Preventive home visits to older home-dwelling people in Denmark: are invitational procedures of importance?

Authors

  • A. Ekmann MScPH,

    1. Danish Aging Research Centre (DARC), University of Aarhus, Odense and Copenhagen
    2. Department of Social Medicine, Institute of Public Health, University of Copenhagen
    Search for more papers by this author
  • M. Vass MD DMedSci,

    1. Department of General Practice and Research Unit for General Practice, Institute of Public Health, University of Copenhagen
    2. Centre for Healthy Aging (CESA), Faculty of Health Science, University of Copenhagen, Copenhagen, Denmark
    Search for more papers by this author
  • K. Avlund PhD DMedSci

    1. Danish Aging Research Centre (DARC), University of Aarhus, Odense and Copenhagen
    2. Department of Social Medicine, Institute of Public Health, University of Copenhagen
    3. Centre for Healthy Aging (CESA), Faculty of Health Science, University of Copenhagen, Copenhagen, Denmark
    Search for more papers by this author

Anette Ekmann
Department of Social Medicine
Institute of Public Health
Faculty of Health Science
University of Copenhagen
Oester Farimagsgade 5
PO Box 2099
DK-1014 Copenhagen
Denmark
E-mail: anek@sund.ku.dk

Abstract

Since 1998 all municipalities in Denmark have been required by law to offer two annual preventive home visits to all home-dwelling citizens aged 75 or over. The influence of invitational procedures on acceptance rates has not been investigated. The aim of this study was to describe and investigate whether different invitational procedures were associated with first preventive home visit acceptance rates. The study was based on secondary analyses of data from the Danish Intervention Study on Preventive Home Visits. Data were collected from 1998 to 2002. Of the 4060 participants in the main study, 3245 reported receiving an offer for an identifiable preventive home visit, of whom 2399 (73.9%) provided complete data for the main analyses in the present study. Invitational procedures were categorised as: (1) a letter with a proposed date and time for the visit, (2) a visitor telephone call, and (3) a letter with encouragement to phone the visitor for appointment (letter without a proposed date). Covariates included sex, age, experience with preventive interventions, functional ability, self rated health, social relations and psychosocial characteristics. Statistical analyses included chi-square tests, and bi- and multivariable logistic regression analyses. Different invitational procedures were associated with first preventive home visit acceptance rates. Significantly more men (75.1%) than women (62.8%) declined the first preventive home visit regardless of the invitational procedure. Compared to ‘letter with a proposed date’, men had an odds ratio of 1.78 (95% CI: 1.16–2.74) for declining visits when ‘telephone call’ was used and an odds ratio 2.81 (95% CI: 1.79–4.40) when ‘letter without a proposed date’ was used as the invitational procedure. In women the odds ratios were 1.23 (95% CI: 0.91–1.68) and 1.87 (95% CI: 1.37–2.55), respectively.

