We were interested to read the McGeoch et al. paper on unmet secondary care need in Canterbury.¹ We have previously published on unmet need in Canterbury and Auckland,² and found a strikingly higher level of unmet secondary care need than was reported by McGeoch et al. (9.3 vs 3.6%). One likely explanation for this difference is that, while the study by McGeoch et al. relied on GP attendance to identify study respondents, our study used online, face-to-face and telephone random population surveys. Of our respondents, 28.7% reported an unmet primary care need and 16.5% reported not attending GPs because of cost. These data are important, as respondents reporting unmet primary need were significantly more likely to report unmet secondary need than those not doing so (17.4 vs 6.1%, P < 0.001). It is therefore very likely that the study by McGeoch et al. will have underestimated unmet secondary care need.

We were intrigued by the Editorial Comment that contrasted the method used in the study by McGeoch et al. with ‘more comfortable methods’.³ While we are not sure what constitutes methodological comfort, we suggest that what is important in study design is whether the chosen method will provide an accurate answer to the study question. If the question is ‘what is the level of unmet need in a population?’, then a study design that excludes a significant proportion of those with the greatest unmet need is clearly problematic. Our pilot study² findings led us to conclude that a larger scale survey of unmet need using random population sampling by address is required; the study findings of McGeoch et al. do not alter this conclusion.

Both P. Bagshaw and B. Hudson were authors of another study (Ref# 2) of unmet health-care need.

Dear Editor,

We share Bagshaw and Hudson's concerns that access to affordable general practice may hide unmet health need. Unmet health need can occur for several reasons around accessability, availability, or acceptability of services. Our survey described in the journal⁴ used a novel method to measure unmet need for referred services in patients attending their general practice. Our survey was not intended to measure total population need and did not include patients who do not attend a general practice. In contrast, the paper of Bagshaw et al.⁵ was based mainly on the findings of a postal survey in a sample of the total population that measured unmet need for secondary care. Direct comparison of the rate of unmet health need measured by our survey and that of Bagshaw et al. is therefore not valid, nor is the conclusion that our method underestimated what we were seeking to measure. Our study was carried out mainly in response to concern that active referral management might be hiding ­unmet health need in the community and to inform service planning.

We also agree with Bagshaw and Hudson that it is important to choose a survey method that will provide accurate answers to study questions but do not believe a single, simple method can answer such a complex question as measuring unmet need. A strength of our survey method was that the joint discussion of the health need between patient and general practitioner minimised any tendency for patients to not recognise their needs or over estimate their needs and to include things that either could not be fixed or did not need intervention. Another strength of our survey method was the use of electronic referral to collect survey data, a system used routinely by nearly all general practices, backed up by support from a liason person who worked with the practices to ensure successful completion of the survey. This approach led to active participation by general practices in Canterbury and succesful data collection of over 2000 patients in a relatively short period of time. In contrast, the general practitioner arm of the Bagshaw et al. survey had a low participation rate of general practices and limited patient data, leading the authors to conclude that such a survey method was not worthwhile. We believe the practicality of the methodology is important and consider the method we used in our survey, with emphasis on conversation and agreement between a patient and their general practitioner, provides an efficient and balanced method for measuring unmet need for referred services in a primary care setting. It does not address the issue of unmet health need in those not attending general practice.

Our experience leads us to believe that general practice involvement can enhance the value and integrity of data about unmet health need but should be combined with other methods to collect data on patients who rarely attend general practice. Perhaps we need to work together with Bagshaw et al. to combine a population survey approach with some general practitioner and independent medical verification of a proportion of the study population. New Zealand needs to develop a method that observes trends in access to public health care over many years.

Dr Graham McGeoch
Clinical Leader, Canterbury Initiative,
Canterbury DHB
on behalf of survey team members