Screening for high utilizing somatizing patients using a prediction rule derived from the management information system of an HMO: a preliminary study

Early identification of persistent somatic symptoms in primary care: data-driven and theory-driven predictive modelling based on electronic medical records of Dutch general practices

OBJECTIVE

The present study aimed to early identify patients with persistent somatic symptoms (PSS) in primary care by exploring routine care data-based approaches.

DESIGN/SETTING

A cohort study based on routine primary care data from 76 general practices in the Netherlands was executed for predictive modelling.

PARTICIPANTS

Inclusion of 94 440 adult patients was based on: at least 7-year general practice enrolment, having more than one symptom/disease registration and >10 consultations.

METHODS

Cases were selected based on the first PSS registration in 2017-2018. Candidate predictors were selected 2-5 years prior to PSS and categorised into data-driven approaches: symptoms/diseases, medications, referrals, sequential patterns and changing lab results; and theory-driven approaches: constructed factors based on literature and terminology in free text. Of these, 12 candidate predictor categories were formed and used to develop prediction models by cross-validated least absolute shrinkage and selection operator regression on 80% of the dataset. Derived models were internally validated on the remaining 20% of the dataset.

RESULTS

All models had comparable predictive values (area under the receiver operating characteristic curves=0.70 to 0.72). Predictors are related to genital complaints, specific symptoms (eg, digestive, fatigue and mood), healthcare utilisation, and number of complaints. Most fruitful predictor categories are literature-based and medications. Predictors often had overlapping constructs, such as digestive symptoms (symptom/disease codes) and drugs for anti-constipation (medication codes), indicating that registration is inconsistent between general practitioners (GPs).

CONCLUSIONS

The findings indicate low to moderate diagnostic accuracy for early identification of PSS based on routine primary care data. Nonetheless, simple clinical decision rules based on structured symptom/disease or medication codes could possibly be an efficient way to support GPs in identifying patients at risk of PSS. A full data-based prediction currently appears to be hampered by inconsistent and missing registrations. Future research on predictive modelling of PSS using routine care data should focus on data enrichment or free-text mining to overcome inconsistent registrations and improve predictive accuracy.

Identifying persistent somatic symptoms in electronic health records: exploring multiple theory-driven methods of identification

OBJECTIVE

Persistent somatic symptoms (PSSs) are defined as symptoms not fully explained by well-established pathophysiological mechanisms and are prevalent in up to 10% of patients in primary care. The present study aimed to explore methods to identify patients with a recognisable risk of having PSS in routine primary care data.

DESIGN

A cross-sectional study to explore four identification methods that each cover part of the broad spectrum of PSS was performed. Cases were selected based on (1) PSS-related syndrome codes, (2) PSS-related symptom codes, (3) PSS-related terminology and (4) Four-Dimensional Symptom Questionnaire scores and all methods combined.

SETTING

Coded electronic health record data were extracted from 76 general practices in the Netherlands.

PARTICIPANTS

Patients who were registered for at least 1 year during 2014-2018, were included (n=169 138).

OUTCOME MEASURES

Identification methods were explored based on (1) PSS sample sizes and demographics, (2) presence of chronic conditions and (3) healthcare utilisation (HCU) variables. Overlap between methods and practice specific differences were examined.

RESULTS

The percentage of cases identified varied between 0.3% and 7.0% across the methods. Over 58.1% of cases had chronic physical condition(s) and over 33.8% had chronic mental condition(s). HCU was generally higher for cases selected by any method compared with the total cohort. HCU was higher for method B compared with the other methods. In 26.7% of cases, cases were selected by multiple methods. Overlap between methods was low.

CONCLUSIONS

Different methods yielded different patient samples which were general practice specific. Therefore, for the most comprehensive data-based selection of PSS cases, a combination of methods A, C and D would be recommended. Advanced (data-driven) methods are needed to create a more sensitive algorithm for identifying the full spectrum of PSS. For clinical purposes, method B could possibly support screening of patients who are currently missed in daily practice.

