Correlating and predicting psychiatric symptom ratings: Spearman's r versus Kendall's tau correlation

Simple correlations play a large role in the analysis of psychiatric data. They are used to predict outcome, validate new instruments, establish treatment efficacy and find symptom patterns. Researchers and data analysts often face a question about which correlation coefficient to use in a study but are often unaware of the strengths and weaknesses of the alternative correlation measures. The presence of outliers, nonconstant variance, skewed distributions and unequal n are common in psychiatric data and this poses severe problems for many classic statistical methods. We compare Pearson, Spearman and Kendall's correlation coefficients using a large sample of subjects with schizophrenia spectrum disorders who were evaluated with 7 different psychiatric rating scales. Samples sizes ranging from 8 to 50 were evaluated using bootstrapping methods. The criteria for evaluation of the correlations were the type I error rates, power, bias and confidence interval width. Pearson's r did not always control for false positives at the nominal rate and was often unstable. Spearman's r performed better than Pearson's but provided a biased estimate of the true correlation. Spearman's r was also difficult to interpret. Our results suggest that Kendall's tau(b) has many advantages over Pearson's and Spearman's r; when applied to psychiatric data, tau(b) maintained adequate control of type I errors, was nearly as powerful as Pearson's r, provided much tighter confidence intervals and had a clear interpretation.     

Prevalence and incidence of clinically diagnosed memory impairments in a geographically defined general population in Sweden. The Piteå Dementia Project

In the Pite? River Valley all persons with memory impairment that interferes with normal life are referred to one hospital department for clinical workup and diagnosis. 619 patients were assessed in the department during the years 1990-1995. Of these, 36.9% had Alzheimer's disease (AD), 30.4% had vascular dementia (VaD), 3.0% had a mixed AD/VaD, 3.2% had frontotemporal dementia and 5.3% had other forms of dementia. Another 7% had memory impairment but no dementia. The overall mean annual incidence rate of clinically relevant dementia was 295/100,000 persons at risk and the mean prevalence rate was 755/100,000 persons. For persons 65 years and older the incidence and prevalence rates were 840 and 2,150/100,000 persons, respectively. This means that annually, approximately 300 persons/100,000 population over the age of 40 need medical attention or social services.     

Pharmacokinetics of sulfadiazine and trimethoprim in man

Summary Sulfadiazine (SDZ) 800 mg and trimethoprim (TMP) 160 mg were given orally to 10 normal subjects and the concentration of SDZ and TMP in serum and urine was followed for 24 h. Both drugs showed a significant negative correlation between individual peak concentrations in serum and the body weight of the subject. Twelve hours after dosing the serum concentration was 12 to 25 µg/ml for SDZ and 0.3 to 1.1 µg/ml for TMP. Individual concentration ratios between SDZ and TMP in serum were 4.8 (1 h) – 145 (24 h), and in the urine the ratio was close to 6 throughout the 24 h collection period. The range of urinary concentrations was from 65 to 400 µg/ml for SDZ and from 13.8 to 93.4 µg/ml for TMP. The fraction acetylated SDZ/acetylated SDZ + SDZ was 21% during the 0–8 h period, 33% during the 8–15 h period and 41% during the 15–24 period. The average values for the notional volume of distribution, V_d, were 0.36±0.13 1/kg for SDZ and 1.39±0.25 1/kg for TMP. The average t_1/2 was 15.2±7.4 h for SDZ and 7.4±1.9 h for TMP. Individual subjects showed a significant correlation between the serum clearance of TMP and SDZ (p<0.01) and also between the renal clearance of the two drugs (p<0.05). The serum clearance was significantly correlated with the renal clearance for TMP but not for SDZ. For SDZ V_d was significantly negatively correlated with the elimination constant; for TMP no such correlation was found. The serum clearance of SDZ was significantly correlated with the percentage of SDZ which was excreted as the (presumably) acetylated compound. The renal clearance of SDZ was independent of the serum concentration of SDZ. There was a highly significant negative correlation between the renal clearance and serum concentration of TMP, as well as for acetylated SDZ. The renal clearance of acetylated SDZ averaged more than six times that of unconjugated SDZ. With increased urine flow the renal clearances of TMP and SDZ were significantly increased.     

The treatment of obesity by carbohydrate deprivation suppresses plasma tryptophan and its ratio to other large neutral amino acids

We measured plasma concentrations of tryptophan (Trp) and the other large neutral amino acids (LNAA) in 6 control and 7 obese subjects before and after they consumed a low-carbohydrate "protein-sparing modified fast" (PSMF) diet; LNAA levels in control subjects were also assessed after supplemental oral Trp. Consumption of the PSMF diet by non-obese subjects, or obesity per se, caused major reductions in the ratio of the plasma Trp concentration to the summed plasma concentrations of the other LNAA (i.e., the "plasma Trp ratio"), and may thus have diminished brain serotonin synthesis. Administration of even 2 g of supplemental Trp did not elevate the plasma Trp ratio beyond the normal range observed previously in subjects consuming carbohydrate-containing meals.     

