Gek Hsiang Lim

Measures of Diagnostic Accuracy: Sensitivity, Specificity, PPV and NPV

INtroDuctIoN In biomedical studies, diagnostic tests are used to determine the presence or absence of diseases in study subjects. Examples include testing for the presence or absence of Alzheimer’s disease and invasive carcinoma. A diagnostic test is validated by comparing test results against a gold standard that establishes the true status of the subject. Test validation is an evaluation method used to determine the fitness of a test for a particular use and through it, one can assess how good the test is at identifying subjects with and without a disease or condition. Validation involves calculating four objective measures of test performance, namely, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV). The ideal diagnostic test would correctly identify subjects with and without the disease with 100% accuracy. Details of the four measures are provided below.

Circadian Dependence of Infarct Size and Acute Heart Failure in ST Elevation Myocardial Infarction

Objectives There are conflicting data on the relationship between the time of symptom onset during the 24-hour cycle (circadian dependence) and infarct size in ST-elevation myocardial infarction (STEMI). Moreover, the impact of this circadian pattern of infarct size on clinical outcomes is unknown. We sought to study the circadian dependence of infarct size and its impact on clinical outcomes in STEMI. Methods We studied 6,710 consecutive patients hospitalized for STEMI from 2006 to 2009 in a tropical climate with non-varying day-night cycles. We categorized the time of symptom onset into four 6-hour intervals: midnight–6:00 A.M., 6:00 A.M.–noon, noon–6:00 P.M. and 6:00 P.M.–midnight. We used peak creatine kinase as a surrogate marker of infarct size. Results Midnight–6:00 A.M patients had the highest prevalence of diabetes mellitus (P = 0.03), more commonly presented with anterior MI (P = 0.03) and received percutaneous coronary intervention less frequently, as compared with other time intervals (P = 0.03). Adjusted mean peak creatine kinase was highest among midnight–6:00 A.M. patients and lowest among 6:00 A.M.–noon patients (2,590.8±2,839.1 IU/L and 2,336.3±2,386.6 IU/L, respectively, P = 0.04). Midnight–6:00 A.M patients were at greatest risk of acute heart failure (P<0.001), 30-day mortality (P = 0.03) and 1-year mortality (P = 0.03), while the converse was observed in 6:00 A.M.–noon patients. After adjusting for diabetes, infarct location and performance of percutaneous coronary intervention, circadian variations in acute heart failure incidence remained strongly significant (P = 0.001). Conclusion We observed a circadian peak and nadir in infarct size during STEMI onset from midnight–6:00A.M and 6:00A.M.–noon respectively. The peak and nadir incidence of acute heart failure paralleled this circadian pattern. Differences in diabetes prevalence, infarct location and mechanical reperfusion may account partly for the observed circadian pattern of infarct size and acute heart failure.

The Low Fall as a Surrogate Marker of Frailty Predicts Long-Term Mortality in Older Trauma Patients.

Background Frailty is associated with adverse outcomes including disability, mortality and risk of falls. Trauma registries capture a broad range of injuries. However, frail patients who fall comprise a large proportion of the injuries occurring in ageing populations and are likely to have different outcomes compared to non-frail injured patients. The effect of frail fallers on mortality is under-explored but potentially significant. Currently, many trauma registries define low falls as less than three metres, a height that is likely to include non-frailty falls. We hypothesized that the low fall from less than 0.5 metres, including same-level falls, is a surrogate marker of frailty and predicts long-term mortality in older trauma patients. Methods Using data from the Singapore National Trauma Registry, 2011–2013, matched till September 2014 to the death registry, we analysed adults aged over 45 admitted via the emergency department in public hospitals sustaining blunt injuries with an injury severity score (ISS) of 9 or more, excluding isolated hip fractures from same-level falls in the over 65. Patients injured by a low fall were compared to patients injured by high fall and other blunt mechanisms. Logistic regression was used to analyze 12-month mortality, controlling for mechanism of injury, ISS, revised trauma score (RTS), co-morbidities, gender, age and age-gender interaction. Different low fall height definitions, adjusting for injury regions, and analyzing the entire adult cohort were used in sensitivity analyses and did not change our findings. Results Of the 8111 adults in our cohort, patients who suffered low falls were more likely to die of causes unrelated to their injuries (p<0.001), compared to other blunt trauma and higher fall heights. They were at higher risk of 12-month mortality (OR 1.75, 95% CI 1.18–2.58, p = 0.005), independent of ISS, RTS, age, gender, age-gender interaction and co-morbidities. Falls that were higher than 0.5m did not show this pattern. Males were at higher risk of mortality after low falls. The effect of age on mortality started at age 55 for males, and age 70 for females, and the difference was attributable to the additional mortality in male low-fallers. Conclusions The low fall mechanism can optimize prediction of long-term mortality after moderate and severe injury, and may be a surrogate marker of frailty, complementing broader-based studies on aging.