Health care utilisation changes among Alaska Native adults after participation in an indigenous community programme to address adverse life experiences: a propensity score-matched analysis

The aim of this study was to investigate whether participation in Family Wellness Warriors Initiative (FWWI), an Alaska Native program that addresses adverse life experiences, is associated with changes in health care utilisation. The study method was a propensity score-matched cohort analysis using retrospective electronic health record data from Southcentral Foundation, a tribal health care system, from 2012 to 2017. Ninety p\articipants in FWWI trainings were identified as the intervention cohort and were propensity matched with 90 people who participated in other emotional wellness-related interventions. The primary outcome was the number of total health system visits. Secondary outcomes included emergency department (ED) visits, substance-use visits and visits with somatisation potential. After adjustment for covariates, FWWI participants showed a 36% reduction in total system visits (incidence rate ratio 0.64, 95% CI 0.49-0.84) and a 70% reduction in substance use visits (incidence rate ratio 0.30, 95% CI 0.10-0.93) when compared to the control. FWWI participants showed a 40% reduction in ED visits (incidence rate ratio 0.60, 95%CI 0.35-1.02) when compared to the control that was borderline significant. No significant differences were found for visits with somatisation potential (incidence rate ratio 1.25, 95% CI 0.79-1.99).

Estimating the prevalence of medically unexplained symptoms from primary care records

OBJECTIVES

To develop models to estimate the likely prevalence of medically unexplained symptoms (MUS) and severe MUS in a primary care practice from existing patient electronic records collected in the previous 2 years for secondary prevention and commissioning of psychological treatment.

STUDY DESIGN

Cross-sectional survey comparing general practitioners' (GPs) assessment of the presence or absence of MUS and severe MUS with clinical, demographic and service use variables associated with MUS or functional somatic syndromes from previous research in the patient's routine electronic record over the previous 2 years.

METHODS

Seventeen GPs from eight practices identified cases of MUS and severe MUS in 828 consecutive consulters in primary care. Models of variables associated with MUS and severe MUS were constructed using multivariate multilevel logistic regression. The predictive validity of the final models was tested, comparing predicted with observed data and expected prevalence rates from the literature.

RESULTS

Models to predict MUS and severe MUS had areas under the receiver operating characteristic curve of 0.70 [95% confidence interval (CI) 0.65-0.74] and 0.76 (95% CI 0.70-0.82), respectively. Both models showed adequate goodness of fit with observed data, and had good predictive validity compared with the expected prevalence of MUS, severe MUS, and anxiety or depression.

CONCLUSION

Models to predict the prevalence of MUS and severe MUS from routine practice records for commissioning purposes were successfully developed, but they require independent validation before general use. The sensitivity of these models was too low for use in clinical screening.

Health care utilization and poor reassurance: potential predictors of somatoform disorders

OBJECTIVE

Somatoform disorders are common conditions, but the current diagnostic criteria are considered to be unreliable, based largely on medically unexplained symptoms. DSM-5 is considering other possible characteristics of somatizers including high utilization, dissatisfaction with care, and poor response to reassurance. This paper reviews the available literature for evidence to support these criteria, and evaluates if distinctive aspects of these characteristics exist in somatizers.

METHODS

The Pubmed database was searched combining terms such as “somatoform disorder” with “reassurance,” “satisfaction,” and “utilization.” Articles were individually inspected.

RESULTS

Many studies report a deficit in long-term response to reassurance in somatizers; there was some evidence that patients respond initially to reassurance, followed by return of anxiety, leading to further reassurance seeking. There was insufficient evidence to support poor satisfaction with care as a characteristic of somatizers. While there is no standard criterion for high utilization, regardless of definition, evidence was found to support over-utilization, particularly in outpatient visits. However, no unique pattern of utilization was found that could identify somatizers within a broader group of high utilizers.

CONCLUSIONS

This review revealed evidence of over-utilization in many areas of healthcare, as well as poor long term response to reassurance in somatizers. Dissatisfaction with care, though, was not a consistent finding. It is difficult to study alternative diagnostic criteria for somatoform patients when the current criteria rest on so many problematic assumptions. Future research should attempt to validate criteria empirically in patient groups, with selection not based on medically unexplained symptoms.

Primary care physicians treat somatization

BACKGROUND

We hypothesized that somatizing patients managed by primary care physicians (PCP) would improve with a relationship-based intervention.

