I want to collect data from 1000 admitted cardiac patients within 03 months. I will collect data from 4 male wards and 4 female wards. Daily patient turnover in these wards is as follows: (Male ward 1=113, Male ward 2=56, Male ward 3=60, Male ward 4=83) and (Female ward 1=88, Female ward 2=69, Female ward 3=87, Female ward 4=71). Which sampling technique will be best to collect data from these male and female wards?
Since a comprehensive list of the admitted cardiac patients is available - put at 1000, the probability sampling technique becomes the appropriate method.
Within the probability sampling method are four basic sampling techniques, namely, simple random, systematic, stratified and clustered sampling techniques.
Out of these four, the stratified sampling technique is the appropriate one. This is os because this technique is meant to ensure all the attributes of a study population are inclusive in the sample size.
Thus, your sampling will cut across the varying attributes of your study population such as Male ward 1, Male ward, Male ward 3, Male ward 4 and Female ward 1, Female ward 2, Female ward 3, Female ward 4
For you to determine the appropriate size from the list of 1000 patients, you need to use the established mathematical sampling model (https://ucanr.edu/sites/ceprogramevaluation/files/143304.pdf; https://hal.science/hal-02557333/document) or digital sample size generating tool (https://www.calculator.net/sample-size-calculator.html; https://uk.surveymonkey.com/mp/sample-size-calculator/).
Once the appropriate number is obtained, determine its proportion (%) from the 1000 list. Whatever percentage you arrive at signifies the proportion of the patients that will be sampled across the eight wards.
@Okanlade Here 1000 is the sample size. I need data of 1000 patients with heart disease from 08 wards.
Do you need to represent each ward accurately? If so, then use stratified sampling. If not, then simply use each patient available until you reach 1000. Report this as obtaining data from 1,000 "consecutive patients."
Do you want to compare makes and females? If so, they use two subsamples of 500 each.
David L Morgan If we apply systematic sampling, will it be acceptable?
To collect data from 1000 cardiac patients within three months, considering the patient turnover in each ward, the Stratified Random Sampling technique would be the best option. Here's why:
Explanation:
Steps to Implement Stratified Sampling:
Advantages of Stratified Sampling:
Mahendra Kumar Verma Kindly clear the sample size calculation. If I collect data from patients with specific types of heart disease (example: ischaemic heart disease), would we apply the sampling you suggested?
Try to collect from different age groups, genders, places of living, types of work that they undertake, religious groups, and types of physical exercises that they undertake.
Most medical journals would be satisfied so long as there is no difference between the wards such as severity
David L Morgan I will collect data from patients with ischemic heart disease.
I have a different take on this. I think it depends on your purpose and what you want to achieve.
if the wards are simply a convenience for obtaining data, then lots of patients in a few wards is fine, and you could control for ward differences by putting in a set of dummies, what I would call a fixed effects analysis.
However, if wards are a real and substantive entity ( that is a target of inference) and you are interested in between ward differences in say clinical practice and outcome (as well as of course individual outcome), I would want lots of wards with different types of practices , say 50 wards with 20 patients in each. I would then do a multilevel random effects analysis to see what individual and ward characteristics are important. I will add some references later when I get to my desktop.
if you look at look at educational research over the last 30 years, pupils are now seen as nested in classes and schools so that populations are (hierarchically)structured not just in the sample, but in the population, and this can be exploited and should not be ignored.
1. Stratification: You have 8 distinct wards (strata) with varying patient turnover rates. Stratified sampling allows you to sample from each ward separately, ensuring representation from each stratum.
2. Randomization: Within each ward, you can randomly select patients to participate in the study. This helps minimize bias and ensures that the sample is representative of the population.
3. Proportionate sampling: You can allocate the sample size to each ward based on their patient turnover rates. For example, Male Ward 1 has the highest turnover rate (113), so you may allocate a larger sample size to this ward.
To determine the sample size for each ward, you can use the following formula:
Sample size for each ward = (Total sample size x Ward's patient turnover rate) / Total patient turnover rate across all wards
Using this formula, you can calculate the sample size for each ward. For example:
Male Ward 1: (1000 x 113) / 637 (total male patient turnover) ≈ 177
Male Ward 2: (1000 x 56) / 637 ≈ 88
...
