By Deborah Turner,2014-08-04 16:44
18 views 0

    Box1: Studies evaluating completeness of tuberculosis notification within the NHSStudy Supplemental Data sources Outcome Measure Authors’ Comments and Quality of Data


     Reported Completeness of

    Tuberculosis Notification: Number.

    (%) or Number of cases notified

    where denominator was not known


    Histopathology A study undertaken in ? 43 patients with positive pathology Under-notification was mostly Slides of all cases reviewed. Accuracy of Department of Pathology in reports notified. from surgical in-patients. diagnosis confirmed: Acid-fast bacilli Hospital Activity Data Edinburgh, Scotland which 26(40%) of patients with convincing Recommended independent present in sections or suggestive of looked at pathology slides combined clinical and pathological notification of pathology reports tuberculosis on morphological grounds. Prescriptions and case notes between diagnosis were not notified. Slides matched with case notes.

    1981-4. Necropsy and Coroners 28 Bradley BL et al 1988Reports

     Death Study undertaken in 1985 - ? ? ? ? ? 426(73%) Chest physicians were more likely Case ascertainment incomplete: limited to Certificates 1989 to identify all cases of to notify cases than clinicians of adult hospital based patients. Does not adult (>16years) tuberculosis all other specialities combined. specify whether population base for

    at two hospitals in East Did not find that positive statutory notifications and supplementary London microbiology or histopathology data sources drawn from same catchment

     increased notification. Patients area 17Sheldon et al. 1992 with a past history of tuberculosis

     and those who had died within 6

    months were less likely to be

    notified. Recommended

    notification of all positive culture

    reports, collating of pharmacy

    scripts, and referral of all cases to

    chest physicians

Above study repeated in ? ? ? ? ? 252(93%) A reported 20% increase in Unclear whether prescriptions for

    1992-3 after changes notifications from 1985-9. chemoprophylaxis were included. These instigated. Data from deaths Attributed to changes introduced: would have counted as false positive cases omitted while hospital based CDSC contacts consultant if included

    prescriptions included responsible for patient if positive

     culture not notified, and collation Case ascertainment incomplete as the 26 Brown JS et al 1995 of prescriptions for izoniazid chest clinic was not used as a data source;

     although one of the recommendations

     made following the above study was that

    all tuberculosis cases were to be referred to

    the chest physician.

    Ascertainment of non-? ? ? ? 43 cases of non-respiratory Positive microbiology increases Low level of confirmation of diagnoses respiratory Tuberculosis in 5 tuberculosis notified notification. Staff to be trained to because only 10% of cases were Welsh boroughs between 38 extra cases of non-respiratory collect histological specimens investigated for accuracy of diagnosis

    1985-94 tuberculosis found (17 from hospital separately for culture

     records, 3 from microbiology and 18 Underascertainment of cases as 50% of 18 Mukerjee AK 1999from histopathology) Direct notification of pathology records from the Patient Episode database

    reports should be instituted to for the study period was missing and

    improve surveillance of patients treated by GPs were also not

    tuberculosis included

    Cases were matched and linked amongst

    data sources.

    Study undertaken in 1993-94 ? ? ? ? 88 cases reported Use of MEMO not timely and Accuracy of diagnosis confirmed by in Tayside to evaluate 43 (49%) cases identified through requires significant resources. Use examination of case notes. Under whether utilising a MEMO that were not previously of routinely available diagnostic ascertainment of cases because 69(36%) of computerised prescriptions reported codes for tuberculosis records were missing and could not be database (MEMO) could surveillance improves case included in the study

    improve case ascertainment ascertainment without additional

    by linkage to cases through resources Cases were matched and linked amongst assignment of unique data sources.

    community health index

    (CHI) to each individual in

    Tayside. Prescriptions for

    two or more anti-

    tuberculosis drugs were

    checked with case notes for

    verification of diagnosis.

     27 Grove A 2001

    Estimated completeness of ? ? 138 (93%) Prescriptions issued by GPs may Case verification incomplete and possible tuberculosis notification in be used to evaluate completeness inclusion of cases treated for Bolton between 1991-93 27 out of 34 cases identified through of notification of tuberculosis, but chemoprophylaxis

    using general practitioner GP prescriptions were notified has several limitations: may

    (GP) prescriptions. include cases treated for Bias in ascertainment of cases as different Combination of Isoniazid chemoprophylaxis and is time periods used for ascertaining of cases with at least one other anti-inappropriate for contact tracing from data sources, and hospital dispensed tuberculosis drug was because of the delay in obtaining prescriptions omitted

    assumed as treatment for prescriptions tuberculosis Cases were matched and linked between

     sources, but small numbers, probable

     inclusion of non-cases and bias in

     ascertainment of cases by using different 29Devine MJ, Aston R 1995 time periods for data sources violated

     principles of using CR

    CR used to determine ? ? 449 (92-93%) Combining several data sources is Case verification and matching of cases completeness of 12 (2.5%)extra cases found by CR useful tool for evaluating between data sources increased validity of Tuberculosis notifications in completeness of tuberculosis CR methodology

    Liverpool between 1989-96 Only 37% identified from in-patient notification

     records Laboratory confirmation improves 22 Tocque et al 2001 Case verification is important to accuracy of diagnosis

     exclude false positive and false

    negative cases. 27% false positive Underascertainment of cases because

    cases in hospital records and cases treated at GPs and hospital outpatient

    0%false positive in departments not included

    microbiological reports

    Introduction of diagnostic register

    for outpatients to increase case


    As data sets become more

    interdependent, CR may not

    produce meaningful results

Report this document

For any questions or suggestions please email