Complex modeling with detailed temporal predictors does not improve health records-based suicide risk prediction.

Suicide risk prediction models can identify individuals for targeted intervention. Discussions of transparency, explainability, and transportability in machine learning presume complex prediction models with many variables outperform simpler models. We compared random forest, artificial neural network, and ensemble models with 1500 temporally defined predictors to logistic regression models. Data from 25,800,888 mental health visits made by 3,081,420 individuals in 7 health systems were used to train and evaluate suicidal behavior prediction models. Model performance was compared across several measures. All models performed well (area under the receiver operating curve [AUC]: 0.794-0.858). Ensemble models performed best, but improvements over a regression model with 100 predictors were minimal (AUC improvements: 0.006-0.020). Results are consistent across performance metrics and subgroups defined by race, ethnicity, and sex. Our results suggest simpler parametric models, which are easier to implement as part of routine clinical practice, perform comparably to more complex machine learning methods.

A value proposition for pragmatic clinical trials.

The authors examine the origin, benefits, and challenges of pragmatic clinical trials to assess the ultimate value of this research design.

Suicide Deaths Before and During the Coronavirus Disease 2019 Pandemic: An Interrupted Time-series Study.

INTRODUCTION: With stressors that are often associated with suicide increasing during the coronavirus disease 2019 (COVID-19) pandemic, there has been concern that suicide mortality rates may also be increasing. Our objective was to determine whether suicide mortality rates increased during the COVID-19 pandemic. METHODS: We conducted an interrupted time-series study using data from January 2019 through December 2020 from 2 large integrated health care systems. The population at risk included all patients or individuals enrolled in a health plan at HealthPartners in Minnesota or Henry Ford Health System in Michigan. The primary outcome was change in suicide mortality rates, expressed as annualized crude rates of suicide death per 100,000 people in 10 months following the start of the pandemic in March 2020 compared with the 14 months prior. RESULTS: There were 6,434,675 people at risk in the sample, with 55% women and a diverse sample across ages, race/ethnicity, and insurance type. From January 2019 through February 2020, there was a slow increase in the suicide mortality rate, with rates then decreasing by 0.45 per 100,000 people per month from March 2020 through December 2020 (SE=0.19, P=0.03). CONCLUSIONS: Overall suicide mortality rates did not increase with the pandemic, and in fact slightly declined from March to December 2020. Our findings should be confirmed across other settings and, when available, using final adjudicated state mortality data.

Connecting Research and Practice: Implementation of Suicide Prevention Strategies in Learning Health Care Systems.

The health care systems affiliated with the Mental Health Research Network strive to be learning health care systems that identify and address evidence gaps of importance to clinicians, patients, and funders. This column describes how research guides clinical care and clinical care guides research in the area of suicide prevention as well as some of the challenges of conducting embedded research.

Association Between Patterns of Alcohol Use and Short-Term Risk of Suicide Attempt Among Patients With and Without Reported Suicidal Ideation.

OBJECTIVE: To evaluate the association between patterns of alcohol use and short-term risk of suicide attempt among patients with and without reported suicidal ideation. METHODS: Kaiser Permanente Washington electronic health record data were used to identify mental health visits (1/1/2010-6/30/2015) with documented assessments for unhealthy alcohol use (AUDIT-C) and suicidal ideation (PHQ-9 ninth question). Logistic regression fit using generalized estimating equations were used to conduct visit-level analyses, accounting for correlation between individuals' assessments. Separate models evaluated the association between (1) level of alcohol consumption and (2) frequency of heavy episodic drinking (HED), in combination with suicidal ideation (any vs none), with suicide attempt within 90 days following each visit. Primary models adjusted for age, gender, race/ethnicity and visit year. RESULTS: Of 59,705 visits (43,706 unique patients), 372 (0.62%) were followed by a suicide attempt within 90 days. The risk of suicide attempt was significantly higher for patients reporting suicidal ideation across all levels of alcohol consumption compared to patients reporting low-level alcohol use and no suicidal ideation, particularly high-level use (OR 9.77, 95% CI, 6.23-15.34). Similarly, risk of suicide attempt was higher for patients reporting suicidal ideation across all levels of HED relative to those reporting no HED or suicidal ideation, particularly HED monthly or more (OR 6.80, 95% CI 4.77-9.72). Among patients reporting no suicidal ideation, no associations were observed. CONCLUSIONS: Findings underscore the potential value of offering alcohol-related care to patient reporting suicidal ideation. Additional strategies are needed to identify suicide risk among those reporting no suicidal ideation.