Female Ward 1: (1000 x 88) / 315 (total female patient turnover) ≈ 280
Female Ward 2: (1000 x 69) / 315 ≈ 220
...
By using stratified random sampling, you'll ensure that your sample is representative of the population, and you'll be able to collect data from 1000 admitted cardiac patients within the 3-month timeframe.
Reasoning:
You have a clearly defined population divided into sub-groups (male and female wards).
You want to ensure representation from each ward, especially those with varying patient volumes.
Stratified sampling allows you to allocate a sample size proportional to the size of each ward, ensuring accurate representation.
How to implement:
Determine the proportion: Calculate the proportion of patients in each ward relative to the total population.
Allocate sample size: Allocate the sample size of 1000 patients to each ward based on the calculated proportion.
Select samples: Randomly select the required number of patients from each ward.
Un Muestreo sistemático, tienes una población fragmentada en sub-grupos y puedes elegir de forma aleatoria dándole un numero a cada participante de cada sub grupos y tienes la posibilidad que todos sean seleccionados, por ejemplo todas las personas que obtuvieron el número 5.
Nuri Masoud Abduali which type of randomization should I apply (simple or systematic) to meet the sample size of each ward if I want to include a specific type of patient with heart disease (ischemic heart disease) excluding the patients with heart failure or cardiomyopathy?
Samuel EKUNDAYO Owoleye which type of randomization should I apply (simple or systematic) to meet the sample size of each ward if I want to include a specific type of patient with heart disease (ischemic heart disease) excluding the patients with heart failure or cardiomyopathy?
Omilcia María Medina Corzo If patient turnover is high (1-3 days on an average), how could you apply systematic random sampling?
The best sampling technique in this case would be stratified random sampling. You can treat each ward as a stratum (male and female wards), and within each ward, randomly select patients based on their proportion in the total turnover. This ensures that the sample represents both the gender and ward-specific variations in patient numbers.
Whether or not you treat each ward as a separate stratum depends on whether you want to maintain their exact representation in the sample. If the wards are essentially inter-changeable, then that level of detail is not necessary.
Lingkan Barua To achieve your sample size requirements while ensuring that only patients with ischemic heart disease (IHD) are included—excluding those with heart failure or cardiomyopathy—you should consider systematic random sampling rather than simple random sampling.
I don’t think there is any issue with limiting the sample to ischemic cases, just as with the earlier limitation of the sample to patients with heart attacks.
I suggest purposive non-probable sampling using nan random inclusion criteria. I agree with David L Morgan that there is no issue with limiting the sample as long as sample meets the inclusion criteria.
Additionally the choice of sampling should consider the research questions that will produce the needed data and purpose of the study that will direct what to find from the data.
Hope this helps. Good luck.
I propose the use of purposive non-probability sampling with non-random inclusion criteria to ensure the selection of participants who meet specific research needs. As David L. Morgan suggests, limiting the sample size is not an issue as long as the selected participants align with the defined inclusion criteria.
Furthermore, the choice of sampling strategy should be guided by the research questions, ensuring the collection of data that directly addresses the study objectives. The purpose of the study should serve as a key determinant in shaping the sampling approach, directing what needs to be extracted from the collected data.
your target population (1,000 admitted cardiac patients in 3 months) and the daily patient turnover in 8 wards, the most suitable sampling technique should ensure:
Recommended Sampling Technique: Stratified Purposive Sampling with Proportional Quotas
Justification:
- Stratified Sampling: Since there are separate male and female wards, you can divide the population into two strata (male and female) and then further into four sub-strata (each ward within male and female groups).
- Purposive Selection: Only admitted cardiac patients will be considered.
- Proportional Quota Sampling: The sample from each ward should be proportional to its daily patient turnover to avoid over- or under-representation.
Steps for Sampling:
Alternative Approach: Systematic Sampling
- If patient admissions follow a regular pattern, you can select every Nth patient based on the required quota per ward.
- This will ensure even distribution without missing any major subgroup.
Final Recommendation:
- Use Stratified Purposive Sampling with Proportional Quotas to ensure fair representation.
- Use Systematic Sampling within each ward to select participants efficiently.
Mahendra Kumar Verma Your posting looks like an uncredited borrowing from an AI. If so, please credit your source just as you would whenever you are citing someone else's work.
- Badges
- Science method