Within-mouth correlations for indicators of the host response in gingival crevicular fluid

Gingival crevicular fluid was collected from multiple sites in patients with chronic adult periodontitis, and analysed for the lysosomal enzymes beta-glucuronidase and arylsulphatase, the cytoplasmic enzyme lactate dehydrogenase, total IgA, IgG and IgM and the protease inhibitor alpha 2-macroglobulin. The within-mouth (intraclass) correlation coefficients were calculated to describe the relationship between samples collected from individual patients. Data collected at baseline and 3 months after root planing and scaling were analysed, as was the change between examinations. Volume of crevicular fluid demonstrated the smallest intraclass correlation coefficient (0.16 at baseline, 0.12 at 3 months; 0.11 change), while probing depth and enzyme activity had moderate intraclass correlations (i.e. 0.36, 0.36, 0.26 for beta-glucuronidase). Immunoglobulin and alpha 2-macroglobulin activity in the fluid had the strongest correlations (i.e. 0.64, 0.57, 0.65 for IgG). The correlations for anatomically related teeth within a quadrant (molar, non-molar) were equivalent to or greater than the correlation for all samples within a mouth. Examined by tooth type, the intraclass correlations for volume of crevicular fluid, probing depth, beta-glucuronidase, arylsulphatase and lactate dehydrogenase were higher for non-molar teeth. In contrast, intraclass correlations for IgA, IgG, IgM and alpha 2-macroglobulin in samples from molar teeth were either equivalent to or greater than the correlations for non-molar samples. Calculation of intraclass correlation coefficients for such data can (1) indicate the degree of variability present in multiple samples of crevicular fluid collected from individual patients, (2) provide information about the source of host mediators in the fluid, and (3) help identify appropriate sampling strategies for the fluid.

Use of the scan statistic to detect time-space clustering

A test for time-space clustering is proposed based on the scan statistic, the maximum number of events in a 365-day period in each of several geographic units. The data under consideration should consist of the exact date and geographic unit for each event, and data should be available for several years for which the risk of disease can be assumed constant. The statistic is the ratio of the excess number of events summed over all the geographic regions, to the square root of the sum of the variances. This statistic is similar in construction to the Ederer-Myers-Mantel statistic (Biometrics 1964;20:626-38), but does not require that attention be limited to calendar years (January 1-December 31). Unlike other tests for time-space clustering, the scan statistic allows one to calculate measures of attributable risk and effect size. Data concerning adolescent suicide are used to illustrate the procedure. The tables and asymptotic formulas given for the mean and variance of the proposed statistic should be useful in the evaluation of both clustering in time and in time-space.

Testing for a pulse in seasonal event data

Most tests for seasonal clustering are sensitive to the peak in incidence in some season (90-day interval) followed by a trough 6 months later. We present two statistics sensitive to a relatively sharp increase in disease incidence for a season superimposed on a constant incidence over the entire year: the ratchet scan, based on the maximum number of events in k consecutive months, and the continuous circular scan, the maximum number of events in d consecutive days. We tabulate the tail distribution of the ratchet scan for N less than or equal to 35 and k = 2 and 3, and find the asymptotic distribution which is applicable for N greater than or equal to 50 when k = 2 or 3. We evaluate an approximation for the distribution of the continuous scan. The statistics are applied to investigate seasonal clustering of adolescent suicide in the United States. An illustration is given of the application of the circular scan, and a small simulation study and the example are used to compare the statistics with others proposed in the literature.

Number of comorbidities as a predictor of nosocomial infection acquisition

Infection control programs are primarily oriented toward chronicling the incidence of nosocomial infections (NI). Intervention programs oriented toward preventing infection would be facilitated by identifying patients at greatest risk of NI acquisition. We studied the number of comorbidities as a risk predictor for NIs in patients admitted to the medical intensive care unit (ICU) for three or more days. In 148 patients, we found by regression analysis that the number of comorbidities varied directly with the development of nosocomial infections, as well as with the appearance of new complications and length of ICU stay. Diagnosis-related groups did not adequately account for the variance in comorbidities observed.

