Link to the University of Pittsburgh Homepage
Link to the University Library System Homepage Link to the Contact Us Form

Prediction of Some Health Outcomes among Advanced Cancer Patient Caregivers Using Gut Microbiome

Song, Ruopu (2020) Prediction of Some Health Outcomes among Advanced Cancer Patient Caregivers Using Gut Microbiome. Master's Thesis, University of Pittsburgh. (Unpublished)

This is the latest version of this item.

Submitted Version

Download (377kB) | Preview


Increasingly researchers are discovering that many health outcomes are associated with gut microbiome. This was the motivation for the research conducted in this thesis with focus on psychosocial and metabolic health of caregivers of cancer patients. The microbiome data used in this study were obtained from cancer caregivers who participated in a study focusing on the relationship between psychosocial and behavioral predictors, and metabolic syndrome. Our goal was to determine how well microbiome alone can predict the psychosocial and metabolic health of a caregiver, as defined by depression, stress, hostility and patient-caregiver relationship, and metabolic syndrome.
We explored two different prediction (or classification) procedures, namely Fisher’s Linear Discriminant Analysis (LDA) and logistic regression. The predictors consisted of the five important bacterial phyla that constitute 95% of the adult gut flora, namely, Actinobacteria, Bacteroidetes, Firmicutes, Proteobacteria, Verrucomicrobia, with the remaining phyla combined as “Other”. Aitchison’s log-ratios were used to transform the data from the simplex to the Euclidean space before applying LDA and logistic regression.
For the five health characteristics, we found that the logistic regression had generally higher total correct prediction/classification rates for classifying subjects into their correct health categories than LDA. We found the overall correct classification rates for depression and patient-caregiver relationship to be 80% and 67%, respectively. Thus, there is an 80% chance of correctly predicting the depression symptom category (low or high) of a caregiver using the microbial phyla. The correct classification rates for caregiver stress, hostility and metabolic syndrome were about 63%, 60% and 53%, respectively.
This appears to be the first study that attempted to predict psychosocial characteristics and metabolic health of a caregiver using the stool microbial phyla in cancer patient caregiver population. Although the overall success is modest, the results are encouraging to conduct a larger follow-up study which can have major clinical implications. If successful, like blood and urine tests, the stool microbiome can potentially be used to diagnose the above noted psychosocial characteristics and metabolic syndrome for caregivers. The public health relevance of this study is that we provided a direction on developing new microbiome-based intervention for helping cancer patient caregivers relieve mental and physical health symptoms.


Social Networking:
Share |


Item Type: University of Pittsburgh ETD
Status: Unpublished
CreatorsEmailPitt UsernameORCID
ETD Committee:
TitleMemberEmail AddressPitt UsernameORCID
Thesis AdvisorPeddada,
Committee MemberPeddada,
Committee MemberSteel,
Committee MemberBuchanich,
Date: 30 July 2020
Date Type: Publication
Defense Date: 17 April 2020
Approval Date: 30 July 2020
Submission Date: 31 March 2020
Access Restriction: No restriction; Release the ETD for access worldwide immediately.
Number of Pages: 40
Institution: University of Pittsburgh
Schools and Programs: School of Public Health > Biostatistics
Degree: MS - Master of Science
Thesis Type: Master's Thesis
Refereed: Yes
Uncontrolled Keywords: Classification, Microbiome, LDA, Caregiver, Machine learning
Date Deposited: 30 Jul 2020 20:40
Last Modified: 30 Jul 2020 20:40

Available Versions of this Item

  • Prediction of Some Health Outcomes among Advanced Cancer Patient Caregivers Using Gut Microbiome. (deposited 30 Jul 2020 20:40) [Currently Displayed]


Monthly Views for the past 3 years

Plum Analytics

Actions (login required)

View Item View Item