+1 Recommend
1 collections
      • Record: found
      • Abstract: found
      • Article: found
      Is Open Access

      Identification of a potential fibromyalgia diagnosis using random forest modeling applied to electronic medical records

      Read this article at

          There is no author summary for this article yet. Authors can add summaries to their articles on ScienceOpen to make them more accessible to a non-specialist audience.



          Diagnosis of fibromyalgia (FM), a chronic musculoskeletal condition characterized by widespread pain and a constellation of symptoms, remains challenging and is often delayed.


          Random forest modeling of electronic medical records was used to identify variables that may facilitate earlier FM identification and diagnosis. Subjects aged ≥18 years with two or more listings of the International Classification of Diseases, Ninth Revision, (ICD-9) code for FM (ICD-9 729.1) ≥30 days apart during the 2012 calendar year were defined as cases among subjects associated with an integrated delivery network and who had one or more health care provider encounter in the Humedica database in calendar years 2011 and 2012. Controls were without the FM ICD-9 codes. Seventy-two demographic, clinical, and health care resource utilization variables were entered into a random forest model with downsampling to account for cohort imbalances (<1% subjects had FM). Importance of the top ten variables was ranked based on normalization to 100% for the variable with the largest loss in predicting performance by its omission from the model. Since random forest is a complex prediction method, a set of simple rules was derived to help understand what factors drive individual predictions.


          The ten variables identified by the model were: number of visits where laboratory/non-imaging diagnostic tests were ordered; number of outpatient visits excluding office visits; age; number of office visits; number of opioid prescriptions; number of medications prescribed; number of pain medications excluding opioids; number of medications administered/ordered; number of emergency room visits; and number of musculoskeletal conditions. A receiver operating characteristic curve confirmed the model’s predictive accuracy using an independent test set (area under the curve, 0.810). To enhance interpretability, nine rules were developed that could be used with good predictive probability of an FM diagnosis and to identify no-FM subjects.


          Random forest modeling may help to quantify the predictive probability of an FM diagnosis. Rules can be developed to simplify interpretability. Further validation of these models may facilitate earlier diagnosis and enhance management.

          Related collections

          Most cited references 27

          • Record: found
          • Abstract: found
          • Article: not found

          Definition, structure, content, use and impacts of electronic health records: a review of the research literature.

          This paper reviews the research literature on electronic health record (EHR) systems. The aim is to find out (1) how electronic health records are defined, (2) how the structure of these records is described, (3) in what contexts EHRs are used, (4) who has access to EHRs, (5) which data components of the EHRs are used and studied, (6) what is the purpose of research in this field, (7) what methods of data collection have been used in the studies reviewed and (8) what are the results of these studies. A systematic review was carried out of the research dealing with the content of EHRs. A literature search was conducted on four electronic databases: Pubmed/Medline, Cinalh, Eval and Cochrane. The concept of EHR comprised a wide range of information systems, from files compiled in single departments to longitudinal collections of patient data. Only very few papers offered descriptions of the structure of EHRs or the terminologies used. EHRs were used in primary, secondary and tertiary care. Data were recorded in EHRs by different groups of health care professionals. Secretarial staff also recorded data from dictation or nurses' or physicians' manual notes. Some information was also recorded by patients themselves; this information is validated by physicians. It is important that the needs and requirements of different users are taken into account in the future development of information systems. Several data components were documented in EHRs: daily charting, medication administration, physical assessment, admission nursing note, nursing care plan, referral, present complaint (e.g. symptoms), past medical history, life style, physical examination, diagnoses, tests, procedures, treatment, medication, discharge, history, diaries, problems, findings and immunization. In the future it will be necessary to incorporate different kinds of standardized instruments, electronic interviews and nursing documentation systems in EHR systems. The aspects of information quality most often explored in the studies reviewed were the completeness and accuracy of different data components. It has been shown in several studies that the use of an information system was conducive to more complete and accurate documentation by health care professionals. The quality of information is particularly important in patient care, but EHRs also provide important information for secondary purposes, such as health policy planning. Studies focusing on the content of EHRs are needed, especially studies of nursing documentation or patient self-documentation. One future research area is to compare the documentation of different health care professionals with the core information about EHRs which has been determined in national health projects. The challenge for ongoing national health record projects around the world is to take into account all the different types of EHRs and the needs and requirements of different health care professionals and consumers in the development of EHRs. A further challenge is the use of international terminologies in order to achieve semantic interoperability.
            • Record: found
            • Abstract: found
            • Article: not found

            The role of psychosocial factors in predicting the onset of chronic widespread pain: results from a prospective population-based study.

