Factors associated with physician agreement on verbal autopsy of over 11,500 injury deaths in India

INTRODUCTION: Worldwide, injuries account for 9.8% of all deaths. The majority of these deaths occur in low- and middle-income countries where vital registration systems are often inadequate. Verbal autopsy (VA) is a tool used to ascertain cause of death in such settings. Validation studies for VA using hospital diagnosed causes of death as comparisons have shown that injury deaths can be reliably diagnosed by VA. However, no study has assessed the factors that may affect physicians' abilities to code specific causes of injury death using VA. METHOD/PRINCIPAL FINDINGS: This study used data from over 11,500 verbal autopsies of injury deaths from the Million Death Study (MDS) in which 6.3 million people in India were monitored from 2001-2003 for vital events. Deaths that occurred in the MDS were coded by two independent physicians. This study focused on whether physician agreement on the classification of injury deaths was affected by characteristics of the deceased and respondent. Agreement was analyzed using three primary methods: 1) kappa statistic; 2) sensitivity and specificity analysis using the final VA diagnosed category of injury death as gold standard; and 3) multivariate logistic regression using a conceptual hierarchical model. The overall agreement for all injury deaths was 77.9% with a kappa of 0.74 (99% CI 0.74-0.75). Deaths in the injury categories of "transport", "falls", "drowning" and "other unintentional injury" occurring outside the home were associated with greater physician agreement than those occurring at home. In contrast, self-inflicted injury deaths that occurred outside the home were associated with lower physician agreement. CONCLUSIONS/SIGNIFICANCE: With few exceptions, most characteristics of the deceased and the respondent did not influence physician agreement on the classification of injury deaths. Physician training and continued adaptation of the VA tool should focus on the reasons these factors influenced physician agreement.     

Validation of the symptom pattern method for analyzing verbal autopsy data

Background

Cause of death data are a critical input to formulating good public health policy. In the absence of reliable vital registration data, information collected after death from household members, called verbal autopsy (VA), is commonly used to study causes of death. VA data are usually analyzed by physician-coded verbal autopsy (PCVA). PCVA is expensive and its comparability across regions is questionable. Nearly all validation studies of PCVA have allowed physicians access to information collected from the household members'' recall of medical records or contact with health services, thus exaggerating accuracy of PCVA in communities where few deaths had any interaction with the health system. In this study we develop and validate a statistical strategy for analyzing VA data that overcomes the limitations of PCVA.

Methods and Findings

We propose and validate a method that combines the advantages of methods proposed by King and Lu, and Byass, which we term the symptom pattern (SP) method. The SP method uses two sources of VA data. First, it requires a dataset for which we know the true cause of death, but which need not be representative of the population of interest; this dataset might come from deaths that occur in a hospital. The SP method can then be applied to a second VA sample that is representative of the population of interest. From the hospital data we compute the properties of each symptom; that is, the probability of responding yes to each symptom, given the true cause of death. These symptom properties allow us first to estimate the population-level cause-specific mortality fractions (CSMFs), and to then use the CSMFs as an input in assigning a cause of death to each individual VA response. Finally, we use our individual cause-of-death assignments to refine our population-level CSMF estimates. The results from applying our method to data collected in China are promising. At the population level, SP estimates the CSMFs with 16% average relative error and 0.7% average absolute error, while PCVA results in 27% average relative error and 1.1% average absolute error. At the individual level, SP assigns the correct cause of death in 83% of the cases, while PCVA does so for 69% of the cases. We also compare the results of SP and PCVA when both methods have restricted access to the information from the medical record recall section of the VA instrument. At the population level, without medical record recall, the SP method estimates the CSMFs with 14% average relative error and 0.6% average absolute error, while PCVA results in 70% average relative error and 3.2% average absolute error. For individual estimates without medical record recall, SP assigns the correct cause of death in 78% of cases, while PCVA does so for 38% of cases.

Conclusions

Our results from the data collected in China suggest that the SP method outperforms PCVA, both at the population and especially at the individual level. Further study is needed on additional VA datasets in order to continue validation of the method, and to understand how the symptom properties vary as a function of culture, language, and other factors. Our results also suggest that PCVA relies heavily on household recall of medical records and related information, limiting its applicability in low-resource settings. SP does not require that additional information to adequately estimate causes of death.     

