Multimorbidity (the presence of two or more chronic conditions in a single patient)¹ is a major issue for primary healthcare² both in New Zealand and internationally.^3,4 There is a limited evidence base on the epidemiology of multimorbidity in New Zealand,⁵ particularly in relation to its prevalence in a primary healthcare population.⁶ Nonetheless, in line with the international literature,^7–9 the New Zealand prevalence increases with age, is more common and occurs earlier in those living in areas with high socioeconomic deprivation and disproportionately affects indigenous people.^5,6 In addition, cohort studies show that multimorbidity leads to poorer health outcomes: it is associated with high mortality, reduced functional status and quality of life, increased use of inpatient and ambulatory healthcare and polypharmacy.^10,11

To date, there has been no research reporting on the prevalence of multimorbidity and its relationship with multiple social disadvantage (MSD) in a New Zealand high-needs primary healthcare population. The term MSD^12,13 is helpful in understanding the pathways by which social inequalities may lead to individuals experiencing disadvantage in multiple areas of life concurrently¹⁴ and how these individuals engage with a range of health and social services aimed at addressing these disadvantages (eg, primary healthcare, housing, criminal justice system).¹³ MSD can be framed as a patient level measure which encompasses a range of social disadvantage domains: health, income, housing, education, employment, material wellbeing, safety and social connectedness.¹⁴ Homelessness, previous imprisonment and problem drug use are some domains that have been highlighted as being important markers for MSD in a high-needs (specialist homeless service) UK primary healthcare population.¹⁵ The extent of MSD was also found to be positively associated with multimorbidity in this study.¹⁵

In New Zealand, patients experiencing MSD will have difficulty accessing mainstream primary healthcare services, due at least in part to the financial barrier to access caused by the co-payment model.^16,17 In the 2016/2017 New Zealand Health Survey, 28% of respondents reported an unmet need for general practice services, with 20% of those living in the most socioeconomically deprived areas indicating cost as a reason.¹⁸ There exist, however, a number of ‘third sector’ (non-government, non-profit)^19,20 New Zealand primary healthcare clinics, which provide care to vulnerable populations including those experiencing MSD.21 One such clinic is Dunedin’s Servants Health Centre (SHC), which provides free healthcare through the voluntary services of doctors, nurses and counsellors to a ‘high needs’ population^22,23 (defined as the general practice having >50% Māori, Pacific and New Zealand deprivation quintile 5 patients).²⁴

The aim of this study was to determine the prevalence of multimorbidity and MSD in relation to age, gender, deprivation and ethnicity in a New Zealand high-needs primary healthcare clinic population (SHC).

Methods

Study design, setting and participants

A cross-sectional study of manually extracted patient-level data from computerised medical records (practice management system in use: MedTech software) was conducted in January 2018. All patients permanently enrolled at SHC as of 1 November 2017 (n=375) were included in the study.

Data collection

The patient-level data was manually extracted from electronic medical records onto an Excel spreadsheet by SS under the supervision of TS. The dataset comprised of administratively and clinically recorded data. This included age, sex, ethnicity, deprivation quintile, chronic medical conditions, multimorbidity, multiple social disadvantage (MSD) and service utilisation (defined as engagement with selected secondary care and community services; emergency department attendances and primary healthcare consultations over the previous calendar year). Deprivation was determined by the area in which the patient lived (NZ Dep2013).^25,26 We selected 38 common morbidities seen in a high-needs primary healthcare population^8,15 derived from the list of 40 long-term conditions presented in a widely cited Scottish primary care multimorbidity cross-sectional study.⁸ These 38 morbidities (See Supplementary Table 1) were selected through discussion between TS, SS and LR. The inclusion criteria were the condition being common in a high-needs population and being able to be recorded through use of general practice disease classification (READ codes) alone. Multimorbidity was defined as two or more long-term health conditions (LTCs).⁸ Social disadvantage domains included previous imprisonment, in receipt of health and disability benefit, homelessness, domestic violence or other reports of violence and lack of food. These social disadvantage domains were based on those used in a high-needs UK primary healthcare population.¹⁵ A detailed description of data collected can be found in Table 1.

Table 1: Collected data variables and data source.

‍

*WINZ = Work and Income New Zealand.

Data analysis

A descriptive statistical analysis was conducted by SS and TS using frequencies, percentages and cross tabulations with reported exact 95% confidence intervals (CI) as appropriate. SPSS Statistics 24 was used for the descriptive analysis. 95% CI were calculated using Epi Info 7. 2. 2. 2.

This study was granted ethical approval by the University of Otago ethics committee (H17/122).

Results

The medical records of all permanently registered patients (n=375) as of 1 November 2017 were reviewed. The practice demographics are presented in Table 2. The mean age of patients was 41.9 years ((median age 42, Interquartile range (IQR) 30-53)).

Table 2: Prevalence of multimorbidity by age, sex, ethnicity and New Zealand deprivation quintile.

‍

*NZDep2013; ** Other: Asian, Other European, African and Middle Eastern.

