Research Flashcards
What was your thesis and elevator pitch?
We questioned the education, training and level of comfort of Ob/Gyn residents in managing shoulder dystocia and sought to determine their clinical experience and competence. We wanted to address the gap that we believed existed between their training and competency. We created a 26-question web-based survey using SurveyMonkey and electronically disseminated them to program directors of 273 ACGME-accredited obstetrics and gynecology residency programs who were then asked to forward the survey to residents in their program. The responses were recorded and data was analyzed.
Why did you do this study?
As fellows, I spent a lot of time supervising the residents both inpatient and outpatient. As I was covering hospitalist calls, there were many shoulder dystocia emergencies, and I came to realize that some of my senior residents did not appear comfortable managing these situations. They would either briefly attempt certain maneuvers then step aside asking for help, get overwhelmed by the heat of the moment or simply not know how to perform maneuvers such as the delivery of the posterior arm. Shoulder dystocia is an unpredictable and often unpreventable obstetrical emergency which not only requires a heightened degree of suspicion but systematic training to be prepared. Therefore, I decided to evaluate the problem and see if a solution would exist.
What statistics did you use and why did you use them?
Independent sample t-test was used to compare the means of two independent (unrelated) groups in order to determine whether there is statistical evidence that the associated population means, on the same continuous, dependent variable are significantly different. For example, we tried to understand whether gender independently affected their experience in training.
Dependent variable – measured on a continuous scale (interval or ratio, ie: IQ level, weight)
Independent variable – consists of two categorical, independent groups (ie: age, gender)
ANOVA (analysis of variance) was used to compare the means of three or more independent groups. For example, we tried to understand whether PGY level independently affected their experience in training.
Chi square (x2) test was used to analyze the nominal (cannot be ordered) qualitative (not in #’s) data to evaluate the differences between groups (pgy level, gender)
Vs ordinal data (in order ie: size)
Vs quantitative data (requires a numerical scale using a bell curve, 1SD 68%, 2SD 95% and 3SD 98%)
A p-value (probability value) of less than 0.05 was used to determine significance to indicate that there is less than a 5% probability the null is correct, thereby accepting our alternative hypothesis.
What was the objective of the study?
To assess the education, training, and comfort of ob/gyn residents in managing shoulder dystocia and determine their clinical experience and competence
What was the hypothesis of the study? Null hypothesis?
The hypothesis was that residents were not comfortable managing shoulder dystocia throughout the different levels of training
The null hypothesis is that residents are comfortable managing shoulder dystocia throughout the different levels of training.
What was the study design? What was your continuous variable?
My research was a survey research. We used quantitative strategies by creating a questionnaire with targeted questions reflecting demographic and personal characteristics as well as numerically rated questions. We decided to use a survey method in order to collect the responses of a large population of residents in training across the nation in a relatively short period of time. I used a Lickert scale with a scale ranging from strongly agree, neutral to strongly disagree to measure their opinion and satisfaction quantitatively. These statistics were used because it allowed me to recruit participants, collect data, and utilize various methods of instrumentation that would best answer my research question – describing and exploring human behavior.
-Continuous variable are variables that can take on any value within a range – age (ie weight, height, bsa, speed and time).
Q’s: PGY level, gender, state of training, single vs multiple site of training, total number of deliveries in the institution, total number of shoulder dystocia cases experienced and whether the program was tracking this number, total number of shoulder dystocia cases managed from start to finish, whether SD and its management were taught within the program and its frequency, whether SD simulation drills were performed and its frequency, whether they went to offsite locations for additional training, and measurement of individual perspective on their performance of the maneuvers (McRoberts Woods/Rubin, posterior arm, Zavanelli), satisfaction of their education, comfort of their management and opinion about the need for additional training and its monitoring through ACGME.
Did you prove your hypothesis?
Yes. Among the 116 respondents, we were able to find that the majority, 97% received formal didactic training and 87% had simulations and drills to exercise shoulder dystocia and 94% of the respondents felt comfortable recognizing the signs of shoulder dystocia with 87% feeling they would be able to independently manage it upon encounter. However, more than half, 53% reported they have not independently managed a shoulder dystocia from start to finish, which included 34% of upper level residents and only half, 48% and 51% of the respondents were able to perform a higher level of maneuver (the Wood’s, Rubin and delivery of the posterior arm). Lastly, half, 49% agreed that it would be helpful for ACGME to track the number of cases during their residency training with some, 24% agreeing that a minimum case number should exist as part of their program requirements for graduation.
What was the population studied?
A total of 116 resident physicians participated in the survey across 23 states. There was a relatively equal distribution of responses by level of training including PGY-1 (25%, n=29), PGY-2 (29%, n= 33), PGY-3 (22%, n= 25) and PGY-4 (24%, n= 28). The mean age of respondents was 29.6 years (SD 2.68) with a range between 24 to 43 years. Approximately 87% of respondents were female (n=99). Response rates were highest in the state of New York (13%, n=21) and Pennsylvania (18%, n= 15). Program sizes varied between 3 and 11 incoming resident physicians per year.
What are other possible sources of sample selection bias?
