Fixed-effects Logistic Regression Research Articles

Shift work, work time control, and informal caregiving as risk factors for sleep disturbances in an ageing municipal workforce.

Objectives:This study aimed to examine the contribution of shift work, work time control (WTC) and informal caregiving, separately and in combination, to sleep disturbances in ageing employees.Methods:Survey data were obtained from two prospective cohort studies with repeated measurements of working conditions, informal caregiving, and sleep disturbances. We used fixed-effect conditional logistic regression analysis to examine whether within-individual changes in shift work, WTC and informal caregiving were associated with changes in sleep. Secondary analyses included between-individuals comparison using standard logistic regression models. Results from the two cohorts were pooled using meta-analysis.Results:Low WTC and informal caregiving were associated with sleep disturbances in within-individual analyses [odds ratios (OR) ranging between 1.13 (95% confidence interval 1.01–1.27) and 1.48 (95% CI 1.29–1.68)] and in between-individuals analyses [OR 1.14 (95% CI 1.03–1.26) to 1.33 (1.19–1.49)]. Shift work alone was not associated with sleep disturbances, but accumulated exposure to shift work, low WTC and informal caregiving was associated with higher risk of sleep disturbances (OR range 1.21–1.76). For some of the sleep outcomes, informal caregiving was related to a higher risk of sleep disturbances when WTC was low and a lower risk when WTC was high.Conclusions:Informal caregiving and low WTC are associated with risk of sleep disturbances among ageing employees. The findings also suggest that low WTC in combination with informal caregiving may increase the risk of sleep disturbances whereas high WTC may alleviate the adverse impact of informal caregiving on sleep.

The use of branding strategies to increase the uptake of modern contraception among Kenyan youth

Building brand equity is a strategic process that can boost normative and behavior change campaigns. However, the use of brand equity within the social and health sciences remains an under-researched field in lower income countries. This work examines Shujaaz, a multi-channel communication platform that uses branding strategies to create positive associations with safe sexual and reproductive health (SRH) practices among Kenyan youth. This study has the following three goals: (1) define and measure Shujaaz brand equity and its components (i.e., loyalty, awareness, perceived quality, popularity and brand personality), (2) analyze if Shujaaz brand equity components are associated with changes in SRH practices among youth, and (3) explore if Shujaaz brand loyalty could be better defined as an outcome, not a component of brand equity. The study uses data from a survey among 450 Kenyan youth interviewed three times; in 2016, 2017, and 2018. The analytical sample includes only those participants who self-reported exposure to Shujaaz. Fixed-effect logistic regression models were used to analyze the association between measures of brand equity and SRH outcomes, and to measure the association between brand equity measures and brand loyalty. Regression estimates indicate that loyalty to the Shujaaz brand was significantly associated with ever using condoms ( OR = 1.79), ever using modern contraception ( OR = 1.95), and with the likelihood of recommending condoms ( OR = 1.88). Shujaaz brand personality was significantly associated with ever using a condom ( OR = 1.87). When analyzing brand loyalty as the outcome of interest, we observe that brand personality, brand awareness, and popularity are positively associated with loyalty. However, no significant association was observed for perceived brand quality. This study underscored the role of brand loyalty as a determinant of health behavior change. However, loyalty to the brand is strongly associated with popularity, awareness, and brand personality, highlighting the need for mediation models to further study the dynamics of brand equity.

The effect of a mystery shopper scheme on prescribing behavior in primary care: Results from a field experiment

BackgroundHealth care systems in many countries are characterized by limited availability of provider performance data that can be used to design and implement welfare improving reforms in the health sector. We question whether a simple mystery shopper scheme can be an effective measure to improve primary care quality in such settings.MethodsUsing a randomized treatment-control design, we conducted a field experiment in primary care clinics in a Chinese city. We investigate whether informing physicians of a forthcoming mystery shopper audit influences their prescribing behavior. The intervention effects are estimated using conditional fixed-effects logistic regression. The estimated coefficients are interpreted as marginal utilities in a choice model.ResultsOur findings suggest that the mystery shopper intervention reduced the probability of prescribing overall. Moreover, the intervention had heterogeneous effects on different types of drugs.ConclusionsThis study provides new evidence suggesting that announced performance auditing of primary care providers could directly affect physician behavior even when it is not combined with pay-for-performance, or measures such as reminders, feedback or educational interventions.

