Describing primary care in rural Western Province, Papua New Guinea

Study design

A retrospective audit was conducted of the records of the presentations to the outreach service doctor in Aiambak, Lake Murray, Samokopa, Suki, Debepare, Dodomona, Habi, Arufi, Weam, Nomad, Wawoi Falls, and Hesalibi villages over 8 months between 13 February 2021 and 30 October 2021 in Middle and South Fly (see Fig. 1 for a map). This outreach service flew into these villages every 6 weeks. It provided primary health services at the health facility in each village, with the primary care team living in the village and providing services for 1–2 weeks at a time. Ambulatory clinics were held daily between 8 am and 6 pm, with an after-hours on-call service for emergencies. Ambulatory outpatients were defined as any patient who presented themselves to the health facility to seek care and was not an admitted inpatient. Consultations were held in English and Tok Pisin, with translation provided by the local healthcare worker when a third language was used.

Fig. 1.

Map showing the villages receiving health services from the outreach program, with the 12 villages in this study highlighted in yellow. Blue boxes indicate towns and circles indicate villages.

The majority of these facilities had a rural aid post with a Community Health Worker, which the outreach team supplemented. Communities were invited to participate in the program if they were not receiving support services from other sources, such as the government, church, or non-government organisations. This is part of a permanent program funded by the SDP, which has been ongoing since 2019.

The visiting team comprised a doctor or health extension officer with experience in rural and remote care and public health, a midwife, two nurses trained in family planning, and two water and sanitation hygiene officers. The teams are managed by a patrol manager who is a public health nurse and who maintains correspondence with their assigned villages throughout the year. The medical superintendent was a supervising Fellow of the Royal Australian College of General Practitioners, and further consultation occurred with the Senior Medical Officers of Daru, Kiunga and Rumginae Hospitals, depending on location. Regarding diagnostics, the team had access to basic rapid diagnostic tests for syphilis and malaria, haemoglobin colour charts and blot paper, pulse oximeters, sphygmomanometers, stethoscopes, otoscopes, thermometers, and urine dipsticks.

The presenting complaints were coded into their appropriate International Category of Disease 11 (ICD-11) labels and subsequently grouped by category. The ICD-11 is an internationally recognised medical coding system that allows conditions to be standardised and compared across facilities and is recommended by the WHO. The study period was during the COVID-19 pandemic, with nationally recognised widespread community transmission. COVID-19 did not affect the outreach schedule in this study.

Sample size, data collection and data source

Over the 8 months, 442 new patient presentations were analysed in 12 villages in Western Province. Records were accessed from SDP Health via a secure organisational spreadsheet. The SDP Health staff entered the data immediately after returning from patrol in each location. The same clinician completed the ICD-11 coding to ensure consistency.

Variable and statistical analysis

Descriptive analysis of data was performed using Microsoft Excel (2024) and RStudio (2024.12.1). Frequencies were used to present descriptive statistics. The five most common ICD-11 categories underwent further analysis using Pearson’s Chi-squared test of independence to estimate association between age group, sex, and district. Variables assessed included sex (male or female), age (<20, 20–40, >40 years), and district (Middle Fly or South Fly). Frequency cross-tabulations were produced to display the ICD-11 categories and subcategories.

Ethics

This is a sub-analysis of a larger project exploring the impacts of COVID-19 on primary care and infectious disease programs in PNG. Ethics approvals for this study were received in PNG by the Papua New Guinea Institute of Medical Research Institutional Review Board (IRB #2015), the Medical Research Advisory Committee (MRAC 20.35), and the University of New South Wales Human Research Ethics Committee (HC210172). Permission for audit was provided by the Western Provincial Health Authority and SDP – Health.

Key findings and interpretation

The key findings in this study were the five main categories of disease with which community members presented. These were ‘signs and symptoms not otherwise specified’, ‘diseases of the respiratory tract’, ‘diseases of the digestive tract’, ‘musculoskeletal and soft tissue diseases’, and ‘infectious and other parasitic conditions’. Further analysis was conducted using variables such as gender, age, and district, which demonstrated statistically significant differences.

Infectious and parasitic diseases were more common in the 0–20-year age group. This is expected as children generally have greater exposure to contagious sources because of proximity to other children, play practices, and poor hygiene. The significance of the greater number of digestive symptoms in the >40-year age group may reflect social behaviours in adulthood, such as use of alcohol and betel nut, which can cause reflux and abdominal discomfort. This is supported by this age group reporting the greatest rate of GORD symptoms. Women represented the majority of presentations, which reflects a gendered predisposition to be more likely to engage in care, which is well reported in the literature.⁹ There was no significant difference between Middle and South Fly presentations. This was surprising as South Fly has greater access to goods and services from both Indonesia and Australia and therefore was expected to have different disease profiles. This is likely explained by selection bias, with villages in the program being chosen because of rurality and lack of services, and therefore having similar characteristics, which was reflected in the results.

