6. PARTICIPANT RESULTS: KOREAN WAR VETERANS' VERSUS COMPARISON GROUP

Download Chapter 6 - PARTICIPANT RESULTS: KOREAN WAR VETERANS' VERSUS COMPARISON GROUP PDF 133kb

6.1 Age distribution, and application of weighting factor to comparison group
participants
6.2 Additional demographic measures
6.3 Military service experience
6.3.1 Korean War service experience and exposures
6.3.2 Other military service
6.4 Health behaviours: smoking and alcohol consumption
6.5 Psychological disorders: Anxiety, Depression and Posttraumatic Stress Disorder
6.6 Life satisfaction and quality of life
6.7 Hospitalisations and self-reported current medical conditions

Participant results are based on the 6,122 Korean War veterans, shown in Table 2, who completed and returned their questionnaire, and the 1,510 comparison group subjects, shown in Table 10, who completed and returned their questionnaire and who were subsequently assessed as being eligible for comparison group membership.

A brief guide to interpreting the statistical analyses results presented in the tables is provided at section 4.12.5 of this report.

6.1 Age distribution, and application of weighting factor to comparison group participants

The age distribution for the Korean War veteran and (unweighted) comparison group participants are shown in the left section of Table 11. The Korean War veteran group had a smaller proportion of participants under the age of 70, and over the age 80, than the comparison group. One half of the Korean War veterans were aged between 70 and 74 years. The mean age in both study groups was close to 75 years, and participants ranged in age from approximately 66 to just under 100 years old.

Table 11 . Age distribution for Korean War veteran participants, unweighted comparison group participants, and weighted comparison group participants

Using the proportions of Korean War veteran and comparison group participants in each age category shown, a weighting factor was calculated and applied to the results of the comparison group participants, as described in section 4.12. The right side of Table 11 shows the redistribution of comparison group participants in to age categories after application of the weighting factor. As expected, the weighted age distribution of the comparison group almost exactly matches that of the Korean War veteran participants.

The weighting factor has subsequently been applied to all descriptive results for the comparison group participants in the remaining results tables in this report.

[back to top]

6.2 Additional demographic measures

The breakdown of Korean War veteran and comparison group participants across additional demographic measures included in the participant questionnaire, are shown in Table 12.

Table 12 . Demographic measures

* Actual N from which each percentage score is derived varies by up to 1% fewer participants depending on the number of respondents to each question.
† Each P value refers to the effect of study group upon the dependent demographic measure, after adjustment for current age (65-69; 70-74; 75-79; 80-84; 85+ years) and adjustment for the other two demographic measures in the table, each entered as categorical variables. When entered as adjustment covariates, country of birth was recoded in to two categories (Australia; other), marital status was recoded in to two categories (married or defacto; other) and education was recoded in to three categories (grade 10 or below; grades 11, 12 or certificate; diploma or university).
± Due to small cell sizes, country of birth as a dependent measure was recoded in to three categories (Australia; New Zealand/UK/Ireland; other) to obtain this P value.

Small, statistically significant P values indicate that Korean War veteran and comparison group participants differ in their overall pattern of country of birth, current marital status and highest education level. The differences, however, appear to be subtle.

Both Korean War veteran (89%) and comparison group participants (86%) were predominantly Australian born. Other veteran participants were typically from the UK or Ireland (9%) or New Zealand (1%), whilst non-Australian born comparison group participants were less likely to be from these countries (4%) and more likely to be from other countries (11%).

The two participating groups demonstrated a similar pattern of marital status, though the Korean War veterans were slightly less likely to be married or in a defacto relationship, and slightly more likely to be widowered, or divorced, or never married, than the comparison group.

The Korean War veterans were also slightly less likely to have post-secondary education qualifications than the comparison group.

The participant questionnaire also included a question about whether participants considered themselves to be Aboriginal or Torres Strait Islanders. The results were unclear; whilst 0.6% of the Korean War veterans (n=36) and 0.5% (weighted) of comparison group participants answered Yes, a further 6% and 26% of the two groups respectively, did not answer the question. The true number of Aboriginal or Torres Strait Islander participants, therefore, cannot be accurately assessed.