Introduction

Since 1998 all municipalities in Denmark have been required by law to offer two annual preventive home visits to all home-dwelling citizens aged 75 or over. The preventive home visit is an elective offer which the older person can accept or decline. The aim is ‘to build confidence, trust and well-being; to offer advice and guidance about activities and possibilities for support; to facilitate older person’s use of own resources and maintain functional ability as long as possible’ (Danish Ministry of Welfare 1995, 2006). Evidence on the effect of preventive home visits is contradictory. Several reviews and meta-analyses have reported beneficial effects of preventive home visits on mortality and nursing home admission rates (Stuck et al. 1993, 2002, Elkan et al.2001, Markle-Reid et al. 2006, Beswick et al. 2008, Huss et al. 2008), but others have failed to show an effect (van Haastregt et al. 2000). In the Danish national home visit programme, a municipality-randomised intervention study demonstrated that education of preventive home visitor and GP was beneficially associated with functional ability (Vass et al. 2004) and nursing home admission rates (Vass et al. 2005). Effects were predominantly seen among the oldest old (Vass et al. 2005, Vass 2009), in those who were non-disabled at baseline (Vass et al. 2005), and among women (Vass et al. 2004, Avlund et al. 2007). The preventive home visit has been described as a complex social process influenced by numerous factors (McNaughton 2000, Clark 2001, Vass et al. 2007a). These can be exemplified by four steps described in the theoretical approach developed by McNaughton (2000). Step 1 (pre-entry) describes factors existing prior to the visit that influences the preventive home visit and the relationship between the visitor and the older person, e.g. experience with preventive home visits, the older person’s interpersonal relationships, perceived need for preventive home visits, and sources and reasons for the home visit referral (McNaughton 2000). Step 2 and 3 describe the relationship between the visitor and the older person, and finally, step 4 describes factors that influence the termination of the relationship (McNaughton 2000). Previous studies have concentrated on measuring effects of preventive home visits and in recent years also on identifying factors which give the highest impact (steps 2 and 3). These studies showed that the effects of the preventive home visits were associated with the home visitors’ knowledge and skills, the number and regularity of visits, and the age, health and functional ability of the older person, (Stuck et al. 2002, Markle-Reid et al. 2006, Vass et al. 2007a). To our knowledge, only one study has paid attention to factors influencing the acceptance of preventive home visits (step 1). Avlund et al. (2008) concluded that older men reporting sadness, low aggressiveness, low life satisfaction and need of help in daily activities were more likely to accept preventive home visits. Among older women a strong sense of coherence and low social participation were associated with higher acceptance rates of preventive home visits (Avlund et al. 2008).

To secure participation in the preventive home visit programme the invitational procedure must aim at reaching the target population and their special needs and preferences. However, the Danish law on preventive home visits gives little guidance to the municipalities on how to offer the preventive home visits. The only claims are that the invitation is to be given individually, twice a year and designed in a way which gives the older person opportunity to accept or decline the offer. This lack of guidance has resulted in a substantial variation of invitational procedures to preventive home visits in daily routine. With detailed information on how the invitational procedures in 34 Danish municipalities were performed in 1998, three distinct procedures were identified. Thus, the aim of this study was to describe and evaluate how these three different invitational procedures were associated with first preventive home visit acceptance rates amongst men and women aged 75 or over in Denmark.

Methods

Study sample

Our study was based on secondary analyses of data from the Danish Intervention Study on Preventive Home Visits. The aim of the primary intervention study was to investigate whether education of home visitors and GPs was associated with functional mobility of older persons (Vass et al. 2002, 2005). The intervention study was designed as a prospective controlled follow-up study with randomisation at municipality level and outcome measured at individual level. Addresses were drawn from the Civil Registration Office, and questionnaires were sent to all non-institutionalised citizens aged 74/75 or 80 years in the participating 34 municipalities. Hence information for study purpose on age, gender, functional ability, psychosocial factors, social relations and experiences with the preventive home visits offer and completion were attained independent of the preventive home visit. In total 5788 non-institutionalised citizens born in 1923/1924 (aged 74/75) and 1918 (aged 80), were invited to participate in the baseline study. In total 4060 persons (70.1%) gave informed consent to take part in the baseline-questionnaire study. Baseline data were collected in 1998/1999 and follow-up questionnaires were administered after 1½, 3 and 4½ years.

In the present study we initially included the 3245 participants who reported having received the first offer of preventive home visit from the municipality, and with whom it was possible to identify the invitational procedure. Due to missing values in covariates included in the multivariable logistic regression, subsequent analyses were conducted within a restricted study sample answering the questions on outcome and the included covariates (n = 2399, 73.9%). Invitational procedures did not differ systematically across intervention and control municipalities. In addition, there were no demographic differences between the primary study sample (n = 4060), our study sample (n = 3245) and our restricted study sample (n = 2399). The primary study design and study sample have been described in detail elsewhere (Vass et al. 2002).

Exposure

Process data from the intervention study provided detailed information on invitational procedures in the municipalities. At baseline copies of the invitational letter were available from 23 municipalities. To obtain information on invitational procedures from the remaining 11 municipalities, they were contacted by phone to recall the exact invitational procedure at baseline. Due to change of the employee in charge, one municipality was not able to provide exact information on how they offered the home visits at baseline.

