Answers to Questions for E-Lesson 1

(Starting Off Right: The Problem Statement)

Suppose you’re a reviewer. . .

Based on the criteria we’ve given you, how would you rate the following problem statements? Which problem statements have too little information, which have too much, which ones speculate about the cause (are “contaminated”), and which ones are ok? (Please note that some may fall into more than one category.) Underline your ratings and tell us why.

1. The calculated batch yield was 62.4% which is below the alert and action limits of 85% and 75%, respectively, according to SOP MFG-23-010 “Reconciliation Limits.”

Too Little


This problem description contains too little information to ground the reader and prepare for the investigation narrative that follows. First, although the material name and batch number may appear in the heading of the investigation form, we recommend that the information be repeated in the problem description. That way the reader doesn’t have to search for the information. This description clearly states the results and the alert and action limits or expected yields, but when did this problem occur? How and when was it discovered? A little more information would prepare the reader for the investigation that follows.

2. The 12-hydroxyducadone impurity result for finished product Lot 987123 was 12 ppm versus a specification of ≤10 ppm. The 12-hydroxyducadone impurity level in the Active Pharmaceutical Ingredient (API) batch used to manufacture the product was also tested and was higher than normal, but within the validation parameters. The batch will be rejected.

Too Little, Too Much, Contaminated 


First, this problem description has too little information because all key elements for a complete problem description are not addressed. To make it more complete, the writer should add how and when the problem was discovered, and when the problem occurred, if different. Although the lot number is identified, what is the finished product?
Next, this problem description has too much information by mentioning testing of the API batch. This information may be pertinent to the investigation, but does not belong in the problem description. The problem description should stick to the single, observable problem.

Finally, how does the investigator know that the batch will be rejected until the investigation is complete? A good problem statement focuses only on the information known when the event occurred. Rejecting the batch is speculation at this point. In fact, in this example, the QA Disposition at the end of the investigation report indicated that the batch would be reprocessed, not rejected. In other words, the investigation resulted in a different outcome than that initially stated in the problem description. So, although there is no conjecture about root cause, the problem description is contaminated with conjecture about batch disposition. If the writer’s intent was to provide the immediate action taken, a better statement would be that the batch was quarantined and an investigation was initiated.

3. USP Purified Water use points #2 and #4 located on the first floor and second floor, respectively, of Building E were sampled on February 4, 2017. Later that day, QC obtained a Total Organic Carbon result of 608 ppb for use point #4 which was greater than the specification of less than 500 ppb. The result for use point #2 met the specification. Facilities locked both use points out of service.



Good work! The writer nailed this problem statement. This clear, concise statement addresses all of the necessary key elements for a complete problem description, including the locations sampled, date the problem was discovered and the date it happened, the results and specification, and how the results were discovered. It also describes the action taken to contain the problem.

4. Each lot meets current specifications. However, the R&D group pulled a sample from the downstream process for development work. The material did not perform as expected. Subsequent analysis of the three commercial lots 213798, 412798 and 512798 indicated that the pH of each lot was ≥ 10, 10 and 7, respectively.

Too Little


Help! Each lot meets current specifications? Which lots? What product? The writer should ground the reader with this information right up front in the first sentence. Other key elements are missing too, such as the dates of occurrence and discovery. Again, don’t make the reader search for the information even though it may be provided in the report header. Write it directly into the problem description so the reader immediately understands the deviation.

Also, without knowing the pH specification the reader really doesn’t know which lots are affected. Is the pH value of 10 acceptable or high, or is the pH value of 7 acceptable or low? Bottom line: remember to interpret the information for the educated, but less informed reader and clearly state the problem.

5. Work Order WO 2017-10-XYZ was generated to install set screws on glass washer nozzles that are position specific. After consulting with the vendor, the engineer determined that the set screws were not needed because the vendor had developed a new nozzle design. The new nozzles were purchased and installed on the referenced washers. The work order was not rerouted to approve the modified scope of work as required by the current procedure MFG-13-423. The Responsible Engineer left the company before the work was completed.

Too Little, Too Much, Contaminated

How many problems can you find with this problem description? First, the statement doesn’t address all the required elements for a complete problem description. When did this problem occur and when was it discovered? Although the nozzles are for a glass washer, which glass washer? How was the problem discovered? Although some of this information may appear in the header, remember to write it directly in the problem description so the reader doesn’t have to refer back and forth to the header to locate the information.

Next, the writer provides a lot of information before getting to the actual problem in the fourth sentence. By not focusing on the single, observable event up front, the problem becomes lost. Also, was all the information known when the problem was discovered or was some of it discovered later during the investigation?

Finally, the last sentence implies that the problem occurred because the Responsible Engineer wasn’t present to oversee the project completion. This contaminates the description with conjecture.



Now use the information given below to write a problem statement that meets the criteria we’ve discussed and that doesn’t say too little or too much, and is not contaminated by conjecture about why the problem occurred.

Product: Ducarase Solution for Injection, 5 mg/mL                     Lot No.: 672344
Date of Occurrence: April 1, 2017

Bioburden in unfiltered product strains the sterilizing filter and contributes to degradations in drug product.  The pre-filtration bioburden limit per SOP-GB-77-556 rev. 2 step 7.4 is <10 CFU/100mL.  However, on April 1, 2017, when analyzing pre-filtration bioburden result for Lot 672344, microbiology analyst noticed data exceeded the specification of <10 CFU/100mL as per SOP-GB-77-556 rev. 2.  The results were 12 CFU/100mL, probably introduced during sampling.  Lots 672345, 672346 and 672347 were also analyzed and found to be within historical trends.

Rewritten Problem Statement:

Product: Ducarase Solution for Injection, 5 mg/mL                     Lot No.: 672344
Date of Occurrence: April 1, 2017

On April 1, 2017 when analyzing pre-filtration bioburden data, the microbiology analyst noticed the result for Ducarase Solution for Injection, 5 mg/mL Lot 672344 was 12 CFU/100mL, exceeding the specification of <10 CFU/100mL.

The first sentence in the original problem description above (that says that bioburden strains the filters) provides basic information that readers would either know or could find out in a background section of the investigation narrative.  Let’s cut it and go straight to what actually happened. 

It’s important to begin with the single, observable event rather than with the specification.  Readers won’t see how relevant the specification is until they know that the results were out of specification, so we start by stating what the microbiology analyst noted at the time of the event, and then provide the specification. 

To make the problem statement clearer and more concise, we also decided not to reference the SOP that contains the specification.  We would mention this later in the investigation.

We left out the phrase, “probably introduced during sampling” because, before you investigate, how do you know that it was introduced during sampling?  If it obviously happened during sampling—because, for instance, the analyst saw something fall into the sample—you would include this in describing the observable event.  However, making an assumption about the cause contaminates the problem description.

Lastly, we removed the details about historical trends.  The writer could investigate and discuss trends later on in the investigation.