Further evidence for the agreement between patients with stroke and their proxies on the Frenchay Activities Index

Objective: To determine item, subscale and total score agreement on the Frenchay Activities Index (FAI) between stroke patients and proxies six months after discharge from rehabilitation. Design: Prospective study design. Setting/subjects: Fifty patient-proxy pairs, interviewed separately, in the patient's residence. Main outcome measures: Modified FAI using 13 items. Individual FAI items, subscales and total score agreement as measured by weighted kappa and intraclass correlation coefficients (ICC). Results: Excellent agreement was found for the total FAI (ICC 0.87, 95% confidence interval (CI) 0.78-0.93), and domestic (ICC 0.85, 95% CI 0.73-0.91) and outdoor (ICC 0.87, 95% CI 0.78-0.95) subscales, with moderate agreement found for the work/leisure subscale (ICC 0.63, 95% CI 0.34-0.78). For the individual FAI items, good, moderate, fair and poor agreement was found for five, three, four and one item, respectively. The best agreement was for objective items of preparing meals, washing-up, washing clothes, shopping and driving. The poorest agreement was for participation in hobbies, social outings and heavy housework. Scoring biases associated with patient or proxy demographic characteristics were found. Female proxies, and those who were spouses, scored patients lower on domestic activities; male patients, and those who were younger, scored themselves higher on outdoor activities and higher patient FIM scores were positively correlated with higher FAI scores. Conclusions: While total and subscale agreement on the FAI was high, individual item agreement varied. Proxy scores should be used with caution due to bias.

Reliability of scores between stroke patients and significant others on the Reintegration to Normal Living (RNL) Index

Purpose: This study measured reliability between stroke patients' and significant others' scores on items on the Reintegration to Normal Living (RNL) Index and whether there were any scoring biases. Method The 11-item RNL Index was administered to 57 pairs of patients and significants six months after stroke rehabilitation. The index was scored using a 10-point visual analogue scale. Patient and significant other demographic information and data on patients' clinical, functional and cognitive status were collected. Reliability was measured using the intra-class correlation coefficient (ICC) and percent agreement. Results: Overall poor reliability was found for the RNL Index total score (ICC=.36, 95% CI. 07 to .59) and the daily functioning subscale (ICC=.24, 95% CI -.003 to .46) and moderate reliability was found for the perception of self subscale (ICC=.55, 95 % CI .28 to .73). There was a moderate bias for patients to rate themselves as achieving better reintegration than was indicated by significant others, although no demographic or clinical factors were associated with this bias. Exact match agreement was best for the subjective items and worse for items reflecting mobility around the community and participation in a work activity. Conclusions: Caution is needed when interpreting patient information reported by significant others on the RNL Index. The use of a shorter scale to rate the RNL Index requires investigation.

Toxicological aspects of treatment to remove cyanobacterial toxins from drinking water determined using the heterozygous P53 transgenic mouse model

The presence of toxic cyanobacteria in drinking water reservoirs renders the need to develop treatment methods for the 'safe' removal of their associated toxins. Chlorine has been shown to successfully remove a range of cyanotoxins including microcystins, cylindrospermopsin and saxitoxins. Each cyanotoxin requires specific treatment parameters, particularly solution pH and free chlorine residual. However, currently there has not been any investigation into the toxicological effect of solutions treated for the removal of these cyanotoxins by chlorine. Using the P53(def) transgenic mouse model mate and female C57BL/6J hybrid mice were used to investigate potential cancer inducing effects from such oral dosing solutions. Both purified cyanotoxins and toxic cell-free extract cyanobacterial solutions were chlorinated and administered over 90 and 170 days (respectively) in drinking water. No increase in cancer was found in any treatment. The parent cyanotoxins, microcystins, cylindrospermopsin and saxitoxins were readily removed by chlorine. There was no significant increase in the disinfection byproducts trihalomethanes or haloacetic acids, levels found were well below guideline values. Histological examination identified no effect of treatment solutions except male mice treated with chlorinated cylindrospermopsin (as a cell free extract). In this instance 40% of males were found to have fatty vacuolation in their livers, cause unknown. It is recommended that further toxicology be undertaken on chlorinated cyanobacterial solutions, particularly for non-genotoxic carcinogenic compounds, for example the Tg. AC transgenic mouse model. (C) 2003 Elsevier Science Ltd. All rights reserved.