Multimorbidity and comorbidity

Most patients had multimorbidity (76.5%, 95% CI 72.0–85.5%) and half (49.9%, 95% CI 44.8–54.9%) had long-term physical and mental health comorbidities. These findings were consistent across all ethnic groups (New Zealand European; Māori; Pacific and other). A majority of patients (58.7%, 95% CI 53.5–63.7%) had three or more long-term conditions and a quarter (24.8%, 95% CI 20.5–29.5%) had five or more long-term conditions. The prevalence of multimorbidity in patients in all ethnic groups increased with age and was high across all deprivation quintiles. Table 2 presents the prevalence of multimorbidity and physical and mental health comorbidity by age, sex, ethnicity and New Zealand deprivation quintile.

Long-term conditions (LTCs)

Both long-term physical and mental health problems were common: most patients had a long-term mental health condition (74.9%, 95% CI 70.3–79.1%) and a majority had a long-term physical health condition (64.3%, 95% CI 59.3–69.0%). The 10 most prevalent long-term conditions were: depression (48.5%, 95% CI 43.4–53.7%), substance misuse (40.8%, 95% CI 35.9–45.8%), anxiety (34.9%, 95% CI 30.1–40.0%), asthma (26.9%, 95% CI 22.5–31.7%), hypertension (13.3%, 95 % CI 10.3–17.2%), severe mental illness (schizophrenia, bipolar disorder, non-organic psychosis) (13.1%, 95% CI 10.0–16.7%), hepatitis C (11.7%, 95% CI 8.9–15.4%), learning disabilities (10.7%, 95% CI 7.9–14.2%), personality disorder (8.8%, 95% CI 6.3–12.1%) and deliberate self-harm (8.5%, 95% 6.1–11.8%). Of these 10 conditions, seven were long-term mental health conditions.

Multiple social disadvantage (MSD)

A majority of patients had at least one MSD domain recorded (54.7%, 95% CI 49.4–59.8%). More Māori had at least one MSD domain recorded than New Zealand Europeans (61.0%, 95% CI 55.9–66.0 and 53.5%, 95% CI 48.1–58.5% respectively). The prevalence for the MSD domains was: in receipt of Work and Income New Zealand (WINZ) health and disability benefit in the previous year (28.5%, 95% CI 24.0–33.4%), previous imprisonment (16.3%, 95% CI 12.7–20.0%), domestic violence or other violence (4.8%, 95% CI 2.9–7.5%), homelessness (2.7%, 95% CI 1.3–4.9%) and lack of food (1.3%, 95% CI 0.4–3.0%).

Service utilisation

Most patients had seen either a GP or a practice nurse in the previous year (31 October 2016–1 November 2017) (80.8% and 78.4% respectively). The consultation rates were: GP (mean 4.5, median 4, IQR 1–7, range 0–45), practice nurse (mean 3.7, median 2, IQR 1–4, range 0–46), GP or practice nurse (mean 8.2, median 5, IQR 3–12, range 0–84). A third of patients (34.4%) had attended the local emergency department (ED) at least once in the previous year. Of those patients with a long-term mental health problem (n=281), a fifth (20.3%) had at least one recorded encounter with community mental health services (CMHS) in the previous year. Of those patients with substance misuse (alcohol, opioids or other psychoactive substance misuse) (n=153), 16.3% had at least one recorded encounter with community alcohol and drug services (CADS) in the previous year.

Discussion

This is the first New Zealand study to report the epidemiology of multimorbidity in a high-needs primary healthcare population. Most patients had multimorbidity (76.5%) and half (49.9%) had long-term physical and mental health comorbidities. These findings were consistent across all ethnic groups (New Zealand European; Māori; Pacific and Other). The prevalence of multimorbidity in patients in all ethnic groups increased with age and was high across all deprivation quintiles. Multimorbidity was the norm for patients aged over 25. Seven of the 10 most prevalent long-term conditions were mental health conditions (depression, substance misuse, anxiety, severe mental illness (schizophrenia, bipolar disorder, non-organic psychosis), learning disabilities, personality disorder and deliberate self-harm). A majority of patients had at least one MSD domain recorded (54.7%), the most common being in receipt of health and disability benefit in the previous year and previous imprisonment.

A key strength of this study is that we conducted a manual review of the clinic electronic medical records, which hold the most complete health data for patients, of all permanently registered patients. This is in contrast to an earlier study of the clinic’s patients, which relied on an opportunistic survey of a sample of clinic attenders.²¹ A limitation of using data from the electronic medical record is that its completeness relies on information being recorded. We consider we have captured all data relating to clinic healthcare utilisation and electronic filing of WINZ medical certificates (health and disability benefits) in the previous year. The clinic has high levels of recording of active and past medical problems through READ codes (the standard clinical terminology system used in general practice in New Zealand). At the time of conducting the study, however, discussion with the clinic team revealed that there was limited use of READ social codes for MSD: there was no standardised method or process for recording MSD in operation. It is therefore likely that the recording of homelessness, violence and lack of food is incomplete and thus the true prevalence of these domains at the clinic is higher than that reported here. We were also not able to explore any association between extent of MSD and degree of multimorbidity due to these data recording limitations. A further limitation of the study is that it is set in a single free clinic study in one New Zealand city, nonetheless it is considered that the findings are likely to be generalisable to other New Zealand free clinics serving similar high-needs populations.²² A final limitation of the study is its cross-sectional design. This does not allow us to explore the relationship between social inequalities, MSD and engagement with health and social services. Recent UK research has explored the role of social deprivation and individual lifestyle risk factors (eg, smoking, alcohol consumption, body mass index) in developing multimorbidity across the lifecourse.²⁷ While lifestyle risk factors partially mediate the relationship between deprivation and multimorbidity, the majority of the relationship between deprivation and multimorbidity remains unexplained.²⁷

