Authors: Hope Linge, Elise Tarbi and Maija Reblin
Posted: August 2022
In a society suffering from a loneliness epidemic, human connections can be rare and special things (Murthy, 2020). Loneliness, though an emotional feeling, can have detrimental impacts on mental and physical health (Badal, 2021). Among older and seriously ill individuals, these effects are magnified, and loneliness has been identified as a significant risk factor for serious illnesses such as heart disease, lung disease, obesity, depression, and anxiety, as well as increased mortality rates (Ong, 2016; Yangus, 2018). Being able to identify expressions of loneliness in serious illness conversations could allow clinicians to better recognize loneliness in patients and intervene to prevent its harmful effects. We sought to create a corpus of loneliness and connection words that can provide a way of identifying and quantifying loneliness using natural language processing (NLP), allowing us to calculate the frequency and proportion of loneliness in conversation.
We started by reading pertinent literature on loneliness in health care settings (meta-analyses and qualitative meta-syntheses on loneliness, as well as existing measurement tools), and by pulling keywords from these articles, we created a preliminary list of 329 words and phrases relevant to loneliness and human connection (Bolmsjö, 2019; Donovan, 2020; Edberg, 2019; Ettema, 2019; Guntuku, 2019; Halam, 2022; Hipson, 2021; Holt-Lunstad, 2015; Holt-Lunstad, 2021; Karhe, 2015; Kitzmüller, 2018; Kotwal, 2021; Mansfield, 2021; Ong, 2016; Reinhardt, 2021; Rokach, 1988; Sjöberg, 2018; Vedder, 2022; Yanguas, 2018). We included a list of relationship terms from Badal and colleagues (2021). All of these words were focused around our central definition of loneliness, that it “is a subjective, unwelcome feeling of lack or loss of companionship that occurs when there is a mismatch between the quantity and quality of social relationships that a person has, and those that person wants” (Reinhardt, 2021). Then, using Word Net, a lexical dictionary of the English language, and thesaurus.com, an online reference work for finding synonyms, we generated the synonyms and antonyms of each word. After adding these related words, our list consisted of 1,055 words and phrases. We then organized this list by grouping words and phrases with the same stem if the stem had an important and relevant meaning (e.g., “connected,” “disconnected,” and “unconnected” were all grouped under the stem “connect”), leaving us with a list of 882 unique stems.
After refining the dictionary, three of the lab members most involved in generating the word list voted on whether each of the 882 words and phrases could be indicative of loneliness or human connection in a clinical conversation using a scale of 0 to 2 (zero being a word or phrase not used to express loneliness or connection, two being a word or phrase that would be used to express loneliness or connection, and one indicating uncertainty). This selection measure was modeled after the rating scale used by Gramling and colleagues in their 2021 uncertainty study. We kept 289 words and phrases whose mode was 2; we excluded 360 words and phrases whose mode was 0. Through small group discussion of the remaining terms, lab members excluded an additional 43 words. The remaining 188 words were brought to a larger group of lab members who voted either yes (this word is an expression of loneliness and/or connection) or no (this word is not an expression of loneliness and/or connection). After this second round of voting, we kept 53 words and phrases that had a majority of yes and excluded 81 words and phrases that had a majority of no. The remaining 49 words that did not reach a majority were brought to a lab meeting where a large group met to adjudicate their relevance. Throughout this meeting, 21 words and phrases were kept and 32 were excluded (4 words had different variations voted on separately due to group disagreement, causing the number of total words to increase from 49 to 53). In summary, out of the preliminary list of 882 words and phrases, 373 were kept in the final lexicon.
While the words and phrases in this lexicon are important indicators as to how loneliness and human connection manifest in clinical conversation, there are other conversational patterns indicative of loneliness that may be worth considering. For instance, how frequently people use plural versus singular pronouns (e.g., “we” versus “I”) (Badal, 2021), if they predominantly talk in the past, present, or future tense, and the tone and proportion of negativity present (Hipson, 2021) (e.g., using phrases like “cannot,” “do not,” “am not,” or “will not”). Additionally, who the speaker is (patient, clinician, or family member), how the transitions between speakers are handled, and the patient’s sentence syntax (complexity, mirroring patterns, and response length) (Badal, 2021) are other potential speech patterns warranting study.