Sublingual absorption of selected opioid analgesics

Ongoing interest in the improvement of pain management with opioid analgesics had led to the investigation of sublingual opioid absorption. The present report determined the percent absorption of selected opioid analgesics from the oral cavity of normal subjects under conditions of controlled pH and swallowing when a 1.0 ml aliquot of the test drug was placed under the tongue for a 10-minute period. Compared with morphine sulfate at pH 6.5 (18% absorption), buprenorphine (55%), fentanyl (51%), and methadone (34%) were absorbed to a significantly greater extent (p less than 0.05), whereas levorphanol, hydromorphone, oxycodone, heroin, and the opioid antagonist naloxone were not. Overall, lipophilic drugs were better absorbed than were hydrophilic drugs. Plasma morphine concentration-time profiles indicate that the apparent sublingual bioavailability of morphine is only 9.0% +/- 11.9% (SD) of that after intramuscular administration. In the same subjects the estimated sublingual absorption was 22.4% +/- 9.2% (SD), indicating that the sublingual absorption method may overestimate apparent bioavailability. When the oral cavity was buffered to pH 8.5, methadone absorption was increased to 75%. Thus, an alkaline pH microenvironment that favors the unionized fraction of opioids increased sublingual drug absorption. Although absorption was found to be independent of drug concentration, it was contact time dependent for methadone and fentanyl but not for buprenorphine. These results indicate that although the sublingual absorption and apparent sublingual bioavailability of morphine are poor, the sublingual absorption of methadone, fentanyl, and buprenorphine under controlled conditions is relatively high.

A re-examination of within-mouth correlations of attachment level and of change in attachment level

Using a data set from the Forsyth Dental Center consisting of repeated measurements of periodontal attachment level on 22 patients over 1 year, we performed correlation analyses at 4 time points (baseline, 1 month, 6 months and 9 months) for 3 units of measurement (sites, teeth and quadrants). Values of the intraclass correlation coefficient were obtained for the cross-sectional measurements and for changes from baseline. The values were positive and often of appreciable magnitude, indicating that measurements of attachment level and of change in attachment level within the same mouth are positively correlated. Taking sites rather than patients as the units of statistical analysis in comparative clinical trials in periodontology will therefore tend to produce underestimated standard errors and overstated statistical significance.

Epidemiology of acute respiratory illness during an influenza outbreak in a nursing home. A prospective study

We observed an influenza epidemic caused by influenza A/Arizona/82 (H3N2) in a nursing home during 1982 to 1983. A survey indicated that 59% of the residents were immunized before the outbreak. The outbreak was observed to begin in November, peak in February, and disappear in April. A significant level of herd immunity may have accounted for the slow progression through the nursing home. In addition, serologic evidence of concurrent infection with respiratory syncytial virus, parainfluenza virus, and Mycoplasma pneumoniae was present in many residents. Epidemics of influenza in a closed, partially immunized population in a nursing home may proceed at a slower rate than in an open, largely unimmunized community. By monitoring for infection with other respiratory agents, the complex nature of the outbreak in this nursing home became evident.

Association of influenza immunization with reduction in mortality in an elderly population. A prospective study

We prospectively studied the efficacy of influenza vaccine during an influenza A/Arizona/80 (H3N2) outbreak at the Jewish Home and Hospital for the Aged in New York in the winter season of 1982 to 1983. All patients had been offered influenza vaccine before the outbreak; 181 chose to be vaccinated and 124 refused vaccination but agreed to participate in the study. Among those with serologic evidence of influenza infection, respiratory illness was significantly more common in the unvaccinated group (six of 14 vs one of 22). The overall mortality was 13 (7.2%) of 181 in the vaccinated group and 22 (17.7%) of 124 in the control group. The vaccinated and the control groups were examined for comparability. A logistic regression analysis, which controlled for differences in sex and level of nursing care, indicated that the difference in mortality was still significant, with a summary odds ratio of 2.7. The relative risk of death in the unvaccinated group was comparable at 2.18. Influenza vaccine reduced the mortality by 59% in the vaccinated group compared with the control group.

A homogeneity test for follow-up studies

A statistical method is presented for detecting random variation in the transition probabilities of certain simple, discrete-time Markov chains that have been found useful for modelling follow-up studies.

