Stata Command Research Articles

Client satisfaction with family planning services in the area of high unmet need: evidence from Tanzania Service Provision Assessment Survey, 2014-2015

BackgroundClient satisfaction has been found to be an important factor for the uptake and continuation of family planning services. This study aimed to examine the current status of and factors associated with client’s satisfaction with family planning services in Tanzania, which has a high unmet need for family planning.MethodsThe study used data from the Tanzania Service Provision Assessment survey of 2014–2015. A facility was classified as having high service readiness for FP if it scored at least 67.7% on a composite score based on three domains (staff training and guidelines, basic diagnostic equipment, and basic medicines), following criteria developed by the World Health Organization. The exit interview questionnaire was used to collect information from women about their level of satisfaction, whether “very satisfied,” “more or less satisfied,” or not satisfied with the services received. The response was dichotomized into “Yes” if the woman reported being very satisfied with services received otherwise coded as “No”. Unadjusted and adjusted logistic regression models were used to assess the association between the client satisfaction and covariate variables; service readiness, facility type, managing authority, location, management meetings, supervision, provider’s sex, and working experience, clients’ age and education. All analyses were weighted to correct for non-response, disproportionate and complex sampling by using the “SVY” command in Stata 14.ResultsOut of the 1188 facilities included in the survey, 427 (35.9%) provided family planning services. A total of 1746 women participated in observations and exit interviews. Few (22%) facilities had a high readiness to provide family planning services. While most facilities had the recommended equipment available, only 42% stocked contraceptives (e.g. oral pills, injectable contraceptives and/or condoms). Further, trained staff and clinical guidelines were present in only 30% of services. Nevertheless, the majority (91%) of clients reported that they were satisfied with services. In the multivariate analysis, a high service readiness score [AOR = 2.5, 95% CI; 1.1–6.0], receiving services from private facilities [AOR = 2.3, 95% CI; 1.1–5.0], and being in the age group 20 to 29 years [AOR = 0.3, 95% CI; 0.1–0.7] were all significantly associated with clients’ satisfaction with family planning services.ConclusionThere is a high level of client satisfaction with family planning services in Tanzania. Maintaining and exceeding this level will require improvements in the provision of staff training and the availability of contraceptives in existing services.

Linear Dynamic Panel-data Estimation Using Maximum Likelihood and Structural Equation Modeling

Panel data make it possible both to control for unobserved confounders and to include lagged, endogenous regressors. However, trying to do both simultaneously leads to serious estimation difficulties. In the econometric literature, these problems have been addressed by using lagged instrumental variables together with the generalized method of moments, while in sociology the same problems have been dealt with using maximum likelihood estimation and structural equation modeling. While both approaches have merit, we show that the maximum likelihood–structural equation models method is substantially more efficient than the generalized method of moments method when the normality assumption is met and that the former also suffers less from finite sample biases. We introduce the command xtdpdml, which has syntax similar to other Stata commands for linear dynamic panel-data estimation. xtdpdml greatly simplifies the structural equation model specification process; makes it possible to test and relax many of the constraints that are typically embodied in dynamic panel models; allows one to include time-invariant variables in the model, unlike most related methods; and takes advantage of Stata's ability to use full-information maximum likelihood for dealing with missing data. The strengths and advantages of xtdpdml are illustrated via examples from both economics and sociology.

How Many Imputations Do You Need? A Two-stage Calculation Using a Quadratic Rule.

When using multiple imputation, users often want to know how many imputations they need. An old answer is that 2-10 imputations usually suffice, but this recommendation only addresses the efficiency of point estimates. You may need more imputations if, in addition to efficient point estimates, you also want standard error (SE) estimates that would not change (much) if you imputed the data again. For replicable SE estimates, the required number of imputations increases quadratically with the fraction of missing information (not linearly, as previous studies have suggested). I recommend a two-stage procedure in which you conduct a pilot analysis using a small-to-moderate number of imputations, then use the results to calculate the number of imputations that are needed for a final analysis whose SE estimates will have the desired level of replicability. I implement the two-stage procedure using a new SAS macro called %mi_combine and a new Stata command called how_many_imputations.