In the future, this dictionary could be combined with gold standard human-annotated loneliness conversations to compare performance. Furthermore, it could be used to compare the frequency of the words in the loneliness lexicon and the trajectory of loneliness within the conversations with other conversational features or self-report loneliness measures. This would allow us to test and improve the library, on the path to building interventions to mitigate the impact of loneliness in the seriously ill population.
Strong subgroup: The strong subgroup consists of the unanimously selected words indicative of loneliness and/or connection in the first round of voting. These words are strong indicators of loneliness and/or connection in conversational settings:
|Associate (associated, disassociated)||Aunt||Belong (belonging)||Betrayed|
|Bleak||Brother (stepbrother, brother-in-law)||Brushed aside||Brushed off|
|Buddy||By my side||By myself||Care (careful, careless, uncared for)|
|Cast away||Cat||Cherished||Child (stepchild, grandchild)|
|Count on||Cousin||Crying out||Cuddling|
|Dad||Daughter (granddaughter, stepdaughter, goddaughter, daughter-in-law)||Dejected||Demoralized|
|Don’t have someone||Drifted apart||Embrace (embraced)||Estrange (estranged)|
|Friend (girlfriend, boyfriend, friendly, unfriendly, friends, friendship)||Get together||Get with||God|
|Has my back||Heard from||Hug||Husband|
|Ignored||In the know||In touch||Inseparable|
|Kept in the dark||Kid (stepkid, grandkid)||Kiss||Lean on|
|Left behind||Left high and dry||Left out||Listened to|
|Lone (lonely, alone, lonesome)||Lost||Love (loved, truelove, unloved)||MA|
|Missing out||Mom||Mother (mother-in-law)||Neglect (neglected)|
|Nobody||Nobody to keep me company||Not a part of something||Not included|
|Not listened to||On my own||Ostracized||OT|
|Out of place||Pal||Parent||Participation|
|Ridiculed||Roommate||Scorned||Share (shared, unshared)|
|Shun (shunned)||Shut out||Sibling (stepsibling)||Sidekick|
|Sister (stepsister, sister-in-law)||Solitary||Solitude||Son (stepson, godson, grandson, son-in-law)|
|Spouse||Support (supporter, unsupported)||Teacher||Team (teammate)|
|There for me||Together (togetherness)||Touch base with||Unaccompanied|
|Uncaring||Uncle||Understood (misunderstood)||Unite (united, reunite, disunite, disunited)|
|With me||With others||Withdrawn|
All other terms: These terms did not meet criteria for being strong indicators of loneliness or connection:
|Meet (meeting, meetup)||Powerless|
|Brought together||Come together||Communal||Cooperation (cooperative)|
|Faithful||Go away||Heartache||Hold hands|
|Household||Identify with||In sync||Just want someone|
|Soul (soulmate)||Split up||Trust||Unapproachable|
|Agony||Apart||Connect (disconnected, unconnected, connected)||Contact (contacts)|
|Couple||Defenseless||Disgraced||Engaged (engagement, disengage, disengaged)|
|Fit in||Going away||Going to leave||Invisible|
|Involve (involved, uninvolved)||Isolated||Just me||Recognized|
|Rejected||Singled out||Thoughtful (unthoughtful, thoughtless)||Turn to|
|Depend (dependable, dependent)||Deserted||Despised||Discouraged|
|Exposed||Fall apart||Familiar (unfamiliar)||Forget (forgotten)|
|Give up||Guarded (unguarded)||Hidden||Hopelessness|
|Known (not known, unknown)||Marooned||Not good enough||Not moving|
|Seclude||Separate (separated)||Set apart||Sheltered|
|Assistance||Attention (attentive)||Home (homelessness)||Social (socializing, antisocial)|
|Conflict (conflicting)||Despondency||Fellow (fellowship)||Guy|
|Hoedown||Hostile (hostility)||Kin||Left being|
|Blackball||Blacklisted||Blocked out||Cut off|
|Mingling||Detached||Part of||Sole (solely)|
|Barred||Help (helper, helpful, helpless, unhelpful)||Reconciliation||Encounter|
|See (unseen)||Popular (unpopular)||Worth (unworthy, worthless)||Divide|
|Drifting||Nothing (nothingness)||On their own||Rudderless|
|Purpose (unpurposed, purposeless)||Satisfied (dissatisfied, unsatisfied)||Vulnerable||Anxious|
|Relate||Passed by||Depressed||Break up|
Badal, V. D., et al. (2021). Do Words matter? Detecting social isolation and loneliness in older adults using natural language processing. Frontiers in psychiatry, 12.