In total three invitational procedures were clearly identified: (1) ‘Letter with a proposed date’; the visitor sends a letter proposing a date and time for the visit. If the appointment has to be cancelled or changed the older person has to phone the visitor; otherwise, the visitor will pay a visit at the stated date and time. (2) ‘Telephone call’; the visitor contacts the older person by telephone to offer the home visit and to set a date and time if the offer is accepted, and (3) ‘Letter without a proposed date’; an invitational letter is used; however, it has no date but a request to phone the home visitor for an appointment. A visit is not arranged if the older person does not phone the visitor.

Outcome

Acceptance of the first preventive home visit was measured at baseline by the question: ‘Have you accepted the invitation from the municipality to receive a preventive home visit?’: yes/no.

The Danish law on preventive home visits was implemented in 1998 at the same time as the baseline study was accomplished. Two years prior to the 1998 law in 1996 all municipalities were imposed to offer preventive home visits to all citizens aged 80 or over. As the municipalities used some time to implement the 1996-law, it is reasonable to assume that the 80-year-old men and women in the study were offered maximum one or two preventive home visits before the study period. Accordingly, the answer on acceptance of the municipality invitation is a good proxy of accepting or declining the first invitation to preventive home visits.

Covariates

Covariates included age, self reported health, functional ability, social relations, psychosocial factors and experience with preventive initiatives. Data on covariates were retrieved from The Danish Personal Registry and the baseline questionnaire, except the questions on experience with preventive initiatives, which were answered at 3 and 4½ year follow-up, respectively. All covariates were selected on basis of knowledge of the association with the outcome. The original codings of the covariates and dichotomised interpretations are described below. We ran analyses with variables in both formats. There were few differences between the results and no differences between conclusions, so results using simpler dichotomised variables are presented in the paper.

Functional ability was measured as ‘tiredness in daily activities’ by the Mobility-Tiredness-scale (Mob-T) and ‘in need of help in daily activities’ by the Mobility-Help-scale (Mob-H) (Avlund 2004). The scales were formed by answers to questions about six activities: transferring from chair/bed, walking indoors, going outdoors, walking outdoors in nice weather, walking outdoors in poor weather and climbing stairs. The Mob-T-scale describes whether participants perform the activities with or without being tired afterwards and counts the number of items performed without tiredness. The Mob-H-scale describes whether participants manage the activities with or without help and counts the number of items managed without help. For analysis, the scales were dichotomised into persons with maximum score (not tired, not needing help) and persons below maximum score (tired, needing help). Self-reported health was measured by the global question: ‘In general, how would you say your health is?’ Responses ‘poor’ and ‘not so good’ were treated as ‘poor’, and responses ‘good’ or ‘very good’ were treated as ‘good’ for analysis. For previous experience of preventive health, participants were asked at the 4½ year follow-up whether they had children and had used a health nurse when the children were newborn. For present use of preventive initiatives, they were asked at the 3-year follow-up if they had received vaccination against influenza.

In terms of social relations, participants were asked whether they lived alone. Social participation as measured by three questions: ‘How often do you participate in leisure time activities outside your home?’, ‘Did you have visitors at home during the previous month?’ and ‘Have you been on visits during the previous month?’ Social participation was considered ‘low’ if there were less than three and ‘high’ if there were three or more social interactions per month. Instrumental social support was measured by four questions covering cleaning, hot meals, shopping and minor house repairs. Participants were considered to have no support if they did all the activities themselves, and to have support if they had help in at least one activity.

Participants were asked whether they felt lonely, whether they sometimes felt sad without reason, whether they sometimes got aggressive and short-tempered without reason, and whether they felt old. Responses such as often, sometimes or seldom were treated as ‘yes’ for analysis with only no treated as ‘no’. Sense of coherence was assessed using three questions: (i) manageability was measured by the question: ‘Do you usually see a solution to problems and difficulties that other people find hopeless?’; (ii) meaningfulness by the question: ‘Do you usually feel that your daily life is a source of personal satisfaction?’; and (iii) comprehensibility by the question: ‘Do you usually feel that the things happening to you in your daily life are hard to understand?’ The response alternatives with were ‘yes, usually’; ‘yes, sometimes’; and ‘no’. Each response was scored either 0 (best response), 1 (yes, sometimes) or 2 (worst response) and the summed score was computed. For the analyses, a score of 0–2 was considered to show high sense of coherence, while a score of 3-6 was considered to show a low sense of coherence.