METHODS

We randomized 30 adults with medically unexplained symptoms to treatment or usual care. Four PCPs were trained to intervene with cognitive-behavioral, pharmacological, and patient-centered management and deployed the intervention with seven scheduled visits over 12 months. Outcomes obtained at baseline and 12 months were: Mental component summary (MCS), the primary endpoint, and measures of physical and psychological symptoms and of satisfaction with the PCP.

RESULTS

Patients averaged 52.5 years; 83.3% were female; 79.6% were black. Using a difference of differences approach, we found that the intervention produced a large effect size (ES) (0.82; CI: 0.08 to 1.57) for the MCS in the predicted direction, similar to the ES for physical (-0.80; CI: -1.55 to -0.04) and psychological (-1.06; CI: -1.83 to -0.28) improvement and for increased satisfaction with the PCP (0.94; CI: 0.15 to 1.74). Using ANCOVA in a sensitivity analysis, we found that the ES fell slightly (0.59), while other measures were unchanged.

CONCLUSIONS

Moderate-large effect sizes support the hypothesis that PCPs can effectively treat somatization. This points to the importance of performing a full RCT.

Diagnostic accuracy of predicting somatization from patients' ICD-9 diagnoses

OBJECTIVES

To hypothesize in a new and different population that administrative database (ADB) screening would identify somatizing patients by increasing numbers of visits, female gender, and greater percent of International Classification of Diseases, 9th Edition (ICD-9) primary diagnosis codes in musculoskeletal, nervous, gastrointestinal (GI), and ill-defined body systems. We labeled these codes as having "somatization potential." Our earlier study demonstrated that ICD-9 codes and other data from the ADB effectively identified somatization.

METHODS

Using a prospective observational design in a staff model health maintenance organization, we evaluated 1364 patients aged 18 to 65 years who had > or =8 visits yearly in the 2 years before study. Clinician raters applied a reliable method of medical chart review to identify patients meeting the criteria for somatization. We randomly selected 2/3 for the derivation set (n = 901) for logistic regression to evaluate the contribution of potential ADB correlates (age, gender, all encounters, primary diagnosis codes (ICD-9), revenue codes, and charges) of a diagnosis of somatization. This prediction rule was then applied to the remaining 1/3 of subjects, the validation set (n = 463).

RESULTS

Patients averaged 47.1 years, 12.8 visits per year, and 71.6% were female; 319 had somatization. Age, visits, and somatization potential were associated with clinician-rated somatization, with a c-statistic 0.72 in the derivation set and 0.68 in the validation set.

CONCLUSIONS

These data support our earlier findings that selected ICD-9 diagnoses in the ADB predict somatization, suggesting their potential in identifying a common, costly, and usually unrecognized problem.

Current somatoform disorders in Norway: prevalence, risk factors and comorbidity with anxiety, depression and musculoskeletal disorders

BACKGROUND

The future existence of somatoform disorders (SDs) has recently been debated. The objectives of this study were to investigate the prevalence of current SDs (defined as the presence of multisomatoform disorder [MSD] or somatoform disorders not otherwise specified [SDnos], without psychosocial impairment) and severe current SDs (MSD or SDnos with psychosocial impairment) in Norway. Differences in markers of severe current SDs, anxiety/depression and self-reported musculoskeletal disorders were explored. In addition, psychological distress and utilization of healthcare in subclasses (defined according to comorbidity with anxiety, depression and musculoskeletal disorders) of severe current SDs were examined.

METHODS

We interviewed 1,247 respondents using the Composite International Diagnostic Interview (CIDI) in the Oslo-Lofoten general population survey in 2000-2001. Six-month prevalence rates (%) and 95% confidence intervals (CIs) for current SDs were investigated by gender and age. Risk factors of disorders, psychological distress, healthcare utilization and use of medication were explored using logistic regression analyses.

RESULTS

The overall prevalence rate for severe current SDs was 10.2%. When psychosocial impairment was excluded as a criterion, the rate increased to 24.6%. Anxiety was strongly correlated with severe current SDs. Comorbidity of severe current SDs with anxiety/depression was 45%, and with musculoskeletal disorders, 43%. Analysis of healthcare utilization and use of medication showed that the presence of a comorbid psychiatric condition was more important than the presence of somatoform disorders alone.

