Automated mortality coding for improved health policy in the Philippines.

In 2016, the Bloomberg Philanthropies Data for Health initiative assisted the Philippine Statistical Authority in implementing Iris, an automated coding software program that enables medical death certificates to be coded according to international standards. Iris was implemented to improve the quality, timeliness, and consistency of coded data as part of broader activities to strengthen the country's civil registration and vital statistics system. This study was conducted as part of the routine implementation of Iris to ensure that automatically coded cause of death data was of sufficient quality to be released and disseminated as national mortality statistics. Data from medical death certificates coded with Iris between 2017 and 2019 were analysed and evaluated for apparent errors and inconsistencies, and trends were examined for plausibility. Cause-specific mortality distributions were calculated for each of the 3 years and compared for consistency, and annual numeric and percentage changes were calculated and compared for all age groups. The typology, reasons, and proportions of records that could not be coded (Iris 'rejects') were also studied. Overall, the study found that the Philippine Statistical Authority successfully operates Iris. The cause-specific mortality fractions for the 20 leading causes of death showed reassuring stability after the introduction of Iris, and the type and proportion of rejects were similar to international experience. Broadly, this study demonstrates how an automated coding system can improve the accuracy and timeliness of cause of death data-providing critical country experiences to help build the evidence base on the topic.

The impact of errors in medical certification on the accuracy of the underlying cause of death.

BACKGROUND: Correct certification of cause of death by physicians (i.e. completing the medical certificate of cause of death or MCCOD) and correct coding according to International Classification of Diseases (ICD) rules are essential to produce quality mortality statistics to inform health policy. Despite clear guidelines, errors in medical certification are common. This study objectively measures the impact of different medical certification errors upon the selection of the underlying cause of death. METHODS: A sample of 1592 error-free MCCODs were selected from the 2017 United States multiple cause of death data. The ten most common types of errors in completing the MCCOD (according to published studies) were individually simulated on the error-free MCCODs. After each simulation, the MCCODs were coded using Iris automated mortality coding software. Chance-corrected concordance (CCC) was used to measure the impact of certification errors on the underlying cause of death. Weights for each error type and Socio-demographic Index (SDI) group (representing different mortality conditions) were calculated from the CCC and categorised (very high, high, medium and low) to describe their effect on cause of death accuracy. FINDINGS: The only very high impact error type was reporting an ill-defined condition as the underlying cause of death. High impact errors were found to be reporting competing causes in Part 1 [of the death certificate] and illegibility, with medium impact errors being reporting underlying cause in Part 2 [of the death certificate], incorrect or absent time intervals and reporting contributory causes in Part 1, and low impact errors comprising multiple causes per line and incorrect sequence. There was only small difference in error importance between SDI groups. CONCLUSIONS: Reporting an ill-defined condition as the underlying cause of death can seriously affect the coding outcome, while other certification errors were mitigated through the correct application of mortality coding rules. Training of physicians in not reporting ill-defined conditions on the MCCOD and mortality coders in correct coding practices and using Iris should be important components of national strategies to improve cause of death data quality.

Are there sex differences in completeness of death registration and quality of cause of death statistics? Results from a global analysis.

INTRODUCTION: Recent studies suggest that more male than female deaths are registered and a higher proportion of female deaths are certified as 'garbage' causes (ie, vague or ill-defined causes of limited policy value). This can reduce the utility of sex-specific mortality statistics for governments to address health problems. To assess whether there are sex differences in completeness and quality of data from civil registration and vital statistics systems, we analysed available global death registration and cause of death data. METHODS: Completeness of death registration for females and males was compared in 112 countries, and in subsets of countries with incomplete death registration. For 64 countries with medical certificate of cause of death data, the level, severity and type of garbage causes was compared between females and males, standardised for the older age distribution and different cause composition of female compared with male deaths. RESULTS: For 42 countries with completeness of less than 95% (both sexes), average female completeness was 1.2 percentage points (p.p.) lower (95% uncertainty interval (UI) -2.5 to -0.2 p.p.) than for males. Aggregate female completeness for these countries was 7.1 p.p. lower (95% UI -12.2 to -2.0 p.p.; female 72.9%, male 80.1%), due to much higher male completeness in nine countries including India. Garbage causes were higher for females than males in 58 of 64 countries (statistically significant in 48 countries), but only by an average 1.4 p.p. (1.3-1.6 p.p.); results were consistent by severity and type of garbage. CONCLUSION: Although in most countries analysed there was no clear bias against females in death registration, there was clear evidence in a few countries of systematic undercounting of female deaths which substantially reduces the utility of mortality data. In countries with cause of death data, it was only of marginally poorer quality for females than males.