Statistical analyses

Analyses were conducted with spss version 15.0 and included Pearson’s chi-square test, bi- and multivariable logistic regression analyses with a significance level of 5%. Several studies have indicated that gender is an effect-modifier (Poulsen et al. 2007, Avlund et al. 2008). Hence, the multivariable logistic regression analyses were stratified by sex.

We conducted the initial descriptive analyses on the initial study sample. In the process of finding a suitable statistical model of the association between invitational procedure and acceptance rate, we grouped the covariates (demographic factors, functional ability and health, health behaviour experience, social relations, and psychosocial factors) and conducted forward group-wise selection. Covariates that were significantly associated with the outcome and at the same time altered the association between invitational procedure and acceptance rate more than 5% were included in the final multivariable logistic regression analyses (data not shown). In both men and women ‘tiredness in daily activities’ and ‘use of influenza vaccination in old age’, and in addition among men, ‘loneliness’ met the criteria and was included in the final multivariable logistic regression analyses. The restricted study sample in the final multivariable analyses included respondents answering the questions on ‘acceptance of preventive home visit’; ‘tiredness in daily activities’, ‘influenza vaccination’ and ‘loneliness’, and with whom it was possible to identify the invitational procedure. Patterns of association for comparable analyses were the same in the initial and the restricted study.

Ethics

The original study complies with the Declaration of Helsinki and was approved by relevant Regional Research Ethical Committees.

Results

The initial study sample comprised 1779 women and 1466 men aged 74/75 or 80. In terms of relative risk (RR) for women compared to men (ratio of the percentage for women compared to the percentage for men in Table 1), women were more likely to live alone (RR = 2.39), require instrumental support (RR = 2.05), feel lonely (RR = 1.61), have a high social participation (RR = 1.20), need help in daily activities (RR = 1.34) and self-rate their health as poor (RR=1.30), while men were more likely to be aggressive (RR = 0.76). The relative risk for feeling sad (RR = 1.17) was about the same as that for being 80 (Table 1).

Table 1.   Characteristics of participants reporting first offer of an identifiable preventive home visit (n = 3245)
CovariatesMenWomenP*
n%n%
  1. * Pearson’s chi-square test.

Demographic factors
 Age1466 1779  
  74/75 years 71.8 66.9 
  80 years 28.2 33.1<0.01
Functional ability and health
 Tiredness in daily activities1466 1778  
  No 67.8 56.8 
  Yes 32.2 43.2<0.001
 Needing help in daily activities1466 1778  
  No 83.2 74.0 
  Yes 16.8 26.0<0.001
 Self rated health1293 1595  
  Good  75.9 68.7 
  Poor  24.1 31.3<0.001
Health behaviour experience
 Prior use of health nurse 625 946  
  Yes 74.7 75.1 
  No 25.3 24.90.88
 Vaccination against the flu in old age1060 1347  
  Yes  56.3 54.7 
  No  43.7 45.30.37
Social relations
 Co-habitation1464 1778  
  Together  75.7 42.0 
  Alone  24.3 58.0<0.001
 Social participation1465 1777  
  Low  58.0 49.6 
  High  42.0 50.4<0.001
 Instrumental support1466 1779  
  No  73.7 46.2 
  Yes  26.3 53.8<0.001
Psychosocial factors
 Loneliness1459 1775  
  No  82.6 71.9 
  Yes  17.4 28.1<0.001
 Sense of coherence1458 1775  
  High  72.4 69.7 
  Low 27.6 30.30.10
 Sadness without reason1444 1761  
  No  48.6 40.1 
  Yes 51.4 59.9<0.001
 Aggressiveness 1450 1757  
  No  54.3 65.3 
  Yes  45.7 34.7<0.001
 Feeling old1454 1770  
  No  68.7 65.6 
  Yes  31.3 34.40.06

We included 1345 women and 1054 men in the restricted study sample. More than half (65.8%) received the first invitation for a preventive home visit as a letter with a proposed date, compared with 15.8% who received a telephone call and 18.3% who received a letter without a proposed date (Table 2).