Short-term risk of suicide attempt associated with patterns of patient-reported alcohol use determined by routine AUDIT-C among adults receiving mental healthcare.

OBJECTIVE: To evaluate the association between alcohol use routinely reported during outpatient mental healthcare visits and short-term risk of subsequent suicide attempt. METHODS: Using a longitudinal retrospective-cohort design, electronic health records identified adult outpatient visits to a mental health provider (1/1/2010-6/30/2015) at Kaiser Permanente Washington with a documented Alcohol Use Disorders Identification Test-Consumption [AUDIT-C]. Suicide attempts within 90 days of AUDIT-C documentation were defined using death certificate cause-of-death and diagnosis codes (non-lethal). Visit-level analyses used generalized estimating equations to account for correlation between multiple AUDIT-Cs for individuals. Separate models evaluated the association between (1) level of consumption and (2) frequency of heavy drinking episodes and suicide attempts, adjusted for visit year, demographics, depressive symptom, and suicidal ideation. RESULTS: Of 59,382 patient visits, 0.62% (N = 371) were followed by a suicide attempt within 90 days. Patients reporting high-level alcohol use were 1.77 times (95% CI, 1.22-2.57) more likely to attempt suicide than those reporting low-level use. Patients reporting daily or almost daily heavy drinking episodes were 2.33 times (95% CI, 1.38-3.93) more likely to attempt suicide than those reporting none. CONCLUSIONS AND RELEVANCE: The AUDIT-C is a valuable tool for assessing patterns of patient-reported alcohol use associated with subsequent suicide attempt.

If You Listen, I Will Talk: the Experience of Being Asked About Suicidality During Routine Primary Care.

BACKGROUND: Routine population-based screening for depression is an essential part of evolving health care models integrating care for mental health in primary care. Depression instruments often include questions about suicidal thoughts, but how patients experience these questions in primary care is not known and may have implications for accurate identification of patients at risk. OBJECTIVES: To explore the patient experience of routine population-based depression screening/assessment followed, for some, by suicide risk assessment and discussions with providers. DESIGN: Qualitative, interview-based study. PARTICIPANTS: Thirty-seven patients from Kaiser Permanente Washington who had recently screened positive for depression on the 2-item Patient Health Questionnaire [PHQ] and completed the full PHQ-9. APPROACH: Criterion sampling identified patients who had recently completed the PHQ-9 ninth question which asks about the frequency of thoughts about self-harm. Patients completed semi-structured interviews by phone, which were recorded and transcribed. Directive and conventional content analyses were used to apply knowledge from prior research and elucidate new information from interviews; thematic analysis was used to organize key content overall and across groups based on endorsement of suicide ideation. KEY RESULTS: Four main organizing themes emerged from analyses: (1) Participants believed being asked about suicidality was contextually appropriate and valuable, (2) some participants described a mismatch between their lived experience and the PHQ-9 ninth question, (3) suicidality disclosures involved weighing hope for help against fears of negative consequences, and (4) provider relationships and acts of listening and caring facilitated discussions about suicidality. CONCLUSIONS: All participants believed being asked questions about suicidal thoughts was appropriate, though some who disclosed suicidal thoughts described experiencing stigma and sometimes distanced themselves from suicidality. Direct communication with trusted providers, who listened and expressed empathy, bolstered comfort with disclosure. Future research should consider strategies for reducing stigma and encouraging fearless disclosure among primary care patients experiencing suicidality.

Predicting Suicide Attempts for Racial and Ethnic Groups of Patients During Routine Clinical Care.

The aim of this study was to examine variation in suicide ideation and its relationship to risk of suicide attempt in the subsequent 90 days by race and ethnicity. Participants were adults who completed the Patient Health Questionnaire depression module (PHQ9) during an outpatient encounter between January 10, 2010, and December 12, 2012 (N = 509,945 patients; N = 1,228,308 completed PHQ9). Data came from the Virtual Data Warehouse from four health care systems in the Mental Health Research Network. The sample was majority female (73.7%), primarily 30-64 years old (60.1%), healthy (64.5% comorbidity index = 0), and over half were non-Hispanic White (52.9%). Only Asian patients (OR: 1.31; 95% CI: 1.24, 1.39) had higher odds of reporting suicide ideation when compared to non-Hispanic White people. All racial and ethnic groups had increased risk for suicide attempt with increased frequency of suicide ideation. The PHQ9 item 9 can be used as an indicator of suicide ideation and risk for suicide attempt up to 90 days after the reported ideation in racial and ethnic minority patients during routine clinical care.