Comparisons of procedures for determining ribosomal ribonucleic acid similarities

Ribosomal ribonucleic acid (rRNA) cistron similarities were determined forClostridium species with labeled preparations of 16S, 23S, or 16S plus 23S rRNA. Similarities were measured by membrane competition and by thermal stability experiments, and the results were correlated. Correlations were also made between 16S rRNA membrane competition homology values and similarities of 16S rRNA oligonucleotides. Very similar results were obtained whether 16S-or 23S-labeled rRNA was used. Clustering patterns were similar for homology values from hybridization experiments incubated at 50° or 60°C, as were the clustering patterns comparing homology values and rRNA hybrid thermal stability values. There was a linear correlation between rRNA homology values and 16S oligonucleotide S_AB values (Tanner et al. [22]) for homology values above 20%, and S_AB values above 0.45.     

Prevention and treatment practices and implications for malaria control in Mukono District Uganda

Available data in Uganda indicate a resurgence of malaria morbidity and mortality countrywide. This study assessed the burden of malaria, treatment and prevention practices in order initiate a policy debate on the scaling-up of current interventions. A triangulation of methods using a cross-sectional survey and key informant interviews was used to assess self-reported malaria at a household level in Mukono District, Uganda. A total of 5583 households were surveyed, and a high proportion (2897, 51.9%) reported a person with malaria two weeks prior to the survey. Only 546 households (9.8%) owned and used insecticide-treated nets (ITNs) for malaria prevention. Similarly, only a few households (86, 1.5%) used indoor residual spraying. Self-treatment with home-stocked drugs was high, yet there was low awareness of the effectiveness of expired drugs on malaria treatment. Self-reported malaria was associated with socioeconomic, behavioural and environmental factors, but more especially with household ownership of ITNs. These results will contribute to the current debate on identifying new approaches for scaling-up prevention interventions and effective case management, as well as selection of priority interventions for malaria control in Uganda.     

Treatment-seeking practices for malaria in pregnancy among rural women in Mukono district, Uganda

Understanding treatment-seeking practices for malaria in pregnancy is necessary in designing effective programmes to address the high malaria morbidity in pregnancy. This study assessed women's perceptions on malaria in pregnancy, recognition of early signs of pregnancy and of malaria, and the cultural context in which treatment seeking takes place in Mukono District. Focus group discussions (FGD) and key informant interviews were conducted among pregnant women, non-pregnant women, adolescents and men. The results showed that malaria, locally known as omusujja, was perceived as the most common cause of ill health among pregnant women. Although malaria commonly presents with fever, some pregnant women feel hot in the womb with or without signs of fever and this illness, locally known as nabuguma, may lead to progressive weakness and occasionally to miscarriage and few respondents associated it with malaria. Primigravidae, adolescents and men were not considered at risk of omusujja or nabuguma. Similarly anaemia and low birth weight were not associated with malaria; in fact paleness was described as a normal sign of pregnancy. There are cultural and social pressures on married women to get pregnant and this forces them to conceal symptoms like feeling feverishness, backache, nausea, general weakness, loss of appetite and vomiting until they are sure these are due to pregnancy. Most women, however, could not differentiate symptoms of malaria from those of early pregnancy. There is a belief that omusujja is a normal sign of pregnancy and this is coupled with a strong cultural practice of using herbs and clays as a first resort to treat pregnancy ailments including malaria. The cultural beliefs and practices regarding delivery of twin and first births, coupled with the high cost of care, prevent women from delivering and using other services at health units.     

Automatic procedures for determining amino acid sequences of peptides.

An automatic procedure is described for determining the amino acid sequences of peptides with various lengths and hydrophobicities in a protein sequenator of the Edman-Begg type. A film consisting of Quadrol salts is left in the cup as a hydrated solid phase on which the peptide partitions during solvent extraction. The partitioning of the peptide is facilitated by using benzene and 1-chlorobutane/acetic acid as the sole extractants after coupling. The reproducibility and efficacy of the procedure is illustrated by the sequences obtained with peptides of from 3–29 residues, including several with a series of hydrophobic residues at the C terminus. The procedure is well suited to the completion of the sequence determination on a large peptide following the normal Edman-Begg procedure for proteins.     

Kinetics of suicide substrates. Practical procedures for determining parameters.