Table 2.   Distribution of invitational procedures for the first invitation to a preventive home visit in men and women
 MenWomenTotal
n%n%n%
  1. Pearson’s χ2 = 1.08, df = 2, P = 0.58.

Letter with date70366.787665.1157965.8
Telephone call15815.022216.538015.8
Letter without date19318.324718.444018.3
Total105410013451002399100

Overall, significantly fewer women (62.8%) than men (75.1%) declined the offer of a preventive home visit (Table 3), equivalent to an odds ratio of 0.56 for women declining compared to men. This pattern was consistent across the three invitational procedures, with significant odds ratios of 0.61, 0.42 and 0.41 for women compared to men declining the invitation following ‘letter with a date’, ‘telephone call’ and ‘letter without a date’ respectively (Table 4).

Table 3.   Distribution of accepting of the first preventive home visits in men and women
 MenWomenTotal
n%n%n%
  1. Pearson’s χ2=41.36, df=1, P < 0.001.

Accept26224.950037.276231.8
Decline79275.184562.8163768.2
Total105410013451002399100
Table 4.   Odds ratio for decline of the first preventive home visit in the three invitational procedures by sex (n = 2399)
 MenWomenP
OR (reference)OR (95% CI)
Total10.56 (0.47–0.67)<0.001
Letter with date10.61 (0.50–0.76)< 0.001
Telephone call10.42 (0.26–0.69)<0.001
Letter without date10.41 (0.25–0.68)<0.001

Table 5 shows the crude and adjusted odds ratios for declining the invitation to a preventive home visit among women. Both crude and adjusted analyses showed the highest acceptance rate when ‘letter with a proposed date’ was used and the lowest acceptance rate when ‘letter without a proposed date’ was used as invitational procedure. When compared to ‘letter with a proposed date’, the crude model showed a non-significant odds ratio of 1.23 (95% CI: 0.91–1.68) and a significant odds ratio of 1.87 (95% CI: 1.37–2.55) for declining visits with ‘telephone call’ and ‘letter without a proposed date’, respectively. Adjustment by covariates did not alter the odds ratio for declining offers received by ‘telephone call’, but adjustment slightly increased the odds ratio for declining visit with ‘letter without a proposed date’.

Table 5.   Crude and adjusted odds ratio for declining first preventive home visit by invitational procedures among women (n = 1345)
 Crude
OR (96% CI)
Model 1
OR (95% CI)
Model 2
OR (95% CI)
Model 3
OR (95% CI)
Invitational procedure
 Letter with a proposed date1111
 Telephone call1.23 (0.91–1.68)1.22 (0.90–1.68)1.23 (0.91–1.68)1.23 (0.90–1.68)
 Letter without a proposed date1.87 (1.37–2.55)1.93 (1.40–2.64)1.91 (1.40–2.62)1.97 (1.43–2.71)
Functional ability and health
Tiredness in daily activities
 Yes vs. No (ref.) 0.50 (0.40–0.63) 0.50 (0.40–0.64)
Health behaviour experience
Use of influenza vaccination in old age
 No vs. Yes (ref.)  1.51 (1.13–2.02)1.42 (1.13–1.80)

Table 6 shows the crude and adjusted odds ratios for declining the invitation to preventive home visits among men. We found the same trend as for women, with the highest acceptance rates when ‘letter with a proposed date’ was used, and the lowest acceptance rates when ‘letter without a proposed date’ was used. When compared to ‘letter with a proposed date’, the crude model showed significant odds ratios of 1.78 (95% CI: 1.16–2.74) and 2.81 (1.79–4.40) for declining visits with ‘telephone call’ and ‘letter without a proposed date’, respectively. Adjustment by covariates did not appreciably alter the odds ratio for declining home visit offers received by ‘telephone call’, but unlike in women, adjustment by covariates among men slightly decreased the odds ratio for declining the visit when offered by ‘letter without a proposed date’.