            Chronic widespread pain (CWP) is strongly associated with psychosocial distress both in a clinical setting and in the community. The aim of this study was to determine the contribution of measures of psychosocial distress, health-seeking behaviour, sleep problems and traumatic life events to the development of new cases of CWP in the community. In a population-based prospective study, 3171 adults aged 25-65 yrs free of CWP were followed-up 15 months later to identify those with new CWP. Baseline data were available on their scores from a number of psychological scales including Illness Attitude Scales (IAS), Somatic Symptom Checklist (SSC), Hospital Anxiety & Depression Scale, Sleep Problems Scale, and Life Events Inventory. 324 subjects [10%, 95% confidence interval (CI) 9.2, 11.3] developed new CWP at follow-up. After adjustment for age and sex, three factors independently predicted the development of CWP: scoring three or more on the SSC [odds ratio (OR) 1.8, 95% CI 1.1, 3.1], scoring eight or more on the Illness Behaviour subscale of the IAS (OR 3.3, 95% CI 2.3, 4.8), and nine or more on the Sleep Problem Scale (OR 2.7, 95% CI 1.6, 3.2). Subjects exposed to all three factors were at 12 times the odds of new CWP than those with low scores on all scales. Subjects are at substantial increased odds of developing CWP if they display features of somatization, health-seeking behaviour and poor sleep. Psychosocial distress has a strong aetiological influence on CWP.
              • Record: found
              • Abstract: found
              • Article: not found

              The health status burden of people with fibromyalgia: a review of studies that assessed health status with the SF-36 or the SF-12

               D. Hoffman,  E Dukes (2008)
              Objective The current review describes how the health status profile of people with fibromyalgia (FM) compares to that of people in the general population and patients with other health conditions. Methods A review of 37 studies of FM that measured health status with the 36-item Medical Outcomes Study Short-Form Health Survey (SF-36) or the 12-item Short-Form Health Survey (SF-12). Results Studies performed worldwide showed that FM groups were significantly more impaired than people in the general population on all eight health status domains assessed. These domains include physical functioning, role functioning difficulties caused by physical problems, bodily pain, general health, vitality (energy vs. fatigue), social functioning, role functioning difficulties caused by emotional problems and mental health. FM groups had mental health summary scores that fell 1 standard deviation (SD) below the general population mean, and physical health summary scores that fell 2 SD below the general population mean. FM groups also had a poorer overall health status compared to those with other specific pain conditions. FM groups had similar or significantly lower (poorer) physical and mental health status scores compared to those with rheumatoid arthritis, osteoarthritis, osteoporosis, systemic lupus erythematosus, myofacial pain syndrome, primary Sjögren's syndrome and others. FM groups scored significantly lower than the pain condition groups mentioned above on domains of bodily pain and vitality. Health status impairments in pain and vitality are consistent with core features of FM. Conclusions People with FM had an overall health status burden that was greater in magnitude compared to people with other specific pain conditions that are widely accepted as impairing. Review Criteria Studies in this review were identified through a search of electronic databases (MEDLINE: 1990–2006; EMBASE: 1990–2006). Search terms included: ‘fibromyalgia’, ‘health status’, ‘quality of life’, ‘SF-36’ and ‘SF-12’. Reference lists from published articles were also searched. Studies were selected if they were published in the English language between 1990 and (March) 2006 and assessed health status with a validated version of the SF-36 or the SF-12. Message for the Clinic Although FM is a controversial construct, studies performed worldwide showed that the health status profile of people with FM was remarkably consistent. People with FM had significant impairments in both mental and physical health status domains. People with FM had a poorer overall health status than people with specific pain conditions that are widely accepted as impairing.

                Author and article information

                J Pain Res
                J Pain Res
                Journal of Pain Research
                Journal of Pain Research
                Dove Medical Press
                10 June 2015
                : 8
                : 277-288
                [1 ]Pfizer Inc., New York, NY, USA
                [2 ]Pfizer Inc., Groton, CT, USA
                [3 ]Cedars-Sinai Medical Center, Los Angeles, CA, USA
                Author notes
                Correspondence: Birol Emir, Pfizer Inc., 235 East 42nd Street, New York, NY 10017-5755, USA, Tel +1 212 733 8581, Fax +1 212 351 1008, Email birol.emir@ 123456pfizer.com
                © 2015 Emir et al. This work is published by Dove Medical Press Limited, and licensed under Creative Commons Attribution – Non Commercial (unported, v3.0) License

                The full terms of the License are available at http://creativecommons.org/licenses/by-nc/3.0/. Non-commercial uses of the work are permitted without any further permission from Dove Medical Press Limited, provided the work is properly attributed.

                Original Research


                Comment on this article