation, this practice of laboratory reporting identified six additional cases that were not ascertained by traditional surveillance and led to institution of appropriate public health interventions. Indeed, laboratory reporting of infectious diseases has been shown to improve completeness and timeliness of surveillance systems.18,19 This mechanism of back reporting of confirmed cases to the relevant health protection unit by the MRU needs to be strengthened. This study has some limitations. Capture–recapture analysis is only reliable when the two datasets are independent of each other, and each individual in the catchment population has an equal chance of capture without any bias towards particular groups. In
807
this instance, the two databases are not entirely independent as the TVHPU and the MRU are important partners in disease control, particularly during outbreaks. The MRU data also showed (data not presented in this paper) that children were more likely to be tested but not notified, and therefore might have a higher likelihood of being on the MRU dataset. Another possibility that might lead to overestimation of completeness is that some cases might have been confirmed at local hospitals without being notified to the TVHPU or samples referred to the MRU. However, detection of such cases would require a complete review of all laboratory results for confirmed IMD cases from all acute hospitals in the region. This study did not include cases that might have been referred to neighbouring hospitals. However, this is unlikely to be significant enough to bias the results and conclusions. Despite these limitations, capture–recapture methods provide a rough estimate of completeness and are useful to understand the utility of single sources of surveillance data. The lack of complete information on the time of case presentation to health services and notification to public health authorities meant that timeliness could only be assessed as the difference in the number of days between the two events. In addition, the study assessed the timeliness of notification of confirmed cases of IMD alone, and therefore suspected but unconfirmed cases were excluded. Conclusions This study has shown that the timeliness and completeness of reporting of confirmed cases of IMD were suboptimal. Despite the risk of failure in achieving full compliance, it is critical to educate clinicians on the need to notify all suspected cases of IMD to the public health authorities in a timely manner. Laboratory reporting of confirmed IMD cases to health protection units identified additional cases that were not notified by the usual methods of case ascertainment. In light of this, it would be appropriate to consider the need for wider implementation of this practice.
Ethical approval None sought.
Funding None declared.
Competing interests None declared. References 1. Health Protection Agency. Laboratory confirmed cases of all invasive meningococcal disease by age and calendar year, England and Wales 1999 to 2008 (provisional). Available at: http://www.hpa.org.uk/webw/HPAweb&HPAwebStandard/HPAweb_C/ 1234510034636?p¼1201094595391; 2009 (last accessed 30/06/2009). 2. Health Protection Agency Meningococcus Forum. Guidance for public health management of meningococcal disease in the UK. Available at: http://www. hpa.org.uk/web/HPAwebFile/HPAweb_C/1194947389261; 2006 (last accessed 30/06/2009). 3. Public Health (Control of Disease) Act. London: HMSO; 1984. 4. Teutsch SM, Churchill RE. Principles and practice of public health surveillance. New York: Oxford University Press; 1994. 5. Tilling K. Capture–recapture methods – useful or misleading? Int J Epidemiol 2001;30:12–4. 6. Petsas A, Sharma A, Aghadiuno O, Abid M, Paranthaman K. A secondary case of meningococcal disease in an ambulance worker, Berkshire, November 2007. Euro Surveill 2008;13. pii ¼ 8020.
808
K. Paranthaman et al. / Public Health 123 (2009) 805–808
7. Davies LA. Assessing the value of different sources of information on meningococcal disease. Community Med 1989;11:239–46. 8. Wylie PA, Stevens D, Drake 3rd W, Stuart J, Cartwright K. Epidemiology and clinical management of meningococcal disease in west Gloucestershire: retrospective, population based study. BMJ 1997;315:774–9. 9. Fortnum H, Mason K. Reliability of notification data for childhood bacterial meningitis. J Public Health Med 1993;15:315–9. 10. Davison KL, Crowcroft NS, Ramsay ME, Begg NT, Kaczmarksi EB, Stuart JM, White JM, Orr H. Enhanced Surveillance of Meningococcal Disease Project Group. Enhanced surveillance scheme for suspected meningococcal disease in five regional health authorities in England. Commun Dis Public Health 1998;2002(5):205–12. 11. Breen E, Ghebrehewet S, Regan M, Thomson AP. How complete and accurate is meningococcal disease notification? Commun Dis Public Health 2004;7:334–8. 12. Trotter C, Samuelsson S, Perrocheau A, de Greeff S, de Melker H, Heuberger S, Ramsay M. Ascertainment of meningococcal disease in Europe. Euro Surveill 2005;10. pii ¼ 586.
13. Cartwright K, Reilly S, White D, Stuart J. Early treatment with parenteral penicillin in meningococcal disease. BMJ 1992;305:143–7. 14. Nadel S, Levin M, Habibi P. Treatment of meningococcal disease in childhood. In: Cartwright K, editor. Meningococcal disease. John Wiley & Sons; 1995. p. 207–43. 15. Bohr V, Rasmussen N, Hansen B, Kjersem H, Jessen O, Johnsen N, Kristensen HS. 875 cases of bacterial meningitis: diagnostic procedures and the impact of preadmission antibiotic therapy. Part III of a three part series. J Infect 1983;7:193–202. 16. Ackman DM, Birkhead G, Flynn M. Assessment of surveillance for meningococcal disease in New York stat. Am J Epidemiol 1991;1996(144):78–82. 17. Fraser A, Gafter-Gvili A, Paul M, Leibovici L. Antibiotics for preventing meningococcal infections. Cochrane Database System Rev 2006;4. CD004785. 18. Nguyen TQ, Thorpe L, Makki HA, Mostashari F. Benefits and barriers to electronic laboratory results reporting for notifiable diseases: the New York City Department of health and mental hygiene experience. Am J Public Health 2007;97(Suppl. 1):S142–5. 19. Wurtz R, Cameron BJ. Clin Infect Dis 2005;40:1638–43.