Sugar-sweetened beverage consumption and risk of hyperuricemia: a longitudinal analysis of the Health Workers Cohort Study participants in Mexico

The elevated consumption of sugar-sweetened beverages (SSBs) in Mexico is an important public health concern. However, the association between SSB consumption and hyperuricemia has been scarcely studied and not well documented. To prospectively evaluate the association between SSB consumption and risk of hyperuricemia in Mexican adults. A longitudinal analysis was conducted using data from the Health Workers Cohort Study. Participants were followed from 2004 to 2018, with measurements every 6 y. The analysis sample consisted of 1300 adults, aged 18 to 85 y. SSB consumption during the previous year was evaluated through a semiquantitative FFQ. Hyperuricemia was defined as a concentration of uric acid ≥7.0 mg/dL in men and ≥5.7 mg/dL in women. We evaluated the association of interest using 2 methodologies: fixed-effects logistic regression and generalized estimating equations (GEEs). Potential confounders were included in both approaches. At baseline, median intake of SSBs was 472.1 mL/wk (IQR: 198.8-1416.4 mL/wk), and 233 participants had hyperuricemia. Uric acid was higher in participants with an SSB intake ≥7 servings/wk, compared with those with an intake <1 serving/wk (P<0.001). Participants who changed from the lowest to the highest category of servings consumption experienced 2.6 increased odds of hyperuricemia (95% CI: 1.27, 5.26). Results from the GEE model indicated the odds of hyperuricemia increased by 44% (OR=1.44; 95% CI: 1.13, 1.84) in the 2-6 servings/wk group, and by 89% (OR=1.89; 95% CI: 1.39, 2.57) in the ≥7 servings/wk categories, compared with the <1 serving/wk category. Diet soft drinks were not associated with hyperuricemia. Our results suggest that the consumption of SSBs is associated with an increased risk of hyperuricemia in Mexican adults, but diet soft drink consumption is not, which supports the need to strengthen existing recommendations to reduce the intake of SSBs.The Health Workers Cohort Study (HWCS) has been approved by the Institutional Review Board of the Mexican Social Security Institute (12CEI 09 006 14), and the National Institute of Public Health of Mexico (13CEI 17 007 36).

Do Online Service-Pitchers’ Words of Certainty and Rapport-Building Boost Transactional Success?

In this study, via a web platform that provides a service-oriented labor market, we examined the communication qualities (certainty- and rapport building-words) of service-pitchers’ online bidding and their accompanying messages in response to job posts that ultimately got contracted during the time period of January 1-December 31, 2010. Focusing our analysis on the 587,547 seller-communications for posted jobs ending in service contracts enabled us to compare, for each contracted project, the communication-qualities of that contract’s selected versus unselected service-pitchers—an impossible comparison to make for job-postings that yielded no contract. Using conditional fixed-effects logistic regression, controlling for service-pitchers’ bidding price and other characteristics, we found that the certainty with which service-pitchers pitch their services is positively associated with being selected by service-seeking buyers up to a threshold, after which online service-pitchers‘ message-certainty harms their contract-acquisitions; and thus, the contract-winning benefits of online service-pitchers’ message-certainty was in the shape of an inverted U-shape, consistent with the “too much of a good thing” (TMGT)-effect. Also as hypothesized, we found that the TMGT- effect of online service-pitchers’ message-certainty was moderated by their extent of rapport-building, such that the online service- pitchers who accompanied words of certainty with more (rather than fewer) rapport-building words tended to be more likely selected and to be less prone to suffering backlash. These two patterns repeated, also as hypothesized, when examining how the message-certainty in selected online service-pitchers’ pitches affected their buyers’ satisfaction with their job-completion. Our findings’ theoretical and practical implications are discussed.

Sociodemographic, Ecological, and Spatiotemporal Factors Associated with Human Immunodeficiency Virus Drug Resistance in Florida: A Retrospective Analysis.

