Concordance between medical records and interview data in correctional facilities.
Concordance between medical records and interview data in correctional facilities.
复制标题
DOI：
10.1186/1471-2288-14-50
复制
发表时间：
2014-04-09
影响因子：
4
通讯作者：
Larson EL
中科院分区：
文献类型：
--
作者：
Bai JR;Mukherjee DV;Befus M;Apa Z;Lowy FD;Larson EL
自我管理的问卷或访谈和医疗记录经常被用作研究数据的来源，因此，评价其一致性和可靠性是必不可少的。本文的目的是评估医疗和行为数据之间的一致性，从医疗记录和访谈问卷在两个惩教设施。对2010年4月至2013年2月期间来自一男一女最高安全监狱的679名囚犯的医疗记录和访谈数据进行了比较。在SPSS中进行性别非分层和性别分层分析，以计算医疗（例如，艾滋病毒、糖尿病、高血压）和行为（例如，吸烟、吸毒、纹身）的情况。在性别未分层数据中计算病历和访谈之间的敏感性/特异性。在性别非分层分析中，HIV、丙型肝炎、糖尿病、哮喘和纹身史的κ评分具有很强或很好的一致性（0.66-0.89）。高血压、肾/肾脏疾病、吸烟、过去6个月内使用抗生素和曾经使用可卡因中度相关（0.49-0.57）。任何非法药物使用史（0.36）和大麻使用史（0.23）的一致性较差。总体而言，女性的κ评分和患病率高于男性。医疗条件更频繁地报告在医疗记录和行为条件有较高的患病率在采访中。在合并的机构数据中，医疗状况的敏感性范围为50.0%至86.0%，行为状况的敏感性范围为48.2%至85.3%，而医疗状况的特异性范围为95.9%至99.5%，行为状况的特异性范围为75.9%至92.8%。医疗记录和自我报告之间的协议水平因因素类型而异。医疗条件更频繁地报告图表审查和行为因素更频繁地自我报告。可能需要根据所寻求的信息类型仔细选择所使用的数据源。
Self- administered questionnaires or interviews and medical records are often used as sources of research data; thus it is essential to evaluate their concordance and reliability. The aim of this paper was to assess the concordance between medical and behavioral data obtained from medical records and interview questionnaires in two correctional facilities. Medical record and interview data were compared for 679 inmates from one male and one female maximum security prison between April 2010 and February 2013. Gender non-stratified and gender-stratified analyses were conducted in SPSS to calculate the prevalence and kappa coefficient scores (κ) for medical (e.g., HIV, diabetes, hypertension) and behavioral (e.g., smoking, drug use, tattoos) conditions. Sensitivity/specificity between medical records and interview were calculated in the gender non-stratified data. In the gender non-stratified analysis, κ score for HIV, hepatitis C, diabetes, asthma, and history of tattoos had strong or good concordance (0.66-0.89). Hypertension, renal/kidney disease, cigarette smoking, antibiotic use in the last 6 months, and cocaine use ever were moderately correlated (0.49-0.57). Both history of any illicit drug use ever (0.36) and marijuana use ever (0.23) had poor concordance. Females had higher κ scores and prevalence rates than males overall. Medical conditions were reported more frequently in medical records and behavioral conditions had higher prevalence in interviews. Sensitivity for medical conditions in the combined facility data ranged from 50.0% to 86.0% and 48.2% to 85.3% for behavioral conditions whereas specificity ranged from 95.9% to 99.5% for medical conditions and 75.9% to 92.8% for behavioral conditions. Levels of agreement between medical records and self-reports varied by type of factor. Medical conditions were more frequently reported by chart review and behavioral factors more frequently by self-report. Data source used may need to be chosen carefully depending upon the type of information sought.
登录
查看更多内容
影响因子：
1.9
作者：
Alves, Elisabete;Lunet, Nuno;Barros, Henrique
通讯作者：
Barros, Henrique
影响因子：
2.1
作者：
Apa, Zoltan L.;Bai, RuoYu;Larson, Elaine L.
通讯作者：
Larson, Elaine L.
影响因子：
1.8
作者：
Fathelrahman, Ahmed I
通讯作者：
Fathelrahman, Ahmed I
影响因子：
2.3
作者：
Mi MY;Collins JE;Lerner V;Losina E;Katz JN
通讯作者：
Katz JN
影响因子：
7.2
作者：
Okura, Y;Urban, LH;Rodeheffer, RJ
通讯作者：
Rodeheffer, RJ