[back to top]

6.3 Military service experience

6.3.1 Korean War service experience and exposures

Self-reported data from the participant questionnaire, and data drawn from the Korean War Nominal Roll, are both presented in Table 13 to describe the participating Korean War veterans in terms of their age when first deployed to the Korean War; number of years of service in the Australian armed forces prior to Korean War deployment; Service branch and highest rank during Korean War service; whether they first deployed during the mobile phase of the war prior to the end of June 1951, during the following static phase from July 1951 to 26 July 1953, or after the 27 July 1953 armistice; and the total duration of their Korean War deployment.

More than one half of the participating Korean War veterans were aged between 21 and 26 years at the time of their first Korean War deployment. The youngest participating veteran was 16, the oldest 47, at the time of deployment. More than 60% of this group deployed to Korea within four years of service with the Australian armed forces. Just under 5% of the group were in their first year of service, and 24% were in their second year.

74% of all participants served in Korea with an enlisted rank, and only 19% and 7% respectively served with a non-commissioned officer, or officer, rank. More than half of the participants served with the Army in Korea, one third served with the Navy, and the Air Force was represented by the smallest proportion of participants.

Just less than 17% of participants first deployed to Korea during the mobile phase of the war prior to 30 June 1951. More than 50% first deployed some time during the static phase between July 1951 and late July 1953. An additional 30% of participating veterans first deployed to Korea after the armistice was signed on 27 July 1953.

The Nominal Roll data showed that participating Korean War veterans averaged a total of 285 days of deployment (approximately nine and a half months), with one day being the least number of days deployed, and 1,188 days (approximately three years and three months) being the largest number of days deployed. Additional analysis (not tabulated) showed that, on average, Army participants were deployed for the largest number of days in total (mean total 323 days, SD 143.6), followed by Navy (mean total 255 days, SD 109.7), and Air Force participants were deployed for the least number of days (mean total 159 days, SD 133.1). Most participating veterans (78%) undertook one tour of duty during the Korean War; the highest number of tours recorded was six for Army veterans, five for Navy veterans, and 17 for Air Force veterans. The average duration of a tour was 245 days for the Army, 218 days for the Navy, and 108 days for the Air Force.

Table 13 . Korean War deployment characteristics

* Actual N from which each percentage or mean score is derived varies by up to 3% fewer participants depending on the number of respondents to each question.
† Age in years at first Korean War deployment is based on self-reported date of birth and Nominal Roll drawn deployment dates. Age categories were derived after rounding age in years to the nearest integer.‡ Derived from Nominal Roll drawn date of first Korean War deployment, and self-reported year of first joining the Australian armed forces.

The Korean War veterans’ participant questionnaire data in relation to Combat Exposure Scale (CES) responses, whether wounded in action and any associated category of evacuation, and any fever associated with the Korean War, are shown in Table 14, Table 15 and Table 16 respectively.

Table 14 shows that approximately one fifth (21%) of the Korean War veterans reported no combat exposure based on the scenarios described in the CES. These veterans reported, for example, no casualties in their unit, never having to fire rounds at the enemy, never seeing others injured by incoming rounds, and never being in danger of being injured or killed in the line of duty. More commonly veterans reported light combat exposure (31%), and large proportions also reported light-moderate (18%) and moderate (18%) combat exposure according to the CES scoring. A small proportion of the veterans (3%) reported heavy combat exposure.

Table 14 . Combat Exposure Scale measures for Korean War service

Additional analysis showed that combat exposure varied noticeably across Service branch (Figure 2), and across deployment era (Figure 4), but that there was little variability across subgroups of rank (Figure 3). Army veterans were much less likely to report light or no combat exposure, and much more likely to report moderate-heavy, or heavy, combat exposure, compared with veterans of the Navy and Air Force Services. In a similar pattern, veterans who first deployed to Korea during the active and static phases of the war were much less likely to report no combat or light combat, and much more likely to report moderate, moderate-heavy, or heavy combat exposure, compared with veterans who first deployed to Korea after the armistice. However across rank, veterans serving as officers or non-commissioned officers were only very slightly more likely than enlisted ranks to report no combat exposure, very slightly less likely to report moderate combat exposure, and equally likely to report heavy combat exposure.

