Table 3 Probabilities used to inform none-medicinal (other household products) decision model
Parameter | Description | Data sources | Derivation of required probabilities | Point estimate (standard error (SE) | Distribution |
|---|---|---|---|---|---|
pSafe | Baseline prevalence of safe storage of non-medicines | Kendrick et al (paper om press) | Number of households with safe other household products/Total number of households selected as community controls in case control of risk and protective factors in children under 5 years old (Kendrick et al, paper om press). | 948/1138 = 0.83 | Beta |
pAccept | Probability of accepting the intervention | Saramago, Cooper et al., 2015 [14] | Assumption based on value in Functional smoke alarm model [14]. Assumed the same for all interventions. | 0.90 | Fixed |
pEff | Probability of safe storage of medicines given the intervention | Achana et al., 2015 [9] | (1) Usual care (2) Education (3) Provision of low cost/free low cost/free equipment (4) Education + provision of low cost/free equipment (5) Education + equipment + home safety inspection (6) Education + equipment + fitting (7) Education + equipment + home safety inspection + fitting | 0.62 (95%CrI 0.34–0.81) 0.66 (95%CrI 0.38–0.87) 0.36 (95%CrI 0.00–1.00) 0.78 (95%CrI 0.48–0.94) 0.80 (95%CrI 0.51–0.94) 0.68 (95%CrI 0.32–0.90) 0.50 (95%CrI 0.00–1.00) | Posterior distribution of absolute intervention effects from network meta-analysis. |
pIngest | Probability of unintentional exposure/ingestion | Orton et al., 2014 [10] Tyrrell et al., 2012 [33] Office for National Statistics, 2013 [34] | Number of unintentional poisoning cases among children under 5 years in 2012 = 12029 (see Table 2 above). Forty percent of poisoning cases is due to ingestion of non-medicinal substance [33]. Hence numerator = 12029*.4 = 4812. | 4812/3996400 = 0.001204 | Beta |
orIngest | Relative risk of exposure to a medicinal substance comparing children with a poisoning to community controls. | Case control study of risk and protective factors for poison injuries in under 5 year olds (unpublished study) | Community controlled adjusted analysis odds ratio for safe storage versus no safe storage = 0.77 (95 % CI 0.59–0.99) | Log odds ratio (SE) = 0.2614 (0.132) | Normal |
pAmb | Probability of using emergency ambulance. | Hospital Episode Statistics, 2012 | Hospital Episode Statistics (2012b): 24.2 % of all cases arrived by emergency transfer (ambulance/helicopter). | 0.242 | Fixed |
pAdmit | Probability of in-patient admission following a medicinal poisoning injury (ICD-10: X40-X44). | Hospital Episode Statistics, 2013 [35] Office for National Statistics, 2013 [34] | Hospital Episode Statistics, 2012–2013) [35]: Number of poisoning cases (X45–X49) admitted in 0-4 years olds (period 2012–2013) in England = 1377. Scaled up by a factor of 1.163 (i.e. 1377*1.16 = 1597 cases for whole of UK) based on mid-2012 population estimates for UK and England, ONS 2012a [34]. | 1597/4812 = 0.3318 | Beta |
pSevere | Probability of severe poisoning injury | Mowry et al., 2013 [36] | NPDS 2012 report [36], Table 13, page 968) 1.91 % of major poisoning cases (across all age groups) resulted in a chronic health condition. Numerator = 0.019* 1597 = 30. | 30/1597 = 0.0191 | Beta |
pFatal | Probability of fatal poisoning injury. | Office for National Statistics, 2012 [37] | UK mortality statistics [37]. 1 fatality from medicinal poisonings in 0–4 years old (assumed fatality occur after a long inpatient stay). | 1/30 = 0.033 | Beta |
pDead | UK mortality statistics | Office for National Statistics, 2010 [38] | UK mortality statistics [38] | Normal |