The majority of ambulatory outpatient presentations fell into the category ‘signs and symptoms not otherwise specified’. This reflects the significant diagnostic challenges faced by clinicians in rural settings. This audit relied on the diagnostic acumen of the doctor employed by the SDP Health program, with limited access to radiology or laboratory services. Most conditions cannot be definitively diagnosed with clinical examination and history alone, and this results in clinicians treating symptoms rather than the cause of the presenting complaint. Although this is a limitation, it reflects the reality of providing primary care in rural healthcare settings.

This audit period was conducted at a time when there was widespread transmission of COVID-19, which may have resulted in a higher proportion of respiratory conditions to be recorded or a lower overall number of presentations due to the avoidance of healthcare providers.¹⁰ The scarcity of COVID-19 tests in remote PNG limited the capacity of the outreach team to diagnose COVID-19, with only a few rapid diagnostic tests becoming available at the end of the reporting period.¹¹ Some of the cases of URTI, pneumonia, or infective exacerbation of chronic obstructive pulmonary disease/asthma may have been COVID-19 infections.

The most common specific diseases that community members presented with were pain in the limb (usually knee pain), tropical skin ulcers, gastro-oesophageal reflux symptoms, lower back pain, and URTIs. Lower back pain was absent in youths less than 20 years of age, slightly more frequent in the 20–40-year age group, and most frequent in those older than 40 years. It is unsurprising that older patients presented with increasing lower back and knee pain, likely due to arthritic conditions after a lifetime of manual labour in the village. Women were more likely to report lower back pain, which likely reflects their manual labour of gardening, which involves extended periods of bending down and carrying heavy bilums (homemade bags) braced against the head and neck.

When compared with the WHO report of ambulatory presentations in PNG, this study’s results differ slightly. There was less malaria (1.8%) when compared with 29% in the WHO report. Skin disease was clinically significant in Western Province, and its frequency was similar to that reported by WHO for the nation of PNG (11.3% vs 10.8%). WHO reports simple cough, pneumonia, and other respiratory conditions collectively accounted for 21.5% of ambulatory presentations, while they only contributed to 11.9% in this study. This may suggest Western Province has fewer respiratory conditions than the national average. Of significance, chronic musculoskeletal pain, including limb and back pain, was absent in the WHO report but was present in 14.7% of the ambulatory patients in this study.

When comparing the outpatient surveillance tally sheet used by the NDoH with the results of this study, 63.8% of presentations in this study would have been categorised as ‘other new cases.’ Although it is possible healthcare workers may classify musculoskeletal pain as ‘other accident and injury’ (seeking to attribute a traumatic cause to the pain), even taking into consideration the possible alternative classifications, current data-collection tools lose a significant amount of information on the primary care burden of disease.

Limitations

The small sample size of only 442 patients limits the power of the conclusions that can be drawn from this paper’s results. Further, the broad range of presentations meant that after categorisation, few patients were in each category, limiting the power of comparison.

The ideal measure for this study would have been to determine the prevalence of disease in the communities. This would have required surveying both the patients who presented to the clinic and those who did not to determine the presence or absence of disease in the community. Unfortunately, as this is a retrospective audit, this information is not available.

Selection bias was likely present as community members self-selected to attend the outpatient clinics. Health-conscious patients may have been more likely to present. Furthermore, those able to physically travel to the clinic were more likely to be seen than those with disabilities who are less mobile or those restricted by COVID-19.¹⁰

The potential subjectivity of the healthcare worker in applying the ICD-11 category to the presenting complaints is also possible. Although ICD-11 offers specific criteria for each categorisation, the clinician must choose which classification best fits the presentation for coding. This was more difficult in rural PNG because of the lack of diagnostic aids such as radiology and laboratory support, as described earlier. However, the diagnosis of these conditions on clinical assessment alone reflects accepted practice within PNG due to resource limitations.

Generalisability

These results would likely be generalisable for rural inland village communities in PNG. The challenges of rural and remote living, including reliance on manual labour and limited access to services and health and hygiene products, are shared across rural and remote PNG communities. It is unlikely they would represent the disease burden of urban populations, who have greater access to healthcare and labour-saving devices such as roads, cars, and electrification.

Furthermore, the isolation of remote communities in Western Province means it is more likely that patients would present with chronic conditions, such as arthritic knee pain, than urban patients, who could likely access simple analgesia from a shop. Urban populations are also likely to have greater access to extended health services, such as radiology and laboratory diagnostics, which could further refine the ICD-11 subcategories, potentially giving a more specific diagnosis.

Future recommendations

PNG healthcare workers already face a significant reporting burden in their clinical practice, and it is not completely clear how much of the information already collected is used effectively.¹² For this reason, updating the current NDoH outpatient tally sheet to collect further information to code presentations by the ICD-11 may not be helpful. One potential recommendation is to encourage doctors or Health Extension Officers conducting rural outreach patrol supervisory visits to conduct sample studies of the prevalence of disease in the rural population to provide detailed primary care data to the NDoH.