Figure 2 . Combat Exposure Scale score category across Service branch

Figure 3. Combat Exposure Scale score category across rank during Korea

Figure 4 . Combat Exposure Scale score category across deployment era

Table 15 . Whether wounded in action during Korean War service, and any evacuation

Veterans who reported being wounded in action during the Korean conflict, and the types (categories) of evacuation required for their worst injury, are shown in Table 15. 871 (14%) participating veterans reported being wounded in action. This group was equally divided in regard to the four types of evacuation reported for their injuries. Approximately one quarter sustained injuries that were treated at a local aid post such as a sick bay or field ambulance before return to duties, a quarter reported evacuation to a local field or ship hospital before return, a quarter reported more extreme evacuation to a hospital in Japan before return, and the final quarter reported evacuation to Japan and then on to Australia, without return to duties in Korea. It is assumed that each increase in category of evacuation, from item 1 through to item 4 above, is likely (though not always) to represent an increase in injury severity.

Additional analysis showed that a large majority (n=782, 90%) of the 871 veterans who reported being wounded in action, had served in the Army during the Korean War (not tabulated). DVA provided a list of all Army veterans officially listed as being Wounded In Action (WIA) during the Korean War. Official WIA qualifying criteria pertained to individuals directly injured by the actions of the enemy, at or within close proximity to the battle line. All other injuries were excluded from these criteria, including accidental injury, transport accidents, frost-bite, trench foot and infections or fevers requiring treatment (such as malaria, influenza and pneumonia). A total of 341 of the Army participants in the study were officially listed by DVA as WIA. Of these, 321 (95%) reported being wounded in action in the participant questionnaire. Compared with the pattern of treatment and evacuation shown in Table 15 for all participants who reported being wounded in action, Army veterans on the DVA official WIA list were less likely to report that their injuries were treated at a local aid post such as a sick bay or field ambulance (12%), or at local field or ship hospital (23%), and more likely to report more extreme evacuation to a hospital in Japan before return (32%), or evacuation to Japan and then on to Australia without return to duties in Korea (34%). Of all Army veteran participants who reported being wounded in action and evacuated to a hospital in Japan before return (n=184), or evacuated to Japan and then on to Australia (n=192), 45% (n=171) were not officially listed by DVA as WIA.

The study participants were not given the official DVA definition for Wounded In Action when completing the questionnaire, and it was not anticipated that their responses would necessarily match with the official DVA WIA records. Instead, these data suggest that study participants may have included various injuries or illnesses, not necessarily just those which were a direct result of enemy action or within close proximity to the battle line, when reporting being wounded in action during the Korea War.

Table 16 . Korean War service related fever

Table 16 shows the responses of 5,968 (97%) veterans who reported whether or not they had ever been told that they had haemorrhagic fever, malaria or another type of fever during, or as a result of, the Korean War conflict. Veterans could report more than one type of fever, and therefore some veterans are represented in more than one of the rows in Table 16. 82% of respondents reported no fever, 1% reported haemorrhagic fever, and 13% reported malaria. Other fevers, reported by 5% of respondents, included glandular fever, dengue fever, hepatitis, pneumonia, or a fever of unknown type. As there is a haemorrhagic form of dengue fever, it is possible that a small number of the dengue fevers reported were also haemorrhagic fevers.

[back to top]

6.3.2 Other military service

Table 17 shows other military career characteristics of the participating Korean War veterans, as reported in the participant questionnaire. One quarter of participating Korean War veterans served for a total period of six years with the Australian armed forces before leaving. Almost half of the group served for 10 years or more. Additional analysis showed that the group reported a median total of eight years service (range 1 to 50).

Approximately 12% reported an officer rank as the highest military rank of their Australian service career, 45% reported a non-commissioned officer rank and 42% reported an enlisted rank.

In addition to their Korean War deployment, 17% also reported involvement in World War II, 26% in British Commonwealth Occupying Force (BCOF) in Japan, 21% in the Malayan Emergency, 9% in the Borneo/Malaysian confrontation and 14% in the Vietnam War. As veterans could report more than one major military conflict in addition to the Korean War, individuals may be represented more than once in the associated percentages presented above and in Table 17. For 45% of veterans, the Korean War was the only major military conflict to which they deployed.

