I'm not clear whether you're going to use an intervention versus control because later you state that you'd allow treatment of hypotension as per clinical judgement and it sounds like this would apply to all patients.

I wouldn't try to work out what volume of different fluids is equivalent. I think it would differ between sites anyway once you take into account other site-related factors that influence outcome such as clinical expertise, culture, treatment preferences, patient type.

Define your outcome(s) for clinically relevant hypotension. Include a linear outcome like systolic BP or MAP so you can do linear regression which provides more power if needed. I would decide on a critical time period in which you consider that fluid volume administered might affect outcome in the patients you're studying. Collect the volume of all fluids administered and lost and fluid balance for that time period. Then using modelling, you can include key fluids (volume) as potential predictors of outcome and see whether they remain independently associated with outcome in multivariable analysis. With this and further analysis you can get a feel for the volumes that might be associated with poorer outcomes.

Depending on your results, from here you might be in a better position to undertake some kind of pilot or feasibility study.

Hope this is useful

Liz

Dear Ms Bains,

I think Ms Moore's answer is directional and will lead you to how to better design and conduct your study. Yes I agree it is very important to decide if this is case/control, cohort or clinical trial. Since you are not planning the study per se, you are not devising the intervention to be the same on all patients, it is not a clinical trial. But rather it is observational.  Here are two a good links to compare Case/Control vs Cohort studies that may help you better classify your study.

http://www.iwh.on.ca/wrmb/cohort-studies-case-control-studies-and-rcts 

www.ciphi.ca/hamilton/Content/content/resources/explore/fb_case_v_cohort.html

Since you will be looking at data after the fact of the natural progression of treatment it is clearly observational and not a true clinical randomized study. However if you are comparing to a control group this sounds more like a case/control study. But you will need to look at all characteristics of your study to truly decide.  

To address your original question about how to find the equivalent-dose of the drugs administered, you might try the following site that utilizes the orange book comparison of drugs within classes.

http://www.medscape.org/viewarticle/416390_3

In addition, you will need to classify the diagnosis of they hypertension: primary essential or secondary [non-essential] along with the specific cause, in order for the study to consider/avoid confounding elements. other confounding factors may include the co-existing disease states of the patients, diet, polypharmacy. Ask yourself: can you truly compare the results of a given drug to another given drug in the same classification with known equivalent - dose information in patients with different comorbidities ? That is a question either you must answer before you begin or design your study OR your study will attempt or be designed to answer.

I do hope this helps you,

Respectfully,

Jeanetta Mastron

In clinical medicine, diagnosis precedes treatment.

Treatments are comparable when the disease state is the same (or similar).

Fluid therapy would be comparable if the cause of hypotension was hypovolemia and the patient had a measurable low cardiac ouput and was volume responsive (preserved myocardial contractility).  Even in this setting, redistribution of the fluid administered and available in the central circulation would depend on capillary permeability among other factors.

Vasopressor therapy would be comparable if the cause of hypotension was a low SVR state in patients with adequate or high cardiac output.

If you only have "hypotension" as the clinical sign determined by BP measurements, without identifying the underlying cause of the hypotension, it would be difficult to make meaningful comparisons.

Dear Vini Bains,

Please note the vein of similarity:

Posted by Jeanetta Mastron: "In addition, you will need to classify the diagnosis of the hypertension: primary essential or secondary [non-essential] along with the specific cause, in order for the study to consider/avoid confounding elements."

Posted by Srikantha L Rao "IF you only have "hypotension" as the clinical sign determined by BP measurements, without identifying the underlying cause of the hypotension, it would be difficult to make meaningful comparisons."

Both are saying that you MUST know the cause.   

To answer your question: "Any thoughts on how to compare these possible treatments for hypotension? " ....The trick is...."

