Evidence Table 18Quality assessment of observational studies

Non-biased selection?High overall loss to follow- up or differential loss to follow-up?Outcomes pre-specified and defined?Ascertainment techniques adequately described?Non-biased and adequate ascertainment methods?Statistical analysis of potential confounders?Adequate duration of follow-up?Overall quality rating
Aarskog 2009
Consecutive sera from patients treated between 2005 and 2007
No missing data reported.YesYesUnclear if specimens blinded to treatment.NA
Descriptive study, no analysis
Baum 2007
Patients were recruited from 76 centers in 13 countriesNo.

Note: there were more patients treated with Betaferon than Rebif.
YesYes, VAS used and patients were telephoned weeklyYes, trained study nurses called patients weekly for FU.Patients who used lower dose of study drug at least once, or in whom VAS assessment were not interpretable were excluded from analyses.Possibly. Observation period was 4 to 5 weeks for evaluation of ISP and ISR.Fair
Boz 2007
Canadian clinic cohort treated with a single IFN β for more than 3 years. Annual visits required by health system for continued prescriptions.
Loss to FU not reported per se. Fewer data years 6 to 10.YesYesYesYesYesGood
Data for this study were obtained from the i3 Lab Rx Database in 2007. Analysis included two distinct cohorts of MS patients with continuous insurance coverage from 6 months before to 24 months after medication start from July 2001 to June 2006.
Retrospective study
Cocco 2008
Patients were selected from 20 MS specialty units.

All women with MS who had 3 cycles of MITO before 45 years of age with at least 3 months of FU after MITO discontinuation were included between June 2005 and 2006.
Retrospective study
YesYes, a standardized questionnaire was usedYesYes but not all confounders were assessed.AE had to have lasted >6 months.Fair
Patients who had discontinued GA for any reason or whose disease had converted to a secondary progressive course were excluded.

Note: neurologists had to submit applications to French Agency for Safety of Healthcare Products (AFSSAPS). Patients approved by AFSSAPS could be prescribed GA.
18% in initial treatment phase were lost to FU and 21% were lost to FU in the long-term extension phaseTolerability was evaluated by spontaneous AE reporting.

No explicit criteria were specified in the protocol for definition of relapse (potential for variance in classification). Progression of disability was defined using EDSS score.
YesUnknown.Intergroup comparisons were performed but analyses of potential confounders were not assessed. (Also, there was no comparator arm.)Initial treatment phase: 4 to 58 months

Long-term extension phase: 2 yrs and 9 months

Total combined FU in both phases: 3.5 to 8 yrs
Durelli 2008
Analysis limited to patients who were treated with IFN β-1b for a full 2-year periodYes.
NAB+ testing on 73% of patients.
YesYesYesYesNo, 2 years (not adequate for relationship of NABs to clinical outcome).Fair
Farrell 2008
Patients receiving care at the National Hospital for Neurology and Neurosurgery at Queen Square in London who received IFN β since 1996 who had serum samples stored were coded into an electronic database. (Not all patients had serum samples stored.)~94% of originally identified patients were included. There was 23% loss of data for persistence.YesYesYes for labs.
Not clearly specified for chart review.
Jordy 2008 companion to another study conducted in BrazilUnclear. 390 patients were selected and then divided into 3 groups.NA
Prospective and retrospective study
YesNo, not described.Unknown.UnclearYesFair-Poor
Patients were selected from a Danish National MS Treatment RegisterNoYesYesYesNoYesFair-Poor
Le Page 2008
First 100 consecutive patients with RRMS who were treated with MITO as induction therapy monthly for 6 months were included.
97% (97/100) were used for effectiveness and harms.
YesYesYes except there were discrepancies with how MRIs were done.Yes but not all confounders were assessed.YesFair-Poor
Lugaresi 2008
Patients were selected from 19 neurological centers. Unclear how these patients were selected/recruited.NA
Single-arm, prospective cohort study which was 1 year in duration.
YesPatient questionnaires were completed monthly under supervision of trained nurses.