Persons living with human immunodeficiency virus (HIV) with resistance to antiretroviral therapy are vulnerable to adverse HIV-related health outcomes and can contribute to transmission of HIV drug resistance (HIVDR) when nonvirally suppressed. The degree to which HIVDR contributes to disease burden in Florida-the US state with the highest HIV incidence- is unknown. We explored sociodemographic, ecological, and spatiotemporal associations of HIVDR. HIV-1 sequences (n = 34 447) collected during 2012-2017 were obtained from the Florida Department of Health. HIVDR was categorized by resistance class, including resistance to nucleoside reverse-transcriptase , nonnucleoside reverse-transcriptase , protease , and integrase inhibitors. Multidrug resistance and transmitted drug resistance were also evaluated. Multivariable fixed-effects logistic regression models were fitted to associate individual- and county-level sociodemographic and ecological health indicators with HIVDR. The HIVDR prevalence was 19.2% (nucleoside reverse-transcriptase inhibitor resistance), 29.7% (nonnucleoside reverse-transcriptase inhibitor resistance), 6.6% (protease inhibitor resistance), 23.5% (transmitted drug resistance), 13.2% (multidrug resistance), and 8.2% (integrase strand transfer inhibitor resistance), with significant variation by Florida county. Individuals who were older, black, or acquired HIV through mother-to-child transmission had significantly higher odds of HIVDR. HIVDR was linked to counties with lower socioeconomic status, higher rates of unemployment, and poor mental health. Our findings indicate that HIVDR prevalence is higher in Florida than aggregate North American estimates with significant geographic and socioecological heterogeneity.

The Effects of International Financial Reporting Standards on Reporting Quality of Financial Institutions in Ghana

This study sought to examine the reporting quality of financial institutions in Ghana after adopting International Financial Reporting Standard (IFRS) as its official national reporting standard. Using a fixed effect logistic regression, the study compares the earnings management of banks and insurance firms before and after IFRS adoption on reporting quality. The data used was drawn from 51 financial institutions made up of 23 universal banks and 28 insurance companies observed over the period 2003 to 2014. The empirical results indicate that financial institutions exhibit more earnings management during the post-adoption era which is interpreted as a decline in the quality of financial reporting among financial institutions in Ghana. The results documented in this study add to the dearth of literature and contributes to the debate on IFRS adoption and its related impact on reporting quality (earnings management) among financial institutions from the perspective of an emerging market. The study is unique in the sense that it includes the insurance industry where the literature is largely silent especially, on the impact of IFRS adoption by countries on the African continent. Furthermore, unlike previous studies, this paper considers both listed and non-listed firms.

Decomposing Urban and Rural Disparities of Preventable ED Visits Among Patients With Alzheimer's Disease and Related Dementias: Evidence of the Availability of Health Care Resources.

The purpose of this study was to examine the urban and rural differences in the frequency of preventable Emergency Department (ED) visits among Alzheimer's Disease and Related Dementias (ADRD) patients, with a focus on the availability of health care resources in urban and rural areas. Linked datasets of 2015 State Emergency Department Databases from the Healthcare Cost and Utilization Project and the Area Health Resource File were used. ED discharges of 7 states were included in our analysis. We performed a state fixed-effect multivariable logistic regression to estimate the variation of preventable EDs by urban and rural areas. Individual characteristics and county-level health care resources were included in the estimation. The Oaxaca decomposition was used to quantify the association of county-level health care resources and urban/rural disparities. Rural patients with ADRD had 1.23 higher adjusted odds (P < .001) of going to the ED for a preventable visit compared to urban counterparts. The decomposition results showed that the model specification explained 49.2% of the differences between urban and rural patients. Patient residence in a mental health professional shortage area is one of the driving factors (contributing to 27%-48%) that explained the urban and rural disparities. Our study demonstrates the importance of improving health care resources in rural areas to improve health care quality and outcomes among ADRD patients who reside in rural areas. Future research and data collection on unobserved factors, such as health care quality, will be helpful in explaining the geographic differences.

Examining the pull, the push, and their simultaneous effects on managerial turnover

PurposeThis paper aims to examine the economic and psychological mechanisms in turnover at the managerial level. The paper investigates how (1) the ease of moving posed by alternative jobs (i.e. the economic mechanism) and (2) the desire to move due to low job satisfaction (i.e. the psychological mechanism) simultaneously influence top management team (TMT) turnover and these managers' subsequent job position and pay.Design/methodology/approachUsing 25 years of panel data on more than 2,000 top managers in the United States, the paper utilizes fixed-effects logistic regressions and the ordinary least squares model to test the hypotheses.FindingsThe authors find that CEO awards (an economic mechanism) and low compensation (a psychological mechanism) independently have positive effects on turnover. Turnover due to the economic mechanism leads to a higher position and pay, whereas turnover due to the psychological mechanism does not guarantee the same outcome. Further, when examining how pay dissatisfaction influences turnover simultaneously with CEO awards, the authors find that managers with the highest pay leave their firm, and not those with the lowest pay.Originality/valueThe paper employs the pull-and-push theory in the employee turnover literature and applies it to the top management team literature. By doing so, this paper contributes original insights to how economic and psychological mechanisms simultaneously affect managerial turnover and its subsequent outcomes.