Dealing with endogeneity bias: The generalized method of moments (GMM) for panel data

Endogeneity bias can lead to inconsistent estimates and incorrect inferences, which may provide misleading conclusions and inappropriate theoretical interpretations. Sometimes, such bias can even lead to coefficients having the wrong sign. Although this is a long-standing issue, it is now emerging in marketing and management science, with high-ranked journals increasingly exploring the issue. In this paper, we methodologically demonstrate how to detect and deal with endogeneity issues in panel data. For illustration purposes, we used a dataset consisting of observations over a 15-year period (i.e., 2002 to 2016) from 101 UK listed companies and examined the direct effect of R&D expenditures, corporate governance, and firms' characteristics on performance. Due to endogeneity bias, the result of our analyses indicates significant differences in findings reported under the ordinary least square (OLS) approach, fixed effects and the generalized method of moments (GMM) estimations. We also provide generic STATA commands that can be utilized by marketing researchers in implementing a GMM model that better controls for the three sources of endogeneity, namely, unobserved heterogeneity, simultaneity and dynamic endogeneity.

Measuring Individual Performance with Comprehensive Bibliometric Reports as an Alternative to h-Index Values.

The h-index is frequently used to measure the performance of single scientists in Korea (and beyond). No single indicator alone, however, is able to provide a stable and complete assessment of performance. The Stata command bibrep.ado is introduced which automatically produces bibliometric reports for single researchers (senior researchers working in the natural or life sciences). The user of the command receives a comprehensive bibliometric report which can be used in research evaluation instead of the h-index.

Estimation of Nested and Zero-Inflated Ordered Probit Models

We introduce three new STATA commands, nop, ziop2 and ziop3, for the estimation of a three-part nested ordered probit model, the two-part zero-inflated ordered probit models of Harris and Zhao (2007, Journal of Econometrics 141: 1073-1099) and Brooks, Harris and Spencer (2012, Economics Letters 117: 683-686), and a three-part zero-inflated ordered probit model for ordinal outcomes, with both exogenous and endogenous switching. The three-part models allow the probabilities of positive, neutral (zero) and negative outcomes to be generated by distinct processes. The zero-inflated models address a preponderance of zeros and allow them to emerge in different latent regimes. We provide postestimation commands to compute probabilistic predictions and various measures of their accuracy, to access the goodness of fit, and to perform model comparison using the Vuong test (Vuong 1989, Econometrica 57: 307-333) with the corrections based on the Akaike and Schwarz information criteria. We investigate the finite-sample performance of the maximum likelihood estimators by Monte Carlo simulations, discuss the relations among the models, and illustrate the new commands with an empirical application to the U.S. federal funds rate target.

Systematic review and meta-analysis

In this review the usual methods applied in systematic reviews and meta-analyses are outlined. The ideal hypothesis for a systematic review should be generated by information not used later in meta-analyses. The selection of studies involves searching in web repertories, and more than one should be consulted. A manual search in the references of articles, editorials, reviews, etc. is mandatory. The selection of studies should be made by two investigators on an independent basis. Data collection on quality of the selected reports is needed, applying validated scales and including specific questions on the main biases which could have a negative impact upon the research question. Such collection also should be carried out by two researchers on an independent basis. The most common procedures for combining studies with binary outcomes are described (inverse of variance, Mantel-Haenszel, and Peto), illustrating how they can be done using Stata commands. Assessment of heterogeneity and publication bias is also illustrated with the same program.

Running Event Studies Using Stata: The Estudy Command

This paper proposes the Stata command estudy and illustrates how it can be used to perform an event study customizing the statistical framework, from the estimates of abnormal returns to the tests for their statistical significance. Our program significantly improves the existing modules both in terms of completeness and users’ comprehension.

Resistance of Staphylococcus aureus to antimicrobial agents in Ethiopia: a meta-analysis