Table 17 . Military career characteristics of participating Korean War veterans

* Actual N from which each percentage score is derived varies by up to 2% fewer participants depending on the number of respondents to each question.

[back to top]

6.4 Health behaviours: smoking and alcohol consumption

Smoking

Postal questionnaire responses were used to categorise Korean War veterans and comparison group participants according to whether they were current smokers, former smokers, or never or occasional smokers. For current smokers and former smokers additional calculations were carried out to estimate number of pack years of smoking.

The results, shown in Table 18, indicate that Korean War veterans (79%) were more likely than the comparison group (60%) to be current or former smokers. The very small accompanying P value indicates that difference in the overall pattern of smoking status between the two groups was statistically significant.

Amongst current or former smokers, Korean War veterans also averaged significantly higher pack-years of cigarette smoking, indicating that they have smoked in higher quantities or for longer durations than comparison group current or former smokers.

Table 18 . Smoking status and median total pack years for current smokers and former smokers

* This P value refers to the effect of study group after adjustment for current age (65-69; 70-74; 75-79; 80-84; 85+ years), education (primary; any secondary up to grade 10; grades 11, 12 or certificate; diploma or university), marital status (married or defacto; widowered; divorced or separated; single, never married) and country of birth (Australia; other).

† This P value for the effect of study group was obtained from median regression using 1,000 bootstrap replications after adjustment for age (<=74 years; >=75 years), marital status (married or defacto; other), education (grade 10 or below; grade 11, 12 or certificate; diploma or university) and country of birth (Australia; other).

Alcohol consumption

Responses to the AUDIT-C questionnaire items, measuring frequency and quantity of current alcohol consumption, are shown in Table 19. Korean War veterans were slightly less likely to be non-drinkers, and slightly more likely to drink four or more times per week, than comparison group participants. Amongst drinkers, Korean War veterans (24%) were much more likely to drink five or more drinks on a standard day of drinking than the comparison group (11%), and veterans were much more likely to binge drink (drink six drinks or more in one sitting) and to do so more frequently than the comparison group. Small P values indicate that the overall pattern of drinking, represented by each AUDIT-C question, did differ statistically significantly between the Korean War veteran and comparison group participants.

Table 19 . Frequency and quantity of alcohol consumption using the AUDIT-C

* Due to small cell sizes these P values were obtained after recoding number of drinks on a standard day to three categories (1 or 2; 3 or 4; 5 or more), and frequency of 6 drinks or more to 4 categories (never; less than once a month; monthly; weekly or more often), and after adjustment for age (<=74 years; >=75 years), marital status (married or defacto; other), education (grade 10 or below; grade 11, 12 or certificate; diploma or university) and country of birth (Australia; other).

The number of participants predicted to be current hazardous drinkers, using an AUDIT-C score of four or more, and also an AUDIT-C score of five or more, are shown in Table 20. At both thresholds, Korean War veterans were more than one and a half times more likely to be current hazardous drinkers than the comparison group.

Also shown in Table 20, and using the recommended cut-off score of two or more on the CAGE questionnaire, Korean War veterans were approximately three times more likely to meet criteria for a history of alcohol related problems indicative of dependence and/or abuse at some time in their lifetime.

Further, Korean War veterans were three times more likely to report ever considering themselves as heavy drinkers, and to report having had treatment for alcoholism.

Table 20 . AUDIT-C current hazardous drinkers, CAGE history of alcohol problems, and self-reported heavy drinking and treatment for alcoholism

* Actual value of N from which each percentage score is derived varies by up to 5% fewer participants depending on the number of respondents to each question.† These odds ratios are adjusted for age (65-69; 70-74; 75-79; 80-84; 85+ years).‡ These odds ratios and associated 95% CIs and P values are adjusted for age (65-69; 70-74; 75-79; 80-84; 85+ years), education (primary; any secondary up to grade 10; grades 11, 12 or certificate; diploma or university), marital status (married or defacto; widowered; divorced or separated; single, never married) and country of birth (Australia; other).

