The heterogeneity of symptom reporting across study sites: a secondary analysis of a randomised placebo-controlled multicentre antimalarial trial

INTRODUCTION

Symptoms reported following the administration of investigational drugs play an important role in decisions for registration and treatment guidelines. However, symptoms are subjective, and interview methods to quantify them are difficult to standardise. We explored differences in symptom reporting across study sites of a multicentre antimalarial trial, with the aim of informing trial design and the interpretation of safety and tolerability data.

METHODS

Data were derived from the IMPROV trial, a randomised, placebo-controlled double blinded trial of high dose primaquine to prevent Plasmodium vivax recurrence conducted in eight study sites in Afghanistan, Ethiopia, Indonesia and Vietnam. At each follow up visit a 13-point symptom questionnaire was completed. The number and percentage of patients with clinically relevant symptoms following the administration of primaquine or placebo, were reported by study site including vomiting, diarrhoea, anorexia, nausea, abdominal pain and dizziness. Multivariable logistic regression was used to estimate the confounder-adjusted site-specific proportion of each symptom.

RESULTS

A total of 2,336 patients were included. The greatest variation between sites in the proportion of patients reporting symptoms was for anorexia between day 0 and day 13: 97.3% (361/371) of patients in Arba Minch, Ethiopia, reported the symptom compared with 4.7% (5/106) of patients in Krong Pa, Vietnam. Differences attenuated slightly after adjusting for treatment arm, age, sex, day 0 parasite density and fever; with the adjusted proportion for anorexia ranging from 4.8% to 97.0%. Differences between sites were greater for symptoms graded as mild or moderate compared to those rated as severe. Differences in symptom reporting were greater between study sites than between treatment arms within the same study site.

CONCLUSION

Despite standardised training, there was large variation in symptom reporting across trial sites. The reporting of severe symptoms was less skewed compared to mild and moderate symptoms, which are likely to be more subjective. Trialists should clearly distinguish between safety and tolerability outcomes. Differences between trial arms were much less variable across sites, suggesting that the relative difference in reported symptoms between intervention and control group is more relevant than absolute numbers and should be reported when possible.

TRIAL REGISTRATION

Clinicaltrials.gov: NCT01814683; March 20th, 2013.

Factors associated with loss to follow-up among TB patients in rural Papua New Guinea

BACKGROUND

Papua New Guinea (PNG) is a lower middle-income country that has struggled to contain TB. The loss of patients to follow-up is a major contributing factor towards the high disease burden.

OBJECTIVE

To describe persons with drug-susceptible TB (DS-TB) registered for treatment at the Gaubin Rural Hospital (GRH) on Karkar Island, Madang Province, PNG, and to investigate factors associated with patient loss to follow-up (LTFU).

DESIGN

A retrospective cohort study was conducted using data from GRH DS-TB registers. Factors associated with LTFU were investigated using univariable and multivariable logistic regression.

RESULTS

A total of 722 patients were registered for DS-TB treatment and eligible for inclusion between 1 January 2014 and 30 June 2018, of whom 97 (13.4%) were lost to follow-up. Male sex was associated with an increased odds of LTFU (adjusted OR [aOR] 1.9, 95% CI 1.2-3.0; P = 0.005), as was travel time to GRH >3 h (aOR 3.7, 95%, CI 2.2-6.3; P < 0.001).

CONCLUSION

A high LTFU rate has been identified in patients with TB in PNG. This study found male sex and increased travel time from treatment location to be associated with unsuccessful treatment adherence, highlighting the need for further interventions to improve adherence.

A retrospective study of tuberculosis outcomes in Gulf Province, Papua New Guinea

Setting

Gulf Province, a rural area of mainland Papua New Guinea, is known to have one of the highest burdens of tuberculosis (TB) in the country.

Objectives

To describe the characteristics and outcomes of TB patients registered for first-line treatment in Kerema General Hospital in Gulf Province between January and December 2016.

Design

This was a retrospective cohort study using routinely collected programme data.

Results

Of 347 cases with a recorded TB site, 54% were male and 32% were aged <15 years. No human immunodeficiency virus (HIV) status was recorded for 51% of cases. TB was bacteriologically confirmed in 23% of cases. Among the cohort, there were 145 extrapulmonary TB cases (42%); the site of disease was unknown in 56% of these cases. Of the 297 cases with treatment outcome evaluated, 56% had a favourable outcome and 26% were lost to follow-up. On multivariable analysis, extrapulmonary TB (adjusted OR [aOR] 0.51, 95%CI 0.30-0.88, P = 0.02) and bacteriologically confirmed TB (aOR 0.40, 95%CI 0.21-0.77, P < 0.01) were associated with decreased odds of an unfavourable treatment outcome.