BackgroundEmergence of antimicrobial resistance by Staphylococcus aureus has limited treatment options against its infections. The purpose of this study was to determine the pooled prevalence of resistance to different antimicrobial agents by S. aureus in Ethiopia.MethodsWeb-based search was conducted in the databases of PubMed, Google Scholar, Hinari, Scopus and the Directory of Open Access Journals (DOAJ) to identify potentially eligible published studies. Required data were extracted and entered into Excel spread sheet. Statistical analyses were performed using Stata version 13.0. The metaprop Stata command was used to pool prevalence values. Twenty-one separate meta-analysis were done to estimate the pooled prevalence of the resistance of S. aureus to twenty-one different antimicrobial agents. Heterogeneity among the studies was assessed using the I2 statistic and chi-square test. Publication bias was assessed using Egger’s test. Because of significant heterogeneity amongst the studies, the random effects model was used to pool prevalence values.ResultsThe electronic database search yielded 1317 studies among which 45 studies met our inclusion criteria. Our analyses demonstrated very high level of resistance to amoxicillin (77% [95% confidence interval (CI): 68%, 0.85%]), penicillin (76% [95% CI: 67%, 84%]), ampicillin (75% [95% CI: 65%, 85%]), tetracycline (62% [95% CI: 55%, 68%]), methicillin (47% [95% CI: 33%, 61%]), cotrimoxaziole (47% [95% CI: 40%, 55%]), doxycycline (43% [95% CI: 26%, 60%]), and erythromycin (41% [95% CI: 29%, 54%]). Relatively low prevalence of resistance was observed with kanamycin (14% [95% CI: 5%, 25%]) and ciprofloxacin (19% [95% CI: 13%, 26%]). The resistance level to vancomycin is 11% 995% CI: (4%, 20%). High heterogeneity was observed for each of the meta-analysis performed (I2 ranging from 79.36% to 95.93%; all p-values ≤0.01). Eggers’ test did not show a significant publication bias for all antimicrobial agents except for erythromycin and ampicillin.ConclusionsS. aureus in Ethiopia has gotten notoriously resistant to almost to all of antimicrobial agents in use including, penicillin, cephalosporins, tetracyclines, chloramphenicol, methicillin, vancomycin and sulphonamides. The resistance level to vancomycin is bothersome and requires a due attention. Continued and multidimensional efforts of antimicrobial stewardship program promoting rational use of antibiotics, infection prevention and containment of AMR are urgently needed.

Rate Decomposition for Aggregate Data Using Das Gupta's Method

Social, behavioral, and health scientists frequently decompose changes or differences in outcome variables into components of change and assess their relative importance. Many Stata commands facilitate this exercise using unit-level data, notably by applying the Blinder–Oaxaca approach. However, none of the comparable user-written commands decompose changes or differences in aggregate data despite their availability and the widespread use of corresponding decomposition techniques. In this article, I present the user-written command rdecompose, which decomposes aggregate or cross-classified data based on Das Gupta's (1993, Standardization and Decomposition of Rates: A User's Manual, Volume 1) approach, and demonstrate its application in multiple settings. This command extends the original method by allowing multiple factors and flexible functional specifications.

The Estimation and Modeling of Cause-specific Cumulative Incidence Functions Using Time-dependent Weights

Competing risks occur in survival analysis when an individual is at risk of more than one type of event and the occurrence of one event precludes the occurrence of any other event. A measure of interest with competing risks data is the cause-specific cumulative incidence function (CIF) which gives the absolute (or crude) risk of having the event by time t, accounting for the fact that it is impossible to have the event if a competing event is experienced first. The user written command, stcompet, calculates non-parametric estimates of the cause-specific CIF and the official Stata command, stcrreg, fits the Fine and Gray model for competing risks data. Geskus (2011) has recently shown that some of the key measures in competing risks can be estimated in standard software by restructuring the data and incorporating weights. This has a number of advantages as any tools developed for standard survival analysis can then be used for the analysis of competing risks data. This paper describes the stcrprep command that restructures the data and calculates the appropriate weights. After using stcrprep a number of standard Stata survival analysis commands can then be used for the analysis of competing risks. For example, sts graph, failure will give a plot of the cause-specific CIF and stcox will fit the Fine and Gray proportional subhazards model. Using stcrprep together with stcox is computationally much more efficient than using stcrreg. In addition, the use of stcrprep opens up new opportunities for competing risk models. This is illustrated by fitting flexible parametric survival models to the expanded data to directly model the cause-specific CIF.

Longitudinal multiple imputation approaches for body mass index or other variables with very low individual-level variability: the mibmi command in Stata

BackgroundIn modern health care systems, the computerization of all aspects of clinical care has led to the development of large data repositories. For example, in the UK, large primary care databases hold millions of electronic medical records, with detailed information on diagnoses, treatments, outcomes and consultations. Careful analyses of these observational datasets of routinely collected data can complement evidence from clinical trials or even answer research questions that cannot been addressed in an experimental setting. However, ‘missingness’ is a common problem for routinely collected data, especially for biological parameters over time. Absence of complete data for the whole of a individual’s study period is a potential bias risk and standard complete-case approaches may lead to biased estimates. However, the structure of the data values makes standard cross-sectional multiple-imputation approaches unsuitable. In this paper we propose and evaluate mibmi, a new command for cleaning and imputing longitudinal body mass index data.ResultsThe regression-based data cleaning aspects of the algorithm can be useful when researchers analyze messy longitudinal data. Although the multiple imputation algorithm is computationally expensive, it performed similarly or even better to existing alternatives, when interpolating observations.ConclusionThe mibmi algorithm can be a useful tool for analyzing longitudinal body mass index data, or other longitudinal data with very low individual-level variability.