Are wishes for death or suicidal ideation symptoms of depression in older adults?

OBJECTIVE: Clinicians may question whether thoughts of being better off dead are normal consequences of aging or symptoms of depression. We examine whether thoughts of suicide are as strongly linked to depression severity in older adults as they are in other age groups. METHODS: Cross-sectional cohort study. Participants included 509,945 outpatients >18 years old from four large integrated healthcare systems in the Mental Health Research Network who completed 1.2 million Patient Health Questionnaires (PHQ) and had data to calculate Charlson Comorbidity Index scores from 2010 through 2012. The PHQ8 estimated depression severity, while suicidal ideation was measured using the 9th item of the PHQ. Data were abstracted from a Virtual Data Warehouse. RESULTS: In older adult patients, suicidal ideation was strongly associated with depression severity. Older adults who had at least moderately-severe depression (PHQ8 ≥15) were 48 times more likely (95% CI: 42.8-53.8) to report suicidal ideation than those with minimal or mild symptoms of depression (PHQ8 <10) after adjustment for all other variables in the model, including medical comorbidity burden. CONCLUSIONS: Depression severity was by far the strongest predictor of suicidal ideation in older adult patients. Older patients with suicidal ideation should be screened for depression.

The impact of race and ethnicity on rates of return to psychotherapy for depression.

BACKGROUND: There are many limitations with the evidence base for the role of race and ethnicity in continuation of psychotherapy for depression. METHODS: The study sample consisted of 242,765 patients ≥ 18 years old from six healthcare systems in the Mental Health Research Network (MHRN) who had a new episode of psychotherapy treatment for depression between 1/1/2010 and 12/31/2013. Data were from electronic medical records and organized in a Virtual Data Warehouse (VDW). The odds of racial and ethnic minority patients returning for a second psychotherapy visit within 45 days of the initial session were examined using multilevel regression. RESULTS: The sample was primarily middle aged (68%, 30-64 years old), female (68.5%), and non-Hispanic white (50.7%), had commercial insurance (81.4%), and a low comorbidity burden (68.8% had no major comorbidities). Return rates within 45 days of the first psychotherapy visit were 47.6%. Compared to their non-Hispanic white counterparts, racial and ethnic minority patients were somewhat less likely to return to psychotherapy for a second visit (adjusted odds ratios [aORs] ranged from 0.80 to 0.90). Healthcare system was a much stronger predictor of return rates (aORs ranged from 0.89 to 5.53), while providers accounted for 21.1% of the variance in return rates. CONCLUSIONS: Provider and healthcare system variation were stronger predictors of patient return to psychotherapy than race and ethnicity. More research is needed to understand why providers and healthcare systems determine psychotherapy return rates for patients of all racial and ethnic groups.

Pragmatic (trial) informatics: a perspective from the NIH Health Care Systems Research Collaboratory.

Pragmatic clinical trials (PCTs) are research investigations embedded in health care settings designed to increase the efficiency of research and its relevance to clinical practice. The Health Care Systems Research Collaboratory, initiated by the National Institutes of Health Common Fund in 2010, is a pioneering cooperative aimed at identifying and overcoming operational challenges to pragmatic research. Drawing from our experience, we present 4 broad categories of informatics-related challenges: (1) using clinical data for research, (2) integrating data from heterogeneous systems, (3) using electronic health records to support intervention delivery or health system change, and (4) assessing and improving data capture to define study populations and outcomes. These challenges impact the validity, reliability, and integrity of PCTs. Achieving the full potential of PCTs and a learning health system will require meaningful partnerships between health system leadership and operations, and federally driven standards and policies to ensure that future electronic health record systems have the flexibility to support research.

Suicidal ideation reported on the PHQ9 and risk of suicidal behavior across age groups.

