Ahmad Navid1, Shahin Hajibandeh2, Jayarama Mohan3, Shahab Hajibandeh4. 1. General Surgery Registrar in the Department of General Surgery, Pilgrim Hospital, Boston, Lincolnshire. 2. Senior House Officer in the Department of General Surgery, Blackpool Victoria Hospital, Blackpool, Lancashire. 3. Consultant Vascular and General Surgeon in the Department of General Surgery, Pilgrim Hospital, Boston, Lincolnshire. 4. Senior House Officer in the Department of General Surgery, Pilgrim Hospital, Boston, Lincolnshire PE21 9QS.
Abstract
BACKGROUND: Poor documentation in medical notes can affect the quality of the source document for coding which can lead to inaccurate coding. This study aimed to determine the accuracy of Hospital Episode Statistics (HES) data for comorbidities and to establish whether better documentation in admission clerking proforma can improve the accuracy of codes for comorbidities in general surgical patients. METHODS: A clinical audit was conducted to assess the accuracy, sensitivity, specificity, positive predictive value and negative predictive value of HES codes for comorbidities in general surgical patients before and after implementing better documentation in admission clerking proforma. The following comorbidities were included: hypertension, ischaemic heart disease, diabetes, asthma, chronic obstructive pulmonary disease, cerebrovascular disease, chronic kidney disease and hypercholesterolaemia. Medical notes were used as reference standard and a target standard of 98% was determined for the above measures. RESULTS: Overall, on the initial audit, HES codes had substandard accuracy (90.5%, kappa = 0.599), sensitivity (47.71%, 95% confidence interval 38.05-57.49%) and negative predictive value (89.60%, 95% confidence interval 86.73-92.03%). HES codes for comorbidities were 100% specific with positive predictive value of 100%. Implementing better documentation in the admission clerking proforma improved the accuracy (99.67%, kappa = 0.985), sensitivity (97.4%, 95% confidence interval 90.93-99.68%) and negative predictive value (99.62%, 95% confidence interval 98.63-99.95%) significantly from the baseline (P<0.0001). CONCLUSIONS: Although HES codes can confidently predict the actual presence of the comorbidities, they have substandard accuracy and ability to rule out the presence of the comorbidities. Better documentation in clerking proforma can improve the accuracy and 'ruling out' ability of the HES codes. This can be achieved by improving knowledge and accountability of clinicians about documenting comorbidities.
BACKGROUND: Poor documentation in medical notes can affect the quality of the source document for coding which can lead to inaccurate coding. This study aimed to determine the accuracy of Hospital Episode Statistics (HES) data for comorbidities and to establish whether better documentation in admission clerking proforma can improve the accuracy of codes for comorbidities in general surgical patients. METHODS: A clinical audit was conducted to assess the accuracy, sensitivity, specificity, positive predictive value and negative predictive value of HES codes for comorbidities in general surgical patients before and after implementing better documentation in admission clerking proforma. The following comorbidities were included: hypertension, ischaemic heart disease, diabetes, asthma, chronic obstructive pulmonary disease, cerebrovascular disease, chronic kidney disease and hypercholesterolaemia. Medical notes were used as reference standard and a target standard of 98% was determined for the above measures. RESULTS: Overall, on the initial audit, HES codes had substandard accuracy (90.5%, kappa = 0.599), sensitivity (47.71%, 95% confidence interval 38.05-57.49%) and negative predictive value (89.60%, 95% confidence interval 86.73-92.03%). HES codes for comorbidities were 100% specific with positive predictive value of 100%. Implementing better documentation in the admission clerking proforma improved the accuracy (99.67%, kappa = 0.985), sensitivity (97.4%, 95% confidence interval 90.93-99.68%) and negative predictive value (99.62%, 95% confidence interval 98.63-99.95%) significantly from the baseline (P<0.0001). CONCLUSIONS: Although HES codes can confidently predict the actual presence of the comorbidities, they have substandard accuracy and ability to rule out the presence of the comorbidities. Better documentation in clerking proforma can improve the accuracy and 'ruling out' ability of the HES codes. This can be achieved by improving knowledge and accountability of clinicians about documenting comorbidities.
Authors: S Hajibandeh; S Hajibandeh; R Deering; D McEleney; J Guirguis; S Dix; A Sreh; E Toner; A El Muntasar; A Kausar; G Sheikh; D OShea; A Shafiq; A Kelly; A Khan; D Arumugam; A Evans Journal: Hernia Date: 2017-12-14 Impact factor: 4.739