CONCLUSION

Somatoform symptoms alone (with no psychiatric comorbidity) should not be considered a psychiatric disorder.

Classification and diagnosis of patients with medically unexplained symptoms

Patients with medically unexplained symptoms (MUS) have little or no demonstrable disease explanation for the symptoms, and comorbid psychiatric disorders are frequent. Although common, costly, distressed, and often receiving ill-advised testing and treatments, most MUS patients go unrecognized, which precludes effective treatment. To enhance recognition, we present an emerging perspective that envisions a unitary classification for the entire spectrum of MUS where this diagnosis comprises severity, duration, and comorbidity. We then present a specific approach for making the diagnosis at each level of severity. Although our disease-based diagnosis system dictates excluding organic disease to diagnose MUS, much exclusion can occur clinically without recourse to laboratory or consultative evaluation because the majority of patients are mild. Only the less common, "difficult" patients with moderate and severe MUS require investigation to exclude organic diseases. By explicitly diagnosing and labeling all severity levels of MUS, we propose that this diagnostic approach cannot only facilitate effective treatment but also reduce the cost and morbidity from unnecessary interventions.

Distinctive Patterns of Medical Care Utilization in Patients Who Somatize

BACKGROUND

Somatizing patients have maladaptive and increased rates of medical care utilization. If there were a way of routinely identifying such patients, one that did not require intensive, case-by-case review, they could be targeted for specific interventions to improve their use of medical care.

OBJECTIVE

We sought to identify patterns of medical care utilization that would distinguish somatizing and nonsomatizing medical outpatients with acceptable sensitivity and specificity.

DESIGN

Subjects completed questionnaires assessing somatization and sociodemographic characteristics. Their medical care utilization was obtained for the 12 months preceding the index visit. We then used multivariable logistic regression and recursive partitioning to identify patients with a provisional diagnosis of somatoform disorder. These exploratory models used various patterns of medical care utilization and sociodemographic characteristics as the independent variables.

SUBJECTS

We studied consecutive adults attending 2 primary care practices on randomly chosen days.

MEASURES

The provisional diagnosis of a somatoform disorder was assessed with a 15-item self-report questionnaire. The number of primary care visits, specialty visits, mental health visits, emergency visits, and inpatient and outpatient costs were obtained for the 12 months preceding the index visit from our hospital's automated medical records, which also provided a rating of aggregate medical morbidity. Self-reported utilization outside our hospital system was obtained from a subsample of patients.

RESULTS

Complete data were obtained on 1440 patients. Somatizing patients had more specialty care than primary care visits, higher outpatient than inpatient costs, and more emergency visits than nonsomatizing patients. A regression model containing 7 measures of utilization and 4 sociodemographic characteristics distinguished somatizing and nonsomatizing patients with a c-statistic = 0.73. Recursive partitioning identified 10 terminal nodes with a very high specificity (99%) but a very low sensitivity (15%).

CONCLUSIONS

We identified 7 discrete patterns of medical care utilization that distinguished somatizing and nonsomatizing patients. However, they did so with only modest specificity and sensitivity. This algorithm might be used effectively as the first step in a 2-step screening procedure whose second step would entail more intensive screening or individual, case-by-case review to identify somatizing patients in primary care practice.

Multisomatoform disorder: agreement between patient and physician report of criterion symptom explanation

OBJECTIVE

To determine the level of agreement between patients and expert physicians in whether criterion multisomatoform (MSD) symptoms are explained.

METHODS

We systematically collected reports from 280 primary care patients about whether they had suffered from any of 15 criterion MSD symptoms in the past month and whether they had received a medical explanation from a physician for positive criterion symptoms. The research team compared MSD symptom diagnoses derived from patient report with MSD symptom diagnoses derived from an expert physician report.

