Clinical and statistical prediction: Where are we and where do we go from here?

In the years following the publication of P.E. Meehl's monograph, our understanding of clinical and statistical prediction has been enriched by both empirical and conceptual analyses. The highlights of this period are reviewed with an emphasis on prediction models, in order to illustrate the basic features held in common by these two literatures. We appear to be on the threshold of an era of parsimonious prediction models. That being the case we can now return our attention to the problems of developing better assessment devices and more meaningful criterion measures and of developing procedures that will increase the accuracy of clinical judgment and prediction.     

Use-effectiveness of the diaphragm in a selected family planning clinic population in the United Kingdom

4052 established users of the diaphragm aged 25–39 years attending family planning clinics in the United Kingdom have been followed during 5909 woman-years of exposure to the risk of pregnancy. 139 accidental pregnancies have been observed, representing a use-effectiveness pregnancy rate of 2.4 per 100 woman-years. Of the factors related to the risk of accidental pregnancy, the age of the user and the duration of use of the diaphragm appear to be the most important.It is concluded that women attending family planning clinics who are already established users of the diaphragm need not be encouraged to change to a more modern method of birth control with its attendant risks.     

Effect of iontophoretic and topical application of antiviral agents in treatment of experimental HSV-1 keratitis in rabbits.

Cathodal (-) iontophoresis of 9-beta-D-arabinofuranosyl-adenine 5'-monophosphate (vidarabine monophosphate; Ara-AMP) was performed once daily for 3 days for the treatment of experimental herpes simplex virus type 1 (HSV-1) keratitis in rabbit eyes, and the therapeutic efficacy was compared with that of topical treatment of Ara-AMP and idoxuridine (IDU) administered five times daily for 4 days. With the treatment initiated 24 hr after viral inoculation, Ara-AMP cathodal iontophoresis resulted in significant suppression of epithelial and anterior segment disease processes. Topical IDU (0.5%) or Ara-AMP (10%) also significantly improved the disease process when compared to the placebo-treated group; however, iontophoresis of Ara-AMP resulted in a more marked improvement. Slit-lamp examination indicated that iontophoresis did not cause any observable pathologic changes in corneal epithelium, stroma, conjunctiva, or iris of rabbit eyes. This experiment suggests that iontophoresis of Ara-AMP is a safe and effective approach for preventing the development of herpes simplex keratitis in rabbits.     

Predictors of a favourable response to icodextrin in peritoneal dialysis patients with ultrafiltration failure

BACKGROUND AND AIMS: Icodextrin is a starch-derived glucose polymer that causes sustained ultrafiltration in long dwells in peritoneal dialysis. The aim of this study was to assess factors that were predictive of an increment in ultrafiltration following the introduction of icodextrin in patients with refractory fluid overload. METHODS: Thirty-nine patients (20 male/19 female, mean age 57.7 +/- 2.4 years) on peritoneal dialysis were enrolled in a prospective pretest/post-test, open-label study. All patients had symptomatic fluid overload refractory to fluid restriction (<800 mL/day), frusemide doses of 250 mg or more daily, optimization of dwell time and use of hypertonic dextrose. An icodextrin exchange was substituted for a 4.25% dextrose exchange for the long-dwell period. RESULTS: After 1 month, median (interquartile range) 24 h ultrafiltration volume increased by 500 mL (interquartile range: 50-1000). An increase in ultrafiltration volume correlated positively with the dialyate : plasma creatinine ratio at 4 h (r = 0.498, P = 0.001) and negatively with the ratio of dialysate glucose concentrations at 4 and 0 h (r = -0.464, P = 0.003). On multivariate regression analysis, high transporter status was predictive of a greater ultrafiltration response to icodextrin relative to dextrose peritoneal dialysis exchanges. Age, sex, race, peritoneal dialysis duration, peritoneal dialysis modality, diabetes mellitus, baseline albumin, and baseline ultrafiltration volume were not significantly correlated with the change in ultrafiltration volume. CONCLUSION: Icodextrin significantly augments ultrafiltration volumes in patients with refractory fluid overload. A high peritoneal membrane transporter status is the best predictor of a favourable ultrafiltration response to icodextrin.     

Elevated white cell count at commencement of peritoneal dialysis predicts overall and cardiac mortality

BACKGROUND: Higher total white blood cell counts (WCC) have been shown in the general population to be strongly and independently predictive of coronary heart disease and all-cause mortality. The aim of the present study was to evaluate the prognostic value of WCC in patients commencing peritoneal dialysis (PD). METHODS: A cohort of 323 patients (mean age 55.1 +/- 17.7 years, 54% male, 81% Caucasian) commencing PD at the Princess Alexandra Hospital between January 1, 1998 and March 31, 2003 were prospectively followed until death, completion of PD therapy, or otherwise to the end of the study (January 2, 2004), at which point data were censored. Individuals with failed renal transplants (N= 17) and those with acute infections at the time of PD onset (N= 12) were not included. A multivariate Cox's proportional hazards model was applied to calculate hazard ratios and adjusted survival curves for time to death or cardiac death, adjusting for baseline demographic, clinical, and laboratory characteristics. RESULTS: Median actuarial patient survival was 3.9 years [95% confidence interval (CI) 3.2-4.7 years]. The highest quartile of WCC (>9.4 x 10(9)/L) was significantly and independently associated with increased risks of both death from all causes [adjusted hazard ratio (HR) 2.27, 95% CI 1.09-4.74, P < 0.05] and cardiac death (HR 3.75, 95% CI 1.2-11.8, P < 0.05). Other adverse risk factors included older age, lower serum albumin, and the presence of coronary artery disease. Similar associations were found between mortality and PMN count, but not lymphocyte count. CONCLUSION: Elevated baseline WCC or PMN count at the commencement of PD (in the absence of acute infection) strongly predicts all-cause and cardiovascular mortality. These data suggest that new PD patients with higher WCC may warrant closer monitoring and extra attention to modifiable cardiovascular risk factors.