[back to top]

6.5 Psychological disorders: Anxiety, Depression and Posttraumatic Stress Disorder

Hospital Anxiety and Depression (HAD) scale

6,016 (98.3%) Korean War veterans and 1,479 (98.2%) comparison group participants responded to all seven of the HAD anxiety subscale questions, and 5,836 (95.3%) Korean War veterans and 1,432 (95.1%) comparison group participants responded to all seven of the HAD depression subscale questions. Total scores are scaled in a negative direction such that higher scores represent poorer health (increased symptom severity). The mean total scores on each subscale for these participants, and the number of participants meeting HAD scale criteria for anxiety or depression (by reaching the cut-off score of 11 or more on either subscale), are shown in Table 21.

Table 21 . HAD scale mean scores, and participants meeting HAD scale criteria for anxiety or depression

* Actual N from which each mean and SD, or percentage, score is derived varies by up to 6% fewer participants depending on the number of respondents to each of the HAD scale’s 14 questions.
† These OR and difference between mean values are both adjusted for current age (65-69; 70-74; 75-79; 80-84; 85+ years).
‡ These OR and difference between mean values, and their associated 95% CIs and P values are adjusted for current age (65-69; 70-74; 75-79; 80-84; 85+ years), education (primary; any secondary up to grade 10; grades 11, 12 or certificate; diploma or university), marital status (married or defacto; widowered; divorced or separated; single, never married) and country of birth (Australia; other).

Korean War veterans recorded significantly higher mean scores, representing poorer health, on both the anxiety and depression subscales of the HAD scale. The difference between groups was very slightly more marked for anxiety than for depression. On both subscales, Korean War veterans were more than five times more likely than the comparison group to be experiencing depression or anxiety respectively, as defined by a total subscale score of 11 or more.

Posttraumatic Stress Disorder Checklist (PCL)

5,564 (91%) Korean War veterans and 1,390 (92%) comparison group participants answered each of the PCL’s 17 symptom questions, and were given a total PCL score. Increasing PCL scores represent increasing number of symptoms (poorer health). Group mean and standard deviation (SD) PCL scores are shown in Table 20.

Korean War veterans recorded significantly higher mean PCL scores than the comparison group, representing markedly higher symptom reporting in the veteran group.

The number of participants predicted to meet criteria for a PTSD diagnosis, using a PCL score of 45 or more, or a more stringent PCL score of 50 or more, are also shown in Table 20. At both thresholds, Korean War veterans were close to six times more likely to meet criteria for a PTSD diagnosis than the comparison group participants.

Table 22. Total PCL scores, and number of participants meeting PCL criteria for PTSD with scores of 45 or greater, or 50 or greater

† These OR and difference between mean values are both adjusted for current age (65-69; 70-74; 75-79; 80-84; 85+ years).
‡ These OR and difference between mean values, and their associated 95% CIs and P values, are adjusted for current age (65-69; 70-74; 75-79; 80-84; 85+ years), education (primary; any secondary up to grade 10; grades 11, 12 or certificate; diploma or university), marital status (married or defacto; widowered; divorced or separated; single, never married) and country of birth (Australia; other).

PCL symptom scores were obtained for participants regardless of whether or not they nominated a stressful life event. Responses by participants who did nominate a stressful life event were categorised as either a Korean War event, other military event, other personal injury/illness/attack, witnessing of trauma to others, fire/flood or natural disaster, domestic event, other civilian event, or another/undetermined event. The category ‘other military event’ included experiences which were either clearly not Korean War events, or experiences which were possibly Korean War events but for which insufficient information was provided to be certain.

Domestic experiences were the most commonly reported stressful life event for both Korean War veterans (35%) and comparison group participants (52%). The second most nominated life event category for both groups was personal injury, illness or attack (assessed as not military related) with 18% of Korean War veterans and 21% of comparison group participants reporting an event of this nature. 18% of Korean War veterans nominated a Korean War event, and 13% nominated another military event. A further 11% of Korean War veterans and 12% of comparison group participants did not specify a stressful life event.

[back to top]

6.6 Life satisfaction and quality of life

Life Satisfaction scale

The Life Satisfaction scale was completed by 6,062 (99.0%) Korean War veterans and 1,506 (99.7%) comparison group participants, and their responses are shown in Figure 5. In relation to how they felt about their life as a whole, and taking in to account what had happened in the last year and what was expected to happen in the future, Korean War veterans were less likely than the comparison group to report feeling delighted, pleased or mostly satisfied, and more likely to report feeling mostly dissatisfied, unhappy or terrible.

