ICD-10 codes generally denote GIB. i.e., Gastrointestinal bleeding in the observational studies. Most studies do not use validated ICD-10 coding algorithms to determine Gastrointestinal Bleeding.
Thus, this raises major concerns about the validity of the results. This test aimed at estimating the accuracy of ICD-10 codes for GIB.
This accuracy was mostly required for patients receiving oral anticoagulation and for developing a new coding algorithm to identify GIB events.
What Were The Findings, And Methods Of The National Cohort Study About GI Bleed ICD 10?
A study was conducted from the year 2014 to 2019 to understand the validity and the authenticity of GI bleed ICD 10. Below are the major findings of the study.
What Was The Method Of The Test For Determining The Accuracy Of GI bleed ICD 10?
The conductors of the test identified all the patients who were receiving oral anticoagulation in Iceland from the year 2014 to 2019.
The organizers of the research study collected the cases related to ICD-10 codes from all five major hospitals in Iceland.
Further, the study team augmented the code search and the results of the endoscopy
registered in the national death registry.
Hence, they reviewed each of the cases related to GI bleed ICD 10 much more thoroughly. They manually confirmed all the diagnoses through a chart review.
The study team conductor the review of the results through gold standards. Moreover, the research team calculated the PPV for each code.
In addition, the research team included the codes with a PPV ≥ 75% in the ICD-10 coding algorithm.
1. What Was The Design Of The Study?
The study team collected the data of all the patients who have been taking OACs. This includes apixaban, dabigatran, rivaroxaban, edoxaban, or even warfarin.
The study team identified the patients from 1st March to 28th of Frbruary 2019 while making use of the Icelandic Medicine Registry.
Then they used a comprehensive “catch-all” method. Thus, the GIB events were identified using a combination of ICD-10 codes, which helped determine the endoscopic results.
The team also reviewed the death registry thoroughly. They confirmed all the GIB results manually with the help of a chart review.
In addition, they collected all the relevant data from the Landspitali University Hospital. This is the only tertiary hospital in Iceland, along with the four regional hospitals in Iceland.
The team collected data on all endoscopic procedures performed on patients using the NOMESCO Classification of Surgical Procedures Code.
The team then manually reviewed to assess all the missing diagnoses of GIB.
The scholars and the researchers tracked the medication of the patients. They followed the patients from filling the OAC prescription for the first time until 28th February 2019.
They also tracked the discontinuation of the treatment or switching to another OAC.
2. What Was The Statistical Analysis Related To GI Bleed ICD 10?
The study team conducted the Welch’s t-test to compare the numerical variables. In addition, they also used the Person’s χ² to compare categorical variables.
They understood that there were at least > 5 observations in each of the groups. Otherwise, they used a Fisher’s exact test.
They calculated the odd ratios while using maximum likelihood estimation and calculated the confidence intervals with the help of normal approximation per Wals’s method.
Furthermore, the team also considered that the value of P less than 0.05 as something that was statistically significant.
They calculated the PPV for the individual ICD 10 codes. They included the ICD-10 codes with PPV ≥ 75% in the final final ICD 10 coding algorithm.
3. What Were The Results Of The Tests Related To GI bleed ICD 10?
In total, 14,611 patients received OACs during the study period. The researches found that almost 745 patients had GIB during this time period.
During the study, the mean age remained at 71.0 years, and 6163 (42%) patients were female.
Moreover, at least 3005 patients received apixaban, while 1220 received dabigatran, and 153 received edoxaban. and 5718 received warfarin.
Moreover, atleast 10,670 patients (73.0%) received OAC for atrial fibrillation, 2813 (19.3%) for venous thrombolism, and 1128 (7.7%) for other indications.
The ICD-10 coding algorithm had 61.3 sensitivity. 99.6% specificity, 98.0% negative predictive value, and 90% PPV.
The CIB events identified using these ICD 10 codings algorithm were more commonly major bleeding events.
During these events, the patients mostly had to go to the hospital. The patients also required anticoagulation reversal and anticoagulation discontinuation, and endoscopic intervention.
These were more in numbers in comparison with the total number of GIB events that were missed.
What Are The Things That We Should Keep In Mind About GI Bleed ICD 10?
The ICD-10 coding algorithms reported here generally had good accuracy in identifying true GIB events, and the events that it identified were clinically much more severe than the ones that were missed.
The sensitivities of the ICD 10 coding algorithm were generally modest. Hence, the doctors and the researchers can improve them by reviewing the endoscopic results and death registries.
During the study, many of the doctors defined GIB as Overt GIB as hematemesis, melena, or hematochezia, without the signs of an alternative source of bleeding.