OBJECTIVE: The Joint Commission recommends all patients be screened for suicide. However, differences in suicide attempt and death rates may affect how well tools predict risk across age groups. Our objective was to determine whether item 9 of the Patient Health Questionnaire (PHQ9) predicts risk for suicide attempts and deaths across age groups. METHODS: PHQ9s completed by adult outpatients treated for mental health conditions in 2010-2012 at four Mental Health Research Network-affiliated healthcare systems were used to measure depression severity and suicidal ideation. Suicide attempts were identified via ICD-9 codes and suicide deaths via ICD-10 codes and state death certificates. RESULTS: In all, 939,268PHQ9s were completed by 297,290 outpatients. Compared to those without, those with nearly daily suicidal ideation were 5-to-8 times more likely to attempt suicide and 3-to-11times more likely to die by suicide within 30 days, and 2-to-4times more likely to attempt suicide and 2-to-6 times more likely to die by suicide within 365 days. The increased risk of suicide death for those with any level of suicidal ideation persisted over two years. The relationships between suicide thoughts and attempts and deaths were similar across age groups. LIMITATIONS: Our sample was limited to outpatients completing a PHQ9 and relied on pre-existing clinical and administrative data. CONCLUSIONS: Suicidal ideation reported on the PHQ9 was a robust predictor of suicide attempts and deaths regardless of age, and this increased risk persisted for two years. Healthcare systems should address both the immediate and sustained risk for suicide for patients of all ages.

ANTIDEPRESSANT ADHERENCE ACROSS DIVERSE POPULATIONS AND HEALTHCARE SETTINGS.

BACKGROUND: Early adherence is key to successful depression treatment, but nearly 60% of patients discontinue antidepressants within 3 months. Our study aimed to determine factors associated with poor early adherence to antidepressants in a large diverse sample of patients. METHODS: Six Mental Health Research Network healthcare systems contributed data for adults with depression and a new antidepressant start, defined by a washout period of at least 270 days, between January 1, 2010 and December 31, 2012. Pharmacy fill and self-reported race/ethnicity data were obtained from the electronic medical record. Patients had early adherence if they had a second antidepressant fill within 180 days of the first. We used logistic regression to investigate the relationship between early adherence and patient characteristics. RESULTS: A total of 177,469 adult patients had 184,967 new episodes of depression with a filled antidepressant prescription. Patients refilled their antidepressants within 180 days in 71% of episodes. Race/ethnicity was a strong predictor of early adherence, with patients from racial/ethnic minorities other than Native Americans/Alaskan Natives less likely (adjusted odd ratios 0.50-0.59) to refill their antidepressants than non-Hispanic whites. Age, neighborhood education, comorbidity burden, provider type and engagement in psychotherapy were also associated with adherence. Other apparent predictors of early adherence, including neighborhood income, gender, and prior mental health hospitalizations, were no longer significant in the fully adjusted model. CONCLUSIONS: Race/ethnicity was a robust predictor of early antidepressant adherence, with minority groups other than Native Americans/Alaskan Natives less likely to be adherent. Further research is needed to determine whether early nonadherence in specific minority populations is intentional, due to side effects or patient preference, or unintentional and appropriate for targeted interventions to improve adherence.

Learning About Missing Data Mechanisms in Electronic Health Records-based Research: A Survey-based Approach.

BACKGROUND: Bias due to missing data is a major concern in electronic health record (EHR)-based research. As part of an ongoing EHR-based study of weight change among patients treated for depression, we conducted a survey to investigate determinants of missingness in the available weight information and to evaluate the missing-at-random assumption. METHODS: We identified 8,345 individuals enrolled in a large EHR-based health care system who had monotherapy treatment for depression from April 2008 to March 2010. A stratified sample of 1,153 individuals completed a detailed survey. Logistic regression was used to investigate determinants of whether a patient (1) had an opportunity to be weighed at treatment initiation (baseline), and (2) had a weight measurement recorded. Parallel analyses were conducted to investigate missingness during follow-up. Throughout, inverse-probability weighting was used to adjust for the design and survey nonresponse. Analyses were also conducted to investigate potential recall bias. RESULTS: Missingness at baseline and during follow-up was associated with numerous factors not routinely collected in the EHR including whether or not the patient had ever chosen not to be weighed, external weight control activities, and self-reported baseline weight. Patient attitudes about their weight and perceptions regarding the potential impact of their depression treatment on weight were not related to missingness. CONCLUSION: Adopting a comprehensive strategy to investigate missingness early in the research process gives researchers information necessary to evaluate key assumptions. While the survey presented focuses on outcome data, the overarching strategy can be applied to any and all data elements subject to missingness.

