Skip to main content

Table 3 Variation in goodness of fit and optimal change point based on different models for the proportion of infective endocarditis cases associated with streptococcal organisms. For each model type, the month of change which gives the best model fit is shown, with best overall models shown in bold italics. AIC measures model goodness of fit (the lower the value the better the fit, within each set of inclusion criteria)

From: Investigation of the impact of the NICE guidelines regarding antibiotic prophylaxis during invasive dental procedures on the incidence of infective endocarditis in England: an electronic health records study

Inclusion criteria

Model type

AIC

Month of change (uncertainty interval)

Proportion of cases with a HES streptococcal code (including mixtures), out of all cases with any HES organism code

Criteria A (Sensitivity/PPV)

One change in trend, with step

1494.8

Oct 2008 (Dec 2005–Oct 2015)

One change in trend, no step

1498.3

Apr 2002 (Oct 1997–Sep 2016)

Single overall trend

1498.7

NA

No trend, with step

1493.4

Oct 2008 (Mar 2006–Apr 2010)

Criteria B (Specificity/PPV)

One change in trend, with step

1415.1

Oct 2015 (Aug 2005–Dec 2015)

One change in trend, no step

1417.0

Sep 2011 (Oct 1997–Sep 2016)

Single overall trend

1416.2

NA

No trend, with step

1413.3

Jun 2013 (Oct 1997–Sep 2016)

Criteria C (Prior study)

One change in trend, with step

1508.4

Dec 2012 (Nov 2012–Oct 2013)

One change in trend, no step

1516.5

Dec 2010 (Oct 1997–Sep 2016)

Single overall trend

1516.6

NA

No trend, with step

1511.6

Dec 2012 (Apr 1999–May 2016)

Proportion of cases matched to an SGSS oral streptococcus sample (including mixtures), out of cases matched to any SGSS organism sample

Criteria A (Sensitivity/PPV)

One change in trend, with step

952.5

Mar 2008 (Mar 2007–Oct 2013)

One change in trend, no step

951.6

Dec 2008 (Jun 2007–Sep 2012)

Single overall trend

970.4

NA

No trend, with step

960.1

Aug 2015 (Apr 2003–Oct 2015)

Criteria B (Specificity/PPV)

One change in trend, with step

894.0

Sep 2013 (Mar 2006–Oct 2015)

One change in trend, no step

892.9

May 2009 (Jan 2007–Feb 2015)

Single overall trend

902.5

NA

No trend, with step

892.6

Aug 2015 (Sep 2013–Nov 2015)

Criteria C (Prior study)

One change in trend, with step

907.3

Oct 2013 (Mar 2007–Oct 2015)

One change in trend, no step

906.6

Jun 2012 (May 2007–Jun 2014)

Single overall trend

919.2

NA

No trend, with step

907.0

Sep 2015 (Sep 2013–Nov 2015)