The only TRICK I can see is stratification, specifically POST - stratification, UNLESS you know in advance how many patients have a specific cause of HTN and subsequently hypotension:

You could stratify the your study by cause, and therefore by pharmaceutical treatment,  but you still must know the cause. However I would say it is best to focus on one cause and one treatment in your study or one cause and differences in outcome due to different treatment by different doctors. But you still need to look at all confounding elements of the subjects or condition of the subjects.

It is common in a text book to look at a specific disease state and how to treat a person with that disease, but the reality is there are many other factors that will contribute to the physician's choice in/of therapy. So prescribed therapy may depend upon level or how far the disease has progressed, co-morbidities, poly- pharmacy, age, overall health and condition of the patient, patient preferences and many, many other factors [in addition to cause].

If you are NEW to stratification, you might consider the following articles:

"Control for Confounding in Case-Control Studies Using the Stratification Score, a Retrospective Balancing Score (2010) by Andrew S. Allen and Glen A. Satten 

"Abstract: The stratification score for a case-control study is the probability of disease modeled as a function of potential confounders. The authors show that the stratification score is a retrospective balancing score and thus plays a similar role in case-control studies as the propensity score plays in prospective studies. The authors further show how standardization using the stratification score can be used to compare the distributions of exposures that would be found among case and control participants if both groups had the same distribution of confounding covariables. The authors illustrate these results using data from a genome-wide association study, the GAIN (Genetic Association Information Network) study of schizophrenia among African Americans (2006–2008)."

The  full article is online at aje.oxfordjournals.org/content/173/7/752.full

The following pdf Case-Control Studies may help you decide when and HOW to stratify.

6.3 Matching and Stratification

6.3.1 Consequences of Matching

6.3.2 Efficiency of Matching

http://depts.washington.edu/epidem/Epi583/Articles/ChapI6-1%28Breslow%29.pdf

IF you select other articles on stratification, be sure that it is NOT for clinical studies, but rather for Case/Controlled Retrospective Observed studies. There is a difference in the application. Some do not agree that stratification should be used for clinical studies. Here is an example of such an article: The benefit of stratification in clinical trials revisited

onlinelibrary.wiley.com/doi/10.1002/sim.4351/abstract

In another RG question I answered a question about the ratio of case to control ratio for the best statistical analysis and study power. You may find the following helpful in the same above document:

6.4.3 How Many Controls per Case?

How Many Control Groups?

----------------------

IF you select other articles to read on stratification, be sure that it is NOT for clinical studies, but rather for Case/Controlled Retrospective Observed studies, unless you will since you are not doing an intervention to all individuals. There is a difference in the application. Some do not agree that stratification should be used for clinical studies. Here is an example of such an article: The benefit of stratification in clinical trials revisited [authors believe that stratification should not be used in clinical trials]  wiley.com/doi/10.1002/sim.4351/abstract

I do hope that this points you in to the right direction...possibly to seek counsel of a biostatistician who understands both medicine, such as Srikantha L Rao does AND statistics in case-control students with/without stratification.

Respectfully,

Jeanetta Mastron

aje.oxfordjournals.org/content/173/7/752.full

http://depts.washington.edu/epidem/Epi583/Articles/ChapI6-1%28Breslow%29.pdf

onlinelibrary.wiley.com/doi/10.1002/sim.4351/abstract

wiley.com/doi/10.1002/sim.4351/abstract

Dear Vini Bains,

You may wish to check out the following thread of Q&A that may help you, where I have posted on ratios in case/control studies

https://www.researchgate.net/post/What_should_be_the_ratio_of_cases_to_controls_in_case_control_clinica_trials#view=556e06ff60614b4ac38b4654

MORE specifically to your question Dr./Professor Ariel Linden also posted  and refers to his RG posted paper that I believe will help you:

"Using balance statistics to determine the optimal number of controls in matching studies"

https://www.researchgate.net/publication/255175789_Using_balance_statistics_to_determine_the_optimal_number_of_controls_in_matching_studies

Respectfully,

Jeanetta Mastron

Article Using Balance Statistics to Determine the Optimal Number of ...