Relationship between Obesity, Depression, and Disability in Middle-Aged Women.

BACKGROUND: Obesity and depression are closely linked, and each has been associated with disability. However, few studies have assessed inter-relationships between these conditions. DESIGN AND METHODS: In this study, 4 641 women aged 40-65 completed a structured telephone interview including self-reported height and weight, the Patient Health Questionnaire (PHQ) assessment of depression, and the World Health Organization Disability Assessment Schedule II (WHODAS II). The survey response rate was 62%. We used multivariable regression models to assess relationships between obesity, depression, and disability. RESULTS: The mean age was 52 years; 82% were white; and 80% were currently employed. One percent were underweight, 39% normal weight, 27% overweight, and 34% obese. Mild depressive symptoms were present in 23% and moderate-to-severe symptoms were present in 13%. After multivariable adjustment, depression was a strong independent predictor of worse disability in all 7 domains (cognition, mobility, self-care, social interaction, role functioning, household, and work), but obesity was only a significant predictor of greater mobility, role-functioning, household, and work limitations (P<0.05) (overweight was not significantly associated with any disability domain). Overall, the effect on disability was stronger and more pervasive for depression than obesity, and there was no significant interaction between the two conditions (P>0.05). Overweight and obesity were associated with 5 760 days of absenteeism per 1 000 person-years, and depression was associated with 18 240 days of absenteeism per 1 000 person-years. CONCLUSIONS: The strong relationships between depression, obesity and disability suggest that these conditions should be routinely screened and treated among middle-aged women.

Relationship between obesity, depression, and disability in middle-aged women.

BACKGROUND: Obesity and depression are closely linked, and each has been associated with disability. However, few studies have assessed inter-relationships between these conditions. DESIGN AND METHODS: In this study, 4641 women aged 40-65 completed a structured telephone interview including self-reported height and weight, the Patient Health Questionnaire (PHQ) assessment of depression, and the World Health Organization Disability Assessment Schedule II (WHODAS II). The survey response rate was 62%. We used multivariable regression models to assess relationships between obesity, depression, and disability. RESULTS: The mean age was 52 years; 82% were White; and 80% were currently employed. One percent were underweight, 39% normal weight, 27% overweight, and 34% obese. Mild depressive symptoms were present in 23% and moderate-to-severe symptoms were present in 13%. After multivariable adjustment, depression was a strong independent predictor of worse disability in all 7 domains (cognition, mobility, self-care, social interaction, role functioning, household, and work), but obesity was only a significant predictor of greater mobility, role-functioning, household, and work limitations (P < 0.05) (overweight was not significantly associated with any disability domain). Overall, the effect on disability was stronger and more pervasive for depression than obesity, and there was no significant interaction between the two conditions (P > 0.05). Overweight and obesity were associated with 5760 days of absenteeism per 1000 person-years, and depression was associated with 18,240 days of absenteeism per 1000 person-years. CONCLUSIONS: The strong relationships between depression, obesity and disability suggest that these conditions should be routinely screened and treated among middle-aged women.

Does telephone care management help Medicaid beneficiaries with depression?

OBJECTIVES: While telephone care management has shown promise as a cost-effective approach to manage patients with depression, there is little evidence on the effectiveness of this method for Medicaid beneficiaries in managed care. This study examines a 1-year telephone care management intervention designed to help this low-income, hard-to-reach population enter and remain engaged with treatment. STUDY DESIGN: A randomized controlled trial of 499 Rhode Island Medicaid managed care beneficiaries with depression (all parents, average age of 35, and 90% women). Care managers conducted telephonic outreach with the intervention group to establish a relationship, initiate treatment, make referrals for in-person psychotherapy and/or medication treatment, and monitor treatment progress. The control group received usual care and was given a referral list of providers participating in the Medicaid program. METHODS: Primary outcomes were the use of health services and depression severity at 6 and 18 months. Administrative claims provided information on medical and mental health services use. Surveys of sample members provided information on depression severity. Analysis controlling for sociodemographic characteristics was done to assess the effectiveness of providing care management. RESULTS: Care managers contacted 91% of those assigned to the intervention group. The intervention was effective in enrolling participants into mental health services (42% in intervention group vs 31% in control; P = .05), but did not successfully reduce average depression severity. CONCLUSIONS: The intervention's lack of success in reducing depression severity for Medicaid beneficiaries suggests the need for more intensive interventions that strengthen telephone care management and potentially include in-person components as well.