Interobserver variability in neonatal cranial ultrasonography

The reliability of cranial ultrasound diagnosis in the premature neonate was examined using data from an ongoing multicentre study of the epidemiology and long-term consequences of neonatal brain haemorrhage. First week ultrasound films (obtained at 4 hours, 24 hours and 7 days) from 60 study subjects were randomly selected for independent review by two groups of experienced interpreters, and results were recorded separately for observations (i.e. presence or absence of an abnormal echodense area on a film) and interpretations (i.e. presence or absence of haemorrhage or ventricular dilatation) in each hemisphere. Because of deaths in the first week of life, the total number of films examined was 138. Concordance on the presence or absence of an abnormal echodensity was examined for each individual film for three areas of interest: the germinal matrix, the ventricles and the parenchyma. Concordance on the presence or absence of haemorrhage or ventricular dilatation was examined only for the seventh-day film, or the final film prior to death. Finally, concordance was analysed with the diagnostic interpretations grouped into categories thought to differ prognostically for long-term outcome. In general, concordance was poorest for germinal matrix lesions and best for parenchymal lesions. Concordance was lower for observations made on each individual film than it was for interpretation of the final film in each case. Fifty-five of 60 cases (92%) were assigned to the same major prognostic category by both readers. Ultrasound review conferences were held periodically and there was evidence that concordance in ultrasound reading and interpretation improved during the course of the study.

Epidemiologic programs for computers and calculators. Inferences on odds ratios, relative risks, and risk differences based on standard regression programs

Statistical analyses of the joint effects of several factors (covariates) on the risk of disease, death, or other dichotomous outcomes, are frequently based on a model that relates the effect of the covariates to some function of the probability of the outcome. The odds ratio, relative risk, and the difference in risks are among the simplest candidates for the outcome function. Each can be specified as a special case of the generalized linear model, but their use has been limited to researchers with access to specialized computer programs that are not yet generally available in standard computer packages. The purpose of this communication is to describe how to implement the maximum likelihood estimation procedures and hypothesis testing associated with the generalized linear model using any computer program that can perform weighted least squares analyses. The procedure is applied specifically to models for relative risks, risk differences, and odds ratios. The techniques are illustrated with SAS and SPSSx programs for data sets previously presented.

An approximation for the distribution of the scan statistic

The scan statistic evaluates whether an apparent cluster of disease in time is due to chance. The statistic employs a 'moving window' of length w and finds the maximum number of cases revealed through the window as it scans or slides over the entire time period T. Computation of the probability of observing a certain size cluster, under the hypothesis of a uniform distribution, is infeasible when N, the total number of events, is large, and w is of moderate or small size relative to T. We give an approximation that is an asymptotic upper bound, easy to compute, and, for the purposes of hypothesis testing, more accurate than other approximations presented in the literature. The approximation applies both when N is fixed, and when N has a Poisson distribution. We illustrate the procedure on a data set of trisomic spontaneous abortions observed in a two year period in New York City.

The choice of place of delivery. Effect of hospital level on mortality in all singleton births in New York City

In an analysis of all singleton births and neonatal deaths with known birth weights and gestational ages in New York City maternity services during a three-year period (1976 to 1978), intensive care services at the hospital of birth were found to influence mortality only in preterm (less than 37 weeks' gestation) or low-birth-weight infants (less than 2251 g). By contrast, for infants who were born at term and of normal birth weight, mortality rates did not differ by level of perinatal care available at the hospital of birth. On the average, preterm and low-birth-weight infants were at a 24% higher risk of death if birth occurred outside of a level 3 center, regardless of whether birth occurred at a level 1 or level 2 hospital. Preterm and low-birth-weight infants, though constituting only 12% of births, accounted for 70% of neonatal deaths in New York City. The remaining infants, ie, those born at term and of normal birth weight, who experienced no measurable mortality advantage when born in a level 3 hospital, accounted for 88% of all births.

Medical care and preterm infants of normal birth weight

Preterm infants of normal birth weight (born before 37 completed weeks of gestation and weighing more than 2,250 g) experience a neonatal mortality risk almost four times higher than do term infants in the same weight range. In an analysis of the effect of hospital level of birth on neonatal mortality, such preterm normal weight infants were found to experience higher mortality if born outside of a Level 3 (tertiary care) center. For all singleton infants in this weight-gestation category born in New York City maternity services during a 3-year period (N = 23,257), the relative mortality risk for Level 1 births (compared with Level 3) was 1.72 (P less than .01) and for Level 2 births 1.47 (P less than .05). The excess mortality at Level 1 and Level 2 units was almost entirely due to a more than twofold higher death rate in black infants born in these units. Several potentially confounding socioeconomic, demographic, and biologic variables entered into a logistic regression model could not account for the higher mortality rates for black infants born in Level 1 and Level 2 units. Among black infants born at Level 1 units, deaths in preterm normal birth weight infants were less likely to occur in a receiving tertiary care center than were either deaths in low birth weight infants or deaths in term normal weight infants, suggesting that the need for special care of preterm normal birth weight infants is underestimated in some hospitals without newborn intensive care units.