Gender and ethnic differences in incidence and survival of lymphoid neoplasm subtypes in an Asian population: Secular trends of a population-based cancer registry from 1998 to 2012

Descriptive epidemiology on incidence and survival by lymphoid neoplasm (LN) subtypes using the 2008 World Health Organisation (WHO) classification remained limited in Asia. The aim of this study was to evaluate whether gender and ethnic differences in incidence and survival of LN subtypes existed using the Singapore Cancer Registry (SCR) from 1998 to 2012. We derived age standardised incidence rates (ASIRs) by the direct standardisation method and 5‐year relative survival (RSR) by the Ederer II method and period approach. Five‐year observed survival (OS) was obtained for each ethnicity. Malays had the highest ASIR of total LNs among the three ethnicities for each time period. The largest increase in 5‐year RSR subtypes was follicular lymphoma from 43.8% in 1998–2002 to 82.3% in 2008–2012; followed by chronic lymphocytic leukaemia (CLL)/small lymphocytic lymphoma (SLL) from 48.1% in 1998–2002 to 77.9% in 2008–2012. Although males had higher incidence than females in each time period, females had greater 5‐year RSR for follicular lymphoma (89.8% in 2008–2012 for females vs. 76.6% in 2008–2012 for males) and CLL/SLL (78.7% in 2008–2012 for females vs. 76.7% in 2008–2012 for males). All three ethnicities experienced an overall increase in 5‐year OS for mature B‐cell lymphoma, with Indians experiencing the greatest increase (37.1% in 1998–2002 to 61.1% in 2008–2012), followed by Malays (30.8% in 1998–2002 to 48.7% in 2008–2012) and then Chinese (36.4% in 1998–2002 to 51.3% in 2008–2012). Our study demonstrated that improved mature B‐cell lymphoma survival was not only observed in the West, but also in Singapore.

Estimating cumulative live-birth rates after IVF treatment with Kaplan–Meier and competing risk methods

OBJECTIVE(S) To explore the use of competing risk (CR) as compared to the commonly used Kaplan-Meier (KM) methodology in estimating the cumulative live-birth rate (CLBR) after IVF Treatment in a context of high dropout rates and informative censoring. STUDY DESIGN We compare the KM and CR methodologies for estimating 2-year CLBR in a retrospective cohort of 2779 patients undergoing 5002 embryo transfers over a period of 9 years, from 2000 to 2008, at KKIVF Centre. RESULTS We observed a total of 1105 LB (39.8%), and a dropout rate of 44.2% (1228 patients). The overall CLBR is lower with CR compared with KM method (39% vs 52%) after up to nine embryo-transfer cycles over a period of two years. The highest CLBR was achieved for ovulation disorders (57% vs 49%, KM vs CR) followed by male factors (54% vs 43%, KM vs CR), with poorer outcomes from patients with decreased ovarian reserve (37% vs 16%, KM vs CR) and endometriosis (36% vs 25%, KM vs CR). As dropouts in our cohort are generally older and more likely to have poorer ovarian reserves, the CR method, which accounted for these dropouts, is likely to give more meaningful estimation of IVF success rates. CONCLUSION(S) The CR method should be considered as a useful alternative in deriving CLBR for IVF treatment where dropout rates are high and when informative censoring is involved.