The association between obesity, depression, and educational attainment in women: the mediating role of body image dissatisfaction.

OBJECTIVE: We examine the mediating role of body image dissatisfaction (BID) on the association between obesity and depression and the variation of this association as a function of years of education among a population-based sample of women aged 40-65 years. METHODS: A series of sample-weighted logistic regression models were used to estimate the associations between obesity, BID, and depression, stratified by educational attainment. Data were obtained from a structured telephone interview of 4543 female health plan enrollees, including self-reported height and weight, the Patient Health Questionnaire assessment of depression, and a single-item measure of BID. RESULTS: Among those with <16 years of education, in both the unadjusted and adjusted models, obesity and BID were significantly associated with depression. Similarly, among those with ≥ 16 years of education, obesity and BID were significantly associated with depression in the unadjusted models. However, in the adjusted model, only BID was associated with depression. A formal test for mediation suggests that the association between obesity and depression was mediated by BID regardless of level of education. CONCLUSIONS: Our data suggest that BID-mediated the obesity-depression association. In addition, obesity and BID may be salient risk factors for depression among middle-aged women as a function of the level of education.

Validity of clinical body weight measures as substitutes for missing data in a randomized trial.

BACKGROUND: Long-term follow-up of weight loss interventions is essential, but collecting weights can be difficult, and self-reports inaccurate. We examined the relationship between weight measures obtained in the context of a weight loss trial and in routine clinical care. METHODS: Body weight data from a trial of behavioral obesity treatment among 88 obese women and 203 women age 40-65 years with comorbid obesity and depression were compared against weight data entered into an electronic medical record (EMR) during routine clinical care. Study and EMR weights and weight changes were then compared at 6 and 12 months using scatterplots, Pearson's correlations, and t-tests. RESULTS: The 12-month follow-up rate for this trial was 77%. Among the 224 12-month completers, 142 women (63%) had an EMR weight within 90 days of their 12-month study weight. Study and EMR weights were highly correlated (0.99), with a mean difference of 0.1 kg. The correlation between two measures of 12-month weight change using study and EMR weights was 0.96. These results were robust to sensitivity analyses that explored the impact of different-sized windows for matching clinical weights with study weights. Among the 67 women who were missing study weights at 12 months, 33 (49%) had an EMR weight available within 90 days of their missed follow-up appointment. CONCLUSIONS: Weight measures routinely obtained in clinical care are highly correlated with those obtained by trained research staff and may be used, without statistical correction, to achieve higher rates of long-term follow-up in weight loss studies.

Validity of Clinical Body Weight Measures as Substitutes for Missing Data in a Randomized Trial.

BACKGROUND: Long-term follow-up of weight loss interventions is essential, but collecting weights can be difficult, and self-reports inaccurate. We examined the relationship between weight measures obtained in the context of a weight loss trial and in routine clinical care. METHODS: Body weight data from a trial of behavioral obesity treatment among 88 obese women and 203 women age 40 to 65 years with comorbid obesity and depression were compared against weight data entered into an electronic medical record (EMR) during routine clinical care. Study and EMR weights and weight changes were then compared at 6 and 12 months using scatterplots, Pearson's correlations, and t-tests. RESULTS: The 12-month follow-up rate for this trial was 77%. Among the 224 12-month completers, 142 women (63%) had an EMR weight within 90 days of their 12-month study weight. Study and EMR weights were highly correlated (0.99), with a mean difference of 0.1 kg. The correlation between two measures of 12-month weight change using study and EMR weights was 0.96. These results were robust to sensitivity analyses that explored the impact of different-sized windows for matching clinical weights with study weights. Among the 67 women who were missing study weights at 12 months, 33 (49%) had an EMR weight available within 90 days of their missed follow-up appointment. CONCLUSIONS: Weight measures routinely obtained in clinical care are highly correlated with those obtained by trained research staff and may be used, without statistical correction, to achieve higher rates of long-term follow-up in weight loss studies.