Characteristics of Inpatient Hypertension Cases and Factors Associated with Admission Outcomes in Ashanti Region, Ghana: An Analytic Cross-Sectional Study.

Background Hypertension remains a cause of morbidity and mortality in the Ashanti Region of Ghana. It has been featured in the top ten causes of OPD attendance, admissions, and deaths since 2012. We investigated the sociodemographic characteristics and spatial distribution of inpatient hypertensives and factors associated with their admission outcomes. Methods A 2014 line list of 1715 inpatient HPT cases aged ≥25 years was used for the cross-sectional analytic study. Accounting for clustering, all analyses were performed using the “svy” command in Stata. Frequencies, Chi-square test, and logistic regression analysis were used in the analysis. Arc view Geographic Information System (ArcGIS) was used to map the density of cases by place of residence and reporting hospital. Results Mean age of cases was 58 (S.D 0.0068). Females constituted 67.6% of the cases. Age, gender, and NHIS status were significantly associated with admission outcomes. Cases were clustered in the regional capital and bordering districts. However, low case densities were recorded in the latter. Conclusion Increasing NHIS access can potentially impact positively on hypertension admission outcomes. Health educational campaigns targeting men are recommended to address hypertension-related issues.

A Simple Command to Calculate Travel Distance and Travel Time

Obtaining the routing distance between two addresses should not be a hassle in the current state of technology. This is unfortunately more complicated than it first seems. Recently, several Stata commands have been implemented for this purpose (traveltime, traveltime3, mqtime, osrmtime), but most of them went out of order only a few months after their introduction or appear as complicated to use. In this paper, we introduce the new command georoute to retrieve travel distance and travel time between two points, defined either by their addresses or by their geographical coordinates. Compared to other existing commands, we argue it is simple to use, efficient in terms of computational speed, and versatile regarding the information that can be provided as input.

Simcadi: Similarity Indices for Categorical Distributions

In this article we introduce the Stata command simcadi. It helps to calculate indicators for the similarity of categorically ordered variables and distribution, respectively. In particular, it permits the calculation of the Cosine index, and indices introduced by Finger and Kreinin (1979), Bray and Curtis (1957), Dice (1945), Sorenson (1948), Jaccard (1912), Grubel and Lloyd (1971), Ruzicka (1958), and Gower (1971). Moreover, it allows us to compute the development of a distribution over time. The command offers various options for an efficient handling of datasets, because it permits the calculation of benchmarks of comparison automatically, and the incoporation of complex weighting schemes.

Cardiovascular risk in patients with psoriasis, psoriatic arthritis, and rheumatoid arthritis: a prospective study of data from the Secure Anonymised Information Linkage databank

BackgroundThe increased risk of cardiovascular disease in individuals with rheumatoid arthritis, psoriatic arthritis, and psoriasis is well known. However, the precise cardiovascular risk in psoriasis and psoriatic arthritis is less understood, particularly the relative contribution of traditional cardiovascular risk factors, such as obesity, smoking, and systemic inflammation as measured by C-reactive protein and erythrocyte sedimentation rate. The objective of this study was to compare the incidence of major adverse cardiac events (MACE) among patients with rheumatoid arthritis, psoriatic arthritis, and psoriasis with population controls adjusting for traditional cardiovascular risk factors, systemic inflammation, and disease-modifying anti-rheumatic drug (DMARD) treatment to better define cardiovascular risk in patients with these conditions. MethodsThe study used linked, routinely collected Welsh health data from 1999 to 2013 available from the Secure Anonymised Information Linkage (SAIL) databank, which includes datasets from general practice (GP), hospitals, and rheumatology clinics. Incidence and first occurrence of a MACE was investigated in individuals who were present in the GP system and positive for Read codes for rheumatoid arthritis (n=8650), psoriatic arthritis (2128), psoriasis (24 630), and controls without codes for the conditions investigated (1 187 706). We controlled for traditional cardiovascular risk factors, such as age, sex, smoking, body-mass index, comorbidities (hyperlipidaemia, diabetes, and hypertension), and medication (non-steroidal anti-inflammatory drugs, DMARDs, and steroids). We used a Cox proportional hazard model with the st command in Stata (version 13). FindingsWhen controlling for traditional risk factors, cardiovascular risk was significantly increased for individuals with rheumatoid arthritis (number of MACE events 548, hazard ratio 1·2, 95% CI 1·0–1·3; p=0·038) and psoriasis (879, 1·1, 1·0–1·3; p=0·025), but not for those with psoriatic arthritis (43, 1·0, 0·7–1·5; p=0·887). No interaction between DMARDs and MACE occurrence was observed. InterpretationThere is an increased incidence of cardiovascular disease for patients with rheumatoid arthritis and psoriasis, but not for those with psoriatic arthritis. This finding demonstrates the varying mediators of cardiovascular risk across these conditions and highlights the importance of cardiovascular risk reduction strategies, especially for patients with rheumatoid arthritis and psoriasis, who may benefit from closer monitoring to help reduce the incidence of MACE. FundingSupported by a Pfizer Inflammation Competitive Research Programme grant. The funder had no involvement in the study.