Figure 5 . Life Satisfaction scale responses

The results for the Life Satisfaction scale were expressed as Percent Life Satisfaction (PLS) scores with higher scores representing greater life satisfaction. PLS mean and SD scores are shown in Table 23 and in Figure 6. Korean War veterans recorded significantly poorer PLS scores than the comparison group, representing poorer life satisfaction.

Table 23 . Mean Percent Life Satisfaction score

* This difference between means is adjusted for age (65-69; 70-74; 75-79; 80-84; 85+ years).
† This difference between means and associated 95% CI and P value is adjusted for age (65-69; 70-74; 75-79; 80-84; 85+ years), education (primary; any secondary up to grade 10; grades 11, 12 or certificate; diploma or university), marital status (married or defacto; widowered; divorced or separated; single, never married) and country of birth (Australia; other).

The multivariate adjusted mean difference value of -12.03 represents an effect size of approximately 0.5 of the pooled group standard deviation. This is defined as a medium effect size by Cohen (1988) ^[160] and, combined with a 95% CI which clearly excludes zero (indicating no difference), and an accompanying very small P value, this difference in mean Percent Life Satisfaction scores would be considered to represent an important or meaningful difference between the two groups.

Figure 6. Percent Life Satisfaction Mean +- 1 SD error bars

World Health Organisation brief Quality of Life questionnaire (WHOQOL-Bref)

Korean War veteran and comparison group responses to the WHOQOL-Bref’s two individual measures of self-rated quality of life, and satisfaction with health, are shown in Figure 7 and Figure 8.

Figure 7 . WHOQOL-Bref question 1 "How would you rate your quality of life?"

Figure 8 . WHOQOL-Bref question 2 "How satisfied are you with your health?"

Korean War veterans were less likely than the comparison group to rate their quality of life as good or very good, and more likely to rate their quality of life as very poor, poor or neither poor nor good. In a similar pattern, Korean War veterans were less likely to report being satisfied or very satisfied with their health, and more likely to report being very dissatisfied, dissatisfied or neither dissatisfied nor satisfied.

Mean scores for the WHOQOL-Bref’s two individual questions on self-rated quality of life, and satisfaction with health, each with a possible range of 1 to 5, and for the instrument’s four Domains, each with a possible range of 4 to 20, are shown in Table 24. All total mean scores are scaled in a positive direction such that higher scores denote higher quality of life.

Table 24 . WHOQOL-Bref scores

* Actual N from which each mean and SD score is derived varies by up to 2% fewer participants depending on the number of respondents to each of the WHOQOL-Bref’s 26 questions.
† This difference between means is adjusted for current age (65-69; 70-74; 75-79; 80-84; 85+ years).
‡ This difference between means and associated 95% CI and P value is adjusted for current age (65-69; 70-74; 75-79; 80-84; 85+ years), education (primary; any secondary up to grade 10; grades 11, 12 or certificate; diploma or university), marital status (married or defacto; widowered; divorced or separated; single, never married) and country of birth (Australia; other).

Korean War veterans receive poorer (lower) quality of life scores on all four Domains of the WHOQOL-Bref questionnaire, and also on the two individual measures of quality of life and satisfaction with health. The difference between groups on the Physical Health Domain was slightly larger than the differences between groups on the other WHOQOL-Bref Domains.

On each of the WHOQOL-Bref measures shown in Table 24 the multivariate adjusted mean difference values represent effect sizes of between 0.6 to 0.8 of the pooled group standard deviations. These are considered medium to large effect sizes ^[160] which suggest important or meaningful differences between the two groups on these quality of life measures.

Figure 9 further demonstrates the differences between the Korean War veteran and weighted comparison group results for each of the WHOQOL-Bref’s four Domains.

Figure 9 . WHOQOL-Bref Domain mean +-1 SD error bars

[back to top]

6.7 Hospitalisations and self-reported current medical conditions

Nights hospitalised in previous 12 months

5,986 (98%) Korean War veterans and 1,492 (99%) comparison group participants provided information about nights hospitalised in the previous 12 months.