Perception of aging in the relation between sport activity and self-rated health in middle and older age - A longitudinal analysis.

BackgroundThis study aims to generate more comprehensive knowledge regarding underlying mechanisms of health production in middle and older age by focusing on perception of aging in the relation between sport activity and self-rated health. More differentiated information can be generated by using more age-specific data and applying advanced empirical methods.MethodsPanel data from the German Ageing Survey (DEAS) (n = 1027) is used, whereby individuals are included, that participated at least in two waves. First, a blockwise fixed effects logistic regression model is carried out, which allows to investigate mediation as well as moderation effects and controls for time invariant unobserved heterogeneity. Additionally, two recursive bivariate probit models are conducted to identify further indirect pathways.FindingsA mediating and moderating effect of perception of aging regarding self-rated health is found in the fixed effects model. Indirect pathways of perception of aging and retirement over sport activity (mediator) are confirmed in the recursive bivariate probit models.ConclusionThe results highlight the complexity of the underlying mechanisms as well as the importance of individual and needs-based designed health promotion. Here, public health policy should not solely focus on sport activity, rather it should include target group specific strategies to shape individual perceptions of aging towards a more positive view on aging, while considering the heterogeneity of middle and older age groups.

Social inequalities in low birthweight outcomes in Sri Lanka: evidence from the Demographic and Health Survey 2016

ObjectiveTo investigate social inequalities underlying low birthweight (LBW) outcomes in Sri Lanka.DesignCross-sectional study.SettingThis study used the Sri Lanka Demographic and Health Survey 2016, the first such survey to cover the...

Uncontested: electoral competition in American school board elections

PurposeThe purpose of this research was to determine the extent to which American school board members faced electoral competition, as well as the factors influencing the likelihood of competition.Design/methodology/approachThe authors utilized original national survey data of American school board members linked with school district demographic data obtained from the National Center for Education Statistics. Several hypotheses were tested using three state-level fixed-effects logistic regression models predicting electoral competition.FindingsThe authors found that 39.6% of American school board members reported not having an opponent in their most recent election. School board members serving larger urban school districts with higher percentages of special needs students and racial minorities were more likely to have faced electoral competition.Originality/valueThe authors highlighted potential flaws in the traditional model of local democratic governance and helped expand understanding of the dissatisfaction theory of American democracy and continuous participation theory. The authors concluded with several suggestions on how the results can be used to inform future local governance reforms that increase electoral competition and/or create more effective governance models.

Beyond Financial Resources: The Role of Parents’ Education in Predicting Children’s Educational Persistence in Mexico

Despite significant educational expansion, Mexico’s educational attainment rates are relatively low. Though primary school enrollment is at nearly 100%, less than half of young adults ages 18-29 have finished upper secondary school (USE). This article examines how family-level factors, particularly parental education and household wealth, are associated with the likelihood of children dropping out of USE early in Mexico – a shift away from the well-established focus on primary education. Using region fixed effects logistic regressions, I examine the role of both mother’s and father’s education in predicting children’s educational persistence – and how this varies for boys and girls. Data is derived from a nationally representative sample of USE-aged youth in Mexico (n = 8,235). Results indicate that increases in parental education decrease the likelihood of children dropping out in upper secondary school, even when controlling for financial resources and other family- and household-level characteristics. Notably, these results vary across boys and girls.

Education elasticities of young women fertility in sub-Saharan Africa: Insights from Ethiopia, Rwanda, and Zimbabwe

ABSTRACT Although the relationship between education and fertility is well established, an understanding of young women’s fertility responses to education over time is needed to enhance a critical appraisal of education elasticities of fertility. This study thus comparatively investigates education effects on fertility in Ethiopia, Rwanda and Zimbabwe. Multivariate fixed effects logistic regression and direct decomposition methods were applied to Demographic and Health Survey data between 1999 and 2011. Results showed that declines in youth fertility in Ethiopia and Rwanda were driven by decrease in fertility rates of those with primary education but in the Zimbabwe youth fertility changes were driven by those with secondary or higher educational attainment. We conclude that education elasticities of fertility are not constant but vary by country and stage of fertility transition. Countries that are more advanced in fertility transition therefore need to place focus on enhancing post-secondary education to sustain fertility transition.

