Hypothesis Testing in a Medical Scenario

• Most of the times in the medical field, decisions are made based on the tests, observations and the patient symptoms.
• However, this may not be perfectly accurate sometimes.
• To manage the risks involved, health workers can use hypothesis testing to evaluate possibilities and make decisions that minimize potential harm and risks involved.

PROBLEM
A lady got a lot of rashes and a very bad stomach upset after eating meat from a restaurant. She went to hospital and the doctor told her that she was allergic to meat. However she has been eating meat her whole life. The following day, she went to another different hospital and she was told she had food poisoning.

NULL HYPOTHESIS
The reaction to the food was caused by an allergy.

ALTERNATIVE HYPOTHESIS
The reaction to the food was not caused by an allergy.

TYPE 1 ERROR
The conclusion was that the reaction to the food was not caused by an allergy but the lady has an allergy.

TYPE 2 ERROR
The conclusion was that the reaction to the food was caused by an allergy but the lady doesn’t have an allergy.

OPINION
The lady has been eating meat all her life but never gotten an allergy reaction. She has grown up without any meat allergies. Perhaps she developed the allergy in her prime years.

The type one error concludes that she doesn’t have allergy whereas she does have it. This might be fatal since she might continue eating meat at the expense of her health. This may lead to her health worsening.

The type two error on the other hand concludes that the reaction to the food was caused by an allergy but she doesn’t have one. This is a lesser consequence because the lady may refrain from eating meat as she has been told she has an allergy. It won’t have an effect on her health. Besides, she may look for other sources of protein apart from meat.

CONCLUSION
The risks of the type two error are better than those of the type one error.

• Hypothesis testing helps us to weigh the outcomes of our decisions by considering the type one and type two errors, and making the most appropriate decision.