Binding of aprindine and moxaprindine to human serum, α1-acid glycoprotein and serum of healthy and diseased humans

Summary A comparison was made between the binding of the anti-arrhythmic agents aprindine and moxaprindine to human serum, to human serum albumin (HSA), to ₁-acid glycoprotein (₁-AGP) and to a mixture of HSA and ₁-AGP. In serum from healthy volunteers (n=4) the binding of aprindine-HCl 5 µg/ml (13.8 µM) was 93.8% (SD±1.0), and that of moxaprindine-HCl 5 µg/ml (12.8 µM) was 94.1% (SD±1.1). Their binding to the mixture of ₁-AGP and albumin approximated their binding to serum. For ₁-AGP, the binding was similar for both compounds, whereas for HSA the binding of aprindine was more pronounced than that of moxaprindine: for both products the affinity coefficient for binding to ₁-AGP was about 100 times greater than that for binding to albumin. In serum from rheumatoid patients and from patients with renal failure a small but significant increase in binding of aprindine and moxaprindine was observed, approximately 1%. Increased and decreased binding was seen in serum from cirrhotic patients; for example, for aprindine the range in cirrhosis was 96.7%–79.8%, and the range in controls was 95.0%–92.4%. Free drug fraction and ₁-AGP concentration were inversely correlated. The results show that ₁-AGP plays an important role in the binding of aprindine and moxaprindine, and that alteration in the binding of the two compounds in disease states to a large extent can be explained by changes in serum ₁-AGP concentration.     

A phase I trial of ZD9331, a water-soluble, nonpolyglutamatable, thymidylate synthase inhibitor.

PURPOSE: ZD9331 is a novel, direct-acting antifolate cytotoxic that does not require polyglutamation for activity, and is a specific thymidylate synthase inhibitor. This Phase I trial aimed to determine the maximum tolerated dose of ZD9331, given as a 30-min i.v. infusion on days 1 and 8 of a 21-day cycle. Pharmacokinetic parameters and tumor response were also assessed. EXPERIMENTAL DESIGN: A total of 71 patients, with a range of solid malignancies and refractory to standard therapies (44% had received > or =3 prior chemotherapy regimens), were treated. The most common malignancies were colorectal cancer (35% of patients) and ovarian cancer (31%). ZD9331 was escalated from 4.8 mg/m(2)/day. RESULTS: Dose-limiting toxicity occurred at 162.5 mg/m(2) ZD9331, with grade 4 thrombocytopenia, grade 4 neutropenia lasting > or =7 days, and grade 3 nonhematologic toxicity. Plasma clearance of ZD9331 was slow and dose-dependent; however, ZD9331 pharmacokinetics were nonlinear. Pharmacodynamics of ZD9331 were determined by measurement of plasma deoxyuridine, which increased at all of the dose levels; dose-related increases in plasma deoxyuridine were significant (P = 0.003) on day 5. Stable disease was observed in 37% of patients; 23% of ovarian cancer patients had a > or =50% reduction in CA125 levels. CONCLUSIONS: The maximum tolerated dose of this schedule was 130 mg/m(2). The toxicity profile at this dose was acceptable, with 7 of 28 patients treated developing grade 3/4 neutropenia and thrombocytopenia, 2 grade 4 diarrhea, and 2 grade 3/4 rash. This schedule was convenient and demonstrated activity in extensively pretreated patients; therefore, this is the recommended dose for study in Phase II trials.     

Cooperative extension initiatives in marriage and couples education

Cooperative extension has a long‐term history of delivering family life education. Further, the organization has developed numerous research‐based resources and undertaken statewide initiatives in marriage and couples education. Most promising of all is the work of a group of extension specialists to create a practical, research‐based curriculum that may reach communities in the United States through the established network of extension educators and their collaborators. We review various family relationship models and describe how the design of the new couples' curriculum is based on those models.     

MATERNAL ANEMIA AND POSTPARTUM WEIGHT CHANGE ASSOCIATED WITH DECREASED MATERNAL-INFANT INTERACTION IN A RURAL KENYAN POPULATION

Associations between maternal nutritional factors including energy intake, body mass index, postpartum weight change, and anemia status and maternal-infant interactions were examined in 124 mother-infant pairs from a marginally malnourished, rural Kenyan population. Anemic mothers spent less time holding and caring for their infants than nonanemic mothers. Mothers who retained their pregnancy weight gain or were able to gain weight postpartum spent more time looking at their infants than mothers who lost weight postpartum. Maternal food intake per se was not associated with patterns of infant interaction. Lower birth weight infants were held more, cared for more, and looked at face-to-face more by their mothers. Older sisters tended to be more involved in infant interactions with higher birth weight infants.