Table 6.   Crude and adjusted odds ratio for declining first preventive home visit by invitational procedures among men (n = 1054)
 Crude
OR (95% CI)
Model 1
OR (95% CI)
Model 2
OR (95%CI)
Model 3
OR (95% CI)
Model 4
OR (95% CI)
Invitational procedure
 Letter with a proposed date11111
 Telephone call1.78 (1.16–2.74)1.83 (1.18–2.82)1.80 (1.17–2.78)1.72 (1.12–2.66)1.78 (1.15–2.76)
 Letter without a proposed date2.81 (1.79–4.40)2.70 (1.71–4.25)2.74 (1.75–4.31)2.76 (1.75–4.33)2.62 (1.66–4.14)
Functional ability and health
Tiredness in daily activities
 Yes vs. No (ref.) 0.49 (0.36–0.67)  0.56 (0.41–0.76)
Health behaviour experience
Use of influenza vaccination in old age
 No vs. Yes (ref.)  1.50 (1.19–1.87) 1.49 (1.11–2.00)
Psychosocial factors
Loneliness
 Yes vs. No (ref.)   0.48 (0.33–0.69)0.54 (0.37–0.80)

To further explain this difference between men and women we investigated the possibility of an epidemiological interaction term between tiredness in daily activities and invitational procedure. Though not significant, we found that tired women who received the invitation as ‘letter without a proposed date’ had an OR = 0.93 (95% CI: 0.60–1.45) of declining the offer compared to women who were not tired and who received the invitation as ‘letter with a proposed date’. Among men we found the opposite pattern. Tired men who received the invitation as ‘letter without a proposed date’ had an OR = 1.14 (95% CI: 0.52–2.46) for declining the offer compared to men who were not tired and who received the invitation as ‘letter with a proposed date’.

Discussion

Three distinct invitational procedures were identified: (1) ‘letter with a proposed date’, (2) ‘telephone call’ and (3) ‘letter without a proposed date’. Our study showed that these three invitational procedures were associated with the first preventive home visit acceptance rates. Regardless of invitational procedure, men had lower acceptance rates compared with women. In both men and women the highest acceptance rate was seen when ‘letter with a proposed date’ was used. A lower acceptance rate was seen with ‘telephone call’ and the lowest acceptance rate was observed when ‘letter without a proposed date’ was used as invitational procedure.

To our knowledge, the relationship between invitational procedure and acceptance of first preventive home visit has never been studied before. The relation between invitational procedure and acceptance has been studied in relation to invitation to population-based cohort studies. These studies indicate that a higher prevalence of participation is attained if the invitation is personal and direct (Eaker et al. 1998, Edwards et al. 2002, Manjer et al. 2002, Brogger et al. 2003), if the study is assessed as relevant by those invited (Minder et al. 2002), and if a follow-up contact to non-respondents is included. In Denmark all invitations to preventive home visits are prescribed to be personal and direct by law (Danish Ministry of Welfare 1995). Invitations by ‘letter with a proposed date’ or ‘telephone call’ are probably more personal and could likely be one explanation of the observed higher acceptance rates. A study of relations between invitational procedure to subsequent preventive home visit acceptance rate (Ekmann 2008) showed that direct and personal follow-up (letter with a proposed date) to persons previously declining invitations to preventive home visits was associated with a higher acceptance rate than follow-up invitations with methods less direct and personal (‘letter without a proposed date’ or ‘telephone call’). However, the same study also showed that a less personal and direct invitation (‘letter without a proposed date’) gave a higher acceptance rate among older persons who had previously accepted four or more preventive home visits. This indicates that the relationship between invitational procedures and acceptance of preventive home visit is complex and probably modified by the number of previous offers and visits.

Another explanation might be the simple fact that people get things done with as little effort as possible in everyday life. When no effort or action follows ‘letter without proposed date’ no visits will be arranged, but no action following ‘letter with proposed date’ results in implemented visits. This simple cost effort given might be another explanation of why more visits are accepted following ‘letter with proposed date’.