There is limited New Zealand6 and international literature15 on the prevalence of multimorbidity in high-needs or vulnerable populations attending primary healthcare clinics. We found, however, that the prevalence of multimorbidity in Māori and Pacific patients attending SHC was higher than that in those enrolled with a large traditional general practice in the same city (53% of Māori and 64% of Pacific patients aged over 35 were multimorbid)⁶ and there were also higher rates of physical and mental health comorbidity in the SHC population. One explanation for this finding is that the patients at SHC had higher socioeconomic deprivation and deprivation is associated with increased multimorbidity.⁸ Our findings of a high multimorbidity prevalence, high physical and mental health comorbidity and the most prevalent conditions being mental health conditions are, however, consistent with a UK study of multimorbidity in a specialist homeless health service.¹⁵ We also found that while the mean age of patients was 41.9 years, the level of multimorbidity was comparable to those aged 65 and over in a general population of primary healthcare attenders.⁸ In contrast to the UK homeless study,¹⁵ however, we found a lower prevalence of MSD. This is likely to be explained by the fact that SHC offers free care to a broader range of patients than a specialist homeless service and is also due to the incomplete recording of MSD data noted above. In line with previous studies the health practitioner consultation rates were higher than those in ‘standard’ New Zealand primary healthcare^19,25 and consistent with the UK homeless study.¹⁵ It is also notable, when the results are compared with a key Scottish primary care multimorbidity cross-sectional study,⁸ that we did not find increasing levels of multimorbidity with increasing levels of area deprivation. In contrast, we found that the prevalence of multimorbidity in patients was high across all deprivation quintiles. This is likely to be explained by the high-needs nature of the SHC population, and we would add that there is considerable uncertainty around the multimorbidity estimates in the lower deprivation quintiles due to small numbers.

This study has a number of implications for healthcare policy and practice. The high prevalence of mental health conditions, physical and mental health comorbidity and use of specialist mental health and addiction services indicates the need for better integration of mental health and social services²⁸ with primary healthcare, particularly in those serving a high-needs population. It is well recognised that geographical, financial, organisational and cultural barriers limit such integration internationally.²⁶ There is scope for increased use of READ social codes for MSD in primary healthcare to better establish the prevalence of MSD in this population. A further issue is the need to ensure that primary healthcare funders and providers explicitly consider health equity when planning and delivering services for a high-needs population (Māori, Pacific and New Zealand deprivation quintile 5 patients), for example, by using the New Zealand Health Equity Assessment Tool (HEAT).²⁹ The study also has implications for improving access to primary healthcare for high-needs populations. One widely cited conceptualisation of access to healthcare sees access as the product of the interaction between accessibility of services (approachability, acceptability, availability and accommodation, affordability, appropriateness) and the individual’s abilities (ability to perceive, ability to seek, ability to reach, ability to pay and ability to engage).³⁰ This conceptualisation emphasises the importance of considering multilevel interventions to improve access that take into account the range of determinants of access. Clearly, improving access involves much more than simply removing a financial barrier to care. We acknowledge, however, that in New Zealand cost is a major barrier to equitable access to primary healthcare for high-needs and vulnerable populations¹⁸ and that ‘third sector’ organisations remain an important provider of care for this population.^19,22 The current New Zealand business model of general practice, with its use of direct patient charges, is a legacy of the compromises struck between the medical profession and the government during the implementation of the 1938 Social Security Act.³¹ Now, 80 years on, the time is surely right for a review of direct patient charges and their impact on access to primary healthcare for vulnerable groups as part of the ongoing review of New Zealand primary healthcare.³²

Conclusions

The high prevalence of multimorbidity and physical and mental health comorbidities in a free primary healthcare population serving a high-needs population experiencing MSD raises important issues relating to health equity in the New Zealand health system.³³ There is a need to re-orient New Zealand primary healthcare delivery around multimorbidity.^4,34,35 There is also a need to further the integration of health and social services.^28,36 In addressing these two key issues the ‘third sector’ is likely to continue to be an important provider of healthcare to those high-needs groups who cannot access mainstream primary healthcare. Further New Zealand research is also required to explore the relationship between MSD, multimorbidity and social inequalities.

Supplementary Table 1: List, data source definitions and prevalence of the 38 long-term conditions included in the multimorbidity analysis.