RESULTS

MSD symptom diagnoses derived from patient report had 98.7% sensitivity, 97.9% specificity, 89.3% positive predictive value, and 99.7% negative predictive value compared with MSD symptom diagnoses derived from an expert physician report. Analysis demonstrated that 15.0% of patients met symptom criteria for MSD, according to patient and physician report; 83.0% failed to meet symptom criteria for MSD, according to patient and physician report; 1.8% of patients met symptom criteria for MSD, according to patient report but not physician report, while 0.2% met symptom criteria for MSD according to physician report but not by patient report.

CONCLUSION

Patients demonstrated high agreement with a physician expert in somatization about whether criterion MSD symptoms are explained, suggesting revised screeners may accurately identify somatizing patients.

A method for rating charts to identify and classify patients with medically unexplained symptoms

BACKGROUND

As part of conducting a randomized control trial (RCT) to treat chronically high utilizing patients with medically unexplained symptoms (MUS), we developed the chart rating method reported here to identify and classify MUS subjects.

METHOD

Intended at this point only as a research tool, the method is comprehensive, uses explicit guidelines, and requires clinician raters. It distinguishes primary organic disease patients from those with primary MUS, quantifies medical comorbidities in primary MUS patients, and also distinguishes subgroups among MUS patients that we call somatization (resembles DSM-IV somatoform disorders) and minor acute illness (MAI) which differs from DSM-IV somatoform definitions. Scoring rules are used to generate the diagnoses above. The rules may be set according to the investigator's needs, from highly sensitive to highly specific.

RESULTS

We found high levels of agreement with the gold standard for MUS vs. organic disease (97.6%) and among raters for the key individual chart elements rated (92-96%). The method identified 206 MUS subjects and the extent of their medical comorbidities for entry into a RCT. It also identified somatization and MAI; the latter supports the validity of this newly reported MAI syndrome.

CONCLUSION

We concluded that this method offered research potential for identifying MUS patients, for quantifying their medical comorbidities, and for classifying MUS subgroups.

Using nurse practitioners to implement an intervention in primary care for high-utilizing patients with medically unexplained symptoms

Patients with medically unexplained symptoms (MUS) often are a source of frustration for clinicians, and despite high quality biomedical attention and frequent diagnostic tests, they have poor health outcomes. Following upon progress in depression treatment approaches, we developed a multidimensional treatment protocol for deployment by primary care personnel. This multi-faceted intervention for MUS patients emphasized cognitive-behavioral principles, the provider-patient relationship, pharmacological management, and treating comorbid medical diseases. We deployed it in an HMO using nurse practitioners (NP) to deliver the intervention to 101 patients, while 102 controls continued to receive medical care from their usual primary care physician. Successful deployment of the intervention required training the NPs, continuing support for the NPs in their management of this difficult population, and establishing strong communication links with the HMO. This paper addresses the practical considerations of using primary care personnel to implement a complex intervention in primary care, and it includes a discussion of special challenges encountered as well as solutions developed to overcome them.

The somatization in primary care study: a tale of three diagnoses

Somatization is a common phenomenon that has been defined in many ways. The two most widely used diagnoses, Somatization Disorder (SD) and Abridged Somatization Disorder (ASD), are based on lifetime unexplained symptoms. However, reports indicate instability in lifetime symptom recall among somatizing patients. Multisomatoform disorder (MSD) is a new diagnosis based on current unexplained symptoms. To understand how knowledge about SD and ASD translates to MSD, we examined the diagnostic concordance, impairment and health care utilization of these groups in a sample from the Somatization in Primary Care Study. The diagnostic concordance was high between MSD and SD, but lower between MSD and ASD. All three groups reported considerable physical impairment (measured using the PCS subscale of the SF-36). The mental health (MCS) scores for the three groups were only slightly lower than those of the general population. Over the course of one year, physical functioning fell significantly for all three groups. Mental functioning did not change significantly for any of the three groups over this period. Utilization patterns were very similar for the three groups. The high prevalence, serious impairment, and worsening physical functioning over the course of one year suggest the importance of developing interventions in primary care to alleviate the impaired physical functioning and reduce utilization in somatizing patients. MSD should be a useful diagnosis for targeting these interventions because it identifies a sizable cohort of somatizing patients reporting impairment of comparable severity to full SD, using a more efficient diagnostic algorithm based on current symptoms.