In addition to a relation between invitational procedure and acceptance of the first preventive home visit, our study indicated a higher acceptance rate among men and women with tiredness in daily activities, and among men feeling lonely. This higher acceptance rate among frail older persons was seen regardless of invitational procedures. However, for ‘letter without a proposed date’ different patterns for men and women were seen when adjusting for tiredness in daily activities, use of influenza vaccination and loneliness. Among women the odds ratio for declining visits increased when we controlled for tiredness in daily activities and the use of influenza vaccination. In contradiction, the findings among men showed decreasing odds ratios for declining visits when controlled for tiredness in daily activities, use of an influenza vaccination and loneliness. Further analysis showed that for an invitation in the form of a letter without a proposed date, tired men were slightly more likely to decline and tired women were slightly less likely to decline than those who were not tired. This difference in declining behaviour among older tired men and women when they received the offer of preventive home visit as ‘letter without a proposed date’, may explain some of the gender differences in acceptance rates.

Our findings clearly demonstrated that it is necessary to carefully consider the possible consequences of invitational procedures for the first preventive home visit. The municipality choice of invitational procedure seems to be associated with the number of visits accepted, the association is different in men and women, and may differ with high and low tiredness in daily activities.

It is possible that factors not included in these analyses influence the decision to accept the preventive home visit. First, studies have shown that the substance of the letter (Edwards et al. 2002), and living in rural or urban areas influence participation rates in cohort studies (Eaker et al. 1998). These factors probably have the same influence on acceptance rate in relation to preventive home visits. Finally, it is possible that both organisational factors in the municipally (Vass et al. 2007b) and social capital (Holstein 2006) influence both the method of invitations chosen by the municipalities and older persons’ acceptance of the preventive home visit.

The large study sample was strength of the study. The findings should be generalisable to the older Danish population aged 75 or more since it is a large sample from different areas and settings in Denmark. The sample may also reflect the inevitable self-selection bias that occurs when adults are invited to provider-initiated health promotion initiatives.

Potential participants were excluded if it was not possible to verify the invitational procedure used. We were in possession of copies of invitational letters and specification on how to handle the telephone call from 23 municipalities. The variance with regard to choice of the words, the length and the design within each of the three categories was small. In addition, the municipalities seldom deviated from their specifications due to minimising administrative working hours regarding first invitation. At the first invitation, the municipally has no prior knowledge of the older persons’ individual needs and preferences. Choice of invitational procedure is thus based on rules and administrative regulations only. The municipality measure of invitational procedure is hence viewed as a good proxy for the actual invitational procedure received by the individual.

Further exclusion of potential participants was done for the more detailed analyses. Here we only included participants who had no missing data on both questions of exposure, outcome and included covariates. We believe that this does not bias our results as there were no demographic differences between the 2399 included and the 846 excluded from these analyses.

The data collection was not designed for the present study, and a potential limitation was the non-validated question that constituted the outcome measure: ‘Have you accepted the invitation from the municipality to receive a preventive home visit?’ This question was also used as exclusion criteria as individuals answering that they had not received an invitation were excluded. Exclusion on this question might lead to a lower sensitivity as we might have excluded individuals that actually did get the invitation but have forgotten it afterwards. It is plausible that individuals getting the invitation as ‘letter without a proposed date’ and who subsequently declined the invitation have a greater probability of not recalling getting the invitation, as declining visits following this invitational procedure does not need further action. With this possible bias it is likely that our estimates are underestimated.

In conclusion, invitational procedure influences older person’s decision of acceptance or decline of municipality-offered preventive home visits. The relationships vary between men and women, and between frail and non-frail older persons. It is recommended to carefully consider the consequences of the invitational procedures used for preventive home visits to older home-dwelling persons.

Acknowledgements

The study was supported by the Danish Medical Research Council, the Research Foundation for General Practice and Primary Care, Eastern Danish Research Forum, the County Value-Added Tax Foundation and the Danish Ministry of Social Affaires. We thank all participating municipalities and Eva Jepsen for following up questionnaires, Christian Cato Holm for data management and development of the municipality registration software, and Carsten Hendriksen for constructive comments.

Conflict of interest

There is no conflict of interest between the funding and the research accomplished in relation to this paper.

Ancillary