Distribution of cases
The complete database included 2,123 unique cases of IMD within the studied period Jan 1st 2004 – Dec 31th 2016. Of these, the NRLBM reported 1,968 cases (93%); 1,995 cases (94%) were reported in the OSIRIS notification database. Of the 2,123 unique cases, 1,840 (87%) cases were reported in both databases and could be linked. One-hundred-twenty-eight (6%) cases were only reported in the NRLBM database and 155 (7%) cases only in the OSIRIS database (figure 2).
[figure 2]
Figure 2: Proportial Venn-diagram showing the distribution of IMD cases from the two data sources, the Netherlands 2004-2016.
Of 155 OSIRIS only cases, in 129 cases (85%) some additional information was available in the comments. In 86 of these (67%), the diagnosis was based on PCR without confirmation by positive culture and therefore no isolate was submitted to the NRLBM. For an additional 15 cases (11%) the diagnosis was based on microscopy only and also no isolate was available. In 12 cases (9%) typing of the isolate was performed in another laboratory (mainly in other countries). For 16 cases (12%) it was commented that typing was performed by the NRLBM, but none of the NRLBM-only cases matched with these OSIRIS-only cases.
The percentage of cases that was reported in both databases, ranged per year from 81% (2014) to 91% (2015), but was not significantly different between the years (p-value = 0.8). The distribution of cases over the databases was significantly different for the different age categories (p<0.001). The percentage of cases that were reported in both systems varied from 82% for the group of 65 years and older to 91% for the 0 to 4 year olds. The highest rate of cases that was only reported by the OSIRIS database was seen in the 5 to 14 year olds (12%), the highest percentage that was only reported in the NRLBM database was seen in the group of 65 years and older (14%) (figure 3).
[figure 3]
Figure 3: relative distribution of cases reported in both or one of the databases per age category, IMD cases in the Netherlands, by age category, 2004 – 2016 (N=2,123)
For 2,107 unique cases (99%), the municipality of residence was available in the databases. All 16 cases without information on the place of residence were reported in the NRLBM database only. On RPHC level, the percentage of cases reported in both systems varied from 76% in the region with the lowest proportion, to 98% in the region with the highest proportion (p<0.001).
Data completeness in OSIRIS-database
Of 1,995 cases in the OSIRIS-database, 1,708 (86%) were reported with information on the serogroup. This percentage was higher among the 1,840 cases that were linked with the NRLBM-database, with 1,671 (91%) cases with information on the serogroup. The percentage with a known serogroup varied significantly (p>0.001) over the years, ranging from 74% (2008) to 95% (2013 and 2015). The percentage of cases with information has increased significantly over the years, from 80% in the years in 2004 – 2008, to 94% in the years 2013 – 2016 (p<0.001). Per RPHC, the percentage differed significantly, ranging from 67% to 94% (p>0.001). In the years 2013 – 2016, nine of the 25 RPHS regions had 100% completeness for the serogroup information.
Vaccination status was reported in 1,760 (88%) of the cases. This percentage ranged from 82% (2013) to 91% (2005) over the years (p=0.55). The completeness showed a decreasing trend from the years 2004 – 2008 (90%) to 2013 – 2016 (86%) but this was not significant (p=0.07). Per RPHC region the percentages were significantly different, ranging from 63% to 97% (p<0.001). Four from the 25 RPHSs had 100% completeness for vaccination status in 2013 – 2016.
Information on mortality was entered for 1,972 (99%) of the cases. This percentage was not significantly different between the years (p=0.21) and varied from 97% (2013) to 100% (several years). On RPHC region level the completeness varied from 86% to 100% (p=0.001), although the RPHC with 86% completeness was the only RPHC below 96%. Twelve RPHCs had 100% completeness for information on mortality. The overall case fatality rate was 6% and varied from 3% to 9% in the different years (p=0.32).
The probable country of infection was entered for 1.934 (97%) cases. This percentage also did not vary significantly between the years (p=0.18), with values ranging from 94% (2014) to 99% (2012). On RPHC level, the completeness varied significantly from 87% to 100% (p=0.01), with four RPHSs with a completeness of 100%. The RPHC with 87% completeness, was the only RPHS with a completeness below 92%. For most patients the probable country of infection was the Netherlands (1,885, 95%); only eight patients (0.4%) contracted the disease outside Europe.
A graphical summary of the data completeness over the three different time periods (2004 – 2008, 2009 – 2012 and 2013 – 2016) is shown in figure 4.
[figure 4]
Figure 4: Data completeness for the four key indicators, stratified by time period. Confirmed IMD-cases in the OSIRIS database, the Netherlands, 2004-2016 (N=1,995).
Timeliness
For 1,953 (98%) OSIRIS cases, the time between diagnosis and notification to the RPHS could be calculated. One-hundred-thirty-five (7%) entries with negative and unusually high numbers were defined as missing values, leaving 1,818 entries for analysis. Of these cases, 1,564 (86%) were notified within one working day. The average notification time was 2.3 days (median 0 days; IQR 1 day). Of the 254 cases that were not notified on time, for 65 cases (26%) a possible explanation was reported in the comment field. Of these, 23 cases were accidently not notified by the hospital, 16 cases were notified with a delay due to late laboratory results or due to human or technical error. In 12 cases no notification was made initially as all laboratory tests remained negative and in 9 cases it was unclear whether the clinical picture matched the notification criteria. The remaining five cases were diagnosed in another country. The timeliness improved significantly over the years from 83% in the period 2004 – 2008 to 93% in the years since 2013 (p<0.001). Per RPHS region, the timeliness varied from 64% to 96% (p<0.001). Thirteen RPHS regions had 100% timeliness in the years since 2013. Overall, 99% of the cases were notified to the RPHS within one week after diagnosis since 2013.
The notification time to the national level could be calculated for all 1,995 OSIRIS cases; 1,982 cases were used for analysis after excluding negative or unusually high numbers. Of these, 1,928 (98%) cases were notified within 3 days to the national level. Per year the timeliness varied from 96% to 100% (p=0.45). The timeliness did not differ significantly between the RPHSs (range 92% to 100%, p=0.15). On average, the RPHSs notified cases to the national level in 0.8 days (median 0 days; IQR 0 days). Based on the information on the date of onset reported in OSIRIS, the median time between date of onset and date of confirmation was 3 days (IQR 3 days). Median time between date of onset and notification to local level was 3 days (IQR 3 days) and median time between date of onset and notification to national level was 4 days (IQR 4 days).