Treating neurologists also completed second questionnaire.
It is unknown whether trained nurses coached patients through questionnaire independently or if they used a common script for uniformity to obtain information from patients.None performed.
Authors combined results from 66% of patients who were naïve to treatment and 34% who switched from other immunomodulatory drugs due to lack of efficacy.
1 year (not adequate FU for serious AEs)Poor
Malucchi 2008
Unclear. Attempted to reduce bias by selecting only patients who were treated for 12 months or longer, but no information on sampling frame.No missing data reported.YesYesUnclear if specimens blinded to treatment.NoNo, 12 months (+/− 3 months of treatment); inadequate to show relationship of NAb status to clinical outcomes.Poor
Mancardi 2008
Patients were selected from medical centers that had experience in MS patient treatment and immunosuppressive therapies, had an MRI scanner available, and expertise in PCR. Centers were to add new patients to the database.NA
Retrospective database study
No. Specific AEs were not prespecified and/or defined.No, not described. Unclear how AEs were identified or organized in the database.Unknown.None performed.
90% of patients included in database had RRMS whose condition did not respond to treatment with IFN or GA. 10% of patients in database had rapidly evolving severe RRMS who were naïve to immunomodifying medications.
Miller 2008Open-label compassionate-use trial, and only patients who received GA or placebo in DB, randomized pilot study in 1978 were eligible. At the time, there were no immunomodulatory drugs.

in a double-blind, randomized pilot study initiated in 1978 [7] could also participat
60.8% discontinued
YesYesTo promote consistency across study sites, attending neurologists were given guidelines to achieve concordance in ambulation index, EDSS, and relapses.YesYes, up to 22 yearsFair
Oturai 2009
Patients were selected from 2 centers that offered treatment from the whole country. Authors included the first 234 consecutive RRMS patients.
Authors included patients who attended the planned visit at 3 months after start of therapy. Authors did not report if all patients from initial inclusion met this criteria.YesYesUnknown.Patients were divided into 3 subgroups. No statistical analyses of potential confounders were assessed in these groups.Median 11.3 months (not long enough for serious AEs like progressive multifocal leukoencephalopathy)Fair-Poor
Portaccio 2008
Patients were identified from database of the Department of Neurology at the Univ. of Florence and were prospectively followed. Unable to determine how patients were enrolled into the database. Data were evaluated from January 1996 to September 2005.No.
2.2% were lost to FU
YesYesYesYesRange 2 to 6 yrsFair
Rio 2007
Patients with SPMS who were started on therapy at the study center from 1998 to 2005 were included.No.
4.1% were lost to FU
YesYesYes, all neurologists participating were trained in EDSS assessment.Yes88.4% were followed for at least 12 months
61% were followed for at least 3 yrs
Sbardella 2009
Unclear if all eligible patients included.Unclear.YesYesUnclear if specimens blinded to treatment.NoNo, therapy duration ranged from 1 to 10 years, but no separate analysis by length of therapy.Poor
Sorensen 2007
Patients were selected from a Danish National MS Treatment RegisterNoYesYesYesNoYesFair-Poor
Tremlett 2008
All adverse drug reaction reporting of MS immunomodulating drugs reported to Health Canada were included.No.
1% of patients (9/888) were excluded due to insufficient information.
YesYesReporting of adverse drug reactions is voluntary except for serious adverse drug reactions which is mandatory for manufacturersYesRange 13 days to 5 yrsFair
Trojano 2009
A cohort of 2570 IFN β-treated RRMS was prospectively followed for up to 7 years in 15 Italian MS Centers.
NRYesYesYesYesYes, up to 7 yearsGood
Weber- Schoendorfer
Patients were identified through the Teratology Information Service of Berlin. Most requests for information in regard to exposure to IFN β or GA came from physicians whose patients inadvertently became pregnant while on disease-modifying therapies. Data were collected between 1996 to 2007 and patients were prospectively enrolled using questionnaires.Unclear.
Possibly there were no loss to FU.
YesYes, a standardized questionnaire was used.YesYes but not all confounders were assessed.YesFair-Poor

From: Evidence Tables

Cover of Drug Class Review: Disease-modifying Drugs for Multiple Sclerosis
Drug Class Review: Disease-modifying Drugs for Multiple Sclerosis: Final Update 1 Report [Internet].
Smith B, Carson S, Fu R, et al.
Portland (OR): Oregon Health & Science University; 2010 Aug.
Copyright © 2010 by Oregon Health & Science University, Portland, Oregon 97239. All rights reserved.

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.