Bolmsjö, I., Tengland, P. A., & Rämgård, M. (2019). Existential loneliness: An attempt at an analysis of the concept and the phenomenon. Nursing ethics, 26(5), 1310-1325.
Donovan, N. J., & Blazer, D. (2020). Social isolation and loneliness in older adults: review and commentary of a National Academies report. The American Journal of Geriatric Psychiatry, 28(12), 1233-1244.
Edberg, A. K., & Bolmsjö, I. (2019). Exploring existential loneliness among frail older people as a basis for an intervention: Protocol for the development phase of the LONE study. JMIR Research Protocols, 8(8), e13607.
Ettema, E. J., Derksen, L. D., & van Leeuwen, E. (2010). Existential loneliness and end-of-life care: A systematic review. Theoretical medicine and bioethics, 31(2), 141-169.
Gramling, R. et al. (2021). Conversational stories & self organizing maps: Innovations for the scalable study of uncertainty in healthcare communication. Patient Education and Counseling, 104(11), 2616-2621.
Guntuku, S. C., et al. (2019). Studying expressions of loneliness in individuals using twitter: an observational study. BMJ open, 9(11), e030355.
Haslam, S. A., et al. (2022). Social identity makes group-based social connection possible: Implications for loneliness and mental health. Current Opinion in Psychology, 43, 161-165.
Hipson, W. E., Kiritchenko, S., Mohammad, S. M., & Coplan, R. J. (2021). Examining the language of solitude versus loneliness in tweets. Journal of Social and Personal Relationships, 38(5), 1596-1610.
Holt-Lunstad, J., Smith, T. B., Baker, M., Harris, T., & Stephenson, D. (2015). Loneliness and social isolation as risk factors for mortality: a meta-analytic review. Perspectives on psychological science, 10(2), 227-237.
Holt-Lunstad, J. (2021). Loneliness and social isolation as risk factors: The power of social connection in prevention. American Journal of Lifestyle Medicine, 15(5), 567-573.
Karhe, L., & Kaunonen, M. (2015). Patient Experiences of Loneliness. Advances in Nursing Science, 38(4), E21-E34.
Kitzmüller, G., Clancy, A., Vaismoradi, M., Wegener, C., & Bondas, T. (2018). “Trapped in an empty waiting room”—the existential human core of loneliness in old age: a meta-synthesis. Qualitative Health Research, 28(2), 213-230.
Kotwal, A. A., et al. (2021). The epidemiology of social isolation and loneliness among older adults during the last years of life. Journal of the American Geriatrics Society, 69(11), 3081-3091.
Mansfield, L., et al. (2021). A conceptual review of loneliness in adults: Qualitative evidence synthesis. International journal of environmental research and public health, 18(21), 11522.
Murthy, V. H. (2020). Together: the healing power of human connection in a sometimes lonely world (First edition.). Harper Wave, an imprint of HarperCollinsPublishers.
Ong, A. D., Uchino, B. N., & Wethington, E. (2016). Loneliness and health in older adults: A mini-review and synthesis. Gerontology, 62(4), 443-449.
Princeton University “About WordNet.” WordNet. Princeton University. 2010.
Reinhardt, G. Y., Vidovic, D., & Hammerton, C. (2021). Understanding loneliness: a systematic review of the impact of social prescribing initiatives on loneliness. Perspectives in public health, 141(4), 204-213.
Rokach, A. (1988). The experience of loneliness: A tri-level model. The Journal of Psychology,122(6), 531-544.
Sjöberg, M., Beck, I., Rasmussen, B. H., & Edberg, A. K. (2018). Being disconnected from life: meanings of existential loneliness as narrated by frail older people. Aging & Mental Health, 22(10), 1357-1364.
Thesaurus.com. (n.d.). Dictionary.com. Retrieved August 12, 2022, from https://www.thesaurus.com/
Vedder, A., Boerner, K., Stokes, J. E., Schut, H. A., Boelen, P. A., & Stroebe, M. S. (2022). A systematic review of loneliness in bereavement: Current research and future directions. Current opinion in psychology, 43, 48-64.
Yanguas, J., Pinazo-Henandis, S., & Tarazona-Santabalbina, F. J. (2018). The complexity of loneliness. Acta Bio Medica: Atenei Parmensis, 89(2), 302.