Effectiveness of training interventions to improve quality of medical certification of cause of death: systematic review and meta-analysis.

BACKGROUND: Valid cause of death data are essential for health policy formation. The quality of medical certification of cause of death (MCCOD) by physicians directly affects the utility of cause of death data for public policy and hospital management. Whilst training in correct certification has been provided for physicians and medical students, the impact of training is often unknown. This study was conducted to systematically review and meta-analyse the effectiveness of training interventions to improve the quality of MCCOD. METHODS: This review was registered in the International Prospective Register of Systematic Reviews (PROSPERO; Registration ID: CRD42020172547) and followed Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. CENTRAL, Ovid MEDLINE and Ovid EMBASE databases were searched using pre-defined search strategies covering the eligibility criteria. Studies were selected using four screening questions using the Distiller-SR software. Risk of bias assessments were conducted with GRADE recommendations and ROBINS-I criteria for randomised and non-randomised interventions, respectively. Study selection, data extraction and bias assessments were performed independently by two reviewers with a third reviewer to resolve conflicts. Clinical, methodological and statistical heterogeneity assessments were conducted. Meta-analyses were performed with Review Manager 5.4 software using the 'generic inverse variance method' with risk difference as the pooled estimate. A 'summary of findings' table was prepared using the 'GRADEproGDT' online tool. Sensitivity analyses and narrative synthesis of the findings were also performed. RESULTS: After de-duplication, 616 articles were identified and 21 subsequently selected for synthesis of findings; four underwent meta-analysis. The meta-analyses indicated that selected training interventions significantly reduced error rates among participants, with pooled risk differences of 15-33%. Robustness was identified with the sensitivity analyses. The findings of the narrative synthesis were similarly suggestive of favourable outcomes for both physicians and medical trainees. CONCLUSIONS: Training physicians in correct certification improves the accuracy and policy utility of cause of death data. Investment in MCCOD training activities should be considered as a key component of strategies to improve vital registration systems given the potential of such training to substantially improve the quality of cause of death data.

Health Information Management Practice in Sri Lanka.

Health information system development in Sri Lanka has generally been a low priority. Weaknesses exist in many areas, including vital registration, and hospital information such as inpatient and outpatient morbidity and mortality data. Lapses in hospital information systems are mainly due to incomplete patient medical records; non-availability, inaccuracy and illegibility of the final diagnoses on the front sheets; shortage of trained statistical staff; lack of supervision at all levels; and lack of facilities at medical record departments. WHO's South-East Asian Regional Organization, in collaboration with the Australian National Centre in Classification in Health, has developed a group of trainers who conduct MR/HIM and ICD-10 training courses within the country.

Quality of Final Diagnosis Coding in the Medical Records of Selected Hospitals in the Colombo District, Sri Lanka.

Clinical coding is a method of translating a clinical description of a disease or procedure into a standard code. Sri Lanka adopted the system of coding recommended by the WHO, The International Classification of Diseases - 10th Revision (ICD-10), in 1997, and this study was undertaken with the objectives of assessing the quality of ICD coding and to identify some factors influencing coding quality in this country. A sample of 1091 medical records was selected from six hospitals in the Colombo District, representing all categories of hospitals in the area. Quality of coding was assessed by using the Australian Coding Benchmark Audit (ACBA), a coding quality assessment tool developed by the National Centre for Classification in Health, Australia. It was found that the availability of the final diagnosis on front sheets of medical records was satisfactory (94.7%), but the accuracy of the diagnostic statement was unsatisfactory (54%). Out of the six hospitals studied, only the Teaching Hospital and the Peripheral Unit practiced coding. The overall rate of accuracy of ICD coding in the Colombo District hospitals was 31%, which is unsatisfactory. It is recommended that training opportunities in ICD-10 and other related subjects should be made available to the coders. Efforts should be made to improve the familiarity of the Medical Officers with the WHO guidelines on recording diagnostic information for ICD coding.