Skip to main content
. 2014 Mar 18;143(1):189–201. doi: 10.1017/S0950268814000533

Table 2.

Count regression model selection for the association between deprivation and incidence of infectious diseases in the North East of England, 2007–2011

Group Dataset Mean count (variance) Zero counts (%) Poisson model ZIP model Negative binomial model ZINB model Final model Residuals
GOF AIC Vuong AIC ΔAIC* LRT AIC Ref. Δ AIC* Vuong AIC Ref. ΔAIC*
BBV HBV infection 0·5 (1·20) 73·13 0·5974 2688·15 0·0129 2646·69 41·46 <0·0001 2640·09 ZIP 6·5950 0·2645 2609·86 NB 30·2310 NB Bimodal
HCV infection 0·6 (1·78) 65·70 <0·0001 3297·41 0·0001 3208·84 88·57 <0·0001 3163·59 ZIP 45·2500 Did not converge NB Bimodal
HIV infection 0·1 (0·17) 89·79 1·0000 1221·36 0·0359 1209·11 12·25 0·0090 1215·70 ZIP −6·5900 <0·0001 1209·11 ZIP 0·0000 ZIP Normal
IBD Invasive GAS infection 0·2 (0·16) 85·08 1·0000 1547·07 0·2220 1546·04 1·03 0·5000 1547·07 Poisson 0·0000 n.a. 1546·04 Poisson 1·0290 Poisson Bimodal
Invasive meningococcal infection 0·2 (0·2) 80·92 1·0000 1833·24 0·4435 1833·14 0·10 0·2410 1897·69 Poisson −64·4500 Did not converge Poisson Bimodal
Invasive pneumococcal infection 0·8 (0·88) 47·77 0·0004 3824·18 0·2869 3820·83 3·35 0·1130 3820·71 Poisson 3·4660 0·5002 3822·71 NB −2·0000 NB Bimodal
IID Campylobacteriosis 9·1 (16·90) 0·06 <0·0001 8968·73 0·5015 8970·73 −2·00 <0·0001 8851·94 Poisson 116·7970 0·6873 8853·94 NB −2·0000 NB Normal
Cryptosporidiosis 0·6 (0·93) 57·67 <0·0001 3701·46 0·0004 3643·18 58·27 <0·0001 3635·16 ZIP 8·0190 Did not converge NB Bimodal
E. coli O157 infection 0·2 (0·31) 84·72 1·0000 1849·14 0·0002 1769·68 79·45 <0·0001 1772·10 ZIP −2·4170 0·1860 1767·69 ZIP 1·9950 ZIP Normal
Giardiasis 0·1 (0·13) 89·79 1·0000 1181·32 0·2797 1179·32 2·01 0·0750 1179·25 Poisson 2·0710 Did not converge Poisson Bimodal
Salmonellosis 1·3 (1·68) 30·56 <0·0001 4954·11 0·0639 4944·17 9·94 <0·0001 4932·55 Poisson 21·5590 0·3544 4931·58 NB 0·9740 NB Bimodal
Shigellosis 0·1 (0·17) 89·19 1·0000 1322·53 0·0375 1303·02 19·51 <0·0001 1304·63 ZIP −1·6160 <0·0001 1303·02 ZIP 0·0000 ZIP Normal
STI Chlamydia infection 16·3 (114·75) 0·12 <0·0001 12286·15 0·2136 12402·45 −116·30 <0·0001 10794·84 Poisson 1491·3100 0·5013 10796·84 NB −2·0000 NB Normal
Genital herpes 2·5 (5·53) 14·67 <0·0001 6519·80 0·0004 6463·26 56·54 <0·0001 6380·17 ZIP 83·0880 0·1768 6378·25 NB 1·9220 NB Bimodal
Genital warts 9·5 (37·40) 0·12 <0·0001 9719·94 0·1753 9712·69 7·25 <0·0001 9208·55 Poisson 511·3900 0·5019 9210·55 NB −2·0000 NB Normal
Gonorrhoea 1·8 (3·88) 28·44 <0·0001 5515·32 0·0006 5463·17 52·15 <0·0001 5436·05 ZIP 27·1220 0·1525 5432·93 NB 3·1230 NB Bimodal
Syphilis 0·2 (0·31) 83·64 1·0000 1820·41 0·0033 1782·72 37·69 <0·0001 1790·35 ZIP −7·6330 Did not converge ZIP Normal
TB Tuberculosis 0·3 (0·68) 80·13 1·0000 1925·97 0·0974 1918·59 7·38 0·0030 1918·35 Poisson 7·6240 0·2148 1975·60 NB −57·2530 NB Bimodal
VPD Measles 0·2 (0·35) 93·78 1·0000 1272·75 Did not converge <0·0001 997·65 Poisson 275·0984 0·2136 994·67 NB 2·9823 NB Normal
Mumps 0·6 (1·36) 59·78 0·0006 3351·60 0·0146 3327·67 23·93 <0·0001 3323·82 ZIP 3·8510 0·1771 3319·82 NB 4·0010 NB Bimodal
Pertussis 0·1 (0·20) 87·98 1·0000 1508·98 0·0042 1464·64 44·34 <0·0001 1462·52 ZIP 2·1230 0·4800 1462·50 ZIP 2·1360 ZIP Normal

AIC, Akaike's Information Criterion; BBV, bloodborne viruses; GAS, group A streptococcal; GOF, goodness of fit; HBV, hepatitis B virus; HCV, hepatitis C virus; IBD, invasive bacterial diseases; IID, infectious intestinal infections; n.a., not available; LRT, likelihood ratio test; NB, negative binomial; STI, sexually-transmitted infections; TB, tuberculosis; VPD, vaccine-preventable diseases; ZINB, zero-inflated negative binomial; ZIP, zero-inflated Poisson.

*

Difference in AIC between the model of interest and the reference model.

Reference model for testing.

Distribution of residuals for the final model assessed by eye.