Table 25 . Nights hospitalised in previous 12 months

*This odds ratio was adjusted for age (65-69; 70-74; 75-79; 80-84; 85+ years).†This odds ratio and associated 95% CI and P value was adjusted for age (65-69; 70-74; 75-79; 80-84; 85+ years), education (primary; any secondary up to grade 10; grades 11, 12 or certificate; diploma or university), marital status (married or defacto; widowered; divorced or separated; single, never married) and country of birth (Australia; other).‡This difference between medians was obtained from median regression after adjustment for age (65-69; 70-74; 75-79; 80-84; 85+ years).
§This difference between medians and associated 95% CI and P value was obtained from median regression using 1000 bootstrap replications after adjustment for age (<=74 years; >=75 years), marital status (married or defacto; other), education (grade 10 or below; grade 11, 12 or certificate; diploma or university) and country of birth (Australia; other).

Table 25 shows that Korean War veterans (35%) were more likely than comparison group participants (26%) to report being hospitalised overnight at least once in the previous 12 months.

Of the 2,111 Korean War veterans and the 383 comparison group participants who reported hospitalisations, 66% of the Korean War veterans and 74% of the comparison group participants reported 1 to 10 nights hospitalisation, 34% and 36% of the groups respectively reported 11 to 30 nights hospitalisation, and 10% and 7% respectively reported more than 30 nights hospitalisation (not tabulated). The total number of nights reported ranged from one to 365 in both groups. Amongst those hospitalised, there was no difference between study groups in the median number of nights reported, after statistical adjustment for the regular covariates age, education, marital status and country of birth (Table 25).

Self-reported current medical conditions

Participant’s self-reported doctor-diagnosed medical conditions, from a list of 15 conditions included in the participant questionnaires, are shown in Table 26. Unlike other health outcomes investigated in the questionnaire, a large number of participants did not fully complete the medical conditions questions; often not ticking either the Yes or No options to a possible condition. For example, 16% of Korean War veterans and 7% of comparison group participants did not answer whether they had asthma or not, and a similar percentage in each group did not indicate whether or not they had suffered stroke. The large percentage of missing responses for each medical condition rendered it difficult to reliably estimate the true group difference in the prevalences of each condition.

Because of the uncertainty in the data, the analysis was conducted using two approaches.

Firstly, the group differences were quantified with the ‘not answered’ participants excluded, and the results for these analyses are presented in the first row of Table 26 for each medical condition. Based on these analyses, Korean War veterans were between 1.5 and 3.5 times more likely to report having each of the listed medical conditions.

This first approach gives the same point estimates (age adj ORs and multiv. adj ORs) as those expected if the ‘not answered’ participants were, in actuality, distributed in to the Yes and No categories in similar proportions to those participants who did answer the questions. However, if the ‘not answered’ participants were added to the analysis in this way, the point estimates would stay the same but the 95% CI’s would become more precise (narrow) and the P values would become smaller (more significant).

Secondly, the group differences were quantified with the ‘not answered’ participants included as non-cases for each medical condition (i.e. the correct answer assumed to be No), and the results for these analyses are presented in the second row of Table 26 for each medical condition. This approach slightly reduced the point estimates observed in the first analysis approach, but still showed that Korean War veterans were between 1.4 and 3.0 times more likely than the comparison group to report having each of the listed medical conditions.

This second approach was based on two theories as to why a large number of participants failed to answer the medical conditions questions fully. The first theory assumes that participants often looked through the list of medical conditions and simply ticked the Yes boxes to the conditions which applied to them, and ignored the conditions which did not apply to them; thus failing to tick the No boxes. The second theory makes the assumption that if a participant does not know whether he has, or does not have, a doctor-diagnosed medical condition, and subsequently does not answer the question on that basis, then it is in fact unlikely that he has that condition; and the subsequent correct answer is No.

Table 26 . Self-reported current medical conditions

* These OR and difference between mean values are adjusted for age (65-69; 70-74; 75-79; 80-84; 85+ years).
† These OR and difference between mean values, and their associated 95% CIs and P values are adjusted for age (65-69; 70-74; 75-79; 80-84; 85+ years), education (primary; any secondary up to grade 10; grades 11, 12 or certificate; diploma or university), marital status (married or defacto; widowered; divorced or separated; single, never married) and country of birth (Australia; other).