Conclusion

The study findings highlight the need to improve TB diagnosis, access to HIV testing, treatment adherence, patient support and the quality of TB programme data in Gulf Province.

A mortality review of adult inpatients with tuberculosis in Mendi, Papua New Guinea

SETTING

Mendi Provincial Hospital, Southern Highlands Province, Papua New Guinea (PNG).

BACKGROUND

PNG is a high burden country for tuberculosis (TB) and TB-human immunodeficiency virus (HIV). TB is the second most common cause of death in PNG.

OBJECTIVE

To identify the number of adult inpatients with TB who died between 1 January 2015 and 30 August 2017; describe these patients' characteristics and identify contributing factors that could be modified.

DESIGN

This was a retrospective case series review.

RESULTS

Among 905 inpatients with TB during the study period, there were 90 deaths. The patients who died were older than those who survived (median age 40 years vs. 32 years, P = 0.011). The majority of patients who died lived less than 3 hours from the hospital (71%), were diagnosed after admission (79%) and were clinically diagnosed (77%). HIV status was not known in 50% of the deaths. Of patients with a known status, 27% (12/45) were HIV-positive. The median symptom duration prior to presentation was 28 days, with females presenting later than males (84 vs. 28 days, P = 0.008).

CONCLUSION

This study highlights areas where community and hospital-based management of TB could be improved to potentially reduce TB mortality, including earlier detection and treatment, improved bacteriological diagnosis and increased HIV testing.

Impact of GxAlert on the management of rifampicin-resistant tuberculosis patients, Port Moresby, Papua New Guinea

Setting

GxAlert is an automatic electronic notification service that provides immediate Xpert^® MTB/RIF testing results. It was implemented for the notification of patients with rifampicin resistant-tuberculosis (RR-TB) at Port Moresby General Hospital, Port Moresby, Papua New Guinea, in May 2015.

Objective

To determine if there were differences in pre-treatment attrition, the time to treatment initiation and patient outcomes in the 12 months pre- and post-introduction of GxAlert for RR-TB patients.

Design

This was a retrospective cohort study.

Results

The median time from Xpert testing to treatment initiation decreased from 35 days [IQR 13-131] prior to GxAlert to 10 days [IQR 3-29] after GxAlert (P = 0.001), with the cumulative proportion of patients initiating treatment within 30 days increasing from 25% (95%CI 17-37) to 54% (95%CI 44-64; P < 0.001) over these periods. However, our analysis of the time to treatment prior to the introduction of GxAlert suggests that a decrease had already occurred prior to implementation. There was no difference in interim clinical outcomes between the periods.

Conclusion

Although a decrease in time to treatment initiation cannot be attributed to GxAlert, there was a significant improvement over the 2-year period, suggesting that considerable improvements have been made in timely RR-TB patient management in Port Moresby.

Melioidosis in a patient on monoclonal antibody therapy for psoriatic arthritis

Melioidosis is caused by the environmental bacterium Burkholderia pseudomallei and can present with severe sepsis. Predisposing risk factors are present in 80% of cases. Monoclonal antibodies are increasingly prescribed for varied medical conditions. This report describes the first known case of melioidosis in a patient whose only risk factor for disease is treatment with a monoclonal antibody. Prescribers of monoclonal antibodies and other immunosuppressants should ensure that their patients are aware of the potential risk of melioidosis prior to travel and the precautions that should be taken.

Triaging pandemic flu: pneumonia severity scores are not the answer

SETTING

The 2009 H1N1 influenza pandemic caused significant strain on health systems worldwide. A tool to triage patients at low risk of requiring intensive care services would assist practitioners in safely reducing hospital admission rates during pandemic influenza outbreaks. Community-acquired pneumonia severity scores have not been validated for use in pandemic influenza.

OBJECTIVE

To assess the accuracy of the pneumonia severity index (PSI), CURB-65 and SMRT-CO severity scores in predicting patients at low risk of requiring intensive care services.

DESIGN

Between May and July 2009, 105 patients admitted with laboratory-confirmed pandemic (H1N1) 2009 influenza to Melbourne public hospitals were assessed on admission to determine their pneumonia severity scores and subsequent need for intensive care unit (ICU) support and length of stay.

RESULTS

SMRT-CO was the most accurate score at predicting ICU admission, with an area under the curve of the receiver operating characteristic of 0.826. No score provided good discrimination of low-risk patients, with respectively 19%, 21% and 12% requiring ICU admission as predicted by PSI, CURB-65 and SMRT-CO.

CONCLUSION

Current pneumonia severity scores have insufficient predictive ability to safely identify low-risk patients with pandemic (H1N1) 2009 influenza.