The association of select medical conditions with road transport and other hospitalised injury among older adults

Introduction: Certain cognitive and physical conditions have been associated with increased risk of injury, particularly risk of vehicle crashes among older car drivers. This study aims to examine the association of seven select medical conditions among hospitalised road users compared to other hospitalised injuries, and to estimate the hospitalised injury rates of car drivers, car passengers and pedestrians with these medical conditions. Method: An examination of road transport and nonroad transport hospitalised injury involving adults aged ≥50 years identified during 2003-2012 in New South Wales, Australia was conducted. Medical conditions were identified from hospital diagnosis records. Conditional fixed effects logistic regression conditioned on the matched cases and comparison-cohort estimated odds ratios for each medical condition by road user type. Results: There were 35,134 road transport injuries (10,664 car drivers and 4,907 pedestrians) and 447,858 nonroad transport injuries. Individuals with vision disorders, cardiovascular disease including stroke, diabetes, and osteoarthritis had higher odds of hospitalisation for an injury as a car driver compared to all other hospitalised injuries. Individuals with diagnoses of dementia or alcohol dependence had a lower odds of an injury hospital admission as a road user (excluding pedestrians) compared to all other hospitalised injuries. Conclusions: As the population ages, there are likely to be more older road users with comorbidities that may affect their ability to drive or safely cross the road. Community mobility strategies need to take into account the influence of comorbid health conditions for older adults.

Developing risk models for multicenter data using standard logistic regression produced suboptimal predictions: A simulation study.

Although multicenter data are common, many prediction model studies ignore this during model development. The objective of this study is to evaluate the predictive performance of regression methods for developing clinical risk prediction models using multicenter data, and provide guidelines for practice. We compared the predictive performance of standard logistic regression, generalized estimating equations, random intercept logistic regression, and fixed effects logistic regression. First, we presented a case study on the diagnosis of ovarian cancer. Subsequently, a simulation study investigated the performance of the different models as a function of the amount of clustering, development sample size, distribution of center‐specific intercepts, the presence of a center‐predictor interaction, and the presence of a dependency between center effects and predictors. The results showed that when sample sizes were sufficiently large, conditional models yielded calibrated predictions, whereas marginal models yielded miscalibrated predictions. Small sample sizes led to overfitting and unreliable predictions. This miscalibration was worse with more heavily clustered data. Calibration of random intercept logistic regression was better than that of standard logistic regression even when center‐specific intercepts were not normally distributed, a center‐predictor interaction was present, center effects and predictors were dependent, or when the model was applied in a new center. Therefore, to make reliable predictions in a specific center, we recommend random intercept logistic regression.

Multiple engagement: the relationship between informal care-giving and formal volunteering among Europe's 50+ population

AbstractThe article investigates the conditions of multiple engagement in the private and public realm in the second half of life. More specifically, I look at the relationship between informal care-giving and formal volunteering in a country-comparative way. Based on longitudinal data from the Survey of Health, Ageing and Retirement, 2004–2015, I investigate the 50+ population in 13 European countries. Controlling for unobserved heterogeneity by using conditional fixed-effect logistic regression models, I confirm earlier findings that care-givers are more likely to volunteer than non-care-givers; this effect is independent of care-giving intensity but only true for those who care outside their own household. As to macro-level influences, I find that both care-in-kind and cash-for-care expenditures increase the likelihood of volunteering among the 50+ population. The effect of cash-for-care expenditure is even stronger for the group of those who give intensive care outside their own households than for non-care-givers. Moreover, I find effects related to family's and women's role in society. First, I show a negative effect of a country's societal norm of family orientation on volunteering participation for those giving sporadic care outside their household but also among non-care-givers. Second, in countries with higher female labour market participation among the middle-aged, the volunteering likelihood is higher for sporadic female care-givers outside their own household but also among female non-care-givers.

Ranking hospitals when performance and risk factors are correlated: A simulation-based comparison of risk adjustment approaches for binary outcomes.