Authors' response

Authors' response

Ownership Increase in International Joint Ventures: The Within- and Across-Country Flexibility Perspective

We employ the real options perspective to examine how multinational corporations (MNCs) increase ownership levels in their international joint ventures (IJVs) to exploit (or utilize) both within- and across-country flexibility. This paper utilize a rich dataset of South Korean firms’ foreign affiliates to test hypotheses derived from the real options theory. It examines the roles of growth options that require small equity stakes under uncertainty as well as switching options that require controlling equity stakes to coordinate switching implementation. Specifically, we examine how the ownership of jointly-owned foreign subsidiaries changes under the influence of changing labor cost uncertainty. We utilize a two-stage model to address the sample selection bias associated with a subsidiary’s choice of joint venture under the influence of high labor cost uncertainty. In the first stage, we run a probit model with a dummy variable for JVs and wholly-owned subsidiaries (WOSs) (1: JV, 0: WOS) as the dependent variable. In the second stage, we run a panel logistic regression model (STATA command, “xtlogit”). We found that MNCs engaged in greenfield IJVs increase their initial ownership when labor cost uncertainty within their host countries is favorably resolved. We also found that MNCs’ tendency to increase their IJV ownership intensifies when their subsidiary networks are exposed to negatively correlated labor cost growth rates across countries. Taking within and across country flexibility perspective, this study adds new insights on MNC ownership changes, especially in IJVs. Based upon our findings, we conclude that MNCs make flexible ownership adjustments for their foreign subsidiaries in response to different uncertainty conditions within and across countries.

Dynsimpie: A Command to Examine Dynamic Compositional Dependent Variables

In this article, we adapt the modeling strategy proposed by Philips, Rutherford, and Whitten (2016, American Journal of Political Science 60: 268–283) and create a user-friendly Stata command, dynsimpie. This command requires the installation of the clarify package of Tomz, Wittenberg, and King (2003, Journal of Statistical Software 8(1): 1–30) and uses the commands in the clarify package to produce estimates from models of compositional dependent variables over time. Users can also examine how counterfactual shocks play through the system with graphs that are easy to interpret. We illustrate this with a model of voter support for the three dominant political parties in the UK.

Versatile Tests for Comparing Survival Curves Based on Weighted Log-rank Statistics

The log-rank test is perhaps the most commonly used nonparametric method for comparing two survival curves and yields maximum power under proportional hazards alternatives. While the assumption of proportional hazards is often reasonable, it need not hold. Several authors have therefore developed versatile tests using combinations of weighted log-rank statistics that are more sensitive to nonproportional hazards. Fleming and Harrington (1991, Counting Processes and Survival Analysis, Wiley) consider the family of Gρ statistics and their supremum versions, while Lee (1996, Biometrics 52: 721-725) and Lee (2007, Computational Statistics and Data Analysis 51: 6557-6564) propose tests based on the more extended Gρ,γ family. In this article, I consider Zm = max(| Z1 |, |Z2|, | Z3|), where Z1, Z2, and Z3 are z statistics obtained from G0,0, G1,0, and G0,1 tests, respectively. G0,0 corresponds to the log-rank test, while G1,0 and G0,1 are more sensitive to early and late-difference alternatives. I conduct a simulation study to compare the performance of Zm with the log-rank test, the more optimally weighted test, and Lee's (2007) tests, under the null hypothesis, proportional hazards, early difference, and late-difference alternatives. Results indicate that the method based on Zm maintains the type I error rate, provides increased power relative to the log-rank test under early difference and late-difference alternatives, and entails only a small to moderate power loss compared with the more optimally chosen test. I apply the procedure to two datasets reported in the literature, both of which exhibit nonproportional hazards. Versatile tests such as Zm may be useful in clinical trial settings where there is concern that the treatment effect may not conform to the proportional hazards assumption. I also describe the syntax for a Stata command, verswlr, to implement the method.