Sampling bias – getting full representation
Non-response bias – timing/length of surveys, distribution methods, anonymity
Survivorship bias – residents from my program were more likely to respond
Acquiescence bias – societal/cultural norms avoiding extremes of answers such that in the lickert scale
Question order bias – started out with general questions and progressed to specific questions
Answer option/primacy bias – they may choose from the first few and not completely read the answers
Social desirability bias – to make them look good
How comparable was the control group?
There was no control group in the survey as the questions were asked of the representative sample
What was the statistical power of the study?
I did not do since this was a survey.
The statistical power, or sensitivity of the study is the likelihood of a significance test detecting an effect when there actually is one. It is used to predict the sample size needed to see a difference between the groups. A high power in a study indicated a large change of a test detecting true effect. Low power means that your test only has a small chance of detecting a true effects or that the results are likely to be distorted by random and systematic error. It is influenced by sample size, effect size and significance level. A power analysis can be used to determine the necessary sample size for a study which is made up of four main components. If you know or have estimates for any three of these, you can calculate the fourth component.
-statistical power – the likelihood that a test will detect an effect of a certain size if there is one, usually set at 80% or higher
= 1-Beta – Type II error, where beta is set at 0.2, so type II error is 20% and power is 80%
If difference exists, the study has an 80% chance of identifying it
- sample size – the minimum number of observations needed to observe an effect of a certain size with a given power level, typically what you are solving for, so must put other 3
- significance level (alpha or Type I error) - the maximum risk of rejecting a true null hypothesis that you are willing to take, usually set at 5%The probability of finding an effect that is not there – false positive
-expected effect size – a standardized way of expressing the magnitude of the expected result of your study, usually based on similar studies or a pilot study
Smaller effect size, more subjects needed
R = correlation coefficienct which ranges from –1 to +1
Low effect size (R) - 0.1, medium 0.3, high 0.5
The effect of a particular variable or set of variables could have a high or low relationship with predicting the outcome
Was the design of the study appropriate for the hypothesis to be tested?
Yes. Qualitative research is typically more exploratory, relying on the collection of verbal, behavioral or observational data that can be interpreted in a subjective manner. It has a broad reach and is commonly used to investigate the reasons of prospective problems. We were able to take qualitative data and analyze them using a quantitative methods.
The advantage of a questionnaire was
- uniformity, all respondents were asked the exact same qquestions
- cost-effectiveness
- collection of data in a short period of time
- minimal bias from the researcher during the data collection process
- gives respondents time to think before answering questions, as opposed to interviews
- ability to reach respondents in distant areas
The disadvantages of a questionnaire was
- respondents may answer randomly without properly reading the questions
- inability of respondents to express their additional thoughts about the matter
- high rate of non-response
- chance of collecting inaccurate information because respondents may not be able to understand the questions correctly
Were the conclusions justified by the findings?
They were justified to a certain degree as most of the residents were enrolled in an institution with greater than 2500 deliveries annually and 50% had a total of 1-5 shoulder dystocia experiences (30% being upper level residents) while 30% had 10-20 cases reflecting that the experience may be limited. Also, up to a half of the respondents agreed with the need for additional measures to improve their training. The suggestion for ACGME case logging and use of alternative educational materials to assess individual performance metrics may ultimately improve their training thereby improving maternal and neonatal outcome.
If the study could be repeated, however these pyramid design be revised to provide better reliability and validity of the conclusions?
I would have attempted other methods of questionnaries including mail with a pre-paid envelope, telephone questionnaire and in-house surveys – especially to decrease my non-responder bias. I would also have like to ask open-ended, dichotomous (yes or no) or scaline (1 to 10) questions to see the difference in ideas on how to implement a more efficient and practical method of training.
Have 1-2 references, know those studies, what inspired your study? Not a primary focus in prep
Deering Shad et al. Green Journal 2004 - Improving resident competency in the management of shoulder dystocia with simulation training.
To determine whether a simulation training scenario improves resident competency in the management of shoulder dystocia. Residents from 2 training program were block-randomized by year-group to a training session on shoulder dystocia management that used an obstetric birthing simulator or to a control group with no specific training. Trained residents and control subjects were subsequently tested on a standardized shoulder dystocia scenario, and the encounters were digitally recorded. A physician grader from an external institution then graded and rated the resident’s performance with a standardized evaluation sheet. Trained residents had significantly higher scores in all evaluation categories, including timelines of their interventions, performance of maneuvers, and overall performance. They also performed the delivery in a shorter time than control subjects
Hill MG, Cohen WR. Womens Health Journal in London 2016 Shoulder dystocia: prediction and management.
In this review, they discussed the risk factors for shoulder dystocia and proposed a framework for the prediction and prevention of the complication. A recommended approach to management when shoulder dystocia occurs is outlined, with review of the maneuvers used to relieve the obstruction with minimal risk of fetal and maternal injury.
History
History of shoulder dystocia or baby with BPI
Maternal diabetes
Maternal obesity
Antepartum factors
Macrosomia (risk increases as fetal weight increases)
Gestational diabetes
Excessive weight gain
Intrapartum factors
Clinical pelvimetry and estimated fetal weight concerning for CPD
Protracted active phase dilatation
Arrest of dilatation
Prolonged deceleration phase
Failed, protracted or arrested descent
Long second stage
Precipitate second stage
Instrumental delivery