BackgroundThe conceptualization of hospital quality indicators usually includes some form of risk adjustment to account for hospital differences in case mix. For binary outcome variables like in-hospital mortality, frequently utilized risk adjusted measures include the standardized mortality ratio (SMR), the risk standardized mortality rate (RSMR), and excess risk (ER). All of these measures require the estimation of expected hospital mortality, which is often based on logistic regression models. In this context, an issue that is often neglected is correlation between hospital performance (e.g. care quality) and patient-specific risk factors. The objective of this study was to investigate the impact of such correlation on the adequacy of hospital rankings based on different measures and methods.MethodsUsing Monte Carlo simulation, the impact of correlation between hospital care quality and patient-specific risk factors on the adequacy of hospital rankings was assessed for SMR/RSMR, and ER based on logistic regression and random effects logistic regression. As an alternative method, fixed effects logistic regression with Firth correction was considered. The adequacies of the resulting hospital rankings were assessed by the shares of hospitals correctly classified into quintiles according to their true (unobserved) care qualities.ResultsThe performance of risk adjustment approaches based on logistic regression and random effects logistic regression declined when correlation between care quality and a risk factor was induced. In contrast, fixed-effects-based estimations proved to be more robust. This was particularly true for fixed-effects-logistic-regression-based ER. In the absence of correlation between risk factors and care quality, all approaches showed similar performance.ConclusionsCorrelation between risk factors and hospital performance may severely bias hospital rankings based on logistic regression and random effects logistic regression. ER based on fixed effects logistic regression with Firth correction should be considered as an alternative approach to assess hospital performance.

Controlling For Self-Selection Bias In Evaluating The Impact of Microcredit Programs: Evidence from Asa Bangladesh

Endogeneity problems such as self-selection and program placement bias are key issues in estimating the impact of microcredit programs. Self-selection problem occurs when borrowers select themselves in the program because they might have entrepreneurial skills, and risk taking abilities. Hence they could do better even without the credit. And, program placement bias occurs when a branch of microfinance institution placed in a developed village in order to become financially sustainable. This study focuses on overcoming these two biases while estimating the role of microcredit in alleviating poverty in Bangladesh. This study empirically estimates Foster-Greer-Thorbecke (FGT) poverty measures which are: incidence of poverty, poverty gap and squared poverty gap. To achieve the objectives of the study, primary data were collected from 2,598 households during the period of June 2014 to September 2014. The survey covers 24 districts out of 64 districts in Bangladesh. In addition to current borrowers, four types of control groups, non-borrowers, drop-outs, refused and pipeline borrowers, were interviewed to control for potential self-selection bias. The results of the FGT poverty measures indicate that the incidence of poverty, poverty gap and squared poverty gap are the lowest among borrowers. Then, this study applies fixed-effect logistic regression for measuring the impact of microcredit on the incidence of poverty and fixed-effect Tobit regression for estimating the poverty gap and squared poverty gap. In the empirical estimations, this study controls for age and education of the head of households, female (spouse) education, and age and gender composition of the household members. The results suggest that microcredit borrowers are less likely to be poor after being involved in the program. The results also indicate that microcredit program reduces poverty gap and squared poverty gap by respectively 3.3 and 1.2 percent even after controlling for self-selection and program placement bias. Therefore, this study claims that microcredit is an effective tool for poverty reduction.

Efficacy and Tolerance of Blended Diets in Children Receiving Gastrostomy Feeds.

To determine the prevalence of gastrointestinal (GI) symptoms in children receiving a blended diet via a gastrostomy tube. This is a single-center, retrospective study of children ages 1-18 years that received a blended diet. We reviewed demographics, anthropometrics, clinical characteristics, and rationale for switching to blended diet. Fixed-effects logistic regression analysis was used to evaluate the changes in patient symptoms over the 12-month follow-up period, and fixed-effects regression was employed to test for changes in anthropometrics. Twenty-three patients (8 female, 15 male) were identified, and data from 89 outpatient visits were analyzed. The most common underlying diagnosis was neurological disorder. Thirty-five percent of patients received commercial whole cow milk formulas, 30% received hydrolysate formulas, and 35% received amino acid-based formulas. After formula switches were made, 65% received homemade blended diets, 17.5% received commercial blended diets, and 17.5% received a combination of both. Median duration of time on a blended diet was 17 months. Ninety-five percent of patients who were previously experiencing upper GI symptoms improved within the first 3 months after blended diet initiation. Twenty-one percent of patients developed mild constipation on the diet, which was managed with increased water intake and/or polyethylene glycol. Only 2 patients discontinued the blended diet, because of inadequate weight gain and worsening of upper GI symptoms. In our study population, blended diets were well tolerated in gastrostomy-fed children and were associated with clinical improvement of upper GI symptoms.