Many clinically important or mechanistically interesting inhibitors react with enzymes by a branched pathway in which inactivation of the enzyme and formation of product are competing reactions. The steady-state kinetics for this pathway [Waley (1980) Biochem. J. 185, 771-773] gave equations for progress curves that were cumbersome. A convenient linear plot is now described. The time (t1/2) for 50% inactivation of the enzyme (this is also the time for 50% formation of product), or for 50% loss of substrate, is measured in a series of experiments in which the concentration of inhibitor, [I]0, is varied; in these experiments the ratio of the concentration of enzyme to the concentration of inhibitor is kept fixed. Then a plot of [I]0 X t1/2 against [I]0 is linear, and the kinetic parameters can be found from the slope and intercept. Furthermore, simplifications of the equations for progress curves are described that are valid when the concentration of inhibitors is high, or is low, or when the extent of reaction is low. The use of simulated data has shown that the recommended methods are not unduly sensitive to experimental error.     

Modelling the epidemiological impact of intermittent preventive treatment against malaria in infants

Background

Trials of intermittent preventive treatment against malaria in infants (IPTi) using sulphadoxine-pyrimethamine (SP) have shown a positive, albeit variable, protective efficacy against clinical malaria episodes. The impact of IPTi in different epidemiological settings and over time is unknown and predictions are hampered by the lack of knowledge about how IPTi works. We investigated mechanisms proposed for the action of IPTi and made predictions of the likely impact on morbidity and mortality.

Methods/Principal Findings

We used a comprehensive, individual-based, stochastic model of malaria epidemiology to simulate recently published trials of IPTi using SP with site-specific characteristics as inputs. This baseline model was then modified to represent hypotheses concerning the duration of action of SP, the temporal pattern of fevers caused by individual infections, potential benefits of avoiding fevers on immunity and the effect of sub-therapeutic levels of SP on parasite dynamics. The baseline model reproduced the pattern of results reasonably well. None of the models based on alternative hypotheses improved the fit between the model predictions and observed data. Predictions suggest that IPTi would have a beneficial effect across a range of transmission intensities. IPTi was predicted to avert a greater number of episodes where IPTi coverage was higher, the health system treatment coverage lower, and for drugs which were more efficacious and had longer prophylactic periods. The predicted cumulative benefits were proportionately slightly greater for severe malaria episodes and malaria-attributable mortality than for acute episodes in the settings modelled. Modest increased susceptibility was predicted between doses and following the last dose, but these were outweighed by the cumulative benefits. The impact on transmission intensity was negligible.

Conclusions

The pattern of trial results can be accounted for by differences between the trial sites together with known features of malaria epidemiology and the action of SP. Predictions suggest that IPTi would have a beneficial impact across a variety of epidemiological settings.     

Validity of an algorithm for determining sleep/wake states using a new actigraph

Background

This study aimed to develop an algorithm for determining sleep/wake states by using chronological data on the amount of physical activity (activity intensity) measured with the FS-750 actigraph, a device that can be worn at the waist, allows for its data to be downloaded at home, and is suitable for use in both sleep research and remote sleep medicine.

Methods

Participants were 34 healthy young adults randomly assigned to two groups, A (n =17) and B (n =17), who underwent an 8-hour polysomnography (PSG) in the laboratory environment. Simultaneous activity data were obtained using the FS-750 attached at the front waist. Sleep/wake state and activity intensity were calculated every 2 minutes (1 epoch). To determine the central epoch of the sleep/wake states (x), a five-variable linear model was developed using the activity intensity of Group A for five epochs (x_-2, x_-1, x, x₊₁, x₊₂; 10 minutes). The optimal coefficients were calculated using discriminant analysis. The agreement rate of the developed algorithm was then retested with Group B, and its validity was examined.

Results

The overall agreement rates for group A and group B calculated using the sleep/wake score algorithm developed were 84.7% and 85.4%, respectively. Mean sensitivity (agreement rate for sleep state) was 88.3% and 90.0% and mean specificity (agreement rate for wakeful state) was 66.0% and 64.9%, respectively. These results confirmed comparable agreement rates between the two groups. Furthermore, when applying an estimation rule developed for the sleep parameters measured by the FS-750, no differences were found in the average values between the calculated scores and PSG results, and we also observed a correlation between the two sets of results. Thus, the validity of these evaluation indices based on measurements from the FS-750 is confirmed.

Conclusions

The developed algorithm could determine sleep/wake states from activity intensity data obtained with the FS-750 with sensitivity and specificity equivalent to that determined with conventional actigraphs. The FS-750, which is smaller, less expensive, and able to take measurements over longer periods than conventional devices, is a promising tool for advancing sleep studies at home and in remote sleep medicine.